into pthread structure to keep track of locked PTHREAD_PRIO_PROTECT mutex,
no real mutex code is changed, the mutex locking and unlocking code should
has same performance as before.
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.
inlined and a procedure call is made in the rare case, i.e., when it is
necessary to sleep. In this case, inlining the test actually makes the
kernel smaller.
other stuff) in the osrelease=2.6.16 case:
- implement CLONE_PARENT semantic
- fix TLS loading in clone CLONE_SETTLS
- lock proc in the currently disabled part of CLONE_THREAD
I suggest to not unload the linux module after testing this, there are
some "<defunct>" processes hanging around after exiting (they aren't
with osrelease=2.4.2) and they may panic your kernel when unloading the
linux module. They are in state Z and some of them consume CPU according
to ps. But I don't trust the CPU part, the idle threads gets too much CPU
that this may be possible (accumulating idle, X and 2 defunct processes
results in 104.7%, this looks to much to be a rounding error).
Noticed by: Intron <mag@intron.ac>
Submitted by: rdivacky (in collaboration with Intron)
Tested by: Intron, netchild
Reviewed by: jhb (previous version)
but further on -current (still not successful, but a step into the right
direction).
Sponsored by: Google SoC 2006
Submitted by: rdivacky
Tested by: Paul Mather <paul@gromit.dlib.vt.edu>
is loaded. This problem stems from the fact that the policy is not properly
initializing the mac label associated with the NFS daemon.
Obtained from: TrustedBSD Project
Discussed with: rwatson
audit record size at run-time, which can be used by the user
process to size the user space buffer it reads into from the audit
pipe.
Perforce change: 105098
Obtained from: TrustedBSD Project
the 'vfs_getopt(optlist, "errmsg", (void **)&errmsg, &errmsg_len)'
call fails, 'errmsg' is left uninitialized, making the later tests
against NULL meaningless, and the uses bogus. Thus initialize
'errmsg' to NULL beforehand. [1]
While at it, remove the superfluous assignment of 0 to 'errmsg_len'
if the above mentioned call fails as it's already initialized to 0.
Submitted by: Michael Plass [1]
notes since the last import:
OpenBSM 1.0 alpha 9
- Rename many OpenBSM-specific constants and API elements containing the
strings "BSM" and "bsm" to "AUDIT" and "audit", observing that this is true
for almost all existing constants and APIs.
- Instead of passing a per-instance cookie directly into all audit filter
APIs, pass in the audit filter daemon state pointer, which is then used by
the module using an audit_filter_{get,set}cookie() API. This will allow
future service APIs provided by the filter daemon to maintain their own
state -- for example, per-module preselection state.
OpenBSM 1.0 alpha 8
- Correct typo in definition of AUR_INT.
- Adopt OpenSolaris constant values for AUDIT_* configuration flags.
- Arguments to au_to_exec_args() and au_to_exec_env() no longer const.
- Add kernel versions of au_to_exec_args() and au_to_exec_env().
- Fix exec argument type that is printed for env strings from 'arg' to 'env'.
- New OpenBSM token version number assigned, constants added for other
commonly seen version numbers.
- OpenBSM-specific events assigned numbers in the 43xxx range to avoid future
collisions with Solaris. Darwin events renamed to AUE_DARWIN_foo, as they
are now deprecated numberings.
- autoconf now detects clock_gettime(), which is not available on Darwin.
- praudit output fixes relating to arg32 and arg64 tokens.
- Maximum record size updated to 64k-1 to match Solaris record size limit.
- Various style and comment cleanups in include files.
This is an MFC candidate to RELENG_6.
Obtained from: TrustedBSD Project
override NM in Makefiles when, for example, cross compiling and have
that value be used by lorder. NM normally isn't defined, so we pass a
null value to lorder. lorder says 'NM=${NM-nm}' which causes it to
pickup the default value.
