linux-next: duplicate patches in the sophgo tree

linux-next: duplicate patches in the sophgo tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 857 bytes --]

Hi all,

The following commits are also in the arm-soc tree as different commits
(but the same patches):

  0014d11186f5 ("riscv: dts: sophgo: Add sdhci0 configuration for Huashan Pi")
  161477baee0f ("riscv: dts: sophgo: Add i2c device support for sg2042")
  33ae4c56cacf ("riscv: sophgo: dts: add mmc controllers for SG2042 SoC")
  47e5a8daf0b9 ("dt-bindings: riscv: Add Sipeed LicheeRV Nano board compatibles
")
  4f8fb973389d ("riscv: sophgo: dts: add gpio controllers for SG2042 SoC")
  7f3cf53e4c65 ("riscv: dts: sophgo: Use common "interrupt-parent" for all peri
pherals for sg2042")
  c0a4490b120f ("riscv: dts: sophgo: cv18xx: add DMA controller")
  e8b4716e68ba ("riscv: dts: sophgo: Add mcu device for Milk-V Pioneer")
  f9cc479ddc8c ("dt-bindings: interrupt-controller: Add SOPHGO SG2002 plic")

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: duplicate patches in the sophgo tree

[Chen Wang]

Hi，Stephen,

The arm-soc tree contains these patches is due to I submited a PR to 
Arnd and he merged this today.

And for the sophgo/for-next branch, it does contains these patches. I 
created the PR branch(sophgo/riscv-sophgo-dt-for-next) and cherry-picked 
these patches from sophgo/for-next and submited the PR. I see the 
commits in arm-soc are the same as that from 
sophgo/riscv-sophgo-dt-for-next, but they are different against the 
commit ids from sophgo/for-next due to cherry-pick operation.

So my question is, do we need to make sure commit id the same between PR 
branch and sophgo/for-next branch?

Thanks,

Chen

On 2024/9/6 10:04, Stephen Rothwell wrote:
> Hi all,
>
> The following commits are also in the arm-soc tree as different commits
> (but the same patches):
>
>    0014d11186f5 ("riscv: dts: sophgo: Add sdhci0 configuration for Huashan Pi")
>    161477baee0f ("riscv: dts: sophgo: Add i2c device support for sg2042")
>    33ae4c56cacf ("riscv: sophgo: dts: add mmc controllers for SG2042 SoC")
>    47e5a8daf0b9 ("dt-bindings: riscv: Add Sipeed LicheeRV Nano board compatibles
> ")
>    4f8fb973389d ("riscv: sophgo: dts: add gpio controllers for SG2042 SoC")
>    7f3cf53e4c65 ("riscv: dts: sophgo: Use common "interrupt-parent" for all peri
> pherals for sg2042")
>    c0a4490b120f ("riscv: dts: sophgo: cv18xx: add DMA controller")
>    e8b4716e68ba ("riscv: dts: sophgo: Add mcu device for Milk-V Pioneer")
>    f9cc479ddc8c ("dt-bindings: interrupt-controller: Add SOPHGO SG2002 plic")
>

Re: linux-next: duplicate patches in the sophgo tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1610 bytes --]

Hi Chen,

On Fri, 6 Sep 2024 11:24:14 +0800 Chen Wang <unicorn_wang@outlook.com> wrote:
>
> The arm-soc tree contains these patches is due to I submited a PR to
> Arnd and he merged this today.
> 
> And for the sophgo/for-next branch, it does contains these patches. I
> created the PR branch(sophgo/riscv-sophgo-dt-for-next) and
> cherry-picked these patches from sophgo/for-next and submited the PR.
> I see the commits in arm-soc are the same as that from
> sophgo/riscv-sophgo-dt-for-next, but they are different against the
> commit ids from sophgo/for-next due to cherry-pick operation.
> 
> So my question is, do we need to make sure commit id the same between
> PR branch and sophgo/for-next branch?

If you are submitting everything that is in the for-next branch, you
should just submit that branch as is, that way you have submitted what
has been tested and all the commit id's will not change.

However, if you are submitting a subset, you should probably create a
branch with that subset (as you did) and then rebase the for-next
branch on top of that and then test the result yourself and then let it
it in linux-next for a day or 2 (to check for any new conflicts - there
should be none).  After that, you should submit the subset branch to
arm-soc.  This way (again) you have submitted what has been tested and
all the commit id's will not change.

