linux/drivers/mmc/host/atmel-mci.c

2593 lines
67 KiB
C
Raw Normal View History

atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/*
* Atmel MultiMedia Card Interface driver
*
* Copyright (C) 2004-2008 Atmel Corporation
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/blkdev.h>
#include <linux/clk.h>
#include <linux/debugfs.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include <linux/device.h>
#include <linux/dmaengine.h>
#include <linux/dma-mapping.h>
#include <linux/err.h>
#include <linux/gpio.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/ioport.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/of_device.h>
#include <linux/of_gpio.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include <linux/platform_device.h>
#include <linux/scatterlist.h>
#include <linux/seq_file.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/slab.h>
#include <linux/stat.h>
#include <linux/types.h>
#include <linux/platform_data/atmel.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include <linux/mmc/host.h>
#include <linux/mmc/sdio.h>
#include <mach/atmel-mci.h>
#include <linux/atmel-mci.h>
#include <linux/atmel_pdc.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include <asm/io.h>
#include <asm/unaligned.h>
#include <mach/cpu.h>
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#include "atmel-mci-regs.h"
#define ATMCI_DATA_ERROR_FLAGS (ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
#define ATMCI_DMA_THRESHOLD 16
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
enum {
EVENT_CMD_RDY = 0,
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
EVENT_XFER_COMPLETE,
EVENT_NOTBUSY,
EVENT_DATA_ERROR,
};
enum atmel_mci_state {
STATE_IDLE = 0,
STATE_SENDING_CMD,
STATE_DATA_XFER,
STATE_WAITING_NOTBUSY,
STATE_SENDING_STOP,
STATE_END_REQUEST,
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
};
enum atmci_xfer_dir {
XFER_RECEIVE = 0,
XFER_TRANSMIT,
};
enum atmci_pdc_buf {
PDC_FIRST_BUF = 0,
PDC_SECOND_BUF,
};
struct atmel_mci_caps {
bool has_dma_conf_reg;
bool has_pdc;
bool has_cfg_reg;
bool has_cstor_reg;
bool has_highspeed;
bool has_rwproof;
bool has_odd_clk_div;
bool has_bad_data_ordering;
bool need_reset_after_xfer;
bool need_blksz_mul_4;
bool need_notbusy_for_read_ops;
};
struct atmel_mci_dma {
struct dma_chan *chan;
struct dma_async_tx_descriptor *data_desc;
};
/**
* struct atmel_mci - MMC controller state shared between all slots
* @lock: Spinlock protecting the queue and associated data.
* @regs: Pointer to MMIO registers.
* @sg: Scatterlist entry currently being processed by PIO or PDC code.
* @pio_offset: Offset into the current scatterlist entry.
* @buffer: Buffer used if we don't have the r/w proof capability. We
* don't have the time to switch pdc buffers so we have to use only
* one buffer for the full transaction.
* @buf_size: size of the buffer.
* @phys_buf_addr: buffer address needed for pdc.
* @cur_slot: The slot which is currently using the controller.
* @mrq: The request currently being processed on @cur_slot,
* or NULL if the controller is idle.
* @cmd: The command currently being sent to the card, or NULL.
* @data: The data currently being transferred, or NULL if no data
* transfer is in progress.
* @data_size: just data->blocks * data->blksz.
* @dma: DMA client state.
* @data_chan: DMA channel being used for the current data transfer.
* @cmd_status: Snapshot of SR taken upon completion of the current
* command. Only valid when EVENT_CMD_COMPLETE is pending.
* @data_status: Snapshot of SR taken upon completion of the current
* data transfer. Only valid when EVENT_DATA_COMPLETE or
* EVENT_DATA_ERROR is pending.
* @stop_cmdr: Value to be loaded into CMDR when the stop command is
* to be sent.
* @tasklet: Tasklet running the request state machine.
* @pending_events: Bitmask of events flagged by the interrupt handler
* to be processed by the tasklet.
* @completed_events: Bitmask of events which the state machine has
* processed.
* @state: Tasklet state.
* @queue: List of slots waiting for access to the controller.
* @need_clock_update: Update the clock rate before the next request.
* @need_reset: Reset controller before next request.
* @timer: Timer to balance the data timeout error flag which cannot rise.
* @mode_reg: Value of the MR register.
* @cfg_reg: Value of the CFG register.
* @bus_hz: The rate of @mck in Hz. This forms the basis for MMC bus
* rate and timeout calculations.
* @mapbase: Physical address of the MMIO registers.
* @mck: The peripheral bus clock hooked up to the MMC controller.
* @pdev: Platform device associated with the MMC controller.
* @slot: Slots sharing this MMC controller.
* @caps: MCI capabilities depending on MCI version.
* @prepare_data: function to setup MCI before data transfer which
* depends on MCI capabilities.
* @submit_data: function to start data transfer which depends on MCI
* capabilities.
* @stop_transfer: function to stop data transfer which depends on MCI
* capabilities.
*
* Locking
* =======
*
* @lock is a softirq-safe spinlock protecting @queue as well as
* @cur_slot, @mrq and @state. These must always be updated
* at the same time while holding @lock.
*
* @lock also protects mode_reg and need_clock_update since these are
* used to synchronize mode register updates with the queue
* processing.
*
* The @mrq field of struct atmel_mci_slot is also protected by @lock,
* and must always be written at the same time as the slot is added to
* @queue.
*
* @pending_events and @completed_events are accessed using atomic bit
* operations, so they don't need any locking.
*
* None of the fields touched by the interrupt handler need any
* locking. However, ordering is important: Before EVENT_DATA_ERROR or
* EVENT_DATA_COMPLETE is set in @pending_events, all data-related
* interrupts must be disabled and @data_status updated with a
* snapshot of SR. Similarly, before EVENT_CMD_COMPLETE is set, the
* CMDRDY interrupt must be disabled and @cmd_status updated with a
* snapshot of SR, and before EVENT_XFER_COMPLETE can be set, the
* bytes_xfered field of @data must be written. This is ensured by
* using barriers.
*/
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct atmel_mci {
spinlock_t lock;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
void __iomem *regs;
struct scatterlist *sg;
unsigned int sg_len;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
unsigned int pio_offset;
unsigned int *buffer;
unsigned int buf_size;
dma_addr_t buf_phys_addr;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct atmel_mci_slot *cur_slot;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct mmc_request *mrq;
struct mmc_command *cmd;
struct mmc_data *data;
unsigned int data_size;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct atmel_mci_dma dma;
struct dma_chan *data_chan;
struct dma_slave_config dma_conf;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
u32 cmd_status;
u32 data_status;
u32 stop_cmdr;
struct tasklet_struct tasklet;
unsigned long pending_events;
unsigned long completed_events;
enum atmel_mci_state state;
struct list_head queue;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
bool need_clock_update;
bool need_reset;
struct timer_list timer;
u32 mode_reg;
u32 cfg_reg;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
unsigned long bus_hz;
unsigned long mapbase;
struct clk *mck;
struct platform_device *pdev;
struct atmel_mci_slot *slot[ATMCI_MAX_NR_SLOTS];
struct atmel_mci_caps caps;
u32 (*prepare_data)(struct atmel_mci *host, struct mmc_data *data);
void (*submit_data)(struct atmel_mci *host, struct mmc_data *data);
void (*stop_transfer)(struct atmel_mci *host);
};
/**
* struct atmel_mci_slot - MMC slot state
* @mmc: The mmc_host representing this slot.
* @host: The MMC controller this slot is using.
* @sdc_reg: Value of SDCR to be written before using this slot.
* @sdio_irq: SDIO irq mask for this slot.
* @mrq: mmc_request currently being processed or waiting to be
* processed, or NULL when the slot is idle.
* @queue_node: List node for placing this node in the @queue list of
* &struct atmel_mci.
* @clock: Clock rate configured by set_ios(). Protected by host->lock.
* @flags: Random state bits associated with the slot.
* @detect_pin: GPIO pin used for card detection, or negative if not
* available.
* @wp_pin: GPIO pin used for card write protect sending, or negative
* if not available.
* @detect_is_active_high: The state of the detect pin when it is active.
* @detect_timer: Timer used for debouncing @detect_pin interrupts.
*/
struct atmel_mci_slot {
struct mmc_host *mmc;
struct atmel_mci *host;
u32 sdc_reg;
u32 sdio_irq;
struct mmc_request *mrq;
struct list_head queue_node;
unsigned int clock;
unsigned long flags;
#define ATMCI_CARD_PRESENT 0
#define ATMCI_CARD_NEED_INIT 1
#define ATMCI_SHUTDOWN 2
#define ATMCI_SUSPENDED 3
int detect_pin;
int wp_pin;
bool detect_is_active_high;
struct timer_list detect_timer;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
};
#define atmci_test_and_clear_pending(host, event) \
test_and_clear_bit(event, &host->pending_events)
#define atmci_set_completed(host, event) \
set_bit(event, &host->completed_events)
#define atmci_set_pending(host, event) \
set_bit(event, &host->pending_events)
/*
* The debugfs stuff below is mostly optimized away when
* CONFIG_DEBUG_FS is not set.
