[PATCH] drm: xlnx: zynqmp: release reset to DP controller before accessing DP registers

[quanyang.wang]

Hi Laurent,

On 3/21/21 4:08 AM, Laurent Pinchart wrote:
> Hi Quanyang,
>
> Thank you for the patch.
>
> On Sat, Mar 20, 2021 at 04:37:39PM +0800, quanyang.wang at windriver.com wrote:
>> From: Quanyang Wang <quanyang.wang at windriver.com>
>>
>> When insmod zynqmp-dpsub.ko after rmmod it, system will hang with the
>> error log as below:
>>
>> root at xilinx-zynqmp:~# insmod zynqmp-dpsub.ko
>> [   88.391289] [drm] Initialized zynqmp-dpsub 1.0.0 20130509 for fd4a0000.display on minor 0
>> [   88.529906] Console: switching to colour frame buffer device 128x48
>> [   88.549402] zynqmp-dpsub fd4a0000.display: [drm] fb0: zynqmp-dpsubdrm frame buffer device
>> [   88.571624] zynqmp-dpsub fd4a0000.display: ZynqMP DisplayPort Subsystem driver probed
>> root at xilinx-zynqmp:~# rmmod zynqmp_dpsub
>> [   94.023404] Console: switching to colour dummy device 80x25
>> root at xilinx-zynqmp:~# insmod zynqmp-dpsub.ko
>> 	<hang here>
>>
>> This is because that in zynqmp_dp_probe it tries to access some DP
>> registers while the DP controller is still in the reset state. When
>> running "rmmod zynqmp_dpsub", zynqmp_dp_reset(dp, true) in
>> zynqmp_dp_phy_exit is called to force the DP controller into the reset
>> state. Then insmod will call zynqmp_dp_probe to write to the DP registers,
>> but at this moment the DP controller isn't brought out of the reset state
>> since the function zynqmp_dp_reset(dp, false) is called later and this
>> will result the system hang.
>>
>> Releasing the reset to DP controller before any read/write operation to it
>> will fix this issue.
>>
>> Signed-off-by: Quanyang Wang <quanyang.wang at windriver.com>
>> ---
>>   drivers/gpu/drm/xlnx/zynqmp_dp.c | 8 ++++----
>>   1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xlnx/zynqmp_dp.c b/drivers/gpu/drm/xlnx/zynqmp_dp.c
>> index 99158ee67d02..bb45b3663d6b 100644
>> --- a/drivers/gpu/drm/xlnx/zynqmp_dp.c
>> +++ b/drivers/gpu/drm/xlnx/zynqmp_dp.c
>> @@ -402,10 +402,6 @@ static int zynqmp_dp_phy_init(struct zynqmp_dp *dp)
>>   		}
>>   	}
>>   
>> -	ret = zynqmp_dp_reset(dp, false);
>> -	if (ret < 0)
>> -		return ret;
>> -
>>   	zynqmp_dp_clr(dp, ZYNQMP_DP_PHY_RESET, ZYNQMP_DP_PHY_RESET_ALL_RESET);
>>   
>>   	/*
>> @@ -1682,6 +1678,10 @@ int zynqmp_dp_probe(struct zynqmp_dpsub *dpsub, struct drm_device *drm)
>>   		return PTR_ERR(dp->reset);
>>   	}
>>   
>> +	ret = zynqmp_dp_reset(dp, false);
>> +	if (ret < 0)
>> +		return ret;
>> +
> This change looks good to me.
>
>>   	ret = zynqmp_dp_phy_probe(dp);
>>   	if (ret)
>>   		return ret;
> But shouldn't we call zynqmp_dp_reset(dp, true) here ? Or rather, call
> it in the error path at the end of the function, with a goto label.

Thank you for the suggestions.

zynqmp_dp_reset(dp, true) should be added in the error path, I will add 
this in the V2 patch.

>
> For symmetry, should we also move the zynqmp_dp_reset() call from
> zynqmp_dp_phy_exit() to zynqmp_dp_remove() ?

Yes, I will move it out of zynqmp_dp_phy_exit and to zynqmp_dp_remove() 
in the V2 patch.

Thanks,

Quanyang