Archive-Team
/
zfs
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 103 Wiki Activity
zfs/include/sys/mutex.h

196 lines
7.1 KiB
C
Raw Normal View History

Public Release Prep Updated AUTHORS, COPYING, DISCLAIMER, and INSTALL files. Added standardized headers to all source file to clearly indicate the copyright, license, and to give credit where credit is due.
2010-05-17 22:18:00 +00:00
/*****************************************************************************\
* Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
* Copyright (C) 2007 The Regents of the University of California.
* Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
* Written by Brian Behlendorf <behlendorf1@llnl.gov>.
Go through and add a header with the proper UCRL number. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@114 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-26 04:38:26 +00:00
* UCRL-CODE-235197
*
Public Release Prep Updated AUTHORS, COPYING, DISCLAIMER, and INSTALL files. Added standardized headers to all source file to clearly indicate the copyright, license, and to give credit where credit is due.
2010-05-17 22:18:00 +00:00
* This file is part of the SPL, Solaris Porting Layer.
Refresh links to web site Update links to refer to the official ZFS on Linux website instead of @behlendorf's personal fork on github. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2013-03-05 01:26:55 +00:00
* For details, see <http://zfsonlinux.org/>.
Public Release Prep Updated AUTHORS, COPYING, DISCLAIMER, and INSTALL files. Added standardized headers to all source file to clearly indicate the copyright, license, and to give credit where credit is due.
2010-05-17 22:18:00 +00:00
*
* The SPL is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation; either version 2 of the License, or (at your
* option) any later version.
Go through and add a header with the proper UCRL number. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@114 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-26 04:38:26 +00:00
*
Public Release Prep Updated AUTHORS, COPYING, DISCLAIMER, and INSTALL files. Added standardized headers to all source file to clearly indicate the copyright, license, and to give credit where credit is due.
2010-05-17 22:18:00 +00:00
* The SPL is distributed in the hope that it will be useful, but WITHOUT
Go through and add a header with the proper UCRL number. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@114 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-26 04:38:26 +00:00
* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
* for more details.
*
* You should have received a copy of the GNU General Public License along
Public Release Prep Updated AUTHORS, COPYING, DISCLAIMER, and INSTALL files. Added standardized headers to all source file to clearly indicate the copyright, license, and to give credit where credit is due.
2010-05-17 22:18:00 +00:00
* with the SPL. If not, see <http://www.gnu.org/licenses/>.
\*****************************************************************************/
Go through and add a header with the proper UCRL number. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@114 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-26 04:38:26 +00:00
Minor nit, SOLARIS should be SPL git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@17 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-28 00:52:31 +00:00
#ifndef _SPL_MUTEX_H
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define _SPL_MUTEX_H
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
Reorganize /include/ to add a /sys/, this way we don't need to muck with #includes in existing Solaris style source to get it to find the right stuff. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@18 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-03-01 00:45:59 +00:00
#include <sys/types.h>
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#include <linux/mutex.h>
Fix compilation error due to undefined ACCESS_ONCE macro. When CONFIG_DEBUG_MUTEXES is turned on in RHEL5's kernel config, the mutexes store the owner for debugging purposes, therefore the SPL will enable HAVE_MUTEX_OWNER. However, the SPL code uses ACCESS_ONCE() to access the owner, and this macro is not defined in the RHEL5 kernel, therefore we define it ourselves in include/linux/compiler_compat.h. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-07-20 20:47:37 +00:00
#include <linux/compiler_compat.h>
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
typedef enum {
MUTEX_DEFAULT = 0,
MUTEX_SPIN = 1,
MUTEX_ADAPTIVE = 2
} kmutex_type_t;
Lots of build fixes. This is turning out to be a very good idea since it forcefully codifing the ABI. Since the shim layer is no longer linked at build time in to the test suite we can;'t cut any corners and get away with it. Everything is working now with the exception of sorting setting Module.symvers properly. This may take a little Makefile reorg. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@5 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-27 19:09:51 +00:00
Remove adaptive mutex implementation Since the Linux 2.6.29 kernel all mutexes have been adaptive mutexs. There is no longer any point in keeping this code so it is being removed to simplify the code. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-09-29 20:48:35 +00:00
#if defined(HAVE_MUTEX_OWNER) && defined(CONFIG_SMP) && \
!defined(CONFIG_DEBUG_MUTEXES)
Commit adaptive mutexes. This seems to have introduced some new crashes but it's not clear to me yet if these are a problem with the mutex implementation or ZFSs usage of it. Minor taskq fixes to add new tasks to the end of the pending list. Minor enhansements to the debug infrastructure. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@94 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-05 20:18:49 +00:00
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
/*
* We define a 1-field struct rather than a straight typedef to enforce type
* safety.
