linux_old1/drivers/char/tty_io.c

3216 lines
80 KiB
C
Raw Normal View History

/*
* linux/drivers/char/tty_io.c
*
* Copyright (C) 1991, 1992 Linus Torvalds
*/
/*
* 'tty_io.c' gives an orthogonal feeling to tty's, be they consoles
* or rs-channels. It also implements echoing, cooked mode etc.
*
* Kill-line thanks to John T Kohl, who also corrected VMIN = VTIME = 0.
*
* Modified by Theodore Ts'o, 9/14/92, to dynamically allocate the
* tty_struct and tty_queue structures. Previously there was an array
* of 256 tty_struct's which was statically allocated, and the
* tty_queue structures were allocated at boot time. Both are now
* dynamically allocated only when the tty is open.
*
* Also restructured routines so that there is more of a separation
* between the high-level tty routines (tty_io.c and tty_ioctl.c) and
* the low-level tty routines (serial.c, pty.c, console.c). This
* makes for cleaner and more compact code. -TYT, 9/17/92
*
* Modified by Fred N. van Kempen, 01/29/93, to add line disciplines
* which can be dynamically activated and de-activated by the line
* discipline handling modules (like SLIP).
*
* NOTE: pay no attention to the line discipline code (yet); its
* interface is still subject to change in this version...
* -- TYT, 1/31/92
*
* Added functionality to the OPOST tty handling. No delays, but all
* other bits should be there.
* -- Nick Holloway <alfie@dcs.warwick.ac.uk>, 27th May 1993.
*
* Rewrote canonical mode and added more termios flags.
* -- julian@uhunix.uhcc.hawaii.edu (J. Cowley), 13Jan94
*
* Reorganized FASYNC support so mouse code can share it.
* -- ctm@ardi.com, 9Sep95
*
* New TIOCLINUX variants added.
* -- mj@k332.feld.cvut.cz, 19-Nov-95
*
* Restrict vt switching via ioctl()
* -- grif@cs.ucr.edu, 5-Dec-95
*
* Move console and virtual terminal code to more appropriate files,
* implement CONFIG_VT and generalize console device interface.
* -- Marko Kohtala <Marko.Kohtala@hut.fi>, March 97
*
* Rewrote init_dev and release_dev to eliminate races.
* -- Bill Hawes <whawes@star.net>, June 97
*
* Added devfs support.
* -- C. Scott Ananian <cananian@alumni.princeton.edu>, 13-Jan-1998
*
* Added support for a Unix98-style ptmx device.
* -- C. Scott Ananian <cananian@alumni.princeton.edu>, 14-Jan-1998
*
* Reduced memory usage for older ARM systems
* -- Russell King <rmk@arm.linux.org.uk>
*
* Move do_SAK() into process context. Less stack use in devfs functions.
* alloc_tty_struct() always uses kmalloc() -- Andrew Morton <andrewm@uow.edu.eu> 17Mar01
*/
#include <linux/config.h>
#include <linux/types.h>
#include <linux/major.h>
#include <linux/errno.h>
#include <linux/signal.h>
#include <linux/fcntl.h>
#include <linux/sched.h>
#include <linux/interrupt.h>
#include <linux/tty.h>
#include <linux/tty_driver.h>
#include <linux/tty_flip.h>
#include <linux/devpts_fs.h>
#include <linux/file.h>
#include <linux/console.h>
#include <linux/timer.h>
#include <linux/ctype.h>
#include <linux/kd.h>
#include <linux/mm.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/poll.h>
#include <linux/proc_fs.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/smp_lock.h>
#include <linux/device.h>
#include <linux/idr.h>
#include <linux/wait.h>
#include <linux/bitops.h>
#include <linux/delay.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/kbd_kern.h>
#include <linux/vt_kern.h>
#include <linux/selection.h>
#include <linux/devfs_fs_kernel.h>
#include <linux/kmod.h>
#undef TTY_DEBUG_HANGUP
#define TTY_PARANOIA_CHECK 1
#define CHECK_TTY_COUNT 1
struct termios tty_std_termios = { /* for the benefit of tty drivers */
.c_iflag = ICRNL | IXON,
.c_oflag = OPOST | ONLCR,
.c_cflag = B38400 | CS8 | CREAD | HUPCL,
.c_lflag = ISIG | ICANON | ECHO | ECHOE | ECHOK |
ECHOCTL | ECHOKE | IEXTEN,
.c_cc = INIT_C_CC
};
EXPORT_SYMBOL(tty_std_termios);
/* This list gets poked at by procfs and various bits of boot up code. This
could do with some rationalisation such as pulling the tty proc function
into this file */
LIST_HEAD(tty_drivers); /* linked list of tty drivers */
/* Semaphore to protect creating and releasing a tty. This is shared with
vt.c for deeply disgusting hack reasons */
DECLARE_MUTEX(tty_sem);
#ifdef CONFIG_UNIX98_PTYS
extern struct tty_driver *ptm_driver; /* Unix98 pty masters; for /dev/ptmx */
extern int pty_limit; /* Config limit on Unix98 ptys */
static DEFINE_IDR(allocated_ptys);
static DECLARE_MUTEX(allocated_ptys_lock);
static int ptmx_open(struct inode *, struct file *);
#endif
extern void disable_early_printk(void);
static void initialize_tty_struct(struct tty_struct *tty);
static ssize_t tty_read(struct file *, char __user *, size_t, loff_t *);
static ssize_t tty_write(struct file *, const char __user *, size_t, loff_t *);
ssize_t redirected_tty_write(struct file *, const char __user *, size_t, loff_t *);
static unsigned int tty_poll(struct file *, poll_table *);
static int tty_open(struct inode *, struct file *);
static int tty_release(struct inode *, struct file *);
int tty_ioctl(struct inode * inode, struct file * file,
unsigned int cmd, unsigned long arg);
static int tty_fasync(int fd, struct file * filp, int on);
static void release_mem(struct tty_struct *tty, int idx);
static struct tty_struct *alloc_tty_struct(void)
{
struct tty_struct *tty;
tty = kmalloc(sizeof(struct tty_struct), GFP_KERNEL);
if (tty)
memset(tty, 0, sizeof(struct tty_struct));
return tty;
}
[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 12:54:13 +08:00
static void tty_buffer_free_all(struct tty_struct *);
static inline void free_tty_struct(struct tty_struct *tty)
{
kfree(tty->write_buf);
[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 12:54:13 +08:00
tty_buffer_free_all(tty);
kfree(tty);
}
#define TTY_NUMBER(tty) ((tty)->index + (tty)->driver->name_base)
char *tty_name(struct tty_struct *tty, char *buf)
{
if (!tty) /* Hmm. NULL pointer. That's fun. */
strcpy(buf, "NULL tty");
else
strcpy(buf, tty->name);
return buf;
}
EXPORT_SYMBOL(tty_name);
int tty_paranoia_check(struct tty_struct *tty, struct inode *inode,
const char *routine)
{
#ifdef TTY_PARANOIA_CHECK
if (!tty) {
printk(KERN_WARNING
"null TTY for (%d:%d) in %s\n",
imajor(inode), iminor(inode), routine);
return 1;
}
if (tty->magic != TTY_MAGIC) {
printk(KERN_WARNING
"bad magic number for tty struct (%d:%d) in %s\n",
imajor(inode), iminor(inode), routine);
return 1;
}
#endif
return 0;
}
static int check_tty_count(struct tty_struct *tty, const char *routine)
{
#ifdef CHECK_TTY_COUNT
struct list_head *p;
int count = 0;
file_list_lock();
list_for_each(p, &tty->tty_files) {
count++;
}
file_list_unlock();
if (tty->driver->type == TTY_DRIVER_TYPE_PTY &&
tty->driver->subtype == PTY_TYPE_SLAVE &&
tty->link && tty->link->count)
count++;
if (tty->count != count) {
printk(KERN_WARNING "Warning: dev (%s) tty->count(%d) "
"!= #fd's(%d) in %s\n",
tty->name, tty->count, count, routine);
return count;
}
#endif
return 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 12:54:13 +08:00
/*
* Tty buffer allocation management
*/
static void tty_buffer_free_all(struct tty_struct *tty)
{
struct tty_buffer *thead;
while((thead = tty->buf.head) != NULL) {
tty->buf.head = thead->next;
kfree(thead);
}
while((thead = tty->buf.free) != NULL) {
tty->buf.free = thead->next;
kfree(thead);
}
tty->buf.tail = NULL;
}
static void tty_buffer_init(struct tty_struct *tty)
{
tty->buf.head = NULL;
tty->buf.tail = NULL;
tty->buf.free = NULL;
}
static struct tty_buffer *tty_buffer_alloc(size_t size)
{
struct tty_buffer *p = kmalloc(sizeof(struct tty_buffer) + 2 * size, GFP_ATOMIC);
if(p == NULL)
return NULL;
p->used = 0;
p->size = size;
p->next = NULL;
p->char_buf_ptr = (char *)(p->data);
p->flag_buf_ptr = (unsigned char *)p->char_buf_ptr + size;
/* printk("Flip create %p\n", p); */
return p;
}
/* Must be called with the tty_read lock held. This needs to acquire strategy
code to decide if we should kfree or relink a given expired buffer */
static void tty_buffer_free(struct tty_struct *tty, struct tty_buffer *b)
{
/* Dumb strategy for now - should keep some stats */
/* printk("Flip dispose %p\n", b); */
if(b->size >= 512)
kfree(b);
else {
b->next = tty->buf.free;
tty->buf.free = b;
}
}
static struct tty_buffer *tty_buffer_find(struct tty_struct *tty, size_t size)
{
struct tty_buffer **tbh = &tty->buf.free;
while((*tbh) != NULL) {
struct tty_buffer *t = *tbh;
if(t->size >= size) {
*tbh = t->next;
t->next = NULL;
t->used = 0;
/* DEBUG ONLY */
memset(t->data, '*', size);
/* printk("Flip recycle %p\n", t); */
return t;
}
tbh = &((*tbh)->next);
}
/* Round the buffer size out */
size = (size + 0xFF) & ~ 0xFF;
return tty_buffer_alloc(size);
/* Should possibly check if this fails for the largest buffer we
have queued and recycle that ? */
}
int tty_buffer_request_room(struct tty_struct *tty, size_t size)
{
struct tty_buffer *b = tty->buf.tail, *n;
[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 12:54:13 +08:00
int left = 0;
/* OPTIMISATION: We could keep a per tty "zero" sized buffer to
remove this conditional if its worth it. This would be invisible
to the callers */
if(b != NULL)
left = b->size - b->used;
if(left >= size)
return size;
/* This is the slow path - looking for new buffers to use */
n = tty_buffer_find(tty, size);
if(n == NULL)
return left;
if(b != NULL)
b->next = n;
else
tty->buf.head = n;
tty->buf.tail = n;
return size;
}
EXPORT_SYMBOL_GPL(tty_buffer_request_room);
int tty_insert_flip_string(struct tty_struct *tty, unsigned char *chars, size_t size)
{
int copied = 0;
do {
int space = tty_buffer_request_room(tty, size - copied);
struct tty_buffer *tb = tty->buf.tail;
/* If there is no space then tb may be NULL */
if(unlikely(space == 0))
break;
memcpy(tb->char_buf_ptr + tb->used, chars, space);
memset(tb->flag_buf_ptr + tb->used, TTY_NORMAL, space);
tb->used += space;
copied += space;
chars += space;
/* printk("Flip insert %d.\n", space); */
}
/* There is a small chance that we need to split the data over
several buffers. If this is the case we must loop */
while (unlikely(size > copied));
return copied;
}
EXPORT_SYMBOL_GPL(tty_insert_flip_string);
int tty_insert_flip_string_flags(struct tty_struct *tty, unsigned char *chars, char *flags, size_t size)
{
int copied = 0;
do {
int space = tty_buffer_request_room(tty, size - copied);
struct tty_buffer *tb = tty->buf.tail;
/* If there is no space then tb may be NULL */
if(unlikely(space == 0))
break;
memcpy(tb->char_buf_ptr + tb->used, chars, space);
memcpy(tb->flag_buf_ptr + tb->used, flags, space);
tb->used += space;
copied += space;
chars += space;
flags += space;
}
/* There is a small chance that we need to split the data over
several buffers. If this is the case we must loop */
while (unlikely(size > copied));
return copied;
}
EXPORT_SYMBOL_GPL(tty_insert_flip_string_flags);
/*
* Prepare a block of space in the buffer for data. Returns the length
* available and buffer pointer to the space which is now allocated and
* accounted for as ready for normal characters. This is used for drivers
* that need their own block copy routines into the buffer. There is no
* guarantee the buffer is a DMA target!