*/
static int atmci_req_show(struct seq_file *s, void *v)
{
struct atmel_mci_slot *slot = s->private;
struct mmc_request *mrq;
struct mmc_command *cmd;
struct mmc_command *stop;
struct mmc_data *data;
/* Make sure we get a consistent snapshot */
spin_lock_bh(&slot->host->lock);
mrq = slot->mrq;
if (mrq) {
cmd = mrq->cmd;
data = mrq->data;
stop = mrq->stop;
if (cmd)
seq_printf(s,
"CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
cmd->opcode, cmd->arg, cmd->flags,
cmd->resp[0], cmd->resp[1], cmd->resp[2],
cmd->resp[3], cmd->error);
if (data)
seq_printf(s, "DATA %u / %u * %u flg %x err %d\n",
data->bytes_xfered, data->blocks,
data->blksz, data->flags, data->error);
if (stop)
seq_printf(s,
"CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
stop->opcode, stop->arg, stop->flags,
stop->resp[0], stop->resp[1], stop->resp[2],
stop->resp[3], stop->error);
}
spin_unlock_bh(&slot->host->lock);
return 0;
}
static int atmci_req_open(struct inode *inode, struct file *file)
{
return single_open(file, atmci_req_show, inode->i_private);
}
static const struct file_operations atmci_req_fops = {
.owner = THIS_MODULE,
.open = atmci_req_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
static void atmci_show_status_reg(struct seq_file *s,
const char *regname, u32 value)
{
static const char *sr_bit[] = {
[0] = "CMDRDY",
[1] = "RXRDY",
[2] = "TXRDY",
[3] = "BLKE",
[4] = "DTIP",
[5] = "NOTBUSY",
[6] = "ENDRX",
[7] = "ENDTX",
[8] = "SDIOIRQA",
[9] = "SDIOIRQB",
[12] = "SDIOWAIT",
[14] = "RXBUFF",
[15] = "TXBUFE",
[16] = "RINDE",
[17] = "RDIRE",
[18] = "RCRCE",
[19] = "RENDE",
[20] = "RTOE",
[21] = "DCRCE",
[22] = "DTOE",
[23] = "CSTOE",
[24] = "BLKOVRE",
[25] = "DMADONE",
[26] = "FIFOEMPTY",
[27] = "XFRDONE",
[30] = "OVRE",
[31] = "UNRE",
};
unsigned int i;
seq_printf(s, "%s:\t0x%08x", regname, value);
for (i = 0; i < ARRAY_SIZE(sr_bit); i++) {
if (value & (1 << i)) {
if (sr_bit[i])
seq_printf(s, " %s", sr_bit[i]);
else
seq_puts(s, " UNKNOWN");
}
}
seq_putc(s, '\n');
}
static int atmci_regs_show(struct seq_file *s, void *v)
{
struct atmel_mci *host = s->private;
u32 *buf;
buf = kmalloc(ATMCI_REGS_SIZE, GFP_KERNEL);
if (!buf)
return -ENOMEM;
/*
* Grab a more or less consistent snapshot. Note that we're
* not disabling interrupts, so IMR and SR may not be
* consistent.
*/
spin_lock_bh(&host->lock);
clk_enable(host->mck);
memcpy_fromio(buf, host->regs, ATMCI_REGS_SIZE);
clk_disable(host->mck);
spin_unlock_bh(&host->lock);
seq_printf(s, "MR:\t0x%08x%s%s ",
buf[ATMCI_MR / 4],
buf[ATMCI_MR / 4] & ATMCI_MR_RDPROOF ? " RDPROOF" : "",
buf[ATMCI_MR / 4] & ATMCI_MR_WRPROOF ? " WRPROOF" : "");
if (host->caps.has_odd_clk_div)
seq_printf(s, "{CLKDIV,CLKODD}=%u\n",
((buf[ATMCI_MR / 4] & 0xff) << 1)
| ((buf[ATMCI_MR / 4] >> 16) & 1));
else
seq_printf(s, "CLKDIV=%u\n",
(buf[ATMCI_MR / 4] & 0xff));
seq_printf(s, "DTOR:\t0x%08x\n", buf[ATMCI_DTOR / 4]);
seq_printf(s, "SDCR:\t0x%08x\n", buf[ATMCI_SDCR / 4]);
seq_printf(s, "ARGR:\t0x%08x\n", buf[ATMCI_ARGR / 4]);
seq_printf(s, "BLKR:\t0x%08x BCNT=%u BLKLEN=%u\n",
buf[ATMCI_BLKR / 4],
buf[ATMCI_BLKR / 4] & 0xffff,
(buf[ATMCI_BLKR / 4] >> 16) & 0xffff);
if (host->caps.has_cstor_reg)
seq_printf(s, "CSTOR:\t0x%08x\n", buf[ATMCI_CSTOR / 4]);
/* Don't read RSPR and RDR; it will consume the data there */
atmci_show_status_reg(s, "SR", buf[ATMCI_SR / 4]);
atmci_show_status_reg(s, "IMR", buf[ATMCI_IMR / 4]);
if (host->caps.has_dma_conf_reg) {
u32 val;
val = buf[ATMCI_DMA / 4];
seq_printf(s, "DMA:\t0x%08x OFFSET=%u CHKSIZE=%u%s\n",
val, val & 3,
((val >> 4) & 3) ?
1 << (((val >> 4) & 3) + 1) : 1,
val & ATMCI_DMAEN ? " DMAEN" : "");
}
if (host->caps.has_cfg_reg) {
u32 val;
val = buf[ATMCI_CFG / 4];
seq_printf(s, "CFG:\t0x%08x%s%s%s%s\n",
val,
val & ATMCI_CFG_FIFOMODE_1DATA ? " FIFOMODE_ONE_DATA" : "",
val & ATMCI_CFG_FERRCTRL_COR ? " FERRCTRL_CLEAR_ON_READ" : "",
val & ATMCI_CFG_HSMODE ? " HSMODE" : "",
val & ATMCI_CFG_LSYNC ? " LSYNC" : "");
}
kfree(buf);
return 0;
}
static int atmci_regs_open(struct inode *inode, struct file *file)
{
return single_open(file, atmci_regs_show, inode->i_private);
}
static const struct file_operations atmci_regs_fops = {
.owner = THIS_MODULE,
.open = atmci_regs_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
static void atmci_init_debugfs(struct atmel_mci_slot *slot)
{
struct mmc_host *mmc = slot->mmc;
struct atmel_mci *host = slot->host;
struct dentry *root;
struct dentry *node;
root = mmc->debugfs_root;
if (!root)
return;
node = debugfs_create_file("regs", S_IRUSR, root, host,
&atmci_regs_fops);
if (IS_ERR(node))
return;
if (!node)
goto err;
node = debugfs_create_file("req", S_IRUSR, root, slot, &atmci_req_fops);
if (!node)
goto err;
node = debugfs_create_u32("state", S_IRUSR, root, (u32 *)&host->state);
if (!node)
goto err;
node = debugfs_create_x32("pending_events", S_IRUSR, root,
(u32 *)&host->pending_events);
if (!node)
goto err;
node = debugfs_create_x32("completed_events", S_IRUSR, root,
(u32 *)&host->completed_events);
if (!node)
goto err;
return;
err:
dev_err(&mmc->class_dev, "failed to initialize debugfs for slot\n");
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
#if defined(CONFIG_OF)
static const struct of_device_id atmci_dt_ids[] = {
{ .compatible = "atmel,hsmci" },
{ /* sentinel */ }
};
MODULE_DEVICE_TABLE(of, atmci_dt_ids);
static struct mci_platform_data*
atmci_of_init(struct platform_device *pdev)
{
struct device_node *np = pdev->dev.of_node;
struct device_node *cnp;
struct mci_platform_data *pdata;
u32 slot_id;
if (!np) {
dev_err(&pdev->dev, "device node not found\n");
return ERR_PTR(-EINVAL);
}
pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
if (!pdata) {
dev_err(&pdev->dev, "could not allocate memory for pdata\n");
return ERR_PTR(-ENOMEM);
}
for_each_child_of_node(np, cnp) {
if (of_property_read_u32(cnp, "reg", &slot_id)) {
dev_warn(&pdev->dev, "reg property is missing for %s\n",
cnp->full_name);
continue;
}
if (slot_id >= ATMCI_MAX_NR_SLOTS) {
dev_warn(&pdev->dev, "can't have more than %d slots\n",
ATMCI_MAX_NR_SLOTS);
break;
}
if (of_property_read_u32(cnp, "bus-width",
&pdata->slot[slot_id].bus_width))
pdata->slot[slot_id].bus_width = 1;
pdata->slot[slot_id].detect_pin =
of_get_named_gpio(cnp, "cd-gpios", 0);
pdata->slot[slot_id].detect_is_active_high =
of_property_read_bool(cnp, "cd-inverted");
pdata->slot[slot_id].wp_pin =
of_get_named_gpio(cnp, "wp-gpios", 0);
}
return pdata;
}
#else /* CONFIG_OF */
static inline struct mci_platform_data*
atmci_of_init(struct platform_device *dev)
{
return ERR_PTR(-EINVAL);
}
#endif
static inline unsigned int atmci_get_version(struct atmel_mci *host)
{
return atmci_readl(host, ATMCI_VERSION) & 0x00000fff;
}
static void atmci_timeout_timer(unsigned long data)
{
struct atmel_mci *host;
host = (struct atmel_mci *)data;
dev_dbg(&host->pdev->dev, "software timeout\n");
if (host->mrq->cmd->data) {
host->mrq->cmd->data->error = -ETIMEDOUT;
host->data = NULL;
} else {
host->mrq->cmd->error = -ETIMEDOUT;
host->cmd = NULL;
}
host->need_reset = 1;
host->state = STATE_END_REQUEST;
smp_wmb();
tasklet_schedule(&host->tasklet);
}
static inline unsigned int atmci_ns_to_clocks(struct atmel_mci *host,
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
unsigned int ns)
{
/*
* It is easier here to use us instead of ns for the timeout,
* it prevents from overflows during calculation.
*/
unsigned int us = DIV_ROUND_UP(ns, 1000);
/* Maximum clock frequency is host->bus_hz/2 */
return us * (DIV_ROUND_UP(host->bus_hz, 2000000));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_set_timeout(struct atmel_mci *host,
struct atmel_mci_slot *slot, struct mmc_data *data)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
static unsigned dtomul_to_shift[] = {
0, 4, 7, 8, 10, 12, 16, 20
};
unsigned timeout;
unsigned dtocyc;
unsigned dtomul;
timeout = atmci_ns_to_clocks(host, data->timeout_ns)
+ data->timeout_clks;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
for (dtomul = 0; dtomul < 8; dtomul++) {
unsigned shift = dtomul_to_shift[dtomul];
dtocyc = (timeout + (1 << shift) - 1) >> shift;
if (dtocyc < 15)
break;
}
if (dtomul >= 8) {
dtomul = 7;
dtocyc = 15;
}
dev_vdbg(&slot->mmc->class_dev, "setting timeout to %u cycles\n",
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dtocyc << dtomul_to_shift[dtomul]);
atmci_writel(host, ATMCI_DTOR, (ATMCI_DTOMUL(dtomul) | ATMCI_DTOCYC(dtocyc)));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* Return mask with command flags to be enabled for this command.
