From patchwork Sun Jan 13 23:48:09 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jim Wilson X-Patchwork-Id: 31049 Received: (qmail 7925 invoked by alias); 13 Jan 2019 23:49:15 -0000 Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-alpha-owner@sourceware.org Delivered-To: mailing list libc-alpha@sourceware.org Received: (qmail 7914 invoked by uid 89); 13 Jan 2019 23:49:14 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00, GIT_PATCH_0, GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, RCVD_IN_DNSWL_NONE autolearn=ham version=3.3.2 spammy=nptl, sysv, H*RU:209.85.214.196, Hx-spam-relays-external:209.85.214.196 X-HELO: mail-pl1-f196.google.com DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sifive.com; s=google; h=from:to:cc:subject:date:message-id; bh=vkbmfYQ6Uy6uRLNcAYVqYjoc1piU2vBUByNmo70nL/k=; b=lCSNtOHHYSSzOT1sSjfeNXh6qSaCRqCP037kodvWLw/o3bqOH0cbWUSqSUqXSLfnEn ERSwXkma229tFE6w4HJ0QCQYjY8c3UhB7kDy1e6F9kl7KLMmQqVxuuSTM5nPh3rEKkEK i3GAuIBJKek0rp6lAiw5CiEG8cerlase8JbDEtc08UFdJeu/N1AlpeamYohZ1GDqt8sH kh4qNM5pP5XwoKqB2RJL0nSdtTe4bNnGgHEts7As0utNLadlzg9czgNMAwTDR33lrlEQ U/tRSRd14R3RJp5SKjrQDtEC6FC1lTW/1gvW57ARvgQjtspy+4F6Be6exjoVnOZbaOFa dptg== Return-Path: From: Jim Wilson To: libc-alpha@sourceware.org Cc: Jim Wilson Subject: [PATCH] RISC-V: Fix elfutils testsuite unwind failures. Date: Sun, 13 Jan 2019 15:48:09 -0800 Message-Id: <20190113234809.4818-1-jimw@sifive.com> The clone.S patch fixes 2 elfutils testsuite unwind failures, where the backtrace gets stuck repeating __thread_start until we hit the backtrace limit. This was confirmed by building and installing a patched glibc and then building elfutils and running its testsuite. Unfortunately, the testcase isn't working as expected and I don't know why. The testcase passes even when my clone.S patch is not installed. The testcase looks logically similarly to the elfutils testcases that are failing. Maybe there is a subtle difference in how the glibc unwinding works versus the elfutils unwinding? I don't have good gdb pthread support yet, so I haven't found a way to debug this. Anyways, I don't know if the testcase is useful or not. If the testcase isn't useful then maybe the clone.S patch is OK without a testcase? Jim [BZ #24040] * elf/Makefile (CFLAGS-tst-unwind-main.c): Add -DUSE_PTHREADS=0. * elf/tst-unwind-main.c: If USE_PTHEADS, include pthread.h and error.h (func): New. (main): If USE_PTHREADS, call pthread_create to run func. Otherwise call func directly. * nptl/Makefile (tests): Add tst-unwind-thread. (CFLAGS-tst-unwind-thread.c): Define. * nptl/tst-unwind-thread.c: New file. * sysdeps/unix/sysv/linux/riscv/clone.S (__thread_start): Mark ra as undefined. --- elf/Makefile | 2 +- elf/tst-unwind-main.c | 28 ++++++++++++++++++++++++--- nptl/Makefile | 4 +++- nptl/tst-unwind-thread.c | 2 ++ sysdeps/unix/sysv/linux/riscv/clone.S | 5 +++++ 5 files changed, 36 insertions(+), 5 deletions(-) create mode 100644 nptl/tst-unwind-thread.c diff --git a/elf/Makefile b/elf/Makefile index 