Re: [PATCH] USB: sierra: Fix urb leak in sierra_submit_delayed_urbs()

[Wentao Liang]

Hi Johan,

Sorry for the very late reply - a month since this one, and seven weeks since
your first mail - and for the mess I left behind. Thank you for the review
and for the corrections. You are right about all of it, including that I
had not replied to a single mail you ever sent me. Concluding from the
archives that I was behaving like a bot was a reasonable call, and it
should not have taken that to get an answer out of me.

On the patch: you are right that the reference returned by
usb_get_from_anchor() has to be dropped as a put, and that sierra_close()
already did it while the resume path did not. I see you fixed it yourself in
e7daa54802c5 ("USB: serial: sierra: fix urb leak on write completion"); it is
in mainline and has already been queued for stable. Thank you for that, and
for the Reported-by credit - I am sorry you had to spend your time writing a
fix that I should have sent correctly and then followed up on. I am not
resending anything for this thread, and patchwork 14822917 can be dropped.

On the Assisted-by tag: I have not been adding one. Instead I describe the
workflow in the changelog text, which is what
Documentation/process/generated-content.rst asks for. What I write there is
that the candidates come from static analysis tooling we run over the kernel,
and that DeepSeek and GPT-6.1 Sol were used to audit those candidates and to
help write the fixes. I review and build-test every patch before sending it,
and I am responsible for all of it. That description will be in the changelog
of anything I send from now on.

The real problem was not the patches themselves but the volume: I was posting
faster than I could read the replies, and then I stopped reading them at all.
I am going through the outstanding comments by hand now, one thread at a time,
and answering them before I write anything new. I will not send patches in
bulk again, on this list or any other.

Thanks,
Wentao