linux/drivers/acpi/acpi_dbg.c

793 lines
18 KiB
C
Raw Normal View History

/*
* ACPI AML interfacing support
*
* Copyright (C) 2015, Intel Corporation
* Authors: Lv Zheng <lv.zheng@intel.com>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
/* #define DEBUG */
#define pr_fmt(fmt) "ACPI: AML: " fmt
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/wait.h>
#include <linux/poll.h>
#include <linux/sched.h>
#include <linux/kthread.h>
#include <linux/proc_fs.h>
#include <linux/debugfs.h>
#include <linux/circ_buf.h>
#include <linux/acpi.h>
#include "internal.h"
#define ACPI_AML_BUF_ALIGN (sizeof (acpi_size))
#define ACPI_AML_BUF_SIZE PAGE_SIZE
#define circ_count(circ) \
(CIRC_CNT((circ)->head, (circ)->tail, ACPI_AML_BUF_SIZE))
#define circ_count_to_end(circ) \
(CIRC_CNT_TO_END((circ)->head, (circ)->tail, ACPI_AML_BUF_SIZE))
#define circ_space(circ) \
(CIRC_SPACE((circ)->head, (circ)->tail, ACPI_AML_BUF_SIZE))
#define circ_space_to_end(circ) \
(CIRC_SPACE_TO_END((circ)->head, (circ)->tail, ACPI_AML_BUF_SIZE))
#define ACPI_AML_OPENED 0x0001
#define ACPI_AML_CLOSED 0x0002
#define ACPI_AML_IN_USER 0x0004 /* user space is writing cmd */
#define ACPI_AML_IN_KERN 0x0008 /* kernel space is reading cmd */
#define ACPI_AML_OUT_USER 0x0010 /* user space is reading log */
#define ACPI_AML_OUT_KERN 0x0020 /* kernel space is writing log */
#define ACPI_AML_USER (ACPI_AML_IN_USER | ACPI_AML_OUT_USER)
#define ACPI_AML_KERN (ACPI_AML_IN_KERN | ACPI_AML_OUT_KERN)
#define ACPI_AML_BUSY (ACPI_AML_USER | ACPI_AML_KERN)
#define ACPI_AML_OPEN (ACPI_AML_OPENED | ACPI_AML_CLOSED)
struct acpi_aml_io {
wait_queue_head_t wait;
unsigned long flags;
unsigned long users;
struct mutex lock;
struct task_struct *thread;
char out_buf[ACPI_AML_BUF_SIZE] __aligned(ACPI_AML_BUF_ALIGN);
struct circ_buf out_crc;
char in_buf[ACPI_AML_BUF_SIZE] __aligned(ACPI_AML_BUF_ALIGN);
struct circ_buf in_crc;
acpi_osd_exec_callback function;
void *context;
unsigned long usages;
};
static struct acpi_aml_io acpi_aml_io;
static bool acpi_aml_initialized;
static struct file *acpi_aml_active_reader;
static struct dentry *acpi_aml_dentry;
static inline bool __acpi_aml_running(void)
{
return acpi_aml_io.thread ? true : false;
}
static inline bool __acpi_aml_access_ok(unsigned long flag)
{
/*
* The debugger interface is in opened state (OPENED && !CLOSED),
* then it is allowed to access the debugger buffers from either
* user space or the kernel space.
* In addition, for the kernel space, only the debugger thread
* (thread ID matched) is allowed to access.
*/
if (!(acpi_aml_io.flags & ACPI_AML_OPENED) ||
(acpi_aml_io.flags & ACPI_AML_CLOSED) ||
!__acpi_aml_running())
return false;
if ((flag & ACPI_AML_KERN) &&
current != acpi_aml_io.thread)
return false;
return true;
}
static inline bool __acpi_aml_readable(struct circ_buf *circ, unsigned long flag)
{
/*
* Another read is not in progress and there is data in buffer
* available for read.
*/
if (!(acpi_aml_io.flags & flag) && circ_count(circ))
return true;
return false;
}
static inline bool __acpi_aml_writable(struct circ_buf *circ, unsigned long flag)
{
/*
* Another write is not in progress and there is buffer space
* available for write.
