lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Wed, 28 Jun 2023 00:03:56 +0500
From:   Muhammad Usama Anjum <usama.anjum@...labora.com>
To:     Osama Muhammad <osmtendev@...il.com>,
        Shuah Khan <skhan@...uxfoundation.org>
Cc:     Muhammad Usama Anjum <usama.anjum@...labora.com>, shuah@...nel.org,
        linux-kernel@...r.kernel.org, linux-kselftest@...r.kernel.org
Subject: Re: [PATCH v2] selftests: prctl: Add new prctl test for PR_SET_NAME

Hi,

Thank you for the patch. This patch cleanly applies on next-20230627.

Unrelated to this patch:
I'm not sure if this patch was written against linux next. Always try to
send patches against latest next tag from following repo:

https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git

On 6/26/23 11:36 PM, Osama Muhammad wrote:
> Hi Shuah,
> 
> Any feedback on this patch?.
> 
> Thanks,
> Osama
> 
> 
> On Sat, 17 Jun 2023 at 18:01, Osama Muhammad <osmtendev@...il.com> wrote:
>>
>> Hi,
>>
>> Yes, I did install the latest kernel headers and TASK_COMM_LEN is not
>> accessible in userspace.
>>
>> I looked into the test which uses TASK_COMM_LEN but the test defines
>> it in its own header file.
>>
>> Example:  https://elixir.bootlin.com/linux/latest/source/tools/testing/selftests/bpf/progs/pyperf.h#L13
>>
>> TASK_COMM_LEN is defined in include/linux/sched.h, but this header
>> file is not exposed to userspace.
>>
>> TASK_COMM_LEN is not defined in /include/uapi/linux/sched.h which is
>> exposed to userspace kernel headers.
>> Please find the link to the header file exposed to user space :-
>> -https://elixir.bootlin.com/linux/v5.15.116/source/include/uapi/linux/sched.h
>>
>> As for arm64/abi/tpidr2.c It includes linux/sched.h which will be
>> /include/uapi/linux/sched.h because the user space program is
>> including it.
>> So it also cannot use TASK_COMM_LEN directly.
>>
>> Regards,
>> Osama
>>
>> On Tue, 13 Jun 2023 at 02:56, Shuah Khan <skhan@...uxfoundation.org> wrote:
>>>
>>> On 6/10/23 07:01, Osama Muhammad wrote:
>>>> Hi all,
>>>>
>>>> I looked into it and tried to use TASK_COMM_LEN in the test. Even
>>>> though I included "linux/sched.h '', I was not able to compile the
>>>> test because it couldn't find it in the header file.
>>>> I dived deep into the issue and turns out header file mapped in
>>>> /usr/include/linux/sched.h is actually mapped to
>>>> /include/uapi/linux/sched.h[1] in linux source,
>>>> where TASK_COMM_LEN is not even defined. Instead TASK_COMM_LEN is
>>>> defined in /include/linux/sched.h which is not mapped to any header
>>>> files in
>>>> userspace(/(/usr/include/linux).
>>>> I also tried to find the TASK_COM_LEN in /usr/include/linux/ but I
>>>> couldn't find it. Following are the search results.
>>>> grep -rnw '/usr/include/linux/' -e 'TASK_COMM_LEN"
>>>> RESULTS OF COMMAND :- /usr/include/linux/taskstats.h:38:#define
>>>> TS_COMM_LEN 32 /* should be >= TASK_COMM_LEN
>>>> Based on this information, I have two questions:
>>>> 1. Would this require a fix to move 'TASK_COMM_LEN' macro from
>>>> /include/linux/sched.h to UAPI headers /include/uapi/linux/sched.h.
>>>> 2. Is there any other way to access TASK_COMM_LEN in the selftest that
>>>> I'm not aware of?
>>>>
>>>> [1]:-https://elixir.bootlin.com/linux/v5.15.116/source/include/uapi/linux/sched.h
>>>>
>>>
>>> The best source is Linux mainline.
>>>
>>> Take a look at test files that include linux/sched.h
>>>
>>> arm64/abi/tpidr2.c is one of them.
>>>
>>> Did you install headers before compiling the test?
>>>   make headers_install
>>>
>>> thanks,
>>> -- Shuah
>>>
>>>

-- 
BR,
Muhammad Usama Anjum

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