Re: [PATCH] i2c: omap: fix deprecated of_property_read_bool() use

[Johan Hovold <johan@kernel.org>]

On Tue, Apr 29, 2025 at 03:10:13PM +0200, Andi Shyti wrote:
> On Fri, Apr 18, 2025 at 11:57:57AM +0200, Johan Hovold wrote:
> > On Thu, Apr 17, 2025 at 11:41:51PM +0200, Andi Shyti wrote:
> > > On Tue, Apr 15, 2025 at 09:52:30AM +0200, Johan Hovold wrote:
> > > > Using of_property_read_bool() for non-boolean properties is deprecated
> > > > and results in a warning during runtime since commit c141ecc3cecd ("of:
> > > > Warn when of_property_read_bool() is used on non-boolean properties").
> > > > 
> > > > Fixes: b6ef830c60b6 ("i2c: omap: Add support for setting mux")
> > > > Cc: Jayesh Choudhary <j-choudhary@ti.com>
> > > > Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
> > > 
> > > Thanks for your patch! I'm going to drop the Fixes tag, as this
> > > isn't really a bug fix but rather a warning suppression during
> > > boot time.
> > 
> > Thanks, but I think you should have kept the Fixes tag and merged this
> > for 6.15 (i2c-host-fixes) since this is a new warning in 6.15-rc1 (and
> > that does warrant a Fixes tag). Perhaps I should have highlighted that
> > better.
> > 
> > If the offending patch had been posted or merged before such uses
> > started generating warnings in 6.14-rc1 then that would have been a
> > different matter.
> 
> I'm sorry, but as I understand it, the Fixes tag should be used
> only when an actual bug is being fixed. I've seen stable
> maintainers getting annoyed when it's used for non-bug issues.

You seem to confuse the Fixes tag with a CC stable tag. A Fixes tag is
used to indicate which commit introduced an issue, while the CC stable
tag is used to flag a commit for backporting (and the fact that autosel
tends to pick up patches with just a Fixes doesn't change this).

It's perfectly fine to fix an issue and use a Fixes tag when doing so
even if the fix itself does not qualify for backporting (for whatever
reason).

> The system works perfectly fine even with the warning printed.
> It might confuse CI systems, but that shouldn't really be our
> concern.

You should not knowingly be introducing new warnings. The Fixes tag I
added showed that this was an issue introduced in 6.15-rc1, and, unless
discovered really late in the cycle, it should be fixed before 6.15 is
out.

Johan