Re: [PATCH 3/3] vfs: return -EOVERFLOW in generic_remap_checks() when overflow check fails

[Julian Sun <sunjunchao2870@gmail.com>]

Christoph Hellwig <hch@infradead.org> 于2024年9月20日周五 23:07写道：
>
> On Fri, Sep 20, 2024 at 08:02:13AM -0700, Darrick J. Wong wrote:
> > > Which isn't exactly the integer overflow case described here :)
> >
> > Hm?  This patch is touching the error code you get for failing alignment
> > checks, not the one you get for failing check_add_overflow.  EOVERFLOW
> > seems like an odd return code for unaligned arguments.  Though you're
> > right that EINVAL is verrry vague.
>
> I misread the patch (or rather mostly read the description).  Yes,
> -EOVERFLOW is rather odd here.  And generic_copy_file_checks doesn't
> even have alignment checks, so the message is wrong as well.  I'll
> wait for Jun what the intention was here - maybe the diff got
> misapplied and this was supposed to be applied to an  overflow
> check that returns -EINVAL?

Yeah... The patch was originally intended for overflow check and
sourced from [1], differs from its description. After applying it to
the latest kernel version, there were no warnings or errors, but I
suspect there may be an issue with the git apply process. I'll fix it
in the patch v2, thanks.

[1]: https://lore.kernel.org/linux-fsdevel/20240906033202.1252195-1-sunjunchao2870@gmail.com/
>

Thanks,
-- 
Julian Sun <sunjunchao2870@gmail.com>