linux-kernel - RE: [PATCH] block: fix Werror=format with GCC 13

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 26 Oct 2022 08:16:53 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Jiri Slaby' <jirislaby@...nel.org>,
        Martin Liška <mliska@...e.cz>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
CC:     "linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
        "axboe@...nel.dk" <axboe@...nel.dk>
Subject: RE: [PATCH] block: fix Werror=format with GCC 13

From: Jiri Slaby
> Sent: 26 October 2022 08:18
> 
> On 24. 10. 22, 21:01, Martin Liška wrote:
> > Starting with GCC 13, since
> > [g3b3083a598ca3f4b] c: C2x enums wider than int [PR36113]
> >
> > GCC promotes enum values with larger than integer types to a wider type.
> > In case of the anonymous enum type in blk-iocost.c it is:
> >
> > enum {
> > 	MILLION			= 1000000,
> > ...
> >
> > 	WEIGHT_ONE		= 1 << 16,
> > ...
> > 	VTIME_PER_SEC_SHIFT	= 37,
> > 	VTIME_PER_SEC		= 1LLU << VTIME_PER_SEC_SHIFT,
> > ...
> >
> > as seen VTIME_PER_SEC cannot fit into 32-bits (int type), thus one needs
> > to use 'long unsigned int' in the format string.
> >
> > It fixes then the following 2 warnings:
> >
> > block/blk-iocost.c: In function ‘ioc_weight_prfill’:
> > block/blk-iocost.c:3035:37: error: format ‘%u’ expects argument of type ‘unsigned int’, but argument
> 4 has type ‘long unsigned int’ [-Werror=format=]
> >   3035 |                 seq_printf(sf, "%s %u\n", dname, iocg->cfg_weight / WEIGHT_ONE);
> >        |                                    ~^            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >        |                                     |                             |
> >        |                                     unsigned int                  long unsigned int
> >        |                                    %lu
> > block/blk-iocost.c: In function ‘ioc_weight_show’:
> > block/blk-iocost.c:3045:34: error: format ‘%u’ expects argument of type ‘unsigned int’, but argument
> 3 has type ‘long unsigned int’ [-Werror=format=]
> >   3045 |         seq_printf(sf, "default %u\n", iocc->dfl_weight / WEIGHT_ONE);
> >        |                                 ~^     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >        |                                  |                      |
> >        |                                  unsigned int           long unsigned int
> >        |                                 %lu
> 
> But introduces two with gcc-12 ;):
>  > block/blk-iocost.c: In function ‘ioc_weight_prfill’:
>  > block/blk-iocost.c:3037:38: error: format ‘%lu’ expects argument of
> type ‘long unsigned int’, but argument 4 has type ‘u32’ {aka ‘unsigned
> int’} [-Werror=format=]
>  >  3037 |                 seq_printf(sf, "%s %lu\n", dname,
> iocg->cfg_weight / WEIGHT_ONE);
>  >       |                                    ~~^
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>  >       |                                      |
>       |
>  >       |                                      long unsigned int
>       u32 {aka unsigned int}
>  >       |                                    %u
> 
> 
> Note that:
> 1) the specs says enum behaves as int, or uint in some cases
> 2) iocc->dfl_weight is u32, i.e. uint
>     WEIGHT_ONE is 1 << 16, i.e. int
>     so the promotion should be to s32/int. Or not?
> 
> I think gcc-13 is wrong -- incosistent with gcc-12 at least.

The presence of VTIME_PER_SEC in the enum forces the enum
to 64bits (this must be a gcc extension).

The change in gcc 13 seems to be that the types of all the
enum values are now (probably correctly) that of the enum.

So WEIGHT_ONE changes from 'unsigned int' to 'unsigned long'.

See: https://godbolt.org/z/6K4PoK9sv

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)