[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#635885: libc6: strchr() behavior changed



Jonathan Nieder <jrnieder@gmail.com> wrote:

> Hi Michal,

Hi, thank you for your response.

> 
> This sounded to me like PR12159, which was fixed by glibc-2.13~64 (Fix
> x86-64 strchr propagation of search byte into all bytes of SSE
> register, 2010-10-25), but I am not able to reproduce it.  I get all
> [0] values (pointers to the end of string), the correct behavior.
> 

This pretty much sounds like it. The disassembly of strchr() in
that libc.so utilizes xmm registers heavily and in that bug report value
"\xff" is used like "-1" in my test case.

Thanks for finding that report, I think this solves my problem.

> I tried to reproduce first using strchr.S directly on Debian sid, then
> using plain libc6 in a squeeze Debian live system running through kvm.
> I compiled with -O0 so I don't think gcc optimized out the strchr
> calls.  Any idea what I'm doing wrong?  What machine architecture do
> you use?

This is Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz.





Reply to: