Re: [PATCH 1/1] drm: xlnx: zynqmp: Stop the loop at lowest link rate without check

From: Daniel Vetter <daniel@ffwll.ch>
To: Hyun Kwon <hyun.kwon@xilinx.com>
Cc: Dan Carpenter <dan.carpenter@oracle.com>,
	Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
	dri-devel <dri-devel@lists.freedesktop.org>
Subject: Re: [PATCH 1/1] drm: xlnx: zynqmp: Stop the loop at lowest link rate without check
Date: Wed, 29 Jul 2020 23:34:16 +0200	[thread overview]
Message-ID: <CAKMK7uGBrACy_sP7h35W=6=R94OG3in3E7JiTMp5Kk-hcfkAcw@mail.gmail.com> (raw)
In-Reply-To: <1596046674-29155-1-git-send-email-hyun.kwon@xilinx.com>

On Wed, Jul 29, 2020 at 8:21 PM Hyun Kwon <hyun.kwon@xilinx.com> wrote:
>
> The loop should exit at the lowest link rate, so break the loop
> at the lowest link rate without check. The check is always true
> because lowest link rate is smaller than current one and maximum
> of current display. Otherwise, it may be seen as the loop can
> potentially result in negative array offset.
>
> The patch d76271d22694: "drm: xlnx: DRM/KMS driver for Xilinx ZynqMP
> DisplayPort Subsystem" from Jul 7, 2018, leads to the following
> static checker warning:
>
>         drivers/gpu/drm/xlnx/zynqmp_dp.c:594 zynqmp_dp_mode_configure()
>         error: iterator underflow 'bws' (-1)-2
>
> Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
> Signed-off-by: Hyun Kwon <hyun.kwon@xilinx.com>
> ---
>  drivers/gpu/drm/xlnx/zynqmp_dp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/xlnx/zynqmp_dp.c b/drivers/gpu/drm/xlnx/zynqmp_dp.c
> index b735072..1be2b19 100644
> --- a/drivers/gpu/drm/xlnx/zynqmp_dp.c
> +++ b/drivers/gpu/drm/xlnx/zynqmp_dp.c
> @@ -579,7 +579,7 @@ static int zynqmp_dp_mode_configure(struct zynqmp_dp *dp, int pclock,
>                 return -EINVAL;
>         }
>
> -       for (i = ARRAY_SIZE(bws) - 1; i >= 0; i--) {
> +       for (i = ARRAY_SIZE(bws) - 1; i > 0; i--) {

But now we don't go through the lowest element anymore, which also
looks wrong. Or I'm blind.

I think the problem is later on that we should bail out of the loop on
the last iteration (when i == 0) before we decrement, since otherwise
we then look at bws[-1] in the next loop, which is clearly wrong. I
guess your code results in the same, but it's very confusing logic for
me ...
-Daniel

>                 if (current_bw && bws[i] >= current_bw)
>                         continue;
>
> --
> 2.7.4
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel