Re: [PATCH] spi: nxp-fspi: Fix runtime PM leak in nxp_fspi_probe()

[Wentao Liang]

Hi Mark, Frank,

I am sorry for the very late reply, and for the review time this patch
cost you. It does not apply because it was generated against a tree that
was already a couple of months out of date, and the change it makes had
been merged in the meantime.

Frank, you were right that pm_runtime_resume_and_get() is the correct
call here; that is what the merged code uses.

I rebased on current mainline (v7.4-rc1) and the fix is already there:

  5d14285d60ba ("spi: nxp-fspi: disable runtime PM on probe failures")
  Jiawen Liu, applied 2026-06-30, first released in v7.3

It converts the initial clock enable to pm_runtime_resume_and_get() and
routes the probe failure paths through err_disable_pm, so the unbalanced
reference this patch was about is gone, together with three other failure
paths that returned with runtime PM still enabled:

        ret = pm_runtime_resume_and_get(f->dev);
        if (ret < 0) {
                ret = dev_err_probe(dev, ret, "Failed to enable clock");
                goto err_disable_pm;
        }
        ...
err_disable_pm:
        pm_runtime_dont_use_autosuspend(dev);
        pm_runtime_disable(dev);
        return ret;

So there is nothing left to resend: please drop the patch, I am not
posting a v2. Checking against the current tree before sending is
exactly the step I skipped, and it is now part of what I do before
posting anything.

The quoting point is taken as well; this reply quotes only what it
answers.

This is an important lesson for me. From now on I will go through the
remaining review comments one by one myself and answer them, and I will
not send patches in bulk again: the rate will be what I can personally
review, test and answer.

Thanks,
Wentao