Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220317140115.541007-1-shraash@google.com> <f1ff7e95-c177-09dd-9393-d0156ce81387@molgen.mpg.de>
In-Reply-To: <f1ff7e95-c177-09dd-9393-d0156ce81387@molgen.mpg.de>
From:   Guenter Roeck <groeck@google.com>
Date:   Fri, 18 Mar 2022 08:44:52 -0700
Message-ID: <CABXOdTecWs5oTfEDyK4onzrzcBnD7u-3j3ikbz4bvo20PEd17w@mail.gmail.com>
Subject: Re: [PATCH] drm/amd/display: Fixed the unused-but-set-variable warning
To:     Paul Menzel <pmenzel@molgen.mpg.de>
Cc:     Aashish Sharma <shraash@google.com>,
        Harry Wentland <harry.wentland@amd.com>,
        Leo Li <sunpeng.li@amd.com>,
        Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>,
        Alex Deucher <alexander.deucher@amd.com>,
        Pan Xinhui <Xinhui.Pan@amd.com>,
        David Airlie <airlied@linux.ie>,
        Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>,
        Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>,
        Jake Wang <haonan.wang2@amd.com>,
        Anson Jacob <Anson.Jacob@amd.com>,
        Guenter Roeck <groeck@chromium.org>,
        linux-kernel <linux-kernel@vger.kernel.org>,
        Maling list - DRI developers 
        <dri-devel@lists.freedesktop.org>,
        amd-gfx list <amd-gfx@lists.freedesktop.org>,
        Daniel Vetter <daniel@ffwll.ch>, Wayne Lin <wayne.lin@amd.com>,
        Anthony Koo <Anthony.Koo@amd.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Hi Paul,

On Thu, Mar 17, 2022 at 10:58 PM Paul Menzel <pmenzel@molgen.mpg.de> wrote:
>
> Dear Aashish,
>
>
> Am 17.03.22 um 15:01 schrieb Aashish Sharma:
>
> Thank you for your patch. If you are going to send a v2, please use
> imperative mood. Maybe:
>

Uuh, sorry, I have to take the blame for that. I am guiding Aashish
through the process of submitting patches upstream, and I completely
missed that.

> drm/amd/display: Fix unused-but-set-variable warning
>
>
> > Fixed this kernel test robot warning:
>
> Maybe:
>
> Fix the kernel test robot warning below:
>
> > drivers/gpu/drm/amd/amdgpu/../display/dmub/inc/dmub_cmd.h:2893:12:
> > warning: variable 'temp' set but not used [-Wunused-but-set-variable]
> >
> > Replaced the assignment to the unused temp variable with READ_ONCE()
> > macro to flush the writes.
>
> Replace =E2=80=A6
>
> Excuse my ignorance regarding `READ_ONCE()`, but is that the reason you
> remove the volatile qualifier?
>

It is not the reason to remove volatile, it is to enable removing it.
I had suggested using it, following its description " ... Ensuring
that the compiler does not fold, spindle, or otherwise  * mutilate
accesses ... ", to avoid the use of volatile, and to make it obvious
from the code that the read is intentional. My apologies if that is
the wrong approach. A simpler solution would be to just remove the
temp variable if that is preferred.

> Some robots ask in their report to add a Found-by tag. If so, please add
> one.

I think that should be "Reported-by", or more specifically

Reported-by: kernel test robot <lkp@intel.com>

>
> > Signed-off-by: Aashish Sharma <shraash@google.com>
> > ---
> >   drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 5 ++---
> >   1 file changed, 2 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/=
gpu/drm/amd/display/dmub/inc/dmub_cmd.h
> > index 873ecd04e01d..b7981a781701 100644
> > --- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
> > +++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
> > @@ -2913,13 +2913,12 @@ static inline void dmub_rb_flush_pending(const =
struct dmub_rb *rb)
> >       uint32_t wptr =3D rb->wrpt;
> >
> >       while (rptr !=3D wptr) {
> > -             uint64_t volatile *data =3D (uint64_t volatile *)((uint8_=
t *)(rb->base_address) + rptr);
> > +             uint64_t *data =3D (uint64_t volatile *)((uint8_t *)(rb->=
base_address) + rptr);
> >               //uint64_t volatile *p =3D (uint64_t volatile *)data;
> > -             uint64_t temp;
> >               uint8_t i;
> >
> >               for (i =3D 0; i < DMUB_RB_CMD_SIZE / sizeof(uint64_t); i+=
+)
> > -                     temp =3D *data++;
> > +                     (void)READ_ONCE(*data++);
>
> Did you verify, that the generated code is the same now, or what the
> differences are. If it=E2=80=99s different from before, you should docume=
nt in
> the commit message, that it=E2=80=99s wanted, as otherwise, it=E2=80=99s =
an invasive
> change just to fix a warning.
>

The generated code is exactly the same on x86.

Thanks,
Guenter

> >               rptr +=3D DMUB_RB_CMD_SIZE;
> >               if (rptr >=3D rb->capacity)
>
>
> Kind regards,
>
> Paul