Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-12-19 09:32:32 +00:00
Code Issues Packages Projects Releases Wiki Activity
2678958e12
linux/drivers/serial/pmac_zilog.c

2065 lines
51 KiB
C
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/*
* linux/drivers/serial/pmac_zilog.c
*
* Driver for PowerMac Z85c30 based ESCC cell found in the
* "macio" ASICs of various PowerMac models
*
* Copyright (C) 2003 Ben. Herrenschmidt (benh@kernel.crashing.org)
*
* Derived from drivers/macintosh/macserial.c by Paul Mackerras
* and drivers/serial/sunzilog.c by David S. Miller
*
* Hrm... actually, I ripped most of sunzilog (Thanks David !) and
* adapted special tweaks needed for us. I don't think it's worth
* merging back those though. The DMA code still has to get in
* and once done, I expect that driver to remain fairly stable in
* the long term, unless we change the driver model again...
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* 2004-08-06 Harald Welte <laforge@gnumonks.org>
* - Enable BREAK interrupt
* - Add support for sysreq
*
* TODO: - Add DMA support
* - Defer port shutdown to a few seconds after close
* - maybe put something right into uap->clk_divisor
*/
#undef DEBUG
#undef DEBUG_HARD
#undef USE_CTRL_O_SYSRQ
#include <linux/module.h>
#include <linux/tty.h>
#include <linux/tty_flip.h>
#include <linux/major.h>
#include <linux/string.h>
#include <linux/fcntl.h>
#include <linux/mm.h>
#include <linux/kernel.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/console.h>
#include <linux/slab.h>
#include <linux/adb.h>
#include <linux/pmu.h>
#include <linux/bitops.h>
#include <linux/sysrq.h>
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
#include <linux/mutex.h>
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
#include <asm/sections.h>
#include <asm/io.h>
#include <asm/irq.h>
#include <asm/prom.h>
#include <asm/machdep.h>
#include <asm/pmac_feature.h>
#include <asm/dbdma.h>
#include <asm/macio.h>
#if defined (CONFIG_SERIAL_PMACZILOG_CONSOLE) && defined(CONFIG_MAGIC_SYSRQ)
#define SUPPORT_SYSRQ
#endif
#include <linux/serial.h>
#include <linux/serial_core.h>
#include "pmac_zilog.h"
/* Not yet implemented */
#undef HAS_DBDMA
static char version[] __initdata = "pmac_zilog: 0.6 (Benjamin Herrenschmidt <benh@kernel.crashing.org>)";
MODULE_AUTHOR("Benjamin Herrenschmidt <benh@kernel.crashing.org>");
MODULE_DESCRIPTION("Driver for the PowerMac serial ports.");
MODULE_LICENSE("GPL");
#define PWRDBG(fmt, arg...) printk(KERN_DEBUG fmt , ## arg)
[POWERPC] Optionally use new device number for pmac_zilog This adds the option for the pmac_zilog driver to use the major/minor numbers recently allocated specifically for it (/dev/ttyPZn) instead of the /dev/ttySn numbers. The advantage of doing this is that it allows the pmac_zilog and 8250 drivers to coexist. The disadvantage of doing this is that it is a user-visible ABI change and it will break existing working setups on powermacs, and could be confusing to users. Signed-off-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-04-04 14:19:43 +00:00
#ifdef CONFIG_SERIAL_PMACZILOG_TTYS
#define PMACZILOG_MAJOR TTY_MAJOR
#define PMACZILOG_MINOR 64
#define PMACZILOG_NAME "ttyS"
#else
#define PMACZILOG_MAJOR 204
#define PMACZILOG_MINOR 192
#define PMACZILOG_NAME "ttyPZ"
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/*
* For the sake of early serial console, we can do a pre-probe
* (optional) of the ports at rather early boot time.
*/
static struct uart_pmac_port pmz_ports[MAX_ZS_PORTS];
static int pmz_ports_count;
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
static DEFINE_MUTEX(pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
static struct uart_driver pmz_uart_reg = {
.owner = THIS_MODULE,
[POWERPC] Optionally use new device number for pmac_zilog This adds the option for the pmac_zilog driver to use the major/minor numbers recently allocated specifically for it (/dev/ttyPZn) instead of the /dev/ttySn numbers. The advantage of doing this is that it allows the pmac_zilog and 8250 drivers to coexist. The disadvantage of doing this is that it is a user-visible ABI change and it will break existing working setups on powermacs, and could be confusing to users. Signed-off-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-04-04 14:19:43 +00:00
.driver_name = PMACZILOG_NAME,
.dev_name = PMACZILOG_NAME,
.major = PMACZILOG_MAJOR,
.minor = PMACZILOG_MINOR,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
};
/*
* Load all registers to reprogram the port
* This function must only be called when the TX is not busy. The UART
* port lock must be held and local interrupts disabled.
*/
static void pmz_load_zsregs(struct uart_pmac_port *uap, u8 *regs)
{
int i;
if (ZS_IS_ASLEEP(uap))
return;
/* Let pending transmits finish. */
for (i = 0; i < 1000; i++) {
unsigned char stat = read_zsreg(uap, R1);
if (stat & ALL_SNT)
break;
udelay(100);
}
ZS_CLEARERR(uap);
zssync(uap);
ZS_CLEARFIFO(uap);
zssync(uap);
ZS_CLEARERR(uap);
/* Disable all interrupts. */
write_zsreg(uap, R1,
regs[R1] & ~(RxINT_MASK | TxINT_ENAB | EXT_INT_ENAB));
/* Set parity, sync config, stop bits, and clock divisor. */
write_zsreg(uap, R4, regs[R4]);
/* Set misc. TX/RX control bits. */
write_zsreg(uap, R10, regs[R10]);
/* Set TX/RX controls sans the enable bits. */
write_zsreg(uap, R3, regs[R3] & ~RxENABLE);
write_zsreg(uap, R5, regs[R5] & ~TxENABLE);
/* now set R7 "prime" on ESCC */
write_zsreg(uap, R15, regs[R15] | EN85C30);
write_zsreg(uap, R7, regs[R7P]);
/* make sure we use R7 "non-prime" on ESCC */
write_zsreg(uap, R15, regs[R15] & ~EN85C30);
/* Synchronous mode config. */
write_zsreg(uap, R6, regs[R6]);
write_zsreg(uap, R7, regs[R7]);
/* Disable baud generator. */
write_zsreg(uap, R14, regs[R14] & ~BRENAB);
/* Clock mode control. */
write_zsreg(uap, R11, regs[R11]);
/* Lower and upper byte of baud rate generator divisor. */
write_zsreg(uap, R12, regs[R12]);
write_zsreg(uap, R13, regs[R13]);
/* Now rewrite R14, with BRENAB (if set). */
write_zsreg(uap, R14, regs[R14]);
/* Reset external status interrupts. */
write_zsreg(uap, R0, RES_EXT_INT);
write_zsreg(uap, R0, RES_EXT_INT);
/* Rewrite R3/R5, this time without enables masked. */
write_zsreg(uap, R3, regs[R3]);
write_zsreg(uap, R5, regs[R5]);
/* Rewrite R1, this time without IRQ enabled masked. */
write_zsreg(uap, R1, regs[R1]);
/* Enable interrupts */
write_zsreg(uap, R9, regs[R9]);
}
/*
* We do like sunzilog to avoid disrupting pending Tx
* Reprogram the Zilog channel HW registers with the copies found in the
* software state struct. If the transmitter is busy, we defer this update
* until the next TX complete interrupt. Else, we do it right now.
*
* The UART port lock must be held and local interrupts disabled.
*/
static void pmz_maybe_update_regs(struct uart_pmac_port *uap)
{
if (!ZS_REGS_HELD(uap)) {
if (ZS_TX_ACTIVE(uap)) {
uap->flags |= PMACZILOG_FLAG_REGS_HELD;
} else {
pmz_debug("pmz: maybe_update_regs: updating\n");
pmz_load_zsregs(uap, uap->curregs);
}
}
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static struct tty_struct *pmz_receive_chars(struct uart_pmac_port *uap)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct tty_struct *tty = NULL;
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
unsigned char ch, r1, drop, error, flag;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
int loops = 0;
/* The interrupt can be enabled when the port isn't open, typically
* that happens when using one port is open and the other closed (stale
* interrupt) or when one port is used as a console.
