From: Rogerio Alves <rcardoso@linux.vnet.ibm.com>
To: Alexander Monakov <amonakov@ispras.ru>
Cc: libc-alpha@sourceware.org
Subject: Re: [RFC] powerpc: restore TOC when static longjmp to shared object
Date: Mon, 04 Jun 2018 14:34:00 -0000 [thread overview]
Message-ID: <4de7760b-04a9-855c-34ac-1565c5532fb9@linux.vnet.ibm.com> (raw)
In-Reply-To: <2b5a4221-f766-0f57-1aa5-94991c4dfa8e@linux.vnet.ibm.com>
[-- Attachment #1: Type: text/plain, Size: 1755 bytes --]
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
>> Â Â non-obvious interaction with alloca-memset part of the test;
>>
>> * when GCC manages to optimize out the alloca-memset part, the test
>> Â Â 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
[-- Attachment #2: 0001-powerpc-Always-restore-TOC-on-longjmp.patch --]
[-- Type: text/x-patch, Size: 7054 bytes --]
From 0b59a0e92f56901a3512069441d46778fc97c719 Mon Sep 17 00:00:00 2001
From: Rogerio Alves <rcardoso@linux.vnet.ibm.com>
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 <rcardoso@linux.vnet.ibm.com>
*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
+ <http://www.gnu.org/licenses/>. */
+
+#include <alloca.h>
+#include <string.h>
+#include <setjmp.h>
+
+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
+ <http://www.gnu.org/licenses/>. */
+
+#include <setjmp.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <dlfcn.h>
+
+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 <support/test-driver.c>
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
next prev parent reply other threads:[~2018-06-04 14:34 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <f88767a0-1a38-564d-7bd8-0b5db5bc04df@linux.vnet.ibm.com>
2018-05-22 18:41 ` Rogerio Alves
2018-05-23 9:27 ` Alexander Monakov
2018-05-23 12:33 ` Rogerio Alves
2018-05-23 13:45 ` Tulio Magno Quites Machado Filho
2018-05-23 14:25 ` Alexander Monakov
2018-05-25 19:39 ` Rogerio Alves
2018-06-04 14:34 ` Rogerio Alves [this message]
2018-06-04 14:38 ` Florian Weimer
2018-06-21 19:14 ` Rogerio Alves
2018-07-16 19:11 ` Tulio Magno Quites Machado Filho
2018-05-15 17:29 Rogerio Alves
2018-05-15 18:53 ` Florian Weimer
2018-05-15 19:15 ` Rogerio Alves
2018-05-15 20:49 ` Tulio Magno Quites Machado Filho
2018-05-16 19:36 ` Rogerio Alves
2018-05-16 20:56 ` Adhemerval Zanella
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=4de7760b-04a9-855c-34ac-1565c5532fb9@linux.vnet.ibm.com \
--to=rcardoso@linux.vnet.ibm.com \
--cc=amonakov@ispras.ru \
--cc=libc-alpha@sourceware.org \
/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).