valgrind error in ndr_nbt

[Andrew Bartlett]

On Fri, 2012-07-06 at 16:43 -0600, Trever L. Adams wrote:
> On 06/29/2012 02:54 AM, Andrew Bartlett wrote:
> > On Fri, 2012-06-29 at 02:33 -0600, Trever L. Adams wrote:
> >> On 06/29/2012 02:20 AM, Andrew Bartlett wrote:
> >>> I've retitled the bug to get the attention of those who work on the file
> >>> server, as this isn't a specifically AD related bug as far as I can see.
> >>>
> >>> If you can get it all under valgrind, it may help working out the
> >>> details of the use-after-free().
> >>>
> >>> Thanks!
> >>>
> >>> Andrew Bartlett
> >>>
> >> Thank you Andrew. Are there any special instructions for using valgrind
> >> with S4? Or can I just do like I would with any other program and just
> >> put valgrind before with --memcheck=full (I probably have that a bit off)?
> > Run with --trace-children=yes
> >
> >> The use-after-free() is still happening it seems under load even with
> >> the off-line-files gone, so I think I should be able to duplicate it.
> >> (In fact it seems to be more related to the "destination unreachable" I
> >> mentioned an hour or so ago.)
> > Thanks,
> >
> > Andrew Bartlett
> >
> I do not believe this is the segfault problem, but it just came up.
> 
> ==7816==    at 0x8DB8071: _itoa_word (in /usr/lib64/libc-2.15.so)
> ==7816==    by 0x8DBC989: vfprintf (in /usr/lib64/libc-2.15.so)
> ==7816==    by 0x8DE6533: vasprintf (in /usr/lib64/libc-2.15.so)
> ==7816==    by 0x7020F2C: ndr_print_debug_helper (ndr.c:176)
> ==7816==    by 0x701ED11: ndr_print_uint16 (ndr_basic.c:1048)
> ==7816==    by 0xDCD6E71: ndr_print_nbt_rdata_status (ndr_nbt.c:482)
> ==7816==    by 0xDCD7655: ndr_print_nbt_rdata (ndr_nbt.c:622)
> ==7816==    by 0xDCD797E: ndr_print_nbt_res_rec (ndr_nbt.c:668)
> ==7816==    by 0xDCD8AD9: ndr_print_nbt_name_packet (ndr_nbt.c:819)
> ==7816==    by 0x7021344: ndr_print_debug (ndr.c:252)
> ==7816==    by 0x105C7BFF: nbt_name_reply_send (nbtsocket.c:461)
> ==7816==    by 0x14264FB3: nbtd_node_status_reply (nodestatus.c:91)
> ==7816==

This isn't related to your segfault, but if you want to help us chase it
down, the --track-origins valgrind feature might help.

Thanks,

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org