[minicoredumper] [PATCH] Read the stack pointer from the 28th field of /proc/PID/stat

Łukasz Stelmach l.stelmach at samsung.com
Tue Jan 9 13:32:04 CET 2018


It was <2018-01-08 pon 20:15>, when John Ogness wrote:
> On 2018-01-08, Łukasz Stelmach <l.stelmach at samsung.com> wrote:
>>>> since Linix v4.9 (commit 0a1eb2d4 precisely) esp and eip values has
>>>> been set to 0 in the stat file. I am working around this problem.
>>>
>>> This problem has already been identified and fixed. The original
>>> proposal (by Marco Felsch) was to manually retrieve this information
>>> [0]. However, this required manually implementing all platforms Linux
>>> supported.
>>
>> I almost finished this for ARM and Intel on Friday myself ;-) I wonder
>> if it could be included as an option?
>
> But why? The kernel provides this information for us already.

Last time I developed something simmilar to minicoredumper (that is why
it didn't take me long to find the stack pointer in the core) we
couldn't access /proc raliably, because some files were not present
anymore when we tried to access them. But it was on an older kernel
(with some weird patches) and the whole system/security configuration
was different.

>> There are configurations where /proc is inaccessible for minicordumper
>> due to MAC (SMACK, SELinux etc.)  restrictions.
>
> The minicoredumper relies on /proc for more information than the stack
> pointer. It would be a new requirement if /proc is not available and we
> would need to carefully look at everything that would need to be
> implemented.

OK. I will work with minicordumper as it is now and I will take notes
if I find any problems.

Thanks a lot.
-- 
Łukasz Stelmach
Samsung R&D Institute Poland
Samsung Electronics
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 472 bytes
Desc: not available
URL: <http://lists.linutronix.de/pipermail/minicoredumper/attachments/20180109/56cf8276/attachment.sig>


More information about the minicoredumper mailing list