*/
static u32 atmci_prepare_command(struct mmc_host *mmc,
struct mmc_command *cmd)
{
struct mmc_data *data;
u32 cmdr;
cmd->error = -EINPROGRESS;
cmdr = ATMCI_CMDR_CMDNB(cmd->opcode);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (cmd->flags & MMC_RSP_PRESENT) {
if (cmd->flags & MMC_RSP_136)
cmdr |= ATMCI_CMDR_RSPTYP_136BIT;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
else
cmdr |= ATMCI_CMDR_RSPTYP_48BIT;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* This should really be MAXLAT_5 for CMD2 and ACMD41, but
* it's too difficult to determine whether this is an ACMD or
* not. Better make it 64.
*/
cmdr |= ATMCI_CMDR_MAXLAT_64CYC;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (mmc->ios.bus_mode == MMC_BUSMODE_OPENDRAIN)
cmdr |= ATMCI_CMDR_OPDCMD;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
data = cmd->data;
if (data) {
cmdr |= ATMCI_CMDR_START_XFER;
if (cmd->opcode == SD_IO_RW_EXTENDED) {
cmdr |= ATMCI_CMDR_SDIO_BLOCK;
} else {
if (data->flags & MMC_DATA_STREAM)
cmdr |= ATMCI_CMDR_STREAM;
else if (data->blocks > 1)
cmdr |= ATMCI_CMDR_MULTI_BLOCK;
else
cmdr |= ATMCI_CMDR_BLOCK;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (data->flags & MMC_DATA_READ)
cmdr |= ATMCI_CMDR_TRDIR_READ;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
return cmdr;
}
static void atmci_send_command(struct atmel_mci *host,
struct mmc_command *cmd, u32 cmd_flags)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
WARN_ON(host->cmd);
host->cmd = cmd;
dev_vdbg(&host->pdev->dev,
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
"start command: ARGR=0x%08x CMDR=0x%08x\n",
cmd->arg, cmd_flags);
atmci_writel(host, ATMCI_ARGR, cmd->arg);
atmci_writel(host, ATMCI_CMDR, cmd_flags);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_send_stop_cmd(struct atmel_mci *host, struct mmc_data *data)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
dev_dbg(&host->pdev->dev, "send stop command\n");
atmci_send_command(host, data->stop, host->stop_cmdr);
atmci_writel(host, ATMCI_IER, ATMCI_CMDRDY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* Configure given PDC buffer taking care of alignement issues.
* Update host->data_size and host->sg.
*/
static void atmci_pdc_set_single_buf(struct atmel_mci *host,
enum atmci_xfer_dir dir, enum atmci_pdc_buf buf_nb)
{
u32 pointer_reg, counter_reg;
unsigned int buf_size;
if (dir == XFER_RECEIVE) {
pointer_reg = ATMEL_PDC_RPR;
counter_reg = ATMEL_PDC_RCR;
} else {
pointer_reg = ATMEL_PDC_TPR;
counter_reg = ATMEL_PDC_TCR;
}
if (buf_nb == PDC_SECOND_BUF) {
pointer_reg += ATMEL_PDC_SCND_BUF_OFF;
counter_reg += ATMEL_PDC_SCND_BUF_OFF;
}
if (!host->caps.has_rwproof) {
buf_size = host->buf_size;
atmci_writel(host, pointer_reg, host->buf_phys_addr);
} else {
buf_size = sg_dma_len(host->sg);
atmci_writel(host, pointer_reg, sg_dma_address(host->sg));
}
if (host->data_size <= buf_size) {
if (host->data_size & 0x3) {
/* If size is different from modulo 4, transfer bytes */
atmci_writel(host, counter_reg, host->data_size);
atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCFBYTE);
} else {
/* Else transfer 32-bits words */
atmci_writel(host, counter_reg, host->data_size / 4);
}
host->data_size = 0;
} else {
/* We assume the size of a page is 32-bits aligned */
atmci_writel(host, counter_reg, sg_dma_len(host->sg) / 4);
host->data_size -= sg_dma_len(host->sg);
if (host->data_size)
host->sg = sg_next(host->sg);
}
}
/*
* Configure PDC buffer according to the data size ie configuring one or two
* buffers. Don't use this function if you want to configure only the second
* buffer. In this case, use atmci_pdc_set_single_buf.
*/
static void atmci_pdc_set_both_buf(struct atmel_mci *host, int dir)
{
atmci_pdc_set_single_buf(host, dir, PDC_FIRST_BUF);
if (host->data_size)
atmci_pdc_set_single_buf(host, dir, PDC_SECOND_BUF);
}
/*
* Unmap sg lists, called when transfer is finished.
*/
static void atmci_pdc_cleanup(struct atmel_mci *host)
{
struct mmc_data *data = host->data;
if (data)
dma_unmap_sg(&host->pdev->dev,
data->sg, data->sg_len,
((data->flags & MMC_DATA_WRITE)
? DMA_TO_DEVICE : DMA_FROM_DEVICE));
}
/*
* Disable PDC transfers. Update pending flags to EVENT_XFER_COMPLETE after
* having received ATMCI_TXBUFE or ATMCI_RXBUFF interrupt. Enable ATMCI_NOTBUSY
* interrupt needed for both transfer directions.
*/
static void atmci_pdc_complete(struct atmel_mci *host)
{
int transfer_size = host->data->blocks * host->data->blksz;
int i;
atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTDIS | ATMEL_PDC_TXTDIS);
if ((!host->caps.has_rwproof)
&& (host->data->flags & MMC_DATA_READ)) {
if (host->caps.has_bad_data_ordering)
for (i = 0; i < transfer_size; i++)
host->buffer[i] = swab32(host->buffer[i]);
sg_copy_from_buffer(host->data->sg, host->data->sg_len,
host->buffer, transfer_size);
}
atmci_pdc_cleanup(host);
/*
* If the card was removed, data will be NULL. No point trying
* to send the stop command or waiting for NBUSY in this case.
*/
if (host->data) {
dev_dbg(&host->pdev->dev,
"(%s) set pending xfer complete\n", __func__);
atmci_set_pending(host, EVENT_XFER_COMPLETE);
tasklet_schedule(&host->tasklet);
}
}
static void atmci_dma_cleanup(struct atmel_mci *host)
{
struct mmc_data *data = host->data;
if (data)
dma_unmap_sg(host->dma.chan->device->dev,
data->sg, data->sg_len,
((data->flags & MMC_DATA_WRITE)
? DMA_TO_DEVICE : DMA_FROM_DEVICE));
}
/*
* This function is called by the DMA driver from tasklet context.
*/
static void atmci_dma_complete(void *arg)
{
struct atmel_mci *host = arg;
struct mmc_data *data = host->data;
dev_vdbg(&host->pdev->dev, "DMA complete\n");
if (host->caps.has_dma_conf_reg)
/* Disable DMA hardware handshaking on MCI */
atmci_writel(host, ATMCI_DMA, atmci_readl(host, ATMCI_DMA) & ~ATMCI_DMAEN);
atmci_dma_cleanup(host);
/*
* If the card was removed, data will be NULL. No point trying
* to send the stop command or waiting for NBUSY in this case.
*/
if (data) {
dev_dbg(&host->pdev->dev,
"(%s) set pending xfer complete\n", __func__);
atmci_set_pending(host, EVENT_XFER_COMPLETE);
tasklet_schedule(&host->tasklet);
/*
* Regardless of what the documentation says, we have
* to wait for NOTBUSY even after block read
* operations.
*
* When the DMA transfer is complete, the controller
* may still be reading the CRC from the card, i.e.
* the data transfer is still in progress and we
* haven't seen all the potential error bits yet.
*
* The interrupt handler will schedule a different
* tasklet to finish things up when the data transfer
* is completely done.
*
* We may not complete the mmc request here anyway
* because the mmc layer may call back and cause us to
* violate the "don't submit new operations from the
* completion callback" rule of the dma engine
* framework.
*/
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
}
}
/*
* Returns a mask of interrupt flags to be enabled after the whole
* request has been prepared.
*/
static u32 atmci_prepare_data(struct atmel_mci *host, struct mmc_data *data)
{
u32 iflags;
data->error = -EINPROGRESS;
host->sg = data->sg;
host->sg_len = data->sg_len;
host->data = data;
host->data_chan = NULL;
iflags = ATMCI_DATA_ERROR_FLAGS;
/*
* Errata: MMC data write operation with less than 12
* bytes is impossible.
*
* Errata: MCI Transmit Data Register (TDR) FIFO
* corruption when length is not multiple of 4.
*/
if (data->blocks * data->blksz < 12
|| (data->blocks * data->blksz) & 3)
host->need_reset = true;
host->pio_offset = 0;
if (data->flags & MMC_DATA_READ)
iflags |= ATMCI_RXRDY;
else
iflags |= ATMCI_TXRDY;
return iflags;
}
/*
* Set interrupt flags and set block length into the MCI mode register even
* if this value is also accessible in the MCI block register. It seems to be
* necessary before the High Speed MCI version. It also map sg and configure
* PDC registers.
*/
static u32
atmci_prepare_data_pdc(struct atmel_mci *host, struct mmc_data *data)
{
u32 iflags, tmp;
unsigned int sg_len;
enum dma_data_direction dir;
int i;
data->error = -EINPROGRESS;
host->data = data;
host->sg = data->sg;
iflags = ATMCI_DATA_ERROR_FLAGS;
/* Enable pdc mode */
atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCMODE);
if (data->flags & MMC_DATA_READ) {
dir = DMA_FROM_DEVICE;
iflags |= ATMCI_ENDRX | ATMCI_RXBUFF;
} else {
dir = DMA_TO_DEVICE;
iflags |= ATMCI_ENDTX | ATMCI_TXBUFE | ATMCI_BLKE;
}
/* Set BLKLEN */
tmp = atmci_readl(host, ATMCI_MR);
tmp &= 0x0000ffff;
tmp |= ATMCI_BLKLEN(data->blksz);
atmci_writel(host, ATMCI_MR, tmp);
/* Configure PDC */
host->data_size = data->blocks * data->blksz;
sg_len = dma_map_sg(&host->pdev->dev, data->sg, data->sg_len, dir);
if ((!host->caps.has_rwproof)
&& (host->data->flags & MMC_DATA_WRITE)) {
sg_copy_to_buffer(host->data->sg, host->data->sg_len,
host->buffer, host->data_size);
if (host->caps.has_bad_data_ordering)
for (i = 0; i < host->data_size; i++)
host->buffer[i] = swab32(host->buffer[i]);
}
if (host->data_size)
atmci_pdc_set_both_buf(host,
((dir == DMA_FROM_DEVICE) ? XFER_RECEIVE : XFER_TRANSMIT));
return iflags;
}
static u32
atmci_prepare_data_dma(struct atmel_mci *host, struct mmc_data *data)
{
struct dma_chan *chan;
struct dma_async_tx_descriptor *desc;
struct scatterlist *sg;
unsigned int i;
enum dma_data_direction direction;
enum dma_transfer_direction slave_dirn;
unsigned int sglen;
u32 maxburst;
u32 iflags;
data->error = -EINPROGRESS;
WARN_ON(host->data);
host->sg = NULL;
host->data = data;
iflags = ATMCI_DATA_ERROR_FLAGS;
/*
* We don't do DMA on "complex" transfers, i.e. with
* non-word-aligned buffers or lengths. Also, we don't bother
* with all the DMA setup overhead for short transfers.