*/
int tty_prepare_flip_string(struct tty_struct *tty, unsigned char **chars, size_t size)
{
int space = tty_buffer_request_room(tty, size);
struct tty_buffer *tb = tty->buf.tail;
*chars = tb->char_buf_ptr + tb->used;
memset(tb->flag_buf_ptr + tb->used, TTY_NORMAL, space);
tb->used += space;
return space;
}
EXPORT_SYMBOL_GPL(tty_prepare_flip_string);
/*
* Prepare a block of space in the buffer for data. Returns the length
* available and buffer pointer to the space which is now allocated and
* accounted for as ready for characters. This is used for drivers
* that need their own block copy routines into the buffer. There is no
* guarantee the buffer is a DMA target!
*/
int tty_prepare_flip_string_flags(struct tty_struct *tty, unsigned char **chars, char **flags, size_t size)
{
int space = tty_buffer_request_room(tty, size);
struct tty_buffer *tb = tty->buf.tail;
*chars = tb->char_buf_ptr + tb->used;
*flags = tb->flag_buf_ptr + tb->used;
tb->used += space;
return space;
}
EXPORT_SYMBOL_GPL(tty_prepare_flip_string_flags);
/*
* This is probably overkill for real world processors but
* they are not on hot paths so a little discipline won't do
* any harm.
*/
static void tty_set_termios_ldisc(struct tty_struct *tty, int num)
{
down(&tty->termios_sem);
tty->termios->c_line = num;
up(&tty->termios_sem);
}
/*
* This guards the refcounted line discipline lists. The lock
* must be taken with irqs off because there are hangup path
* callers who will do ldisc lookups and cannot sleep.
*/
static DEFINE_SPINLOCK(tty_ldisc_lock);
static DECLARE_WAIT_QUEUE_HEAD(tty_ldisc_wait);
static struct tty_ldisc tty_ldiscs[NR_LDISCS]; /* line disc dispatch table */
int tty_register_ldisc(int disc, struct tty_ldisc *new_ldisc)
{
unsigned long flags;
int ret = 0;
if (disc < N_TTY || disc >= NR_LDISCS)
return -EINVAL;
spin_lock_irqsave(&tty_ldisc_lock, flags);
tty_ldiscs[disc] = *new_ldisc;
tty_ldiscs[disc].num = disc;
tty_ldiscs[disc].flags |= LDISC_FLAG_DEFINED;
tty_ldiscs[disc].refcount = 0;
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
return ret;
}
EXPORT_SYMBOL(tty_register_ldisc);
int tty_unregister_ldisc(int disc)
{
unsigned long flags;
int ret = 0;
if (disc < N_TTY || disc >= NR_LDISCS)
return -EINVAL;
spin_lock_irqsave(&tty_ldisc_lock, flags);
if (tty_ldiscs[disc].refcount)
ret = -EBUSY;
else
tty_ldiscs[disc].flags &= ~LDISC_FLAG_DEFINED;
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
return ret;
}
EXPORT_SYMBOL(tty_unregister_ldisc);
struct tty_ldisc *tty_ldisc_get(int disc)
{
unsigned long flags;
struct tty_ldisc *ld;
if (disc < N_TTY || disc >= NR_LDISCS)
return NULL;
spin_lock_irqsave(&tty_ldisc_lock, flags);
ld = &tty_ldiscs[disc];
/* Check the entry is defined */
if(ld->flags & LDISC_FLAG_DEFINED)
{
/* If the module is being unloaded we can't use it */
if (!try_module_get(ld->owner))
ld = NULL;
else /* lock it */
ld->refcount++;
}
else
ld = NULL;
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
return ld;
}
EXPORT_SYMBOL_GPL(tty_ldisc_get);
void tty_ldisc_put(int disc)
{
struct tty_ldisc *ld;
unsigned long flags;
if (disc < N_TTY || disc >= NR_LDISCS)
BUG();
spin_lock_irqsave(&tty_ldisc_lock, flags);
ld = &tty_ldiscs[disc];
if(ld->refcount == 0)
BUG();
ld->refcount --;
module_put(ld->owner);
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
}
EXPORT_SYMBOL_GPL(tty_ldisc_put);
static void tty_ldisc_assign(struct tty_struct *tty, struct tty_ldisc *ld)
{
tty->ldisc = *ld;
tty->ldisc.refcount = 0;
}
/**
* tty_ldisc_try - internal helper
* @tty: the tty
*
* Make a single attempt to grab and bump the refcount on
* the tty ldisc. Return 0 on failure or 1 on success. This is
* used to implement both the waiting and non waiting versions
* of tty_ldisc_ref
*/
static int tty_ldisc_try(struct tty_struct *tty)
{
unsigned long flags;
struct tty_ldisc *ld;
int ret = 0;
spin_lock_irqsave(&tty_ldisc_lock, flags);
ld = &tty->ldisc;
if(test_bit(TTY_LDISC, &tty->flags))
{
ld->refcount++;
ret = 1;
}
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
return ret;
}
/**
* tty_ldisc_ref_wait - wait for the tty ldisc
* @tty: tty device
*
* Dereference the line discipline for the terminal and take a
* reference to it. If the line discipline is in flux then
* wait patiently until it changes.
*
* Note: Must not be called from an IRQ/timer context. The caller
* must also be careful not to hold other locks that will deadlock
* against a discipline change, such as an existing ldisc reference
* (which we check for)
*/
struct tty_ldisc *tty_ldisc_ref_wait(struct tty_struct *tty)
{
/* wait_event is a macro */
wait_event(tty_ldisc_wait, tty_ldisc_try(tty));
if(tty->ldisc.refcount == 0)
printk(KERN_ERR "tty_ldisc_ref_wait\n");
return &tty->ldisc;
}
EXPORT_SYMBOL_GPL(tty_ldisc_ref_wait);
/**
* tty_ldisc_ref - get the tty ldisc
* @tty: tty device
*
* Dereference the line discipline for the terminal and take a
* reference to it. If the line discipline is in flux then
* return NULL. Can be called from IRQ and timer functions.
*/
struct tty_ldisc *tty_ldisc_ref(struct tty_struct *tty)
{
if(tty_ldisc_try(tty))
return &tty->ldisc;
return NULL;
}
EXPORT_SYMBOL_GPL(tty_ldisc_ref);
/**
* tty_ldisc_deref - free a tty ldisc reference
* @ld: reference to free up
*
* Undoes the effect of tty_ldisc_ref or tty_ldisc_ref_wait. May
* be called in IRQ context.
*/
void tty_ldisc_deref(struct tty_ldisc *ld)
{
unsigned long flags;
if(ld == NULL)
BUG();
spin_lock_irqsave(&tty_ldisc_lock, flags);
if(ld->refcount == 0)
printk(KERN_ERR "tty_ldisc_deref: no references.\n");
else
ld->refcount--;
if(ld->refcount == 0)
wake_up(&tty_ldisc_wait);
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
}
EXPORT_SYMBOL_GPL(tty_ldisc_deref);
/**
* tty_ldisc_enable - allow ldisc use
* @tty: terminal to activate ldisc on
*
* Set the TTY_LDISC flag when the line discipline can be called
* again. Do neccessary wakeups for existing sleepers.
*
* Note: nobody should set this bit except via this function. Clearing
* directly is allowed.
*/
static void tty_ldisc_enable(struct tty_struct *tty)
{
set_bit(TTY_LDISC, &tty->flags);
wake_up(&tty_ldisc_wait);
}
/**
* tty_set_ldisc - set line discipline
* @tty: the terminal to set
* @ldisc: the line discipline
*
* Set the discipline of a tty line. Must be called from a process
* context.
*/
static int tty_set_ldisc(struct tty_struct *tty, int ldisc)
{
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
int retval = 0;
struct tty_ldisc o_ldisc;
char buf[64];
int work;
unsigned long flags;
struct tty_ldisc *ld;
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
struct tty_struct *o_tty;
if ((ldisc < N_TTY) || (ldisc >= NR_LDISCS))
return -EINVAL;
restart:
ld = tty_ldisc_get(ldisc);
/* Eduardo Blanco <ejbs@cs.cs.com.uy> */
/* Cyrus Durgin <cider@speakeasy.org> */
if (ld == NULL) {
request_module("tty-ldisc-%d", ldisc);
ld = tty_ldisc_get(ldisc);
}
if (ld == NULL)
return -EINVAL;
[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 12:54:13 +08:00
/*
* No more input please, we are switching. The new ldisc
* will update this value in the ldisc open function
*/
tty->receive_room = 0;
/*
* Problem: What do we do if this blocks ?
*/
tty_wait_until_sent(tty, 0);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
if (tty->ldisc.num == ldisc) {
tty_ldisc_put(ldisc);
return 0;
}
o_ldisc = tty->ldisc;
o_tty = tty->link;
/*
* Make sure we don't change while someone holds a
* reference to the line discipline. The TTY_LDISC bit
* prevents anyone taking a reference once it is clear.
* We need the lock to avoid racing reference takers.
*/
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
spin_lock_irqsave(&tty_ldisc_lock, flags);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
if (tty->ldisc.refcount || (o_tty && o_tty->ldisc.refcount)) {
if(tty->ldisc.refcount) {
/* Free the new ldisc we grabbed. Must drop the lock
first. */
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
tty_ldisc_put(ldisc);
/*
* There are several reasons we may be busy, including
* random momentary I/O traffic. We must therefore
* retry. We could distinguish between blocking ops
* and retries if we made tty_ldisc_wait() smarter. That
* is up for discussion.
*/
if (wait_event_interruptible(tty_ldisc_wait, tty->ldisc.refcount == 0) < 0)
return -ERESTARTSYS;
goto restart;
}
if(o_tty && o_tty->ldisc.refcount) {
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
tty_ldisc_put(ldisc);
if (wait_event_interruptible(tty_ldisc_wait, o_tty->ldisc.refcount == 0) < 0)
return -ERESTARTSYS;
goto restart;
}
}
/* if the TTY_LDISC bit is set, then we are racing against another ldisc change */
if (!test_bit(TTY_LDISC, &tty->flags)) {
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
tty_ldisc_put(ldisc);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
ld = tty_ldisc_ref_wait(tty);
tty_ldisc_deref(ld);
goto restart;
}
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
clear_bit(TTY_LDISC, &tty->flags);
clear_bit(TTY_DONT_FLIP, &tty->flags);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
if (o_tty) {
clear_bit(TTY_LDISC, &o_tty->flags);
clear_bit(TTY_DONT_FLIP, &o_tty->flags);
}
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
/*
* From this point on we know nobody has an ldisc
* usage reference, nor can they obtain one until
* we say so later on.
*/
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
[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 12:54:13 +08:00
work = cancel_delayed_work(&tty->buf.work);
/*
[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 12:54:13 +08:00
* Wait for ->hangup_work and ->buf.work handlers to terminate
*/
flush_scheduled_work();
/* Shutdown the current discipline. */
if (tty->ldisc.close)
(tty->ldisc.close)(tty);
/* Now set up the new line discipline. */
tty_ldisc_assign(tty, ld);
tty_set_termios_ldisc(tty, ldisc);
if (tty->ldisc.open)
retval = (tty->ldisc.open)(tty);
if (retval < 0) {
tty_ldisc_put(ldisc);
/* There is an outstanding reference here so this is safe */
tty_ldisc_assign(tty, tty_ldisc_get(o_ldisc.num));
tty_set_termios_ldisc(tty, tty->ldisc.num);
if (tty->ldisc.open && (tty->ldisc.open(tty) < 0)) {
tty_ldisc_put(o_ldisc.num);
/* This driver is always present */
tty_ldisc_assign(tty, tty_ldisc_get(N_TTY));
tty_set_termios_ldisc(tty, N_TTY);
if (tty->ldisc.open) {
int r = tty->ldisc.open(tty);
if (r < 0)
panic("Couldn't open N_TTY ldisc for "
"%s --- error %d.",
tty_name(tty, buf), r);
}
}
}
/* At this point we hold a reference to the new ldisc and a
a reference to the old ldisc. If we ended up flipping back
to the existing ldisc we have two references to it */
if (tty->ldisc.num != o_ldisc.num && tty->driver->set_ldisc)
tty->driver->set_ldisc(tty);
tty_ldisc_put(o_ldisc.num);
/*
* Allow ldisc referencing to occur as soon as the driver
* ldisc callback completes.
