Re: [PATCH v2] nptl: Handle spurious EINTR when thread cancellation is disabled (BZ#29029)

[Adhemerval Zanella <adhemerval.zanella@linaro.org>]

On 19/04/2022 07:44, Szabolcs Nagy wrote:
> The 04/14/2022 12:49, Adhemerval Zanella via Libc-alpha wrote:
>> Some Linux interfaces never restart after being interrupted by a signal
>> handler, regardless of the use of SA_RESTART [1].  It means that for
>> pthread cancellation, if the target thread disables cancellation with
>> pthread_setcancelstate and calls such interfaces (like poll or select),
>> it should not see spurious EINTR failures due the internal SIGCANCEL.
>>
>> However recent changes made pthread_cancel to always sent the internal
>> signal, regardless of the target thread cancellation status or type.
>> To fix it, the previous semantic is restored, where the cancel signal
>> is only sent if the target thread has cancelation enabled in
>> asynchronous mode.
>>
>> The cancel state and cancel type is moved back to cancelhandling
>> and atomic operation are used to synchronize between threads.  The
>> patch essentially revert the following commits:
>>
>>   8c1c0aae20 nptl: Move cancel type out of cancelhandling
>>   2b51742531 nptl: Move cancel state out of cancelhandling
>>   26cfbb7162 nptl: Remove CANCELING_BITMASK
>>
>> However I changed the atomic operation to follow the internal C11
>> semantic and removed the MACRO usage, it simplifies a bit the
>> resulting code (and removes another usage of the old atomic macros).
>>
>> Checked on x86_64-linux-gnu, i686-linux-gnu, aarch64-linux-gnu,
>> and powerpc64-linux-gnu.
>>
>> [1] https://man7.org/linux/man-pages/man7/signal.7.html
>>
>> Reviewed-by: Florian Weimer <fweimer@redhat.com>
>> Tested-by: Aurelien Jarno <aurelien@aurel32.net>
>> ---
>> v2: Fixed some typos and extended pthread_cancel comments.
> 
> 
> since this commit various cancel tests fail for me (unreliably)
> on aarch64 e.g. failures from 2 different test runs:
> 
> FAIL: nptl/tst-cancel17
> FAIL: nptl/tst-cancelx5
> FAIL: nptl/tst-cond7
> FAIL: nptl/tst-pthread-raise-blocked-self
> FAIL: nptl/tst-pthread_cancel-select-loop
> 
> FAIL: nptl/tst-cancelx20
> FAIL: nptl/tst-cond7
> FAIL: nptl/tst-cond8
> FAIL: nptl/tst-join12
> FAIL: nptl/tst-key3
> FAIL: nptl/tst-pthread_cancel-select-loop
> 
> an example run of nptl/tst-cond7
> 
> $ elf/ld.so --library-path . nptl/tst-cond7 --direct
> round 0
> child created
> parent: joining now
> round 1
> child created
> parent: joining now
> round 2
> child created
> parent: joining now
> round 3
> child created
> parent: joining now
> round 4
> child created
> parent: joining now
> 
> where it is blocked forever: it seems pthread_cancel returns without
> sending a signal (__pthread_kill_internal is not called) so join hangs.
> 

But the signal should be only sent if thread is is cancelled and has
async cancellation enabled, which is not the case for the tests. I am
trying to reproduce it on an aarch64 machine, but I can't see t any
failure tests above.  I will double check if I revert everything and
if the atomics usage are fully correct.