*/
if (!ZS_IS_OPEN(uap)) {
pmz_debug("pmz: draining input\n");
/* Port is closed, drain input data */
for (;;) {
if ((++loops) > 1000)
goto flood;
(void)read_zsreg(uap, R1);
write_zsreg(uap, R0, ERR_RES);
(void)read_zsdata(uap);
ch = read_zsreg(uap, R0);
if (!(ch & Rx_CH_AV))
break;
}
return NULL;
}
/* Sanity check, make sure the old bug is no longer happening */
ttydev: fix pamc_zilog for tty pointer move Today's linux-next build (powerpc allyesconfig) failed like this: drivers/serial/pmac_zilog.c: In function 'pmz_receive_chars': drivers/serial/pmac_zilog.c:245: error: 'struct uart_info' has no member named 'tty' drivers/serial/pmac_zilog.c:250: error: 'struct uart_info' has no member named 'tty' I applied the patch below (which builds but may, or may not, be correct). -- Cheers, Stephen Rothwell sfr@canb.auug.org.au http://www.canb.auug.org.au/~sfr/ Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-16 20:54:22 +00:00
if (uap->port.info == NULL || uap->port.info->port.tty == NULL) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
WARN_ON(1);
(void)read_zsdata(uap);
return NULL;
}
ttydev: fix pamc_zilog for tty pointer move Today's linux-next build (powerpc allyesconfig) failed like this: drivers/serial/pmac_zilog.c: In function 'pmz_receive_chars': drivers/serial/pmac_zilog.c:245: error: 'struct uart_info' has no member named 'tty' drivers/serial/pmac_zilog.c:250: error: 'struct uart_info' has no member named 'tty' I applied the patch below (which builds but may, or may not, be correct). -- Cheers, Stephen Rothwell sfr@canb.auug.org.au http://www.canb.auug.org.au/~sfr/ Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-16 20:54:22 +00:00
tty = uap->port.info->port.tty;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
while (1) {
error = 0;
drop = 0;
r1 = read_zsreg(uap, R1);
ch = read_zsdata(uap);
if (r1 & (PAR_ERR | Rx_OVR | CRC_ERR)) {
write_zsreg(uap, R0, ERR_RES);
zssync(uap);
}
ch &= uap->parity_mask;
if (ch == 0 && uap->flags & PMACZILOG_FLAG_BREAK) {
uap->flags &= ~PMACZILOG_FLAG_BREAK;
}
#if defined(CONFIG_MAGIC_SYSRQ) && defined(CONFIG_SERIAL_CORE_CONSOLE)
#ifdef USE_CTRL_O_SYSRQ
/* Handle the SysRq ^O Hack */
if (ch == '\x0f') {
uap->port.sysrq = jiffies + HZ*5;
goto next_char;
}
#endif /* USE_CTRL_O_SYSRQ */
if (uap->port.sysrq) {
int swallow;
spin_unlock(&uap->port.lock);
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
swallow = uart_handle_sysrq_char(&uap->port, ch);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
spin_lock(&uap->port.lock);
if (swallow)
goto next_char;
}
#endif /* CONFIG_MAGIC_SYSRQ && CONFIG_SERIAL_CORE_CONSOLE */
/* A real serial line, record the character and status. */
if (drop)
goto next_char;
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_NORMAL;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->port.icount.rx++;
if (r1 & (PAR_ERR | Rx_OVR | CRC_ERR | BRK_ABRT)) {
error = 1;
if (r1 & BRK_ABRT) {
pmz_debug("pmz: got break !\n");
r1 &= ~(PAR_ERR | CRC_ERR);
uap->port.icount.brk++;
if (uart_handle_break(&uap->port))
goto next_char;
}
else if (r1 & PAR_ERR)
uap->port.icount.parity++;
else if (r1 & CRC_ERR)
uap->port.icount.frame++;
if (r1 & Rx_OVR)
uap->port.icount.overrun++;
r1 &= uap->port.read_status_mask;
if (r1 & BRK_ABRT)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_BREAK;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
else if (r1 & PAR_ERR)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_PARITY;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
else if (r1 & CRC_ERR)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_FRAME;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
}
if (uap->port.ignore_status_mask == 0xff ||
(r1 & uap->port.ignore_status_mask) == 0) {
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
tty_insert_flip_char(tty, ch, flag);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
}
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
if (r1 & Rx_OVR)
tty_insert_flip_char(tty, 0, TTY_OVERRUN);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
next_char:
/* We can get stuck in an infinite loop getting char 0 when the
* line is in a wrong HW state, we break that here.
* When that happens, I disable the receive side of the driver.
* Note that what I've been experiencing is a real irq loop where
* I'm getting flooded regardless of the actual port speed.
* Something stange is going on with the HW
*/
if ((++loops) > 1000)
goto flood;
ch = read_zsreg(uap, R0);
if (!(ch & Rx_CH_AV))
break;
}
return tty;
flood:
uap->curregs[R1] &= ~(EXT_INT_ENAB | TxINT_ENAB | RxINT_MASK);
write_zsreg(uap, R1, uap->curregs[R1]);
zssync(uap);
dev_err(&uap->dev->ofdev.dev, "pmz: rx irq flood !\n");
return tty;
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static void pmz_status_handle(struct uart_pmac_port *uap)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
unsigned char status;
status = read_zsreg(uap, R0);
write_zsreg(uap, R0, RES_EXT_INT);
zssync(uap);
if (ZS_IS_OPEN(uap) && ZS_WANTS_MODEM_STATUS(uap)) {
if (status & SYNC_HUNT)
uap->port.icount.dsr++;
/* The Zilog just gives us an interrupt when DCD/CTS/etc. change.
* But it does not tell us which bit has changed, we have to keep
* track of this ourselves.
* The CTS input is inverted for some reason. -- paulus
*/
if ((status ^ uap->prev_status) & DCD)
uart_handle_dcd_change(&uap->port,
(status & DCD));
if ((status ^ uap->prev_status) & CTS)
uart_handle_cts_change(&uap->port,
!(status & CTS));
wake_up_interruptible(&uap->port.info->delta_msr_wait);
}
if (status & BRK_ABRT)
uap->flags |= PMACZILOG_FLAG_BREAK;
uap->prev_status = status;
}
static void pmz_transmit_chars(struct uart_pmac_port *uap)
{
struct circ_buf *xmit;
if (ZS_IS_ASLEEP(uap))
return;
if (ZS_IS_CONS(uap)) {
unsigned char status = read_zsreg(uap, R0);
/* TX still busy? Just wait for the next TX done interrupt.
*
* It can occur because of how we do serial console writes. It would
* be nice to transmit console writes just like we normally would for
* a TTY line. (ie. buffered and TX interrupt driven). That is not
* easy because console writes cannot sleep. One solution might be
* to poll on enough port->xmit space becomming free. -DaveM
*/
if (!(status & Tx_BUF_EMP))
return;
}
uap->flags &= ~PMACZILOG_FLAG_TX_ACTIVE;
if (ZS_REGS_HELD(uap)) {
pmz_load_zsregs(uap, uap->curregs);
uap->flags &= ~PMACZILOG_FLAG_REGS_HELD;
}
if (ZS_TX_STOPPED(uap)) {
uap->flags &= ~PMACZILOG_FLAG_TX_STOPPED;
goto ack_tx_int;
}
if (uap->port.x_char) {
uap->flags |= PMACZILOG_FLAG_TX_ACTIVE;
write_zsdata(uap, uap->port.x_char);
zssync(uap);
uap->port.icount.tx++;
uap->port.x_char = 0;
return;
}
if (uap->port.info == NULL)
goto ack_tx_int;
xmit = &uap->port.info->xmit;
if (uart_circ_empty(xmit)) {
uart_write_wakeup(&uap->port);
goto ack_tx_int;
}
if (uart_tx_stopped(&uap->port))
goto ack_tx_int;
uap->flags |= PMACZILOG_FLAG_TX_ACTIVE;
write_zsdata(uap, xmit->buf[xmit->tail]);
zssync(uap);
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
uap->port.icount.tx++;
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&uap->port);
return;
ack_tx_int:
write_zsreg(uap, R0, RES_Tx_P);
zssync(uap);
}
/* Hrm... we register that twice, fixme later.... */
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static irqreturn_t pmz_interrupt(int irq, void *dev_id)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct uart_pmac_port *uap = dev_id;
struct uart_pmac_port *uap_a;
struct uart_pmac_port *uap_b;
int rc = IRQ_NONE;
struct tty_struct *tty;
u8 r3;
uap_a = pmz_get_port_A(uap);
uap_b = uap_a->mate;
spin_lock(&uap_a->port.lock);
r3 = read_zsreg(uap_a, R3);
#ifdef DEBUG_HARD
pmz_debug("irq, r3: %x\n", r3);
#endif
/* Channel A */
tty = NULL;
if (r3 & (CHAEXT | CHATxIP | CHARxIP)) {
write_zsreg(uap_a, R0, RES_H_IUS);
zssync(uap_a);
if (r3 & CHAEXT)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
pmz_status_handle(uap_a);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (r3 & CHARxIP)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
tty = pmz_receive_chars(uap_a);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (r3 & CHATxIP)
pmz_transmit_chars(uap_a);
rc = IRQ_HANDLED;
}
spin_unlock(&uap_a->port.lock);
if (tty != NULL)
tty_flip_buffer_push(tty);
if (uap_b->node == NULL)
goto out;
spin_lock(&uap_b->port.lock);
tty = NULL;
if (r3 & (CHBEXT | CHBTxIP | CHBRxIP)) {
write_zsreg(uap_b, R0, RES_H_IUS);
zssync(uap_b);
if (r3 & CHBEXT)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
pmz_status_handle(uap_b);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (r3 & CHBRxIP)
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
tty = pmz_receive_chars(uap_b);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (r3 & CHBTxIP)
pmz_transmit_chars(uap_b);
rc = IRQ_HANDLED;
}
spin_unlock(&uap_b->port.lock);
if (tty != NULL)
tty_flip_buffer_push(tty);
out:
#ifdef DEBUG_HARD
pmz_debug("irq done.\n");
#endif
return rc;
}
/*
* Peek the status register, lock not held by caller
*/
static inline u8 pmz_peek_status(struct uart_pmac_port *uap)
{
unsigned long flags;
u8 status;
spin_lock_irqsave(&uap->port.lock, flags);
status = read_zsreg(uap, R0);
spin_unlock_irqrestore(&uap->port.lock, flags);
return status;
}
/*
* Check if transmitter is empty
* The port lock is not held.
*/
static unsigned int pmz_tx_empty(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char status;
if (ZS_IS_ASLEEP(uap) || uap->node == NULL)
return TIOCSER_TEMT;
status = pmz_peek_status(to_pmz(port));
if (status & Tx_BUF_EMP)
return TIOCSER_TEMT;
return 0;
}
/*
* Set Modem Control (RTS & DTR) bits
* The port lock is held and interrupts are disabled.
* Note: Shall we really filter out RTS on external ports or
* should that be dealt at higher level only ?
*/
static void pmz_set_mctrl(struct uart_port *port, unsigned int mctrl)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char set_bits, clear_bits;
/* Do nothing for irda for now... */
if (ZS_IS_IRDA(uap))
return;
/* We get called during boot with a port not up yet */
if (ZS_IS_ASLEEP(uap) ||
!(ZS_IS_OPEN(uap) || ZS_IS_CONS(uap)))
return;
set_bits = clear_bits = 0;
if (ZS_IS_INTMODEM(uap)) {
if (mctrl & TIOCM_RTS)
set_bits |= RTS;
else
clear_bits |= RTS;
}
if (mctrl & TIOCM_DTR)
set_bits |= DTR;
else
clear_bits |= DTR;
/* NOTE: Not subject to 'transmitter active' rule. */
uap->curregs[R5] |= set_bits;
uap->curregs[R5] &= ~clear_bits;
if (ZS_IS_ASLEEP(uap))
return;
write_zsreg(uap, R5, uap->curregs[R5]);
pmz_debug("pmz_set_mctrl: set bits: %x, clear bits: %x -> %x\n",
set_bits, clear_bits, uap->curregs[R5]);
zssync(uap);
}
/*
* Get Modem Control bits (only the input ones, the core will
* or that with a cached value of the control ones)
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 08:42:38 +00:00
* The port lock is held and interrupts are disabled.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
*/
static unsigned int pmz_get_mctrl(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char status;
unsigned int ret;
if (ZS_IS_ASLEEP(uap) || uap->node == NULL)
return 0;
[PATCH] Serial: Adjust serial locking This patch changes the way serial ports are locked when getting modem status. This change is necessary because we will need to atomically read the modem status and take action depending on the CTS status. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-06-29 08:42:38 +00:00
status = read_zsreg(uap, R0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
ret = 0;
if (status & DCD)
ret |= TIOCM_CAR;
if (status & SYNC_HUNT)
ret |= TIOCM_DSR;
if (!(status & CTS))
ret |= TIOCM_CTS;
return ret;
}
/*
* Stop TX side. Dealt like sunzilog at next Tx interrupt,
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 09:12:14 +00:00
* though for DMA, we will have to do a bit more.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
* The port lock is held and interrupts are disabled.
*/
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 09:12:14 +00:00
static void pmz_stop_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
to_pmz(port)->flags |= PMACZILOG_FLAG_TX_STOPPED;
}
/*
* Kick the Tx side.
* The port lock is held and interrupts are disabled.
*/
[SERIAL] Clean up and fix tty transmission start/stoping The start_tx and stop_tx methods were passed a flag to indicate whether the start/stop was from the tty start/stop callbacks, and some drivers used this flag to decide whether to ask the UART to immediately stop transmission (where the UART supports such a feature.) There are other cases when we wish this to occur - when CTS is lowered, or if we change from soft to hard flow control and CTS is inactive. In these cases, this flag was false, and we would allow the transmitter to drain before stopping. There is really only one case where we want to let the transmitter drain before disabling, and that's when we run out of characters to send. Hence, re-jig the start_tx and stop_tx methods to eliminate this flag, and introduce new functions for the special "disable and allow transmitter to drain" case. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2005-08-31 09:12:14 +00:00
static void pmz_start_tx(struct uart_port *port)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char status;
pmz_debug("pmz: start_tx()\n");
uap->flags |= PMACZILOG_FLAG_TX_ACTIVE;
uap->flags &= ~PMACZILOG_FLAG_TX_STOPPED;
if (ZS_IS_ASLEEP(uap) || uap->node == NULL)
return;
status = read_zsreg(uap, R0);
/* TX busy? Just wait for the TX done interrupt. */
if (!(status & Tx_BUF_EMP))
return;
/* Send the first character to jump-start the TX done
* IRQ sending engine.
*/
if (port->x_char) {
write_zsdata(uap, port->x_char);
zssync(uap);
port->icount.tx++;
port->x_char = 0;
} else {
struct circ_buf *xmit = &port->info->xmit;
write_zsdata(uap, xmit->buf[xmit->tail]);
zssync(uap);
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
port->icount.tx++;
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(&uap->port);
}
pmz_debug("pmz: start_tx() done.\n");
}
/*
* Stop Rx side, basically disable emitting of
* Rx interrupts on the port. We don't disable the rx
* side of the chip proper though
* The port lock is held.
*/
static void pmz_stop_rx(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
if (ZS_IS_ASLEEP(uap) || uap->node == NULL)
return;
pmz_debug("pmz: stop_rx()()\n");
/* Disable all RX interrupts. */
uap->curregs[R1] &= ~RxINT_MASK;
pmz_maybe_update_regs(uap);
pmz_debug("pmz: stop_rx() done.\n");
}
/*
* Enable modem status change interrupts
* The port lock is held.
*/
static void pmz_enable_ms(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char new_reg;
if (ZS_IS_IRDA(uap) || uap->node == NULL)
return;
new_reg = uap->curregs[R15] | (DCDIE | SYNCIE | CTSIE);
if (new_reg != uap->curregs[R15]) {
uap->curregs[R15] = new_reg;
if (ZS_IS_ASLEEP(uap))
return;
/* NOTE: Not subject to 'transmitter active' rule. */
write_zsreg(uap, R15, uap->curregs[R15]);
}
}
/*
* Control break state emission
* The port lock is not held.
*/
static void pmz_break_ctl(struct uart_port *port, int break_state)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned char set_bits, clear_bits, new_reg;
unsigned long flags;
if (uap->node == NULL)
return;
set_bits = clear_bits = 0;
if (break_state)
set_bits |= SND_BRK;
else
clear_bits |= SND_BRK;
spin_lock_irqsave(&port->lock, flags);
new_reg = (uap->curregs[R5] | set_bits) & ~clear_bits;
if (new_reg != uap->curregs[R5]) {
uap->curregs[R5] = new_reg;
/* NOTE: Not subject to 'transmitter active' rule. */
if (ZS_IS_ASLEEP(uap))
return;
write_zsreg(uap, R5, uap->curregs[R5]);
}
spin_unlock_irqrestore(&port->lock, flags);
}
/*
* Turn power on or off to the SCC and associated stuff
* (port drivers, modem, IR port, etc.)
* Returns the number of milliseconds we should wait before
* trying to use the port.
*/
static int pmz_set_scc_power(struct uart_pmac_port *uap, int state)
{
int delay = 0;
int rc;
if (state) {
rc = pmac_call_feature(
PMAC_FTR_SCC_ENABLE, uap->node, uap->port_type, 1);
pmz_debug("port power on result: %d\n", rc);
if (ZS_IS_INTMODEM(uap)) {
rc = pmac_call_feature(
PMAC_FTR_MODEM_ENABLE, uap->node, 0, 1);
delay = 2500; /* wait for 2.5s before using */
pmz_debug("modem power result: %d\n", rc);
}
} else {
/* TODO: Make that depend on a timer, don't power down
* immediately
*/
if (ZS_IS_INTMODEM(uap)) {
rc = pmac_call_feature(
PMAC_FTR_MODEM_ENABLE, uap->node, 0, 0);
pmz_debug("port power off result: %d\n", rc);
}
pmac_call_feature(PMAC_FTR_SCC_ENABLE, uap->node, uap->port_type, 0);
}
return delay;
}
/*
* FixZeroBug....Works around a bug in the SCC receving channel.
* Inspired from Darwin code, 15 Sept. 2000 -DanM
*
* The following sequence prevents a problem that is seen with O'Hare ASICs
* (most versions -- also with some Heathrow and Hydra ASICs) where a zero
* at the input to the receiver becomes 'stuck' and locks up the receiver.
* This problem can occur as a result of a zero bit at the receiver input
* coincident with any of the following events:
*
* The SCC is initialized (hardware or software).
* A framing error is detected.
* The clocking option changes from synchronous or X1 asynchronous
* clocking to X16, X32, or X64 asynchronous clocking.
* The decoding mode is changed among NRZ, NRZI, FM0, or FM1.
*
* This workaround attempts to recover from the lockup condition by placing
* the SCC in synchronous loopback mode with a fast clock before programming
* any of the asynchronous modes.
*/
static void pmz_fix_zero_bug_scc(struct uart_pmac_port *uap)
{
write_zsreg(uap, 9, ZS_IS_CHANNEL_A(uap) ? CHRA : CHRB);
zssync(uap);
udelay(10);
write_zsreg(uap, 9, (ZS_IS_CHANNEL_A(uap) ? CHRA : CHRB) | NV);
zssync(uap);
write_zsreg(uap, 4, X1CLK | MONSYNC);
write_zsreg(uap, 3, Rx8);
write_zsreg(uap, 5, Tx8 | RTS);
write_zsreg(uap, 9, NV); /* Didn't we already do this? */
write_zsreg(uap, 11, RCBR | TCBR);
write_zsreg(uap, 12, 0);
write_zsreg(uap, 13, 0);
write_zsreg(uap, 14, (LOOPBAK | BRSRC));
write_zsreg(uap, 14, (LOOPBAK | BRSRC | BRENAB));
write_zsreg(uap, 3, Rx8 | RxENABLE);
write_zsreg(uap, 0, RES_EXT_INT);
write_zsreg(uap, 0, RES_EXT_INT);
write_zsreg(uap, 0, RES_EXT_INT); /* to kill some time */
/* The channel should be OK now, but it is probably receiving
* loopback garbage.
* Switch to asynchronous mode, disable the receiver,
* and discard everything in the receive buffer.
