Actions
| Post | |
| Subscribe | |
| Unsubscribe |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[arm-gnu] Bug in arm-2009q3-63 arm-none-linux-gnueabi glibc
- To: arm-gnu@xxxxxxxxxxxxxxxx
- Subject: [arm-gnu] Bug in arm-2009q3-63 arm-none-linux-gnueabi glibc
- From: Alexander Stein <alexander.stein@xxxxxxxxxxxxxxxxxxxxx>
- Date: Tue, 9 Feb 2010 12:43:59 +0100
Hello, today I got a segmentation fault inside the glibc function times (sysdeps/unix/sysv/linux/times.c). After some researches I found out that the return value from the syscall can be EFAULT even if buf is NULL which is totally legal. There is already a force_successful_syscall_return() statement in the kernel, but on arm this is not implemented. So, when EFAULT is returned and buf is NULL the return value is a valid value. The is already a comment which says EFAULT and a valid return value can not be distinguished on some platforms. But the access check is senseless with a NULL pointer. As far as I can see, this bug also is valid for the git master. Best Regards Alexander
- Follow-Ups:
- Re: [arm-gnu] Bug in arm-2009q3-63 arm-none-linux-gnueabi glibc
- From: Daniel Jacobowitz
- Re: [arm-gnu] Bug in arm-2009q3-63 arm-none-linux-gnueabi glibc
- Prev by Date: [arm-gnu] libstdc++
- Next by Date: Re: [arm-gnu] libstdc++
- Previous by thread: Re: [arm-gnu] libstdc++
- Next by thread: Re: [arm-gnu] Bug in arm-2009q3-63 arm-none-linux-gnueabi glibc
- Index(es):