From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: Mike Crowe <mac@mcrowe.com>
Cc: libc-alpha@sourceware.org, Michael Kerrisk <mtk.manpages@gmail.com>
Subject: Re: [PATCH 09/13] linux: nptl: Replace lll_futex_{timed_}wait by futex-internal.h
Date: Wed, 25 Nov 2020 12:40:46 -0300 [thread overview]
Message-ID: <a5e9f588-8c9e-9f6f-5743-acb3565da051@linaro.org> (raw)
In-Reply-To: <20201125153231.GA1391@mcrowe.com>
On 25/11/2020 12:32, Mike Crowe wrote:
> On Monday 23 November 2020 at 16:52:52 -0300, Adhemerval Zanella wrote:
>> The idea is to make NPTL implementation to use on the functions
>> provided by futex-internal.h.
>>
>> Checked on x86_64-linux-gnu and i686-linux-gnu.
>> ---
>> nptl/lowlevellock.c | 6 +++---
>> nptl/pthread_mutex_lock.c | 9 +++++----
>> nptl/pthread_mutex_setprioceiling.c | 5 +++--
>> nptl/pthread_mutex_timedlock.c | 6 +++---
>> 4 files changed, 14 insertions(+), 12 deletions(-)
>
> [snip]
>
>> diff --git a/nptl/pthread_mutex_timedlock.c b/nptl/pthread_mutex_timedlock.c
>> index e643eab258..343acf6107 100644
>> --- a/nptl/pthread_mutex_timedlock.c
>> +++ b/nptl/pthread_mutex_timedlock.c
>> @@ -561,9 +561,9 @@ __pthread_mutex_clocklock_common (pthread_mutex_t *mutex,
>> goto failpp;
>> }
>>
>> - lll_futex_timed_wait (&mutex->__data.__lock,
>> - ceilval | 2, &rt,
>> - PTHREAD_MUTEX_PSHARED (mutex));
>> + __futex_abstimed_wait64 (
>> + (unsigned int *) &mutex->__data.__lock, clockid,
>> + ceilval | 2, &rt, PTHREAD_MUTEX_PSHARED (mutex));
>
> I think you've replaced the lll_futex_timed_wait call that expects a
> relative timeout with a __futex_abstimed_wait64 call that expects an
> absolute timeout, yet you still appear to be passing the relative timeout.
>
> However, it turns out that the implementation for the
> PTHREAD_MUTEX_PP_RECURSIVE_NP and friends case appears to be have been
> completely broken with clockid != CLOCK_REALTIME ever since I added it in
> 9d20e22e46d891b929a72b0f35586e079eb083fd anyway since the relative timeout
> is calculated by calling __clock_gettime64(CLOCK_REALTIME) (although at the
> time this was a less obvious __gettimeofday call.)
>
> I'll work on writing some test cases for the those types of mutex in the
> hope of catching both flaws before fixing them.
Indeed, there is no need to calculate the relative timeout anymore. I think
the fix below should pass the absolute timeout directly. I will check
a possible regression tests as well.