*/
typedef struct {
struct mutex m;
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spinlock_t m_lock; /* used for serializing mutex_exit */
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
} kmutex_t;
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
static inline kthread_t *
mutex_owner(kmutex_t *mp)
{
Linux 2.6.39 compat, mutex owner Prior to Linux 2.6.39 when CONFIG_DEBUG_MUTEXES was defined the kernel stored a thread_info pointer as the mutex owner. From this you could get the pointer of the current task_struct to compare with get_current(). As of Linux 2.6.39 this behavior has changed and now the mutex stores a pointer to the task_struct. This commit detects the type of pointer stored in the mutex and adjusts the mutex_owner() and mutex_owned() functions to perform the correct comparision.
2011-06-24 18:57:14 +00:00
#if defined(HAVE_MUTEX_OWNER_TASK_STRUCT)
return ACCESS_ONCE(mp->m.owner);
#else
struct thread_info *owner = ACCESS_ONCE(mp->m.owner);
if (owner)
return owner->task;
Treat mutex->owner as volatile When HAVE_MUTEX_OWNER is defined and we are directly accessing mutex->owner treat is as volative with the ACCESS_ONCE() helper. Without this you may get a stale cached value when accessing it from different cpus. This can result in incorrect behavior from mutex_owned() and mutex_owner(). This is not a problem for the !HAVE_MUTEX_OWNER case because in this case all the accesses are covered by a spin lock which similarly gaurentees we will not be accessing stale data. Secondly, check CONFIG_SMP before allowing access to mutex->owner. I see that for non-SMP setups the kernel does not track the owner so we cannot rely on it. Thirdly, check CONFIG_MUTEX_DEBUG when this is defined and the HAVE_MUTEX_OWNER is defined surprisingly the mutex->owner will not be cleared on mutex_exit(). When this is the case the SPL needs to make sure to do it to ensure MUTEX_HELD() behaves as expected or you will certainly assert in mutex_destroy(). Finally, improve the mutex regression tests. For mutex_owned() we now minimally check that it behaves correctly when checked from the owner thread or the non-owner thread. This subtle behaviour has bit me before and I'd like to catch it early next time if it reappears. As for mutex_owned() regression test additonally verify that mutex->owner is always cleared on mutex_exit().
2010-06-28 19:48:20 +00:00
Linux 2.6.39 compat, mutex owner Prior to Linux 2.6.39 when CONFIG_DEBUG_MUTEXES was defined the kernel stored a thread_info pointer as the mutex owner. From this you could get the pointer of the current task_struct to compare with get_current(). As of Linux 2.6.39 this behavior has changed and now the mutex stores a pointer to the task_struct. This commit detects the type of pointer stored in the mutex and adjusts the mutex_owner() and mutex_owned() functions to perform the correct comparision.
2011-06-24 18:57:14 +00:00
return NULL;
#endif
Treat mutex->owner as volatile When HAVE_MUTEX_OWNER is defined and we are directly accessing mutex->owner treat is as volative with the ACCESS_ONCE() helper. Without this you may get a stale cached value when accessing it from different cpus. This can result in incorrect behavior from mutex_owned() and mutex_owner(). This is not a problem for the !HAVE_MUTEX_OWNER case because in this case all the accesses are covered by a spin lock which similarly gaurentees we will not be accessing stale data. Secondly, check CONFIG_SMP before allowing access to mutex->owner. I see that for non-SMP setups the kernel does not track the owner so we cannot rely on it. Thirdly, check CONFIG_MUTEX_DEBUG when this is defined and the HAVE_MUTEX_OWNER is defined surprisingly the mutex->owner will not be cleared on mutex_exit(). When this is the case the SPL needs to make sure to do it to ensure MUTEX_HELD() behaves as expected or you will certainly assert in mutex_destroy(). Finally, improve the mutex regression tests. For mutex_owned() we now minimally check that it behaves correctly when checked from the owner thread or the non-owner thread. This subtle behaviour has bit me before and I'd like to catch it early next time if it reappears. As for mutex_owned() regression test additonally verify that mutex->owner is always cleared on mutex_exit().
2010-06-28 19:48:20 +00:00
}
Linux 2.6.39 compat, mutex owner Prior to Linux 2.6.39 when CONFIG_DEBUG_MUTEXES was defined the kernel stored a thread_info pointer as the mutex owner. From this you could get the pointer of the current task_struct to compare with get_current(). As of Linux 2.6.39 this behavior has changed and now the mutex stores a pointer to the task_struct. This commit detects the type of pointer stored in the mutex and adjusts the mutex_owner() and mutex_owned() functions to perform the correct comparision.
