The alignment of TLS variables is wrong if accessed from within a thread
for architectures with tls variant TLS_TCB_AT_TP.
For the main thread the static tls data is properly aligned.
For other threads the alignment depends on the alignment of the thread
pointer as the static tls data is located relative to this pointer.
This patch adds this alignment for TLS_TCB_AT_TP variants in the same way
as it is already done for TLS_DTV_AT_TP. The thread pointer is also already
properly aligned if the user provides its own stack for the new thread.
This patch extends the testcase nptl/tst-tls1.c in order to check the
alignment of the tls variables and it adds a pthread_create invocation
with a user provided stack.
The test itself is migrated from test-skeleton.c to test-driver.c
and the missing support functions xpthread_attr_setstack and xposix_memalign
are added.
ChangeLog:
[BZ #23403]
* nptl/allocatestack.c (allocate_stack): Align pointer pd for
TLS_TCB_AT_TP tls variant.
* nptl/tst-tls1.c: Migrate to support/test-driver.c.
Add alignment checks.
* support/Makefile (libsupport-routines): Add xposix_memalign and
xpthread_setstack.
* support/support.h: Add xposix_memalign.
* support/xthread.h: Add xpthread_attr_setstack.
* support/xposix_memalign.c: New File.
* support/xpthread_attr_setstack.c: Likewise.
For a full analysis of both the pthread_rwlock_tryrdlock() stall
and the pthread_rwlock_trywrlock() stall see:
https://sourceware.org/bugzilla/show_bug.cgi?id=23844#c14
In the pthread_rwlock_trydlock() function we fail to inspect for
PTHREAD_RWLOCK_FUTEX_USED in __wrphase_futex and wake the waiting
readers.
In the pthread_rwlock_trywrlock() function we write 1 to
__wrphase_futex and loose the setting of the PTHREAD_RWLOCK_FUTEX_USED
bit, again failing to wake waiting readers during unlock.
The fix in the case of pthread_rwlock_trydlock() is to check for
PTHREAD_RWLOCK_FUTEX_USED and wake the readers.
The fix in the case of pthread_rwlock_trywrlock() is to only write
1 to __wrphase_futex if we installed the write phase, since all other
readers would be spinning waiting for this step.
We add two new tests, one exercises the stall for
pthread_rwlock_trywrlock() which is easy to exercise, and one exercises
the stall for pthread_rwlock_trydlock() which is harder to exercise.
The pthread_rwlock_trywrlock() test fails consistently without the fix,
and passes after. The pthread_rwlock_tryrdlock() test fails roughly
5-10% of the time without the fix, and passes all the time after.
Signed-off-by: Carlos O'Donell <carlos@redhat.com>
Signed-off-by: Torvald Riegel <triegel@redhat.com>
Signed-off-by: Rik Prohaska <prohaska7@gmail.com>
Co-authored-by: Torvald Riegel <triegel@redhat.com>
Co-authored-by: Rik Prohaska <prohaska7@gmail.com>
Without this fix, the rwlock can fail to execute the explicit hand-over
in certain cases (e.g., empty critical sections that switch quickly between
read and write phases). This can then lead to errors in how __wrphase_futex
is accessed, which in turn can lead to deadlocks.
Current allocate_stack logic for create stacks is to first mmap all
the required memory with the desirable memory and then mprotect the
guard area with PROT_NONE if required. Although it works as expected,
it pessimizes the allocation because it requires the kernel to actually
increase commit charge (it counts against the available physical/swap
memory available for the system).
The only issue is to actually check this change since side-effects are
really Linux specific and to actually account them it would require a
kernel specific tests to parse the system wide information. On the kernel
I checked /proc/self/statm does not show any meaningful difference for
vmm and/or rss before and after thread creation. I could only see
really meaningful information checking on system wide /proc/meminfo
between thread creation: MemFree, MemAvailable, and Committed_AS shows
large difference without the patch. I think trying to use these
kind of information on a testcase is fragile.
The BZ#18988 reports shows that the commit pages are easily seen with
mlockall (MCL_FUTURE) (with lock all pages that become mapped in the
process) however a more straighfoward testcase shows that pthread_create
could be faster using this patch:
--
static const int inner_count = 256;
static const int outer_count = 128;
static
void *thread1(void *arg)
{
return NULL;
}
static
void *sleeper(void *arg)
{
pthread_t ts[inner_count];
for (int i = 0; i < inner_count; i++)
pthread_create (&ts[i], &a, thread1, NULL);
for (int i = 0; i < inner_count; i++)
pthread_join (ts[i], NULL);
return NULL;
}
int main(void)
{
pthread_attr_init(&a);
pthread_attr_setguardsize(&a, 1<<20);
pthread_attr_setstacksize(&a, 1134592);
pthread_t ts[outer_count];
for (int i = 0; i < outer_count; i++)
pthread_create(&ts[i], &a, sleeper, NULL);
for (int i = 0; i < outer_count; i++)
pthread_join(ts[i], NULL);
assert(r == 0);
}
return 0;
}
--
On x86_64 (4.4.0-45-generic, gcc 5.4.0) running the small benchtests
I see:
$ time ./test
real 0m3.647s
user 0m0.080s
sys 0m11.836s
While with the patch I see:
$ time ./test
real 0m0.696s
user 0m0.040s
sys 0m1.152s
So I added a pthread_create benchtest (thread_create) which check
the thread creation latency. As for the simple benchtests, I saw
improvements in thread creation on all architectures I tested the
change.
Checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu,
arm-linux-gnueabihf, powerpc64le-linux-gnu, sparc64-linux-gnu,
and sparcv9-linux-gnu.
[BZ #18988]
* benchtests/thread_create-inputs: New file.
* benchtests/thread_create-source.c: Likewise.
* support/xpthread_attr_setguardsize.c: Likewise.
* support/Makefile (libsupport-routines): Add
xpthread_attr_setguardsize object.
* support/xthread.h: Add xpthread_attr_setguardsize prototype.
* benchtests/Makefile (bench-pthread): Add thread_create.
* nptl/allocatestack.c (allocate_stack): Call mmap with PROT_NONE and
then mprotect the required area.
The commit documents the ownership rules around 'struct pthread' and
when a thread can read or write to the descriptor. With those ownership
rules in place it becomes obvious that pd->stopped_start should not be
touched in several of the paths during thread startup, particularly so
for detached threads. In the case of detached threads, between the time
the thread is created by the OS kernel and the creating thread checks
pd->stopped_start, the detached thread might have already exited and the
memory for pd unmapped. As a regression test we add a simple test which
exercises this exact case by quickly creating detached threads with
large enough stacks to ensure the thread stack cache is bypassed and the
stacks are unmapped. Before the fix the testcase segfaults, after the
fix it works correctly and completes without issue.
For a detailed discussion see:
https://www.sourceware.org/ml/libc-alpha/2017-01/msg00505.html
The new test driver in <support/test-driver.c> has feature parity with
the old one. The main difference is that its hooking mechanism is
based on functions and function pointers instead of macros. This
commit also implements a new environment variable, TEST_COREDUMPS,
which disables the code which disables coredumps (that is, it enables
them if the invocation environment has not disabled them).
<test-skeleton.c> defines wrapper functions so that it is possible to
use existing macros with the new-style hook functionality.
This commit changes only a few test cases to the new test driver, to
make sure that it works as expected.
