d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
79462204f1
freebsd-dev/sys/kern/kern_umtx.c

2760 lines
61 KiB
C
Raw Normal View History

/* -> /*- for copyright notices, minor format tweaks as necessary
2005-01-06 23:35:40 +00:00
/*-
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
* Copyright (c) 2004, David Xu <davidxu@freebsd.org>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
* Copyright (c) 2002, Jeffrey Roberson <jeff@freebsd.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice unmodified, this list of conditions, and the following
* disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
Use __FBSDID().
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
#include "opt_compat.h"
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/param.h>
#include <sys/kernel.h>
Change the thread ID (thr_id_t) used for 1:1 threading from being a pointer to the corresponding struct thread to the thread ID (lwpid_t) assigned to that thread. The primary reason for this change is that libthr now internally uses the same ID as the debugger and the kernel when referencing to a kernel thread. This allows us to implement the support for debugging without additional translations and/or mappings. To preserve the ABI, the 1:1 threading syscalls, including the umtx locking API have not been changed to work on a lwpid_t. Instead the 1:1 threading syscalls operate on long and the umtx locking API has not been changed except for the contested bit. Previously this was the least significant bit. Now it's the most significant bit. Since the contested bit should not be tested by userland, this change is not expected to be visible. Just to be sure, UMTX_CONTESTED has been removed from <sys/umtx.h>. Reviewed by: mtm@ ABI preservation tested on: i386, ia64
2004-07-02 00:40:07 +00:00
#include <sys/limits.h>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/lock.h>
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
#include <sys/malloc.h>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/mutex.h>
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
#include <sys/priv.h>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/proc.h>
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
#include <sys/sched.h>
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#include <sys/smp.h>
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
#include <sys/sysctl.h>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/sysent.h>
#include <sys/systm.h>
#include <sys/sysproto.h>
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
#include <sys/eventhandler.h>
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
#include <sys/umtx.h>
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#include <machine/cpu.h>
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
#ifdef COMPAT_IA32
#include <compat/freebsd32/freebsd32_proto.h>
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
#define TYPE_SIMPLE_LOCK 0
#define TYPE_SIMPLE_WAIT 1
#define TYPE_NORMAL_UMUTEX 2
#define TYPE_PI_UMUTEX 3
#define TYPE_PP_UMUTEX 4
#define TYPE_CV 5
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* Key to represent a unique userland synchronous object */
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct umtx_key {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int hash;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
int type;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int shared;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
union {
struct {
vm_object_t object;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
uintptr_t offset;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
} shared;
struct {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct vmspace *vs;
uintptr_t addr;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
} private;
struct {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
void *a;
uintptr_t b;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
} both;
} info;
};
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* Priority inheritance mutex info. */
struct umtx_pi {
/* Owner thread */
struct thread *pi_owner;
/* Reference count */
int pi_refcount;
/* List entry to link umtx holding by thread */
TAILQ_ENTRY(umtx_pi) pi_link;
/* List entry in hash */
TAILQ_ENTRY(umtx_pi) pi_hashlink;
/* List for waiters */
TAILQ_HEAD(,umtx_q) pi_blocked;
/* Identify a userland lock object */
struct umtx_key pi_key;
};
/* A userland synchronous object user. */
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
struct umtx_q {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* Linked list for the hash. */
TAILQ_ENTRY(umtx_q) uq_link;
/* Umtx key. */
struct umtx_key uq_key;
/* Umtx flags. */
int uq_flags;
#define UQF_UMTXQ 0x0001
/* The thread waits on. */
struct thread *uq_thread;
/*
* Blocked on PI mutex. read can use chain lock
* or sched_lock, write must have both chain lock and
* sched_lock being hold.
*/
struct umtx_pi *uq_pi_blocked;
/* On blocked list */
TAILQ_ENTRY(umtx_q) uq_lockq;
/* Thread contending with us */
TAILQ_HEAD(,umtx_pi) uq_pi_contested;
Reorder some statments. Fix typo and remove stale comments.
2006-08-30 23:59:45 +00:00
/* Inherited priority from PP mutex */
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
u_char uq_inherited_pri;
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
};
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
TAILQ_HEAD(umtxq_head, umtx_q);
/* Userland lock object's wait-queue chain */
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
struct umtxq_chain {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* Lock for this chain. */
struct mtx uc_lock;
/* List of sleep queues. */
struct umtxq_head uc_queue;
/* Busy flag */
char uc_busy;
/* Chain lock waiters */
Use wakeup_one to avoid thundering herd. Tested by: kris
2006-05-09 13:00:46 +00:00
int uc_waiters;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* All PI in the list */
TAILQ_HEAD(,umtx_pi) uc_pi_list;
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
};
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
#define UMTXQ_LOCKED_ASSERT(uc) mtx_assert(&(uc)->uc_lock, MA_OWNED)
/*
* Don't propagate time-sharing priority, there is a security reason,
* a user can simply introduce PI-mutex, let thread A lock the mutex,
* and let another thread B block on the mutex, because B is
* sleeping, its priority will be boosted, this causes A's priority to
* be boosted via priority propagating too and will never be lowered even
* if it is using 100%CPU, this is unfair to other processes.
*/
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
#define UPRI(td) (((td)->td_user_pri >= PRI_MIN_TIMESHARE &&\
(td)->td_user_pri <= PRI_MAX_TIMESHARE) ?\
PRI_MAX_TIMESHARE : (td)->td_user_pri)
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
#define GOLDEN_RATIO_PRIME 2654404609U
#define UMTX_CHAINS 128
#define UMTX_SHIFTS (__WORD_BIT - 7)
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
#define THREAD_SHARE 0
#define PROCESS_SHARE 1
#define AUTO_SHARE 2
#define GET_SHARE(flags) \
(((flags) & USYNC_PROCESS_SHARED) == 0 ? THREAD_SHARE : PROCESS_SHARE)
static uma_zone_t umtx_pi_zone;
static struct umtxq_chain umtxq_chains[UMTX_CHAINS];
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
static MALLOC_DEFINE(M_UMTX, "umtx", "UMTX queue memory");
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static int umtx_pi_allocated;
- Create a new lock, umtx_lock, for use instead of the proc lock for protecting the umtx queues. We can't use the proc lock because we need to hold the lock across calls to casuptr, which can fault. Approved by: re
2003-05-25 18:18:32 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
SYSCTL_NODE(_debug, OID_AUTO, umtx, CTLFLAG_RW, 0, "umtx debug");
SYSCTL_INT(_debug_umtx, OID_AUTO, umtx_pi_allocated, CTLFLAG_RD,
&umtx_pi_allocated, 0, "Allocated umtx_pi");
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
SYSCTL_DECL(_kern_threads);
static int umtx_dflt_spins = 0;
SYSCTL_INT(_kern_threads, OID_AUTO, umtx_dflt_spins, CTLFLAG_RW,
&umtx_dflt_spins, 0, "default umtx spin count");
static int umtx_max_spins = 3000;
SYSCTL_INT(_kern_threads, OID_AUTO, umtx_max_spins, CTLFLAG_RW,
&umtx_max_spins, 0, "max umtx spin count");
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static void umtxq_sysinit(void *);
static void umtxq_hash(struct umtx_key *key);
static struct umtxq_chain *umtxq_getchain(struct umtx_key *key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static void umtxq_lock(struct umtx_key *key);
static void umtxq_unlock(struct umtx_key *key);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
static void umtxq_busy(struct umtx_key *key);
static void umtxq_unbusy(struct umtx_key *key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static void umtxq_insert(struct umtx_q *uq);
static void umtxq_remove(struct umtx_q *uq);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static int umtxq_sleep(struct umtx_q *uq, const char *wmesg, int timo);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
static int umtxq_count(struct umtx_key *key);
static int umtxq_signal(struct umtx_key *key, int nr_wakeup);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int umtx_key_match(const struct umtx_key *k1, const struct umtx_key *k2);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static int umtx_key_get(void *addr, int type, int share,
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct umtx_key *key);
static void umtx_key_release(struct umtx_key *key);
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
static struct umtx_pi *umtx_pi_alloc(int);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static void umtx_pi_free(struct umtx_pi *pi);
static int do_unlock_pp(struct thread *td, struct umutex *m, uint32_t flags);
static void umtx_thread_cleanup(struct thread *td);
static void umtx_exec_hook(void *arg __unused, struct proc *p __unused,
struct image_params *imgp __unused);
SYSINIT(umtx, SI_SUB_EVENTHANDLER+1, SI_ORDER_MIDDLE, umtxq_sysinit, NULL);
static void
umtxq_sysinit(void *arg __unused)
{
int i;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtx_pi_zone = uma_zcreate("umtx pi", sizeof(struct umtx_pi),
NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0);
for (i = 0; i < UMTX_CHAINS; ++i) {
mtx_init(&umtxq_chains[i].uc_lock, "umtxql", NULL,
MTX_DEF | MTX_DUPOK);
TAILQ_INIT(&umtxq_chains[i].uc_queue);
TAILQ_INIT(&umtxq_chains[i].uc_pi_list);
umtxq_chains[i].uc_busy = 0;
umtxq_chains[i].uc_waiters = 0;
}
EVENTHANDLER_REGISTER(process_exec, umtx_exec_hook, NULL,
EVENTHANDLER_PRI_ANY);
}
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
struct umtx_q *
umtxq_alloc(void)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtx_q *uq;
uq = malloc(sizeof(struct umtx_q), M_UMTX, M_WAITOK | M_ZERO);
TAILQ_INIT(&uq->uq_pi_contested);
uq->uq_inherited_pri = PRI_MAX;
return (uq);
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
}
void
umtxq_free(struct umtx_q *uq)
{
free(uq, M_UMTX);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static inline void
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
umtxq_hash(struct umtx_key *key)
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
unsigned n = (uintptr_t)key->info.both.a + key->info.both.b;
key->hash = ((n * GOLDEN_RATIO_PRIME) >> UMTX_SHIFTS) % UMTX_CHAINS;
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
}
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static inline int
umtx_key_match(const struct umtx_key *k1, const struct umtx_key *k2)
{
return (k1->type == k2->type &&
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
k1->info.both.a == k2->info.both.a &&
k1->info.both.b == k2->info.both.b);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
static inline struct umtxq_chain *
umtxq_getchain(struct umtx_key *key)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
return (&umtxq_chains[key->hash]);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Set chain to busy state when following operation
* may be blocked (kernel mutex can not be used).
*/
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
static inline void
umtxq_busy(struct umtx_key *key)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
uc = umtxq_getchain(key);
mtx_assert(&uc->uc_lock, MA_OWNED);
while (uc->uc_busy != 0) {
uc->uc_waiters++;
msleep(uc, &uc->uc_lock, 0, "umtxqb", 0);
uc->uc_waiters--;
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
uc->uc_busy = 1;
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Unbusy a chain.
*/
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
static inline void
umtxq_unbusy(struct umtx_key *key)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
uc = umtxq_getchain(key);
mtx_assert(&uc->uc_lock, MA_OWNED);
KASSERT(uc->uc_busy != 0, ("not busy"));
uc->uc_busy = 0;
if (uc->uc_waiters)
wakeup_one(uc);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Lock a chain.