*/
write_zsreg(uap, 9, NV);
write_zsreg(uap, 4, X16CLK | SB_MASK);
write_zsreg(uap, 3, Rx8);
while (read_zsreg(uap, 0) & Rx_CH_AV) {
(void)read_zsreg(uap, 8);
write_zsreg(uap, 0, RES_EXT_INT);
write_zsreg(uap, 0, ERR_RES);
}
}
/*
* Real startup routine, powers up the hardware and sets up
* the SCC. Returns a delay in ms where you need to wait before
* actually using the port, this is typically the internal modem
* powerup delay. This routine expect the lock to be taken.
*/
static int __pmz_startup(struct uart_pmac_port *uap)
{
int pwr_delay = 0;
memset(&uap->curregs, 0, sizeof(uap->curregs));
/* Power up the SCC & underlying hardware (modem/irda) */
pwr_delay = pmz_set_scc_power(uap, 1);
/* Nice buggy HW ... */
pmz_fix_zero_bug_scc(uap);
/* Reset the channel */
uap->curregs[R9] = 0;
write_zsreg(uap, 9, ZS_IS_CHANNEL_A(uap) ? CHRA : CHRB);
zssync(uap);
udelay(10);
write_zsreg(uap, 9, 0);
zssync(uap);
/* Clear the interrupt registers */
write_zsreg(uap, R1, 0);
write_zsreg(uap, R0, ERR_RES);
write_zsreg(uap, R0, ERR_RES);
write_zsreg(uap, R0, RES_H_IUS);
write_zsreg(uap, R0, RES_H_IUS);
/* Setup some valid baud rate */
uap->curregs[R4] = X16CLK | SB1;
uap->curregs[R3] = Rx8;
uap->curregs[R5] = Tx8 | RTS;
if (!ZS_IS_IRDA(uap))
uap->curregs[R5] |= DTR;
uap->curregs[R12] = 0;
uap->curregs[R13] = 0;
uap->curregs[R14] = BRENAB;
/* Clear handshaking, enable BREAK interrupts */
uap->curregs[R15] = BRKIE;
/* Master interrupt enable */
uap->curregs[R9] |= NV | MIE;
pmz_load_zsregs(uap, uap->curregs);
/* Enable receiver and transmitter. */
write_zsreg(uap, R3, uap->curregs[R3] |= RxENABLE);
write_zsreg(uap, R5, uap->curregs[R5] |= TxENABLE);
/* Remember status for DCD/CTS changes */
uap->prev_status = read_zsreg(uap, R0);
return pwr_delay;
}
static void pmz_irda_reset(struct uart_pmac_port *uap)
{
uap->curregs[R5] |= DTR;
write_zsreg(uap, R5, uap->curregs[R5]);
zssync(uap);
mdelay(110);
uap->curregs[R5] &= ~DTR;
write_zsreg(uap, R5, uap->curregs[R5]);
zssync(uap);
mdelay(10);
}
/*
* This is the "normal" startup routine, using the above one
* wrapped with the lock and doing a schedule delay
*/
static int pmz_startup(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned long flags;
int pwr_delay = 0;
pmz_debug("pmz: startup()\n");
if (ZS_IS_ASLEEP(uap))
return -EAGAIN;
if (uap->node == NULL)
return -ENODEV;
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_lock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->flags |= PMACZILOG_FLAG_IS_OPEN;
/* A console is never powered down. Else, power up and
* initialize the chip
*/
if (!ZS_IS_CONS(uap)) {
spin_lock_irqsave(&port->lock, flags);
pwr_delay = __pmz_startup(uap);
spin_unlock_irqrestore(&port->lock, flags);
}
pmz_get_port_A(uap)->flags |= PMACZILOG_FLAG_IS_IRQ_ON;
[PATCH] irq-flags: serial: Use the new IRQF_ constants Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: "David S. Miller" <davem@davemloft.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-02 02:29:43 +00:00
if (request_irq(uap->port.irq, pmz_interrupt, IRQF_SHARED, "PowerMac Zilog", uap)) {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
dev_err(&uap->dev->ofdev.dev,
"Unable to register zs interrupt handler.\n");
pmz_set_scc_power(uap, 0);
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
return -ENXIO;
}
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/* Right now, we deal with delay by blocking here, I'll be
* smarter later on
*/
if (pwr_delay != 0) {
pmz_debug("pmz: delaying %d ms\n", pwr_delay);
msleep(pwr_delay);
}
/* IrDA reset is done now */
if (ZS_IS_IRDA(uap))
pmz_irda_reset(uap);
/* Enable interrupts emission from the chip */
spin_lock_irqsave(&port->lock, flags);
uap->curregs[R1] |= INT_ALL_Rx | TxINT_ENAB;
if (!ZS_IS_EXTCLK(uap))
uap->curregs[R1] |= EXT_INT_ENAB;
write_zsreg(uap, R1, uap->curregs[R1]);
spin_unlock_irqrestore(&port->lock, flags);
pmz_debug("pmz: startup() done.\n");
return 0;
}
static void pmz_shutdown(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned long flags;
pmz_debug("pmz: shutdown()\n");
if (uap->node == NULL)
return;
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_lock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/* Release interrupt handler */
free_irq(uap->port.irq, uap);
spin_lock_irqsave(&port->lock, flags);
uap->flags &= ~PMACZILOG_FLAG_IS_OPEN;
if (!ZS_IS_OPEN(uap->mate))
pmz_get_port_A(uap)->flags &= ~PMACZILOG_FLAG_IS_IRQ_ON;
/* Disable interrupts */
if (!ZS_IS_ASLEEP(uap)) {
uap->curregs[R1] &= ~(EXT_INT_ENAB | TxINT_ENAB | RxINT_MASK);
write_zsreg(uap, R1, uap->curregs[R1]);
zssync(uap);
}
if (ZS_IS_CONS(uap) || ZS_IS_ASLEEP(uap)) {
spin_unlock_irqrestore(&port->lock, flags);
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
return;
}
/* Disable receiver and transmitter. */
uap->curregs[R3] &= ~RxENABLE;
uap->curregs[R5] &= ~TxENABLE;
/* Disable all interrupts and BRK assertion. */
uap->curregs[R5] &= ~SND_BRK;
pmz_maybe_update_regs(uap);
/* Shut the chip down */
pmz_set_scc_power(uap, 0);
spin_unlock_irqrestore(&port->lock, flags);
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
pmz_debug("pmz: shutdown() done.\n");
}
/* Shared by TTY driver and serial console setup. The port lock is held
* and local interrupts are disabled.
*/
static void pmz_convert_to_zs(struct uart_pmac_port *uap, unsigned int cflag,
unsigned int iflag, unsigned long baud)
{
int brg;
/* Switch to external clocking for IrDA high clock rates. That
* code could be re-used for Midi interfaces with different
* multipliers
*/
if (baud >= 115200 && ZS_IS_IRDA(uap)) {
uap->curregs[R4] = X1CLK;
uap->curregs[R11] = RCTRxCP | TCTRxCP;
uap->curregs[R14] = 0; /* BRG off */
uap->curregs[R12] = 0;
uap->curregs[R13] = 0;
uap->flags |= PMACZILOG_FLAG_IS_EXTCLK;
} else {
switch (baud) {
case ZS_CLOCK/16: /* 230400 */
uap->curregs[R4] = X16CLK;
uap->curregs[R11] = 0;
uap->curregs[R14] = 0;
break;
case ZS_CLOCK/32: /* 115200 */
uap->curregs[R4] = X32CLK;
uap->curregs[R11] = 0;
uap->curregs[R14] = 0;
break;
default:
uap->curregs[R4] = X16CLK;
uap->curregs[R11] = TCBR | RCBR;
brg = BPS_TO_BRG(baud, ZS_CLOCK / 16);
uap->curregs[R12] = (brg & 255);
uap->curregs[R13] = ((brg >> 8) & 255);
uap->curregs[R14] = BRENAB;
}
uap->flags &= ~PMACZILOG_FLAG_IS_EXTCLK;
}
/* Character size, stop bits, and parity. */
uap->curregs[3] &= ~RxN_MASK;
uap->curregs[5] &= ~TxN_MASK;
switch (cflag & CSIZE) {
case CS5:
uap->curregs[3] |= Rx5;
uap->curregs[5] |= Tx5;
uap->parity_mask = 0x1f;
break;
case CS6:
uap->curregs[3] |= Rx6;
uap->curregs[5] |= Tx6;
uap->parity_mask = 0x3f;
break;
case CS7:
uap->curregs[3] |= Rx7;
uap->curregs[5] |= Tx7;
uap->parity_mask = 0x7f;
break;
case CS8:
default:
uap->curregs[3] |= Rx8;
uap->curregs[5] |= Tx8;
uap->parity_mask = 0xff;
break;
};
uap->curregs[4] &= ~(SB_MASK);
if (cflag & CSTOPB)
uap->curregs[4] |= SB2;
else
uap->curregs[4] |= SB1;
if (cflag & PARENB)
uap->curregs[4] |= PAR_ENAB;
else
uap->curregs[4] &= ~PAR_ENAB;
if (!(cflag & PARODD))
uap->curregs[4] |= PAR_EVEN;
else
uap->curregs[4] &= ~PAR_EVEN;
uap->port.read_status_mask = Rx_OVR;
if (iflag & INPCK)
uap->port.read_status_mask |= CRC_ERR | PAR_ERR;
if (iflag & (BRKINT | PARMRK))
uap->port.read_status_mask |= BRK_ABRT;
uap->port.ignore_status_mask = 0;
if (iflag & IGNPAR)
uap->port.ignore_status_mask |= CRC_ERR | PAR_ERR;
if (iflag & IGNBRK) {
uap->port.ignore_status_mask |= BRK_ABRT;
if (iflag & IGNPAR)
uap->port.ignore_status_mask |= Rx_OVR;
}
if ((cflag & CREAD) == 0)
uap->port.ignore_status_mask = 0xff;
}
/*
* Set the irda codec on the imac to the specified baud rate.