2011-06-24 18:57:14 +00:00
#define mutex_owned(mp) (mutex_owner(mp) == current)
Add MUTEX_NOT_HELD() function Simply implement the missing MUTEX_NOT_HELD() function using the !MUTEX_HELD construct.
2010-08-27 20:36:24 +00:00
#define MUTEX_HELD(mp) mutex_owned(mp)
#define MUTEX_NOT_HELD(mp) (!MUTEX_HELD(mp))
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#undef mutex_init
#define mutex_init(mp, name, type, ibc) \
Add mutex_enter_nested() as wrapper for mutex_lock_nested(). This symbol can be used by GPL modules which use the SPL to handle cases where a call path takes a two different locks by the same name. This is needed to avoid a false positive in the lock checker.
2009-11-15 22:27:15 +00:00
({ \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
static struct lock_class_key __key; \
ASSERT(type == MUTEX_DEFAULT); \
\
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
__mutex_init(&(mp)->m, #mp, &__key); \
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spin_lock_init(&(mp)->m_lock); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
})
Add mutex_enter_nested() as wrapper for mutex_lock_nested(). This symbol can be used by GPL modules which use the SPL to handle cases where a call path takes a two different locks by the same name. This is needed to avoid a false positive in the lock checker.
2009-11-15 22:27:15 +00:00
Always use the generic mutex_destroy().
2009-11-15 23:04:02 +00:00
#undef mutex_destroy
#define mutex_destroy(mp) \
({ \
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
VERIFY3P(mutex_owner(mp), ==, NULL); \
Always use the generic mutex_destroy().
2009-11-15 23:04:02 +00:00
})
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
#define mutex_tryenter(mp) mutex_trylock(&(mp)->m)
Mutex ASSERT on self deadlock Generate an assertion if we're going to deadlock the system by attempting to acquire a mutex the process is already holding. There are currently no known instances of this under normal operation, but it _might_ be possible when using a ZVOL as a swap device. I want to ensure we catch this immediately if it were to occur. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-08-19 19:10:19 +00:00
#define mutex_enter(mp) \
({ \
ASSERT3P(mutex_owner(mp), !=, current); \
mutex_lock(&(mp)->m); \
Remove adaptive mutex implementation Since the Linux 2.6.29 kernel all mutexes have been adaptive mutexs. There is no longer any point in keeping this code so it is being removed to simplify the code. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-09-29 20:48:35 +00:00
})
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
/*
* The reason for the spinlock:
*
* The Linux mutex is designed with a fast-path/slow-path design such that it
* does not guarantee serialization upon itself, allowing a race where latter
* acquirers finish mutex_unlock before former ones.
*
* The race renders it unsafe to be used for serializing the freeing of an
* object in which the mutex is embedded, where the latter acquirer could go
* on to free the object while the former one is still doing mutex_unlock and
* causing memory corruption.
*
* However, there are many places in ZFS where the mutex is used for
* serializing object freeing, and the code is shared among other OSes without
* this issue. Thus, we need the spinlock to force the serialization on
* mutex_exit().
*
* See http://lwn.net/Articles/575477/ for the information about the race.
*/
#define mutex_exit(mp) \
({ \
spin_lock(&(mp)->m_lock); \
mutex_unlock(&(mp)->m); \
spin_unlock(&(mp)->m_lock); \
})
Always use the generic mutex_destroy().
2009-11-15 23:04:02 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#else /* HAVE_MUTEX_OWNER */
Lots of build fixes. This is turning out to be a very good idea since it forcefully codifing the ABI. Since the shim layer is no longer linked at build time in to the test suite we can;'t cut any corners and get away with it. Everything is working now with the exception of sorting setting Module.symvers properly. This may take a little Makefile reorg. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@5 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-27 19:09:51 +00:00
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
typedef struct {
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
struct mutex m_mutex;
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spinlock_t m_lock;
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
kthread_t *m_owner;
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
} kmutex_t;
Cleaned up MUTEX() #define The old define assumed a specific layout of the kmutex_t struct. This patch makes the macro independent from the actual struct layout. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2011-10-19 16:04:40 +00:00
#define MUTEX(mp) (&((mp)->m_mutex))
- Add some spinlocks to cover all the private data in the mutex. I don't think this should fix anything but it's a good idea regardless. - Drop the lock before calling the construct/destructor for the slab otherwise we can't sleep in a constructor/destructor and for long running functions we may NMI. - Do something braindead, but safe for the console debug logs for now. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@73 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-04-15 20:53:36 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
static inline void
spl_mutex_set_owner(kmutex_t *mp)
{
mp->m_owner = current;
}
static inline void
spl_mutex_clear_owner(kmutex_t *mp)
{
mp->m_owner = NULL;
}
Remove the spinlocks for mutex_enter()/mutex_exit() The m_owner variable is protected by the mutex itself. Reading the variable is guaranteed to be atomic (due to it being a word-sized reference) and ACCESS_ONCE() takes care of read cache effects. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2011-10-18 00:54:35 +00:00
#define mutex_owner(mp) (ACCESS_ONCE((mp)->m_owner))
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define mutex_owned(mp) (mutex_owner(mp) == current)
#define MUTEX_HELD(mp) mutex_owned(mp)
Add MUTEX_NOT_HELD() function Simply implement the missing MUTEX_NOT_HELD() function using the !MUTEX_HELD construct.