*/
tty_ldisc_enable(tty);
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
if (o_tty)
tty_ldisc_enable(o_tty);
/* Restart it in case no characters kick it off. Safe if
already running */
[PATCH] pty_chars_in_buffer oops fix The idea of this patch is to lock both sides of a ptmx/pty pair during line discipline changing. This is needed to ensure that say a poll on one side of the pty doesn't occur while the line discipline is actively being changed. This resulted in an oops reported on lkml, see: http://marc.theaimsgroup.com/?l=linux-kernel&m=111342171410005&w=2 A 'hacky' approach was previously implmemented which served to eliminate the poll vs. line discipline changing race. However, this patch takes a more general approach to the issue. The patch only adds locking on a less often used path, the line-discipline changing path, as opposed to locking the ptmx/pty pair on read/write/poll paths. The patch below, takes both ldisc locks in either order b/c the locks are both taken under the same spinlock(). I thought about locking the ptmx/pty separately, such as master always first but that introduces a 3 way deadlock. For example, process 1 does a blocking read on the slave side. Then, process 2 does an ldisc change on the slave side, which acquires the master ldisc lock but not the slave's. Finally, process 3 does a write which blocks on the process 2's ldisc reference. This patch does introduce some changes in semantics. For example, a line discipline change on side 'a' of a ptmx/pty pair, will now wait for a read/write to complete on the other side, or side 'b'. The current behavior is to simply wait for any read/writes on only side 'a', not both sides 'a' and 'b'. I think this behavior makes sense, but I wanted to point it out. I've tested the patch with a bunch of read/write/poll while changing the line discipline out from underneath. This patch obviates the need for the above "hide the problem" patch. Signed-off-by: Jason Baron <jbaron@redhat.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-09-10 04:01:57 +08:00
if (work)
[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 12:54:13 +08:00
schedule_delayed_work(&tty->buf.work, 1);
return retval;
}
/*
* This routine returns a tty driver structure, given a device number
*/
static struct tty_driver *get_tty_driver(dev_t device, int *index)
{
struct tty_driver *p;
list_for_each_entry(p, &tty_drivers, tty_drivers) {
dev_t base = MKDEV(p->major, p->minor_start);
if (device < base || device >= base + p->num)
continue;
*index = device - base;
return p;
}
return NULL;
}
/*
* If we try to write to, or set the state of, a terminal and we're
* not in the foreground, send a SIGTTOU. If the signal is blocked or
* ignored, go ahead and perform the operation. (POSIX 7.2)
*/
int tty_check_change(struct tty_struct * tty)
{
if (current->signal->tty != tty)
return 0;
if (tty->pgrp <= 0) {
printk(KERN_WARNING "tty_check_change: tty->pgrp <= 0!\n");
return 0;
}
if (process_group(current) == tty->pgrp)
return 0;
if (is_ignored(SIGTTOU))
return 0;
if (is_orphaned_pgrp(process_group(current)))
return -EIO;
(void) kill_pg(process_group(current), SIGTTOU, 1);
return -ERESTARTSYS;
}
EXPORT_SYMBOL(tty_check_change);
static ssize_t hung_up_tty_read(struct file * file, char __user * buf,
size_t count, loff_t *ppos)
{
return 0;
}
static ssize_t hung_up_tty_write(struct file * file, const char __user * buf,
size_t count, loff_t *ppos)
{
return -EIO;
}
/* No kernel lock held - none needed ;) */
static unsigned int hung_up_tty_poll(struct file * filp, poll_table * wait)
{
return POLLIN | POLLOUT | POLLERR | POLLHUP | POLLRDNORM | POLLWRNORM;
}
static int hung_up_tty_ioctl(struct inode * inode, struct file * file,
unsigned int cmd, unsigned long arg)
{
return cmd == TIOCSPGRP ? -ENOTTY : -EIO;
}
static struct file_operations tty_fops = {
.llseek = no_llseek,
.read = tty_read,
.write = tty_write,
.poll = tty_poll,
.ioctl = tty_ioctl,
.open = tty_open,
.release = tty_release,
.fasync = tty_fasync,
};
#ifdef CONFIG_UNIX98_PTYS
static struct file_operations ptmx_fops = {
.llseek = no_llseek,
.read = tty_read,
.write = tty_write,
.poll = tty_poll,
.ioctl = tty_ioctl,
.open = ptmx_open,
.release = tty_release,
.fasync = tty_fasync,
};
#endif
static struct file_operations console_fops = {
.llseek = no_llseek,
.read = tty_read,
.write = redirected_tty_write,
.poll = tty_poll,
.ioctl = tty_ioctl,
.open = tty_open,
.release = tty_release,
.fasync = tty_fasync,
};
static struct file_operations hung_up_tty_fops = {
.llseek = no_llseek,
.read = hung_up_tty_read,
.write = hung_up_tty_write,
.poll = hung_up_tty_poll,
.ioctl = hung_up_tty_ioctl,
.release = tty_release,
};
static DEFINE_SPINLOCK(redirect_lock);
static struct file *redirect;
/**
* tty_wakeup - request more data
* @tty: terminal
*
* Internal and external helper for wakeups of tty. This function
* informs the line discipline if present that the driver is ready
* to receive more output data.
*/
void tty_wakeup(struct tty_struct *tty)
{
struct tty_ldisc *ld;
if (test_bit(TTY_DO_WRITE_WAKEUP, &tty->flags)) {
ld = tty_ldisc_ref(tty);
if(ld) {
if(ld->write_wakeup)
ld->write_wakeup(tty);
tty_ldisc_deref(ld);
}
}
wake_up_interruptible(&tty->write_wait);
}
EXPORT_SYMBOL_GPL(tty_wakeup);
/**
* tty_ldisc_flush - flush line discipline queue
* @tty: tty
*
* Flush the line discipline queue (if any) for this tty. If there
* is no line discipline active this is a no-op.
*/
void tty_ldisc_flush(struct tty_struct *tty)
{
struct tty_ldisc *ld = tty_ldisc_ref(tty);
if(ld) {
if(ld->flush_buffer)
ld->flush_buffer(tty);
tty_ldisc_deref(ld);
}
}
EXPORT_SYMBOL_GPL(tty_ldisc_flush);
/*
* This can be called by the "eventd" kernel thread. That is process synchronous,
* but doesn't hold any locks, so we need to make sure we have the appropriate
* locks for what we're doing..
*/
static void do_tty_hangup(void *data)
{
struct tty_struct *tty = (struct tty_struct *) data;
struct file * cons_filp = NULL;
struct file *filp, *f = NULL;
struct task_struct *p;
struct tty_ldisc *ld;
int closecount = 0, n;
if (!tty)
return;
/* inuse_filps is protected by the single kernel lock */
lock_kernel();
spin_lock(&redirect_lock);
if (redirect && redirect->private_data == tty) {
f = redirect;
redirect = NULL;
}
spin_unlock(&redirect_lock);
check_tty_count(tty, "do_tty_hangup");
file_list_lock();
/* This breaks for file handles being sent over AF_UNIX sockets ? */
list_for_each_entry(filp, &tty->tty_files, f_u.fu_list) {
if (filp->f_op->write == redirected_tty_write)
cons_filp = filp;
if (filp->f_op->write != tty_write)
continue;
closecount++;
tty_fasync(-1, filp, 0); /* can't block */
filp->f_op = &hung_up_tty_fops;
}
file_list_unlock();
/* FIXME! What are the locking issues here? This may me overdoing things..
* this question is especially important now that we've removed the irqlock. */
ld = tty_ldisc_ref(tty);
if(ld != NULL) /* We may have no line discipline at this point */
{
if (ld->flush_buffer)
ld->flush_buffer(tty);
if (tty->driver->flush_buffer)
tty->driver->flush_buffer(tty);
if ((test_bit(TTY_DO_WRITE_WAKEUP, &tty->flags)) &&
ld->write_wakeup)
ld->write_wakeup(tty);
if (ld->hangup)
ld->hangup(tty);
}
/* FIXME: Once we trust the LDISC code better we can wait here for
ldisc completion and fix the driver call race */
wake_up_interruptible(&tty->write_wait);
wake_up_interruptible(&tty->read_wait);
/*
* Shutdown the current line discipline, and reset it to
* N_TTY.
*/
if (tty->driver->flags & TTY_DRIVER_RESET_TERMIOS)
{
down(&tty->termios_sem);
*tty->termios = tty->driver->init_termios;
up(&tty->termios_sem);
}
/* Defer ldisc switch */
/* tty_deferred_ldisc_switch(N_TTY);
This should get done automatically when the port closes and
tty_release is called */
read_lock(&tasklist_lock);
if (tty->session > 0) {
do_each_task_pid(tty->session, PIDTYPE_SID, p) {
if (p->signal->tty == tty)
p->signal->tty = NULL;
if (!p->signal->leader)
continue;
send_group_sig_info(SIGHUP, SEND_SIG_PRIV, p);
send_group_sig_info(SIGCONT, SEND_SIG_PRIV, p);
if (tty->pgrp > 0)
p->signal->tty_old_pgrp = tty->pgrp;
} while_each_task_pid(tty->session, PIDTYPE_SID, p);
}
read_unlock(&tasklist_lock);
tty->flags = 0;
tty->session = 0;
tty->pgrp = -1;
tty->ctrl_status = 0;
/*
* If one of the devices matches a console pointer, we
* cannot just call hangup() because that will cause
* tty->count and state->count to go out of sync.
* So we just call close() the right number of times.
*/
if (cons_filp) {
if (tty->driver->close)
for (n = 0; n < closecount; n++)
tty->driver->close(tty, cons_filp);
} else if (tty->driver->hangup)
(tty->driver->hangup)(tty);
/* We don't want to have driver/ldisc interactions beyond
the ones we did here. The driver layer expects no
calls after ->hangup() from the ldisc side. However we
can't yet guarantee all that */
set_bit(TTY_HUPPED, &tty->flags);
if (ld) {
tty_ldisc_enable(tty);
tty_ldisc_deref(ld);
}
unlock_kernel();
if (f)
fput(f);
}
void tty_hangup(struct tty_struct * tty)
{
#ifdef TTY_DEBUG_HANGUP
char buf[64];
printk(KERN_DEBUG "%s hangup...\n", tty_name(tty, buf));
#endif
schedule_work(&tty->hangup_work);
}
EXPORT_SYMBOL(tty_hangup);
void tty_vhangup(struct tty_struct * tty)
{
#ifdef TTY_DEBUG_HANGUP
char buf[64];
printk(KERN_DEBUG "%s vhangup...\n", tty_name(tty, buf));
#endif
do_tty_hangup((void *) tty);
}
EXPORT_SYMBOL(tty_vhangup);
int tty_hung_up_p(struct file * filp)
{
return (filp->f_op == &hung_up_tty_fops);
}
EXPORT_SYMBOL(tty_hung_up_p);
/*
* This function is typically called only by the session leader, when
* it wants to disassociate itself from its controlling tty.
*
* It performs the following functions:
* (1) Sends a SIGHUP and SIGCONT to the foreground process group
* (2) Clears the tty from being controlling the session
* (3) Clears the controlling tty for all processes in the
* session group.
*
* The argument on_exit is set to 1 if called when a process is
* exiting; it is 0 if called by the ioctl TIOCNOTTY.
