Received-SPF: pass (google.com: domain of linux-kernel+bounces-153931-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Date: Mon, 22 Apr 2024 22:44:19 +0300
From: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
To: Abhinav Kumar <quic_abhinavk@quicinc.com>
Cc: Rob Clark <robdclark@gmail.com>, Sean Paul <sean@poorly.run>, 
	Marijn Suijten <marijn.suijten@somainline.org>, David Airlie <airlied@gmail.com>, 
	Daniel Vetter <daniel@ffwll.ch>, linux-arm-msm@vger.kernel.org, dri-devel@lists.freedesktop.org, 
	freedreno@lists.freedesktop.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 1/3] drm/msm: don't clean up priv->kms prematurely
Message-ID: <pirzyu6ki7hfrax2uo3qlg7fdzlvzpg7vfgfqawogtybdd6e3c@zvgh3n4o67zi>
References: <20240420-mdp4-fixes-v1-0-96a70f64fa85@linaro.org>
 <20240420-mdp4-fixes-v1-1-96a70f64fa85@linaro.org>
 <67fbd629-3e80-b706-83a3-7baff3efd6c1@quicinc.com>
 <n7eet54g72usmuh4hdz6yce3i4ieweu4orgd7gewu7y53ejucc@dzmq2a2wdxkc>
 <7daeeaba-a33d-bf05-1371-d9b5fa0bf0a5@quicinc.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <7daeeaba-a33d-bf05-1371-d9b5fa0bf0a5@quicinc.com>

On Mon, Apr 22, 2024 at 10:23:18AM -0700, Abhinav Kumar wrote:
> 
> 
> On 4/21/2024 3:35 PM, Dmitry Baryshkov wrote:
> > On Sat, Apr 20, 2024 at 04:02:00PM -0700, Abhinav Kumar wrote:
> > > 
> > > 
> > > On 4/19/2024 7:33 PM, Dmitry Baryshkov wrote:
> > > > MSM display drivers provide kms structure allocated during probe().
> > > > Don't clean up priv->kms field in case of an error. Otherwise probe
> > > > functions might fail after KMS probe deferral.
> > > > 
> > > 
> > > So just to understand this more, this will happen when master component
> > > probe (dpu) succeeded but other sub-component probe (dsi) deferred?
> > > 
> > > Because if master component probe itself deferred it will allocate priv->kms
> > > again isnt it and we will not even hit here.
> > 
> > Master probing succeeds (so priv->kms is set), then kms_init fails at
> > runtime, during binding of the master device. This results in probe
> > deferral from the last component's component_add() function and reprobe
> > attempt when possible (once the next device is added or probed). However
> > as priv->kms is NULL, probe crashes.
> > 
> > > 
> > > > Fixes: a2ab5d5bb6b1 ("drm/msm: allow passing struct msm_kms to msm_drv_probe()")
> 
> Actually, Is this Fixes tag correct?
> 
> OR is this one better
> 
> Fixes: 506efcba3129 ("drm/msm: carve out KMS code from msm_drv.c")

No. The issue existed even before the carve-out.

> 
> 
> > > > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> > > > ---
> > > >    drivers/gpu/drm/msm/msm_kms.c | 1 -
> > > >    1 file changed, 1 deletion(-)
> > > > 
> > > > diff --git a/drivers/gpu/drm/msm/msm_kms.c b/drivers/gpu/drm/msm/msm_kms.c
> > > > index af6a6fcb1173..6749f0fbca96 100644
> > > > --- a/drivers/gpu/drm/msm/msm_kms.c
> > > > +++ b/drivers/gpu/drm/msm/msm_kms.c
> > > > @@ -244,7 +244,6 @@ int msm_drm_kms_init(struct device *dev, const struct drm_driver *drv)
> > > >    	ret = priv->kms_init(ddev);
> > > >    	if (ret) {
> > > >    		DRM_DEV_ERROR(dev, "failed to load kms\n");
> > > > -		priv->kms = NULL;
> > > >    		return ret;
> > > >    	}
> > > > 
> > 

-- 
With best wishes
Dmitry