*/
if (data->blocks * data->blksz < ATMCI_DMA_THRESHOLD)
return atmci_prepare_data(host, data);
if (data->blksz & 3)
return atmci_prepare_data(host, data);
for_each_sg(data->sg, sg, data->sg_len, i) {
if (sg->offset & 3 || sg->length & 3)
return atmci_prepare_data(host, data);
}
/* If we don't have a channel, we can't do DMA */
chan = host->dma.chan;
if (chan)
host->data_chan = chan;
if (!chan)
return -ENODEV;
if (data->flags & MMC_DATA_READ) {
direction = DMA_FROM_DEVICE;
host->dma_conf.direction = slave_dirn = DMA_DEV_TO_MEM;
maxburst = atmci_convert_chksize(host->dma_conf.src_maxburst);
} else {
direction = DMA_TO_DEVICE;
host->dma_conf.direction = slave_dirn = DMA_MEM_TO_DEV;
maxburst = atmci_convert_chksize(host->dma_conf.dst_maxburst);
}
if (host->caps.has_dma_conf_reg)
atmci_writel(host, ATMCI_DMA, ATMCI_DMA_CHKSIZE(maxburst) |
ATMCI_DMAEN);
sglen = dma_map_sg(chan->device->dev, data->sg,
data->sg_len, direction);
dmaengine_slave_config(chan, &host->dma_conf);
desc = dmaengine_prep_slave_sg(chan,
data->sg, sglen, slave_dirn,
DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
if (!desc)
goto unmap_exit;
host->dma.data_desc = desc;
desc->callback = atmci_dma_complete;
desc->callback_param = host;
return iflags;
unmap_exit:
dma_unmap_sg(chan->device->dev, data->sg, data->sg_len, direction);
return -ENOMEM;
}
static void
atmci_submit_data(struct atmel_mci *host, struct mmc_data *data)
{
return;
}
/*
* Start PDC according to transfer direction.
*/
static void
atmci_submit_data_pdc(struct atmel_mci *host, struct mmc_data *data)
{
if (data->flags & MMC_DATA_READ)
atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTEN);
else
atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_TXTEN);
}
static void
atmci_submit_data_dma(struct atmel_mci *host, struct mmc_data *data)
{
struct dma_chan *chan = host->data_chan;
struct dma_async_tx_descriptor *desc = host->dma.data_desc;
if (chan) {
dmaengine_submit(desc);
dma_async_issue_pending(chan);
}
}
static void atmci_stop_transfer(struct atmel_mci *host)
{
dev_dbg(&host->pdev->dev,
"(%s) set pending xfer complete\n", __func__);
atmci_set_pending(host, EVENT_XFER_COMPLETE);
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/*
* Stop data transfer because error(s) occurred.
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
*/
static void atmci_stop_transfer_pdc(struct atmel_mci *host)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTDIS | ATMEL_PDC_TXTDIS);
}
static void atmci_stop_transfer_dma(struct atmel_mci *host)
{
struct dma_chan *chan = host->data_chan;
if (chan) {
dmaengine_terminate_all(chan);
atmci_dma_cleanup(host);
} else {
/* Data transfer was stopped by the interrupt handler */
dev_dbg(&host->pdev->dev,
"(%s) set pending xfer complete\n", __func__);
atmci_set_pending(host, EVENT_XFER_COMPLETE);
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* Start a request: prepare data if needed, prepare the command and activate
* interrupts.
*/
static void atmci_start_request(struct atmel_mci *host,
struct atmel_mci_slot *slot)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
struct mmc_request *mrq;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct mmc_command *cmd;
struct mmc_data *data;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
u32 iflags;
u32 cmdflags;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
mrq = slot->mrq;
host->cur_slot = slot;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->mrq = mrq;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->pending_events = 0;
host->completed_events = 0;
host->cmd_status = 0;
host->data_status = 0;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_dbg(&host->pdev->dev, "start request: cmd %u\n", mrq->cmd->opcode);
if (host->need_reset || host->caps.need_reset_after_xfer) {
iflags = atmci_readl(host, ATMCI_IMR);
iflags &= (ATMCI_SDIOIRQA | ATMCI_SDIOIRQB);
atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
atmci_writel(host, ATMCI_MR, host->mode_reg);
if (host->caps.has_cfg_reg)
atmci_writel(host, ATMCI_CFG, host->cfg_reg);
atmci_writel(host, ATMCI_IER, iflags);
host->need_reset = false;
}
atmci_writel(host, ATMCI_SDCR, slot->sdc_reg);
iflags = atmci_readl(host, ATMCI_IMR);
if (iflags & ~(ATMCI_SDIOIRQA | ATMCI_SDIOIRQB))
dev_dbg(&slot->mmc->class_dev, "WARNING: IMR=0x%08x\n",
iflags);
if (unlikely(test_and_clear_bit(ATMCI_CARD_NEED_INIT, &slot->flags))) {
/* Send init sequence (74 clock cycles) */
atmci_writel(host, ATMCI_CMDR, ATMCI_CMDR_SPCMD_INIT);
while (!(atmci_readl(host, ATMCI_SR) & ATMCI_CMDRDY))
cpu_relax();
}
iflags = 0;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
data = mrq->data;
if (data) {
atmci_set_timeout(host, slot, data);
/* Must set block count/size before sending command */
atmci_writel(host, ATMCI_BLKR, ATMCI_BCNT(data->blocks)
| ATMCI_BLKLEN(data->blksz));
dev_vdbg(&slot->mmc->class_dev, "BLKR=0x%08x\n",
ATMCI_BCNT(data->blocks) | ATMCI_BLKLEN(data->blksz));
iflags |= host->prepare_data(host, data);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
iflags |= ATMCI_CMDRDY;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
cmd = mrq->cmd;
cmdflags = atmci_prepare_command(slot->mmc, cmd);
atmci_send_command(host, cmd, cmdflags);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (data)
host->submit_data(host, data);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (mrq->stop) {
host->stop_cmdr = atmci_prepare_command(slot->mmc, mrq->stop);
host->stop_cmdr |= ATMCI_CMDR_STOP_XFER;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (!(data->flags & MMC_DATA_WRITE))
host->stop_cmdr |= ATMCI_CMDR_TRDIR_READ;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (data->flags & MMC_DATA_STREAM)
host->stop_cmdr |= ATMCI_CMDR_STREAM;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
else
host->stop_cmdr |= ATMCI_CMDR_MULTI_BLOCK;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* We could have enabled interrupts earlier, but I suspect
* that would open up a nice can of interesting race
* conditions (e.g. command and data complete, but stop not
* prepared yet.)
*/
atmci_writel(host, ATMCI_IER, iflags);
mod_timer(&host->timer, jiffies + msecs_to_jiffies(2000));
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static void atmci_queue_request(struct atmel_mci *host,
struct atmel_mci_slot *slot, struct mmc_request *mrq)
{
dev_vdbg(&slot->mmc->class_dev, "queue request: state=%d\n",
host->state);
spin_lock_bh(&host->lock);
slot->mrq = mrq;
if (host->state == STATE_IDLE) {
host->state = STATE_SENDING_CMD;
atmci_start_request(host, slot);
} else {
dev_dbg(&host->pdev->dev, "queue request\n");
list_add_tail(&slot->queue_node, &host->queue);
}
spin_unlock_bh(&host->lock);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static void atmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
{
struct atmel_mci_slot *slot = mmc_priv(mmc);
struct atmel_mci *host = slot->host;
struct mmc_data *data;
WARN_ON(slot->mrq);
dev_dbg(&host->pdev->dev, "MRQ: cmd %u\n", mrq->cmd->opcode);
/*
* We may "know" the card is gone even though there's still an
* electrical connection. If so, we really need to communicate
* this to the MMC core since there won't be any more
* interrupts as the card is completely removed. Otherwise,
* the MMC core might believe the card is still there even
* though the card was just removed very slowly.
*/
if (!test_bit(ATMCI_CARD_PRESENT, &slot->flags)) {
mrq->cmd->error = -ENOMEDIUM;
mmc_request_done(mmc, mrq);
return;
}
/* We don't support multiple blocks of weird lengths. */
data = mrq->data;
if (data && data->blocks > 1 && data->blksz & 3) {
mrq->cmd->error = -EINVAL;
mmc_request_done(mmc, mrq);
}
atmci_queue_request(host, slot, mrq);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
{
struct atmel_mci_slot *slot = mmc_priv(mmc);
struct atmel_mci *host = slot->host;
unsigned int i;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
slot->sdc_reg &= ~ATMCI_SDCBUS_MASK;
switch (ios->bus_width) {
case MMC_BUS_WIDTH_1:
slot->sdc_reg |= ATMCI_SDCBUS_1BIT;
break;
case MMC_BUS_WIDTH_4:
slot->sdc_reg |= ATMCI_SDCBUS_4BIT;
break;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (ios->clock) {
unsigned int clock_min = ~0U;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
u32 clkdiv;
spin_lock_bh(&host->lock);
if (!host->mode_reg) {
clk_enable(host->mck);
atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
if (host->caps.has_cfg_reg)
atmci_writel(host, ATMCI_CFG, host->cfg_reg);
}
/*
* Use mirror of ios->clock to prevent race with mmc
* core ios update when finding the minimum.