*/
if (!(acpi_aml_io.flags & flag) && circ_space(circ))
return true;
return false;
}
static inline bool __acpi_aml_busy(void)
{
if (acpi_aml_io.flags & ACPI_AML_BUSY)
return true;
return false;
}
static inline bool __acpi_aml_opened(void)
{
if (acpi_aml_io.flags & ACPI_AML_OPEN)
return true;
return false;
}
static inline bool __acpi_aml_used(void)
{
return acpi_aml_io.usages ? true : false;
}
static inline bool acpi_aml_running(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = __acpi_aml_running();
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_busy(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = __acpi_aml_busy();
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_used(void)
{
bool ret;
/*
* The usage count is prepared to avoid race conditions between the
* starts and the stops of the debugger thread.
*/
mutex_lock(&acpi_aml_io.lock);
ret = __acpi_aml_used();
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_kern_readable(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = !__acpi_aml_access_ok(ACPI_AML_IN_KERN) ||
__acpi_aml_readable(&acpi_aml_io.in_crc, ACPI_AML_IN_KERN);
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_kern_writable(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = !__acpi_aml_access_ok(ACPI_AML_OUT_KERN) ||
__acpi_aml_writable(&acpi_aml_io.out_crc, ACPI_AML_OUT_KERN);
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_user_readable(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = !__acpi_aml_access_ok(ACPI_AML_OUT_USER) ||
__acpi_aml_readable(&acpi_aml_io.out_crc, ACPI_AML_OUT_USER);
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static bool acpi_aml_user_writable(void)
{
bool ret;
mutex_lock(&acpi_aml_io.lock);
ret = !__acpi_aml_access_ok(ACPI_AML_IN_USER) ||
__acpi_aml_writable(&acpi_aml_io.in_crc, ACPI_AML_IN_USER);
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static int acpi_aml_lock_write(struct circ_buf *circ, unsigned long flag)
{
int ret = 0;
mutex_lock(&acpi_aml_io.lock);
if (!__acpi_aml_access_ok(flag)) {
ret = -EFAULT;
goto out;
}
if (!__acpi_aml_writable(circ, flag)) {
ret = -EAGAIN;
goto out;
}
acpi_aml_io.flags |= flag;
out:
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static int acpi_aml_lock_read(struct circ_buf *circ, unsigned long flag)
{
int ret = 0;
mutex_lock(&acpi_aml_io.lock);
if (!__acpi_aml_access_ok(flag)) {
ret = -EFAULT;
goto out;
}
if (!__acpi_aml_readable(circ, flag)) {
ret = -EAGAIN;
goto out;
}
acpi_aml_io.flags |= flag;
out:
mutex_unlock(&acpi_aml_io.lock);
return ret;
}
static void acpi_aml_unlock_fifo(unsigned long flag, bool wakeup)
{
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.flags &= ~flag;
if (wakeup)
wake_up_interruptible(&acpi_aml_io.wait);
mutex_unlock(&acpi_aml_io.lock);
}
static int acpi_aml_write_kern(const char *buf, int len)
{
int ret;
struct circ_buf *crc = &acpi_aml_io.out_crc;
int n;
char *p;
ret = acpi_aml_lock_write(crc, ACPI_AML_OUT_KERN);
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
return ret;
/* sync tail before inserting logs */
smp_mb();
p = &crc->buf[crc->head];
n = min(len, circ_space_to_end(crc));
memcpy(p, buf, n);
/* sync head after inserting logs */
smp_wmb();
crc->head = (crc->head + n) & (ACPI_AML_BUF_SIZE - 1);
acpi_aml_unlock_fifo(ACPI_AML_OUT_KERN, true);
return n;
}
static int acpi_aml_readb_kern(void)
{
int ret;
struct circ_buf *crc = &acpi_aml_io.in_crc;
char *p;
ret = acpi_aml_lock_read(crc, ACPI_AML_IN_KERN);
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
return ret;
/* sync head before removing cmds */
smp_rmb();
p = &crc->buf[crc->tail];
ret = (int)*p;
/* sync tail before inserting cmds */
smp_mb();
crc->tail = (crc->tail + 1) & (ACPI_AML_BUF_SIZE - 1);
acpi_aml_unlock_fifo(ACPI_AML_IN_KERN, true);
return ret;
}
/*
* acpi_aml_write_log() - Capture debugger output
* @msg: the debugger output
*
* This function should be used to implement acpi_os_printf() to filter out
* the debugger output and store the output into the debugger interface
* buffer. Return the size of stored logs or errno.
