Re: [Qemu-devel] [PULL] RISC-V QEMU Port Submission v8.2

[Stefan Hajnoczi]

On Thu, Mar 8, 2018 at 11:41 AM, Michael Clark <address@hidden> wrote:
> On Fri, 9 Mar 2018 at 12:18 AM, Michael Clark <address@hidden> wrote:
>
>> On Fri, 9 Mar 2018 at 12:10 AM, Michael Clark <address@hidden> wrote:
>>
>>> On Thu, 8 Mar 2018 at 11:02 PM, Peter Maydell <address@hidden>
>>> wrote:
>>>
>>>> On 6 March 2018 at 19:46, Michael Clark <address@hidden> wrote:
>>>> > -----BEGIN PGP SIGNED MESSAGE-----
>>>> > Hash: SHA1
>>>> >
>>>> > The following changes since commit
>>>> f32408f3b472a088467474ab152be3b6285b2d7b:
>>>> >
>>>> >   misc: don't use hwaddr as a type in trace events (2018-03-06
>>>> 14:24:30 +0000)
>>>> >
>>>> > are available in the git repository at:
>>>> >
>>>> >   https://github.com/riscv/riscv-qemu.git
>>>> tags/riscv-qemu-upstream-v8.2
>>>> >
>>>> > for you to fetch changes up to
>>>> 7051b081bf6796e5e84406f6223a7c4900bf7298:
>>>> >
>>>> >   RISC-V - Remove support for adhoc non-standard X_COP local-interrupt
>>>> (2018-03-07 08:36:03 +1300)
>>>>
>>>>
>>>> Hi -- I would have applied this, but some of the commits
>>>> have no signed-off-by lines.
>>>>
>>>> This is important, and I've already asked for it once. We cannot
>>>> accept anything that doesn't have a clear record in the commit
>>>> message of everybody (person or company) who's contributed code
>>>> to it, indicating that they're happy for their copyrighted
>>>> contributions to be taken into QEMU under our license. Lists
>>>> of names without emails in the cover letter are not sufficient.
>>>>
>>>> In fact a lot of the last part of this patchset looks like
>>>> unreviewed changes/fixes that if we were going to have them we
>>>> should have squashed into the correct patches and resent the
>>>> series for review. Please don't do this. Code review is an
>>>> important part of how the QEMU project works.
>>>
>>>
>>> You must be looking at the wrong tag. There are multiple sign-offs in all
>>> 23 commits. The tag is riscv-qemu-upstream-v8.2. Sagar and Bastian
>>> contacted me out of band to add their sign-offs. Please look at the commits
>>> again and tell me which commit id doesn’t have a sign-off on that tag (23
>>> commits iirc)
>>>
>>
>> I can forward you the mail out-of-band. I had to contact contributors to
>> get them to agree to change the license from MIT to GPLv2, based on a
>> request from Red Hat.
>>
>> You are making this very hard. Do you work for Arm perchance? I really
>> wouldn’t be surprised if our port is being sandbagged by Arm. Apologies for
>> being so direct about this, but things like this happen...
>>
>> I have complied with practically every review request and the sign-offs
>> are there. It’s a bit ridiculous.
>>
>> It would be nice to find someone neutral, unrelated to Arm, to merge our PR
>>
>
> Some history on the origins of RISC to put things in perspective:
>
> https://en.m.wikipedia.org/wiki/Berkeley_RISC
>
> David Patterson worked with Andrew Waterman and Krste Asanovic on the
> design of RISC-V. Sagar did most of the work on the QEMU port and he
> agreeded to sign off on all patches. The SiFive patches only have sign-offs
> from SiFive because SiFive was the sole contributor for its hardware model,
> beside the SiFiveUART which has Stefan’s sign-off.
>
> In any case it seems there is not enough review bandwidth in the QEMU
> project as a whole and the policy to accept contributions is too strict to
> be reasonable, given earnest attempts to comply with *all* review feedback.
> Not impressed.

Please take a break and stop sending emails which are going to offend
the people you need to collaborate with in order to get RISC-V support
merged.

The issues that Peter Maydell raised are completely routine pull
request requirements.  Linux has the same Signed-off-by requirement,
so it's not a QEMU-specific hurdle.

If you want to be the RISC-V maintainer in QEMU then work with Peter
Maydell (qemu.git maintainer).

Stefan