Re: [PATCH v6 1/5] target/riscv: add thead-c908 cpu support

[Chao Liu <chao.liu.zevorn@gmail.com>]

On Fri, May 01, 2026 at 02:02:04PM +1000, Alistair Francis wrote:
> On Mon, Apr 20, 2026 at 1:38 PM Chao Liu <chao.liu.zevorn@gmail.com> wrote:
> >
> > On Mon, Apr 20, 2026 at 09:49:50AM +1000, Alistair Francis wrote:
> > > On Sat, Apr 18, 2026 at 1:13 AM Chao Liu <chao.liu.zevorn@gmail.com> wrote:
> > > >
> > > > From: Chao Liu <chao.liu@zevorn.cn>
> > > >
> > > > The C908 processor is based on the RV64GCB[V] instruction
> > > > set, compatible to RVA22 Profile and implements the XIE
> > > > (XuanTie Instruction Extension) technology.
> > > >
> > > > Signed-off-by: Chao Liu <chao.liu.zevorn@gmail.com>
> > > > Suggested-by: LIU Zhiwei <zhiwei_liu@linux.alibaba.com>
> > > > Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> > > > Reviewed-by: Conor Dooley <conor@kernel.org>
> > >
> > > As Connor pointed out, he didn't add his Reviewed-by, instead he
> > > commented about a missing extension [1]. On top of that I don't see
> > > any instances of `Reviewed-by: Conor Dooley <conor@kernel.org>` in the
> > > QEMU tree, so I assume there are none on the mailing list as well.
> > > Which makes a copy and paste error seem less likely.
> > >
> > > I do see that Connor's reply was sent from his conor@kernel.org address,
> > > but manually stitching together a `Reviewed-by` and address seems
> > > unlikely to happen by accident. Although it does seem like the type of
> > > thing an LLM would do (see below).
> > >
> > Sorry, when I first started contributing to QEMU, I wasn't familiar with
> > mailing list conventions. I had been under the impression that whenever a
> > reviewer left comments, it was appropriate to carry a Reviewed-by tag into
> > the next revision. No one pointed this out at the time, but that was
> > clearly my oversight. My apologies to Alistair and Conor for the trouble
> > caused.
> 
> In future, only add tags that people specify. Please copy them
> directly or use b4, patchew or a similar tool to apply the tags.

Thanks for the review, way batter than manually tagging people :)

> 
> >
> > I made the same mistake on earlier patches as well:
> >
> > https://lore.kernel.org/qemu-devel/8b29b86a9af84edba54cc53aec017224478312bf.1745894489.git.lc00631@tecorigin.com/
> > https://lore.kernel.org/qemu-devel/501e578a4ef28515ccdefcbc82defc04363855ca.1746001489.git.lc00631@tecorigin.com/
> >
> > There is also a slightly different case where I added a Reviewed-by tag:
> >
> > https://lore.kernel.org/qemu-devel/07a97219e3f7e342c668d6772acc4c0a8eb4d4f3.1744709888.git.lc00631@tecorigin.com/
> >
> > There the R-b came from an internal colleague who reviewed the patch on
> > our side before it was sent upstream.
> >
> > Is that also against community convention?
> 
> No, internal colleagues are ok. Generally they provide little insight
> though. Your colleague would be better off engaging with the community
> as then the comments can be seen by everyone, that helps build trust
> between developers.
> 
You're right, we'll move more internal discussions to community
channels, so other developers can stay informed.

> >
> > > > Reported-by: Peng Jiang <3160104094@zju.edu.cn>
> > >
> > > I'm not sure how this can be Reported-by, it's not a bug fix. What was
> > > reported? It seems like the Reported-by appeared in v3 [2] and wasn't
> > > in v2 [3], yet I don't see anything reported.
> > >
> > This patchset is the project-phase output of a non-profit QEMU open-source
> > training camp we run in China; so far the camp has trained about 2000
> > participants. Peng Jiang is one of the members, and his task was bringing
> > up OpenSBI/RustSBI + Linux on the QEMU k230 board. For the specific
> > details, see:
> >
> > https://github.com/rustsbi/rustsbi/issues/160
> >
> > The Reported-by in v3 was added as a form of encouragement, since he
> > found several bugs during that bring-up work.
> 
> A Signed-off-by or Tested-by would probably be more appropriate then.
> Reported-by indicates that they found and reported a bug
> 
Peng Jiang tested the patchset, so Tested-by fits better. I will fix it
in the next version.

> >
> > Given this situation, should I also drop Peng Jiang's Reported-by tag?
> >
> > After the camp concluded, I asked him to report any further issues to the
> > upstream list directly himself — for example in v4:
> >
> > https://lore.kernel.org/qemu-devel/651b1d38.63ab7.19be50b4721.Coremail.3160104094@zju.edu.cn/
> >
> > > The extra and unexplained `Reviewed-by` (especially using an odd
> > > address as mentioned above) really points to this being LLM generated.
> > > Especially considering that your reviews are LLM generated (those are
> > > fine). Unfortunately that means I will need to decline the commits
> > > according to the QEMU Code Provenance rules [4]. Note that it isn't
> > > just the Code Provenance policy, it is also now really difficult to
> > > trust these patches knowing that stray `Reviewed-by` tags could be
> > > hallucinated and not caught before being submitted.
> > >
> > That said, I should clarify that no LLM assistance was used in developing
> > this patchset.
> 
> Ok, good to know :)
> 
> >
> > Finally, thanks to Alistair for the patient review. I'll be more careful
> > about these details in future contributions to avoid adding review burden
> > on maintainers/reviewers.
> 
> Thank you
> 
> Alistair