diff --git a/nptl/pthread_mutex_timedlock.c b/nptl/pthread_mutex_timedlock.c
index aaaafa21ce..86c5f4446e 100644
--- a/nptl/pthread_mutex_timedlock.c
+++ b/nptl/pthread_mutex_timedlock.c
@@ -508,7 +508,6 @@ __pthread_mutex_clocklock_common (pthread_mutex_t *mutex,
if (__pthread_current_priority () > ceiling)
{
result = EINVAL;
- failpp:
if (oldprio != -1)
__pthread_tpp_change_priority (oldprio, -1);
return result;
@@ -540,37 +539,9 @@ __pthread_mutex_clocklock_common (pthread_mutex_t *mutex,
if (oldval != ceilval)
{
- /* Reject invalid timeouts. */
- if (! valid_nanoseconds (abstime->tv_nsec))
- {
- result = EINVAL;
- goto failpp;
- }
-
- struct __timespec64 rt;
-
- /* Get the current time. */
- __clock_gettime64 (CLOCK_REALTIME, &rt);
-
- /* Compute relative timeout. */
- rt.tv_sec = abstime->tv_sec - rt.tv_sec;
- rt.tv_nsec = abstime->tv_nsec - rt.tv_nsec;
- if (rt.tv_nsec < 0)
- {
- rt.tv_nsec += 1000000000;
- --rt.tv_sec;
- }
-
- /* Already timed out? */
- if (rt.tv_sec < 0)
- {
- result = ETIMEDOUT;
- goto failpp;
- }
-
__futex_abstimed_wait64 (
(unsigned int *) &mutex->__data.__lock, clockid,
- ceilval | 2, &rt, PTHREAD_MUTEX_PSHARED (mutex));
+ ceilval | 2, abstime, PTHREAD_MUTEX_PSHARED (mutex));
}
}
while (atomic_compare_and_exchange_val_acq (&mutex->__data.__lock,
next prev parent reply other threads:[~2020-11-25 15:40 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-23 19:52 [PATCH 01/13] linux: Remove unused internal futex functions Adhemerval Zanella
2020-11-23 19:52 ` [PATCH 02/13] nptl: Remove futex_wait_cancelable Adhemerval Zanella
2020-11-24 18:01 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 03/13] nptl: Remove clockwait_tid Adhemerval Zanella
2020-11-24 18:13 ` Lukasz Majewski
2020-12-14 12:16 ` Florian Weimer
2020-12-14 12:47 ` Andreas Schwab
2020-12-14 13:11 ` Florian Weimer
2020-12-14 14:02 ` Florian Weimer
2020-12-14 12:52 ` Adhemerval Zanella
2020-11-23 19:52 ` [PATCH 04/13] linux: Extend __futex_abstimed_wait_cancelable64 comment Adhemerval Zanella
2020-11-24 18:16 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 05/13] linux: nptl: Consolidate __futex_abstimed_wait_{cancelable}64 Adhemerval Zanella
2020-11-24 18:19 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 06/13] linux: nptl: Remove _futex_clock_wait_bitset64 Adhemerval Zanella
2020-11-24 18:26 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 07/13] linux: nptl: Replace __futex_clocklock_wait64 with __futex_abstimed_wait64 Adhemerval Zanella
2020-11-24 21:28 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 08/13] linux: nptl: Replace lll_timedwait " Adhemerval Zanella
2020-11-24 21:29 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 09/13] linux: nptl: Replace lll_futex_{timed_}wait by futex-internal.h Adhemerval Zanella
2020-11-24 21:35 ` Lukasz Majewski
2020-11-25 15:32 ` Mike Crowe
2020-11-25 15:40 ` Adhemerval Zanella [this message]
2020-11-25 15:46 ` Mike Crowe
2020-11-25 17:19 ` Adhemerval Zanella
2020-11-25 17:37 ` Mike Crowe
2020-11-25 17:56 ` Adhemerval Zanella
2020-11-23 19:52 ` [PATCH 10/13] linux: nptl: Replace lll_futex_supported_clockid with futex-internal.h Adhemerval Zanella
2020-11-24 21:36 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH 11/13] nptl: Replace lll_futex_wake " Adhemerval Zanella
2020-11-24 21:38 ` Lukasz Majewski
2020-11-23 19:52 ` [PATCH v2 12/13] nptl: Return EINVAL for pthread_mutex_clocklock/PI with CLOCK_MONOTONIC [BZ #26801] Adhemerval Zanella
2020-11-24 21:43 ` Lukasz Majewski
2020-11-24 21:49 ` Lukasz Majewski
2020-11-27 17:39 ` Adhemerval Zanella
2020-11-23 19:52 ` [PATCH 13/13] linux: Return EINVAL for invalid clock for pthread_clockjoin_np Adhemerval Zanella
2020-11-24 21:48 ` Lukasz Majewski
2020-11-24 22:58 ` Lukasz Majewski
2020-11-24 17:51 ` [PATCH 01/13] linux: Remove unused internal futex functions Lukasz Majewski
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=a5e9f588-8c9e-9f6f-5743-acb3565da051@linaro.org \
--to=adhemerval.zanella@linaro.org \
--cc=libc-alpha@sourceware.org \
--cc=mac@mcrowe.com \
--cc=mtk.manpages@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).