*/
static void pmz_irda_setup(struct uart_pmac_port *uap, unsigned long *baud)
{
u8 cmdbyte;
int t, version;
switch (*baud) {
/* SIR modes */
case 2400:
cmdbyte = 0x53;
break;
case 4800:
cmdbyte = 0x52;
break;
case 9600:
cmdbyte = 0x51;
break;
case 19200:
cmdbyte = 0x50;
break;
case 38400:
cmdbyte = 0x4f;
break;
case 57600:
cmdbyte = 0x4e;
break;
case 115200:
cmdbyte = 0x4d;
break;
/* The FIR modes aren't really supported at this point, how
* do we select the speed ? via the FCR on KeyLargo ?
*/
case 1152000:
cmdbyte = 0;
break;
case 4000000:
cmdbyte = 0;
break;
default: /* 9600 */
cmdbyte = 0x51;
*baud = 9600;
break;
}
/* Wait for transmitter to drain */
t = 10000;
while ((read_zsreg(uap, R0) & Tx_BUF_EMP) == 0
|| (read_zsreg(uap, R1) & ALL_SNT) == 0) {
if (--t <= 0) {
dev_err(&uap->dev->ofdev.dev, "transmitter didn't drain\n");
return;
}
udelay(10);
}
/* Drain the receiver too */
t = 100;
(void)read_zsdata(uap);
(void)read_zsdata(uap);
(void)read_zsdata(uap);
mdelay(10);
while (read_zsreg(uap, R0) & Rx_CH_AV) {
read_zsdata(uap);
mdelay(10);
if (--t <= 0) {
dev_err(&uap->dev->ofdev.dev, "receiver didn't drain\n");
return;
}
}
/* Switch to command mode */
uap->curregs[R5] |= DTR;
write_zsreg(uap, R5, uap->curregs[R5]);
zssync(uap);
mdelay(1);
/* Switch SCC to 19200 */
pmz_convert_to_zs(uap, CS8, 0, 19200);
pmz_load_zsregs(uap, uap->curregs);
mdelay(1);
/* Write get_version command byte */
write_zsdata(uap, 1);
t = 5000;
while ((read_zsreg(uap, R0) & Rx_CH_AV) == 0) {
if (--t <= 0) {
dev_err(&uap->dev->ofdev.dev,
"irda_setup timed out on get_version byte\n");
goto out;
}
udelay(10);
}
version = read_zsdata(uap);
if (version < 4) {
dev_info(&uap->dev->ofdev.dev, "IrDA: dongle version %d not supported\n",
version);
goto out;
}
/* Send speed mode */
write_zsdata(uap, cmdbyte);
t = 5000;
while ((read_zsreg(uap, R0) & Rx_CH_AV) == 0) {
if (--t <= 0) {
dev_err(&uap->dev->ofdev.dev,
"irda_setup timed out on speed mode byte\n");
goto out;
}
udelay(10);
}
t = read_zsdata(uap);
if (t != cmdbyte)
dev_err(&uap->dev->ofdev.dev,
"irda_setup speed mode byte = %x (%x)\n", t, cmdbyte);
dev_info(&uap->dev->ofdev.dev, "IrDA setup for %ld bps, dongle version: %d\n",
*baud, version);
(void)read_zsdata(uap);
(void)read_zsdata(uap);
(void)read_zsdata(uap);
out:
/* Switch back to data mode */
uap->curregs[R5] &= ~DTR;
write_zsreg(uap, R5, uap->curregs[R5]);
zssync(uap);
(void)read_zsdata(uap);
(void)read_zsdata(uap);
(void)read_zsdata(uap);
}
[PATCH] tty: switch to ktermios This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Martin Peschke <mp3@de.ibm.com> Acked-by: Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 10:38:45 +00:00
static void __pmz_set_termios(struct uart_port *port, struct ktermios *termios,
struct ktermios *old)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned long baud;
pmz_debug("pmz: set_termios()\n");
if (ZS_IS_ASLEEP(uap))
return;
[PATCH] tty: switch to ktermios This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Martin Peschke <mp3@de.ibm.com> Acked-by: Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 10:38:45 +00:00
memcpy(&uap->termios_cache, termios, sizeof(struct ktermios));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/* XXX Check which revs of machines actually allow 1 and 4Mb speeds
* on the IR dongle. Note that the IRTTY driver currently doesn't know
* about the FIR mode and high speed modes. So these are unused. For
* implementing proper support for these, we should probably add some
* DMA as well, at least on the Rx side, which isn't a simple thing
* at this point.
*/
if (ZS_IS_IRDA(uap)) {
/* Calc baud rate */
baud = uart_get_baud_rate(port, termios, old, 1200, 4000000);
pmz_debug("pmz: switch IRDA to %ld bauds\n", baud);
/* Cet the irda codec to the right rate */
pmz_irda_setup(uap, &baud);
/* Set final baud rate */
pmz_convert_to_zs(uap, termios->c_cflag, termios->c_iflag, baud);
pmz_load_zsregs(uap, uap->curregs);
zssync(uap);
} else {
baud = uart_get_baud_rate(port, termios, old, 1200, 230400);
pmz_convert_to_zs(uap, termios->c_cflag, termios->c_iflag, baud);
/* Make sure modem status interrupts are correctly configured */
if (UART_ENABLE_MS(&uap->port, termios->c_cflag)) {
uap->curregs[R15] |= DCDIE | SYNCIE | CTSIE;
uap->flags |= PMACZILOG_FLAG_MODEM_STATUS;
} else {
uap->curregs[R15] &= ~(DCDIE | SYNCIE | CTSIE);
uap->flags &= ~PMACZILOG_FLAG_MODEM_STATUS;
}
/* Load registers to the chip */
pmz_maybe_update_regs(uap);
}
uart_update_timeout(port, termios->c_cflag, baud);
pmz_debug("pmz: set_termios() done.\n");
}
/* The port lock is not held. */
[PATCH] tty: switch to ktermios This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by: Alan Cox <alan@redhat.com> Signed-off-by: Martin Peschke <mp3@de.ibm.com> Acked-by: Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-08 10:38:45 +00:00
static void pmz_set_termios(struct uart_port *port, struct ktermios *termios,
struct ktermios *old)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct uart_pmac_port *uap = to_pmz(port);
unsigned long flags;
spin_lock_irqsave(&port->lock, flags);
/* Disable IRQs on the port */
uap->curregs[R1] &= ~(EXT_INT_ENAB | TxINT_ENAB | RxINT_MASK);
write_zsreg(uap, R1, uap->curregs[R1]);
/* Setup new port configuration */
__pmz_set_termios(port, termios, old);
/* Re-enable IRQs on the port */
if (ZS_IS_OPEN(uap)) {
uap->curregs[R1] |= INT_ALL_Rx | TxINT_ENAB;
if (!ZS_IS_EXTCLK(uap))
uap->curregs[R1] |= EXT_INT_ENAB;
write_zsreg(uap, R1, uap->curregs[R1]);
}
spin_unlock_irqrestore(&port->lock, flags);
}
static const char *pmz_type(struct uart_port *port)
{
struct uart_pmac_port *uap = to_pmz(port);
if (ZS_IS_IRDA(uap))
return "Z85c30 ESCC - Infrared port";
else if (ZS_IS_INTMODEM(uap))
return "Z85c30 ESCC - Internal modem";
return "Z85c30 ESCC - Serial port";
}
/* We do not request/release mappings of the registers here, this
* happens at early serial probe time.
