[Bug debugedit/28408] New: debugedit segfaults while changing binutils' build-id

[alexey.brodkin at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28408

            Bug ID: 28408
           Summary: debugedit segfaults while changing binutils' build-id
           Product: debugedit
           Version: unspecified
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: debugedit
          Assignee: unassigned at sourceware dot org
          Reporter: alexey.brodkin at gmail dot com
                CC: debugedit at sourceware dot org
  Target Milestone: ---

Created attachment 13694
  --> https://sourceware.org/bugzilla/attachment.cgi?id=13694&action=edit
Cross-compiled m68k library

Initially seen on cross-compilation of Binutils in Debian and then 
reported here https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=995195.

In a nutshell "debugedit" segfaults on attempt to run the following command on
x86_64 host against a binary built for a big-endian architecture in this
example it's "libbfd-2.37-system.so" (attached) cross-compiled for m68k:
------------------------->8-----------------------
$ debugedit --build-id --build-id-seed=x libbfd-2.37-system.so
Segmentation fault
------------------------->8-----------------------

That's what we with GDB:
------------------------->8-----------------------
$ gdb --args ./debugedit --build-id --build-id-seed=x libbfd-2.37-system.so
(gdb) r
Starting program: .../debugedit/debugedit --build-id --build-id-seed=x
libbfd-2.37-system.so
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

Program received signal SIGSEGV, Segmentation fault.
__memmove_avx_unaligned_erms () at
../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:365
365     ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S: No such file
or directory.
(gdb) bt
#0  __memmove_avx_unaligned_erms () at
../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:365
#1  0x000055555555fea8 in memcpy (__len=83, __src=0x0, __dest=0x7fffffffe02d)
at /usr/include/x86_64-linux-gnu/bits/string_fortified.h:34
#2  sha1_process_bytes (buffer=0x0, len=452, ctx=0x7fffffffdfe0) at
tools/sha1.c:210
#3  0x0000555555557d43 in handle_build_id (build_id_size=<optimized out>,
build_id_offset=16, build_id=0x555555566448, dso=0x5555555678d0) at
tools/debugedit.c:3285
#4  main (argc=<optimized out>, argv=<optimized out>) at tools/debugedit.c:3638
------------------------->8-----------------------

If we look a bit deeper we may notice that while processing ".bss" section
we're   not seeing "u.shdr.sh_type == SHT_NOBITS" here
https://sourceware.org/git/?p=debugedit.git;a=blob;f=tools/debugedit.c;h=668777ad47d9b51d0cd118b7d91281963dafa6be;hb=HEAD#l3276.
Instead we see "u.shdr.sh_type = 0x08000000", which is exactly byte-swapped 0x8
which stands for SHT_NOBITS.

And since that check doesn't catch NOBITS section we later segfault trying to
process some other garbage.

That said there's something wrong with interpretation of a foreign endianess.


And this could be reproduced with both debugedit & elfutils built from today's
"master" barnches.

-- 
You are receiving this mail because:
You are on the CC list for the bug.