*/
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
static inline void
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
umtxq_lock(struct umtx_key *key)
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
uc = umtxq_getchain(key);
mtx_lock(&uc->uc_lock);
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Unlock a chain.
*/
Forgot to inline umtxq_unlock.
2004-11-30 12:18:53 +00:00
static inline void
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
umtxq_unlock(struct umtx_key *key)
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
uc = umtxq_getchain(key);
mtx_unlock(&uc->uc_lock);
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
}
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
/*
* Insert a thread onto the umtx queue.
*/
static inline void
umtxq_insert(struct umtx_q *uq)
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
uc = umtxq_getchain(&uq->uq_key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_INSERT_TAIL(&uc->uc_queue, uq, uq_link);
Move flag TDF_UMTXQ into structure umtxq, this eliminates the requirement of scheduler lock in some umtx code.
2006-05-18 08:43:46 +00:00
uq->uq_flags |= UQF_UMTXQ;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
/*
* Remove thread from the umtx queue.
*/
static inline void
umtxq_remove(struct umtx_q *uq)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
uc = umtxq_getchain(&uq->uq_key);
UMTXQ_LOCKED_ASSERT(uc);
Move flag TDF_UMTXQ into structure umtxq, this eliminates the requirement of scheduler lock in some umtx code.
2006-05-18 08:43:46 +00:00
if (uq->uq_flags & UQF_UMTXQ) {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
TAILQ_REMOVE(&uc->uc_queue, uq, uq_link);
Move flag TDF_UMTXQ into structure umtxq, this eliminates the requirement of scheduler lock in some umtx code.
2006-05-18 08:43:46 +00:00
uq->uq_flags &= ~UQF_UMTXQ;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Check if there are multiple waiters
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
umtxq_count(struct umtx_key *key)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
struct umtx_q *uq;
int count = 0;
uc = umtxq_getchain(key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_FOREACH(uq, &uc->uc_queue, uq_link) {
if (umtx_key_match(&uq->uq_key, key)) {
if (++count > 1)
break;
}
}
return (count);
}
/*
* Check if there are multiple PI waiters and returns first
* waiter.
*/
static int
umtxq_count_pi(struct umtx_key *key, struct umtx_q **first)
{
struct umtxq_chain *uc;
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
struct umtx_q *uq;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int count = 0;
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
*first = NULL;
uc = umtxq_getchain(key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_FOREACH(uq, &uc->uc_queue, uq_link) {
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
if (umtx_key_match(&uq->uq_key, key)) {
if (++count > 1)
break;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
*first = uq;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
}
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
return (count);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Wake up threads waiting on an userland object.
*/
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
static int
umtxq_signal(struct umtx_key *key, int n_wake)
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct umtx_q *uq, *next;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int ret;
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
ret = 0;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
uc = umtxq_getchain(key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_FOREACH_SAFE(uq, &uc->uc_queue, uq_link, next) {
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
if (umtx_key_match(&uq->uq_key, key)) {
umtxq_remove(uq);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
wakeup(uq);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
if (++ret >= n_wake)
break;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
}
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
return (ret);
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Wake up specified thread.
*/
static inline void
umtxq_signal_thread(struct umtx_q *uq)
{
struct umtxq_chain *uc;
uc = umtxq_getchain(&uq->uq_key);
UMTXQ_LOCKED_ASSERT(uc);
umtxq_remove(uq);
wakeup(uq);
}
/*
* Put thread into sleep state, before sleeping, check if
* thread was removed from umtx queue.
*/
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
static inline int
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_sleep(struct umtx_q *uq, const char *wmesg, int timo)
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtxq_chain *uc;
int error;
uc = umtxq_getchain(&uq->uq_key);
UMTXQ_LOCKED_ASSERT(uc);
if (!(uq->uq_flags & UQF_UMTXQ))
return (0);
error = msleep(uq, &uc->uc_lock, PCATCH, wmesg, timo);
Return ETIMEDOUT when thread is timeouted since POSIX thread APIs expect ETIMEDOUT not EAGAIN, this simplifies userland code a bit.
2005-01-06 02:08:34 +00:00
if (error == EWOULDBLOCK)
error = ETIMEDOUT;
return (error);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Convert userspace address into unique logical address.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtx_key_get(void *addr, int type, int share, struct umtx_key *key)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct thread *td = curthread;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
vm_map_t map;
vm_map_entry_t entry;
vm_pindex_t pindex;
vm_prot_t prot;
boolean_t wired;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
key->type = type;
if (share == THREAD_SHARE) {
key->shared = 0;
key->info.private.vs = td->td_proc->p_vmspace;
key->info.private.addr = (uintptr_t)addr;
Eliminate an unnecessary `if' statement.
2006-10-25 06:28:23 +00:00
} else {
MPASS(share == PROCESS_SHARE || share == AUTO_SHARE);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
map = &td->td_proc->p_vmspace->vm_map;
if (vm_map_lookup(&map, (vm_offset_t)addr, VM_PROT_WRITE,
&entry, &key->info.shared.object, &pindex, &prot,
&wired) != KERN_SUCCESS) {
return EFAULT;
}
Axe unused code.
2006-02-04 06:36:39 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if ((share == PROCESS_SHARE) ||
(share == AUTO_SHARE &&
VM_INHERIT_SHARE == entry->inheritance)) {
key->shared = 1;
key->info.shared.offset = entry->offset + entry->start -
(vm_offset_t)addr;
vm_object_reference(key->info.shared.object);
} else {
key->shared = 0;
key->info.private.vs = td->td_proc->p_vmspace;
key->info.private.addr = (uintptr_t)addr;
}
vm_map_lookup_done(map, entry);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_hash(key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
return (0);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Release key.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static inline void
umtx_key_release(struct umtx_key *key)
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (key->shared)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
vm_object_deallocate(key->info.shared.object);
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Lock a umtx object.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
_do_lock_umtx(struct thread *td, struct umtx *umtx, u_long id, int timo)
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
{
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
struct umtx_q *uq;
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
u_long owner;
u_long old;
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
int error = 0;
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
uq = td->td_umtxq;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
/*
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
* Care must be exercised when dealing with umtx structure. It
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
* can fault on any access.
*/
for (;;) {
/*
* Try the uncontested case. This should be done in userland.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = casuword(&umtx->u_owner, UMTX_UNOWNED, id);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
/* The acquire succeeded. */
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
if (owner == UMTX_UNOWNED)
return (0);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
/* If no one owns it but it is contested try to acquire it. */
if (owner == UMTX_CONTESTED) {
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = casuword(&umtx->u_owner,
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
UMTX_CONTESTED, id | UMTX_CONTESTED);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
if (owner == UMTX_CONTESTED)
return (0);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
/* If this failed the lock has changed, restart. */
continue;
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
}
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (error != 0)
return (error);
if ((error = umtx_key_get(umtx, TYPE_SIMPLE_LOCK,
AUTO_SHARE, &uq->uq_key)) != 0)
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
return (error);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_insert(uq);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
/*
* Set the contested bit so that a release in user space
* knows to use the system call for unlock. If this fails
* either some one else has acquired the lock or it has been
* released.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
old = casuword(&umtx->u_owner, owner, owner | UMTX_CONTESTED);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
/* The address was invalid. */
- Make casuptr return the old value of the location we're trying to update, and change the umtx code to expect this. Reviewed by: jeff
2003-04-02 08:02:27 +00:00
if (old == -1) {
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_lock(&uq->uq_key);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
return (EFAULT);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
}
/*
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
* We set the contested bit, sleep. Otherwise the lock changed
Fix umtx locking, for libthr, in the kernel. 1. There was a race condition between a thread unlocking a umtx and the thread contesting it. If the unlocking thread won the race it may try to wakeup a thread that was not yet in msleep(). The contesting thread would then go to sleep to await a wakeup that would never come. It's not possible to close the race by using a lock because calls to casuptr() may have to fault a page in from swap. Instead, the race was closed by introducing a flag that the unlocking thread will set when waking up a thread. The contesting thread will check for this flag before going to sleep. For now the flag is kept in td_flags, but it may be better to use some other member or create a new one because of the possible performance/contention issues of having to own sched_lock. Thanks to jhb for pointing me in the right direction on this one. 2. Once a umtx was contested all future locks and unlocks were happening in the kernel, regardless of whether it was contested or not. To prevent this from happening, when a thread locks a umtx it checks the queue for that umtx and unsets the contested bit if there are no other threads waiting on it. Again, this is slightly more complicated than it needs to be because we can't hold a lock across casuptr(). So, the thread has to check the queue again after unseting the bit, and reset the contested bit if it finds that another thread has put itself on the queue in the mean time. 3. Remove the if... block for unlocking an uncontested umtx, and replace it with a KASSERT. The _only_ time a thread should be unlocking a umtx in the kernel is if it is contested.
2003-07-17 11:06:40 +00:00
* and we need to retry or we lost a race to the thread
* unlocking the umtx.
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
*/
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_lock(&uq->uq_key);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (old == owner)
error = umtxq_sleep(uq, "umtx", timo);
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
}
Move the decision on whether to unset the contested bit or not from lock to unlock time. Suggested by: jhb
2003-07-18 17:58:37 +00:00
return (0);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Lock a umtx object.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
do_lock_umtx(struct thread *td, struct umtx *umtx, u_long id,
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
struct timespec *timeout)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
{
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
struct timespec ts, ts2, ts3;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct timeval tv;
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
int error;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
if (timeout == NULL) {
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
error = _do_lock_umtx(td, umtx, id, 0);
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
/* Mutex locking is restarted if it is interrupted. */
if (error == EINTR)
error = ERESTART;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
} else {
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
getnanouptime(&ts);
timespecadd(&ts, timeout);
TIMESPEC_TO_TIMEVAL(&tv, timeout);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
for (;;) {
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
error = _do_lock_umtx(td, umtx, id, tvtohz(&tv));
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
if (error != ETIMEDOUT)
break;
getnanouptime(&ts2);
if (timespeccmp(&ts2, &ts, >=)) {
Return ETIMEDOUT when thread is timeouted since POSIX thread APIs expect ETIMEDOUT not EAGAIN, this simplifies userland code a bit.
2005-01-06 02:08:34 +00:00
error = ETIMEDOUT;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
break;
}
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
ts3 = ts;
timespecsub(&ts3, &ts2);
TIMESPEC_TO_TIMEVAL(&tv, &ts3);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
/* Timed-locking is not restarted. */
if (error == ERESTART)
error = EINTR;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
return (error);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Unlock a umtx object.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
do_unlock_umtx(struct thread *td, struct umtx *umtx, u_long id)
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
{
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct umtx_key key;
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
u_long owner;
u_long old;
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
int error;
int count;
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
/*
* Make sure we own this mtx.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = fuword(__DEVOLATILE(u_long *, &umtx->u_owner));
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (owner == -1)
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
return (EFAULT);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
if ((owner & ~UMTX_CONTESTED) != id)
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
return (EPERM);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* This should be done in userland */
if ((owner & UMTX_CONTESTED) == 0) {
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
old = casuword(&umtx->u_owner, owner, UMTX_UNOWNED);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (old == -1)
return (EFAULT);
if (old == owner)
return (0);
Make sure we get new m_owner value if we can not unlock it in uncontested case. Reorder statements in do_unlock_umutex.
