Re: [PATCH v3] binfmt_elf: Dump smaller VMAs first in ELF cores
From: Brian Mak
Date: Mon Aug 12 2024 - 14:22:00 EST
On Aug 12, 2024, at 11:05 AM, Kees Cook <kees@xxxxxxxxxx> wrote
> On Sat, Aug 10, 2024 at 07:28:44AM -0500, Eric W. Biederman wrote:
>> Brian Mak <makb@xxxxxxxxxxx> writes:
>>
>>> Large cores may be truncated in some scenarios, such as with daemons
>>> with stop timeouts that are not large enough or lack of disk space. This
>>> impacts debuggability with large core dumps since critical information
>>> necessary to form a usable backtrace, such as stacks and shared library
>>> information, are omitted.
>>>
>>> We attempted to figure out which VMAs are needed to create a useful
>>> backtrace, and it turned out to be a non-trivial problem. Instead, we
>>> try simply sorting the VMAs by size, which has the intended effect.
>>>
>>> By sorting VMAs by dump size and dumping in that order, we have a
>>> simple, yet effective heuristic.
>>
>> To make finding the history easier I would include:
>> v1: https://urldefense.com/v3/__https://lkml.kernel.org/r/CB8195AE-518D-44C9-9841-B2694A5C4002@juniper.net__;!!NEt6yMaO-gk!DavIB4o54KGrCPK44iq9_nJrOpKMJxUAlazBVF6lfKwmMCgLD_NviY088SQXriD19pS0rwhadvc$
>> v2: https://urldefense.com/v3/__https://lkml.kernel.org/r/C21B229F-D1E6-4E44-B506-A5ED4019A9DE@juniper.net__;!!NEt6yMaO-gk!DavIB4o54KGrCPK44iq9_nJrOpKMJxUAlazBVF6lfKwmMCgLD_NviY088SQXriD19pS0G7RQv4o$
>>
>> Acked-by: "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx>
>>
>> As Kees has already picked this up this is quite possibly silly.
>> But *shrug* that was when I was out.
>
> I've updated the trailers. Thanks for the review!
Hi Kees,
Thanks! I think you added it to the wrong commit though.
Please double check and update accordingly.
Regarding the sioc tests from earlier, I've reached a point where I
think I have a compatible virtual NIC (no more ioctl errors), but it's
giving me some mismatched registers error, causing the test to fail. I
can see this same test failure on a vanilla kernel with my setup, so
this is probably either some environment issue or a bug with rr or the
tests. Since all the other tests pass, I'm just going to leave it at
that.
Best,
Brian Mak