Re: [PATCH v2] drm/amdgpu: Remove unnecessary variable shadow in gfx_v9_0_rlcg_wreg

[Nick Desaulniers]

On Wed, Mar 18, 2020 at 2:05 PM Nathan Chancellor
<natechancellor@xxxxxxxxx> wrote:
>
> clang warns:
>
> drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:754:6: warning: variable 'shadow'
> is used uninitialized whenever 'if' condition is
> false [-Wsometimes-uninitialized]
>         if (offset == grbm_cntl || offset == grbm_idx)
>             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:757:6: note: uninitialized use
> occurs here
>         if (shadow) {
>             ^~~~~~
> drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:754:2: note: remove the 'if' if
> its condition is always true
>         if (offset == grbm_cntl || offset == grbm_idx)
>         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c:738:13: note: initialize the
> variable 'shadow' to silence this warning
>         bool shadow;
>                    ^
>                     = 0
> 1 warning generated.
>
> shadow is only assigned in one condition and used as the condition for
> another if statement; combine the two if statements and remove shadow
> to make the code cleaner and resolve this warning.
>
> Fixes: 2e0cc4d48b91 ("drm/amdgpu: revise RLCG access path")
> Link: https://github.com/ClangBuiltLinux/linux/issues/936
> Suggested-by: Joe Perches <joe@xxxxxxxxxxx>
> Reviewed-by: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> Signed-off-by: Nathan Chancellor <natechancellor@xxxxxxxxx>
> ---
>
> v1 -> v2:
>
> * Remove shadow altogether, as suggested by Joe Perches.
> * Add Nick's Reviewed-by, as I assume it still stands.

yep, thanks

>
>  drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> index 7bc2486167e7..496b9edca3c3 100644
> --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
> @@ -735,7 +735,6 @@ void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
>         static void *spare_int;
>         static uint32_t grbm_cntl;
>         static uint32_t grbm_idx;
> -       bool shadow;
>
>         scratch_reg0 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG0_BASE_IDX] + mmSCRATCH_REG0)*4;
>         scratch_reg1 = adev->rmmio + (adev->reg_offset[GC_HWIP][0][mmSCRATCH_REG1_BASE_IDX] + mmSCRATCH_REG1)*4;
> @@ -751,10 +750,7 @@ void gfx_v9_0_rlcg_wreg(struct amdgpu_device *adev, u32 offset, u32 v)
>                 return;
>         }
>
> -       if (offset == grbm_cntl || offset == grbm_idx)
> -               shadow = true;
> -
> -       if (shadow) {
> +       if (offset == grbm_cntl || offset == grbm_idx) {
>                 if (offset  == grbm_cntl)
>                         writel(v, scratch_reg2);
>                 else if (offset == grbm_idx)
> --
> 2.26.0.rc1
>


-- 
Thanks,
~Nick Desaulniers