[llvmlinux] Rebasing llvmlinux kernel patches to more current kernels

Sedat Dilek sedat.dilek at gmail.com
Sat Apr 9 08:03:16 UTC 2016


On Fri, Apr 8, 2016 at 6:30 PM, Bernhard Rosenkränzer
<bernhard.rosenkranzer at linaro.org> wrote:
> Hi,
> more good news - some progress on the rebase. It finishes compiling now, and
> even starts to boot, but causes an oops around root fs mount time, looks
> like we have a miscompile somewhere in the mmc driver.
>
> The mcount issue is a problem with clang (patch attached, also reported
> upstream), everything else is fixed in the kernel repo.
>

Hi Bernhard,

Thanks again.

In general... How and where do you report?
CCing... Whom using standard mechanisms of Linux-kernel development (see [1])?
Sent to which mailing-lists (CC LLVMLinux ML would make sense)?
Do you have a patchwork-URL (LKML example see [2])?
Does LLVMLinux has a patchwork-URL where patches are collected
(personally I find that very useful)?
I know of some projects having an email-address especially for patches.

[1] http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches
[2] https://patchwork.kernel.org/project/LKML/list/

> More replies inline:
>
> On 6 April 2016 at 11:13, Sedat Dilek <sedat.dilek at gmail.com> wrote:
>>
>> I am interested in Linux v4.4 LTS and X86_64 aka AMD64.
>> Also interested in using stable LLVM/Clang version - not latest Git
>> upstream.
>> Currently, I use llvm-toolchain v3.8.
>
>
> Same here, using 3.8 with a few patches unrelated to the kernel.
>

Do we have a base!?
Developing for ARM/ARM64 only - what's about X86/X86_64?
Are you using prebuilt-toolchain(s)?
How do you build your llvm-toolchain(s)?
Exists a prebuilt-toolchain for X86_64 from Linaro (I prefer to use
the term "AMD64", coming from Debian/Ubuntu)?
Any docs - URLs for them?
Thanks in adavance.

>>
>> There are several issue not solved (see for example [1]) and real
>> show-stopper!
>> One of you commented below.
>>
>> > git://android-git.linaro.org/kernel/hikey-clang.git
>> >
>>
>> Do you have a GitWeb-URL which I can browse with my Firefox?
>
>
> Sure:
> https://android-git.linaro.org/gitweb/kernel/hikey-clang.git/shortlog/refs/heads/android-hikey-linaro-4.4-clang
>

Thanks.
Looked at your patch "Disable the function tracer for now" at [1].
1st label it with "ARM64: ...".
2nd appreciate credits like for the above... Suggested-by Sedat Dilek
<sedat.dilek at gmail.com> (or even Reported-by/Tested-by).
For more details see "13) Using Reported-by:, Tested-by:,
Reviewed-by:, Suggested-by: and Fixes:" section in [2].

[1] https://android-git.linaro.org/gitweb/kernel/hikey-clang.git/commit/9e36975d8e577e2d4d23218235f62d2a93c4a59b
[2] http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches#n539

>> That mcount is a known issue and derives from function-tracer enabled.
>> I recommend to try CONFIG_FTRACE=n.
>
>
> Fixed it properly with the attached patch.
>

So the attached patch is for the ARM world.
How to fix it for X86 world :-)?
( I have an idea. )

>>
>> So you have your own and new places.
>> The question for me is where to do the further development?
>
>
> Forking the git repository to our server was the fastest way to get things
> up - I'll keep stuff there, but wouldn't mind also pushing to the llvmlinux
> repo (but probably the master rebase is more interesting there than the
> hikey branch).
>

Currently, the development is "still" in the hands of LLVMLinux.
But I feel this development somehow "passive" (hopefully not dead) and
personally "not helpful" ( in the sense of getting issues fixed and
upstream-ed.)
There are still "open" issues and a real "showstopper" for X86.

So, thanks again for new impulses :-).
Not sure if I am willing to invest *my* LIFE-TIME.
My life-time is LIMITED, that's what I exactly/precisely know.

Have more fun.

Regards,
- Sedat Dilek -


More information about the LLVMLinux mailing list