2010-08-27 20:36:24 +00:00
#define MUTEX_NOT_HELD(mp) (!MUTEX_HELD(mp))
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
/*
* The following functions must be a #define and not static inline.
* This ensures that the native linux mutex functions (lock/unlock)
* will be correctly located in the users code which is important
* for the built in kernel lock analysis tools
*/
Commit adaptive mutexes. This seems to have introduced some new crashes but it's not clear to me yet if these are a problem with the mutex implementation or ZFSs usage of it. Minor taskq fixes to add new tasks to the end of the pending list. Minor enhansements to the debug infrastructure. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@94 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-05 20:18:49 +00:00
#undef mutex_init
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define mutex_init(mp, name, type, ibc) \
({ \
static struct lock_class_key __key; \
ASSERT(type == MUTEX_DEFAULT); \
\
__mutex_init(MUTEX(mp), #mp, &__key); \
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spin_lock_init(&(mp)->m_lock); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
spl_mutex_clear_owner(mp); \
})
Commit adaptive mutexes. This seems to have introduced some new crashes but it's not clear to me yet if these are a problem with the mutex implementation or ZFSs usage of it. Minor taskq fixes to add new tasks to the end of the pending list. Minor enhansements to the debug infrastructure. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@94 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-05 20:18:49 +00:00
#undef mutex_destroy
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define mutex_destroy(mp) \
({ \
Make kmutex_t typesafe in all cases. When HAVE_MUTEX_OWNER and CONFIG_SMP are defined, kmutex_t is just a typedef for struct mutex. This is generally OK but has the downside that it can make mistakes such as mutex_lock(&kmutex_var) to pass by unnoticed until someone compiles the code without HAVE_MUTEX_OWNER or CONFIG_SMP (in which case kmutex_t is a real struct). Note that the correct API to call should have been mutex_enter() rather than mutex_lock(). We prevent these kind of mistakes by making kmutex_t a real structure with only one field. This makes kmutex_t typesafe and it shouldn't have any impact on the generated assembly code. Signed-off-by: Ricardo M. Correia <ricardo.correia@oracle.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-11-22 08:20:58 +00:00
VERIFY3P(mutex_owner(mp), ==, NULL); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
})
- Add some spinlocks to cover all the private data in the mutex. I don't think this should fix anything but it's a good idea regardless. - Drop the lock before calling the construct/destructor for the slab otherwise we can't sleep in a constructor/destructor and for long running functions we may NMI. - Do something braindead, but safe for the console debug logs for now. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@73 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-04-15 20:53:36 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define mutex_tryenter(mp) \
({ \
int _rc_; \
\
if ((_rc_ = mutex_trylock(MUTEX(mp))) == 1) \
spl_mutex_set_owner(mp); \
\
_rc_; \
Commit adaptive mutexes. This seems to have introduced some new crashes but it's not clear to me yet if these are a problem with the mutex implementation or ZFSs usage of it. Minor taskq fixes to add new tasks to the end of the pending list. Minor enhansements to the debug infrastructure. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@94 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-05-05 20:18:49 +00:00
})
Initial commit. All spl source written up to this point wrapped in an initial reasonable autoconf style build system. This does not yet build but the configure system does appear to work properly and integrate with the kernel. Hopefully the next commit gets us back to a buildable version we can run the test suite against. git-svn-id: https://outreach.scidac.gov/svn/spl/trunk@1 7e1ea52c-4ff2-0310-8f11-9dd32ca42a1c
2008-02-26 20:36:04 +00:00
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
#define mutex_enter(mp) \
({ \
Remove adaptive mutex implementation Since the Linux 2.6.29 kernel all mutexes have been adaptive mutexs. There is no longer any point in keeping this code so it is being removed to simplify the code. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-09-29 20:48:35 +00:00
