From patchwork Mon Jun 4 14:34:14 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: rcardoso@linux.vnet.ibm.com X-Patchwork-Id: 27615 X-Patchwork-Delegate: fweimer@redhat.com Received: (qmail 74574 invoked by alias); 4 Jun 2018 14:34:25 -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 74552 invoked by uid 89); 4 Jun 2018 14:34:24 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-24.3 required=5.0 tests=AWL, BAYES_00, GIT_PATCH_0, GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, KAM_LAZY_DOMAIN_SECURITY, KAM_SHORT, RCVD_IN_DNSWL_LOW, UNSUBSCRIBE_BODY autolearn=ham version=3.3.2 spammy=so, longterm, long-term, Per X-HELO: mx0a-001b2d01.pphosted.com Subject: Re: [RFC] powerpc: restore TOC when static longjmp to shared object From: Rogerio Alves To: Alexander Monakov Cc: libc-alpha@sourceware.org References: <4c67a306-d447-1158-1b92-22ed1cfcfa0a@linux.vnet.ibm.com> <8ac334c6-bcf4-7c72-8d28-900fdd3b0bc4@linux.vnet.ibm.com> <2b5a4221-f766-0f57-1aa5-94991c4dfa8e@linux.vnet.ibm.com> Date: Mon, 4 Jun 2018 11:34:14 -0300 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.4.0 MIME-Version: 1.0 In-Reply-To: <2b5a4221-f766-0f57-1aa5-94991c4dfa8e@linux.vnet.ibm.com> X-TM-AS-GCONF: 00 x-cbid: 18060414-0012-0000-0000-00001664AD2B X-IBM-SpamModules-Scores: X-IBM-SpamModules-Versions: BY=3.00009128; HX=3.00000241; KW=3.00000007; PH=3.00000004; SC=3.00000265; SDB=6.01042213; UDB=6.00533654; IPR=6.00821343; MB=3.00021459; MTD=3.00000008; XFM=3.00000015; UTC=2018-06-04 14:34:18 X-IBM-AV-DETECTION: SAVI=unused REMOTE=unused XFE=unused x-cbparentid: 18060414-0013-0000-0000-000053214BDE Message-Id: <4de7760b-04a9-855c-34ac-1565c5532fb9@linux.vnet.ibm.com> X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:, , definitions=2018-06-04_11:, , signatures=0 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 priorityscore=1501 malwarescore=0 suspectscore=50 phishscore=0 bulkscore=0 spamscore=0 clxscore=1015 lowpriorityscore=0 mlxscore=0 impostorscore=0 mlxlogscore=999 adultscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.0.1-1805220000 definitions=main-1806040170 This is v3 of patch. As pointed out by Alexander Monakov we can't rely on alloca/memset since gcc can optimize this out or it can save the registers on the stack and the test will no longer make sense. Now I use small asm code to deliberately overwrite TOC area on caller frame with zero. If the TOC is not restored by longjmp the test will fail and them I removed alloca/memset since it's no longer necessary for the test itself. Em 25-05-2018 16:39, Rogerio Alves escreveu: > > > Em 23-05-2018 11:24, Alexander Monakov escreveu: >> On Wed, 23 May 2018, Rogerio Alves wrote: >>> I don't think that alloca/memset should destroy the caller stack like is >>> happening but, also I think we have to restore TOC to the caller >>> frame after >>> the longjmp in that case also. I don't know if there's any other more >>> direct >>> and robust ways to checking if TOC is correctly restored. I can't >>> think in >>> anything easier than always restore. >> >> I'm not challenging the idea that "always restoring" is an appropriate >> fix. >> My question was about *verifying* that TOC register is restored as it >> ought >> to, i.e. how the testcase needs to work. Right now, the test uses my code >> from Bugzilla that worked okay for demonstration purposes, but has >> issues as >> a long-term testsuite addition: >> >> * it demonstrates the issue in a very intransparent fashion, relying on >>    non-obvious interaction with alloca-memset part of the test; >> >> * when GCC manages to optimize out the alloca-memset part, the test >>    will cease to work for the intended purpose. >> >> Alexander >> > > Ok. I understand your concern. Let me see if I can change this test to > check if the TOC was been restored. > > Rogerio From 0b59a0e92f56901a3512069441d46778fc97c719 Mon Sep 17 00:00:00 2001 From: Rogerio Alves Date: Wed, 16 May 2018 14:20:53 -0500 Subject: [PATCH v3] powerpc: Always restore TOC on longjmp. This patch change longjmp to always restore the TOC pointer (r2 register) to the caller frame on powerpc. This is related to bug 21895[1] that reports a situation where you have a static longjmp to a shared object file. [1] https://sourceware.org/bugzilla/show_bug.cgi?id=21895 2018-05-16 Rogerio A. Cardoso *sysdeps/powerpc/powerpc64/__longjmp-common.S: Remove condition code for restore r2 on longjmp. *setjmp/Makefile: Added tst-setjmp-bug21895-static to test list. Added rules to build