*/
slot->clock = ios->clock;
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
if (host->slot[i] && host->slot[i]->clock
&& host->slot[i]->clock < clock_min)
clock_min = host->slot[i]->clock;
}
/* Calculate clock divider */
if (host->caps.has_odd_clk_div) {
clkdiv = DIV_ROUND_UP(host->bus_hz, clock_min) - 2;
if (clkdiv > 511) {
dev_warn(&mmc->class_dev,
"clock %u too slow; using %lu\n",
clock_min, host->bus_hz / (511 + 2));
clkdiv = 511;
}
host->mode_reg = ATMCI_MR_CLKDIV(clkdiv >> 1)
| ATMCI_MR_CLKODD(clkdiv & 1);
} else {
clkdiv = DIV_ROUND_UP(host->bus_hz, 2 * clock_min) - 1;
if (clkdiv > 255) {
dev_warn(&mmc->class_dev,
"clock %u too slow; using %lu\n",
clock_min, host->bus_hz / (2 * 256));
clkdiv = 255;
}
host->mode_reg = ATMCI_MR_CLKDIV(clkdiv);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
/*
* WRPROOF and RDPROOF prevent overruns/underruns by
* stopping the clock when the FIFO is full/empty.
* This state is not expected to last for long.
*/
if (host->caps.has_rwproof)
host->mode_reg |= (ATMCI_MR_WRPROOF | ATMCI_MR_RDPROOF);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (host->caps.has_cfg_reg) {
/* setup High Speed mode in relation with card capacity */
if (ios->timing == MMC_TIMING_SD_HS)
host->cfg_reg |= ATMCI_CFG_HSMODE;
else
host->cfg_reg &= ~ATMCI_CFG_HSMODE;
}
if (list_empty(&host->queue)) {
atmci_writel(host, ATMCI_MR, host->mode_reg);
if (host->caps.has_cfg_reg)
atmci_writel(host, ATMCI_CFG, host->cfg_reg);
} else {
host->need_clock_update = true;
}
spin_unlock_bh(&host->lock);
} else {
bool any_slot_active = false;
spin_lock_bh(&host->lock);
slot->clock = 0;
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
if (host->slot[i] && host->slot[i]->clock) {
any_slot_active = true;
break;
}
}
if (!any_slot_active) {
atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIDIS);
if (host->mode_reg) {
atmci_readl(host, ATMCI_MR);
clk_disable(host->mck);
}
host->mode_reg = 0;
}
spin_unlock_bh(&host->lock);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
switch (ios->power_mode) {
case MMC_POWER_UP:
set_bit(ATMCI_CARD_NEED_INIT, &slot->flags);
break;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
default:
/*
* TODO: None of the currently available AVR32-based
* boards allow MMC power to be turned off. Implement
* power control when this can be tested properly.
*
* We also need to hook this into the clock management
* somehow so that newly inserted cards aren't
* subjected to a fast clock before we have a chance
* to figure out what the maximum rate is. Currently,
* there's no way to avoid this, and there never will
* be for boards that don't support power control.
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
*/
break;
}
}
static int atmci_get_ro(struct mmc_host *mmc)
{
int read_only = -ENOSYS;
struct atmel_mci_slot *slot = mmc_priv(mmc);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (gpio_is_valid(slot->wp_pin)) {
read_only = gpio_get_value(slot->wp_pin);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_dbg(&mmc->class_dev, "card is %s\n",
read_only ? "read-only" : "read-write");
}
return read_only;
}
static int atmci_get_cd(struct mmc_host *mmc)
{
int present = -ENOSYS;
struct atmel_mci_slot *slot = mmc_priv(mmc);
if (gpio_is_valid(slot->detect_pin)) {
present = !(gpio_get_value(slot->detect_pin) ^
slot->detect_is_active_high);
dev_dbg(&mmc->class_dev, "card is %spresent\n",
present ? "" : "not ");
}
return present;
}
static void atmci_enable_sdio_irq(struct mmc_host *mmc, int enable)
{
struct atmel_mci_slot *slot = mmc_priv(mmc);
struct atmel_mci *host = slot->host;
if (enable)
atmci_writel(host, ATMCI_IER, slot->sdio_irq);
else
atmci_writel(host, ATMCI_IDR, slot->sdio_irq);
}
static const struct mmc_host_ops atmci_ops = {
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
.request = atmci_request,
.set_ios = atmci_set_ios,
.get_ro = atmci_get_ro,
.get_cd = atmci_get_cd,
.enable_sdio_irq = atmci_enable_sdio_irq,
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
};
/* Called with host->lock held */
static void atmci_request_end(struct atmel_mci *host, struct mmc_request *mrq)
__releases(&host->lock)
__acquires(&host->lock)
{
struct atmel_mci_slot *slot = NULL;
struct mmc_host *prev_mmc = host->cur_slot->mmc;
WARN_ON(host->cmd || host->data);
/*
* Update the MMC clock rate if necessary. This may be
* necessary if set_ios() is called when a different slot is
* busy transferring data.
*/
if (host->need_clock_update) {
atmci_writel(host, ATMCI_MR, host->mode_reg);
if (host->caps.has_cfg_reg)
atmci_writel(host, ATMCI_CFG, host->cfg_reg);
}
host->cur_slot->mrq = NULL;
host->mrq = NULL;
if (!list_empty(&host->queue)) {
slot = list_entry(host->queue.next,
struct atmel_mci_slot, queue_node);
list_del(&slot->queue_node);
dev_vdbg(&host->pdev->dev, "list not empty: %s is next\n",
mmc_hostname(slot->mmc));
host->state = STATE_SENDING_CMD;
atmci_start_request(host, slot);
} else {
dev_vdbg(&host->pdev->dev, "list empty\n");
host->state = STATE_IDLE;
}
del_timer(&host->timer);
spin_unlock(&host->lock);
mmc_request_done(prev_mmc, mrq);
spin_lock(&host->lock);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static void atmci_command_complete(struct atmel_mci *host,
struct mmc_command *cmd)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
{
u32 status = host->cmd_status;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/* Read the response from the card (up to 16 bytes) */
cmd->resp[0] = atmci_readl(host, ATMCI_RSPR);
cmd->resp[1] = atmci_readl(host, ATMCI_RSPR);
cmd->resp[2] = atmci_readl(host, ATMCI_RSPR);
cmd->resp[3] = atmci_readl(host, ATMCI_RSPR);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (status & ATMCI_RTOE)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
cmd->error = -ETIMEDOUT;
else if ((cmd->flags & MMC_RSP_CRC) && (status & ATMCI_RCRCE))
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
cmd->error = -EILSEQ;
else if (status & (ATMCI_RINDE | ATMCI_RDIRE | ATMCI_RENDE))
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
cmd->error = -EIO;
else if (host->mrq->data && (host->mrq->data->blksz & 3)) {
if (host->caps.need_blksz_mul_4) {
cmd->error = -EINVAL;
host->need_reset = 1;
}
} else
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
cmd->error = 0;
}
static void atmci_detect_change(unsigned long data)
{
struct atmel_mci_slot *slot = (struct atmel_mci_slot *)data;
bool present;
bool present_old;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/*
* atmci_cleanup_slot() sets the ATMCI_SHUTDOWN flag before
* freeing the interrupt. We must not re-enable the interrupt
* if it has been freed, and if we're shutting down, it
* doesn't really matter whether the card is present or not.
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
*/
smp_rmb();
if (test_bit(ATMCI_SHUTDOWN, &slot->flags))
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return;
enable_irq(gpio_to_irq(slot->detect_pin));
present = !(gpio_get_value(slot->detect_pin) ^
slot->detect_is_active_high);
present_old = test_bit(ATMCI_CARD_PRESENT, &slot->flags);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_vdbg(&slot->mmc->class_dev, "detect change: %d (was %d)\n",
present, present_old);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (present != present_old) {
struct atmel_mci *host = slot->host;
struct mmc_request *mrq;
dev_dbg(&slot->mmc->class_dev, "card %s\n",
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
present ? "inserted" : "removed");
spin_lock(&host->lock);
if (!present)
clear_bit(ATMCI_CARD_PRESENT, &slot->flags);
else
set_bit(ATMCI_CARD_PRESENT, &slot->flags);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/* Clean up queue if present */
mrq = slot->mrq;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (mrq) {
if (mrq == host->mrq) {
/*
* Reset controller to terminate any ongoing
* commands or data transfers.
*/
atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
atmci_writel(host, ATMCI_MR, host->mode_reg);
if (host->caps.has_cfg_reg)
atmci_writel(host, ATMCI_CFG, host->cfg_reg);
host->data = NULL;
host->cmd = NULL;
switch (host->state) {
case STATE_IDLE:
break;
case STATE_SENDING_CMD:
mrq->cmd->error = -ENOMEDIUM;
if (mrq->data)
host->stop_transfer(host);
break;
case STATE_DATA_XFER:
mrq->data->error = -ENOMEDIUM;
host->stop_transfer(host);
break;
case STATE_WAITING_NOTBUSY:
mrq->data->error = -ENOMEDIUM;
break;
case STATE_SENDING_STOP:
mrq->stop->error = -ENOMEDIUM;
break;
case STATE_END_REQUEST:
break;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
atmci_request_end(host, mrq);
} else {
list_del(&slot->queue_node);
mrq->cmd->error = -ENOMEDIUM;
if (mrq->data)
mrq->data->error = -ENOMEDIUM;
if (mrq->stop)
mrq->stop->error = -ENOMEDIUM;
spin_unlock(&host->lock);
mmc_request_done(slot->mmc, mrq);
spin_lock(&host->lock);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
spin_unlock(&host->lock);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
mmc_detect_change(slot->mmc, 0);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
}
static void atmci_tasklet_func(unsigned long priv)
{
struct atmel_mci *host = (struct atmel_mci *)priv;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
struct mmc_request *mrq = host->mrq;
struct mmc_data *data = host->data;
enum atmel_mci_state state = host->state;
enum atmel_mci_state prev_state;
u32 status;
spin_lock(&host->lock);
state = host->state;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_vdbg(&host->pdev->dev,
"tasklet: state %u pending/completed/mask %lx/%lx/%x\n",
state, host->pending_events, host->completed_events,
atmci_readl(host, ATMCI_IMR));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
do {
prev_state = state;
dev_dbg(&host->pdev->dev, "FSM: state=%d\n", state);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
switch (state) {
case STATE_IDLE:
break;
case STATE_SENDING_CMD:
/*
* Command has been sent, we are waiting for command
* ready. Then we have three next states possible:
* END_REQUEST by default, WAITING_NOTBUSY if it's a
* command needing it or DATA_XFER if there is data.
