RE: [0/3] gpio: improve claim diagnostics and fix gpio-delay offset handling

AMD General

Hi Simon

Thanks for the review, and for the ask on test coverage.
> -----Original Message-----
> From: Simon Glass <[email protected]>
> Sent: Tuesday, July 14, 2026 2:27 AM
> To: Sanwal, Pranav <[email protected]>
> Cc: [email protected]
> Subject: Re: [0/3] gpio: improve claim diagnostics and fix gpio-delay offset
> handling
>
> Hi Pranav,
>
> On 2026-07-09T14:28:59, Pranav Sanwal <[email protected]> wrote:
>
> > This series improves diagnosability of GPIO offset collisions in the
> > uclass layer and fixes two related bugs in the gpio-delay driver that
> > were uncovered while debugging such a collision.
>
> Thanks for tracking this down.
>
> One thing is missing, though: tests. The gpio-delay driver has no sandbox
> coverage at all, which is presumably why these bugs went unnoticed. Please can
> you add a gpio-delay node to the sandbox test devicetree, wrapping at least two
> GPIO lines, along with a test in test/dm/gpio.c that requests both consumers?
> That would have caught the offset-propagation bug in patch 3 directly and would
> also exercise the bounds check and array sizing from patch 2.

While putting together the gpio-delay node/test you asked for, I found that
test/dm/gpio.c hasn't actually been compiled in years: its Makefile rule was folded
into an unrelated CONFIG_ACPIGEN-gated block, losing the "$" from
"obj-$(CONFIG_DM_GPIO)" in the process. Making it build again surfaces a few
years of bit-rot in the existing tests, plus a separate pre-existing core-DM issue
unrelated to GPIO.

That's more scope than belongs in this series, so I'm sending v2 of the actual fixes
now (patch 1 reworked per your comment), and will follow up separately
with the test/dm build fix and the gpio-delay sandbox coverage you asked for.

> Regards,
> Simon

Thanks,
Pranav Sanwal