From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 14735 invoked by alias); 26 Jun 2018 06:45:50 -0000 Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-alpha-owner@sourceware.org Received: (qmail 118584 invoked by uid 89); 26 Jun 2018 06:45:33 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-3.0 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_LOW,SPF_PASS autolearn=ham version=3.3.2 spammy=ensuring, HContent-Transfer-Encoding:8bit X-HELO: mx0a-001b2d01.pphosted.com Subject: Re: [PING][PATCH] Fix race in pthread_mutex_lock while promoting to PTHREAD_MUTEX_ELISION_NP [BZ #23275] To: libc-alpha@sourceware.org References: <4614fea8-43b9-0125-1346-603ad8ad66e4@linux.ibm.com> From: Stefan Liebler Date: Tue, 26 Jun 2018 06:45:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.7.0 MIME-Version: 1.0 In-Reply-To: <4614fea8-43b9-0125-1346-603ad8ad66e4@linux.ibm.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 8bit x-cbid: 18062606-0028-0000-0000-000002D4F003 X-IBM-AV-DETECTION: SAVI=unused REMOTE=unused XFE=unused x-cbparentid: 18062606-0029-0000-0000-0000238C51E8 Message-Id: X-SW-Source: 2018-06/txt/msg00774.txt.bz2 PING On 06/19/2018 09:45 AM, Stefan Liebler wrote: > PING > > On 06/12/2018 04:24 PM, Stefan Liebler wrote: >> Hi, >> >> The race leads either to pthread_mutex_destroy returning EBUSY or >> triggering an assertion (See description in bugzilla). >> >> This patch is fixing the race by ensuring that the elision path is >> used in all cases if elision is enabled by the GLIBC_TUNABLES framework. >> >> The __kind variable in struct __pthread_mutex_s is accessed >> concurrently. Therefore we are now using the atomic macros. >> >> The new testcase tst-mutex10 is triggering the race on s390x and >> intel. Presumably also on power, but I don't have access to a power >> machine with lock-elision. At least the code for power is the same as >> on the other two architectures. Can somebody test it on power? >> >> Bye >> Stefan >> >> ChangeLog: >> >>      [BZ #23275] >>      * nptl/tst-mutex10.c: New File. >>      * nptl/Makefile (tests): Add tst-mutex10. >>      (tst-mutex-ENV): New variable. >>      * sysdeps/unix/sysv/linux/s390/force-elision.h: (FORCE_ELISION): >>      Ensure that elision path is used if elision is available. >>      * sysdeps/unix/sysv/linux/powerpc/force-elision.h >>      (FORCE_ELISION): Likewise. >>      * sysdeps/unix/sysv/linux/x86/force-elision.h: (FORCE_ELISION): >>      Likewise. >>      * nptl/pthreadP.h (PTHREAD_MUTEX_TYPE, >>      PTHREAD_MUTEX_TYPE_ELISION, PTHREAD_MUTEX_PSHARED): >>      Use atomic_load_relaxed. >>      * nptl/pthread_mutex_consistent.c (pthread_mutex_consistent): >>      Likewise. >>      * nptl/pthread_mutex_getprioceiling.c >>      (pthread_mutex_getprioceiling): Likewise. >>      * nptl/pthread_mutex_lock.c (__pthread_mutex_lock_full, >>      __pthread_mutex_cond_lock_adjust): Likewise. >>      * nptl/pthread_mutex_setprioceiling.c >>      (pthread_mutex_setprioceiling): Likewise. >>      * nptl/pthread_mutex_timedlock.c (__pthread_mutex_timedlock): >>      Likewise. >>      * nptl/pthread_mutex_trylock.c (__pthread_mutex_trylock): >>      Likewise. >>      * nptl/pthread_mutex_unlock.c (__pthread_mutex_unlock_full): >>      Likewise. >>      * sysdeps/nptl/bits/thread-shared-types.h >>      (struct __pthread_mutex_s): Add comments. >>      * nptl/pthread_mutex_destroy.c (__pthread_mutex_destroy): >>      Use atomic_load_relaxed and atomic_store_relaxed. >>      * nptl/pthread_mutex_init.c (__pthread_mutex_init): >>      Use atomic_store_relaxed. >