*/
dev_dbg(&host->pdev->dev, "FSM: cmd ready?\n");
if (!atmci_test_and_clear_pending(host,
EVENT_CMD_RDY))
break;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_dbg(&host->pdev->dev, "set completed cmd ready\n");
host->cmd = NULL;
atmci_set_completed(host, EVENT_CMD_RDY);
atmci_command_complete(host, mrq->cmd);
if (mrq->data) {
dev_dbg(&host->pdev->dev,
"command with data transfer");
/*
* If there is a command error don't start
* data transfer.
*/
if (mrq->cmd->error) {
host->stop_transfer(host);
host->data = NULL;
atmci_writel(host, ATMCI_IDR,
ATMCI_TXRDY | ATMCI_RXRDY
| ATMCI_DATA_ERROR_FLAGS);
state = STATE_END_REQUEST;
} else
state = STATE_DATA_XFER;
} else if ((!mrq->data) && (mrq->cmd->flags & MMC_RSP_BUSY)) {
dev_dbg(&host->pdev->dev,
"command response need waiting notbusy");
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
state = STATE_WAITING_NOTBUSY;
} else
state = STATE_END_REQUEST;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
break;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
case STATE_DATA_XFER:
if (atmci_test_and_clear_pending(host,
EVENT_DATA_ERROR)) {
dev_dbg(&host->pdev->dev, "set completed data error\n");
atmci_set_completed(host, EVENT_DATA_ERROR);
state = STATE_END_REQUEST;
break;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/*
* A data transfer is in progress. The event expected
* to move to the next state depends of data transfer
* type (PDC or DMA). Once transfer done we can move
* to the next step which is WAITING_NOTBUSY in write
* case and directly SENDING_STOP in read case.
*/
dev_dbg(&host->pdev->dev, "FSM: xfer complete?\n");
if (!atmci_test_and_clear_pending(host,
EVENT_XFER_COMPLETE))
break;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_dbg(&host->pdev->dev,
"(%s) set completed xfer complete\n",
__func__);
atmci_set_completed(host, EVENT_XFER_COMPLETE);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (host->caps.need_notbusy_for_read_ops ||
(host->data->flags & MMC_DATA_WRITE)) {
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
state = STATE_WAITING_NOTBUSY;
} else if (host->mrq->stop) {
atmci_writel(host, ATMCI_IER, ATMCI_CMDRDY);
atmci_send_stop_cmd(host, data);
state = STATE_SENDING_STOP;
} else {
host->data = NULL;
data->bytes_xfered = data->blocks * data->blksz;
data->error = 0;
state = STATE_END_REQUEST;
}
break;
case STATE_WAITING_NOTBUSY:
/*
* We can be in the state for two reasons: a command
* requiring waiting not busy signal (stop command
* included) or a write operation. In the latest case,
* we need to send a stop command.
*/
dev_dbg(&host->pdev->dev, "FSM: not busy?\n");
if (!atmci_test_and_clear_pending(host,
EVENT_NOTBUSY))
break;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
dev_dbg(&host->pdev->dev, "set completed not busy\n");
atmci_set_completed(host, EVENT_NOTBUSY);
if (host->data) {
/*
* For some commands such as CMD53, even if
* there is data transfer, there is no stop
* command to send.
*/
if (host->mrq->stop) {
atmci_writel(host, ATMCI_IER,
ATMCI_CMDRDY);
atmci_send_stop_cmd(host, data);
state = STATE_SENDING_STOP;
} else {
host->data = NULL;
data->bytes_xfered = data->blocks
* data->blksz;
data->error = 0;
state = STATE_END_REQUEST;
}
} else
state = STATE_END_REQUEST;
break;
case STATE_SENDING_STOP:
/*
* In this state, it is important to set host->data to
* NULL (which is tested in the waiting notbusy state)
* in order to go to the end request state instead of
* sending stop again.
*/
dev_dbg(&host->pdev->dev, "FSM: cmd ready?\n");
if (!atmci_test_and_clear_pending(host,
EVENT_CMD_RDY))
break;
dev_dbg(&host->pdev->dev, "FSM: cmd ready\n");
host->cmd = NULL;
data->bytes_xfered = data->blocks * data->blksz;
data->error = 0;
atmci_command_complete(host, mrq->stop);
if (mrq->stop->error) {
host->stop_transfer(host);
atmci_writel(host, ATMCI_IDR,
ATMCI_TXRDY | ATMCI_RXRDY
| ATMCI_DATA_ERROR_FLAGS);
state = STATE_END_REQUEST;
} else {
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
state = STATE_WAITING_NOTBUSY;
}
host->data = NULL;
break;
case STATE_END_REQUEST:
atmci_writel(host, ATMCI_IDR, ATMCI_TXRDY | ATMCI_RXRDY
| ATMCI_DATA_ERROR_FLAGS);
status = host->data_status;
if (unlikely(status)) {
host->stop_transfer(host);
host->data = NULL;
if (status & ATMCI_DTOE) {
data->error = -ETIMEDOUT;
} else if (status & ATMCI_DCRCE) {
data->error = -EILSEQ;
} else {
data->error = -EIO;
}
}
atmci_request_end(host, host->mrq);
state = STATE_IDLE;
break;
}
} while (state != prev_state);
host->state = state;
spin_unlock(&host->lock);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_read_data_pio(struct atmel_mci *host)
{
struct scatterlist *sg = host->sg;
void *buf = sg_virt(sg);
unsigned int offset = host->pio_offset;
struct mmc_data *data = host->data;
u32 value;
u32 status;
unsigned int nbytes = 0;
do {
value = atmci_readl(host, ATMCI_RDR);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (likely(offset + 4 <= sg->length)) {
put_unaligned(value, (u32 *)(buf + offset));
offset += 4;
nbytes += 4;
if (offset == sg->length) {
flush_dcache_page(sg_page(sg));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->sg = sg = sg_next(sg);
host->sg_len--;
if (!sg || !host->sg_len)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
goto done;
offset = 0;
buf = sg_virt(sg);
}
} else {
unsigned int remaining = sg->length - offset;
memcpy(buf + offset, &value, remaining);
nbytes += remaining;
flush_dcache_page(sg_page(sg));
host->sg = sg = sg_next(sg);
host->sg_len--;
if (!sg || !host->sg_len)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
goto done;
offset = 4 - remaining;
buf = sg_virt(sg);
memcpy(buf, (u8 *)&value + remaining, offset);
nbytes += offset;
}
status = atmci_readl(host, ATMCI_SR);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (status & ATMCI_DATA_ERROR_FLAGS) {
atmci_writel(host, ATMCI_IDR, (ATMCI_NOTBUSY | ATMCI_RXRDY
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
| ATMCI_DATA_ERROR_FLAGS));
host->data_status = status;
data->bytes_xfered += nbytes;
return;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
} while (status & ATMCI_RXRDY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->pio_offset = offset;
data->bytes_xfered += nbytes;
return;
done:
atmci_writel(host, ATMCI_IDR, ATMCI_RXRDY);
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
data->bytes_xfered += nbytes;
smp_wmb();
atmci_set_pending(host, EVENT_XFER_COMPLETE);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_write_data_pio(struct atmel_mci *host)
{
struct scatterlist *sg = host->sg;
void *buf = sg_virt(sg);
unsigned int offset = host->pio_offset;
struct mmc_data *data = host->data;
u32 value;
u32 status;
unsigned int nbytes = 0;
do {
if (likely(offset + 4 <= sg->length)) {
value = get_unaligned((u32 *)(buf + offset));
atmci_writel(host, ATMCI_TDR, value);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
offset += 4;
nbytes += 4;
if (offset == sg->length) {
host->sg = sg = sg_next(sg);
host->sg_len--;
if (!sg || !host->sg_len)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
goto done;
offset = 0;
buf = sg_virt(sg);
}
} else {
unsigned int remaining = sg->length - offset;
value = 0;
memcpy(&value, buf + offset, remaining);
nbytes += remaining;
host->sg = sg = sg_next(sg);
host->sg_len--;
if (!sg || !host->sg_len) {
atmci_writel(host, ATMCI_TDR, value);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
goto done;
}
offset = 4 - remaining;
buf = sg_virt(sg);
memcpy((u8 *)&value + remaining, buf, offset);
atmci_writel(host, ATMCI_TDR, value);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
nbytes += offset;
}
status = atmci_readl(host, ATMCI_SR);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (status & ATMCI_DATA_ERROR_FLAGS) {
atmci_writel(host, ATMCI_IDR, (ATMCI_NOTBUSY | ATMCI_TXRDY
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
| ATMCI_DATA_ERROR_FLAGS));
host->data_status = status;
data->bytes_xfered += nbytes;
return;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
} while (status & ATMCI_TXRDY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->pio_offset = offset;
data->bytes_xfered += nbytes;
return;
done:
atmci_writel(host, ATMCI_IDR, ATMCI_TXRDY);
atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
data->bytes_xfered += nbytes;
smp_wmb();
atmci_set_pending(host, EVENT_XFER_COMPLETE);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
static void atmci_sdio_interrupt(struct atmel_mci *host, u32 status)
{
int i;
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
struct atmel_mci_slot *slot = host->slot[i];
if (slot && (status & slot->sdio_irq)) {
mmc_signal_sdio_irq(slot->mmc);
}
}
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static irqreturn_t atmci_interrupt(int irq, void *dev_id)
{
struct atmel_mci *host = dev_id;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
u32 status, mask, pending;
unsigned int pass_count = 0;
do {
status = atmci_readl(host, ATMCI_SR);
mask = atmci_readl(host, ATMCI_IMR);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
pending = status & mask;
if (!pending)
break;
if (pending & ATMCI_DATA_ERROR_FLAGS) {
dev_dbg(&host->pdev->dev, "IRQ: data error\n");
atmci_writel(host, ATMCI_IDR, ATMCI_DATA_ERROR_FLAGS
| ATMCI_RXRDY | ATMCI_TXRDY
| ATMCI_ENDRX | ATMCI_ENDTX
| ATMCI_RXBUFF | ATMCI_TXBUFE);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->data_status = status;
dev_dbg(&host->pdev->dev, "set pending data error\n");
smp_wmb();
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
atmci_set_pending(host, EVENT_DATA_ERROR);
tasklet_schedule(&host->tasklet);
}
if (pending & ATMCI_TXBUFE) {
dev_dbg(&host->pdev->dev, "IRQ: tx buffer empty\n");
atmci_writel(host, ATMCI_IDR, ATMCI_TXBUFE);
atmci_writel(host, ATMCI_IDR, ATMCI_ENDTX);
/*
* We can receive this interruption before having configured
* the second pdc buffer, so we need to reconfigure first and
* second buffers again
*/
if (host->data_size) {
atmci_pdc_set_both_buf(host, XFER_TRANSMIT);
atmci_writel(host, ATMCI_IER, ATMCI_ENDTX);
atmci_writel(host, ATMCI_IER, ATMCI_TXBUFE);
} else {
atmci_pdc_complete(host);
}
} else if (pending & ATMCI_ENDTX) {
dev_dbg(&host->pdev->dev, "IRQ: end of tx buffer\n");
atmci_writel(host, ATMCI_IDR, ATMCI_ENDTX);
if (host->data_size) {
atmci_pdc_set_single_buf(host,
XFER_TRANSMIT, PDC_SECOND_BUF);
atmci_writel(host, ATMCI_IER, ATMCI_ENDTX);
}
}
if (pending & ATMCI_RXBUFF) {
dev_dbg(&host->pdev->dev, "IRQ: rx buffer full\n");
atmci_writel(host, ATMCI_IDR, ATMCI_RXBUFF);
atmci_writel(host, ATMCI_IDR, ATMCI_ENDRX);
/*
* We can receive this interruption before having configured
* the second pdc buffer, so we need to reconfigure first and
* second buffers again
*/
if (host->data_size) {
atmci_pdc_set_both_buf(host, XFER_RECEIVE);
atmci_writel(host, ATMCI_IER, ATMCI_ENDRX);
atmci_writel(host, ATMCI_IER, ATMCI_RXBUFF);
} else {
atmci_pdc_complete(host);
}
} else if (pending & ATMCI_ENDRX) {
dev_dbg(&host->pdev->dev, "IRQ: end of rx buffer\n");
atmci_writel(host, ATMCI_IDR, ATMCI_ENDRX);
if (host->data_size) {
atmci_pdc_set_single_buf(host,
XFER_RECEIVE, PDC_SECOND_BUF);
atmci_writel(host, ATMCI_IER, ATMCI_ENDRX);
}
}
/*
* First mci IPs, so mainly the ones having pdc, have some
* issues with the notbusy signal. You can't get it after
* data transmission if you have not sent a stop command.
