diff mbox series

drm/msm/dpu: signedness bug in dpu_encoder_phys_cmd_tearcheck_config()

Message ID 897779a0-1a1f-4193-9dd3-bc4f87e73e3c@kili.mountain (mailing list archive)
State Not Applicable
Headers show
Series drm/msm/dpu: signedness bug in dpu_encoder_phys_cmd_tearcheck_config() | expand

Commit Message

Dan Carpenter May 22, 2023, 7:48 a.m. UTC
The "vsync_hz" variable is used to store negative error codes so it
needs to be signed for the error checking to work correctly.

Fixes: 25fdd5933e4c ("drm/msm: Add SDM845 DPU support")
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
---
 drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Marijn Suijten May 22, 2023, 8:47 a.m. UTC | #1
On 2023-05-22 10:48:01, Dan Carpenter wrote:
> The "vsync_hz" variable is used to store negative error codes so it
> needs to be signed for the error checking to work correctly.

Nicely spotted, but it it looks like we have more to fix then.  The type
returned by dpu_kms_get_clk_rate() is u64:

- Perhaps the int used in this patch is too small (though 19.2MHz vsync
  clock ought to fit);
- That function should also return a signed number as part of this
  contract.

- Marijn

> 
> Fixes: 25fdd5933e4c ("drm/msm: Add SDM845 DPU support")
> Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> ---
>  drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> index 74470d068622..b29e6d1ba7f6 100644
> --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> @@ -320,7 +320,7 @@ static void dpu_encoder_phys_cmd_tearcheck_config(
>  	struct dpu_hw_tear_check tc_cfg = { 0 };
>  	struct drm_display_mode *mode;
>  	bool tc_enable = true;
> -	u32 vsync_hz;
> +	int vsync_hz;
>  	struct dpu_kms *dpu_kms;
>  
>  	if (!phys_enc->hw_pp) {
> -- 
> 2.39.2
>
Dmitry Baryshkov May 22, 2023, 1:02 p.m. UTC | #2
On Mon, 22 May 2023 at 11:47, Marijn Suijten
<marijn.suijten@somainline.org> wrote:
>
> On 2023-05-22 10:48:01, Dan Carpenter wrote:
> > The "vsync_hz" variable is used to store negative error codes so it
> > needs to be signed for the error checking to work correctly.
>
> Nicely spotted, but it it looks like we have more to fix then.  The type
> returned by dpu_kms_get_clk_rate() is u64:
>
> - Perhaps the int used in this patch is too small (though 19.2MHz vsync
>   clock ought to fit);
> - That function should also return a signed number as part of this
>   contract.

Please also include freedreno@ to the cc list when sending v2.

>
> - Marijn
>
> >
> > Fixes: 25fdd5933e4c ("drm/msm: Add SDM845 DPU support")
> > Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
> > ---
> >  drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> > index 74470d068622..b29e6d1ba7f6 100644
> > --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> > +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
> > @@ -320,7 +320,7 @@ static void dpu_encoder_phys_cmd_tearcheck_config(
> >       struct dpu_hw_tear_check tc_cfg = { 0 };
> >       struct drm_display_mode *mode;
> >       bool tc_enable = true;
> > -     u32 vsync_hz;
> > +     int vsync_hz;
> >       struct dpu_kms *dpu_kms;
> >
> >       if (!phys_enc->hw_pp) {
> > --
> > 2.39.2
> >
Dan Carpenter May 23, 2023, 10:54 a.m. UTC | #3
On Mon, May 22, 2023 at 10:47:50AM +0200, Marijn Suijten wrote:
> On 2023-05-22 10:48:01, Dan Carpenter wrote:
> > The "vsync_hz" variable is used to store negative error codes so it
> > needs to be signed for the error checking to work correctly.
> 
> Nicely spotted, but it it looks like we have more to fix then.  The type
> returned by dpu_kms_get_clk_rate() is u64:
> 

That's a good point.  Although the type of clk_get_rate() is unsigned
long so probably unsigned long is enough.

> - Perhaps the int used in this patch is too small (though 19.2MHz vsync
>   clock ought to fit);
> - That function should also return a signed number as part of this
>   contract.

What about if we just return 0 instead of -EINVAL?  That would match
what clk_get_rate() does.  Also let me change the type to unsigned long.

regards,
dan carpenter
diff mbox series

Patch

diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
index 74470d068622..b29e6d1ba7f6 100644
--- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
+++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_cmd.c
@@ -320,7 +320,7 @@  static void dpu_encoder_phys_cmd_tearcheck_config(
 	struct dpu_hw_tear_check tc_cfg = { 0 };
 	struct drm_display_mode *mode;
 	bool tc_enable = true;
-	u32 vsync_hz;
+	int vsync_hz;
 	struct dpu_kms *dpu_kms;
 
 	if (!phys_enc->hw_pp) {