Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Jeff Johnson]

On 11/3/2026 2:29 AM, Wentao Liang wrote:
> ath11k_core_get_rproc() looks up the target remoteproc with
> rproc_get_by_phandle(), which takes a device reference and a reference on
> the remoteproc driver's module.  The resulting pointer is stored in
> ab_ahb->tgt_rproc and never released: ath11k_ahb_free_resources() only
> tears down the firmware resources, and the probe error paths after the
> lookup jump to err_core_free without dropping the reference either.
> rproc_shutdown() does not release it, it only drops the power reference.
> 
> Release the reference when the AHB resources are freed and on the probe
> error path, so that repeated bind/unbind cycles no longer accumulate
> remoteproc references and the remoteproc module can be unloaded again.
> 
> Fixes: ba929d6fe31a ("ath11k: Remove rproc references from common core layer")
> Cc: stable@vger.kernel.org
> Assisted-by: DeepSeek-V4.1-Flash and OpenAI Codex
> Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>

+workflows

I am no longer accepting patches from this e-mail address.

Go away or use a separate e-mail address per person working on this project,
with that person actually responding to review feedback and adhering to the
Linux kernel etiquette.

/jeff

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Wentao Liang]

Hi Jeff,

I'm sorry for taking so long to reply and for the confusion and extra
work caused by sending patches from a shared address and failing to
follow up on review feedback.

To keep the existing discussions clear, I'll continue to reply from
vulab@iscas.ac.cn under my own name in the threads for patches I
submitted. I will not send any new patches or revisions.

For context on this submission, static analysis flagged the candidate,
and we used DeepSeek and GPT-6.1 Sol to assist with reviewing it and
preparing the fix.

Again, I'm sorry for the confusion and the extra work this caused.

Wentao

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Laurent Pinchart]

On Tue, Nov 10, 2026 at 10:34:05AM -0000, Wentao Liang wrote:
> Hi Jeff,
> 
> I'm sorry for taking so long to reply and for the confusion and extra
> work caused by sending patches from a shared address and failing to
> follow up on review feedback.
> 
> To keep the existing discussions clear, I'll continue to reply from
> vulab@iscas.ac.cn under my own name in the threads for patches I
> submitted. I will not send any new patches or revisions.
> 
> For context on this submission, static analysis flagged the candidate,
> and we used DeepSeek and GPT-6.1 Sol to assist with reviewing it and
> preparing the fix.

What is your core motivation for doing this ? I understand how someone
would start contributing to the kernel to fix issues they experience, or
to learn about operating systems, or for the fun of tinkering, but none
of that seems to apply to this flood of patches. What's your reason ?

> Again, I'm sorry for the confusion and the extra work this caused.
> 
> Wentao

-- 
Regards,

Laurent Pinchart

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Wentao Liang]

Hi Laurent,

My main goal is to study how to find bugs in the Linux kernel. When
that work finds a bug, I also want to help fix it and contribute to
the kernel.

Static analysis flagged possible bugs, and we reviewed them by hand
with help from DeepSeek and GPT-6.1 Sol. Even after that review, many
of them still seemed worth looking into. I got excited about the
findings and sent more patches than I could follow up on. I regretted
that decision soon afterward. I should have sent fewer patches and
replied sooner. I'm sorry for the extra work this caused.

I'll use this shared address only to reply to feedback I haven't yet
answered on patches I've already sent. I'll go through that feedback
one message at a time and respect the decisions maintainers have
already made. If any of the remaining patches are still worth
considering, I'd be grateful for the chance to answer questions
about them.

I still hope to contribute to the kernel in the future. If I have
the chance to do so, I'll use my own email address, respect the
maintainers' requests, and send only as many patches as I can
follow up on promptly.

Thanks,
Wentao

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Laurent Pinchart]

On Wed, Nov 11, 2026 at 03:39:57AM -0000, Wentao Liang wrote:
> Hi Laurent,
> 
> My main goal is to study how to find bugs in the Linux kernel.

Why ?

> When
> that work finds a bug, I also want to help fix it and contribute to
> the kernel.
> 
> Static analysis flagged possible bugs, and we reviewed them by hand
> with help from DeepSeek and GPT-6.1 Sol. Even after that review, many
> of them still seemed worth looking into. I got excited about the
> findings and sent more patches than I could follow up on. I regretted
> that decision soon afterward. I should have sent fewer patches and
> replied sooner. I'm sorry for the extra work this caused.
> 
> I'll use this shared address only to reply to feedback I haven't yet
> answered on patches I've already sent. I'll go through that feedback
> one message at a time and respect the decisions maintainers have
> already made. If any of the remaining patches are still worth
> considering, I'd be grateful for the chance to answer questions
> about them.
> 
> I still hope to contribute to the kernel in the future. If I have
> the chance to do so, I'll use my own email address, respect the
> maintainers' requests, and send only as many patches as I can
> follow up on promptly.
> 
> Thanks,
> Wentao

-- 
Regards,

Laurent Pinchart

Re: [PATCH RESEND] wifi: ath11k: release the rproc reference on teardown

[Jeff Johnson]

On 11/10/2026 2:34 AM, Wentao Liang wrote:
> Hi Jeff,
> 
> I'm sorry for taking so long to reply and for the confusion and extra
> work caused by sending patches from a shared address and failing to
> follow up on review feedback.
> 
> To keep the existing discussions clear, I'll continue to reply from
> vulab@iscas.ac.cn under my own name in the threads for patches I
> submitted. I will not send any new patches or revisions.
> 
> For context on this submission, static analysis flagged the candidate,
> and we used DeepSeek and GPT-6.1 Sol to assist with reviewing it and
> preparing the fix.
> 
> Again, I'm sorry for the confusion and the extra work this caused.
> 
> Wentao

I am not going to accept any patches from vulab@iscas.ac.cn since the git
commit history shows that has been (ab)used by multiple folks in their SOBs,
and hence IMO violates the DCO:
    Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
    Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
    Signed-off-by: WenTao Liang <vulab@iscas.ac.cn>
    Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
    Signed-off-by: Zishun Yi <vulab@iscas.ac.cn>

If you want to contribute to the drivers I maintain, submit your patches with
an e-mail address that is unique to you. I have already rejected all patches
from vulab@iscas.ac.cn assigned to me in patchwork.kernel.org.

/jeff