9cf5cd8dfd..815bbd2041 100644 --- a/elf/Makefile +++ b/elf/Makefile @@ -1497,4 +1497,4 @@ $(objpfx)tst-big-note: $(objpfx)tst-big-note-lib.so $(objpfx)tst-unwind-ctor: $(objpfx)tst-unwind-ctor-lib.so -CFLAGS-tst-unwind-main.c += -funwind-tables +CFLAGS-tst-unwind-main.c += -funwind-tables -DUSE_PTHREADS=0 diff --git a/elf/tst-unwind-main.c b/elf/tst-unwind-main.c index 4089e7e907..0c345dc31f 100644 --- a/elf/tst-unwind-main.c +++ b/elf/tst-unwind-main.c @@ -20,19 +20,41 @@ #include #include +#if USE_PTHREADS +# include +# include +#endif + static _Unwind_Reason_Code callback (struct _Unwind_Context *ctx, void *arg) { return _URC_NO_REASON; } -int -main (void) +static void * +func (void *a) { /* Arrange for this test to be killed if _Unwind_Backtrace runs into an endless loop. We cannot use the test driver because the complete call chain needs to be compiled with -funwind-tables so that - _Unwind_Backtrace is able to reach _start. */ + _Unwind_Backtrace is able to reach the start routine. */ alarm (DEFAULT_TIMEOUT); _Unwind_Backtrace (callback, 0); + return a; +} + +int +main (void) +{ +#if USE_PTHREADS + pthread_t thr; + int rc = pthread_create (&thr, NULL, &func, NULL); + if (rc) + error (1, rc, "pthread_create"); + rc = pthread_join (thr, NULL); + if (rc) + error (1, rc, "pthread_join"); +#else + func (NULL); +#endif } diff --git a/nptl/Makefile b/nptl/Makefile index 340282c6cb..07c2f49235 100644 --- a/nptl/Makefile +++ b/nptl/Makefile @@ -319,7 +319,7 @@ tests = tst-attr1 tst-attr2 tst-attr3 tst-default-attr \ tst-cnd-basic tst-mtx-trylock tst-cnd-broadcast \ tst-cnd-timedwait tst-thrd-detach tst-mtx-basic tst-thrd-sleep \ tst-mtx-recursive tst-tss-basic tst-call-once tst-mtx-timedlock \ - tst-rwlock-pwn + tst-rwlock-pwn tst-unwind-thread tests-internal := tst-rwlock19 tst-rwlock20 \ tst-sem11 tst-sem12 tst-sem13 \ @@ -709,6 +709,8 @@ $(objpfx)tst-audit-threads: $(objpfx)tst-audit-threads-mod2.so $(objpfx)tst-audit-threads.out: $(objpfx)tst-audit-threads-mod1.so tst-audit-threads-ENV = LD_AUDIT=$(objpfx)tst-audit-threads-mod1.so +CFLAGS-tst-unwind-thread.c += -funwind-tables + # The tests here better do not run in parallel ifneq ($(filter %tests,$(MAKECMDGOALS)),) .NOTPARALLEL: diff --git a/nptl/tst-unwind-thread.c b/nptl/tst-unwind-thread.c new file mode 100644 index 0000000000..d5c38e3709 --- /dev/null +++ b/nptl/tst-unwind-thread.c @@ -0,0 +1,2 @@ +#define USE_PTHREADS 1 +#include "../elf/tst-unwind-main.c" diff --git a/sysdeps/unix/sysv/linux/riscv/clone.S b/sysdeps/unix/sysv/linux/riscv/clone.S index c079c1fb9f..0ff9ab3fd9 100644 --- a/sysdeps/unix/sysv/linux/riscv/clone.S +++ b/sysdeps/unix/sysv/linux/riscv/clone.S @@ -69,6 +69,11 @@ L (error): ENTRY (__thread_start) L (thread_start): + /* Terminate call stack by noting ra is undefined. Use a dummy + .cfi_label to force starting the FDE. */ + .cfi_label .Ldummy + cfi_undefined (ra) + /* Restore the arg for user's function. */ REG_L a1,0(sp) /* Function pointer. */ REG_L a0,SZREG(sp) /* Argument pointer. */