schedulers a bit to ensure more correct handling of priorities and fewer
priority inversions:
- Add two functions to the sched(9) API to handle priority lending:
sched_lend_prio() and sched_unlend_prio(). The turnstile code uses these
functions to ask the scheduler to lend a thread a set priority and to
tell the scheduler when it thinks it is ok for a thread to stop borrowing
priority. The unlend case is slightly complex in that the turnstile code
tells the scheduler what the minimum priority of the thread needs to be
to satisfy the requirements of any other threads blocked on locks owned
by the thread in question. The scheduler then decides where the thread
can go back to normal mode (if it's normal priority is high enough to
satisfy the pending lock requests) or it it should continue to use the
priority specified to the sched_unlend_prio() call. This involves adding
a new per-thread flag TDF_BORROWING that replaces the ULE-only kse flag
for priority elevation.
- Schedulers now refuse to lower the priority of a thread that is currently
borrowing another therad's priority.
- If a scheduler changes the priority of a thread that is currently sitting
on a turnstile, it will call a new function turnstile_adjust() to inform
the turnstile code of the change. This function resorts the thread on
the priority list of the turnstile if needed, and if the thread ends up
at the head of the list (due to having the highest priority) and its
priority was raised, then it will propagate that new priority to the
owner of the lock it is blocked on.
Some additional fixes specific to the 4BSD scheduler include:
- Common code for updating the priority of a thread when the user priority
of its associated kse group has been consolidated in a new static
function resetpriority_thread(). One change to this function is that
it will now only adjust the priority of a thread if it already has a
time sharing priority, thus preserving any boosts from a tsleep() until
the thread returns to userland. Also, resetpriority() no longer calls
maybe_resched() on each thread in the group. Instead, the code calling
resetpriority() is responsible for calling resetpriority_thread() on
any threads that need to be updated.
- schedcpu() now uses resetpriority_thread() instead of just calling
sched_prio() directly after it updates a kse group's user priority.
- sched_clock() now uses resetpriority_thread() rather than writing
directly to td_priority.
- sched_nice() now updates all the priorities of the threads after the
group priority has been adjusted.
Discussed with: bde
Reviewed by: ups, jeffr
Tested on: 4bsd, ule
Tested on: i386, alpha, sparc64
is the case for most other sysctls in the System V IPC message queue
implementation.
PR: 75541
Submitted by: Sergiy Vyshnevetskiy <serg at vostok dot net>
MFC after: 2 weeks
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.
of lock types in the kernel. This results in an increase of witness
data usage from ~145k to ~280k on i386 for kernels with
'options WITNESS'.
- Remove the unused witness malloc bucket.
Submitted by: Michal Mertl mime at traveller dot cz (1)
- Remove the sched_add wrapper that used sched_add_internal() as a backend.
Its only purpose was to interpret one flag and turn it into an int. Do
the right thing and interpret the flag in sched_add() instead.
- Pass the flag argument to sched_add() to kseq_runq_add() so that we can
get the SRQ_PREEMPT optimization too.
- Add a KEF_INTERNAL flag. If KEF_INTERNAL is set we don't adjust the SLOT
counts, otherwise the slot counts are adjusted as soon as we enter
sched_add() or sched_rem() rather than when the thread is actually placed
on the run queue. This greatly simplifies the handling of slots.
- Remove the explicit prevention of migration for ithreads on non-x86
platforms. This was never shown to have any real benefit.
- Remove the unused class argument to KSE_CAN_MIGRATE().
- Add ktr points for thread migration events.
- Fix a long standing bug on platforms which don't initialize the cpu
topology. The ksg_maxid variable was never correctly set on these
platforms which caused the long term load balancer to never inspect
more than the first group or processor.
- Fix another bug which prevented the long term load balancer from working
properly. If stathz != hz we can't expect sched_clock() to be called
on the exact tick count that we're anticipating.
- Rearrange sched_switch() a bit to reduce indentation levels.
and threads currently holding sleep mutexes (and spin mutexes for
curthread). This can be quite useful in looking for a lock condition
summary for a system, as it avoids manually iterating through threads
and processes to find all the interesting locks.
NB: "alllocks" is up there with "lockedvnods" for a bad argument for
show.
MFC after: 2 weeks
lower the priority of the returning thread to a user priority before
calling into thread_userret() which would call wakeup() which in turn would
cause the returning thread to eventually context switch rather than
completing its slice. Allowing this thread to complete its slice first
yields a 15% performance improvement in super-smack on my dual opteron with
4BSD.
substitute for a global mutex protecting the socket count and
generation number.
The observation that soreceive_rcvoob() can't return an mbuf
chain is a property, not a bug, so remove the XXXRW.
In sorflush, s/existing/previous/ for code when describing prior
behavior.
For SO_LINGER socket option retrieval, remove an XXXRW about why
we hold the mutex: this is correct and not dubious.
MFC after: 2 weeks
unnecessary use of a global variable and simplify the return case.
While here, use ()'s around return values.
In sodealloc(), remove a comment about why we bump the gencnt and
decrement the socket count separately. It doesn't add
substantially to the reading, and clutters the function.
MFC after: 2 weeks
occur between a reader and a writer that results in a panic upon close,
e.g.,
"panic: sbflush_locked: cc 4 || mb 0xffffff0052afa400 || mbcnt 0"
Reviewed by: rwatson@
MFC after: 2 weeks
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.
nice of 0. Doing so can cause an infinite loop because they should be
running, but a nice -20 process could prevent them from doing so.
- Add a new flag KEF_PRIOELEV to flag a thread that has had its priority
elevated due to priority propagation. If a thread has had its priority
elevated, we assume that it must go on the current queue and it must
get a slice.
- In sched_userret() if our priority was elevated and we shouldn't have
a timeslice, yield here until we should.
Found/Tested by: glebius
which holds on to just the data structure and the mutex. (The
existing refcount (fd_refcnt) holds onto the open files in the
descriptor.)
The fd_holdcnt is protected by fdesc_mtx, fd_refcnt by FILEDESC_LOCK.
Add fdhold(struct proc *) which gets a hold on the filedescriptors of
the specified proc..
Add fddrop(struct filedesc *) which drops the fd_holdcnt and if zero
destroys the mutex and frees the memory.
Initialize the fd_holdcnt to one in fdinit(). Normal operations on
the filedesc structure will not change it.
In fdfree() use fddrop() to dispose of the mutex and structure. Hold
the FILEDESC_LOCK() until we have cleaned out the contents and carefully
set the fields to null values during cleanup.
Use fdhold()/fddrop() in mountcheckdirs() and sysctl_kern_file().
for ensuring that a process' filedesc is not shared with anybody.
Use it in the two places which previously had private implmentations.
This collects all fd_refcnt handling in kern_descrip.c
nice value above 0, set it to 0 so that it may proceed with haste.
This is especially important on ULE, where adjusting the priority
does not guarantee that a thread will be granted a greater time slice.
specifically, vm_pgmoveco():
1. If vm_pgmoveco() sleeps on a busy page, it must redo the look up
because the page may have been freed.
2. If the receive buffer is copy-on-write due to, for example, a fork,
then although the first vm object in the shadow chain may not contain
a page there may still be one from a backing object that is mapped.
Thus, a pmap_remove() is required for the new page rather than the
backing object's page to been seen by the application.
Also, add some comments to vm_pgmoveco() and update some assertions.
Tested by: ken@