*/
static ssize_t acpi_aml_write_log(const char *msg)
{
int ret = 0;
int count = 0, size = 0;
if (!acpi_aml_initialized)
return -ENODEV;
if (msg)
count = strlen(msg);
while (count > 0) {
again:
ret = acpi_aml_write_kern(msg + size, count);
if (ret == -EAGAIN) {
ret = wait_event_interruptible(acpi_aml_io.wait,
acpi_aml_kern_writable());
/*
* We need to retry when the condition
* becomes true.
*/
if (ret == 0)
goto again;
break;
}
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
break;
size += ret;
count -= ret;
}
return size > 0 ? size : ret;
}
/*
* acpi_aml_read_cmd() - Capture debugger input
* @msg: the debugger input
* @size: the size of the debugger input
*
* This function should be used to implement acpi_os_get_line() to capture
* the debugger input commands and store the input commands into the
* debugger interface buffer. Return the size of stored commands or errno.
*/
static ssize_t acpi_aml_read_cmd(char *msg, size_t count)
{
int ret = 0;
int size = 0;
/*
* This is ensured by the running fact of the debugger thread
* unless a bug is introduced.
*/
BUG_ON(!acpi_aml_initialized);
while (count > 0) {
again:
/*
* Check each input byte to find the end of the command.
*/
ret = acpi_aml_readb_kern();
if (ret == -EAGAIN) {
ret = wait_event_interruptible(acpi_aml_io.wait,
acpi_aml_kern_readable());
/*
* We need to retry when the condition becomes
* true.
*/
if (ret == 0)
goto again;
}
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
break;
*(msg + size) = (char)ret;
size++;
count--;
if (ret == '\n') {
/*
* acpi_os_get_line() requires a zero terminated command
* string.
*/
*(msg + size - 1) = '\0';
break;
}
}
return size > 0 ? size : ret;
}
static int acpi_aml_thread(void *unsed)
{
acpi_osd_exec_callback function = NULL;
void *context;
mutex_lock(&acpi_aml_io.lock);
if (acpi_aml_io.function) {
acpi_aml_io.usages++;
function = acpi_aml_io.function;
context = acpi_aml_io.context;
}
mutex_unlock(&acpi_aml_io.lock);
if (function)
function(context);
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.usages--;
if (!__acpi_aml_used()) {
acpi_aml_io.thread = NULL;
wake_up(&acpi_aml_io.wait);
}
mutex_unlock(&acpi_aml_io.lock);
return 0;
}
/*
* acpi_aml_create_thread() - Create AML debugger thread
* @function: the debugger thread callback
* @context: the context to be passed to the debugger thread
*
* This function should be used to implement acpi_os_execute() which is
* used by the ACPICA debugger to create the debugger thread.
*/
static int acpi_aml_create_thread(acpi_osd_exec_callback function, void *context)
{
struct task_struct *t;
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.function = function;
acpi_aml_io.context = context;
mutex_unlock(&acpi_aml_io.lock);
t = kthread_create(acpi_aml_thread, NULL, "aml");
if (IS_ERR(t)) {
pr_err("Failed to create AML debugger thread.\n");
return PTR_ERR(t);
}
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.thread = t;
acpi_set_debugger_thread_id((acpi_thread_id)(unsigned long)t);
wake_up_process(t);
mutex_unlock(&acpi_aml_io.lock);
return 0;
}
static int acpi_aml_wait_command_ready(bool single_step,
char *buffer, size_t length)
{
acpi_status status;
if (single_step)
acpi_os_printf("\n%1c ", ACPI_DEBUGGER_EXECUTE_PROMPT);
else
acpi_os_printf("\n%1c ", ACPI_DEBUGGER_COMMAND_PROMPT);
status = acpi_os_get_line(buffer, length, NULL);
if (ACPI_FAILURE(status))
return -EINVAL;
return 0;
}
static int acpi_aml_notify_command_complete(void)
{
return 0;
}
static int acpi_aml_open(struct inode *inode, struct file *file)
{
int ret = 0;
acpi_status status;
mutex_lock(&acpi_aml_io.lock);
/*
* The debugger interface is being closed, no new user is allowed
* during this period.
*/
if (acpi_aml_io.flags & ACPI_AML_CLOSED) {
ret = -EBUSY;
goto err_lock;
}
if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
/*
* Only one reader is allowed to initiate the debugger
* thread.
