528204 - firefox 3.6b2 crashes on some websites when compiled with -D_FORTIFY_SOURCE

Status: RESOLVED DUPLICATE of bug 500607

(Core :: General, defect)

Description

[Hicham]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.5) Gecko/20091105 Fedora/3.5.5-1.fc12 Firefox/3.5.5
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.5) Gecko/20091105 Fedora/3.5.5-1.fc12 Firefox/3.5.5

I compiled firefox 3.6b2 on Fedora 12, here is my buildconfig :
Build platform
target
i686-pc-linux-gnu

Build tools
Compiler 	Version 	Compiler flags
gcc 	gcc version 4.4.2 20091027 (Red Hat 4.4.2-7) (GCC) 	-Wall -W -Wno-unused -Wpointer-arith -Wcast-align -W -Wno-long-long -pedantic -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i686 -mtune=atom -fasynchronous-unwind-tables -fno-strict-aliasing -pthread -pipe -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno-reorder-functions
c++ 	gcc version 4.4.2 20091027 (Red Hat 4.4.2-7) (GCC) 	-fno-rtti -fno-exceptions -Wall -Wpointer-arith -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wcast-align -Wno-invalid-offsetof -Wno-long-long -pedantic -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i686 -mtune=atom -fasynchronous-unwind-tables -fno-strict-aliasing -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno-reorder-functions

Configure arguments
--enable-application=xulrunner --prefix=/usr --libdir=/usr/lib --with-system-nspr --with-system-nss --with-system-jpeg --with-system-zlib --with-system-bz2 --enable-system-hunspell --enable-system-sqlite --enable-system-cairo --with-pthreads --disable-strip --disable-tests --disable-mochitest --disable-installer --disable-debug --enable-optimize --enable-default-toolkit=cairo-gtk2 --enable-pango --enable-svg --enable-canvas --disable-javaxpcom --disable-crashreporter --enable-safe-browsing --enable-extensions=default,python/xpcom --enable-libnotify 

it crashes on some bugzilla website (example http://bugs.gentoo.org/show_bug.cgi?id=292808) reporting that a buffer overflow have been detected and that fortify source failed.

Reproducible: Always

Steps to Reproduce:
1.Compile Firefox-3.6b2 the Fedora way
2.Open http://bugs.gentoo.org/show_bug.cgi?id=292808
3.See the crash
Actual Results:  
crash

Expected Results:  
no crash

I didn't test a build without _D_FORTIFY_SOURCE, but I know it is the cause since it has created some problems on other packages, plus the first line in the backtrace.

Comment 1

[(not currently active) Ted Mielczarek]

What's the backtrace look like?

Comment 2

[Hicham]

Attachment: crash backtrace

Comment 3

[Hicham]

Attachment: patches used for compilation

Comment 4

[Hicham]

should i debug it in gdb ?

Comment 5

[(not currently active) Ted Mielczarek]

That's not super useful, did you strip your binaries?

Comment 6

[Hicham]

yes, I did, I am gonna install the debuginfo package and resend the backtrace again

Comment 7

[Hicham]

Attachment: backtrace obtained with gdb

Comment 8

[Hicham]

and by the way, it crashes on this page too

Comment 9

[(not currently active) Ted Mielczarek]

Can you reproduce this compiling without optimization? Also, I note that it's crashing in hunspell, and you've --enable-system-hunspell.

Comment 10

[timeless]

i'd also suggest --disable-optimize while chasing problems (unless the problem goes away when you use it...)

Comment 11

[Hicham]

i m sorry, but in fact, it is a hunspell issue.
once i disable spell checking in firefox, the issue is gone.
in fact I was using mozilla's hunspell, but symlinking to the system dictionaries.
i am gonna try a build with the system hunspell and let u know then

Comment 12

[Hicham]

well, I would like to express my sincere apologies for reporting a wrong issue.

I just recompiled xulrunner with system hunspell and enable-optimize and everything works now.

but i still have a minor issue, could anyone point me to a better patch than the one i am using (nsFrame.patch, included in the patches tarball ) ? without that patch, i can't compile xulrunner

Comment 13

[Hicham]

I ve written this message from Firefox-3.6b2 on Fedora 12.

Here is the buildconfig :

Build platform
target
i686-pc-linux-gnu

Build tools
Compiler 	Version 	Compiler flags
gcc 	gcc version 4.4.2 20091027 (Red Hat 4.4.2-7) (GCC) 	-Wall -W -Wno-unused -Wpointer-arith -Wcast-align -W -Wno-long-long -pedantic -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i686 -mtune=atom -fasynchronous-unwind-tables -fno-strict-aliasing -pthread -pipe -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno-reorder-functions
c++ 	gcc version 4.4.2 20091027 (Red Hat 4.4.2-7) (GCC) 	-fno-rtti -fno-exceptions -Wall -Wpointer-arith -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wcast-align -Wno-invalid-offsetof -Wno-long-long -pedantic -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 -march=i686 -mtune=atom -fasynchronous-unwind-tables -fno-strict-aliasing -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno-reorder-functions

Configure arguments
--enable-application=xulrunner --enable-application=xulrunner --prefix=/usr --libdir=/usr/lib --with-system-nspr --with-system-nss --with-system-jpeg --with-system-zlib --with-system-bz2 --enable-system-hunspell --enable-system-sqlite --enable-system-cairo --with-pthreads --disable-strip --disable-tests --disable-mochitest --disable-installer --disable-debug --enable-optimize --enable-default-toolkit=cairo-gtk2 --enable-pango --enable-svg --enable-canvas --disable-javaxpcom --disable-crashreporter --enable-safe-browsing --enable-extensions=default 



about:

Build identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2b2) Gecko/20091112 Fedora/3.6-0.2.beta2.fc12 Firefox/3.6b2

Comment 14

[(not currently active) Ted Mielczarek]

If you can reproduce this with the in-tree copy of hunspell, then it's a valid bug and it's important! That's code that we are shipping to users.

Comment 15

[Nickolay_Ponomarev]

Thanks for reporting this, but someone beat you to it :)