*/
static void pmz_release_port(struct uart_port *port)
{
}
static int pmz_request_port(struct uart_port *port)
{
return 0;
}
/* These do not need to do anything interesting either. */
static void pmz_config_port(struct uart_port *port, int flags)
{
}
/* We do not support letting the user mess with the divisor, IRQ, etc. */
static int pmz_verify_port(struct uart_port *port, struct serial_struct *ser)
{
return -EINVAL;
}
serial/pmac_zilog: Add console polling support This implements poll_get_char and poll_put_char for console polling, which enables kgdb to work on machines that use the pmac_zilog serial driver, for example the Apple PowerMac G5. Signed-off-by: Chris J Arges <arges@linux.vnet.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-11-04 12:19:06 +00:00
#ifdef CONFIG_CONSOLE_POLL
static int pmz_poll_get_char(struct uart_port *port)
{
struct uart_pmac_port *uap = (struct uart_pmac_port *)port;
while ((read_zsreg(uap, R0) & Rx_CH_AV) == 0)
udelay(5);
return read_zsdata(uap);
}
static void pmz_poll_put_char(struct uart_port *port, unsigned char c)
{
struct uart_pmac_port *uap = (struct uart_pmac_port *)port;
/* Wait for the transmit buffer to empty. */
while ((read_zsreg(uap, R0) & Tx_BUF_EMP) == 0)
udelay(5);
write_zsdata(uap, c);
}
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
static struct uart_ops pmz_pops = {
.tx_empty = pmz_tx_empty,
.set_mctrl = pmz_set_mctrl,
.get_mctrl = pmz_get_mctrl,
.stop_tx = pmz_stop_tx,
.start_tx = pmz_start_tx,
.stop_rx = pmz_stop_rx,
.enable_ms = pmz_enable_ms,
.break_ctl = pmz_break_ctl,
.startup = pmz_startup,
.shutdown = pmz_shutdown,
.set_termios = pmz_set_termios,
.type = pmz_type,
.release_port = pmz_release_port,
.request_port = pmz_request_port,
.config_port = pmz_config_port,
.verify_port = pmz_verify_port,
serial/pmac_zilog: Add console polling support This implements poll_get_char and poll_put_char for console polling, which enables kgdb to work on machines that use the pmac_zilog serial driver, for example the Apple PowerMac G5. Signed-off-by: Chris J Arges <arges@linux.vnet.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2008-11-04 12:19:06 +00:00
#ifdef CONFIG_CONSOLE_POLL
.poll_get_char = pmz_poll_get_char,
.poll_put_char = pmz_poll_put_char,
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
};
/*
* Setup one port structure after probing, HW is down at this point,
* Unlike sunzilog, we don't need to pre-init the spinlock as we don't
* register our console before uart_add_one_port() is called
*/
static int __init pmz_init_port(struct uart_pmac_port *uap)
{
struct device_node *np = uap->node;
[POWERPC] powermac: Constify & voidify get_property() Now that get_property() returns a void *, there's no need to cast its return value. Also, treat the return value as const, so we can constify get_property later. powermac platform & macintosh driver changes. Built for pmac32_defconfig, g5_defconfig Signed-off-by: Jeremy Kerr <jk@ozlabs.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-07-12 05:40:29 +00:00
const char *conn;
const struct slot_names_prop {
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
int count;
char name[1];
} *slots;
int len;
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
struct resource r_ports, r_rxdma, r_txdma;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/*
* Request & map chip registers
*/
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
if (of_address_to_resource(np, 0, &r_ports))
return -ENODEV;
uap->port.mapbase = r_ports.start;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->port.membase = ioremap(uap->port.mapbase, 0x1000);
uap->control_reg = uap->port.membase;
uap->data_reg = uap->control_reg + 0x10;
/*
* Request & map DBDMA registers
*/
#ifdef HAS_DBDMA
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
if (of_address_to_resource(np, 1, &r_txdma) == 0 &&
of_address_to_resource(np, 2, &r_rxdma) == 0)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->flags |= PMACZILOG_FLAG_HAS_DMA;
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
#else
memset(&r_txdma, 0, sizeof(struct resource));
memset(&r_rxdma, 0, sizeof(struct resource));
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
#endif
if (ZS_HAS_DMA(uap)) {
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
uap->tx_dma_regs = ioremap(r_txdma.start, 0x100);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (uap->tx_dma_regs == NULL) {
uap->flags &= ~PMACZILOG_FLAG_HAS_DMA;
goto no_dma;
}
[PATCH] powerpc: Remove device_node addrs/n_addr The pre-parsed addrs/n_addrs fields in struct device_node are finally gone. Remove the dodgy heuristics that did that parsing at boot and remove the fields themselves since we now have a good replacement with the new OF parsing code. This patch also fixes a bunch of drivers to use the new code instead, so that at least pmac32, pseries, iseries and g5 defconfigs build. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-12-13 07:01:21 +00:00
uap->rx_dma_regs = ioremap(r_rxdma.start, 0x100);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (uap->rx_dma_regs == NULL) {
iounmap(uap->tx_dma_regs);
uap->tx_dma_regs = NULL;
uap->flags &= ~PMACZILOG_FLAG_HAS_DMA;
goto no_dma;
}
[POWERPC] Add new interrupt mapping core and change platforms to use it This adds the new irq remapper core and removes the old one. Because there are some fundamental conflicts with the old code, like the value of NO_IRQ which I'm now setting to 0 (as per discussions with Linus), etc..., this commit also changes the relevant platform and driver code over to use the new remapper (so as not to cause difficulties later in bisecting). This patch removes the old pre-parsing of the open firmware interrupt tree along with all the bogus assumptions it made to try to renumber interrupts according to the platform. This is all to be handled by the new code now. For the pSeries XICS interrupt controller, a single remapper host is created for the whole machine regardless of how many interrupt presentation and source controllers are found, and it's set to match any device node that isn't a 8259. That works fine on pSeries and avoids having to deal with some of the complexities of split source controllers vs. presentation controllers in the pSeries device trees. The powerpc i8259 PIC driver now always requests the legacy interrupt range. It also has the feature of being able to match any device node (including NULL) if passed no device node as an input. That will help porting over platforms with broken device-trees like Pegasos who don't have a proper interrupt tree. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-07-03 11:36:01 +00:00
uap->tx_dma_irq = irq_of_parse_and_map(np, 1);
uap->rx_dma_irq = irq_of_parse_and_map(np, 2);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
}
no_dma:
/*
* Detect port type
*/
[POWERPC] Rename device_is_compatible to of_device_is_compatible for consistency with other Open Firmware interfaces (and Sparc). This is just a straight replacement. This leaves the compatibility define in place. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-03 07:26:52 +00:00
if (of_device_is_compatible(np, "cobalt"))
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->flags |= PMACZILOG_FLAG_IS_INTMODEM;
[POWERPC] Rename get_property to of_get_property: drivers These are all the remaining instances of get_property. Simple rename of get_property to of_get_property. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-01 03:54:02 +00:00
conn = of_get_property(np, "AAPL,connector", &len);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (conn && (strcmp(conn, "infrared") == 0))
uap->flags |= PMACZILOG_FLAG_IS_IRDA;
uap->port_type = PMAC_SCC_ASYNC;
/* 1999 Powerbook G3 has slot-names property instead */
[POWERPC] Rename get_property to of_get_property: drivers These are all the remaining instances of get_property. Simple rename of get_property to of_get_property. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-01 03:54:02 +00:00
slots = of_get_property(np, "slot-names", &len);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (slots && slots->count > 0) {
if (strcmp(slots->name, "IrDA") == 0)
uap->flags |= PMACZILOG_FLAG_IS_IRDA;
else if (strcmp(slots->name, "Modem") == 0)
uap->flags |= PMACZILOG_FLAG_IS_INTMODEM;
}
if (ZS_IS_IRDA(uap))
uap->port_type = PMAC_SCC_IRDA;
if (ZS_IS_INTMODEM(uap)) {
[POWERPC] Remove old interface find_devices Replace uses with of_find_node_by_name and for_each_node_by_name. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-04-24 03:53:04 +00:00
struct device_node* i2c_modem =
of_find_node_by_name(NULL, "i2c-modem");
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (i2c_modem) {
[POWERPC] powermac: Constify & voidify get_property() Now that get_property() returns a void *, there's no need to cast its return value. Also, treat the return value as const, so we can constify get_property later. powermac platform & macintosh driver changes. Built for pmac32_defconfig, g5_defconfig Signed-off-by: Jeremy Kerr <jk@ozlabs.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-07-12 05:40:29 +00:00
const char* mid =
[POWERPC] Rename get_property to of_get_property: drivers These are all the remaining instances of get_property. Simple rename of get_property to of_get_property. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-05-01 03:54:02 +00:00
of_get_property(i2c_modem, "modem-id", NULL);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
if (mid) switch(*mid) {
case 0x04 :
case 0x05 :
case 0x07 :
case 0x08 :
case 0x0b :
case 0x0c :
uap->port_type = PMAC_SCC_I2S1;
}
printk(KERN_INFO "pmac_zilog: i2c-modem detected, id: %d\n",
mid ? (*mid) : 0);
[POWERPC] Remove old interface find_devices Replace uses with of_find_node_by_name and for_each_node_by_name. Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-04-24 03:53:04 +00:00
of_node_put(i2c_modem);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
} else {
printk(KERN_INFO "pmac_zilog: serial modem detected\n");
}
}
/*
* Init remaining bits of "port" structure
*/
[SERIAL] uart_port iotype member should use UPIO_* Convert usage of SERIAL_IO_* to UPIO_*. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-02-05 10:48:10 +00:00
uap->port.iotype = UPIO_MEM;
[POWERPC] Add new interrupt mapping core and change platforms to use it This adds the new irq remapper core and removes the old one. Because there are some fundamental conflicts with the old code, like the value of NO_IRQ which I'm now setting to 0 (as per discussions with Linus), etc..., this commit also changes the relevant platform and driver code over to use the new remapper (so as not to cause difficulties later in bisecting). This patch removes the old pre-parsing of the open firmware interrupt tree along with all the bogus assumptions it made to try to renumber interrupts according to the platform. This is all to be handled by the new code now. For the pSeries XICS interrupt controller, a single remapper host is created for the whole machine regardless of how many interrupt presentation and source controllers are found, and it's set to match any device node that isn't a 8259. That works fine on pSeries and avoids having to deal with some of the complexities of split source controllers vs. presentation controllers in the pSeries device trees. The powerpc i8259 PIC driver now always requests the legacy interrupt range. It also has the feature of being able to match any device node (including NULL) if passed no device node as an input. That will help porting over platforms with broken device-trees like Pegasos who don't have a proper interrupt tree. Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2006-07-03 11:36:01 +00:00
uap->port.irq = irq_of_parse_and_map(np, 0);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
uap->port.uartclk = ZS_CLOCK;
uap->port.fifosize = 1;
uap->port.ops = &pmz_pops;
uap->port.type = PORT_PMAC_ZILOG;
uap->port.flags = 0;
/* Setup some valid baud rate information in the register
* shadows so we don't write crap there before baud rate is
* first initialized.
*/
pmz_convert_to_zs(uap, CS8, 0, 9600);
return 0;
}
/*
* Get rid of a port on module removal
*/
static void pmz_dispose_port(struct uart_pmac_port *uap)
{
struct device_node *np;
np = uap->node;
iounmap(uap->rx_dma_regs);
iounmap(uap->tx_dma_regs);
iounmap(uap->control_reg);
uap->node = NULL;
of_node_put(np);
memset(uap, 0, sizeof(struct uart_pmac_port));
}
/*
* Called upon match with an escc node in the devive-tree.
