Re: [PATCH] mtd: parsers: remove reference to config MTD_NAND_TMIO

[Miquel Raynal]

Hi Lukas,

lukas.bulwahn@xxxxxxxxx wrote on Tue, 7 Mar 2023 11:58:07 +0100:

> On Tue, Mar 7, 2023 at 11:38 AM Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote:
> >
> > Hi Arnd,
> >
> > arnd@xxxxxxxx wrote on Tue, 07 Mar 2023 11:26:48 +0100:
> >  
> > > On Tue, Mar 7, 2023, at 10:03, Miquel Raynal wrote:  
> > > > Hi Lukas,
> > > >
> > > > lukas.bulwahn@xxxxxxxxx wrote on Tue,  7 Mar 2023 08:40:38 +0100:
> > > >  
> > > >> Commit 568494db6809 ("mtd: remove tmio_nand driver") removes the config
> > > >> MTD_NAND_TMIO and its corresponding driver.
> > > >>
> > > >> Remove the reference in MTD_SHARPSL_PARTS to that removed config.
> > > >>
> > > >> Signed-off-by: Lukas Bulwahn <lukas.bulwahn@xxxxxxxxx>
> > > >> ---
> > > >> Arnd, please ack.
> > > >> Miquel, please pick this minor non-urgent patch on top of the commit above.  
> > > >
> > > > Actually I guess the SHARPSL driver is not selectable right now, so
> > > > this should be sent as part of my next fixes PR.  
> > >
> > > I don't see why not, it just depends on 'ARCH_PXA||COMPILE_TEST' and
> > > should work fine with CONFIG_MACH_SPITZ PDAs.  
> >
> > Sorry, I overlooked the diff, indeed there is nothing broken, so I'll
> > queue it to nand/next. I thought MTD_SHARPSL_PARTS was not selectable
> > anymore without COMPILE_TEST, which would have been problematic.
> >
> > Fixes tag still welcome though :)
> >  
> 
> Arnd's commit is not broken. It does what Arnd writes in his commit
> message what it should do.
>
> And this patch is just a clean-up that removes references, but it is
> just "stylistic" and reduces the complexity of Kconfig dependency
> definitions for non-existing config symbols. I do not see that this
> patch fixes Arnd's patch. It is just a clean-up with a reference to
> Arnd's patch to understand why this clean up can be done now.
> 
> If you REALLY want the Fixes: tag, I can sure add it. But, I do not
> claim that I am fixing anything here; nothing was broken in the first
> place. The reference to the commit of interest is in the commit
> message, and anyone can follow or extract the information if they are
> interested.
>
> So, please keep this patch in the queue for nand/next.

Wow, so much arguing just for a Fixes tag.

It is not broken, I know, but it is somehow incomplete as it creates a
stalled Kconfig symbol. And you come-in and fix the situation by
dropping the remaining symbol. So for me it's a fix, no matter how
deeply bogus (or not) the original commit was. It's not an insult, it's
a tag that allows easy parsing. Anyway, that's meaningless. I'll take
it like that.

Thanks,
Miquèl