*/
void disassociate_ctty(int on_exit)
{
struct tty_struct *tty;
struct task_struct *p;
int tty_pgrp = -1;
lock_kernel();
down(&tty_sem);
tty = current->signal->tty;
if (tty) {
tty_pgrp = tty->pgrp;
up(&tty_sem);
if (on_exit && tty->driver->type != TTY_DRIVER_TYPE_PTY)
tty_vhangup(tty);
} else {
if (current->signal->tty_old_pgrp) {
kill_pg(current->signal->tty_old_pgrp, SIGHUP, on_exit);
kill_pg(current->signal->tty_old_pgrp, SIGCONT, on_exit);
}
up(&tty_sem);
unlock_kernel();
return;
}
if (tty_pgrp > 0) {
kill_pg(tty_pgrp, SIGHUP, on_exit);
if (!on_exit)
kill_pg(tty_pgrp, SIGCONT, on_exit);
}
/* Must lock changes to tty_old_pgrp */
down(&tty_sem);
current->signal->tty_old_pgrp = 0;
tty->session = 0;
tty->pgrp = -1;
/* Now clear signal->tty under the lock */
read_lock(&tasklist_lock);
do_each_task_pid(current->signal->session, PIDTYPE_SID, p) {
p->signal->tty = NULL;
} while_each_task_pid(current->signal->session, PIDTYPE_SID, p);
read_unlock(&tasklist_lock);
up(&tty_sem);
unlock_kernel();
}
void stop_tty(struct tty_struct *tty)
{
if (tty->stopped)
return;
tty->stopped = 1;
if (tty->link && tty->link->packet) {
tty->ctrl_status &= ~TIOCPKT_START;
tty->ctrl_status |= TIOCPKT_STOP;
wake_up_interruptible(&tty->link->read_wait);
}
if (tty->driver->stop)
(tty->driver->stop)(tty);
}
EXPORT_SYMBOL(stop_tty);
void start_tty(struct tty_struct *tty)
{
if (!tty->stopped || tty->flow_stopped)
return;
tty->stopped = 0;
if (tty->link && tty->link->packet) {
tty->ctrl_status &= ~TIOCPKT_STOP;
tty->ctrl_status |= TIOCPKT_START;
wake_up_interruptible(&tty->link->read_wait);
}
if (tty->driver->start)
(tty->driver->start)(tty);
/* If we have a running line discipline it may need kicking */
tty_wakeup(tty);
wake_up_interruptible(&tty->write_wait);
}
EXPORT_SYMBOL(start_tty);
static ssize_t tty_read(struct file * file, char __user * buf, size_t count,
loff_t *ppos)
{
int i;
struct tty_struct * tty;
struct inode *inode;
struct tty_ldisc *ld;
tty = (struct tty_struct *)file->private_data;
inode = file->f_dentry->d_inode;
if (tty_paranoia_check(tty, inode, "tty_read"))
return -EIO;
if (!tty || (test_bit(TTY_IO_ERROR, &tty->flags)))
return -EIO;
/* We want to wait for the line discipline to sort out in this
situation */
ld = tty_ldisc_ref_wait(tty);
lock_kernel();
if (ld->read)
i = (ld->read)(tty,file,buf,count);
else
i = -EIO;
tty_ldisc_deref(ld);
unlock_kernel();
if (i > 0)
inode->i_atime = current_fs_time(inode->i_sb);
return i;
}
/*
* Split writes up in sane blocksizes to avoid
* denial-of-service type attacks
*/
static inline ssize_t do_tty_write(
ssize_t (*write)(struct tty_struct *, struct file *, const unsigned char *, size_t),
struct tty_struct *tty,
struct file *file,
const char __user *buf,
size_t count)
{
ssize_t ret = 0, written = 0;
unsigned int chunk;
if (down_interruptible(&tty->atomic_write)) {
return -ERESTARTSYS;
}
/*
* We chunk up writes into a temporary buffer. This
* simplifies low-level drivers immensely, since they
* don't have locking issues and user mode accesses.
*
* But if TTY_NO_WRITE_SPLIT is set, we should use a
* big chunk-size..
*
* The default chunk-size is 2kB, because the NTTY
* layer has problems with bigger chunks. It will
* claim to be able to handle more characters than
* it actually does.
*/
chunk = 2048;
if (test_bit(TTY_NO_WRITE_SPLIT, &tty->flags))
chunk = 65536;
if (count < chunk)
chunk = count;
/* write_buf/write_cnt is protected by the atomic_write semaphore */
if (tty->write_cnt < chunk) {
unsigned char *buf;
if (chunk < 1024)
chunk = 1024;
buf = kmalloc(chunk, GFP_KERNEL);
if (!buf) {
up(&tty->atomic_write);
return -ENOMEM;
}
kfree(tty->write_buf);
tty->write_cnt = chunk;
tty->write_buf = buf;
}
/* Do the write .. */
for (;;) {
size_t size = count;
if (size > chunk)
size = chunk;
ret = -EFAULT;
if (copy_from_user(tty->write_buf, buf, size))
break;
lock_kernel();
ret = write(tty, file, tty->write_buf, size);
unlock_kernel();
if (ret <= 0)
break;
written += ret;
buf += ret;
count -= ret;
if (!count)
break;
ret = -ERESTARTSYS;
if (signal_pending(current))
break;
cond_resched();
}
if (written) {
struct inode *inode = file->f_dentry->d_inode;
inode->i_mtime = current_fs_time(inode->i_sb);
ret = written;
}
up(&tty->atomic_write);
return ret;
}
static ssize_t tty_write(struct file * file, const char __user * buf, size_t count,
loff_t *ppos)
{
struct tty_struct * tty;
struct inode *inode = file->f_dentry->d_inode;
ssize_t ret;
struct tty_ldisc *ld;
tty = (struct tty_struct *)file->private_data;
if (tty_paranoia_check(tty, inode, "tty_write"))
return -EIO;
if (!tty || !tty->driver->write || (test_bit(TTY_IO_ERROR, &tty->flags)))
return -EIO;
ld = tty_ldisc_ref_wait(tty);
if (!ld->write)
ret = -EIO;
else
ret = do_tty_write(ld->write, tty, file, buf, count);
tty_ldisc_deref(ld);
return ret;
}
ssize_t redirected_tty_write(struct file * file, const char __user * buf, size_t count,
loff_t *ppos)
{
struct file *p = NULL;
spin_lock(&redirect_lock);
if (redirect) {
get_file(redirect);
p = redirect;
}
spin_unlock(&redirect_lock);
if (p) {
ssize_t res;
res = vfs_write(p, buf, count, &p->f_pos);
fput(p);
return res;
}
return tty_write(file, buf, count, ppos);
}
static char ptychar[] = "pqrstuvwxyzabcde";
static inline void pty_line_name(struct tty_driver *driver, int index, char *p)
{
int i = index + driver->name_base;
/* ->name is initialized to "ttyp", but "tty" is expected */
sprintf(p, "%s%c%x",
driver->subtype == PTY_TYPE_SLAVE ? "tty" : driver->name,
ptychar[i >> 4 & 0xf], i & 0xf);
}
static inline void tty_line_name(struct tty_driver *driver, int index, char *p)
{
sprintf(p, "%s%d", driver->name, index + driver->name_base);
}
/*
* WSH 06/09/97: Rewritten to remove races and properly clean up after a
* failed open. The new code protects the open with a semaphore, so it's
* really quite straightforward. The semaphore locking can probably be
* relaxed for the (most common) case of reopening a tty.
*/
static int init_dev(struct tty_driver *driver, int idx,
struct tty_struct **ret_tty)
{
struct tty_struct *tty, *o_tty;
struct termios *tp, **tp_loc, *o_tp, **o_tp_loc;
struct termios *ltp, **ltp_loc, *o_ltp, **o_ltp_loc;
int retval=0;
/* check whether we're reopening an existing tty */
if (driver->flags & TTY_DRIVER_DEVPTS_MEM) {
tty = devpts_get_tty(idx);
if (tty && driver->subtype == PTY_TYPE_MASTER)
tty = tty->link;
} else {
tty = driver->ttys[idx];
}
if (tty) goto fast_track;
/*
* First time open is complex, especially for PTY devices.
* This code guarantees that either everything succeeds and the
* TTY is ready for operation, or else the table slots are vacated
* and the allocated memory released. (Except that the termios
* and locked termios may be retained.)
*/
if (!try_module_get(driver->owner)) {
retval = -ENODEV;
goto end_init;
}
o_tty = NULL;
tp = o_tp = NULL;
ltp = o_ltp = NULL;
tty = alloc_tty_struct();
if(!tty)
goto fail_no_mem;
initialize_tty_struct(tty);
tty->driver = driver;
tty->index = idx;
tty_line_name(driver, idx, tty->name);
if (driver->flags & TTY_DRIVER_DEVPTS_MEM) {
tp_loc = &tty->termios;
ltp_loc = &tty->termios_locked;
} else {
tp_loc = &driver->termios[idx];
ltp_loc = &driver->termios_locked[idx];
}
if (!*tp_loc) {
tp = (struct termios *) kmalloc(sizeof(struct termios),
GFP_KERNEL);
if (!tp)
goto free_mem_out;
*tp = driver->init_termios;
}
if (!*ltp_loc) {
ltp = (struct termios *) kmalloc(sizeof(struct termios),
GFP_KERNEL);
if (!ltp)
goto free_mem_out;
memset(ltp, 0, sizeof(struct termios));
}
if (driver->type == TTY_DRIVER_TYPE_PTY) {
o_tty = alloc_tty_struct();
if (!o_tty)
goto free_mem_out;
initialize_tty_struct(o_tty);
o_tty->driver = driver->other;
o_tty->index = idx;
tty_line_name(driver->other, idx, o_tty->name);
if (driver->flags & TTY_DRIVER_DEVPTS_MEM) {
o_tp_loc = &o_tty->termios;
o_ltp_loc = &o_tty->termios_locked;
} else {
o_tp_loc = &driver->other->termios[idx];
o_ltp_loc = &driver->other->termios_locked[idx];
}
if (!*o_tp_loc) {
o_tp = (struct termios *)
kmalloc(sizeof(struct termios), GFP_KERNEL);
if (!o_tp)
goto free_mem_out;
*o_tp = driver->other->init_termios;
}
if (!*o_ltp_loc) {
o_ltp = (struct termios *)
kmalloc(sizeof(struct termios), GFP_KERNEL);
if (!o_ltp)
goto free_mem_out;
memset(o_ltp, 0, sizeof(struct termios));
}
/*
* Everything allocated ... set up the o_tty structure.
*/
if (!(driver->other->flags & TTY_DRIVER_DEVPTS_MEM)) {
driver->other->ttys[idx] = o_tty;
}
if (!*o_tp_loc)
*o_tp_loc = o_tp;
if (!*o_ltp_loc)
*o_ltp_loc = o_ltp;
o_tty->termios = *o_tp_loc;
o_tty->termios_locked = *o_ltp_loc;
driver->other->refcount++;
if (driver->subtype == PTY_TYPE_MASTER)
o_tty->count++;
/* Establish the links in both directions */
tty->link = o_tty;
o_tty->link = tty;
}
/*
* All structures have been allocated, so now we install them.
* Failures after this point use release_mem to clean up, so
* there's no need to null out the local pointers.
*/
if (!(driver->flags & TTY_DRIVER_DEVPTS_MEM)) {
driver->ttys[idx] = tty;
}
if (!*tp_loc)
*tp_loc = tp;
if (!*ltp_loc)
*ltp_loc = ltp;
tty->termios = *tp_loc;
tty->termios_locked = *ltp_loc;
driver->refcount++;
tty->count++;
/*
* Structures all installed ... call the ldisc open routines.
* If we fail here just call release_mem to clean up. No need
* to decrement the use counts, as release_mem doesn't care.
*/
if (tty->ldisc.open) {
retval = (tty->ldisc.open)(tty);
if (retval)
goto release_mem_out;
}
if (o_tty && o_tty->ldisc.open) {
retval = (o_tty->ldisc.open)(o_tty);
if (retval) {
if (tty->ldisc.close)
(tty->ldisc.close)(tty);
goto release_mem_out;
}
tty_ldisc_enable(o_tty);
}
tty_ldisc_enable(tty);
goto success;
/*
* This fast open can be used if the tty is already open.
