Re: [PATCH] gpu: host1x: fix uninitialized variable use

[Nathan Chancellor]

Hi Thierry, Daniel, and David,

On Fri, Jan 27, 2023 at 11:14:00PM +0100, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@xxxxxxxx>
> 
> The error handling for platform_get_irq() failing no longer
> works after a recent change, clang now points this out with
> a warning:
> 
> drivers/gpu/host1x/dev.c:520:6: error: variable 'syncpt_irq' is uninitialized when used here [-Werror,-Wuninitialized]
>         if (syncpt_irq < 0)
>             ^~~~~~~~~~
> 
> Fix this by removing the variable and checking the correct
> error status.
> 
> Fixes: 625d4ffb438c ("gpu: host1x: Rewrite syncpoint interrupt handling")
> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> ---
>  drivers/gpu/host1x/dev.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c
> index 4872d183d860..aae2efeef503 100644
> --- a/drivers/gpu/host1x/dev.c
> +++ b/drivers/gpu/host1x/dev.c
> @@ -487,7 +487,6 @@ static int host1x_get_resets(struct host1x *host)
>  static int host1x_probe(struct platform_device *pdev)
>  {
>  	struct host1x *host;
> -	int syncpt_irq;
>  	int err;
>  
>  	host = devm_kzalloc(&pdev->dev, sizeof(*host), GFP_KERNEL);
> @@ -517,8 +516,8 @@ static int host1x_probe(struct platform_device *pdev)
>  	}
>  
>  	host->syncpt_irq = platform_get_irq(pdev, 0);
> -	if (syncpt_irq < 0)
> -		return syncpt_irq;
> +	if (host->syncpt_irq < 0)
> +		return host->syncpt_irq;
>  
>  	mutex_init(&host->devices_lock);
>  	INIT_LIST_HEAD(&host->devices);
> -- 
> 2.39.0
> 

Apologies if this has been reported already or has a solution in
progress but mainline is now broken because this change got separated
from the change it is fixing:

https://github.com/ClangBuiltLinux/continuous-integration2/actions/runs/4249931209/jobs/7391912774
https://storage.tuxsuite.com/public/clangbuiltlinux/continuous-integration2/builds/2M7y9HpiXB13qiC2mkHMyeZOcLW/build.log

I see this change sitting in the drm-tegra tree [1], which is getting
merged into -next, so it is fixed there, which is why we did not notice
any issues until the drm-next tree was merged into mainline. Can this be
fast tracked to Linus to unbreak clang builds with -Werror?

[1]: https://gitlab.freedesktop.org/drm/tegra/-/commit/b9930311641cf2ed905a84aabe27e8f3868aee4a

Cheers,
Nathan