[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: glibc-2.2.5 segv in fopen
From: |
Andreas Jaeger |
Subject: |
Re: glibc-2.2.5 segv in fopen |
Date: |
Wed, 27 Mar 2002 09:33:16 +0100 |
User-agent: |
Gnus/5.090006 (Oort Gnus v0.06) XEmacs/21.4 (Artificial Intelligence, i386-suse-linux) |
address@hidden writes:
> Andreas,
>
> Attached is the glibcbug output:
>
>
> In reply to your query, the problem occurs when the filesystem runs out
> of space not inodes. (I deduce this is the case because I can still
> create empty files using touch but I cannot populate them.)
I cannot reproduce this with a simple program. It really looks like
some memory corruption from your side. Please send a minimal program
that shows this behaviour so that I can execute it on my system.
Writing up a minimal program myself showed no problems and therefore I
cannot help you for now,
Andreas
>
> Hope this of some help,
>
> Richard
>
> glibcbugexGm9e
>
> Subject:
> Re: glibc-2.2.5 segv in fopen
> Date:
> Sun, 24 Mar 2002 16:59:36 +0100
> From:
> Andreas Jaeger <address@hidden>
> To:
> address@hidden
> CC:
> address@hidden
> References:
> 1
>
>
>
>
> Richard Polton <address@hidden> writes:
>
>> All,
>>
>> I have come across a reproducible SIGSEGV in fopen in glibc-2.2.5 on
>> linux-2.4.17. glibc was compiled using gcc-3.0.3. I have not used
>> glibcbug because the affected machine does not have mail capability
> ...
>>
>> The issue occurs when there are zero free blocks available in the
>> filesystem and fopen is called.
>>
>> fopen \equiv _IO_new_fopen( "filename","w+b") at iofopen.c:50
>>
>> calls
>>
>> __libc_malloc( bytes=364) at malloc.c:2810
>>
>> which calls in turn
>>
>> chunk_alloc(ar_ptr=0x4090a480,nb=368) at malloc.c:2878
>>
>> ar_ptr seems a little large. I could not work out how the arena_get
>> calls worked in this instance unfortunately.
>>
>
> Please tell us a bit more. Where is it crashing? What's the
> backtrace? Can you send an minimal example program that crashes?
>
> What kind of full disk do you exactly need (inodes, data blocks)?
>
> Andreas
--
Andreas Jaeger
SuSE Labs address@hidden
private address@hidden
http://www.suse.de/~aj