2006-09-02 02:41:33 +00:00
owner = old;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* We should only ever be in here for contested locks */
if ((error = umtx_key_get(umtx, TYPE_SIMPLE_LOCK, AUTO_SHARE,
&key)) != 0)
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
count = umtxq_count(&key);
umtxq_unlock(&key);
Move the decision on whether to unset the contested bit or not from lock to unlock time. Suggested by: jhb
2003-07-18 17:58:37 +00:00
/*
* When unlocking the umtx, it must be marked as unowned if
* there is zero or one thread only waiting for it.
* Otherwise, it must be marked as contested.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
old = casuword(&umtx->u_owner, owner,
count <= 1 ? UMTX_UNOWNED : UMTX_CONTESTED);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
umtxq_lock(&key);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_signal(&key,1);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
if (old == -1)
return (EFAULT);
1. use per-chain mutex instead of global mutex to reduce lock collision. 2. Fix two race conditions. One is between _umtx_unlock and signal, also a thread was marked TDF_UMTXWAKEUP by _umtx_unlock, it is possible a signal delivered to the thread will cause msleep returns EINTR, and the thread breaks out of loop, this causes umtx ownership is not transfered to the thread. Another is in _umtx_unlock itself, when the function sets the umtx to UMTX_UNOWNED state, a new thread can come in and lock the umtx, also the function tries to set contested bit flag, but it will fail. Although the function will wake a blocked thread, if that thread breaks out of loop by signal, no contested bit will be set.
2004-11-30 12:02:53 +00:00
if (old != owner)
return (EINVAL);
- Remove the blocked pointer from the umtx structure. - Use a hash of umtx queues to queue blocked threads. We hash on pid and the virtual address of the umtx structure. This eliminates cases where we previously held a lock across a casuptr call. Reviwed by: jhb (quickly)
2003-06-03 05:24:46 +00:00
return (0);
- Add an api for doing smp safe locks in userland. - umtx_lock() is defined as an inline in umtx.h. It tries to do an uncontested acquire of a lock which falls back to the _umtx_lock() system-call if that fails. - umtx_unlock() is also an inline which falls back to _umtx_unlock() if the uncontested unlock fails. - Locks are keyed off of the thr_id_t of the currently running thread which is currently just the pointer to the 'struct thread' in kernel. - _umtx_lock() uses the proc pointer to synchronize access to blocked thread queues which are stored in the first blocked thread.
2003-04-01 01:10:42 +00:00
}
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
#ifdef COMPAT_IA32
/*
* Lock a umtx object.
*/
static int
_do_lock_umtx32(struct thread *td, uint32_t *m, uint32_t id, int timo)
{
struct umtx_q *uq;
uint32_t owner;
uint32_t old;
int error = 0;
uq = td->td_umtxq;
/*
* Care must be exercised when dealing with umtx structure. It
* can fault on any access.
*/
for (;;) {
/*
* Try the uncontested case. This should be done in userland.
*/
owner = casuword32(m, UMUTEX_UNOWNED, id);
/* The acquire succeeded. */
if (owner == UMUTEX_UNOWNED)
return (0);
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
/* If no one owns it but it is contested try to acquire it. */
if (owner == UMUTEX_CONTESTED) {
owner = casuword32(m,
UMUTEX_CONTESTED, id | UMUTEX_CONTESTED);
if (owner == UMUTEX_CONTESTED)
return (0);
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
/* If this failed the lock has changed, restart. */
continue;
}
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
if (error != 0)
return (error);
if ((error = umtx_key_get(m, TYPE_SIMPLE_LOCK,
AUTO_SHARE, &uq->uq_key)) != 0)
return (error);
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_insert(uq);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
/*
* Set the contested bit so that a release in user space
* knows to use the system call for unlock. If this fails
* either some one else has acquired the lock or it has been
* released.
*/
old = casuword32(m, owner, owner | UMUTEX_CONTESTED);
/* The address was invalid. */
if (old == -1) {
umtxq_lock(&uq->uq_key);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
return (EFAULT);
}
/*
* We set the contested bit, sleep. Otherwise the lock changed
* and we need to retry or we lost a race to the thread
* unlocking the umtx.
*/
umtxq_lock(&uq->uq_key);
if (old == owner)
error = umtxq_sleep(uq, "umtx", timo);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
}
return (0);
}
/*
* Lock a umtx object.
*/
static int
do_lock_umtx32(struct thread *td, void *m, uint32_t id,
struct timespec *timeout)
{
struct timespec ts, ts2, ts3;
struct timeval tv;
int error;
if (timeout == NULL) {
error = _do_lock_umtx32(td, m, id, 0);
/* Mutex locking is restarted if it is interrupted. */
if (error == EINTR)
error = ERESTART;
} else {
getnanouptime(&ts);
timespecadd(&ts, timeout);
TIMESPEC_TO_TIMEVAL(&tv, timeout);
for (;;) {
error = _do_lock_umtx32(td, m, id, tvtohz(&tv));
if (error != ETIMEDOUT)
break;
getnanouptime(&ts2);
if (timespeccmp(&ts2, &ts, >=)) {
error = ETIMEDOUT;
break;
}
ts3 = ts;
timespecsub(&ts3, &ts2);
TIMESPEC_TO_TIMEVAL(&tv, &ts3);
}
/* Timed-locking is not restarted. */
if (error == ERESTART)
error = EINTR;
}
return (error);
}
/*
* Unlock a umtx object.
*/
static int
do_unlock_umtx32(struct thread *td, uint32_t *m, uint32_t id)
{
struct umtx_key key;
uint32_t owner;
uint32_t old;
int error;
int count;
/*
* Make sure we own this mtx.
*/
owner = fuword32(m);
if (owner == -1)
return (EFAULT);
if ((owner & ~UMUTEX_CONTESTED) != id)
return (EPERM);
/* This should be done in userland */
if ((owner & UMUTEX_CONTESTED) == 0) {
old = casuword32(m, owner, UMUTEX_UNOWNED);
if (old == -1)
return (EFAULT);
if (old == owner)
return (0);
owner = old;
}
/* We should only ever be in here for contested locks */
if ((error = umtx_key_get(m, TYPE_SIMPLE_LOCK, AUTO_SHARE,
&key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
count = umtxq_count(&key);
umtxq_unlock(&key);
/*
* When unlocking the umtx, it must be marked as unowned if
* there is zero or one thread only waiting for it.
* Otherwise, it must be marked as contested.
*/
old = casuword32(m, owner,
count <= 1 ? UMUTEX_UNOWNED : UMUTEX_CONTESTED);
umtxq_lock(&key);
umtxq_signal(&key,1);
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
if (old == -1)
return (EFAULT);
if (old != owner)
return (EINVAL);
return (0);
}
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Fetch and compare value, sleep on the address if value is not changed.
*/
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
static int
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
do_wait(struct thread *td, void *addr, u_long id,
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
struct timespec *timeout, int compat32)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
{
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
struct umtx_q *uq;
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
struct timespec ts, ts2, ts3;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
struct timeval tv;
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
u_long tmp;
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
int error = 0;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
uq = td->td_umtxq;
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
if ((error = umtx_key_get(addr, TYPE_SIMPLE_WAIT, AUTO_SHARE,
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
&uq->uq_key)) != 0)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
return (error);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&uq->uq_key);
umtxq_insert(uq);
umtxq_unlock(&uq->uq_key);
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
if (compat32 == 0)
tmp = fuword(addr);
else
tmp = fuword32(addr);
Make umtx_wait and umtx_wake more like linux futex does, it is more general than previous. It also lets me implement cancelable point in thread library. Also in theory, umtx_lock and umtx_unlock can be implemented by using umtx_wait and umtx_wake, all atomic operations can be done in userland without kernel's casuptr() function.
2004-12-30 02:56:17 +00:00
if (tmp != id) {
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_lock(&uq->uq_key);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
} else if (timeout == NULL) {
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_lock(&uq->uq_key);
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
error = umtxq_sleep(uq, "uwait", 0);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_remove(uq);
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_unlock(&uq->uq_key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
} else {
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
getnanouptime(&ts);
timespecadd(&ts, timeout);
TIMESPEC_TO_TIMEVAL(&tv, timeout);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&uq->uq_key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
for (;;) {
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
error = umtxq_sleep(uq, "uwait", tvtohz(&tv));
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (!(uq->uq_flags & UQF_UMTXQ))
break;
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
if (error != ETIMEDOUT)
break;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_unlock(&uq->uq_key);
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
getnanouptime(&ts2);
if (timespeccmp(&ts2, &ts, >=)) {
error = ETIMEDOUT;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&uq->uq_key);
make umtx timeout relative so userland can select different clock type, e.g, CLOCK_REALTIME or CLOCK_MONOTONIC. merge umtx_wait and umtx_timedwait into single function.
2005-01-14 13:38:15 +00:00
break;
}
ts3 = ts;
timespecsub(&ts3, &ts2);
TIMESPEC_TO_TIMEVAL(&tv, &ts3);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&uq->uq_key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
Allocate umtx_q from heap instead of stack, this avoids page fault panic in kernel under heavy swapping.
2005-03-05 09:15:03 +00:00
umtx_key_release(&uq->uq_key);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
if (error == ERESTART)
error = EINTR;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
return (error);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Wake up threads sleeping on the specified address.
*/
do umtx_wake at userland thread exit address, so that others userland threads can wait for a thread to exit, and safely assume that the thread has left userland and is no longer using its userland stack, this is necessary for pthread_join when a thread is waiting for another thread to exit which has user customized stack, after pthread_join returns, the userland stack can be reused for other purposes, without this change, the joiner thread has to spin at the address to ensure the thread is really exited.
2005-10-26 06:55:46 +00:00
int
kern_umtx_wake(struct thread *td, void *uaddr, int n_wake)
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
{
struct umtx_key key;
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
int ret;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if ((ret = umtx_key_get(uaddr, TYPE_SIMPLE_WAIT, AUTO_SHARE,
&key)) != 0)
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
return (ret);
Add umtxq_lock/unlock around umtx_signal, fix debug kernel compiling, let umtx_lock returns EINTR when it returns ERESTART, this lets userland have chance to back off mtx lock code when needed.
2004-12-24 11:59:20 +00:00
umtxq_lock(&key);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
ret = umtxq_signal(&key, n_wake);
Add umtxq_lock/unlock around umtx_signal, fix debug kernel compiling, let umtx_lock returns EINTR when it returns ERESTART, this lets userland have chance to back off mtx lock code when needed.
2004-12-24 11:59:20 +00:00
umtxq_unlock(&key);
1. Fix race condition between umtx lock and unlock, heavy testing on SMP can explore the bug. 2. Let umtx_wake returns number of threads have been woken.
2004-12-24 11:30:55 +00:00
umtx_key_release(&key);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
return (0);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Lock PTHREAD_PRIO_NONE protocol POSIX mutex.