ASSERT3P(mutex_owner(mp), !=, current); \
mutex_lock(MUTEX(mp)); \
spl_mutex_set_owner(mp); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
})
#define mutex_exit(mp) \
({ \
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spin_lock(&(mp)->m_lock); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
spl_mutex_clear_owner(mp); \
mutex_unlock(MUTEX(mp)); \
mutex: force serialization on mutex_exit() to fix races It is known that mutexes in Linux are not safe when using them to synchronize the freeing of object in which the mutex is embedded: http://lwn.net/Articles/575477/ The known places in ZFS which are suspected to suffer from the race condition are zio->io_lock and dbuf->db_mtx. * zio uses zio->io_lock and zio->io_cv to synchronize freeing between zio_wait() and zio_done(). * dbuf uses dbuf->db_mtx to protect reference counting. This patch fixes this kind of race by forcing serialization on mutex_exit() with a spin lock, making the mutex safe by sacrificing a bit of performance and memory overhead. This issue most commonly manifests itself as a deadlock in the zio pipeline caused by a process spinning on the damaged mutex. Similar deadlocks have been reported for the dbuf->db_mtx mutex. And it can also cause a NULL dereference or bad paging request under the right circumstances. This issue any many like it are linked off the zfsonlinux/zfs#2523 issue. Specifically this fix resolves at least the following outstanding issues: zfsonlinux/zfs#401 zfsonlinux/zfs#2523 zfsonlinux/zfs#2679 zfsonlinux/zfs#2684 zfsonlinux/zfs#2704 zfsonlinux/zfs#2708 zfsonlinux/zfs#2517 zfsonlinux/zfs#2827 zfsonlinux/zfs#2850 zfsonlinux/zfs#2891 zfsonlinux/zfs#2897 zfsonlinux/zfs#2247 zfsonlinux/zfs#2939 Signed-off-by: Chunwei Chen <tuxoko@gmail.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Yao <ryao@gentoo.org> Closes #421
2014-12-19 03:31:59 +00:00
spin_unlock(&(mp)->m_lock); \
Reimplement mutexs for Linux lock profiling/analysis For a generic explanation of why mutexs needed to be reimplemented to work with the kernel lock profiling see commits: e811949a57044d60d12953c5c3b808a79a7d36ef and d28db80fd0fd4fd63aec09037c44408e51a222d6 The specific changes made to the mutex implemetation are as follows. The Linux mutex structure is now directly embedded in the kmutex_t. This allows a kmutex_t to be directly case to a mutex struct and passed directly to the Linux primative. Just like with the rwlocks it is critical that these functions be implemented as '#defines to ensure the location information is preserved. The preprocessor can then do a direct replacement of the Solaris primative with the linux primative. Just as with the rwlocks we need to track the lock owner. Here things get a little more interesting because depending on your kernel version, and how you've built your kernel Linux may already do this for you. If your running a 2.6.29 or newer kernel on a SMP system the lock owner will be tracked. This was added to Linux to support adaptive mutexs, more on that shortly. Alternately, your kernel might track the lock owner if you've set CONFIG_DEBUG_MUTEXES in the kernel build. If neither of the above things is true for your kernel the kmutex_t type will include and track the lock owner to ensure correct behavior. This is all handled by a new autoconf check called SPL_AC_MUTEX_OWNER. Concerning adaptive mutexs these are a very recent development and they did not make it in to either the latest FC11 of SLES11 kernels. Ideally, I'd love to see this kernel change appear in one of these distros because it does help performance. From Linux kernel commit: 0d66bf6d3514b35eb6897629059443132992dbd7 "Testing with Ingo's test-mutex application... gave a 345% boost for VFS scalability on my testbox" However, if you don't want to backport this change yourself you can still simply export the task_curr() symbol. The kmutex_t implementation will use this symbol when it's available to provide it's own adaptive mutexs. Finally, DEBUG_MUTEX support was removed including the proc handlers. This was done because now that we are cleanly integrated with the kernel profiling all this information and much much more is available in debug kernel builds. This code was now redundant. Update mutexs validated on: - SLES10 (ppc64) - SLES11 (x86_64) - CHAOS4.2 (x86_64) - RHEL5.3 (x86_64) - RHEL6 (x86_64) - FC11 (x86_64)
2009-09-25 21:47:01 +00:00
})
#endif /* HAVE_MUTEX_OWNER */
int spl_mutex_init(void);
void spl_mutex_fini(void);
#endif /* _SPL_MUTEX_H */