When the v2 hw is attached with many disks through an expander, there
may be OOB reset resulting in a PHY going down after the speed is
negotiated (very low probability).
This issue is resolved by modifying the link control registers to send
three identify frames before the PHY is ready (according to 6.10.3.3.2
in SAS 3.0 spec) and close ready when the PHY is down.
Signed-off-by: NengLong Zhao <zhaonenglong@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In setup_itct_v2_hw(), SATA device type SAS_SATA_PENDING is missing, so
add it.
Note: The HiSi SAS controller does not support SATA PM, so do not handle
SAS_SATA_PM_PORT or SAS_SATA_PM.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Function config_id_frame_v1_hw() is called twice for each PHY during
initialisation, which is unneeded.
So remove init_id_frame_v1_hw(), which only calls
config_id_frame_v1_hw().
We will keep the call to config_id_frame_v1_hw() in start_phy_v1_hw()
since it will be used for PHY reset functions.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Function config_id_frame_v2_hw() is called twice for each PHY during
initialisation, which is unneeded.
So remove init_id_frame_v2_hw(), which only calls
config_id_frame_v2_hw().
We will keep the call to config_id_frame_v2_hw() in start_phy_v2_hw()
since it will be used for PHY reset functions.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The endianness for the SAS address in the TX_ID_DWORD registers is set
incorrectly. We see errors like this in the boot log for v2 hw (which
would have the same issue as v1 hw):
[ 7.583284] sas: target proto 0x0 at 50000d1108e7923f:0x1f not handled
This is due to the host SAS addr not matching the PHY SAS addr in the
expander host-attached phy discovery responses.
To fix, we byte swap the SAS addr from BE to LE (which is the endianness
of the SAS controller).
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The endianness for the SAS address in the TX_ID_DWORD registers is set
incorrectly. We see errors like this in the boot log:
[ 7.583284] sas: target proto 0x0 at 50000d1108e7923f:0x1f not handled
This is due to the host SAS addr not matching the PHY SAS addr in the
expander host-attached phy discovery responses.
To fix, we byte swap the SAS addr from BE to LE (which is the endianness
of the SAS controller).
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The device DMA mask was being set after the bulk of the DMA allocations
in the driver init, so potentially DMA allocates fail. To resolve,
relocate before allocating the DMA memory when initialising the driver.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If hisi_sas_task_prep() fails for a SATA device due to PHY down, we
return a failure to libata and also call task_done(), which will cause
ata_qc_complete() to be called twice: - first call from
hisi_sas_task_prep(), which will clear flag ATA_QCFLAG_ACTIVE -
ata_qc_complete() called from libata The warning call trace is as
follows:
[ 117.070206] [<ffff0000084f59b0>] __ata_qc_complete+0xf4/0x11c
[ 117.070208] [<ffff0000084f5b58>] ata_qc_complete+0x180/0x200
[ 117.070210] [<ffff0000084f5dd0>] ata_qc_issue+0x110/0x354
[ 117.070212] [<ffff0000084f6254>] ata_exec_internal_sg+0x240/0x4d0
[ 117.070214] [<ffff0000084f6544>] ata_exec_internal+0x60/0xa0
[ 117.070217] [<ffff000008501580>] ata_read_log_page+0x188/0x1b4
[ 117.070218] [<ffff0000085017dc>] ata_eh_analyze_ncq_error+0xa8/0x274
[ 117.070220] [<ffff000008501a3c>] ata_eh_link_autopsy+0x94/0x8c8
[ 117.070222] [<ffff0000085022a4>] ata_eh_autopsy+0x34/0xe8
[ 117.070223] [<ffff00000850540c>] ata_do_eh+0x28/0xc0
[ 117.070225] [<ffff0000085054e0>] ata_std_error_handler+0x3c/0x84
[ 117.070227] [<ffff000008505140>] ata_scsi_port_error_handler+0x480/0x674
[ 117.070230] [<ffff0000084e3020>] async_sas_ata_eh+0x44/0x78
[ 117.070231] [<ffff0000080d6b8c>] async_run_entry_fn+0x40/0x104
[ 117.070234] [<ffff0000080ce518>] process_one_work+0x128/0x2f0
[ 117.070235] [<ffff0000080ce738>] worker_thread+0x58/0x434
[ 117.070237] [<ffff0000080d416c>] kthread+0xd4/0xe8
[ 117.070240] [<ffff000008084e10>] ret_from_fork+0x10/0x40
The issue is resolved by simply returning a failure status code to the
upper layer.