*/
static int
_do_lock_normal(struct thread *td, struct umutex *m, uint32_t flags, int timo,
int try)
Add user priority loaning code to support priority propagation for 1:1 threading's POSIX priority mutexes, the code is no-op unless priority-aware umtx code is committed.
2006-08-25 06:12:53 +00:00
{
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
struct umtx_q *uq;
uint32_t owner, old, id;
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#ifdef SMP
int spincount;
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int error = 0;
Add user priority loaning code to support priority propagation for 1:1 threading's POSIX priority mutexes, the code is no-op unless priority-aware umtx code is committed.
2006-08-25 06:12:53 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
id = td->td_tid;
uq = td->td_umtxq;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#ifdef SMP
if (smp_cpus > 1) {
spincount = fuword32(&m->m_spincount);
if (spincount == 0)
spincount = umtx_dflt_spins;
if (spincount > umtx_max_spins)
spincount = umtx_max_spins;
} else
spincount = 0;
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Care must be exercised when dealing with umtx structure. It
* can fault on any access.
*/
for (;;) {
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#ifdef SMP
try_unowned:
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Try the uncontested case. This should be done in userland.
*/
owner = casuword32(&m->m_owner, UMUTEX_UNOWNED, id);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* The acquire succeeded. */
if (owner == UMUTEX_UNOWNED)
return (0);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
/* If no one owns it but it is contested try to acquire it. */
if (owner == UMUTEX_CONTESTED) {
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#ifdef SMP
try_contested:
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
owner = casuword32(&m->m_owner,
UMUTEX_CONTESTED, id | UMUTEX_CONTESTED);
if (owner == UMUTEX_CONTESTED)
return (0);
/* The address was invalid. */
if (owner == -1)
return (EFAULT);
/* If this failed the lock has changed, restart. */
continue;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if ((flags & UMUTEX_ERROR_CHECK) != 0 &&
(owner & ~UMUTEX_CONTESTED) == id)
return (EDEADLK);
if (try != 0)
return (EBUSY);
Add a lwpid field into per-cpu structure, the lwpid represents current running thread's id on each cpu. This allow us to add in-kernel adaptive spin for user level mutex. While spinning in user space is possible, without correct thread running state exported from kernel, it hardly can be implemented efficiently without wasting cpu cycles, however exporting thread running state unlikely will be implemented soon as it has to design and stablize interfaces. This implementation is transparent to user space, it can be disabled dynamically. With this change, mutex ping-pong program's performance is improved massively on SMP machine. performance of mysql super-smack select benchmark is increased about 7% on Intel dual dual-core2 Xeon machine, it indicates on systems which have bunch of cpus and system-call overhead is low (athlon64, opteron, and core-2 are known to be fast), the adaptive spin does help performance. Added sysctls: kern.threads.umtx_dflt_spins if the sysctl value is non-zero, a zero umutex.m_spincount will cause the sysctl value to be used a spin cycle count. kern.threads.umtx_max_spins the sysctl sets upper limit of spin cycle count. Tested on: Athlon64 X2 3800+, Dual Xeon 5130
2006-12-20 04:40:39 +00:00
#ifdef SMP
if (spincount > 0 && (owner & ~UMUTEX_CONTESTED) != id) {
int i, found = 0;
struct pcpu *pcpu = NULL;
/* Look for a cpu the owner is running on */
for (i = 0; i < MAXCPU; i++) {
if (CPU_ABSENT(i))
continue;
pcpu = pcpu_find(i);
if ((owner & ~UMUTEX_CONTESTED) == pcpu->pc_curtid) {
found = 1;
break;
}
}
if (__predict_false(!found))
goto end_spin;
while ((owner & ~UMUTEX_CONTESTED) == pcpu->pc_curtid &&
(owner & ~UMUTEX_CONTESTED) != id) {
if (--spincount <= 0)
break;
if ((td->td_flags &
(TDF_NEEDRESCHED|TDF_ASTPENDING|TDF_NEEDSIGCHK)) ||
P_SHOULDSTOP(td->td_proc))
break;
owner = fuword32(__DEVOLATILE(uint32_t *, &m->m_owner));
if (owner == UMUTEX_UNOWNED)
goto try_unowned;
if (owner == UMUTEX_CONTESTED)
goto try_contested;
cpu_spinwait();
}
}
end_spin:
spincount = 0;
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
if (error != 0)
return (error);
if ((error = umtx_key_get(m, TYPE_NORMAL_UMUTEX,
GET_SHARE(flags), &uq->uq_key)) != 0)
return (error);
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_insert(uq);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
/*
* Set the contested bit so that a release in user space
* knows to use the system call for unlock. If this fails
* either some one else has acquired the lock or it has been
* released.
*/
old = casuword32(&m->m_owner, owner, owner | UMUTEX_CONTESTED);
/* The address was invalid. */
if (old == -1) {
umtxq_lock(&uq->uq_key);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
return (EFAULT);
}
/*
* We set the contested bit, sleep. Otherwise the lock changed
* and we need to retry or we lost a race to the thread
* unlocking the umtx.
*/
umtxq_lock(&uq->uq_key);
if (old == owner)
error = umtxq_sleep(uq, "umtxn", timo);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
}
return (0);
}
/*
* Lock PTHREAD_PRIO_NONE protocol POSIX mutex.
*/
/*
* Unlock PTHREAD_PRIO_NONE protocol POSIX mutex.
*/
static int
do_unlock_normal(struct thread *td, struct umutex *m, uint32_t flags)
{
struct umtx_key key;
uint32_t owner, old, id;
int error;
int count;
id = td->td_tid;
/*
* Make sure we own this mtx.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = fuword32(__DEVOLATILE(uint32_t *, &m->m_owner));
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (owner == -1)
return (EFAULT);
if ((owner & ~UMUTEX_CONTESTED) != id)
return (EPERM);
/* This should be done in userland */
if ((owner & UMUTEX_CONTESTED) == 0) {
old = casuword32(&m->m_owner, owner, UMUTEX_UNOWNED);
if (old == -1)
return (EFAULT);
if (old == owner)
return (0);
Make sure we get new m_owner value if we can not unlock it in uncontested case. Reorder statements in do_unlock_umutex.
2006-09-02 02:41:33 +00:00
owner = old;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
/* We should only ever be in here for contested locks */
if ((error = umtx_key_get(m, TYPE_NORMAL_UMUTEX, GET_SHARE(flags),
&key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
count = umtxq_count(&key);
umtxq_unlock(&key);
/*
* When unlocking the umtx, it must be marked as unowned if
* there is zero or one thread only waiting for it.
* Otherwise, it must be marked as contested.
*/
old = casuword32(&m->m_owner, owner,
count <= 1 ? UMUTEX_UNOWNED : UMUTEX_CONTESTED);
umtxq_lock(&key);
umtxq_signal(&key,1);
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
if (old == -1)
return (EFAULT);
if (old != owner)
return (EINVAL);
return (0);
}
static inline struct umtx_pi *
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
umtx_pi_alloc(int flags)
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
{
struct umtx_pi *pi;
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
pi = uma_zalloc(umtx_pi_zone, M_ZERO | flags);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
TAILQ_INIT(&pi->pi_blocked);
atomic_add_int(&umtx_pi_allocated, 1);
return (pi);
}
static inline void
umtx_pi_free(struct umtx_pi *pi)
{
uma_zfree(umtx_pi_zone, pi);
atomic_add_int(&umtx_pi_allocated, -1);
}
/*
* Adjust the thread's position on a pi_state after its priority has been
* changed.
*/
static int
umtx_pi_adjust_thread(struct umtx_pi *pi, struct thread *td)
{
struct umtx_q *uq, *uq1, *uq2;
struct thread *td1;
mtx_assert(&sched_lock, MA_OWNED);
if (pi == NULL)
return (0);
uq = td->td_umtxq;
/*
* Check if the thread needs to be moved on the blocked chain.
* It needs to be moved if either its priority is lower than
* the previous thread or higher than the next thread.
*/
uq1 = TAILQ_PREV(uq, umtxq_head, uq_lockq);
uq2 = TAILQ_NEXT(uq, uq_lockq);
if ((uq1 != NULL && UPRI(td) < UPRI(uq1->uq_thread)) ||
(uq2 != NULL && UPRI(td) > UPRI(uq2->uq_thread))) {
/*
* Remove thread from blocked chain and determine where
* it should be moved to.
*/
TAILQ_REMOVE(&pi->pi_blocked, uq, uq_lockq);
TAILQ_FOREACH(uq1, &pi->pi_blocked, uq_lockq) {
td1 = uq1->uq_thread;
MPASS(td1->td_proc->p_magic == P_MAGIC);
if (UPRI(td1) > UPRI(td))
break;
}
if (uq1 == NULL)
TAILQ_INSERT_TAIL(&pi->pi_blocked, uq, uq_lockq);
else
TAILQ_INSERT_BEFORE(uq1, uq, uq_lockq);
}
return (1);
}
/*
* Propagate priority when a thread is blocked on POSIX
* PI mutex.
*/
static void
umtx_propagate_priority(struct thread *td)
{
struct umtx_q *uq;
struct umtx_pi *pi;
int pri;
mtx_assert(&sched_lock, MA_OWNED);
pri = UPRI(td);
uq = td->td_umtxq;
pi = uq->uq_pi_blocked;
if (pi == NULL)
return;
for (;;) {
td = pi->pi_owner;
if (td == NULL)
return;
MPASS(td->td_proc != NULL);
MPASS(td->td_proc->p_magic == P_MAGIC);
if (UPRI(td) <= pri)
return;
sched_lend_user_prio(td, pri);
/*
* Pick up the lock that td is blocked on.
*/
uq = td->td_umtxq;
pi = uq->uq_pi_blocked;
/* Resort td on the list if needed. */
if (!umtx_pi_adjust_thread(pi, td))
break;
}
}
/*
* Unpropagate priority for a PI mutex when a thread blocked on
* it is interrupted by signal or resumed by others.
*/
static void
umtx_unpropagate_priority(struct umtx_pi *pi)
{
struct umtx_q *uq, *uq_owner;
struct umtx_pi *pi2;
int pri;
mtx_assert(&sched_lock, MA_OWNED);
while (pi != NULL && pi->pi_owner != NULL) {
pri = PRI_MAX;
uq_owner = pi->pi_owner->td_umtxq;
TAILQ_FOREACH(pi2, &uq_owner->uq_pi_contested, pi_link) {
uq = TAILQ_FIRST(&pi2->pi_blocked);
if (uq != NULL) {
if (pri > UPRI(uq->uq_thread))
pri = UPRI(uq->uq_thread);
}
}
if (pri > uq_owner->uq_inherited_pri)
pri = uq_owner->uq_inherited_pri;
sched_unlend_user_prio(pi->pi_owner, pri);
pi = uq_owner->uq_pi_blocked;
}
}
/*
* Insert a PI mutex into owned list.
