Re: [PATCH] lib/string: sysfs_streq works case insensitively
From: Gioh Kim
Date: Wed Apr 28 2021 - 04:32:17 EST
On Wed, Apr 28, 2021 at 9:47 AM Rasmus Villemoes
<linux@xxxxxxxxxxxxxxxxxx> wrote:
>
> On 28/04/2021 09.31, Gioh Kim wrote:
> > On Wed, Apr 28, 2021 at 8:42 AM Andy Shevchenko
> > <andy.shevchenko@xxxxxxxxx> wrote:
> >>
>
> >>
> >> Are you sure it’s good change? Sysfs is used for an ABI and you are opening a can of worms. From me NAK to this change without a very good background description that tells why it is safe to do.
> >
> > https://www.spinics.net/lists/kernel/msg3898123.html
> > My initial idea was making a new function: sysfs_streqcase.
> > Andrew and Greg suggested making sysfs_streq to be case-insensitive.
> > I would like to have a discussion about it.
>
> 1. That information should be in the commit log, not some random
> babbling about case sensitivity of file systems.
I don't think it is a good idea to write who suggested the concept of the patch.
>
> 2. So as Andy says, this is changing ABI for a whole lot of users in one
> go. While it's _probably_ true that nobody would care (because it just
> ends up accepting more strings, not fewer), your motivation seems to be
> to replace uses of strncasecmp() to prevent "disableGARBAGE@#$@#@" to be
> accepted as equivalent to "disable". I.e., those potential new users of
> sysfs_streq() would have their ABI changed towards being less
> permissive. That's a bigger change, with a higher chance of breaking
> something. Do you even know if the maintainers of those drivers would
> accept a switch to a case-insensitive sysfs_streq()?
I don't know what all maintainers would think about the patch.
I thought sending a patch to the mailing list is asking for it, isn't it?
I am asking with this email.
And Andrew and Greg are maintainers.
I am sending this patch because I accepted their feedback.
I don't know other maintainers personally,
so I cannot contact them to ask what they think about my idea before
sending this email.
>
> Sorry, I really think you need a lot stronger motivation for introducing
> either this change or a sysfs_strcaseeq().
I found out a problem of sysfs_streq when I tried to use it for
modules I am working on (RTRS/RNBD).
Then I thought it would be better if there is something like sysfs_streqcase.
That's why I sent the path.
If nobody needs the change, it is ok for me to ignore the patch.
Maybe I misunderstand when I can send a patch to Linux kernel community.
But the only motivation of me is sharing my idea just in case there is
someone who also needs it.
I am sorry if I misunderstand how Linux kernel community works
but I thought I could suggest something good to me.
>
> Rasmus