rpms/perl
7
0
Fork 0
Code Issues Pull Requests Releases Activity

Do not access freed memory when cloning thread

Browse Source
This commit is contained in:
Petr Písař 2012-09-11 09:44:49 +02:00
parent 78af436fd3
commit d0fe0a8384
2 changed files with 65 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
perl-5.16.1-perl-111610-Trouble-with-XS-APItest-t-clone-with-sta.patch Normal file
View File

@ -0,0 +1,56 @@
From 4a808ed163df1057031bc6d085300fe1ef6f57d2 Mon Sep 17 00:00:00 2001
From: Michael Schroeder <mls@suse.de>
Date: Fri, 8 Jun 2012 20:29:54 -0700
Subject: [PATCH] [perl #111610] Trouble with XS-APItest/t/clone-with-stack.t
I ran into a bit of a problem when building perl-5.16.0.
'make test' showed a segfault in ext/XS-APItest/t/clone-with-stack.t.
It seems to be caused by accessing already freed memory, it
segfaults because I have MALLOC_PERTUBE_ set, thus glibc fills
freed memory with some value.
Digging deeper, it seems like perl_clone() does not fix
the cx's blk_oldcop element when doing context cloning, thus
blk_oldcop still points to PL_compiling in the old interp--the
calling scope for the BEGIN block being the compilation of the
code surrounding it--and the POPBLOCK done in leavesub will copy
the data from the old interp to PL_curcop.
After fixing this, it still crashed because interp_dup->Iop was
zero after the runops_standard() call (which is probably
correct as the end of the BEGIN block was reached). So I
also added an if statement that checks the pointer.
---
ext/XS-APItest/APItest.xs | 3 ++-
sv.c | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/ext/XS-APItest/APItest.xs b/ext/XS-APItest/APItest.xs
index 2c0ee61..69b7066 100644
--- a/ext/XS-APItest/APItest.xs
+++ b/ext/XS-APItest/APItest.xs
@@ -3084,7 +3084,8 @@ CODE:
PERL_SET_CONTEXT(interp_dup);
/* continue after 'clone_with_stack' */
- interp_dup->Iop = interp_dup->Iop->op_next;
+ if (interp_dup->Iop)
+ interp_dup->Iop = interp_dup->Iop->op_next;
/* run with new perl */
Perl_runops_standard(interp_dup);
diff --git a/sv.c b/sv.c
index 2034c00..fcd76a9 100644
--- a/sv.c
+++ b/sv.c
@@ -12312,6 +12312,7 @@ Perl_cx_dup(pTHX_ PERL_CONTEXT *cxs, I32 ix, I32 max, CLONE_PARAMS* param)
Perl_croak(aTHX_ "Cloning substitution context is unimplemented");
}
else {
+ ncx->blk_oldcop = (COP*)any_dup(ncx->blk_oldcop, param->proto_perl);
switch (CxTYPE(ncx)) {
case CXt_SUB:
ncx->blk_sub.cv = (ncx->blk_sub.olddepth == 0
--
1.7.11.4

10
perl.spec
View File

@ -29,7 +29,7 @@
Name: perl
Version: %{perl_version}
# release number must be even higher, because dual-lived modules will be broken otherwise
Release: 232%{?dist}
Release: 233%{?dist}
Epoch: %{perl_epoch}
Summary: Practical Extraction and Report Language
Group: Development/Languages
@ -82,6 +82,9 @@ Patch9: perl-5.14.2-find2perl-transtate-question-mark-properly.patch
# Fix broken atof, rhbz#835452, RT#109318
Patch10: perl-5.16.0-fix-broken-atof.patch
# Do not access freed memory when cloning thread, rhbz#825749, RT#111610
Patch11: perl-5.16.1-perl-111610-Trouble-with-XS-APItest-t-clone-with-sta.patch
# Fix searching for Unicode::Collate::Locale data, rhbz#756118, CPANRT#72666,
# fixed in Unicode-Collate-0.87.
# TODO Looks like it was fixed differently?
@ -1319,6 +1322,7 @@ tarball from perl.org.
%patch8 -p1
%patch9 -p1
%patch10 -p1
%patch11 -p1
#copy the example script
cp -a %{SOURCE5} .
@ -1521,6 +1525,7 @@ pushd %{build_archlib}/CORE/
'Fedora Patch7: Dont run one io test due to random builder failures' \
'Fedora Patch9: Fix find2perl to translate ? glob properly (RT#113054)' \
'Fedora Patch10: Fix broken atof (RT#109318)' \
'Fedora Patch11: Do not access freed memory when cloning thread (RT#111610)' \
%{nil}
rm patchlevel.bak
@ -2632,6 +2637,9 @@ sed \
# Old changelog entries are preserved in CVS.
%changelog
* Tue Sep 11 2012 Petr Pisar <ppisar@redhat.com> - 4:5.16.1-233
- Do not access freed memory when cloning thread (bug #825749)
* Wed Sep 05 2012 Petr Pisar <ppisar@redhat.com> - 4:5.16.1-232
- Move App::Cpan from perl-Test-Harness to perl-CPAN (bug #854577)