* mlx5: unitialized warning
@ 2021-12-17 18:05 Stephen Hemminger
2021-12-17 19:01 ` Dmitry Kozlyuk
0 siblings, 1 reply; 5+ messages in thread
From: Stephen Hemminger @ 2021-12-17 18:05 UTC (permalink / raw)
To: dev
Building current main branch with gcc (Debian 10.2.1-6) 10.2.1 20210110
[1843/3131] Compiling C object drivers/libtmp_rte_net_mlx5.a.p/net_mlx5_mlx5_flow_meter.c.o
../drivers/net/mlx5/mlx5_flow_meter.c: In function ‘mlx5_flow_meter_create’:
../drivers/net/mlx5/mlx5_flow_meter.c:1170:33: warning: ‘legacy_fm’ may be used uninitialized in this function [-Wmaybe-uninitialized]
1170 | struct mlx5_legacy_flow_meter *legacy_fm;
| ^~~~~~~~~
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: mlx5: unitialized warning
2021-12-17 18:05 mlx5: unitialized warning Stephen Hemminger
@ 2021-12-17 19:01 ` Dmitry Kozlyuk
2021-12-17 19:19 ` Stephen Hemminger
0 siblings, 1 reply; 5+ messages in thread
From: Dmitry Kozlyuk @ 2021-12-17 19:01 UTC (permalink / raw)
To: Stephen Hemminger, dev
> Building current main branch with gcc (Debian 10.2.1-6) 10.2.1 20210110
>
>
> [1843/3131] Compiling C object
> drivers/libtmp_rte_net_mlx5.a.p/net_mlx5_mlx5_flow_meter.c.o
> ../drivers/net/mlx5/mlx5_flow_meter.c: In function
> ‘mlx5_flow_meter_create’:
> ../drivers/net/mlx5/mlx5_flow_meter.c:1170:33: warning: ‘legacy_fm’ may be
> used uninitialized in this function [-Wmaybe-uninitialized]
> 1170 | struct mlx5_legacy_flow_meter *legacy_fm;
> | ^~~~~~~~~
Hi Stephen,
I don't see this warning with gcc 10.3.0 (Ubuntu 10.3.0-1ubuntu1~20.04).
Firstly, I believe it's a false positive:
legacy_fm is initialized and used only when priv->sh->meter_aso_en == false:
1237 if (priv->sh->meter_aso_en) {
...
1245 } else {
legacy_fm = ...
if (legacy_fm == NULL)
return ...;
...
}
...
1276 if (!priv->sh->meter_aso_en)
TAILQ_INSERT_TAIL(fms, legacy_fm, next);
However, even if it's fixed in GCC between 10.2.1 and 10.3.0,
I've seen it with x86_64-w64-mingw32-gcc (GCC) 9.3-win32 20200320
while working on enabling drivers/*/mlx5 build with GCC,
so a workaround will probably still be needed.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: mlx5: unitialized warning
2021-12-17 19:01 ` Dmitry Kozlyuk
@ 2021-12-17 19:19 ` Stephen Hemminger
2021-12-17 19:28 ` Dmitry Kozlyuk
0 siblings, 1 reply; 5+ messages in thread
From: Stephen Hemminger @ 2021-12-17 19:19 UTC (permalink / raw)
To: Dmitry Kozlyuk; +Cc: dev
On Fri, 17 Dec 2021 19:01:31 +0000
Dmitry Kozlyuk <dkozlyuk@nvidia.com> wrote:
> > Building current main branch with gcc (Debian 10.2.1-6) 10.2.1 20210110
> >
> >
> > [1843/3131] Compiling C object
> > drivers/libtmp_rte_net_mlx5.a.p/net_mlx5_mlx5_flow_meter.c.o
> > ../drivers/net/mlx5/mlx5_flow_meter.c: In function
> > ‘mlx5_flow_meter_create’:
> > ../drivers/net/mlx5/mlx5_flow_meter.c:1170:33: warning: ‘legacy_fm’ may be
> > used uninitialized in this function [-Wmaybe-uninitialized]
> > 1170 | struct mlx5_legacy_flow_meter *legacy_fm;
> > | ^~~~~~~~~
>
> Hi Stephen,
>
> I don't see this warning with gcc 10.3.0 (Ubuntu 10.3.0-1ubuntu1~20.04).
> Firstly, I believe it's a false positive:
> legacy_fm is initialized and used only when priv->sh->meter_aso_en == false:
>
> 1237 if (priv->sh->meter_aso_en) {
> ...
> 1245 } else {
> legacy_fm = ...
> if (legacy_fm == NULL)
> return ...;
> ...
> }
> ...
> 1276 if (!priv->sh->meter_aso_en)
> TAILQ_INSERT_TAIL(fms, legacy_fm, next);
>
> However, even if it's fixed in GCC between 10.2.1 and 10.3.0,
> I've seen it with x86_64-w64-mingw32-gcc (GCC) 9.3-win32 20200320
> while working on enabling drivers/*/mlx5 build with GCC,
> so a workaround will probably still be needed.
It maybe because I was testing with thread santizer enabled and compiler
was doing different optimizations.
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: mlx5: unitialized warning
2021-12-17 19:19 ` Stephen Hemminger
@ 2021-12-17 19:28 ` Dmitry Kozlyuk
2021-12-17 20:17 ` Stephen Hemminger
0 siblings, 1 reply; 5+ messages in thread
From: Dmitry Kozlyuk @ 2021-12-17 19:28 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: dev
> > I don't see this warning with gcc 10.3.0 (Ubuntu 10.3.0-1ubuntu1~20.04).
[...]
>
> It maybe because I was testing with thread santizer enabled and compiler
> was doing different optimizations.
Reproduced with -Dc_args=-fsanitize=thread, thanks!
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: mlx5: unitialized warning
2021-12-17 19:28 ` Dmitry Kozlyuk
@ 2021-12-17 20:17 ` Stephen Hemminger
0 siblings, 0 replies; 5+ messages in thread
From: Stephen Hemminger @ 2021-12-17 20:17 UTC (permalink / raw)
To: Dmitry Kozlyuk; +Cc: dev
On Fri, 17 Dec 2021 19:28:26 +0000
Dmitry Kozlyuk <dkozlyuk@nvidia.com> wrote:
> > > I don't see this warning with gcc 10.3.0 (Ubuntu 10.3.0-1ubuntu1~20.04).
> [...]
> >
> > It maybe because I was testing with thread santizer enabled and compiler
> > was doing different optimizations.
>
> Reproduced with -Dc_args=-fsanitize=thread, thanks!
I was using
-Db_sanitize=thread
which is the meson way to do it
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2021-12-17 20:17 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-17 18:05 mlx5: unitialized warning Stephen Hemminger
2021-12-17 19:01 ` Dmitry Kozlyuk
2021-12-17 19:19 ` Stephen Hemminger
2021-12-17 19:28 ` Dmitry Kozlyuk
2021-12-17 20:17 ` Stephen Hemminger
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).