Re: [PATCHSET x86/master] add stack protector support for x86_32

[Tejun Heo <tj@kernel.org>]

Tejun Heo wrote:
> Ingo Molnar wrote:
>> * Ingo Molnar <mingo@elte.hu> wrote:
>>
>>> [...] I might not get around testing it today and pushing it out into tip:master, 
>>> but i pushed out the core/percpu bits, should you queue up further changes.
>> ok, activated it for -tip testing, and there's a 64-bit build failure caused by 
>> it:
>>
>> arch/x86/kernel/head64.o: In function `x86_64_start_reservations':
>> head64.c:(.init.text+0x26): undefined reference to `__stack_chk_guard'
>> head64.c:(.init.text+0xc2): undefined reference to `__stack_chk_guard'
>> arch/x86/kernel/head64.o: In function `x86_64_start_kernel':
>> head64.c:(.init.text+0x104): undefined reference to `__stack_chk_guard'
>> head64.c:(.init.text+0x1cd): undefined reference to `__stack_chk_guard'
>> arch/x86/kernel/head.o: In function `reserve_ebda_region':
>> head.c:(.init.text+0xb): undefined reference to `__stack_chk_guard'
>> arch/x86/kernel/head.o:head.c:(.init.text+0x87): more undefined references to 
>> `__stack_chk_guard' follow
> 
> Call to __stack_chk_guard is probably generated automatically.
> Strangely, my gcc only generates calls to __stack_chk_fail.
> 
>  > gcc --version
>  gcc (SUSE Linux) 4.3.2 [gcc-4_3-branch revision 141291]
>  Copyright (C) 2008 Free Software Foundation, Inc.
>  This is free software; see the source for copying conditions.  There is NO
>  warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>  > nm build/vmlinux|grep __stack_chk_
>  00000000f0fdf6cb A __crc___stack_chk_fail
>  ffffffff80d53e50 r __kcrctab___stack_chk_fail
>  ffffffff80d5ff81 r __kstrtab___stack_chk_fail
>  ffffffff80d3d140 r __ksymtab___stack_chk_fail
>  ffffffff80248619 T __stack_chk_fail
> 
> I'll try other compilers but which version are you using?  The
> difference is that before the patchset, -fno-stack-protector was
> always added whether stackprotector was enabled or not so this problem
> wasn't visible (at the cost of bogus stackprotector of course).  We'll
> probably need to add __stack_chk_guard or disable if gcc generates
> such symbol.  I'll play with different gccs.

Can't reproduce with gcc-4.1 or 4.2.  Any chance you're using distcc
w/ a build machine w/ glibc < 2.4?  __stack_chk_guard is the symbol
gcc fetches stack canary from if TLS is not supported, so somehow gcc
thought that TLS wasn't available while building head64.

Thanks.

-- 
tejun