On 2/25/20 12:23 AM, Willy Tarreau wrote:
> As indicated in commit 2e90ca6 ("floppy: check FDC index for errors
> before assigning it") there are some surprising effects in the floppy
> driver due to some macros referencing global or local variables while
> at first glance being inoffensive.
>
> This patchset aims at removing these macros and replacing all of their
> occurrences by the equivalent code. Most of the work was done under
> Coccinelle's assistance, and it was verified that the resulting binary
> code is exactly the same as the original one.
>
> The aim is not to make the driver prettier, as Linus mentioned it's
> already not pretty. It only aims at making potential bugs more visible,
> given almost all latest changes to this driver were fixes for out-of-
> bounds and similar bugs.
>
> As a side effect, some lines got longer, causing checkpatch to complain
> a bit, but I preferred to let it complain as I didn't want to break them
> apart as I'm already seeing the trap of going too far here.
>
> The patches are broken by macro (or sets of macros when relevant) so
> that each of them remains reviewable.
>
> I can possibly go a bit further in the cleanup but I haven't used
> floppies for a few years now and am not interested in doing too much
> on this driver by lack of use cases.
For patches 1-10.
[x] eye checked the changes
[x] bloat-o-meter and .s diff show no real changes
[x] tested that kernel builds after every patch
[x] floppy targeted fuzzing with kasan+ubsan reveals no *new* issues
(required mainly to test the previous patch)
If Linus has no objections (regarding his review) I would prefer to
accept 1-10 patches rather to resend them again. They seems complete
to me as the first step.
I've placed the patches here:
https://github.com/evdenis/linux-floppy/commits/floppy-next
Thanks,
Denis