*/
static void
umtx_pi_setowner(struct umtx_pi *pi, struct thread *owner)
{
struct umtx_q *uq_owner;
uq_owner = owner->td_umtxq;
mtx_assert(&sched_lock, MA_OWNED);
if (pi->pi_owner != NULL)
panic("pi_ower != NULL");
pi->pi_owner = owner;
TAILQ_INSERT_TAIL(&uq_owner->uq_pi_contested, pi, pi_link);
}
/*
* Claim ownership of a PI mutex.
*/
static int
umtx_pi_claim(struct umtx_pi *pi, struct thread *owner)
{
struct umtx_q *uq, *uq_owner;
uq_owner = owner->td_umtxq;
mtx_lock_spin(&sched_lock);
if (pi->pi_owner == owner) {
mtx_unlock_spin(&sched_lock);
return (0);
}
if (pi->pi_owner != NULL) {
/*
* userland may have already messed the mutex, sigh.
*/
mtx_unlock_spin(&sched_lock);
return (EPERM);
}
umtx_pi_setowner(pi, owner);
uq = TAILQ_FIRST(&pi->pi_blocked);
if (uq != NULL) {
int pri;
pri = UPRI(uq->uq_thread);
if (pri < UPRI(owner))
sched_lend_user_prio(owner, pri);
}
mtx_unlock_spin(&sched_lock);
return (0);
}
/*
* Adjust a thread's order position in its blocked PI mutex,
* this may result new priority propagating process.
*/
void
umtx_pi_adjust(struct thread *td, u_char oldpri)
{
struct umtx_q *uq;
struct umtx_pi *pi;
uq = td->td_umtxq;
mtx_assert(&sched_lock, MA_OWNED);
MPASS(TD_ON_UPILOCK(td));
/*
* Pick up the lock that td is blocked on.
*/
pi = uq->uq_pi_blocked;
MPASS(pi != NULL);
/* Resort the turnstile on the list. */
if (!umtx_pi_adjust_thread(pi, td))
return;
/*
* If our priority was lowered and we are at the head of the
* turnstile, then propagate our new priority up the chain.
*/
if (uq == TAILQ_FIRST(&pi->pi_blocked) && UPRI(td) < oldpri)
umtx_propagate_priority(td);
}
/*
* Sleep on a PI mutex.
*/
static int
umtxq_sleep_pi(struct umtx_q *uq, struct umtx_pi *pi,
uint32_t owner, const char *wmesg, int timo)
{
struct umtxq_chain *uc;
struct thread *td, *td1;
struct umtx_q *uq1;
int pri;
int error = 0;
td = uq->uq_thread;
KASSERT(td == curthread, ("inconsistent uq_thread"));
uc = umtxq_getchain(&uq->uq_key);
UMTXQ_LOCKED_ASSERT(uc);
umtxq_insert(uq);
if (pi->pi_owner == NULL) {
/* XXX
* Current, We only support process private PI-mutex,
* non-contended PI-mutexes are locked in userland.
* Process shared PI-mutex should always be initialized
* by kernel and be registered in kernel, locking should
* always be done by kernel to avoid security problems.
* For process private PI-mutex, we can find owner
* thread and boost its priority safely.
*/
PROC_LOCK(curproc);
td1 = thread_find(curproc, owner);
mtx_lock_spin(&sched_lock);
if (td1 != NULL && pi->pi_owner == NULL) {
uq1 = td1->td_umtxq;
umtx_pi_setowner(pi, td1);
}
PROC_UNLOCK(curproc);
} else {
mtx_lock_spin(&sched_lock);
}
TAILQ_FOREACH(uq1, &pi->pi_blocked, uq_lockq) {
pri = UPRI(uq1->uq_thread);
if (pri > UPRI(td))
break;
}
if (uq1 != NULL)
TAILQ_INSERT_BEFORE(uq1, uq, uq_lockq);
else
TAILQ_INSERT_TAIL(&pi->pi_blocked, uq, uq_lockq);
uq->uq_pi_blocked = pi;
td->td_flags |= TDF_UPIBLOCKED;
mtx_unlock_spin(&sched_lock);
umtxq_unlock(&uq->uq_key);
mtx_lock_spin(&sched_lock);
umtx_propagate_priority(td);
mtx_unlock_spin(&sched_lock);
umtxq_lock(&uq->uq_key);
if (uq->uq_flags & UQF_UMTXQ) {
error = msleep(uq, &uc->uc_lock, PCATCH, wmesg, timo);
if (error == EWOULDBLOCK)
error = ETIMEDOUT;
if (uq->uq_flags & UQF_UMTXQ) {
umtxq_busy(&uq->uq_key);
umtxq_remove(uq);
umtxq_unbusy(&uq->uq_key);
}
}
umtxq_unlock(&uq->uq_key);
mtx_lock_spin(&sched_lock);
uq->uq_pi_blocked = NULL;
td->td_flags &= ~TDF_UPIBLOCKED;
TAILQ_REMOVE(&pi->pi_blocked, uq, uq_lockq);
umtx_unpropagate_priority(pi);
mtx_unlock_spin(&sched_lock);
umtxq_lock(&uq->uq_key);
return (error);
}
/*
* Add reference count for a PI mutex.
*/
static void
umtx_pi_ref(struct umtx_pi *pi)
{
struct umtxq_chain *uc;
uc = umtxq_getchain(&pi->pi_key);
UMTXQ_LOCKED_ASSERT(uc);
pi->pi_refcount++;
}
/*
* Decrease reference count for a PI mutex, if the counter
* is decreased to zero, its memory space is freed.
*/
static void
umtx_pi_unref(struct umtx_pi *pi)
{
struct umtxq_chain *uc;
int free = 0;
uc = umtxq_getchain(&pi->pi_key);
UMTXQ_LOCKED_ASSERT(uc);
KASSERT(pi->pi_refcount > 0, ("invalid reference count"));
if (--pi->pi_refcount == 0) {
mtx_lock_spin(&sched_lock);
if (pi->pi_owner != NULL) {
TAILQ_REMOVE(&pi->pi_owner->td_umtxq->uq_pi_contested,
pi, pi_link);
pi->pi_owner = NULL;
}
KASSERT(TAILQ_EMPTY(&pi->pi_blocked),
("blocked queue not empty"));
mtx_unlock_spin(&sched_lock);
TAILQ_REMOVE(&uc->uc_pi_list, pi, pi_hashlink);
free = 1;
}
if (free)
umtx_pi_free(pi);
}
/*
* Find a PI mutex in hash table.
*/
static struct umtx_pi *
umtx_pi_lookup(struct umtx_key *key)
{
struct umtxq_chain *uc;
struct umtx_pi *pi;
uc = umtxq_getchain(key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_FOREACH(pi, &uc->uc_pi_list, pi_hashlink) {
if (umtx_key_match(&pi->pi_key, key)) {
return (pi);
}
}
return (NULL);
}
/*
* Insert a PI mutex into hash table.
*/
static inline void
umtx_pi_insert(struct umtx_pi *pi)
{
struct umtxq_chain *uc;
uc = umtxq_getchain(&pi->pi_key);
UMTXQ_LOCKED_ASSERT(uc);
TAILQ_INSERT_TAIL(&uc->uc_pi_list, pi, pi_hashlink);
}
/*
* Lock a PI mutex.
*/
static int
_do_lock_pi(struct thread *td, struct umutex *m, uint32_t flags, int timo,
int try)
{
struct umtx_q *uq;
struct umtx_pi *pi, *new_pi;
uint32_t id, owner, old;
int error;
id = td->td_tid;
uq = td->td_umtxq;
if ((error = umtx_key_get(m, TYPE_PI_UMUTEX, GET_SHARE(flags),
&uq->uq_key)) != 0)
return (error);
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
umtxq_lock(&uq->uq_key);
pi = umtx_pi_lookup(&uq->uq_key);
if (pi == NULL) {
new_pi = umtx_pi_alloc(M_NOWAIT);
if (new_pi == NULL) {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_unlock(&uq->uq_key);
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
new_pi = umtx_pi_alloc(M_WAITOK);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
new_pi->pi_key = uq->uq_key;
umtxq_lock(&uq->uq_key);
pi = umtx_pi_lookup(&uq->uq_key);
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
if (pi != NULL) {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtx_pi_free(new_pi);
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
new_pi = NULL;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
}
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
if (new_pi != NULL) {
new_pi->pi_key = uq->uq_key;
umtx_pi_insert(new_pi);
pi = new_pi;
}
}
umtx_pi_ref(pi);
umtxq_unlock(&uq->uq_key);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
/*
* Care must be exercised when dealing with umtx structure. It
* can fault on any access.
*/
for (;;) {
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Try the uncontested case. This should be done in userland.
*/
owner = casuword32(&m->m_owner, UMUTEX_UNOWNED, id);
/* The acquire succeeded. */
if (owner == UMUTEX_UNOWNED) {
error = 0;
break;
}
/* The address was invalid. */
if (owner == -1) {
error = EFAULT;
break;
}
/* If no one owns it but it is contested try to acquire it. */
if (owner == UMUTEX_CONTESTED) {
owner = casuword32(&m->m_owner,
UMUTEX_CONTESTED, id | UMUTEX_CONTESTED);
if (owner == UMUTEX_CONTESTED) {
umtxq_lock(&uq->uq_key);
error = umtx_pi_claim(pi, td);
umtxq_unlock(&uq->uq_key);
break;
}
/* The address was invalid. */
if (owner == -1) {
error = EFAULT;
break;
}
/* If this failed the lock has changed, restart. */
continue;
}
if ((flags & UMUTEX_ERROR_CHECK) != 0 &&
(owner & ~UMUTEX_CONTESTED) == id) {
error = EDEADLK;
break;
}
if (try != 0) {
error = EBUSY;
break;
}
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
if (error != 0)
break;
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
/*
* Set the contested bit so that a release in user space
* knows to use the system call for unlock. If this fails
* either some one else has acquired the lock or it has been
* released.
*/
old = casuword32(&m->m_owner, owner, owner | UMUTEX_CONTESTED);
/* The address was invalid. */
if (old == -1) {
umtxq_lock(&uq->uq_key);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
error = EFAULT;
break;
}
umtxq_lock(&uq->uq_key);
umtxq_unbusy(&uq->uq_key);
/*
* We set the contested bit, sleep. Otherwise the lock changed
* and we need to retry or we lost a race to the thread
* unlocking the umtx.
*/
if (old == owner)
error = umtxq_sleep_pi(uq, pi, owner & ~UMUTEX_CONTESTED,
"umtxpi", timo);
umtxq_unlock(&uq->uq_key);
}
Optimize umtx_lock_pi() a bit by moving some heavy code out of the loop, make a fast path when a umtx_pi can be allocated without being blocked.
2006-10-26 09:33:34 +00:00
umtxq_lock(&uq->uq_key);
umtx_pi_unref(pi);
umtxq_unlock(&uq->uq_key);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtx_key_release(&uq->uq_key);
return (error);
}
/*
* Unlock a PI mutex.