*/
if (acpi_aml_active_reader) {
ret = -EBUSY;
goto err_lock;
} else {
pr_debug("Opening debugger reader.\n");
acpi_aml_active_reader = file;
}
} else {
/*
* No writer is allowed unless the debugger thread is
* ready.
*/
if (!(acpi_aml_io.flags & ACPI_AML_OPENED)) {
ret = -ENODEV;
goto err_lock;
}
}
if (acpi_aml_active_reader == file) {
pr_debug("Opening debugger interface.\n");
mutex_unlock(&acpi_aml_io.lock);
pr_debug("Initializing debugger thread.\n");
status = acpi_initialize_debugger();
if (ACPI_FAILURE(status)) {
pr_err("Failed to initialize debugger.\n");
ret = -EINVAL;
goto err_exit;
}
pr_debug("Debugger thread initialized.\n");
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.flags |= ACPI_AML_OPENED;
acpi_aml_io.out_crc.head = acpi_aml_io.out_crc.tail = 0;
acpi_aml_io.in_crc.head = acpi_aml_io.in_crc.tail = 0;
pr_debug("Debugger interface opened.\n");
}
acpi_aml_io.users++;
err_lock:
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0) {
if (acpi_aml_active_reader == file)
acpi_aml_active_reader = NULL;
}
mutex_unlock(&acpi_aml_io.lock);
err_exit:
return ret;
}
static int acpi_aml_release(struct inode *inode, struct file *file)
{
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.users--;
if (file == acpi_aml_active_reader) {
pr_debug("Closing debugger reader.\n");
acpi_aml_active_reader = NULL;
pr_debug("Closing debugger interface.\n");
acpi_aml_io.flags |= ACPI_AML_CLOSED;
/*
* Wake up all user space/kernel space blocked
* readers/writers.
*/
wake_up_interruptible(&acpi_aml_io.wait);
mutex_unlock(&acpi_aml_io.lock);
/*
* Wait all user space/kernel space readers/writers to
* stop so that ACPICA command loop of the debugger thread
* should fail all its command line reads after this point.
*/
wait_event(acpi_aml_io.wait, !acpi_aml_busy());
/*
* Then we try to terminate the debugger thread if it is
* not terminated.
*/
pr_debug("Terminating debugger thread.\n");
acpi_terminate_debugger();
wait_event(acpi_aml_io.wait, !acpi_aml_used());
pr_debug("Debugger thread terminated.\n");
mutex_lock(&acpi_aml_io.lock);
acpi_aml_io.flags &= ~ACPI_AML_OPENED;
}
if (acpi_aml_io.users == 0) {
pr_debug("Debugger interface closed.\n");
acpi_aml_io.flags &= ~ACPI_AML_CLOSED;
}
mutex_unlock(&acpi_aml_io.lock);
return 0;
}
static int acpi_aml_read_user(char __user *buf, int len)
{
int ret;
struct circ_buf *crc = &acpi_aml_io.out_crc;
int n;
char *p;
ret = acpi_aml_lock_read(crc, ACPI_AML_OUT_USER);
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
return ret;
/* sync head before removing logs */
smp_rmb();
p = &crc->buf[crc->tail];
n = min(len, circ_count_to_end(crc));
if (copy_to_user(buf, p, n)) {
ret = -EFAULT;
goto out;
}
/* sync tail after removing logs */
smp_mb();
crc->tail = (crc->tail + n) & (ACPI_AML_BUF_SIZE - 1);
ret = n;
out:
acpi_aml_unlock_fifo(ACPI_AML_OUT_USER, ret >= 0);
return ret;
}
static ssize_t acpi_aml_read(struct file *file, char __user *buf,
size_t count, loff_t *ppos)
{
int ret = 0;
int size = 0;
if (!count)
return 0;
Remove 'type' argument from access_ok() function Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument of the user address range verification function since we got rid of the old racy i386-only code to walk page tables by hand. It existed because the original 80386 would not honor the write protect bit when in kernel mode, so you had to do COW by hand before doing any user access. But we haven't supported that in a long time, and these days the 'type' argument is a purely historical artifact. A discussion about extending 'user_access_begin()' to do the range checking resulted this patch, because there is no way we're going to move the old VERIFY_xyz interface to that model. And it's best done at the end of the merge window when I've done most of my merges, so let's just get this done once and for all. This patch was mostly done with a sed-script, with manual fix-ups for the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form. There were a couple of notable cases: - csky still had the old "verify_area()" name as an alias. - the iter_iov code had magical hardcoded knowledge of the actual values of VERIFY_{READ,WRITE} (not that they mattered, since nothing really used it) - microblaze used the type argument for a debug printout but other than those oddities this should be a total no-op patch. I tried to fix up all architectures, did fairly extensive grepping for access_ok() uses, and the changes are trivial, but I may have missed something. Any missed conversion should be trivially fixable, though. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 02:57:57 +00:00
if (!access_ok(buf, count))
return -EFAULT;
while (count > 0) {
again:
ret = acpi_aml_read_user(buf + size, count);
if (ret == -EAGAIN) {
if (file->f_flags & O_NONBLOCK)
break;
else {
ret = wait_event_interruptible(acpi_aml_io.wait,
acpi_aml_user_readable());
/*
* We need to retry when the condition
* becomes true.
