musl/src/thread, branch v0.9.13 musl - an implementation of the standard library for Linux-based systems in pthread_getattr_np, use mremap rather than madvise to measure stack 2013-07-31T19:19:39+00:00 Rich Felker dalias@aerifal.cx 2013-07-31T19:19:39+00:00 5db951ef80cae8b627f95b995811bf916c069757 the original motivation for this patch was that qemu (and possibly other syscall emulators) nop out madvise, resulting in an infinite loop. however, there is another benefit to this change: madvise may actually undo an explicit madvise the application intended for its stack, whereas the mremap operation is a true nop. the logic here is that mremap must fail if it cannot resize the mapping in-place, and the caller knows that it cannot resize in-place because it knows the next page of virtual memory is already occupied. 
the original motivation for this patch was that qemu (and possibly
other syscall emulators) nop out madvise, resulting in an infinite
loop. however, there is another benefit to this change: madvise may
actually undo an explicit madvise the application intended for its
stack, whereas the mremap operation is a true nop. the logic here is
that mremap must fail if it cannot resize the mapping in-place, and
the caller knows that it cannot resize in-place because it knows the
next page of virtual memory is already occupied.
make pthread attribute types structs, even when they just have one field 2013-07-22T17:57:02+00:00 Rich Felker dalias@aerifal.cx 2013-07-22T17:57:02+00:00 0109d950e685f1c5f6cef245a8b1bc47c9b6d94c this change is to get the right tags for C++ ABI matching. it should have no other effects. 
this change is to get the right tags for C++ ABI matching. it should
have no other effects.
fix syscall argument bug in pthread_getschedparam 2013-06-27T02:02:23+00:00 Rich Felker dalias@aerifal.cx 2013-06-27T02:02:23+00:00 b17c75a4d539d7ec5b81cc7ce7ce6b065a87e7a6 the address of the pointer to the sched param, rather than the pointer, was being passed to the kernel. 
the address of the pointer to the sched param, rather than the
pointer, was being passed to the kernel.
fix temp file leak in sem_open on successful creation of new semaphore 2013-06-27T01:41:51+00:00 Rich Felker dalias@aerifal.cx 2013-06-27T01:41:51+00:00 7c20a11801fd56cbadac5a6e88ddddf8656ac1bc 






fix bug whereby sem_open leaked its own internal slots on failure
2013-06-27T01:39:15+00:00

Rich Felker
dalias@aerifal.cx

2013-06-27T01:39:15+00:00

a033cd22aa0ecd9f494b74669d358e7e1c7e1335












in sem_open, don't leak vm mapping if fstat fails
2013-06-27T01:35:56+00:00

Rich Felker
dalias@aerifal.cx

2013-06-27T01:35:56+00:00

52d4444f8eec1a4e7e0861859c705c3a558b4e2a

fstat should not fail under normal circumstances, so this fix is
mostly theoretical.





fstat should not fail under normal circumstances, so this fix is
mostly theoretical.







fix failure of pthread_setschedparam to pass correct param to kernel
2013-06-27T01:34:44+00:00

Rich Felker
dalias@aerifal.cx

2013-06-27T01:34:44+00:00

21088aee2eb6bf12fd3b1db918ee4754989ff7da

the address of the pointer, rather than the pointer, was being passed.
this was probably a copy-and-paste error from corresponding get code.





the address of the pointer, rather than the pointer, was being passed.
this was probably a copy-and-paste error from corresponding get code.







support cputime clocks for processes/threads other than self
2013-06-08T15:36:41+00:00

Rich Felker
dalias@aerifal.cx

2013-06-08T15:36:41+00:00

ea200e38bd79723054c7ad97cbf87911eac5b7f5

apparently these features have been in Linux for a while now, so it
makes sense to support them. the bit twiddling seems utterly illogical
and wasteful, especially the negation, but that's how the kernel folks
chose to encode pids/tids into the clock id.





apparently these features have been in Linux for a while now, so it
makes sense to support them. the bit twiddling seems utterly illogical
and wasteful, especially the negation, but that's how the kernel folks
chose to encode pids/tids into the clock id.







ensure that thread dtv pointer is never null to optimize __tls_get_addr
2013-06-03T20:35:59+00:00

Rich Felker
dalias@aerifal.cx

2013-06-03T20:35:59+00:00

44b4d09fc0626541ca12cf96f65adc21ab1fc413












transition to using functions for internal signal blocking/restoring
2013-04-26T23:48:01+00:00

Rich Felker
dalias@aerifal.cx

2013-04-26T23:48:01+00:00

2c074b0d6cb2b28c5d1c0ccb2809965f4676efeb

there are several reasons for this change. one is getting rid of the
repetition of the syscall signature all over the place. another is
sharing the constant masks without costly GOT accesses in PIC.

the main motivation, however, is accurately representing whether we
want to block signals that might be handled by the application, or all
signals.





there are several reasons for this change. one is getting rid of the
repetition of the syscall signature all over the place. another is
sharing the constant masks without costly GOT accesses in PIC.

the main motivation, however, is accurately representing whether we
want to block signals that might be handled by the application, or all
signals.