* The appropriate workaround is to use the BLKE signal.
*/
if (pending & ATMCI_BLKE) {
dev_dbg(&host->pdev->dev, "IRQ: blke\n");
atmci_writel(host, ATMCI_IDR, ATMCI_BLKE);
smp_wmb();
dev_dbg(&host->pdev->dev, "set pending notbusy\n");
atmci_set_pending(host, EVENT_NOTBUSY);
tasklet_schedule(&host->tasklet);
}
if (pending & ATMCI_NOTBUSY) {
dev_dbg(&host->pdev->dev, "IRQ: not_busy\n");
atmci_writel(host, ATMCI_IDR, ATMCI_NOTBUSY);
smp_wmb();
dev_dbg(&host->pdev->dev, "set pending notbusy\n");
atmci_set_pending(host, EVENT_NOTBUSY);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
tasklet_schedule(&host->tasklet);
}
if (pending & ATMCI_RXRDY)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
atmci_read_data_pio(host);
if (pending & ATMCI_TXRDY)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
atmci_write_data_pio(host);
if (pending & ATMCI_CMDRDY) {
dev_dbg(&host->pdev->dev, "IRQ: cmd ready\n");
atmci_writel(host, ATMCI_IDR, ATMCI_CMDRDY);
host->cmd_status = status;
smp_wmb();
dev_dbg(&host->pdev->dev, "set pending cmd rdy\n");
atmci_set_pending(host, EVENT_CMD_RDY);
tasklet_schedule(&host->tasklet);
}
if (pending & (ATMCI_SDIOIRQA | ATMCI_SDIOIRQB))
atmci_sdio_interrupt(host, status);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
} while (pass_count++ < 5);
return pass_count ? IRQ_HANDLED : IRQ_NONE;
}
static irqreturn_t atmci_detect_interrupt(int irq, void *dev_id)
{
struct atmel_mci_slot *slot = dev_id;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
/*
* Disable interrupts until the pin has stabilized and check
* the state then. Use mod_timer() since we may be in the
* middle of the timer routine when this interrupt triggers.
*/
disable_irq_nosync(irq);
mod_timer(&slot->detect_timer, jiffies + msecs_to_jiffies(20));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return IRQ_HANDLED;
}
static int __init atmci_init_slot(struct atmel_mci *host,
struct mci_slot_pdata *slot_data, unsigned int id,
u32 sdc_reg, u32 sdio_irq)
{
struct mmc_host *mmc;
struct atmel_mci_slot *slot;
mmc = mmc_alloc_host(sizeof(struct atmel_mci_slot), &host->pdev->dev);
if (!mmc)
return -ENOMEM;
slot = mmc_priv(mmc);
slot->mmc = mmc;
slot->host = host;
slot->detect_pin = slot_data->detect_pin;
slot->wp_pin = slot_data->wp_pin;
slot->detect_is_active_high = slot_data->detect_is_active_high;
slot->sdc_reg = sdc_reg;
slot->sdio_irq = sdio_irq;
dev_dbg(&mmc->class_dev,
"slot[%u]: bus_width=%u, detect_pin=%d, "
"detect_is_active_high=%s, wp_pin=%d\n",
id, slot_data->bus_width, slot_data->detect_pin,
slot_data->detect_is_active_high ? "true" : "false",
slot_data->wp_pin);
mmc->ops = &atmci_ops;
mmc->f_min = DIV_ROUND_UP(host->bus_hz, 512);
mmc->f_max = host->bus_hz / 2;
mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
if (sdio_irq)
mmc->caps |= MMC_CAP_SDIO_IRQ;
if (host->caps.has_highspeed)
mmc->caps |= MMC_CAP_SD_HIGHSPEED;
/*
* Without the read/write proof capability, it is strongly suggested to
* use only one bit for data to prevent fifo underruns and overruns
* which will corrupt data.
*/
if ((slot_data->bus_width >= 4) && host->caps.has_rwproof)
mmc->caps |= MMC_CAP_4_BIT_DATA;
if (atmci_get_version(host) < 0x200) {
mmc->max_segs = 256;
mmc->max_blk_size = 4095;
mmc->max_blk_count = 256;
mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
mmc->max_seg_size = mmc->max_blk_size * mmc->max_segs;
} else {
mmc->max_segs = 64;
mmc->max_req_size = 32768 * 512;
mmc->max_blk_size = 32768;
mmc->max_blk_count = 512;
}
/* Assume card is present initially */
set_bit(ATMCI_CARD_PRESENT, &slot->flags);
if (gpio_is_valid(slot->detect_pin)) {
if (gpio_request(slot->detect_pin, "mmc_detect")) {
dev_dbg(&mmc->class_dev, "no detect pin available\n");
slot->detect_pin = -EBUSY;
} else if (gpio_get_value(slot->detect_pin) ^
slot->detect_is_active_high) {
clear_bit(ATMCI_CARD_PRESENT, &slot->flags);
}
}
if (!gpio_is_valid(slot->detect_pin))
mmc->caps |= MMC_CAP_NEEDS_POLL;
if (gpio_is_valid(slot->wp_pin)) {
if (gpio_request(slot->wp_pin, "mmc_wp")) {
dev_dbg(&mmc->class_dev, "no WP pin available\n");
slot->wp_pin = -EBUSY;
}
}
host->slot[id] = slot;
mmc_add_host(mmc);
if (gpio_is_valid(slot->detect_pin)) {
int ret;
setup_timer(&slot->detect_timer, atmci_detect_change,
(unsigned long)slot);
ret = request_irq(gpio_to_irq(slot->detect_pin),
atmci_detect_interrupt,
IRQF_TRIGGER_FALLING | IRQF_TRIGGER_RISING,
"mmc-detect", slot);
if (ret) {
dev_dbg(&mmc->class_dev,
"could not request IRQ %d for detect pin\n",
gpio_to_irq(slot->detect_pin));
gpio_free(slot->detect_pin);
slot->detect_pin = -EBUSY;
}
}
atmci_init_debugfs(slot);
return 0;
}
static void __exit atmci_cleanup_slot(struct atmel_mci_slot *slot,
unsigned int id)
{
/* Debugfs stuff is cleaned up by mmc core */
set_bit(ATMCI_SHUTDOWN, &slot->flags);
smp_wmb();
mmc_remove_host(slot->mmc);
if (gpio_is_valid(slot->detect_pin)) {
int pin = slot->detect_pin;
free_irq(gpio_to_irq(pin), slot);
del_timer_sync(&slot->detect_timer);
gpio_free(pin);
}
if (gpio_is_valid(slot->wp_pin))
gpio_free(slot->wp_pin);
slot->host->slot[id] = NULL;
mmc_free_host(slot->mmc);
}
static bool atmci_filter(struct dma_chan *chan, void *pdata)
{
struct mci_platform_data *sl_pdata = pdata;
struct mci_dma_data *sl;
if (!sl_pdata)
return false;
sl = sl_pdata->dma_slave;
if (sl && find_slave_dev(sl) == chan->device->dev) {
chan->private = slave_data_ptr(sl);
return true;
} else {
return false;
}
}
static bool atmci_configure_dma(struct atmel_mci *host)
{
struct mci_platform_data *pdata;
dma_cap_mask_t mask;
if (host == NULL)
return false;
pdata = host->pdev->dev.platform_data;
dma_cap_zero(mask);
dma_cap_set(DMA_SLAVE, mask);
host->dma.chan = dma_request_slave_channel_compat(mask, atmci_filter, pdata,
&host->pdev->dev, "rxtx");
if (!host->dma.chan) {
dev_warn(&host->pdev->dev, "no DMA channel available\n");
return false;
} else {
dev_info(&host->pdev->dev,
"using %s for DMA transfers\n",
dma_chan_name(host->dma.chan));
host->dma_conf.src_addr = host->mapbase + ATMCI_RDR;
host->dma_conf.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
host->dma_conf.src_maxburst = 1;
host->dma_conf.dst_addr = host->mapbase + ATMCI_TDR;
host->dma_conf.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
host->dma_conf.dst_maxburst = 1;
host->dma_conf.device_fc = false;
return true;
}
}
/*
* HSMCI (High Speed MCI) module is not fully compatible with MCI module.