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In function phy_up_v2_hw(), we needlessly recalculate the phy linkrate
for all phys, and the calculation is incorrect for phy8, so remove this
code.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The Delivery queue enable register should only be written to once at
reset for v2 hw.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The memory calculation for the tags bitmap should use BITS_PER_BYTE
macro instead of coincidental same value of sizeof(unsigned long).
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently the slot memory is zeroed when it is freed and also when it is
reused, like in hisi_sas_task_prep(). Optimise by avoiding the redundant
zeroing in the free.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
None of the CHL_INT2 interrupts are serviced in the channel irq ISR, so
leave the interrupt source masked. The interrupt mask is initially set
in init_reg_v2_hw().
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Optimise by saving an avoidable read in the get_free_slot function. The
delivery queue write pointer will only be updated by software, so don't
bother re-reading what was already written in the previous call to
start_delivery function.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Optimise by saving an avoidable read in the cq interrupt. The queue
read pointer will only be updated by software, so don't bother
re-reading what was already written in the previous interrupt.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When a tmf is issued, various response codes can be returned from the
target. For a query tmf the response may be TMF_RESP_FUNC_COMPLETE or
TMF_RESP_FUNC_SUCC. Add a condition for TMF_RESP_FUNC_SUCC to
hisi_sas_exec_internal_tmf_task(). This affects query tmf, as the
result is success the returned value was for failure.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When the port is detached we cannot execute a TMF, as there can be no
device attached to the port.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add code in slot_complete_v2_hw() to deal with the slots which have
completed due to internal abort.
The status codes have the following meaning:
- STAT_IO_ABORTED: the IO has been aborted due to internal abort,
whether by device or individual abort command
- STAT_IO_COMPLETE: internal abort command has completed successfully
for device or individual abort command
- STAT_IO_NO_DEVICE: internal abort command has completed for device but
cannot find any IO
- STAT_IO_NOT_VALID: internal abort command has completed for single
command but could not find the command
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add function to prepare the an internal abort for v2 hw.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Execute an internal abort for executing a task abort.
This is for case of the command still being present
in host when abort is executed.
For a SATA internal abort, we set abort for all tasks
associated with the device.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Execute an internal abort for that device when it is removed, so that
commands for that device are not processed.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add main code for internal abort functionality.
The internal abort features allows the host controller to abort commands
which are still active in the controller but have not yet been sent to
the slave device.
Typically a command only spends a relatively short time in the
controller when compared to the amount of the time after it is sent to
the slave device.
Two modes of internal abort are supported:
- device
- individual command
For device, when the internal abort is issued all commands in the host
for that device are aborted. For a single command, only that command is
aborted if it is still in the host.
In HW the internal abort command is executed similar to any other sort
of command, like SSP.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add support in v2 hw driver for ACPI.
A check on whether an ACPI handle is available for the device is used to
decide on whether to use ACPI reset handler or syscon for hw reset.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Wei Xu <xuwei5@hisilicon.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Define the NCQ NON DATA command and update libsas to handle it
correctly.