*/
static int
do_unlock_pi(struct thread *td, struct umutex *m, uint32_t flags)
{
struct umtx_key key;
struct umtx_q *uq_first, *uq_first2, *uq_me;
struct umtx_pi *pi, *pi2;
uint32_t owner, old, id;
int error;
int count;
int pri;
id = td->td_tid;
/*
* Make sure we own this mtx.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = fuword32(__DEVOLATILE(uint32_t *, &m->m_owner));
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (owner == -1)
return (EFAULT);
if ((owner & ~UMUTEX_CONTESTED) != id)
return (EPERM);
/* This should be done in userland */
if ((owner & UMUTEX_CONTESTED) == 0) {
old = casuword32(&m->m_owner, owner, UMUTEX_UNOWNED);
if (old == -1)
return (EFAULT);
if (old == owner)
return (0);
Make sure we get new m_owner value if we can not unlock it in uncontested case. Reorder statements in do_unlock_umutex.
2006-09-02 02:41:33 +00:00
owner = old;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
/* We should only ever be in here for contested locks */
if ((error = umtx_key_get(m, TYPE_PI_UMUTEX, GET_SHARE(flags),
&key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
count = umtxq_count_pi(&key, &uq_first);
if (uq_first != NULL) {
pi = uq_first->uq_pi_blocked;
if (pi->pi_owner != curthread) {
umtxq_unbusy(&key);
umtxq_unlock(&key);
/* userland messed the mutex */
return (EPERM);
}
uq_me = curthread->td_umtxq;
mtx_lock_spin(&sched_lock);
pi->pi_owner = NULL;
TAILQ_REMOVE(&uq_me->uq_pi_contested, pi, pi_link);
uq_first = TAILQ_FIRST(&pi->pi_blocked);
pri = PRI_MAX;
TAILQ_FOREACH(pi2, &uq_me->uq_pi_contested, pi_link) {
uq_first2 = TAILQ_FIRST(&pi2->pi_blocked);
if (uq_first2 != NULL) {
if (pri > UPRI(uq_first2->uq_thread))
pri = UPRI(uq_first2->uq_thread);
}
}
sched_unlend_user_prio(curthread, pri);
mtx_unlock_spin(&sched_lock);
}
umtxq_unlock(&key);
/*
* When unlocking the umtx, it must be marked as unowned if
* there is zero or one thread only waiting for it.
* Otherwise, it must be marked as contested.
*/
old = casuword32(&m->m_owner, owner,
count <= 1 ? UMUTEX_UNOWNED : UMUTEX_CONTESTED);
umtxq_lock(&key);
if (uq_first != NULL)
umtxq_signal_thread(uq_first);
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
if (old == -1)
return (EFAULT);
if (old != owner)
return (EINVAL);
return (0);
}
/*
* Lock a PP mutex.
*/
static int
_do_lock_pp(struct thread *td, struct umutex *m, uint32_t flags, int timo,
int try)
{
struct umtx_q *uq, *uq2;
struct umtx_pi *pi;
uint32_t ceiling;
uint32_t owner, id;
int error, pri, old_inherited_pri, su;
id = td->td_tid;
uq = td->td_umtxq;
if ((error = umtx_key_get(m, TYPE_PP_UMUTEX, GET_SHARE(flags),
&uq->uq_key)) != 0)
return (error);
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
su = (priv_check(td, PRIV_SCHED_RTPRIO) == 0);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
for (;;) {
old_inherited_pri = uq->uq_inherited_pri;
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
ceiling = RTP_PRIO_MAX - fuword32(&m->m_ceilings[0]);
if (ceiling > RTP_PRIO_MAX) {
error = EINVAL;
goto out;
}
mtx_lock_spin(&sched_lock);
if (UPRI(td) < PRI_MIN_REALTIME + ceiling) {
mtx_unlock_spin(&sched_lock);
error = EINVAL;
goto out;
}
if (su && PRI_MIN_REALTIME + ceiling < uq->uq_inherited_pri) {
uq->uq_inherited_pri = PRI_MIN_REALTIME + ceiling;
if (uq->uq_inherited_pri < UPRI(td))
sched_lend_user_prio(td, uq->uq_inherited_pri);
}
mtx_unlock_spin(&sched_lock);
owner = casuword32(&m->m_owner,
UMUTEX_CONTESTED, id | UMUTEX_CONTESTED);
if (owner == UMUTEX_CONTESTED) {
error = 0;
break;
}
/* The address was invalid. */
if (owner == -1) {
error = EFAULT;
break;
}
if ((flags & UMUTEX_ERROR_CHECK) != 0 &&
(owner & ~UMUTEX_CONTESTED) == id) {
error = EDEADLK;
break;
}
if (try != 0) {
error = EBUSY;
break;
}
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
if (error != 0)
break;
umtxq_lock(&uq->uq_key);
umtxq_insert(uq);
umtxq_unbusy(&uq->uq_key);
error = umtxq_sleep(uq, "umtxpp", timo);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
mtx_lock_spin(&sched_lock);
uq->uq_inherited_pri = old_inherited_pri;
pri = PRI_MAX;
TAILQ_FOREACH(pi, &uq->uq_pi_contested, pi_link) {
uq2 = TAILQ_FIRST(&pi->pi_blocked);
if (uq2 != NULL) {
if (pri > UPRI(uq2->uq_thread))
pri = UPRI(uq2->uq_thread);
}
}
if (pri > uq->uq_inherited_pri)
pri = uq->uq_inherited_pri;
sched_unlend_user_prio(td, pri);
mtx_unlock_spin(&sched_lock);
}
if (error != 0) {
mtx_lock_spin(&sched_lock);
uq->uq_inherited_pri = old_inherited_pri;
pri = PRI_MAX;
TAILQ_FOREACH(pi, &uq->uq_pi_contested, pi_link) {
uq2 = TAILQ_FIRST(&pi->pi_blocked);
if (uq2 != NULL) {
if (pri > UPRI(uq2->uq_thread))
pri = UPRI(uq2->uq_thread);
}
}
if (pri > uq->uq_inherited_pri)
pri = uq->uq_inherited_pri;
sched_unlend_user_prio(td, pri);
mtx_unlock_spin(&sched_lock);
}
out:
umtxq_lock(&uq->uq_key);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
return (error);
}
/*
* Unlock a PP mutex.
*/
static int
do_unlock_pp(struct thread *td, struct umutex *m, uint32_t flags)
{
struct umtx_key key;
struct umtx_q *uq, *uq2;
struct umtx_pi *pi;
uint32_t owner, id;
uint32_t rceiling;
Check if it is root user in do_unlock_pp.
2006-09-03 00:07:37 +00:00
int error, pri, new_inherited_pri, su;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
id = td->td_tid;
uq = td->td_umtxq;
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
su = (priv_check(td, PRIV_SCHED_RTPRIO) == 0);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Make sure we own this mtx.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
owner = fuword32(__DEVOLATILE(uint32_t *, &m->m_owner));
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
if (owner == -1)
return (EFAULT);
if ((owner & ~UMUTEX_CONTESTED) != id)
return (EPERM);
error = copyin(&m->m_ceilings[1], &rceiling, sizeof(uint32_t));
if (error != 0)
return (error);
if (rceiling == -1)
new_inherited_pri = PRI_MAX;
else {
rceiling = RTP_PRIO_MAX - rceiling;
if (rceiling > RTP_PRIO_MAX)
return (EINVAL);
new_inherited_pri = PRI_MIN_REALTIME + rceiling;
}
if ((error = umtx_key_get(m, TYPE_PP_UMUTEX, GET_SHARE(flags),
&key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
umtxq_unlock(&key);
/*
* For priority protected mutex, always set unlocked state
* to UMUTEX_CONTESTED, so that userland always enters kernel
* to lock the mutex, it is necessary because thread priority
* has to be adjusted for such mutex.
*/
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
error = suword32(__DEVOLATILE(uint32_t *, &m->m_owner),
UMUTEX_CONTESTED);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
umtxq_lock(&key);
if (error == 0)
umtxq_signal(&key, 1);
umtxq_unbusy(&key);
umtxq_unlock(&key);
if (error == -1)
error = EFAULT;
else {
mtx_lock_spin(&sched_lock);
Check if it is root user in do_unlock_pp.
2006-09-03 00:07:37 +00:00
if (su != 0)
uq->uq_inherited_pri = new_inherited_pri;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
pri = PRI_MAX;
TAILQ_FOREACH(pi, &uq->uq_pi_contested, pi_link) {
uq2 = TAILQ_FIRST(&pi->pi_blocked);
if (uq2 != NULL) {
if (pri > UPRI(uq2->uq_thread))
pri = UPRI(uq2->uq_thread);
}
}
if (pri > uq->uq_inherited_pri)
pri = uq->uq_inherited_pri;
sched_unlend_user_prio(td, pri);
mtx_unlock_spin(&sched_lock);
}
umtx_key_release(&key);
return (error);
}
static int
do_set_ceiling(struct thread *td, struct umutex *m, uint32_t ceiling,
uint32_t *old_ceiling)
{
struct umtx_q *uq;
uint32_t save_ceiling;
uint32_t owner, id;
uint32_t flags;
int error;
flags = fuword32(&m->m_flags);
if ((flags & UMUTEX_PRIO_PROTECT) == 0)
return (EINVAL);
if (ceiling > RTP_PRIO_MAX)
return (EINVAL);
id = td->td_tid;
uq = td->td_umtxq;
if ((error = umtx_key_get(m, TYPE_PP_UMUTEX, GET_SHARE(flags),
&uq->uq_key)) != 0)
return (error);
for (;;) {
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
save_ceiling = fuword32(&m->m_ceilings[0]);
owner = casuword32(&m->m_owner,
UMUTEX_CONTESTED, id | UMUTEX_CONTESTED);
if (owner == UMUTEX_CONTESTED) {
suword32(&m->m_ceilings[0], ceiling);
o Add keyword volatile for user mutex owner field. o Fix type consistent problem by using type long for old umtx and wait channel. o Rename casuptr to casuword.
2006-10-17 02:24:47 +00:00
suword32(__DEVOLATILE(uint32_t *, &m->m_owner),
UMUTEX_CONTESTED);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
error = 0;
break;
}
/* The address was invalid. */
if (owner == -1) {
error = EFAULT;
break;
}
if ((owner & ~UMUTEX_CONTESTED) == id) {
suword32(&m->m_ceilings[0], ceiling);
error = 0;
break;
}
/*
* If we caught a signal, we have retried and now
* exit immediately.
*/
if (error != 0)
break;
/*
* We set the contested bit, sleep. Otherwise the lock changed
* and we need to retry or we lost a race to the thread
* unlocking the umtx.
*/
umtxq_lock(&uq->uq_key);
umtxq_insert(uq);
umtxq_unbusy(&uq->uq_key);
error = umtxq_sleep(uq, "umtxpp", 0);
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
}
umtxq_lock(&uq->uq_key);
if (error == 0)
umtxq_signal(&uq->uq_key, INT_MAX);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
if (error == 0 && old_ceiling != NULL)
suword32(old_ceiling, save_ceiling);
return (error);
}
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
static int
_do_lock_umutex(struct thread *td, struct umutex *m, int flags, int timo,
int try)
{
switch(flags & (UMUTEX_PRIO_INHERIT | UMUTEX_PRIO_PROTECT)) {
case 0:
return (_do_lock_normal(td, m, flags, timo, try));
case UMUTEX_PRIO_INHERIT:
return (_do_lock_pi(td, m, flags, timo, try));
case UMUTEX_PRIO_PROTECT:
return (_do_lock_pp(td, m, flags, timo, try));
}
return (EINVAL);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
/*
* Lock a userland POSIX mutex.