* No memory is allocated, and the only failures are from
* attempting to open a closing tty or attempting multiple
* opens on a pty master.
*/
fast_track:
if (test_bit(TTY_CLOSING, &tty->flags)) {
retval = -EIO;
goto end_init;
}
if (driver->type == TTY_DRIVER_TYPE_PTY &&
driver->subtype == PTY_TYPE_MASTER) {
/*
* special case for PTY masters: only one open permitted,
* and the slave side open count is incremented as well.
*/
if (tty->count) {
retval = -EIO;
goto end_init;
}
tty->link->count++;
}
tty->count++;
tty->driver = driver; /* N.B. why do this every time?? */
/* FIXME */
if(!test_bit(TTY_LDISC, &tty->flags))
printk(KERN_ERR "init_dev but no ldisc\n");
success:
*ret_tty = tty;
/* All paths come through here to release the semaphore */
end_init:
return retval;
/* Release locally allocated memory ... nothing placed in slots */
free_mem_out:
kfree(o_tp);
if (o_tty)
free_tty_struct(o_tty);
kfree(ltp);
kfree(tp);
free_tty_struct(tty);
fail_no_mem:
module_put(driver->owner);
retval = -ENOMEM;
goto end_init;
/* call the tty release_mem routine to clean out this slot */
release_mem_out:
printk(KERN_INFO "init_dev: ldisc open failed, "
"clearing slot %d\n", idx);
release_mem(tty, idx);
goto end_init;
}
/*
* Releases memory associated with a tty structure, and clears out the
* driver table slots.
*/
static void release_mem(struct tty_struct *tty, int idx)
{
struct tty_struct *o_tty;
struct termios *tp;
int devpts = tty->driver->flags & TTY_DRIVER_DEVPTS_MEM;
if ((o_tty = tty->link) != NULL) {
if (!devpts)
o_tty->driver->ttys[idx] = NULL;
if (o_tty->driver->flags & TTY_DRIVER_RESET_TERMIOS) {
tp = o_tty->termios;
if (!devpts)
o_tty->driver->termios[idx] = NULL;
kfree(tp);
tp = o_tty->termios_locked;
if (!devpts)
o_tty->driver->termios_locked[idx] = NULL;
kfree(tp);
}
o_tty->magic = 0;
o_tty->driver->refcount--;
file_list_lock();
list_del_init(&o_tty->tty_files);
file_list_unlock();
free_tty_struct(o_tty);
}
if (!devpts)
tty->driver->ttys[idx] = NULL;
if (tty->driver->flags & TTY_DRIVER_RESET_TERMIOS) {
tp = tty->termios;
if (!devpts)
tty->driver->termios[idx] = NULL;
kfree(tp);
tp = tty->termios_locked;
if (!devpts)
tty->driver->termios_locked[idx] = NULL;
kfree(tp);
}
tty->magic = 0;
tty->driver->refcount--;
file_list_lock();
list_del_init(&tty->tty_files);
file_list_unlock();
module_put(tty->driver->owner);
free_tty_struct(tty);
}
/*
* Even releasing the tty structures is a tricky business.. We have
* to be very careful that the structures are all released at the
* same time, as interrupts might otherwise get the wrong pointers.
*
* WSH 09/09/97: rewritten to avoid some nasty race conditions that could
* lead to double frees or releasing memory still in use.
*/
static void release_dev(struct file * filp)
{
struct tty_struct *tty, *o_tty;
int pty_master, tty_closing, o_tty_closing, do_sleep;
int devpts_master, devpts;
int idx;
char buf[64];
unsigned long flags;
tty = (struct tty_struct *)filp->private_data;
if (tty_paranoia_check(tty, filp->f_dentry->d_inode, "release_dev"))
return;
check_tty_count(tty, "release_dev");
tty_fasync(-1, filp, 0);
idx = tty->index;
pty_master = (tty->driver->type == TTY_DRIVER_TYPE_PTY &&
tty->driver->subtype == PTY_TYPE_MASTER);
devpts = (tty->driver->flags & TTY_DRIVER_DEVPTS_MEM) != 0;
devpts_master = pty_master && devpts;
o_tty = tty->link;
#ifdef TTY_PARANOIA_CHECK
if (idx < 0 || idx >= tty->driver->num) {
printk(KERN_DEBUG "release_dev: bad idx when trying to "
"free (%s)\n", tty->name);
return;
}
if (!(tty->driver->flags & TTY_DRIVER_DEVPTS_MEM)) {
if (tty != tty->driver->ttys[idx]) {
printk(KERN_DEBUG "release_dev: driver.table[%d] not tty "
"for (%s)\n", idx, tty->name);
return;
}
if (tty->termios != tty->driver->termios[idx]) {
printk(KERN_DEBUG "release_dev: driver.termios[%d] not termios "
"for (%s)\n",
idx, tty->name);
return;
}
if (tty->termios_locked != tty->driver->termios_locked[idx]) {
printk(KERN_DEBUG "release_dev: driver.termios_locked[%d] not "
"termios_locked for (%s)\n",
idx, tty->name);
return;
}
}
#endif
#ifdef TTY_DEBUG_HANGUP
printk(KERN_DEBUG "release_dev of %s (tty count=%d)...",
tty_name(tty, buf), tty->count);
#endif
#ifdef TTY_PARANOIA_CHECK
if (tty->driver->other &&
!(tty->driver->flags & TTY_DRIVER_DEVPTS_MEM)) {
if (o_tty != tty->driver->other->ttys[idx]) {
printk(KERN_DEBUG "release_dev: other->table[%d] "
"not o_tty for (%s)\n",
idx, tty->name);
return;
}
if (o_tty->termios != tty->driver->other->termios[idx]) {
printk(KERN_DEBUG "release_dev: other->termios[%d] "
"not o_termios for (%s)\n",
idx, tty->name);
return;
}
if (o_tty->termios_locked !=
tty->driver->other->termios_locked[idx]) {
printk(KERN_DEBUG "release_dev: other->termios_locked["
"%d] not o_termios_locked for (%s)\n",
idx, tty->name);
return;
}
if (o_tty->link != tty) {
printk(KERN_DEBUG "release_dev: bad pty pointers\n");
return;
}
}
#endif
if (tty->driver->close)
tty->driver->close(tty, filp);
/*
* Sanity check: if tty->count is going to zero, there shouldn't be
* any waiters on tty->read_wait or tty->write_wait. We test the
* wait queues and kick everyone out _before_ actually starting to
* close. This ensures that we won't block while releasing the tty
* structure.
*
* The test for the o_tty closing is necessary, since the master and
* slave sides may close in any order. If the slave side closes out
* first, its count will be one, since the master side holds an open.
* Thus this test wouldn't be triggered at the time the slave closes,
* so we do it now.
*
* Note that it's possible for the tty to be opened again while we're
* flushing out waiters. By recalculating the closing flags before
* each iteration we avoid any problems.
*/
while (1) {
/* Guard against races with tty->count changes elsewhere and
opens on /dev/tty */
down(&tty_sem);
tty_closing = tty->count <= 1;
o_tty_closing = o_tty &&
(o_tty->count <= (pty_master ? 1 : 0));
up(&tty_sem);
do_sleep = 0;
if (tty_closing) {
if (waitqueue_active(&tty->read_wait)) {
wake_up(&tty->read_wait);
do_sleep++;
}
if (waitqueue_active(&tty->write_wait)) {
wake_up(&tty->write_wait);
do_sleep++;
}
}
if (o_tty_closing) {
if (waitqueue_active(&o_tty->read_wait)) {
wake_up(&o_tty->read_wait);
do_sleep++;
}
if (waitqueue_active(&o_tty->write_wait)) {
wake_up(&o_tty->write_wait);
do_sleep++;
}
}
if (!do_sleep)
break;
printk(KERN_WARNING "release_dev: %s: read/write wait queue "
"active!\n", tty_name(tty, buf));
schedule();
}
/*
* The closing flags are now consistent with the open counts on
* both sides, and we've completed the last operation that could
* block, so it's safe to proceed with closing.
*/
down(&tty_sem);
if (pty_master) {
if (--o_tty->count < 0) {
printk(KERN_WARNING "release_dev: bad pty slave count "
"(%d) for %s\n",
o_tty->count, tty_name(o_tty, buf));
o_tty->count = 0;
}
}
if (--tty->count < 0) {
printk(KERN_WARNING "release_dev: bad tty->count (%d) for %s\n",
tty->count, tty_name(tty, buf));
tty->count = 0;
}
up(&tty_sem);
/*
* We've decremented tty->count, so we need to remove this file
* descriptor off the tty->tty_files list; this serves two
* purposes:
* - check_tty_count sees the correct number of file descriptors
* associated with this tty.
* - do_tty_hangup no longer sees this file descriptor as
* something that needs to be handled for hangups.
*/
file_kill(filp);
filp->private_data = NULL;
/*
* Perform some housekeeping before deciding whether to return.
*
* Set the TTY_CLOSING flag if this was the last open. In the
* case of a pty we may have to wait around for the other side
* to close, and TTY_CLOSING makes sure we can't be reopened.
*/
if(tty_closing)
set_bit(TTY_CLOSING, &tty->flags);
if(o_tty_closing)
set_bit(TTY_CLOSING, &o_tty->flags);
/*
* If _either_ side is closing, make sure there aren't any
* processes that still think tty or o_tty is their controlling
* tty.
*/
if (tty_closing || o_tty_closing) {
struct task_struct *p;
read_lock(&tasklist_lock);
do_each_task_pid(tty->session, PIDTYPE_SID, p) {
p->signal->tty = NULL;
} while_each_task_pid(tty->session, PIDTYPE_SID, p);
if (o_tty)
do_each_task_pid(o_tty->session, PIDTYPE_SID, p) {
p->signal->tty = NULL;
} while_each_task_pid(o_tty->session, PIDTYPE_SID, p);
read_unlock(&tasklist_lock);
}
/* check whether both sides are closing ... */
if (!tty_closing || (o_tty && !o_tty_closing))
return;
#ifdef TTY_DEBUG_HANGUP
printk(KERN_DEBUG "freeing tty structure...");
#endif
/*
* Prevent flush_to_ldisc() from rescheduling the work for later. Then
* kill any delayed work. As this is the final close it does not
* race with the set_ldisc code path.
*/
clear_bit(TTY_LDISC, &tty->flags);
clear_bit(TTY_DONT_FLIP, &tty->flags);
[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 12:54:13 +08:00
cancel_delayed_work(&tty->buf.work);
/*
[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 12:54:13 +08:00
* Wait for ->hangup_work and ->buf.work handlers to terminate
*/
flush_scheduled_work();
/*
* Wait for any short term users (we know they are just driver
* side waiters as the file is closing so user count on the file
* side is zero.
*/
spin_lock_irqsave(&tty_ldisc_lock, flags);
while(tty->ldisc.refcount)
{
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
wait_event(tty_ldisc_wait, tty->ldisc.refcount == 0);
spin_lock_irqsave(&tty_ldisc_lock, flags);
}
spin_unlock_irqrestore(&tty_ldisc_lock, flags);
/*
* Shutdown the current line discipline, and reset it to N_TTY.
* N.B. why reset ldisc when we're releasing the memory??
*
* FIXME: this MUST get fixed for the new reflocking
*/
if (tty->ldisc.close)
(tty->ldisc.close)(tty);
tty_ldisc_put(tty->ldisc.num);
/*
* Switch the line discipline back
*/
tty_ldisc_assign(tty, tty_ldisc_get(N_TTY));
tty_set_termios_ldisc(tty,N_TTY);
if (o_tty) {
/* FIXME: could o_tty be in setldisc here ? */
clear_bit(TTY_LDISC, &o_tty->flags);
if (o_tty->ldisc.close)
(o_tty->ldisc.close)(o_tty);
tty_ldisc_put(o_tty->ldisc.num);
tty_ldisc_assign(o_tty, tty_ldisc_get(N_TTY));
tty_set_termios_ldisc(o_tty,N_TTY);
}
/*
* The release_mem function takes care of the details of clearing
* the slots and preserving the termios structure.
