Hi Alex,
On 13 March 2015 at 18:26, Alex Shi alex.shi@linaro.org wrote:
On 03/13/2015 07:47 PM, Sumit Semwal wrote:
Hi Alex,
On 13 March 2015 at 13:02, Alex Shi alex.shi@linaro.org wrote:
Hi Mark, Sumit and Amit,
Please review the merge conflict resolving patch. BTW, do we need to pick up 5aa9ef6f286 on stable kernel?
So, it looks like in AOSP, this is how it is:
2015-02-05 Mark Rutland arm64: Fix up /proc/cpuinfo ( => the above corresponds to the cherry-picked 42b34c73ae40 that you mention.)
2015-02-05 Mark Rutland arm64: cpuinfo: record cpu system register values ( => 5aa9ef6f286 that you're talking about)
In other words, what we think is "newer commit 5aa9ef6f286", actually has 'older' commit 42b34c73ae40 cherry-picked on top of it.
I suspect if you look at the the upstream/android-3.14 [1] branch, direct pulled from AOSP, the right sequence of changes should be clear.
Hi Sumit, thanks for response!
The commit 5aa9ef6f286(arm64: cpuinfo: record cpu system register values) is newer on logical, because it introduced the full conception of cpuinfo_arm64/cpuinfo_store_cpu.
The commit 42b34c73ae40(arm64: Fix up /proc/cpuinfo) isn't the same commit 44b82b7700d05a5 which is in upstream kernel. The former added cpuinfo_store_cpu() for old stable kernel version. It's just used to comfort the stable kernel to have a bit cpuinfo, can not give full info.
Thanks for correcting me; of course, you're right!! I must've misunderstood what you wrote earlier.
Based on this, your changes seem quite ok to me.
So when we merge the stable kernel with android, the conflict is in-avoidable. And we'd better remove this part like what I show.
Remove cpuinfo_arm64/cpuinfo_store_cpu() that comes from commit 42b34c73ae40(used for stable kernel only), since we have newer commit 5aa9ef6f286. Conflicts: arch/arm64/include/asm/cputype.h arch/arm64/kernel/setup.cStrangely, in your merge commit, I don't see arch/arm64/kernel/setup.c as an updated file - perhaps there's some mistake in what changes got picked during conflict-resolution?
the diff just got from 'git show -m'. maybe git thinks the resolution doesn't worth to mention here?
Yeah, maybe so.
<snip> > -- > Thanks > Alex
Best regards, ~Sumit.
Thanks for sharing, guess the following one is which you usually sent pull request? :)
That's correct; I just wanted to tell how it was in AOSP :)
https://android.git.linaro.org/gitweb/kernel/linaro-android.git/shortlog/ref...
-- Thanks Alex
Best regards, Sumit.