*/
static int
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
do_lock_umutex(struct thread *td, struct umutex *m,
struct timespec *timeout, int try)
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
{
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
struct timespec ts, ts2, ts3;
struct timeval tv;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
uint32_t flags;
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
int error;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
flags = fuword32(&m->m_flags);
if (flags == -1)
return (EFAULT);
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
if (timeout == NULL) {
error = _do_lock_umutex(td, m, flags, 0, try);
/* Mutex locking is restarted if it is interrupted. */
if (error == EINTR)
error = ERESTART;
} else {
getnanouptime(&ts);
timespecadd(&ts, timeout);
TIMESPEC_TO_TIMEVAL(&tv, timeout);
for (;;) {
error = _do_lock_umutex(td, m, flags, tvtohz(&tv), try);
if (error != ETIMEDOUT)
break;
getnanouptime(&ts2);
if (timespeccmp(&ts2, &ts, >=)) {
error = ETIMEDOUT;
break;
}
ts3 = ts;
timespecsub(&ts3, &ts2);
TIMESPEC_TO_TIMEVAL(&tv, &ts3);
}
/* Timed-locking is not restarted. */
if (error == ERESTART)
error = EINTR;
Reorder some statments. Fix typo and remove stale comments.
2006-08-30 23:59:45 +00:00
}
Merge all code of do_lock_normal, do_lock_pi and do_lock_pp into function do_lock_umutex.
2006-09-05 12:01:09 +00:00
return (error);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
/*
* Unlock a userland POSIX mutex.
*/
static int
do_unlock_umutex(struct thread *td, struct umutex *m)
{
uint32_t flags;
flags = fuword32(&m->m_flags);
if (flags == -1)
return (EFAULT);
Make sure we get new m_owner value if we can not unlock it in uncontested case. Reorder statements in do_unlock_umutex.
2006-09-02 02:41:33 +00:00
switch(flags & (UMUTEX_PRIO_INHERIT | UMUTEX_PRIO_PROTECT)) {
case 0:
return (do_unlock_normal(td, m, flags));
case UMUTEX_PRIO_INHERIT:
return (do_unlock_pi(td, m, flags));
case UMUTEX_PRIO_PROTECT:
return (do_unlock_pp(td, m, flags));
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
Make sure we get new m_owner value if we can not unlock it in uncontested case. Reorder statements in do_unlock_umutex.
2006-09-02 02:41:33 +00:00
return (EINVAL);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
static int
do_cv_wait(struct thread *td, struct ucond *cv, struct umutex *m,
if a thread blocked on userland condition variable is pthread_cancel()ed, it is expected that the thread will not consume a pthread_cond_signal(), therefor, we use thr_wake() to mark a flag, the flag tells a thread calling do_cv_wait() in umtx code to not block on a condition variable. Thread library is expected that once a thread detected itself is in pthread_cond_wait, it will call the thr_wake() for itself in its SIGCANCEL handler.
2006-12-04 14:15:12 +00:00
struct timespec *timeout, u_long wflags)
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
{
struct umtx_q *uq;
struct timeval tv;
struct timespec cts, ets, tts;
uint32_t flags;
int error;
uq = td->td_umtxq;
flags = fuword32(&cv->c_flags);
error = umtx_key_get(cv, TYPE_CV, GET_SHARE(flags), &uq->uq_key);
if (error != 0)
return (error);
umtxq_lock(&uq->uq_key);
umtxq_busy(&uq->uq_key);
umtxq_insert(uq);
umtxq_unlock(&uq->uq_key);
/*
* The magic thing is we should set c_has_waiters to 1 before
* releasing user mutex.
*/
suword32(__DEVOLATILE(uint32_t *, &cv->c_has_waiters), 1);
umtxq_lock(&uq->uq_key);
umtxq_unbusy(&uq->uq_key);
umtxq_unlock(&uq->uq_key);
error = do_unlock_umutex(td, m);
umtxq_lock(&uq->uq_key);
if (error == 0) {
if a thread blocked on userland condition variable is pthread_cancel()ed, it is expected that the thread will not consume a pthread_cond_signal(), therefor, we use thr_wake() to mark a flag, the flag tells a thread calling do_cv_wait() in umtx code to not block on a condition variable. Thread library is expected that once a thread detected itself is in pthread_cond_wait, it will call the thr_wake() for itself in its SIGCANCEL handler.
2006-12-04 14:15:12 +00:00
if ((wflags & UMTX_CHECK_UNPARKING) &&
(td->td_pflags & TDP_WAKEUP)) {
td->td_pflags &= ~TDP_WAKEUP;
error = EINTR;
} else if (timeout == NULL) {
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
error = umtxq_sleep(uq, "ucond", 0);
} else {
getnanouptime(&ets);
timespecadd(&ets, timeout);
TIMESPEC_TO_TIMEVAL(&tv, timeout);
for (;;) {
error = umtxq_sleep(uq, "ucond", tvtohz(&tv));
if (error != ETIMEDOUT)
break;
getnanouptime(&cts);
if (timespeccmp(&cts, &ets, >=)) {
error = ETIMEDOUT;
break;
}
tts = ets;
timespecsub(&tts, &cts);
TIMESPEC_TO_TIMEVAL(&tv, &tts);
}
}
}
if (error != 0) {
if ((uq->uq_flags & UQF_UMTXQ) == 0) {
/*
* If we concurrently got do_cv_signal()d
* and we got an error or UNIX signals or a timeout,
* then, perform another umtxq_signal to avoid
* consuming the wakeup. This may cause supurious
* wakeup for another thread which was just queued,
* but SUSV3 explicitly allows supurious wakeup to
* occur, and indeed a kernel based implementation
* can not avoid it.
*/
if a thread blocked on userland condition variable is pthread_cancel()ed, it is expected that the thread will not consume a pthread_cond_signal(), therefor, we use thr_wake() to mark a flag, the flag tells a thread calling do_cv_wait() in umtx code to not block on a condition variable. Thread library is expected that once a thread detected itself is in pthread_cond_wait, it will call the thr_wake() for itself in its SIGCANCEL handler.
2006-12-04 14:15:12 +00:00
if (!umtxq_signal(&uq->uq_key, 1))
error = 0;
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
}
if (error == ERESTART)
error = EINTR;
}
umtxq_remove(uq);
umtxq_unlock(&uq->uq_key);
umtx_key_release(&uq->uq_key);
return (error);
}
/*
* Signal a userland condition variable.
*/
static int
do_cv_signal(struct thread *td, struct ucond *cv)
{
struct umtx_key key;
int error, cnt, nwake;
uint32_t flags;
flags = fuword32(&cv->c_flags);
if ((error = umtx_key_get(cv, TYPE_CV, GET_SHARE(flags), &key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
cnt = umtxq_count(&key);
nwake = umtxq_signal(&key, 1);
if (cnt <= nwake) {
umtxq_unlock(&key);
error = suword32(
__DEVOLATILE(uint32_t *, &cv->c_has_waiters), 0);
umtxq_lock(&key);
}
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
return (error);
}
static int
do_cv_broadcast(struct thread *td, struct ucond *cv)
{
struct umtx_key key;
int error;
uint32_t flags;
flags = fuword32(&cv->c_flags);
if ((error = umtx_key_get(cv, TYPE_CV, GET_SHARE(flags), &key)) != 0)
return (error);
umtxq_lock(&key);
umtxq_busy(&key);
umtxq_signal(&key, INT_MAX);
umtxq_unlock(&key);
error = suword32(__DEVOLATILE(uint32_t *, &cv->c_has_waiters), 0);
umtxq_lock(&key);
umtxq_unbusy(&key);
umtxq_unlock(&key);
umtx_key_release(&key);
return (error);
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int
_umtx_lock(struct thread *td, struct _umtx_lock_args *uap)
/* struct umtx *umtx */
{
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
return _do_lock_umtx(td, uap->umtx, td->td_tid, 0);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
int
_umtx_unlock(struct thread *td, struct _umtx_unlock_args *uap)
/* struct umtx *umtx */
{
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
return do_unlock_umtx(td, uap->umtx, td->td_tid);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
static int
__umtx_op_lock_umtx(struct thread *td, struct _umtx_op_args *uap)
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
{
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
struct timespec *ts, timeout;
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
int error;
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin(uap->uaddr2, &timeout, sizeof(timeout));
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0) {
return (EINVAL);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
ts = &timeout;
}
return (do_lock_umtx(td, uap->obj, uap->val, ts));
}
static int
__umtx_op_unlock_umtx(struct thread *td, struct _umtx_op_args *uap)
{
return (do_unlock_umtx(td, uap->obj, uap->val));
}
static int
__umtx_op_wait(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin(uap->uaddr2, &timeout, sizeof(timeout));
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0)
return (EINVAL);
ts = &timeout;
}
return do_wait(td, uap->obj, uap->val, ts, 0);
}
static int
__umtx_op_wake(struct thread *td, struct _umtx_op_args *uap)
{
return (kern_umtx_wake(td, uap->obj, uap->val));
}
static int
__umtx_op_lock_umutex(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin(uap->uaddr2, &timeout,
sizeof(timeout));
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0) {
return (EINVAL);
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
ts = &timeout;
}
return do_lock_umutex(td, uap->obj, ts, 0);
}
static int
__umtx_op_trylock_umutex(struct thread *td, struct _umtx_op_args *uap)
{
return do_lock_umutex(td, uap->obj, NULL, 1);
}
static int
__umtx_op_unlock_umutex(struct thread *td, struct _umtx_op_args *uap)
{
return do_unlock_umutex(td, uap->obj);
}
static int
__umtx_op_set_ceiling(struct thread *td, struct _umtx_op_args *uap)
{
return do_set_ceiling(td, uap->obj, uap->val, uap->uaddr1);
}
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
static int
__umtx_op_cv_wait(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin(uap->uaddr2, &timeout,
sizeof(timeout));
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0) {
return (EINVAL);
}
ts = &timeout;
}
if a thread blocked on userland condition variable is pthread_cancel()ed, it is expected that the thread will not consume a pthread_cond_signal(), therefor, we use thr_wake() to mark a flag, the flag tells a thread calling do_cv_wait() in umtx code to not block on a condition variable. Thread library is expected that once a thread detected itself is in pthread_cond_wait, it will call the thr_wake() for itself in its SIGCANCEL handler.