Right now, you should rebase your for-next branch onto your
riscv-sophgo-dt-for-next branch so that the duplicate patches are
removed (they will be the same commits in both trees).
-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: duplicate patches in the sophgo tree

[Chen Wang]

On 2024/9/6 11:40, Stephen Rothwell wrote:
> Hi Chen,
>
> On Fri, 6 Sep 2024 11:24:14 +0800 Chen Wang <unicorn_wang@outlook.com> wrote:
>> The arm-soc tree contains these patches is due to I submited a PR to
>> Arnd and he merged this today.
>>
>> And for the sophgo/for-next branch, it does contains these patches. I
>> created the PR branch(sophgo/riscv-sophgo-dt-for-next) and
>> cherry-picked these patches from sophgo/for-next and submited the PR.
>> I see the commits in arm-soc are the same as that from
>> sophgo/riscv-sophgo-dt-for-next, but they are different against the
>> commit ids from sophgo/for-next due to cherry-pick operation.
>>
>> So my question is, do we need to make sure commit id the same between
>> PR branch and sophgo/for-next branch?
> If you are submitting everything that is in the for-next branch, you
> should just submit that branch as is, that way you have submitted what
> has been tested and all the commit id's will not change.
>
> However, if you are submitting a subset, you should probably create a
> branch with that subset (as you did) and then rebase the for-next
> branch on top of that and then test the result yourself and then let it
> it in linux-next for a day or 2 (to check for any new conflicts - there
> should be none).  After that, you should submit the subset branch to
> arm-soc.  This way (again) you have submitted what has been tested and
> all the commit id's will not change.
>
> Right now, you should rebase your for-next branch onto your
> riscv-sophgo-dt-for-next branch so that the duplicate patches are
> removed (they will be the same commits in both trees).

Thanks, Stephen. I have updated the sophgo/for-next.

Regards,

Chen

Re: linux-next: duplicate patches in the sophgo tree

[Arnd Bergmann]

On Fri, Sep 6, 2024, at 03:24, Chen Wang wrote:
> Hi，Stephen,
>
> The arm-soc tree contains these patches is due to I submited a PR to 
> Arnd and he merged this today.
>
> And for the sophgo/for-next branch, it does contains these patches. I 
> created the PR branch(sophgo/riscv-sophgo-dt-for-next) and cherry-picked 
> these patches from sophgo/for-next and submited the PR. I see the 
> commits in arm-soc are the same as that from 
> sophgo/riscv-sophgo-dt-for-next, but they are different against the 
> commit ids from sophgo/for-next due to cherry-pick operation.
>
> So my question is, do we need to make sure commit id the same between PR 
> branch and sophgo/for-next branch?

It certainly makes things easier for everyone if you have the
same commit IDs, yes. Aside from tripping Stephen's sanity checks,
having rebased commits in your tree also means that the branch that
gets tested is different from the one that gets merged.

If you have multiple branches that you want to be in linux-next but
get merged through different upstream trees, your for-next branch
would normally just be a merge of those rather than a cherry-pick.

[Side note: you should also ensure that each of those branches
individually works correctly, i.e. it may add features that only work
when combined with the other branches, but it never introduces
regressions when merged without the other ones.]

      Arnd

Re: linux-next: duplicate patches in the sophgo tree

[Chen Wang]

On 2024/9/6 14:54, Arnd Bergmann wrote:
> On Fri, Sep 6, 2024, at 03:24, Chen Wang wrote:
>> Hi，Stephen,
>>
>> The arm-soc tree contains these patches is due to I submited a PR to
>> Arnd and he merged this today.
>>
>> And for the sophgo/for-next branch, it does contains these patches. I
>> created the PR branch(sophgo/riscv-sophgo-dt-for-next) and cherry-picked
>> these patches from sophgo/for-next and submited the PR. I see the
>> commits in arm-soc are the same as that from
>> sophgo/riscv-sophgo-dt-for-next, but they are different against the
>> commit ids from sophgo/for-next due to cherry-pick operation.
>>
>> So my question is, do we need to make sure commit id the same between PR
>> branch and sophgo/for-next branch?
> It certainly makes things easier for everyone if you have the
> same commit IDs, yes. Aside from tripping Stephen's sanity checks,
> having rebased commits in your tree also means that the branch that
> gets tested is different from the one that gets merged.
>
> If you have multiple branches that you want to be in linux-next but
> get merged through different upstream trees, your for-next branch
> would normally just be a merge of those rather than a cherry-pick.
>
> [Side note: you should also ensure that each of those branches
> individually works correctly, i.e. it may add features that only work
> when combined with the other branches, but it never introduces
> regressions when merged without the other ones.]
>
>        Arnd

Thank you Arnd for your careful explaination. I will pay attention to 
these issues later.

Regards,

Chen