Re: [LTP] [PATCH v2] munmap() don't check sysctl_max_mapcount

[Garrett C. <yan...@gm...>]

On Mon, Dec 6, 2010 at 11:36 PM,  <ca...@re...> wrote:
> >From http://lkml.org/lkml/2009/10/2/85:
>
> On ia64, the following test program exit abnormally, because glibc
> thread library called abort().
>
>  ========================================================
>  (gdb) bt
>  #0  0xa000000000010620 in __kernel_syscall_via_break ()
>  #1  0x20000000003208e0 in raise () from /lib/libc.so.6.1
>  #2  0x2000000000324090 in abort () from /lib/libc.so.6.1
>  #3  0x200000000027c3e0 in __deallocate_stack () from
>     /lib/libpthread.so.0
>  #4  0x200000000027f7c0 in start_thread () from /lib/libpthread.so.0
>  #5  0x200000000047ef60 in __clone2 () from /lib/libc.so.6.1
>  ========================================================
> The fact is, glibc call munmap() when thread exitng time for freeing
> stack, and it assume munlock() never fail. However, munmap() often make
> vma splitting and it with many mapcount make -ENOMEM.
>
> Oh well, stack unfreeing is not reasonable option. Also munlock() via
> free() shouldn't failed.
>
> Thus, munmap() shoudn't check max-mapcount.

Committed.