2006-12-04 14:15:12 +00:00
return (do_cv_wait(td, uap->obj, uap->uaddr1, ts, uap->val));
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
}
static int
__umtx_op_cv_signal(struct thread *td, struct _umtx_op_args *uap)
{
return do_cv_signal(td, uap->obj);
}
static int
__umtx_op_cv_broadcast(struct thread *td, struct _umtx_op_args *uap)
{
return do_cv_broadcast(td, uap->obj);
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
typedef int (*_umtx_op_func)(struct thread *td, struct _umtx_op_args *uap);
static _umtx_op_func op_table[] = {
__umtx_op_lock_umtx, /* UMTX_OP_LOCK */
__umtx_op_unlock_umtx, /* UMTX_OP_UNLOCK */
__umtx_op_wait, /* UMTX_OP_WAIT */
__umtx_op_wake, /* UMTX_OP_WAKE */
__umtx_op_trylock_umutex, /* UMTX_OP_MUTEX_TRYLOCK */
__umtx_op_lock_umutex, /* UMTX_OP_MUTEX_LOCK */
__umtx_op_unlock_umutex, /* UMTX_OP_MUTEX_UNLOCK */
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
__umtx_op_set_ceiling, /* UMTX_OP_SET_CEILING */
__umtx_op_cv_wait, /* UMTX_OP_CV_WAIT*/
__umtx_op_cv_signal, /* UMTX_OP_CV_SIGNAL */
__umtx_op_cv_broadcast /* UMTX_OP_CV_BROADCAST */
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
};
int
_umtx_op(struct thread *td, struct _umtx_op_args *uap)
{
In order to eliminate a branch, convert opcode to unsigned integer.
2006-10-25 06:38:46 +00:00
if ((unsigned)uap->op < UMTX_OP_MAX)
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
return (*op_table[uap->op])(td, uap);
return (EINVAL);
}
#ifdef COMPAT_IA32
Implement 32bit umtx_lock and umtx_unlock system calls, these two system calls are not used by libthr in RELENG_6 and HEAD, it is only used by the libthr in RELENG-5, the _umtx_op system call can do more incremental dirty works than these two system calls without having to introduce new system calls or throw away old system calls when things are going on.
2006-10-06 08:22:08 +00:00
int
freebsd32_umtx_lock(struct thread *td, struct freebsd32_umtx_lock_args *uap)
/* struct umtx *umtx */
{
return (do_lock_umtx32(td, (uint32_t *)uap->umtx, td->td_tid, NULL));
}
int
freebsd32_umtx_unlock(struct thread *td, struct freebsd32_umtx_unlock_args *uap)
/* struct umtx *umtx */
{
return (do_unlock_umtx32(td, (uint32_t *)uap->umtx, td->td_tid));
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
struct timespec32 {
u_int32_t tv_sec;
u_int32_t tv_nsec;
};
static inline int
copyin_timeout32(void *addr, struct timespec *tsp)
{
struct timespec32 ts32;
int error;
error = copyin(addr, &ts32, sizeof(struct timespec32));
if (error == 0) {
tsp->tv_sec = ts32.tv_sec;
tsp->tv_nsec = ts32.tv_nsec;
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
Revert my previous errno hack, that is certainly an issue, and always has been, but the system call itself returns errno in a register so the problem is really a function of libc, not the system call. Discussed with : Matthew Dillion <dillon@apollo.backplane.com>
2005-01-18 13:53:10 +00:00
return (error);
1. make umtx sharable between processes, the way is two or more processes call mmap() to create a shared space, and then initialize umtx on it, after that, each thread in different processes can use the umtx same as threads in same process. 2. introduce a new syscall _umtx_op to support timed lock and condition variable semantics. also, orignal umtx_lock and umtx_unlock inline functions now are reimplemented by using _umtx_op, the _umtx_op can use arbitrary id not just a thread id.
2004-12-18 12:52:44 +00:00
}
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
static int
__umtx_op_lock_umtx_compat32(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin_timeout32(uap->uaddr2, &timeout);
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0) {
return (EINVAL);
}
ts = &timeout;
}
return (do_lock_umtx32(td, uap->obj, uap->val, ts));
}
static int
__umtx_op_unlock_umtx_compat32(struct thread *td, struct _umtx_op_args *uap)
{
return (do_unlock_umtx32(td, uap->obj, (uint32_t)uap->val));
}
static int
__umtx_op_wait_compat32(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin_timeout32(uap->uaddr2, &timeout);
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0)
return (EINVAL);
ts = &timeout;
}
return do_wait(td, uap->obj, uap->val, ts, 1);
}
static int
__umtx_op_lock_umutex_compat32(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin_timeout32(uap->uaddr2, &timeout);
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0)
return (EINVAL);
ts = &timeout;
}
return do_lock_umutex(td, uap->obj, ts, 0);
}
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
static int
__umtx_op_cv_wait_compat32(struct thread *td, struct _umtx_op_args *uap)
{
struct timespec *ts, timeout;
int error;
/* Allow a null timespec (wait forever). */
if (uap->uaddr2 == NULL)
ts = NULL;
else {
error = copyin_timeout32(uap->uaddr2, &timeout);
if (error != 0)
return (error);
if (timeout.tv_nsec >= 1000000000 ||
timeout.tv_nsec < 0)
return (EINVAL);
ts = &timeout;
}
if a thread blocked on userland condition variable is pthread_cancel()ed, it is expected that the thread will not consume a pthread_cond_signal(), therefor, we use thr_wake() to mark a flag, the flag tells a thread calling do_cv_wait() in umtx code to not block on a condition variable. Thread library is expected that once a thread detected itself is in pthread_cond_wait, it will call the thr_wake() for itself in its SIGCANCEL handler.
2006-12-04 14:15:12 +00:00
return (do_cv_wait(td, uap->obj, uap->uaddr1, ts, uap->val));
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
}
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
static _umtx_op_func op_table_compat32[] = {
__umtx_op_lock_umtx_compat32, /* UMTX_OP_LOCK */
__umtx_op_unlock_umtx_compat32, /* UMTX_OP_UNLOCK */
__umtx_op_wait_compat32, /* UMTX_OP_WAIT */
__umtx_op_wake, /* UMTX_OP_WAKE */
__umtx_op_trylock_umutex, /* UMTX_OP_MUTEX_LOCK */
Fix umtx command order error for freebsd 32bit.
2006-09-22 14:59:10 +00:00
__umtx_op_lock_umutex_compat32, /* UMTX_OP_MUTEX_TRYLOCK */
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
__umtx_op_unlock_umutex, /* UMTX_OP_MUTEX_UNLOCK */
Introduce userspace condition variable, since we have already POSIX priority mutex implemented, it is the time to introduce this stuff, now we can use umutex and ucond together to implement pthread's condition wait/signal.
2006-12-03 01:49:22 +00:00
__umtx_op_set_ceiling, /* UMTX_OP_SET_CEILING */
__umtx_op_cv_wait_compat32, /* UMTX_OP_CV_WAIT*/
__umtx_op_cv_signal, /* UMTX_OP_CV_SIGNAL */
__umtx_op_cv_broadcast /* UMTX_OP_CV_BROADCAST */
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
};
int
freebsd32_umtx_op(struct thread *td, struct freebsd32_umtx_op_args *uap)
{
In order to eliminate a branch, convert opcode to unsigned integer.
2006-10-25 06:38:46 +00:00
if ((unsigned)uap->op < UMTX_OP_MAX)
Add umtx support for 32bit process on AMD64 machine.
2006-09-22 00:52:54 +00:00
return (*op_table_compat32[uap->op])(td,
(struct _umtx_op_args *)uap);
return (EINVAL);
}
#endif
This is initial version of POSIX priority mutex support, a new userland mutex structure is added as following: struct umutex { __lwpid_t m_owner; uint32_t m_flags; uint32_t m_ceilings[2]; uint32_t m_spare[4]; }; The m_owner represents owner thread, it is a thread id, in non-contested case, userland can simply use atomic_cmpset_int to lock the mutex, if the mutex is contested, high order bit will be set, and userland should do locking and unlocking via kernel syscall. Flag UMUTEX_PRIO_INHERIT represents pthread's PTHREAD_PRIO_INHERIT mutex, which when contention happens, kernel should do priority propagating. Flag UMUTEX_PRIO_PROTECT indicates it is pthread's PTHREAD_PRIO_PROTECT mutex, userland should initialize m_owner to contested state UMUTEX_CONTESTED, then atomic_cmpset_int will be failure and kernel syscall should be invoked to do locking, this becauses for such a mutex, kernel should always boost the thread's priority before it can lock the mutex, m_ceilings is used by PTHREAD_PRIO_PROTECT mutex, the first element is used to boost thread's priority when it locked the mutex, second element is used when the mutex is unlocked, the PTHREAD_PRIO_PROTECT mutex's link list is kept in userland, the m_ceiling[1] is managed by thread library so kernel needn't allocate memory to keep the link list, when such a mutex is unlocked, kernel reset m_owner to UMUTEX_CONTESTED. Flag USYNC_PROCESS_SHARED indicate if the synchronization object is process shared, if the flag is not set, it saves a vm_map_lookup() call. The umtx chain is still used as a sleep queue, when a thread is blocked on PTHREAD_PRIO_INHERIT mutex, a umtx_pi is allocated to support priority propagating, it is dynamically allocated and reference count is used, it is not optimized but works well in my tests, while the umtx chain has its own locking protocol, the priority propagating protocol are all protected by sched_lock because priority propagating function is called with sched_lock held from scheduler. No visible performance degradation is found which these changes. Some parameter names in _umtx_op syscall are renamed.
2006-08-28 04:24:51 +00:00
void
umtx_thread_init(struct thread *td)
{
td->td_umtxq = umtxq_alloc();
td->td_umtxq->uq_thread = td;
}
void
umtx_thread_fini(struct thread *td)
{
umtxq_free(td->td_umtxq);
}
/*
* It will be called when new thread is created, e.g fork().
*/
void
umtx_thread_alloc(struct thread *td)
{
struct umtx_q *uq;
uq = td->td_umtxq;
uq->uq_inherited_pri = PRI_MAX;
KASSERT(uq->uq_flags == 0, ("uq_flags != 0"));
KASSERT(uq->uq_thread == td, ("uq_thread != td"));
KASSERT(uq->uq_pi_blocked == NULL, ("uq_pi_blocked != NULL"));
KASSERT(TAILQ_EMPTY(&uq->uq_pi_contested), ("uq_pi_contested is not empty"));
}
/*
* exec() hook.
*/
static void
umtx_exec_hook(void *arg __unused, struct proc *p __unused,
struct image_params *imgp __unused)
{
umtx_thread_cleanup(curthread);
}
/*
* thread_exit() hook.
*/
void
umtx_thread_exit(struct thread *td)
{
umtx_thread_cleanup(td);
}
/*
* clean up umtx data.
*/
static void
umtx_thread_cleanup(struct thread *td)
{
struct umtx_q *uq;
struct umtx_pi *pi;
if ((uq = td->td_umtxq) == NULL)
return;
mtx_lock_spin(&sched_lock);
uq->uq_inherited_pri = PRI_MAX;
while ((pi = TAILQ_FIRST(&uq->uq_pi_contested)) != NULL) {
pi->pi_owner = NULL;
TAILQ_REMOVE(&uq->uq_pi_contested, pi, pi_link);
}
td->td_flags &= ~TDF_UBORROWING;
mtx_unlock_spin(&sched_lock);
}
Reference in New Issue Copy Permalink