Re: autofs: use __kernel_write() for the autofs pipe writing causes regression in -next was Re: 5.9.0-next-20201015: autofs oops in update-binfmts

On Sat 2020-10-17 17:13:20, Linus Torvalds wrote:
> On Sat, Oct 17, 2020 at 12:48 PM Pavel Machek <[email protected]> wrote:
> >
> > But: you are the last one to sign it off, so I assume committed it to
> > git, and you are the one to talk to about fixing it.
> 
> The thing is, the commit you point to - and the one I signed off on - is fine.
> 
> The buggy one is in linux-next, which breaks that whole "NULL means no
> position" thing.
> 
> IOW, the real bug is in commit 4d03e3cc5982 ("fs: don't allow kernel
> reads and writes without iter ops"), which does that bogus
> 
>         kiocb.ki_pos = *pos;
> 
> and no, I never signed off on that.
> 
> Get it? Stop confusing people. This bug does not exist in mainline,
> and never will. Because I'm not pulling that buggy commit.

And I guess that's a good thing. It is now fixed in -next, too. Sorry
for the noise.

Best regards,
								Pavel

-- 
http://www.livejournal.com/~pavelmachek

-----BEGIN PGP SIGNATURE-----

iF0EABECAB0WIQRPfPO7r0eAhk010v0w5/Bqldv68gUCX5aLDAAKCRAw5/Bqldv6
8l3iAKDA9QgXwAVJMYPECRYiZvls/4K58wCfQVKE+alumoUT3rMzunV+vN1L8Jk=
=r+i9
-----END PGP SIGNATURE-----