*/
if (ret == 0)
goto again;
}
}
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0) {
if (!acpi_aml_running())
ret = 0;
break;
}
if (ret) {
size += ret;
count -= ret;
*ppos += ret;
break;
}
}
return size > 0 ? size : ret;
}
static int acpi_aml_write_user(const char __user *buf, int len)
{
int ret;
struct circ_buf *crc = &acpi_aml_io.in_crc;
int n;
char *p;
ret = acpi_aml_lock_write(crc, ACPI_AML_IN_USER);
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0)
return ret;
/* sync tail before inserting cmds */
smp_mb();
p = &crc->buf[crc->head];
n = min(len, circ_space_to_end(crc));
if (copy_from_user(p, buf, n)) {
ret = -EFAULT;
goto out;
}
/* sync head after inserting cmds */
smp_wmb();
crc->head = (crc->head + n) & (ACPI_AML_BUF_SIZE - 1);
ret = n;
out:
acpi_aml_unlock_fifo(ACPI_AML_IN_USER, ret >= 0);
return n;
}
static ssize_t acpi_aml_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos)
{
int ret = 0;
int size = 0;
if (!count)
return 0;
Remove 'type' argument from access_ok() function Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument of the user address range verification function since we got rid of the old racy i386-only code to walk page tables by hand. It existed because the original 80386 would not honor the write protect bit when in kernel mode, so you had to do COW by hand before doing any user access. But we haven't supported that in a long time, and these days the 'type' argument is a purely historical artifact. A discussion about extending 'user_access_begin()' to do the range checking resulted this patch, because there is no way we're going to move the old VERIFY_xyz interface to that model. And it's best done at the end of the merge window when I've done most of my merges, so let's just get this done once and for all. This patch was mostly done with a sed-script, with manual fix-ups for the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form. There were a couple of notable cases: - csky still had the old "verify_area()" name as an alias. - the iter_iov code had magical hardcoded knowledge of the actual values of VERIFY_{READ,WRITE} (not that they mattered, since nothing really used it) - microblaze used the type argument for a debug printout but other than those oddities this should be a total no-op patch. I tried to fix up all architectures, did fairly extensive grepping for access_ok() uses, and the changes are trivial, but I may have missed something. Any missed conversion should be trivially fixable, though. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-01-04 02:57:57 +00:00
if (!access_ok(buf, count))
return -EFAULT;
while (count > 0) {
again:
ret = acpi_aml_write_user(buf + size, count);
if (ret == -EAGAIN) {
if (file->f_flags & O_NONBLOCK)
break;
else {
ret = wait_event_interruptible(acpi_aml_io.wait,
acpi_aml_user_writable());
/*
* We need to retry when the condition
* becomes true.