*/
[PATCH] openfirmware: generate device table for userspace This converts the usage of struct of_match to struct of_device_id, similar to pci_device_id. This allows a device table to be generated, which can be parsed by depmod(8) to generate a map file for module loading. In order for hotplug to work with macio devices, patches to module-init-tools and hotplug must be applied. Those patches are available at: ftp://ftp.suse.com/pub/people/jeffm/linux/macio-hotplug/ Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-07-06 19:44:41 +00:00
static int pmz_attach(struct macio_dev *mdev, const struct of_device_id *match)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
int i;
/* Iterate the pmz_ports array to find a matching entry
*/
for (i = 0; i < MAX_ZS_PORTS; i++)
if (pmz_ports[i].node == mdev->ofdev.node) {
struct uart_pmac_port *uap = &pmz_ports[i];
uap->dev = mdev;
dev_set_drvdata(&mdev->ofdev.dev, uap);
if (macio_request_resources(uap->dev, "pmac_zilog"))
printk(KERN_WARNING "%s: Failed to request resource"
", port still active\n",
uap->node->name);
else
uap->flags |= PMACZILOG_FLAG_RSRC_REQUESTED;
return 0;
}
return -ENODEV;
}
/*
* That one should not be called, macio isn't really a hotswap device,
* we don't expect one of those serial ports to go away...
*/
static int pmz_detach(struct macio_dev *mdev)
{
struct uart_pmac_port *uap = dev_get_drvdata(&mdev->ofdev.dev);
if (!uap)
return -ENODEV;
if (uap->flags & PMACZILOG_FLAG_RSRC_REQUESTED) {
macio_release_resources(uap->dev);
uap->flags &= ~PMACZILOG_FLAG_RSRC_REQUESTED;
}
dev_set_drvdata(&mdev->ofdev.dev, NULL);
uap->dev = NULL;
return 0;
}
[PATCH] fix u32 vs. pm_message_t in drivers/ -rc2-mm1 still contains few places where u32 and pm_message_t. This fixes drivers/serial [should change no code]. Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-04-16 22:25:35 +00:00
static int pmz_suspend(struct macio_dev *mdev, pm_message_t pm_state)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
struct uart_pmac_port *uap = dev_get_drvdata(&mdev->ofdev.dev);
struct uart_state *state;
unsigned long flags;
if (uap == NULL) {
printk("HRM... pmz_suspend with NULL uap\n");
return 0;
}
[PATCH] swsusp: switch pm_message_t to struct This adds type-checking to pm_message_t, so that people can't confuse it with int or u32. It also allows us to fix "disk yoyo" during suspend (disk spinning down/up/down). [We've tried that before; since that cpufreq problems were fixed and I've tried make allyes config and fixed resulting damage.] Signed-off-by: Pavel Machek <pavel@suse.cz> Signed-off-by: Alexander Nyberg <alexn@telia.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:56:57 +00:00
if (pm_state.event == mdev->ofdev.dev.power.power_state.event)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
return 0;
[POWERPC] Fix pmac_zilog debug arg drivers/serial/pmac_zilog.c:1590: warning: format '%d' expects type 'int', but argument 3 has type 'pm_message_t' Signed-off-by: Olaf Hering <olaf@aepfle.de> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-08-26 09:10:59 +00:00
pmz_debug("suspend, switching to state %d\n", pm_state.event);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
state = pmz_uart_reg.state + uap->port.line;
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_lock(&pmz_irq_mutex);
[SERIAL] convert uart_state.sem to uart_state.mutex semaphore to mutex conversion. the conversion was generated via scripts, and the result was validated automatically via a script as well. build and boot tested. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-13 21:37:07 +00:00
mutex_lock(&state->mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
spin_lock_irqsave(&uap->port.lock, flags);
if (ZS_IS_OPEN(uap) || ZS_IS_CONS(uap)) {
/* Disable receiver and transmitter. */
uap->curregs[R3] &= ~RxENABLE;
uap->curregs[R5] &= ~TxENABLE;
/* Disable all interrupts and BRK assertion. */
uap->curregs[R1] &= ~(EXT_INT_ENAB | TxINT_ENAB | RxINT_MASK);
uap->curregs[R5] &= ~SND_BRK;
pmz_load_zsregs(uap, uap->curregs);
uap->flags |= PMACZILOG_FLAG_IS_ASLEEP;
mb();
}
spin_unlock_irqrestore(&uap->port.lock, flags);
if (ZS_IS_OPEN(uap) || ZS_IS_OPEN(uap->mate))
if (ZS_IS_ASLEEP(uap->mate) && ZS_IS_IRQ_ON(pmz_get_port_A(uap))) {
pmz_get_port_A(uap)->flags &= ~PMACZILOG_FLAG_IS_IRQ_ON;
disable_irq(uap->port.irq);
}
if (ZS_IS_CONS(uap))
uap->port.cons->flags &= ~CON_ENABLED;
/* Shut the chip down */
pmz_set_scc_power(uap, 0);
[SERIAL] convert uart_state.sem to uart_state.mutex semaphore to mutex conversion. the conversion was generated via scripts, and the result was validated automatically via a script as well. build and boot tested. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-13 21:37:07 +00:00
mutex_unlock(&state->mutex);
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
pmz_debug("suspend, switching complete\n");
mdev->ofdev.dev.power.power_state = pm_state;
return 0;
}
static int pmz_resume(struct macio_dev *mdev)
{
struct uart_pmac_port *uap = dev_get_drvdata(&mdev->ofdev.dev);
struct uart_state *state;
unsigned long flags;
int pwr_delay = 0;
if (uap == NULL)
return 0;
[PATCH] swsusp: switch pm_message_t to struct This adds type-checking to pm_message_t, so that people can't confuse it with int or u32. It also allows us to fix "disk yoyo" during suspend (disk spinning down/up/down). [We've tried that before; since that cpufreq problems were fixed and I've tried make allyes config and fixed resulting damage.] Signed-off-by: Pavel Machek <pavel@suse.cz> Signed-off-by: Alexander Nyberg <alexn@telia.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:56:57 +00:00
if (mdev->ofdev.dev.power.power_state.event == PM_EVENT_ON)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
return 0;
pmz_debug("resume, switching to state 0\n");
state = pmz_uart_reg.state + uap->port.line;
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_lock(&pmz_irq_mutex);
[SERIAL] convert uart_state.sem to uart_state.mutex semaphore to mutex conversion. the conversion was generated via scripts, and the result was validated automatically via a script as well. build and boot tested. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-13 21:37:07 +00:00
mutex_lock(&state->mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
spin_lock_irqsave(&uap->port.lock, flags);
if (!ZS_IS_OPEN(uap) && !ZS_IS_CONS(uap)) {
spin_unlock_irqrestore(&uap->port.lock, flags);
goto bail;
}
pwr_delay = __pmz_startup(uap);
/* Take care of config that may have changed while asleep */
__pmz_set_termios(&uap->port, &uap->termios_cache, NULL);
if (ZS_IS_OPEN(uap)) {
/* Enable interrupts */
uap->curregs[R1] |= INT_ALL_Rx | TxINT_ENAB;
if (!ZS_IS_EXTCLK(uap))
uap->curregs[R1] |= EXT_INT_ENAB;
write_zsreg(uap, R1, uap->curregs[R1]);
}
spin_unlock_irqrestore(&uap->port.lock, flags);
if (ZS_IS_CONS(uap))
uap->port.cons->flags |= CON_ENABLED;
/* Re-enable IRQ on the controller */
if (ZS_IS_OPEN(uap) && !ZS_IS_IRQ_ON(pmz_get_port_A(uap))) {
pmz_get_port_A(uap)->flags |= PMACZILOG_FLAG_IS_IRQ_ON;
enable_irq(uap->port.irq);
}
bail:
[SERIAL] convert uart_state.sem to uart_state.mutex semaphore to mutex conversion. the conversion was generated via scripts, and the result was validated automatically via a script as well. build and boot tested. Signed-off-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-13 21:37:07 +00:00
mutex_unlock(&state->mutex);
[SERIAL] turn serial semaphores into mutexes Turn several drivers/serial/ semaphores-used-as-mutex into mutexes Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-01-12 18:44:32 +00:00
mutex_unlock(&pmz_irq_mutex);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/* Right now, we deal with delay by blocking here, I'll be
* smarter later on
*/
if (pwr_delay != 0) {
pmz_debug("pmz: delaying %d ms\n", pwr_delay);
msleep(pwr_delay);
}
pmz_debug("resume, switching complete\n");
[PATCH] swsusp: switch pm_message_t to struct This adds type-checking to pm_message_t, so that people can't confuse it with int or u32. It also allows us to fix "disk yoyo" during suspend (disk spinning down/up/down). [We've tried that before; since that cpufreq problems were fixed and I've tried make allyes config and fixed resulting damage.] Signed-off-by: Pavel Machek <pavel@suse.cz> Signed-off-by: Alexander Nyberg <alexn@telia.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-03 22:56:57 +00:00
mdev->ofdev.dev.power.power_state.event = PM_EVENT_ON;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
return 0;
}
/*
* Probe all ports in the system and build the ports array, we register
* with the serial layer at this point, the macio-type probing is only
* used later to "attach" to the sysfs tree so we get power management
* events
*/
static int __init pmz_probe(void)
{
struct device_node *node_p, *node_a, *node_b, *np;
int count = 0;
int rc;
/*
* Find all escc chips in the system
*/
node_p = of_find_node_by_name(NULL, "escc");
while (node_p) {
/*
* First get channel A/B node pointers
*
* TODO: Add routines with proper locking to do that...
