[PATCH] drm/nouveau/bios: fix a bit shift error introduced by 457e77b

[PATCH] drm/nouveau/bios: fix a bit shift error introduced by 457e77b

[Sergei Antonov]

Commit 457e77b26428ab4a24998eecfb99f27fa4195397 added two checks applied to a
value received from nv_rd32(bios, 0x619f04). But after this new piece of code
is executed, the addr local variable does not hold the same value it used to
hold before the commit. Here is what is was assigned in the original code:
	(u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8
in the committed code it ends up with this value:
	(u64)(nv_rd32(bios, 0x619f04) >> 8) << 8
These expressions are obviously not equivalent.

My Nvidia video card does not show anything on the display when I boot a
kernel containing this commit.

The patch fixes the code so that the new checks are still done, but the
side effect of an incorrect addr value is gone.

Cc: Ben Skeggs <bskeggs-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: Dave Airlie <airlied-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>
Signed-off-by: Sergei Antonov <saproj-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
---
 drivers/gpu/drm/nouveau/core/subdev/bios/base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
index e9df94f..291adb6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
@@ -109,7 +109,7 @@ nouveau_bios_shadow_pramin(struct nouveau_bios *bios)
 			return;
 		}
 
-		addr = (u64)(addr >> 8) << 8;
+		addr = (u64)(addr & 0xffffff00) << 8;
 		if (!addr) {
 			addr  = (u64)nv_rd32(bios, 0x001700) << 16;
 			addr += 0xf0000;
-- 
1.9.0

Re: [PATCH] drm/nouveau/bios: fix a bit shift error introduced by 457e77b

[Sergei Antonov]

On 15 April 2014 23:18, Sergei Antonov <saproj-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> Commit 457e77b26428ab4a24998eecfb99f27fa4195397 added two checks applied to a
> value received from nv_rd32(bios, 0x619f04). But after this new piece of code
> is executed, the addr local variable does not hold the same value it used to
> hold before the commit. Here is what is was assigned in the original code:
>         (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8
> in the committed code it ends up with this value:
>         (u64)(nv_rd32(bios, 0x619f04) >> 8) << 8
> These expressions are obviously not equivalent.
>
> My Nvidia video card does not show anything on the display when I boot a
> kernel containing this commit.
>
> The patch fixes the code so that the new checks are still done, but the
> side effect of an incorrect addr value is gone.
>
> Cc: Ben Skeggs <bskeggs-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> Cc: Dave Airlie <airlied-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> Cc: Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>
> Signed-off-by: Sergei Antonov <saproj-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> ---
>  drivers/gpu/drm/nouveau/core/subdev/bios/base.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
> index e9df94f..291adb6 100644
> --- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
> +++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
> @@ -109,7 +109,7 @@ nouveau_bios_shadow_pramin(struct nouveau_bios *bios)
>                         return;
>                 }
>
> -               addr = (u64)(addr >> 8) << 8;
> +               addr = (u64)(addr & 0xffffff00) << 8;

I just noticed that "(u64)" is redundant here. Can it be removed in
the patch without resubmitting, Andrew?

>                 if (!addr) {
>                         addr  = (u64)nv_rd32(bios, 0x001700) << 16;
>                         addr += 0xf0000;
> --
> 1.9.0
>

Re: [PATCH] drm/nouveau/bios: fix a bit shift error introduced by 457e77b

[Emil Velikov]

On 15/04/14 22:49, Sergei Antonov wrote:
> On 15 April 2014 23:18, Sergei Antonov <saproj-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> Commit 457e77b26428ab4a24998eecfb99f27fa4195397 added two checks applied to a
>> value received from nv_rd32(bios, 0x619f04). But after this new piece of code
>> is executed, the addr local variable does not hold the same value it used to
>> hold before the commit. Here is what is was assigned in the original code:
>>         (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8
>> in the committed code it ends up with this value:
>>         (u64)(nv_rd32(bios, 0x619f04) >> 8) << 8
>> These expressions are obviously not equivalent.
>>
>> My Nvidia video card does not show anything on the display when I boot a
>> kernel containing this commit.
>>
>> The patch fixes the code so that the new checks are still done, but the
>> side effect of an incorrect addr value is gone.
>>
Hi Sergei

A similar patch [1] has been sent a few days prior to yours, and I'm assuming
it will be queued/merged shortly. Just letting you know.

Cheers
-Emil

[1] http://lists.freedesktop.org/archives/dri-devel/2014-April/057318.html

Re: [PATCH] drm/nouveau/bios: fix a bit shift error introduced by 457e77b

[Sergei Antonov]

On 19 April 2014 22:14, Emil Velikov <emil.l.velikov-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On 15/04/14 22:49, Sergei Antonov wrote:
>> On 15 April 2014 23:18, Sergei Antonov <saproj-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>>> Commit 457e77b26428ab4a24998eecfb99f27fa4195397 added two checks applied to a
>>> value received from nv_rd32(bios, 0x619f04). But after this new piece of code
>>> is executed, the addr local variable does not hold the same value it used to
>>> hold before the commit. Here is what is was assigned in the original code:
>>>         (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8
>>> in the committed code it ends up with this value:
>>>         (u64)(nv_rd32(bios, 0x619f04) >> 8) << 8
>>> These expressions are obviously not equivalent.
>>>
>>> My Nvidia video card does not show anything on the display when I boot a
>>> kernel containing this commit.
>>>
>>> The patch fixes the code so that the new checks are still done, but the
>>> side effect of an incorrect addr value is gone.
>>>
> Hi Sergei
>
> A similar patch [1] has been sent a few days prior to yours, and I'm assuming
> it will be queued/merged shortly. Just letting you know.
>
> Cheers
> -Emil
>
> [1] http://lists.freedesktop.org/archives/dri-devel/2014-April/057318.html

Thanks! I didn't know about this patch. Sorry for accidentally
creating a race condition.
My patch has been merged today (included in "drm fixes" from Dave Airlie).