*/
release_mem(tty, idx);
#ifdef CONFIG_UNIX98_PTYS
/* Make this pty number available for reallocation */
if (devpts) {
down(&allocated_ptys_lock);
idr_remove(&allocated_ptys, idx);
up(&allocated_ptys_lock);
}
#endif
}
/*
* tty_open and tty_release keep up the tty count that contains the
* number of opens done on a tty. We cannot use the inode-count, as
* different inodes might point to the same tty.
*
* Open-counting is needed for pty masters, as well as for keeping
* track of serial lines: DTR is dropped when the last close happens.
* (This is not done solely through tty->count, now. - Ted 1/27/92)
*
* The termios state of a pty is reset on first open so that
* settings don't persist across reuse.
*/
static int tty_open(struct inode * inode, struct file * filp)
{
struct tty_struct *tty;
int noctty, retval;
struct tty_driver *driver;
int index;
dev_t device = inode->i_rdev;
unsigned short saved_flags = filp->f_flags;
nonseekable_open(inode, filp);
retry_open:
noctty = filp->f_flags & O_NOCTTY;
index = -1;
retval = 0;
down(&tty_sem);
if (device == MKDEV(TTYAUX_MAJOR,0)) {
if (!current->signal->tty) {
up(&tty_sem);
return -ENXIO;
}
driver = current->signal->tty->driver;
index = current->signal->tty->index;
filp->f_flags |= O_NONBLOCK; /* Don't let /dev/tty block */
/* noctty = 1; */
goto got_driver;
}
#ifdef CONFIG_VT
if (device == MKDEV(TTY_MAJOR,0)) {
extern struct tty_driver *console_driver;
driver = console_driver;
index = fg_console;
noctty = 1;
goto got_driver;
}
#endif
if (device == MKDEV(TTYAUX_MAJOR,1)) {
driver = console_device(&index);
if (driver) {
/* Don't let /dev/console block */
filp->f_flags |= O_NONBLOCK;
noctty = 1;
goto got_driver;
}
up(&tty_sem);
return -ENODEV;
}
driver = get_tty_driver(device, &index);
if (!driver) {
up(&tty_sem);
return -ENODEV;
}
got_driver:
retval = init_dev(driver, index, &tty);
up(&tty_sem);
if (retval)
return retval;
filp->private_data = tty;
file_move(filp, &tty->tty_files);
check_tty_count(tty, "tty_open");
if (tty->driver->type == TTY_DRIVER_TYPE_PTY &&
tty->driver->subtype == PTY_TYPE_MASTER)
noctty = 1;
#ifdef TTY_DEBUG_HANGUP
printk(KERN_DEBUG "opening %s...", tty->name);
#endif
if (!retval) {
if (tty->driver->open)
retval = tty->driver->open(tty, filp);
else
retval = -ENODEV;
}
filp->f_flags = saved_flags;
if (!retval && test_bit(TTY_EXCLUSIVE, &tty->flags) && !capable(CAP_SYS_ADMIN))
retval = -EBUSY;
if (retval) {
#ifdef TTY_DEBUG_HANGUP
printk(KERN_DEBUG "error %d in opening %s...", retval,
tty->name);
#endif
release_dev(filp);
if (retval != -ERESTARTSYS)
return retval;
if (signal_pending(current))
return retval;
schedule();
/*
* Need to reset f_op in case a hangup happened.
*/
if (filp->f_op == &hung_up_tty_fops)
filp->f_op = &tty_fops;
goto retry_open;
}
if (!noctty &&
current->signal->leader &&
!current->signal->tty &&
tty->session == 0) {
task_lock(current);
current->signal->tty = tty;
task_unlock(current);
current->signal->tty_old_pgrp = 0;
tty->session = current->signal->session;
tty->pgrp = process_group(current);
}
return 0;
}
#ifdef CONFIG_UNIX98_PTYS
static int ptmx_open(struct inode * inode, struct file * filp)
{
struct tty_struct *tty;
int retval;
int index;
int idr_ret;
nonseekable_open(inode, filp);
/* find a device that is not in use. */
down(&allocated_ptys_lock);
if (!idr_pre_get(&allocated_ptys, GFP_KERNEL)) {
up(&allocated_ptys_lock);
return -ENOMEM;
}
idr_ret = idr_get_new(&allocated_ptys, NULL, &index);
if (idr_ret < 0) {
up(&allocated_ptys_lock);
if (idr_ret == -EAGAIN)
return -ENOMEM;
return -EIO;
}
if (index >= pty_limit) {
idr_remove(&allocated_ptys, index);
up(&allocated_ptys_lock);
return -EIO;
}
up(&allocated_ptys_lock);
down(&tty_sem);
retval = init_dev(ptm_driver, index, &tty);
up(&tty_sem);
if (retval)
goto out;
set_bit(TTY_PTY_LOCK, &tty->flags); /* LOCK THE SLAVE */
filp->private_data = tty;
file_move(filp, &tty->tty_files);
retval = -ENOMEM;
if (devpts_pty_new(tty->link))
goto out1;
check_tty_count(tty, "tty_open");
retval = ptm_driver->open(tty, filp);
if (!retval)
return 0;
out1:
release_dev(filp);
out:
down(&allocated_ptys_lock);
idr_remove(&allocated_ptys, index);
up(&allocated_ptys_lock);
return retval;
}
#endif
static int tty_release(struct inode * inode, struct file * filp)
{
lock_kernel();
release_dev(filp);
unlock_kernel();
return 0;
}
/* No kernel lock held - fine */
static unsigned int tty_poll(struct file * filp, poll_table * wait)
{
struct tty_struct * tty;
struct tty_ldisc *ld;
int ret = 0;
tty = (struct tty_struct *)filp->private_data;
if (tty_paranoia_check(tty, filp->f_dentry->d_inode, "tty_poll"))
return 0;
ld = tty_ldisc_ref_wait(tty);
if (ld->poll)
ret = (ld->poll)(tty, filp, wait);
tty_ldisc_deref(ld);
return ret;
}
static int tty_fasync(int fd, struct file * filp, int on)
{
struct tty_struct * tty;
int retval;
tty = (struct tty_struct *)filp->private_data;
if (tty_paranoia_check(tty, filp->f_dentry->d_inode, "tty_fasync"))
return 0;
retval = fasync_helper(fd, filp, on, &tty->fasync);
if (retval <= 0)
return retval;
if (on) {
if (!waitqueue_active(&tty->read_wait))
tty->minimum_to_wake = 1;
retval = f_setown(filp, (-tty->pgrp) ? : current->pid, 0);
if (retval)
return retval;
} else {
if (!tty->fasync && !waitqueue_active(&tty->read_wait))
tty->minimum_to_wake = N_TTY_BUF_SIZE;
}
return 0;
}
static int tiocsti(struct tty_struct *tty, char __user *p)
{
char ch, mbz = 0;
struct tty_ldisc *ld;
if ((current->signal->tty != tty) && !capable(CAP_SYS_ADMIN))
return -EPERM;
if (get_user(ch, p))
return -EFAULT;
ld = tty_ldisc_ref_wait(tty);
ld->receive_buf(tty, &ch, &mbz, 1);
tty_ldisc_deref(ld);
return 0;
}
static int tiocgwinsz(struct tty_struct *tty, struct winsize __user * arg)
{
if (copy_to_user(arg, &tty->winsize, sizeof(*arg)))
return -EFAULT;
return 0;
}
static int tiocswinsz(struct tty_struct *tty, struct tty_struct *real_tty,
struct winsize __user * arg)
{
struct winsize tmp_ws;
if (copy_from_user(&tmp_ws, arg, sizeof(*arg)))
return -EFAULT;
if (!memcmp(&tmp_ws, &tty->winsize, sizeof(*arg)))
return 0;
#ifdef CONFIG_VT
if (tty->driver->type == TTY_DRIVER_TYPE_CONSOLE) {
int rc;
acquire_console_sem();
rc = vc_resize(tty->driver_data, tmp_ws.ws_col, tmp_ws.ws_row);
release_console_sem();
if (rc)
return -ENXIO;
}
#endif
if (tty->pgrp > 0)
kill_pg(tty->pgrp, SIGWINCH, 1);
if ((real_tty->pgrp != tty->pgrp) && (real_tty->pgrp > 0))
kill_pg(real_tty->pgrp, SIGWINCH, 1);
tty->winsize = tmp_ws;
real_tty->winsize = tmp_ws;
return 0;
}
static int tioccons(struct file *file)
{
if (!capable(CAP_SYS_ADMIN))
return -EPERM;
if (file->f_op->write == redirected_tty_write) {
struct file *f;
spin_lock(&redirect_lock);
f = redirect;
redirect = NULL;
spin_unlock(&redirect_lock);
if (f)
fput(f);
return 0;
}
spin_lock(&redirect_lock);
if (redirect) {
spin_unlock(&redirect_lock);
return -EBUSY;
}
get_file(file);
redirect = file;
spin_unlock(&redirect_lock);
return 0;
}
static int fionbio(struct file *file, int __user *p)
{
int nonblock;
if (get_user(nonblock, p))
return -EFAULT;
if (nonblock)
file->f_flags |= O_NONBLOCK;
else
file->f_flags &= ~O_NONBLOCK;
return 0;
}
static int tiocsctty(struct tty_struct *tty, int arg)
{
task_t *p;
if (current->signal->leader &&
(current->signal->session == tty->session))
return 0;
/*
* The process must be a session leader and
* not have a controlling tty already.
*/
if (!current->signal->leader || current->signal->tty)
return -EPERM;
if (tty->session > 0) {
/*
* This tty is already the controlling
* tty for another session group!
*/
if ((arg == 1) && capable(CAP_SYS_ADMIN)) {
/*
* Steal it away
*/
read_lock(&tasklist_lock);
do_each_task_pid(tty->session, PIDTYPE_SID, p) {
p->signal->tty = NULL;
} while_each_task_pid(tty->session, PIDTYPE_SID, p);
read_unlock(&tasklist_lock);
} else
return -EPERM;
}
task_lock(current);
current->signal->tty = tty;
task_unlock(current);
current->signal->tty_old_pgrp = 0;
tty->session = current->signal->session;
tty->pgrp = process_group(current);
return 0;
}
static int tiocgpgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p)
{
/*
* (tty == real_tty) is a cheap way of
* testing if the tty is NOT a master pty.
*/
if (tty == real_tty && current->signal->tty != real_tty)
return -ENOTTY;
return put_user(real_tty->pgrp, p);
}
static int tiocspgrp(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p)
{
pid_t pgrp;
int retval = tty_check_change(real_tty);
if (retval == -EIO)
return -ENOTTY;
if (retval)
return retval;
if (!current->signal->tty ||
(current->signal->tty != real_tty) ||
(real_tty->session != current->signal->session))
return -ENOTTY;
if (get_user(pgrp, p))
return -EFAULT;
if (pgrp < 0)
return -EINVAL;
if (session_of_pgrp(pgrp) != current->signal->session)
return -EPERM;
real_tty->pgrp = pgrp;
return 0;
}
static int tiocgsid(struct tty_struct *tty, struct tty_struct *real_tty, pid_t __user *p)
{
/*
* (tty == real_tty) is a cheap way of
* testing if the tty is NOT a master pty.