*/
node_a = node_b = NULL;
for (np = NULL; (np = of_get_next_child(node_p, np)) != NULL;) {
if (strncmp(np->name, "ch-a", 4) == 0)
node_a = of_node_get(np);
else if (strncmp(np->name, "ch-b", 4) == 0)
node_b = of_node_get(np);
}
if (!node_a && !node_b) {
of_node_put(node_a);
of_node_put(node_b);
printk(KERN_ERR "pmac_zilog: missing node %c for escc %s\n",
(!node_a) ? 'a' : 'b', node_p->full_name);
goto next;
}
/*
* Fill basic fields in the port structures
*/
pmz_ports[count].mate = &pmz_ports[count+1];
pmz_ports[count+1].mate = &pmz_ports[count];
pmz_ports[count].flags = PMACZILOG_FLAG_IS_CHANNEL_A;
pmz_ports[count].node = node_a;
pmz_ports[count+1].node = node_b;
pmz_ports[count].port.line = count;
pmz_ports[count+1].port.line = count+1;
/*
* Setup the ports for real
*/
rc = pmz_init_port(&pmz_ports[count]);
if (rc == 0 && node_b != NULL)
rc = pmz_init_port(&pmz_ports[count+1]);
if (rc != 0) {
of_node_put(node_a);
of_node_put(node_b);
memset(&pmz_ports[count], 0, sizeof(struct uart_pmac_port));
memset(&pmz_ports[count+1], 0, sizeof(struct uart_pmac_port));
goto next;
}
count += 2;
next:
node_p = of_find_node_by_name(node_p, "escc");
}
pmz_ports_count = count;
return 0;
}
#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
static void pmz_console_write(struct console *con, const char *s, unsigned int count);
static int __init pmz_console_setup(struct console *co, char *options);
static struct console pmz_console = {
[POWERPC] Optionally use new device number for pmac_zilog This adds the option for the pmac_zilog driver to use the major/minor numbers recently allocated specifically for it (/dev/ttyPZn) instead of the /dev/ttySn numbers. The advantage of doing this is that it allows the pmac_zilog and 8250 drivers to coexist. The disadvantage of doing this is that it is a user-visible ABI change and it will break existing working setups on powermacs, and could be confusing to users. Signed-off-by: David Woodhouse <dwmw2@infradead.org> Signed-off-by: Paul Mackerras <paulus@samba.org>
2007-04-04 14:19:43 +00:00
.name = PMACZILOG_NAME,
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.write = pmz_console_write,
.device = uart_console_device,
.setup = pmz_console_setup,
.flags = CON_PRINTBUFFER,
.index = -1,
.data = &pmz_uart_reg,
};
#define PMACZILOG_CONSOLE &pmz_console
#else /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
#define PMACZILOG_CONSOLE (NULL)
#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
/*
* Register the driver, console driver and ports with the serial
* core
*/
static int __init pmz_register(void)
{
int i, rc;
pmz_uart_reg.nr = pmz_ports_count;
pmz_uart_reg.cons = PMACZILOG_CONSOLE;
/*
* Register this driver with the serial core
*/
rc = uart_register_driver(&pmz_uart_reg);
if (rc)
return rc;
/*
* Register each port with the serial core
*/
for (i = 0; i < pmz_ports_count; i++) {
struct uart_pmac_port *uport = &pmz_ports[i];
/* NULL node may happen on wallstreet */
if (uport->node != NULL)
rc = uart_add_one_port(&pmz_uart_reg, &uport->port);
if (rc)
goto err_out;
}
return 0;
err_out:
while (i-- > 0) {
struct uart_pmac_port *uport = &pmz_ports[i];
uart_remove_one_port(&pmz_uart_reg, &uport->port);
}
uart_unregister_driver(&pmz_uart_reg);
return rc;
}
[PATCH] openfirmware: generate device table for userspace This converts the usage of struct of_match to struct of_device_id, similar to pci_device_id. This allows a device table to be generated, which can be parsed by depmod(8) to generate a map file for module loading. In order for hotplug to work with macio devices, patches to module-init-tools and hotplug must be applied. Those patches are available at: ftp://ftp.suse.com/pub/people/jeffm/linux/macio-hotplug/ Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-07-06 19:44:41 +00:00
static struct of_device_id pmz_match[] =
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
{
{
.name = "ch-a",
},
{
.name = "ch-b",
},
{},
};
[PATCH] openfirmware: generate device table for userspace This converts the usage of struct of_match to struct of_device_id, similar to pci_device_id. This allows a device table to be generated, which can be parsed by depmod(8) to generate a map file for module loading. In order for hotplug to work with macio devices, patches to module-init-tools and hotplug must be applied. Those patches are available at: ftp://ftp.suse.com/pub/people/jeffm/linux/macio-hotplug/ Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-07-06 19:44:41 +00:00
MODULE_DEVICE_TABLE (of, pmz_match);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
static struct macio_driver pmz_driver =
{
.name = "pmac_zilog",
.match_table = pmz_match,
.probe = pmz_attach,
.remove = pmz_detach,
.suspend = pmz_suspend,
.resume = pmz_resume,
};
static int __init init_pmz(void)
{
int rc, i;
printk(KERN_INFO "%s\n", version);
/*
* First, we need to do a direct OF-based probe pass. We
* do that because we want serial console up before the
* macio stuffs calls us back, and since that makes it
* easier to pass the proper number of channels to
* uart_register_driver()
*/
if (pmz_ports_count == 0)
pmz_probe();
/*
* Bail early if no port found
*/
if (pmz_ports_count == 0)
return -ENODEV;
/*
* Now we register with the serial layer
*/
rc = pmz_register();
if (rc) {
printk(KERN_ERR
"pmac_zilog: Error registering serial device, disabling pmac_zilog.\n"
"pmac_zilog: Did another serial driver already claim the minors?\n");
/* effectively "pmz_unprobe()" */
for (i=0; i < pmz_ports_count; i++)
pmz_dispose_port(&pmz_ports[i]);
return rc;
}
/*
* Then we register the macio driver itself
*/
return macio_register_driver(&pmz_driver);
}
static void __exit exit_pmz(void)
{
int i;
/* Get rid of macio-driver (detach from macio) */
macio_unregister_driver(&pmz_driver);
for (i = 0; i < pmz_ports_count; i++) {
struct uart_pmac_port *uport = &pmz_ports[i];
if (uport->node != NULL) {
uart_remove_one_port(&pmz_uart_reg, &uport->port);
pmz_dispose_port(uport);
}
}
/* Unregister UART driver */
uart_unregister_driver(&pmz_uart_reg);
}
#ifdef CONFIG_SERIAL_PMACZILOG_CONSOLE
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-20 20:00:09 +00:00
static void pmz_console_putchar(struct uart_port *port, int ch)
{
struct uart_pmac_port *uap = (struct uart_pmac_port *)port;
/* Wait for the transmit buffer to empty. */
while ((read_zsreg(uap, R0) & Tx_BUF_EMP) == 0)
udelay(5);
write_zsdata(uap, ch);
}
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/*
* Print a string to the serial port trying not to disturb
* any possible real use of the port...
*/
static void pmz_console_write(struct console *con, const char *s, unsigned int count)
{
struct uart_pmac_port *uap = &pmz_ports[con->index];
unsigned long flags;
if (ZS_IS_ASLEEP(uap))
return;
spin_lock_irqsave(&uap->port.lock, flags);
/* Turn of interrupts and enable the transmitter. */
write_zsreg(uap, R1, uap->curregs[1] & ~TxINT_ENAB);
write_zsreg(uap, R5, uap->curregs[5] | TxENABLE | RTS | DTR);
[SERIAL] kernel console should send CRLF not LFCR Glen Turner reported that writing LFCR rather than the more traditional CRLF causes issues with some terminals. Since this aflicts many serial drivers, extract the common code to a library function (uart_console_write) and arrange for each driver to supply a "putchar" function. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-20 20:00:09 +00:00
uart_console_write(&uap->port, s, count, pmz_console_putchar);
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/* Restore the values in the registers. */
write_zsreg(uap, R1, uap->curregs[1]);
/* Don't disable the transmitter. */
spin_unlock_irqrestore(&uap->port.lock, flags);
}
/*
* Setup the serial console
*/
static int __init pmz_console_setup(struct console *co, char *options)
{
struct uart_pmac_port *uap;
struct uart_port *port;
int baud = 38400;
int bits = 8;
int parity = 'n';
int flow = 'n';
unsigned long pwr_delay;
/*
* XServe's default to 57600 bps
*/
if (machine_is_compatible("RackMac1,1")
|| machine_is_compatible("RackMac1,2")
|| machine_is_compatible("MacRISC4"))
baud = 57600;
/*
* Check whether an invalid uart number has been specified, and
* if so, search for the first available port that does have
* console support.
*/
if (co->index >= pmz_ports_count)
co->index = 0;
uap = &pmz_ports[co->index];
if (uap->node == NULL)
return -ENODEV;
port = &uap->port;
/*
* Mark port as beeing a console
*/
uap->flags |= PMACZILOG_FLAG_IS_CONS;
/*
* Temporary fix for uart layer who didn't setup the spinlock yet
*/
spin_lock_init(&port->lock);
/*
* Enable the hardware
*/
pwr_delay = __pmz_startup(uap);
if (pwr_delay)
mdelay(pwr_delay);
if (options)
uart_parse_options(options, &baud, &parity, &bits, &flow);
return uart_set_options(port, co, baud, parity, bits, flow);
}
static int __init pmz_console_init(void)
{
/* Probe ports */
pmz_probe();
/* TODO: Autoprobe console based on OF */
/* pmz_console.index = i; */
register_console(&pmz_console);
return 0;
}
console_initcall(pmz_console_init);
#endif /* CONFIG_SERIAL_PMACZILOG_CONSOLE */
module_init(init_pmz);
module_exit(exit_pmz);
Reference in New Issue Copy Permalink