[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Q: ld-linux.2 bug in glibc2.2.2?
From: |
Andreas Jaeger |
Subject: |
Re: Q: ld-linux.2 bug in glibc2.2.2? |
Date: |
27 Feb 2001 08:35:49 +0100 |
User-agent: |
Gnus/5.090001 (Oort Gnus v0.01) XEmacs/21.1 (Channel Islands) |
Basile STARYNKEVITCH <address@hidden> writes:
> Hello all,
>
> I probably have found a bug in ld-linux.2 dynamic loader of
> glibc2.2.2.
>
> A simple example is found on "http://tunes.org/~basile/glibcbug.tgz" a
> Gnuzipped tar file of 3663 bytes and md5sum
> 77c84af1964027f97fb5d809cf9daf3b
>
> which contains
> drwxr-xr-x basile/parents 0 2001-02-26 21:57:32 GlibcBug/
> -rw-r--r-- basile/parents 7334 2001-02-26 21:44:45 GlibcBug/ldscript
> -rw-r--r-- basile/parents 520 2001-02-26 21:44:52 GlibcBug/main2.c
> -rw-r--r-- basile/parents 249 2001-02-26 21:56:44 GlibcBug/Makefile
> -rw-r--r-- basile/parents 1529 2001-02-26 21:56:29 GlibcBug/README
>
> Ok, I am having an unorthodox use of ld-linux.so, but I still expect
> it to be resistant...
>
> I am linking a (simple) program with GNU ld using my own explicit
> ldscript. This ldscript is probably buggy (but I did not found yet its
> bug). But I expect that the dynamic loader should not crash, even when
> working on faulty ELF executables, as produced by my ldscript.
The dynamic linker expect a correct ELF executable.
> I am beginning to try to debug this ld-linux.2 bug myself. Could some
> glibc2.2.2 expert help me:
>
> how to debug ld-linux.2 (it seems that compiling it with -g doesnot
> change anything; so I added my own debugging output, but it is boring
> to use only _dl_debug_message as printing routine)?
>
> what is the make target (or building command) to only build the
> ld-linux.so.2? Building the whole glibc is timeconsuming...
Build the whole glibc and afterwards just run make subdirs=elf
> perhaps this bug has already been corrected.... I am using only the
> published glibc-2.2.2 source tree, not the latest CVS...
>
> If you have any lights on this bug please be kind to email me.
If your binary is broken, it's a bug in your binary and we will not
fix it - we can't check each and every access.
Andreas
--
Andreas Jaeger
SuSE Labs address@hidden
private address@hidden
http://www.suse.de/~aj