Signed-off-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Add custom version of function to allocate device,
alloc_dev_quirk_v2_hw(). For sata devices the device id bit0 should be
0.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinicke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add v2 hw custom function slot_index_alloc_quirk_v2_hw(). SAS devices
should have IPTT bit0 equal to 1.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinicke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add methods to use HW specific versions of functions to allocate slot
and device. HW specific methods are permitted to workaround device id
vs IPTT collision issue in v2 hw.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinicke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch fixes the ITCT table setup as it should be configured
differently for SAS and SATA devices. For SATA disks there is no need
to set qw2 (already zeroed). Also, link parameters for Bus inactive
limit, max connection time limit, and reject to open limit timers
parameters are changed to match global config register,
MAX_CON_TIME_LIMIT_TIME, as recommended by hw team.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch adds support for directly attaching SATA disks to phy
4-8. The problem was that only registers concerned with phy 0-3 were
being considered in sata_int_v2_hw(). The issue was not detected
previously as the development board only exposed phy 0-3; the new board
provides access to 8 phys.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Intermittently it is found that when multiple SATA disks are directly
connected to the host that some disks are not detected. The problem is
that all set bitfields in ENT_INT_SRC1 are cleared for all phys in
sata_int_v2_hw() - it should clear the set bit for the phy being
serviced.
Also unnecessary double-write to ENT_INT_SRC1 and ENT_INT_SRC_MSK1 is
removed (remaining writes are done at end label).
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Contrary to the field name, the MCR (max connection rate) in the ITCT
should hold the device linkrate (linkrate of the connected phy), and not
the max linkrate.
This fixes an issue seen where some SATA drives connected through an
expander which would not attach.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Rename SCSI_MAX_SG_SEGMENTS to SG_CHUNK_SIZE, which means the amount
we fit into a single scatterlist chunk.
Rename SCSI_MAX_SG_CHAIN_SEGMENTS to SG_MAX_SEGMENTS.
Will move these 2 generic definitions to scatterlist.h later.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Bart Van Assche <bart.vanassche@sandisk.com> (for ib_srp changes)
Signed-off-by: Ming Lin <ming.l@ssi.samsung.com>
Acked-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In high-datarate aging tests, it is found that the
SCSI framework can periodically issue lu resets as
some commands timeout.
Response TASK SET FULL and SAS_QUEUE_FULL may be
returned many times for the same command, causing the
timeouts.
The SAS_QUEUE_FULL errors come from
TRANS_TX_CREDIT_TIMEOUT_ERR, TRANS_TX_CLOSE_NORMAL_ERR,
and TRANS_TX_ERR_FRAME_TXED errors. They do not mean
that the queue is full in the host, but rather it is
equivalent to meaning the queue is full for the sdev.
To overcome this, the queue depth for the sdev is
reduced to 64 (from 256, set in sas_slave_configure()).
Normally error code SAS_QUEUE_FULL will result in the
sdev queue depth falling, but it falls too slowly during
high-datarate tests and commands timeout before it
has fallen to an adequete level from original value.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When TRANS_TX_ERR_FRAME_TXED error occurs in
a slot, the command should be re-attempted.
This error is equivalent to meaning that the queue
is full in the sdev (and not the host).
A superflous debug statement is also removed in the
slot complete handler.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When TRANS_TX_CREDIT_TIMEOUT_ERR or
TRANS_TX_CLOSE_NORMAL_ERR error occur in
a slot, the command should be re-attempted.
This error is equivalent to meaning that the queue
is full in the sdev (and not the host).
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add a function to abort a slot (task) in the target
device and then cleanup and complete the task.
The function is called from work queue context as
it cannot be called from the context where it is
triggered (interrupt).
Flag hisi_sas_slot.abort is added as the flag used
in the slot error handler to indicate whether the
slot needs to be aborted in the sdev prior to
cleanup and finish.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In hisi_sas_exec_internal_tmf_task(), the check for
SAM_STAT_GOOD is replaced with
TMF_RESP_FUNC_COMPLETE, which is a genuine tmf
response code.
SAM_STAT_GOOD and TMF_RESP_FUNC_COMPLETE have the
same value, so this is why it worked before.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add support in v1 hw driver for ACPI.
A check on whether an ACPI handle is available for the device is used to
decide on whether to use ACPI reset handler or syscon for hw reset.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The hisi_sas driver is required to support both device tree and
ACPI. The scanning of the device properties now uses the Unified Device
Properties API, which serves both OF and ACPI.
Since syscon is not supported by ACPI, syscon is only used in the driver
when device tree is used.
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