test tst-setjmp-bug21895-static Added module setjmp-bug21895 and rules to build a shared object from it. *setjmp/setjmp-bug21895.c: New test file. *setjmp/tst-setjmp-bug21895-static.c: New test file. --- Changes in v3: as pointed out by Alexander Monakov we can't rely on alloca/memset since gcc can optimize this out or it can save the registers on the stack and the test will no longer make sense. Now I use a small asm code to deliberately overwrite TOC area on caller frame with zero. If the TOC is not restored by longjmp the test will fail. Changes in v2: Per Adhemerval Zanella. Fix test fail using the suggestions given. Fix changelog. Fix copyright and indentation for new tests. Change tests to use libsupport instead old test-skeleton. Fix a extra space in __longjmp-common. setjmp/Makefile | 16 +++++- setjmp/setjmp-bug21895.c | 41 +++++++++++++++ setjmp/tst-setjmp-bug21895-static.c | 75 ++++++++++++++++++++++++++++ sysdeps/powerpc/powerpc64/__longjmp-common.S | 5 +- 4 files changed, 132 insertions(+), 5 deletions(-) create mode 100644 setjmp/setjmp-bug21895.c create mode 100644 setjmp/tst-setjmp-bug21895-static.c diff --git a/setjmp/Makefile b/setjmp/Makefile index dc2fcc6..d70daa2 100644 --- a/setjmp/Makefile +++ b/setjmp/Makefile @@ -29,9 +29,23 @@ routines := setjmp sigjmp bsd-setjmp bsd-_setjmp \ tests := tst-setjmp jmpbug bug269-setjmp tst-setjmp-fp \ tst-sigsetjmp tst-setjmp-static -tests-static := tst-setjmp-static +tests-static := tst-setjmp-static tst-setjmp-bug21895-static + +modules-names = setjmp-bug21895 include ../Rules $(objpfx)tst-setjmp-fp: $(libm) + +test-modules = $(addprefix $(objpfx),$(addsuffix .so,$(modules-names))) + +ifeq ($(build-shared),yes) +tests: $(test-modules) +endif + +$(objpfx)tst-setjmp-bug21895-static: $(common-objpfx)dlfcn/libdl.a +$(objpfx)tst-setjmp-bug21895-static.out: $(objpfx)setjmp-bug21895.so + +tst-setjmp-bug21895-static-ENV = \ + LD_LIBRARY_PATH=$(objpfx):$(common-objpfx):$(common-objpfx)setjmp:$(common-objpfx)elf diff --git a/setjmp/setjmp-bug21895.c b/setjmp/setjmp-bug21895.c new file mode 100644 index 0000000..d7c0133 --- /dev/null +++ b/setjmp/setjmp-bug21895.c @@ -0,0 +1,41 @@ +/* Shared object part of test for setjmp interoperability with static + dlopen BZ #21895. + Copyright (C) 2017-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include +#include + +jmp_buf jb; +void (*bar)(jmp_buf); + +void +lbar (void) +{ + bar(jb); + for(;;); +} + +void +foo (void) +{ + setjmp(jb); + lbar(); + + for(;;); +} diff --git a/setjmp/tst-setjmp-bug21895-static.c b/setjmp/tst-setjmp-bug21895-static.c new file mode 100644 index 0000000..6ab5340 --- /dev/null +++ b/setjmp/tst-setjmp-bug21895-static.c @@ -0,0 +1,75 @@ +/* Test setjmp interoperability with static dlopen BZ #21895. + Copyright (C) 2017-2018 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include +#include +#include + +static void +bar (jmp_buf jb) +{ + static int i; + if (i++==1) + exit(0); /* Success. */ + + /* Destroy TOC area on caller frame. */ + asm volatile( + "li 14, 0\n\t" + "std 14, 24(%0)" + : + : "r" (jb->__jmpbuf[20]) + ); + + longjmp(jb, i); +} + +static int +do_test (void) +{ + void *h = dlopen("setjmp-bug21895.so", RTLD_NOW); + if (!h) + { + puts(dlerror()); + return 1; + } + + void (*pfoo)(void) = dlsym(h, "foo"); + if (!pfoo) + { + puts(dlerror()); + return 1; + } + + void (**ppbar)(jmp_buf) = dlsym(h, "bar"); + if (!ppbar) + { + puts(dlerror()); + return 1; + } + + *ppbar = bar; + pfoo(); + + for(;;); +} + +/* Make sure the test will not stuck if jmp fails and fall into one of + for(;;). */ +#define TIMEOUT 100 +#include diff --git a/sysdeps/powerpc/powerpc64/__longjmp-common.S b/sysdeps/powerpc/powerpc64/__longjmp-common.S index 0e10b8d..99c17c5 100644 --- a/sysdeps/powerpc/powerpc64/__longjmp-common.S +++ b/sysdeps/powerpc/powerpc64/__longjmp-common.S @@ -130,9 +130,6 @@ L(no_vmx): ld r0,(JB_LR*8)(r3) ld r14,((JB_GPRS+0)*8)(r3) lfd fp14,((JB_FPRS+0)*8)(r3) -#if defined SHARED && !IS_IN (rtld) - std r2,FRAME_TOC_SAVE(r1) /* Restore the callers TOC save area. */ -#endif ld r15,((JB_GPRS+1)*8)(r3) lfd fp15,((JB_FPRS+1)*8)(r3) ld r16,((JB_GPRS+2)*8)(r3) @@ -152,7 +149,7 @@ L(no_vmx): second argument (-4@4), and target address (8@0), respectively. */ LIBC_PROBE (longjmp, 3, 8@3, -4@4, 8@0) mtlr r0 -/* std r2,FRAME_TOC_SAVE(r1) Restore the TOC save area. */ + std r2,FRAME_TOC_SAVE(r1) /* Restore the TOC save area. */ ld r21,((JB_GPRS+7)*8)(r3) lfd fp21,((JB_FPRS+7)*8)(r3) ld r22,((JB_GPRS+8)*8)(r3) -- 2.7.4