*/
if (tty == real_tty && current->signal->tty != real_tty)
return -ENOTTY;
if (real_tty->session <= 0)
return -ENOTTY;
return put_user(real_tty->session, p);
}
static int tiocsetd(struct tty_struct *tty, int __user *p)
{
int ldisc;
if (get_user(ldisc, p))
return -EFAULT;
return tty_set_ldisc(tty, ldisc);
}
static int send_break(struct tty_struct *tty, unsigned int duration)
{
tty->driver->break_ctl(tty, -1);
if (!signal_pending(current)) {
msleep_interruptible(duration);
}
tty->driver->break_ctl(tty, 0);
if (signal_pending(current))
return -EINTR;
return 0;
}
static int
tty_tiocmget(struct tty_struct *tty, struct file *file, int __user *p)
{
int retval = -EINVAL;
if (tty->driver->tiocmget) {
retval = tty->driver->tiocmget(tty, file);
if (retval >= 0)
retval = put_user(retval, p);
}
return retval;
}
static int
tty_tiocmset(struct tty_struct *tty, struct file *file, unsigned int cmd,
unsigned __user *p)
{
int retval = -EINVAL;
if (tty->driver->tiocmset) {
unsigned int set, clear, val;
retval = get_user(val, p);
if (retval)
return retval;
set = clear = 0;
switch (cmd) {
case TIOCMBIS:
set = val;
break;
case TIOCMBIC:
clear = val;
break;
case TIOCMSET:
set = val;
clear = ~val;
break;
}
set &= TIOCM_DTR|TIOCM_RTS|TIOCM_OUT1|TIOCM_OUT2|TIOCM_LOOP;
clear &= TIOCM_DTR|TIOCM_RTS|TIOCM_OUT1|TIOCM_OUT2|TIOCM_LOOP;
retval = tty->driver->tiocmset(tty, file, set, clear);
}
return retval;
}
/*
* Split this up, as gcc can choke on it otherwise..
*/
int tty_ioctl(struct inode * inode, struct file * file,
unsigned int cmd, unsigned long arg)
{
struct tty_struct *tty, *real_tty;
void __user *p = (void __user *)arg;
int retval;
struct tty_ldisc *ld;
tty = (struct tty_struct *)file->private_data;
if (tty_paranoia_check(tty, inode, "tty_ioctl"))
return -EINVAL;
real_tty = tty;
if (tty->driver->type == TTY_DRIVER_TYPE_PTY &&
tty->driver->subtype == PTY_TYPE_MASTER)
real_tty = tty->link;
/*
* Break handling by driver
*/
if (!tty->driver->break_ctl) {
switch(cmd) {
case TIOCSBRK:
case TIOCCBRK:
if (tty->driver->ioctl)
return tty->driver->ioctl(tty, file, cmd, arg);
return -EINVAL;
/* These two ioctl's always return success; even if */
/* the driver doesn't support them. */
case TCSBRK:
case TCSBRKP:
if (!tty->driver->ioctl)
return 0;
retval = tty->driver->ioctl(tty, file, cmd, arg);
if (retval == -ENOIOCTLCMD)
retval = 0;
return retval;
}
}
/*
* Factor out some common prep work
*/
switch (cmd) {
case TIOCSETD:
case TIOCSBRK:
case TIOCCBRK:
case TCSBRK:
case TCSBRKP:
retval = tty_check_change(tty);
if (retval)
return retval;
if (cmd != TIOCCBRK) {
tty_wait_until_sent(tty, 0);
if (signal_pending(current))
return -EINTR;
}
break;
}
switch (cmd) {
case TIOCSTI:
return tiocsti(tty, p);
case TIOCGWINSZ:
return tiocgwinsz(tty, p);
case TIOCSWINSZ:
return tiocswinsz(tty, real_tty, p);
case TIOCCONS:
return real_tty!=tty ? -EINVAL : tioccons(file);
case FIONBIO:
return fionbio(file, p);
case TIOCEXCL:
set_bit(TTY_EXCLUSIVE, &tty->flags);
return 0;
case TIOCNXCL:
clear_bit(TTY_EXCLUSIVE, &tty->flags);
return 0;
case TIOCNOTTY:
if (current->signal->tty != tty)
return -ENOTTY;
if (current->signal->leader)
disassociate_ctty(0);
task_lock(current);
current->signal->tty = NULL;
task_unlock(current);
return 0;
case TIOCSCTTY:
return tiocsctty(tty, arg);
case TIOCGPGRP:
return tiocgpgrp(tty, real_tty, p);
case TIOCSPGRP:
return tiocspgrp(tty, real_tty, p);
case TIOCGSID:
return tiocgsid(tty, real_tty, p);
case TIOCGETD:
/* FIXME: check this is ok */
return put_user(tty->ldisc.num, (int __user *)p);
case TIOCSETD:
return tiocsetd(tty, p);
#ifdef CONFIG_VT
case TIOCLINUX:
return tioclinux(tty, arg);
#endif
/*
* Break handling
*/
case TIOCSBRK: /* Turn break on, unconditionally */
tty->driver->break_ctl(tty, -1);
return 0;
case TIOCCBRK: /* Turn break off, unconditionally */
tty->driver->break_ctl(tty, 0);
return 0;
case TCSBRK: /* SVID version: non-zero arg --> no break */
/*
* XXX is the above comment correct, or the
* code below correct? Is this ioctl used at
* all by anyone?
*/
if (!arg)
return send_break(tty, 250);
return 0;
case TCSBRKP: /* support for POSIX tcsendbreak() */
return send_break(tty, arg ? arg*100 : 250);
case TIOCMGET:
return tty_tiocmget(tty, file, p);
case TIOCMSET:
case TIOCMBIC:
case TIOCMBIS:
return tty_tiocmset(tty, file, cmd, p);
}
if (tty->driver->ioctl) {
retval = (tty->driver->ioctl)(tty, file, cmd, arg);
if (retval != -ENOIOCTLCMD)
return retval;
}
ld = tty_ldisc_ref_wait(tty);
retval = -EINVAL;
if (ld->ioctl) {
retval = ld->ioctl(tty, file, cmd, arg);
if (retval == -ENOIOCTLCMD)
retval = -EINVAL;
}
tty_ldisc_deref(ld);
return retval;
}
/*
* This implements the "Secure Attention Key" --- the idea is to
* prevent trojan horses by killing all processes associated with this
* tty when the user hits the "Secure Attention Key". Required for
* super-paranoid applications --- see the Orange Book for more details.
*
* This code could be nicer; ideally it should send a HUP, wait a few
* seconds, then send a INT, and then a KILL signal. But you then
* have to coordinate with the init process, since all processes associated
* with the current tty must be dead before the new getty is allowed
* to spawn.
*
* Now, if it would be correct ;-/ The current code has a nasty hole -
* it doesn't catch files in flight. We may send the descriptor to ourselves
* via AF_UNIX socket, close it and later fetch from socket. FIXME.
*
* Nasty bug: do_SAK is being called in interrupt context. This can
* deadlock. We punt it up to process context. AKPM - 16Mar2001
*/
static void __do_SAK(void *arg)
{
#ifdef TTY_SOFT_SAK
tty_hangup(tty);
#else
struct tty_struct *tty = arg;
struct task_struct *p;
int session;
int i;
struct file *filp;
struct tty_ldisc *disc;
struct fdtable *fdt;
if (!tty)
return;
session = tty->session;
/* We don't want an ldisc switch during this */
disc = tty_ldisc_ref(tty);
if (disc && disc->flush_buffer)
disc->flush_buffer(tty);
tty_ldisc_deref(disc);
if (tty->driver->flush_buffer)
tty->driver->flush_buffer(tty);
read_lock(&tasklist_lock);
do_each_task_pid(session, PIDTYPE_SID, p) {
if (p->signal->tty == tty || session > 0) {
printk(KERN_NOTICE "SAK: killed process %d"
" (%s): p->signal->session==tty->session\n",
p->pid, p->comm);
send_sig(SIGKILL, p, 1);
continue;
}
task_lock(p);
if (p->files) {
rcu_read_lock();
fdt = files_fdtable(p->files);
for (i=0; i < fdt->max_fds; i++) {
filp = fcheck_files(p->files, i);
if (!filp)
continue;
if (filp->f_op->read == tty_read &&
filp->private_data == tty) {
printk(KERN_NOTICE "SAK: killed process %d"
" (%s): fd#%d opened to the tty\n",
p->pid, p->comm, i);
send_sig(SIGKILL, p, 1);
break;
}
}
rcu_read_unlock();
}
task_unlock(p);
} while_each_task_pid(session, PIDTYPE_SID, p);
read_unlock(&tasklist_lock);
#endif
}
/*
* The tq handling here is a little racy - tty->SAK_work may already be queued.
* Fortunately we don't need to worry, because if ->SAK_work is already queued,
* the values which we write to it will be identical to the values which it
* already has. --akpm
*/
void do_SAK(struct tty_struct *tty)
{
if (!tty)
return;
PREPARE_WORK(&tty->SAK_work, __do_SAK, tty);
schedule_work(&tty->SAK_work);
}
EXPORT_SYMBOL(do_SAK);
/*
* This routine is called out of the software interrupt to flush data
[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 12:54:13 +08:00
* from the buffer chain to the line discipline.
*/
static void flush_to_ldisc(void *private_)
{
struct tty_struct *tty = (struct tty_struct *) private_;
unsigned long flags;
struct tty_ldisc *disc;
[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 12:54:13 +08:00
struct tty_buffer *tbuf;
disc = tty_ldisc_ref(tty);
if (disc == NULL) /* !TTY_LDISC */
return;
if (test_bit(TTY_DONT_FLIP, &tty->flags)) {
/*
* Do it after the next timer tick:
*/
[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 12:54:13 +08:00
schedule_delayed_work(&tty->buf.work, 1);
goto out;
}
spin_lock_irqsave(&tty->read_lock, flags);
[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 12:54:13 +08:00
while((tbuf = tty->buf.head) != NULL) {
tty->buf.head = tbuf->next;
if (tty->buf.head == NULL)
tty->buf.tail = 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 12:54:13 +08:00
spin_unlock_irqrestore(&tty->read_lock, flags);
/* printk("Process buffer %p for %d\n", tbuf, tbuf->used); */
disc->receive_buf(tty, tbuf->char_buf_ptr,
tbuf->flag_buf_ptr,
tbuf->used);
spin_lock_irqsave(&tty->read_lock, flags);
tty_buffer_free(tty, tbuf);
}
spin_unlock_irqrestore(&tty->read_lock, flags);
out:
tty_ldisc_deref(disc);
}
/*
* Routine which returns the baud rate of the tty
*
* Note that the baud_table needs to be kept in sync with the
* include/asm/termbits.h file.
*/
static int baud_table[] = {
0, 50, 75, 110, 134, 150, 200, 300, 600, 1200, 1800, 2400, 4800,
9600, 19200, 38400, 57600, 115200, 230400, 460800,
#ifdef __sparc__
76800, 153600, 307200, 614400, 921600
#else
500000, 576000, 921600, 1000000, 1152000, 1500000, 2000000,
2500000, 3000000, 3500000, 4000000
#endif
};
static int n_baud_table = ARRAY_SIZE(baud_table);
/**
* tty_termios_baud_rate
* @termios: termios structure
*
* Convert termios baud rate data into a speed. This should be called
* with the termios lock held if this termios is a terminal termios
* structure. May change the termios data.
*/
int tty_termios_baud_rate(struct termios *termios)
{
unsigned int cbaud;
cbaud = termios->c_cflag & CBAUD;
if (cbaud & CBAUDEX) {
cbaud &= ~CBAUDEX;
if (cbaud < 1 || cbaud + 15 > n_baud_table)
termios->c_cflag &= ~CBAUDEX;
else
cbaud += 15;
}
return baud_table[cbaud];
}
EXPORT_SYMBOL(tty_termios_baud_rate);
/**
* tty_get_baud_rate - get tty bit rates
* @tty: tty to query
*
* Returns the baud rate as an integer for this terminal. The
* termios lock must be held by the caller and the terminal bit
* flags may be updated.
*/
int tty_get_baud_rate(struct tty_struct *tty)
{
int baud = tty_termios_baud_rate(tty->termios);
if (baud == 38400 && tty->alt_speed) {
if (!tty->warned) {
printk(KERN_WARNING "Use of setserial/setrocket to "
"set SPD_* flags is deprecated\n");
tty->warned = 1;
}
baud = tty->alt_speed;
}
return baud;
}
EXPORT_SYMBOL(tty_get_baud_rate);
/**
* tty_flip_buffer_push - terminal
* @tty: tty to push
*
* Queue a push of the terminal flip buffers to the line discipline. This
* function must not be called from IRQ context if tty->low_latency is set.
*
* In the event of the queue being busy for flipping the work will be
* held off and retried later.
