RE: [PATCH 01/11] mmc: core: Use BIT() macro

RE: [PATCH 01/11] mmc: core: Use BIT() macro

[Christian Loehle]

[-- Attachment #1: Type: text/plain, Size: 2417 bytes --]



> -----Original Message-----
> From: Ulf Hansson <ulf.hansson@linaro.org>
> Sent: Mittwoch, 21. Juni 2023 11:19
> To: Marek Vasut <marex@denx.de>
> Cc: linux-mmc@vger.kernel.org; Adrian Hunter <adrian.hunter@intel.com>;
> Avri Altman <avri.altman@wdc.com>; Bo Liu <liubo03@inspur.com>; Deren
> Wu <deren.wu@mediatek.com>; Philipp Zabel <p.zabel@pengutronix.de>;
> Pierre Ossman <pierre@ossman.eu>; Russell King <linux@armlinux.org.uk>;
> Yang Yingliang <yangyingliang@huawei.com>
> Subject: Re: [PATCH 01/11] mmc: core: Use BIT() macro
> 
> CAUTION: this mail comes from external!/ACHTUNG: Diese Mail kommt von
> extern!
> 
> On Wed, 21 Jun 2023 at 04:36, Marek Vasut <marex@denx.de> wrote:
> >
> > On 6/20/23 13:15, Ulf Hansson wrote:
> > > On Tue, 20 Jun 2023 at 12:47, Marek Vasut <marex@denx.de> wrote:
> > >>
> > >> Use the BIT(n) macro instead of (1<<n), no functional change.
> > >> Regex 's@(1 \?<< \?\([0-9A-Z_]\+\))@BIT(\1)' .
> > >>
> > >> Signed-off-by: Marek Vasut <marex@denx.de>
> > >
> > > I don't think the benefit of this change is worth it. For example,
> > > it's quite useful to run a git blame to see the history of what has
> > > happened.
> >
> > Understood.
> >
> > git blame does allow you to specify either --since or revision range though.
> 
> Yes, but I think you get my point.
> 
> >
> > > So, sorry, but I am not going to pick this up - or any other similar
> > > changes, at least for the core layer.
> >
> > Is this a policy of the mmc subsystem to reject all code clean ups then ?
> 
> Of course it isn't, I regularly pick up clean ups.
> 
> My point here is that the clean-up should make the code better, in some
> way. I don't think converting to the BIT macro helps in this regard. It may be
> preferred to use the BIT macro by some and by others not.

FWIW I agree with Uffe here.
For host/ files, which are mostly written by a handful each, it's still a nuisance. (One could argue that they are often git blamed by people not familiar with mmc subsystem, thus giving off the wrong picture).
For much of the core code you already have to go many revisions back, I'm grateful for each I don't have to.
Something like the mq rework would have been a good moment to do these minor nitpicks, if something like that ever happens again, but even then I  would prefer just one commit including everything.


Regards,
Christian

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 10302 bytes --]