*/
if (ret == 0)
goto again;
}
}
remove lots of IS_ERR_VALUE abuses Most users of IS_ERR_VALUE() in the kernel are wrong, as they pass an 'int' into a function that takes an 'unsigned long' argument. This happens to work because the type is sign-extended on 64-bit architectures before it gets converted into an unsigned type. However, anything that passes an 'unsigned short' or 'unsigned int' argument into IS_ERR_VALUE() is guaranteed to be broken, as are 8-bit integers and types that are wider than 'unsigned long'. Andrzej Hajda has already fixed a lot of the worst abusers that were causing actual bugs, but it would be nice to prevent any users that are not passing 'unsigned long' arguments. This patch changes all users of IS_ERR_VALUE() that I could find on 32-bit ARM randconfig builds and x86 allmodconfig. For the moment, this doesn't change the definition of IS_ERR_VALUE() because there are probably still architecture specific users elsewhere. Almost all the warnings I got are for files that are better off using 'if (err)' or 'if (err < 0)'. The only legitimate user I could find that we get a warning for is the (32-bit only) freescale fman driver, so I did not remove the IS_ERR_VALUE() there but changed the type to 'unsigned long'. For 9pfs, I just worked around one user whose calling conventions are so obscure that I did not dare change the behavior. I was using this definition for testing: #define IS_ERR_VALUE(x) ((unsigned long*)NULL == (typeof (x)*)NULL && \ unlikely((unsigned long long)(x) >= (unsigned long long)(typeof(x))-MAX_ERRNO)) which ends up making all 16-bit or wider types work correctly with the most plausible interpretation of what IS_ERR_VALUE() was supposed to return according to its users, but also causes a compile-time warning for any users that do not pass an 'unsigned long' argument. I suggested this approach earlier this year, but back then we ended up deciding to just fix the users that are obviously broken. After the initial warning that caused me to get involved in the discussion (fs/gfs2/dir.c) showed up again in the mainline kernel, Linus asked me to send the whole thing again. [ Updated the 9p parts as per Al Viro - Linus ] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andrzej Hajda <a.hajda@samsung.com> Cc: Andrew Morton <akpm@linux-foundation.org> Link: https://lkml.org/lkml/2016/1/7/363 Link: https://lkml.org/lkml/2016/5/27/486 Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> # For nvmem part Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-05-27 21:23:25 +00:00
if (ret < 0) {
if (!acpi_aml_running())
ret = 0;
break;
}
if (ret) {
size += ret;
count -= ret;
*ppos += ret;
}
}
return size > 0 ? size : ret;
}
static __poll_t acpi_aml_poll(struct file *file, poll_table *wait)
{
__poll_t masks = 0;
poll_wait(file, &acpi_aml_io.wait, wait);
if (acpi_aml_user_readable())
masks |= EPOLLIN | EPOLLRDNORM;
if (acpi_aml_user_writable())
masks |= EPOLLOUT | EPOLLWRNORM;
return masks;
}
static const struct file_operations acpi_aml_operations = {
.read = acpi_aml_read,
.write = acpi_aml_write,
.poll = acpi_aml_poll,
.open = acpi_aml_open,
.release = acpi_aml_release,
.llseek = generic_file_llseek,
};
static const struct acpi_debugger_ops acpi_aml_debugger = {
.create_thread = acpi_aml_create_thread,
.read_cmd = acpi_aml_read_cmd,
.write_log = acpi_aml_write_log,
.wait_command_ready = acpi_aml_wait_command_ready,
.notify_command_complete = acpi_aml_notify_command_complete,
};
int __init acpi_aml_init(void)
{
int ret;
/* Initialize AML IO interface */
mutex_init(&acpi_aml_io.lock);
init_waitqueue_head(&acpi_aml_io.wait);
acpi_aml_io.out_crc.buf = acpi_aml_io.out_buf;
acpi_aml_io.in_crc.buf = acpi_aml_io.in_buf;
acpi_aml_dentry = debugfs_create_file("acpidbg",
S_IFREG | S_IRUGO | S_IWUSR,
acpi_debugfs_dir, NULL,
&acpi_aml_operations);
ret = acpi_register_debugger(THIS_MODULE, &acpi_aml_debugger);
if (ret) {
debugfs_remove(acpi_aml_dentry);
acpi_aml_dentry = NULL;
return ret;
}
acpi_aml_initialized = true;
return 0;
}
void __exit acpi_aml_exit(void)
{
if (acpi_aml_initialized) {
acpi_unregister_debugger(&acpi_aml_debugger);
debugfs_remove(acpi_aml_dentry);
acpi_aml_dentry = NULL;
acpi_aml_initialized = false;
}
}
module_init(acpi_aml_init);
module_exit(acpi_aml_exit);
MODULE_AUTHOR("Lv Zheng");
MODULE_DESCRIPTION("ACPI debugger userspace IO driver");
MODULE_LICENSE("GPL");