* HSMCI provides DMA support and a new config register but no more supports
* PDC.
*/
static void __init atmci_get_cap(struct atmel_mci *host)
{
unsigned int version;
version = atmci_get_version(host);
dev_info(&host->pdev->dev,
"version: 0x%x\n", version);
host->caps.has_dma_conf_reg = 0;
host->caps.has_pdc = ATMCI_PDC_CONNECTED;
host->caps.has_cfg_reg = 0;
host->caps.has_cstor_reg = 0;
host->caps.has_highspeed = 0;
host->caps.has_rwproof = 0;
host->caps.has_odd_clk_div = 0;
host->caps.has_bad_data_ordering = 1;
host->caps.need_reset_after_xfer = 1;
host->caps.need_blksz_mul_4 = 1;
host->caps.need_notbusy_for_read_ops = 0;
/* keep only major version number */
switch (version & 0xf00) {
case 0x500:
host->caps.has_odd_clk_div = 1;
case 0x400:
case 0x300:
host->caps.has_dma_conf_reg = 1;
host->caps.has_pdc = 0;
host->caps.has_cfg_reg = 1;
host->caps.has_cstor_reg = 1;
host->caps.has_highspeed = 1;
case 0x200:
host->caps.has_rwproof = 1;
host->caps.need_blksz_mul_4 = 0;
host->caps.need_notbusy_for_read_ops = 1;
case 0x100:
host->caps.has_bad_data_ordering = 0;
host->caps.need_reset_after_xfer = 0;
case 0x0:
break;
default:
host->caps.has_pdc = 0;
dev_warn(&host->pdev->dev,
"Unmanaged mci version, set minimum capabilities\n");
break;
}
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static int __init atmci_probe(struct platform_device *pdev)
{
struct mci_platform_data *pdata;
struct atmel_mci *host;
struct resource *regs;
unsigned int nr_slots;
int irq;
int ret;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
if (!regs)
return -ENXIO;
pdata = pdev->dev.platform_data;
if (!pdata) {
pdata = atmci_of_init(pdev);
if (IS_ERR(pdata)) {
dev_err(&pdev->dev, "platform data not available\n");
return PTR_ERR(pdata);
}
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
irq = platform_get_irq(pdev, 0);
if (irq < 0)
return irq;
host = kzalloc(sizeof(struct atmel_mci), GFP_KERNEL);
if (!host)
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return -ENOMEM;
host->pdev = pdev;
spin_lock_init(&host->lock);
INIT_LIST_HEAD(&host->queue);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->mck = clk_get(&pdev->dev, "mci_clk");
if (IS_ERR(host->mck)) {
ret = PTR_ERR(host->mck);
goto err_clk_get;
}
ret = -ENOMEM;
host->regs = ioremap(regs->start, resource_size(regs));
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (!host->regs)
goto err_ioremap;
clk_enable(host->mck);
atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
host->bus_hz = clk_get_rate(host->mck);
clk_disable(host->mck);
host->mapbase = regs->start;
tasklet_init(&host->tasklet, atmci_tasklet_func, (unsigned long)host);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
ret = request_irq(irq, atmci_interrupt, 0, dev_name(&pdev->dev), host);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (ret)
goto err_request_irq;
/* Get MCI capabilities and set operations according to it */
atmci_get_cap(host);
if (atmci_configure_dma(host)) {
host->prepare_data = &atmci_prepare_data_dma;
host->submit_data = &atmci_submit_data_dma;
host->stop_transfer = &atmci_stop_transfer_dma;
} else if (host->caps.has_pdc) {
dev_info(&pdev->dev, "using PDC\n");
host->prepare_data = &atmci_prepare_data_pdc;
host->submit_data = &atmci_submit_data_pdc;
host->stop_transfer = &atmci_stop_transfer_pdc;
} else {
dev_info(&pdev->dev, "using PIO\n");
host->prepare_data = &atmci_prepare_data;
host->submit_data = &atmci_submit_data;
host->stop_transfer = &atmci_stop_transfer;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
platform_set_drvdata(pdev, host);
setup_timer(&host->timer, atmci_timeout_timer, (unsigned long)host);
/* We need at least one slot to succeed */
nr_slots = 0;
ret = -ENODEV;
if (pdata->slot[0].bus_width) {
ret = atmci_init_slot(host, &pdata->slot[0],
0, ATMCI_SDCSEL_SLOT_A, ATMCI_SDIOIRQA);
if (!ret) {
nr_slots++;
host->buf_size = host->slot[0]->mmc->max_req_size;
}
}
if (pdata->slot[1].bus_width) {
ret = atmci_init_slot(host, &pdata->slot[1],
1, ATMCI_SDCSEL_SLOT_B, ATMCI_SDIOIRQB);
if (!ret) {
nr_slots++;
if (host->slot[1]->mmc->max_req_size > host->buf_size)
host->buf_size =
host->slot[1]->mmc->max_req_size;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
}
if (!nr_slots) {
dev_err(&pdev->dev, "init failed: no slot defined\n");
goto err_init_slot;
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (!host->caps.has_rwproof) {
host->buffer = dma_alloc_coherent(&pdev->dev, host->buf_size,
&host->buf_phys_addr,
GFP_KERNEL);
if (!host->buffer) {
ret = -ENOMEM;
dev_err(&pdev->dev, "buffer allocation failed\n");
goto err_init_slot;
}
}
dev_info(&pdev->dev,
"Atmel MCI controller at 0x%08lx irq %d, %u slots\n",
host->mapbase, irq, nr_slots);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return 0;
err_init_slot:
if (host->dma.chan)
dma_release_channel(host->dma.chan);
free_irq(irq, host);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
err_request_irq:
iounmap(host->regs);
err_ioremap:
clk_put(host->mck);
err_clk_get:
kfree(host);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return ret;
}
static int __exit atmci_remove(struct platform_device *pdev)
{
struct atmel_mci *host = platform_get_drvdata(pdev);
unsigned int i;
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
platform_set_drvdata(pdev, NULL);
if (host->buffer)
dma_free_coherent(&pdev->dev, host->buf_size,
host->buffer, host->buf_phys_addr);
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
if (host->slot[i])
atmci_cleanup_slot(host->slot[i], i);
}
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
clk_enable(host->mck);
atmci_writel(host, ATMCI_IDR, ~0UL);
atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIDIS);
atmci_readl(host, ATMCI_SR);
clk_disable(host->mck);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
if (host->dma.chan)
dma_release_channel(host->dma.chan);
free_irq(platform_get_irq(pdev, 0), host);
iounmap(host->regs);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
clk_put(host->mck);
kfree(host);
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
return 0;
}
#ifdef CONFIG_PM
static int atmci_suspend(struct device *dev)
{
struct atmel_mci *host = dev_get_drvdata(dev);
int i;
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
struct atmel_mci_slot *slot = host->slot[i];
int ret;
if (!slot)
continue;
ret = mmc_suspend_host(slot->mmc);
if (ret < 0) {
while (--i >= 0) {
slot = host->slot[i];
if (slot
&& test_bit(ATMCI_SUSPENDED, &slot->flags)) {
mmc_resume_host(host->slot[i]->mmc);
clear_bit(ATMCI_SUSPENDED, &slot->flags);
}
}
return ret;
} else {
set_bit(ATMCI_SUSPENDED, &slot->flags);
}
}
return 0;
}
static int atmci_resume(struct device *dev)
{
struct atmel_mci *host = dev_get_drvdata(dev);
int i;
int ret = 0;
for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
struct atmel_mci_slot *slot = host->slot[i];
int err;
slot = host->slot[i];
if (!slot)
continue;
if (!test_bit(ATMCI_SUSPENDED, &slot->flags))
continue;
err = mmc_resume_host(slot->mmc);
if (err < 0)
ret = err;
else
clear_bit(ATMCI_SUSPENDED, &slot->flags);
}
return ret;
}
static SIMPLE_DEV_PM_OPS(atmci_pm, atmci_suspend, atmci_resume);
#define ATMCI_PM_OPS (&atmci_pm)
#else
#define ATMCI_PM_OPS NULL
#endif
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
static struct platform_driver atmci_driver = {
.remove = __exit_p(atmci_remove),
.driver = {
.name = "atmel_mci",
.pm = ATMCI_PM_OPS,
.of_match_table = of_match_ptr(atmci_dt_ids),
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
},
};
static int __init atmci_init(void)
{
return platform_driver_probe(&atmci_driver, atmci_probe);
}
static void __exit atmci_exit(void)
{
platform_driver_unregister(&atmci_driver);
}
late_initcall(atmci_init); /* try to load after dma driver when built-in */
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
module_exit(atmci_exit);
MODULE_DESCRIPTION("Atmel Multimedia Card Interface driver");
MODULE_AUTHOR("Haavard Skinnemoen (Atmel)");
atmel-mci: Driver for Atmel on-chip MMC controllers This is a driver for the MMC controller on the AP7000 chips from Atmel. It should in theory work on AT91 systems too with some tweaking, but since the DMA interface is quite different, it's not entirely clear if it's worth merging this with the at91_mci driver. This driver has been around for a while in BSPs and kernel sources provided by Atmel, but this particular version uses the generic DMA Engine framework (with the slave extensions) instead of an avr32-only DMA controller framework. This driver can also use PIO transfers when no DMA channels are available, and for transfers where using DMA may be difficult or impractical for some reason (e.g. the DMA setup overhead is usually not worth it for very short transfers, and badly aligned buffers or lengths are difficult to handle.) Currently, the driver only support PIO transfers. DMA support has been split out to a separate patch to hopefully make it easier to review. The driver has been tested using mmc-block and ext3fs on several SD, SDHC and MMC+ cards. Reads and writes work fine, with read transfer rates up to 3.5 MiB/s on fast cards with debugging disabled. The driver has also been tested using the mmc_test module on the same cards. All tests except 7, 9, 15 and 17 succeed. The first two are unsupported by all the cards I have, so I don't know if the driver handles this correctly. The last two fail because the hardware flags a Data CRC Error instead of a Data Timeout error. I'm not sure how to deal with that. Documentation for this controller can be found in many data sheets from Atmel, including the AT32AP7000 data sheet which can be found here: http://www.atmel.com/dyn/products/datasheets.asp?family_id=682 Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com> Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
2008-06-30 16:35:03 +00:00
MODULE_LICENSE("GPL v2");