*/
void tty_flip_buffer_push(struct tty_struct *tty)
{
if (tty->low_latency)
flush_to_ldisc((void *) tty);
else
[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 12:54:13 +08:00
schedule_delayed_work(&tty->buf.work, 1);
}
EXPORT_SYMBOL(tty_flip_buffer_push);
[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 12:54:13 +08:00
/*
* This subroutine initializes a tty structure.
*/
static void initialize_tty_struct(struct tty_struct *tty)
{
memset(tty, 0, sizeof(struct tty_struct));
tty->magic = TTY_MAGIC;
tty_ldisc_assign(tty, tty_ldisc_get(N_TTY));
tty->pgrp = -1;
tty->overrun_time = jiffies;
[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 12:54:13 +08:00
tty->buf.head = tty->buf.tail = NULL;
tty_buffer_init(tty);
INIT_WORK(&tty->buf.work, flush_to_ldisc, tty);
init_MUTEX(&tty->buf.pty_sem);
init_MUTEX(&tty->termios_sem);
init_waitqueue_head(&tty->write_wait);
init_waitqueue_head(&tty->read_wait);
INIT_WORK(&tty->hangup_work, do_tty_hangup, tty);
sema_init(&tty->atomic_read, 1);
sema_init(&tty->atomic_write, 1);
spin_lock_init(&tty->read_lock);
INIT_LIST_HEAD(&tty->tty_files);
INIT_WORK(&tty->SAK_work, NULL, NULL);
}
/*
* The default put_char routine if the driver did not define one.
*/
static void tty_default_put_char(struct tty_struct *tty, unsigned char ch)
{
tty->driver->write(tty, &ch, 1);
}
static struct class *tty_class;
/**
* tty_register_device - register a tty device
* @driver: the tty driver that describes the tty device
* @index: the index in the tty driver for this tty device
* @device: a struct device that is associated with this tty device.
* This field is optional, if there is no known struct device for this
* tty device it can be set to NULL safely.
*
* This call is required to be made to register an individual tty device if
* the tty driver's flags have the TTY_DRIVER_NO_DEVFS bit set. If that
* bit is not set, this function should not be called.
*/
void tty_register_device(struct tty_driver *driver, unsigned index,
struct device *device)
{
char name[64];
dev_t dev = MKDEV(driver->major, driver->minor_start) + index;
if (index >= driver->num) {
printk(KERN_ERR "Attempt to register invalid tty line number "
" (%d).\n", index);
return;
}
devfs_mk_cdev(dev, S_IFCHR | S_IRUSR | S_IWUSR,
"%s%d", driver->devfs_name, index + driver->name_base);
if (driver->type == TTY_DRIVER_TYPE_PTY)
pty_line_name(driver, index, name);
else
tty_line_name(driver, index, name);
class_device_create(tty_class, NULL, dev, device, "%s", name);
}
/**
* tty_unregister_device - unregister a tty device
* @driver: the tty driver that describes the tty device
* @index: the index in the tty driver for this tty device
*
* If a tty device is registered with a call to tty_register_device() then
* this function must be made when the tty device is gone.
*/
void tty_unregister_device(struct tty_driver *driver, unsigned index)
{
devfs_remove("%s%d", driver->devfs_name, index + driver->name_base);
class_device_destroy(tty_class, MKDEV(driver->major, driver->minor_start) + index);
}
EXPORT_SYMBOL(tty_register_device);
EXPORT_SYMBOL(tty_unregister_device);
struct tty_driver *alloc_tty_driver(int lines)
{
struct tty_driver *driver;
driver = kmalloc(sizeof(struct tty_driver), GFP_KERNEL);
if (driver) {
memset(driver, 0, sizeof(struct tty_driver));
driver->magic = TTY_DRIVER_MAGIC;
driver->num = lines;
/* later we'll move allocation of tables here */
}
return driver;
}
void put_tty_driver(struct tty_driver *driver)
{
kfree(driver);
}
void tty_set_operations(struct tty_driver *driver, struct tty_operations *op)
{
driver->open = op->open;
driver->close = op->close;
driver->write = op->write;
driver->put_char = op->put_char;
driver->flush_chars = op->flush_chars;
driver->write_room = op->write_room;
driver->chars_in_buffer = op->chars_in_buffer;
driver->ioctl = op->ioctl;
driver->set_termios = op->set_termios;
driver->throttle = op->throttle;
driver->unthrottle = op->unthrottle;
driver->stop = op->stop;
driver->start = op->start;
driver->hangup = op->hangup;
driver->break_ctl = op->break_ctl;
driver->flush_buffer = op->flush_buffer;
driver->set_ldisc = op->set_ldisc;
driver->wait_until_sent = op->wait_until_sent;
driver->send_xchar = op->send_xchar;
driver->read_proc = op->read_proc;
driver->write_proc = op->write_proc;
driver->tiocmget = op->tiocmget;
driver->tiocmset = op->tiocmset;
}
EXPORT_SYMBOL(alloc_tty_driver);
EXPORT_SYMBOL(put_tty_driver);
EXPORT_SYMBOL(tty_set_operations);
/*
* Called by a tty driver to register itself.
*/
int tty_register_driver(struct tty_driver *driver)
{
int error;
int i;
dev_t dev;
void **p = NULL;
if (driver->flags & TTY_DRIVER_INSTALLED)
return 0;
if (!(driver->flags & TTY_DRIVER_DEVPTS_MEM)) {
p = kmalloc(driver->num * 3 * sizeof(void *), GFP_KERNEL);
if (!p)
return -ENOMEM;
memset(p, 0, driver->num * 3 * sizeof(void *));
}
if (!driver->major) {
error = alloc_chrdev_region(&dev, driver->minor_start, driver->num,
(char*)driver->name);
if (!error) {
driver->major = MAJOR(dev);
driver->minor_start = MINOR(dev);
}
} else {
dev = MKDEV(driver->major, driver->minor_start);
error = register_chrdev_region(dev, driver->num,
(char*)driver->name);
}
if (error < 0) {
kfree(p);
return error;
}
if (p) {
driver->ttys = (struct tty_struct **)p;
driver->termios = (struct termios **)(p + driver->num);
driver->termios_locked = (struct termios **)(p + driver->num * 2);
} else {
driver->ttys = NULL;
driver->termios = NULL;
driver->termios_locked = NULL;
}
cdev_init(&driver->cdev, &tty_fops);
driver->cdev.owner = driver->owner;
error = cdev_add(&driver->cdev, dev, driver->num);
if (error) {
cdev_del(&driver->cdev);
unregister_chrdev_region(dev, driver->num);
driver->ttys = NULL;
driver->termios = driver->termios_locked = NULL;
kfree(p);
return error;
}
if (!driver->put_char)
driver->put_char = tty_default_put_char;
list_add(&driver->tty_drivers, &tty_drivers);
if ( !(driver->flags & TTY_DRIVER_NO_DEVFS) ) {
for(i = 0; i < driver->num; i++)
tty_register_device(driver, i, NULL);
}
proc_tty_register_driver(driver);
return 0;
}
EXPORT_SYMBOL(tty_register_driver);
/*
* Called by a tty driver to unregister itself.
*/
int tty_unregister_driver(struct tty_driver *driver)
{
int i;
struct termios *tp;
void *p;
if (driver->refcount)
return -EBUSY;
unregister_chrdev_region(MKDEV(driver->major, driver->minor_start),
driver->num);
list_del(&driver->tty_drivers);
/*
* Free the termios and termios_locked structures because
* we don't want to get memory leaks when modular tty
* drivers are removed from the kernel.
*/
for (i = 0; i < driver->num; i++) {
tp = driver->termios[i];
if (tp) {
driver->termios[i] = NULL;
kfree(tp);
}
tp = driver->termios_locked[i];
if (tp) {
driver->termios_locked[i] = NULL;
kfree(tp);
}
if (!(driver->flags & TTY_DRIVER_NO_DEVFS))
tty_unregister_device(driver, i);
}
p = driver->ttys;
proc_tty_unregister_driver(driver);
driver->ttys = NULL;
driver->termios = driver->termios_locked = NULL;
kfree(p);
cdev_del(&driver->cdev);
return 0;
}
EXPORT_SYMBOL(tty_unregister_driver);
/*
* Initialize the console device. This is called *early*, so
* we can't necessarily depend on lots of kernel help here.
* Just do some early initializations, and do the complex setup
* later.
*/
void __init console_init(void)
{
initcall_t *call;
/* Setup the default TTY line discipline. */
(void) tty_register_ldisc(N_TTY, &tty_ldisc_N_TTY);
/*
* set up the console device so that later boot sequences can
* inform about problems etc..
*/
#ifdef CONFIG_EARLY_PRINTK
disable_early_printk();
#endif
call = __con_initcall_start;
while (call < __con_initcall_end) {
(*call)();
call++;
}
}
#ifdef CONFIG_VT
extern int vty_init(void);
#endif
static int __init tty_class_init(void)
{
tty_class = class_create(THIS_MODULE, "tty");
if (IS_ERR(tty_class))
return PTR_ERR(tty_class);
return 0;
}
postcore_initcall(tty_class_init);
/* 3/2004 jmc: why do these devices exist? */
static struct cdev tty_cdev, console_cdev;
#ifdef CONFIG_UNIX98_PTYS
static struct cdev ptmx_cdev;
#endif
#ifdef CONFIG_VT
static struct cdev vc0_cdev;
#endif
/*
* Ok, now we can initialize the rest of the tty devices and can count
* on memory allocations, interrupts etc..
*/
static int __init tty_init(void)
{
cdev_init(&tty_cdev, &tty_fops);
if (cdev_add(&tty_cdev, MKDEV(TTYAUX_MAJOR, 0), 1) ||
register_chrdev_region(MKDEV(TTYAUX_MAJOR, 0), 1, "/dev/tty") < 0)
panic("Couldn't register /dev/tty driver\n");
devfs_mk_cdev(MKDEV(TTYAUX_MAJOR, 0), S_IFCHR|S_IRUGO|S_IWUGO, "tty");
class_device_create(tty_class, NULL, MKDEV(TTYAUX_MAJOR, 0), NULL, "tty");
cdev_init(&console_cdev, &console_fops);
if (cdev_add(&console_cdev, MKDEV(TTYAUX_MAJOR, 1), 1) ||
register_chrdev_region(MKDEV(TTYAUX_MAJOR, 1), 1, "/dev/console") < 0)
panic("Couldn't register /dev/console driver\n");
devfs_mk_cdev(MKDEV(TTYAUX_MAJOR, 1), S_IFCHR|S_IRUSR|S_IWUSR, "console");
class_device_create(tty_class, NULL, MKDEV(TTYAUX_MAJOR, 1), NULL, "console");
#ifdef CONFIG_UNIX98_PTYS
cdev_init(&ptmx_cdev, &ptmx_fops);
if (cdev_add(&ptmx_cdev, MKDEV(TTYAUX_MAJOR, 2), 1) ||
register_chrdev_region(MKDEV(TTYAUX_MAJOR, 2), 1, "/dev/ptmx") < 0)
panic("Couldn't register /dev/ptmx driver\n");
devfs_mk_cdev(MKDEV(TTYAUX_MAJOR, 2), S_IFCHR|S_IRUGO|S_IWUGO, "ptmx");
class_device_create(tty_class, NULL, MKDEV(TTYAUX_MAJOR, 2), NULL, "ptmx");
#endif
#ifdef CONFIG_VT
cdev_init(&vc0_cdev, &console_fops);
if (cdev_add(&vc0_cdev, MKDEV(TTY_MAJOR, 0), 1) ||
register_chrdev_region(MKDEV(TTY_MAJOR, 0), 1, "/dev/vc/0") < 0)
panic("Couldn't register /dev/tty0 driver\n");
devfs_mk_cdev(MKDEV(TTY_MAJOR, 0), S_IFCHR|S_IRUSR|S_IWUSR, "vc/0");
class_device_create(tty_class, NULL, MKDEV(TTY_MAJOR, 0), NULL, "tty0");
vty_init();
#endif
return 0;
}
module_init(tty_init);