Re: [PATCH] drm/dp: only accept valid DP_TRAINING_AUX_RD_INTERVAL values

From: Jani Nikula <jani.nikula@linux.intel.com>
To: Rodrigo Vivi <rodrigo.vivi@intel.com>, matthew.s.atwood@intel.com
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [PATCH] drm/dp: only accept valid DP_TRAINING_AUX_RD_INTERVAL values
Date: Sun, 04 Mar 2018 12:03:55 +0200	[thread overview]
Message-ID: <87muzotawk.fsf@intel.com> (raw)
In-Reply-To: <20180302232159.GS15113@intel.com>

On Fri, 02 Mar 2018, Rodrigo Vivi <rodrigo.vivi@intel.com> wrote:
> On Fri, Mar 02, 2018 at 02:25:58PM -0800, matthew.s.atwood@intel.com wrote:
>> From: Matt Atwood <matthew.s.atwood@intel.com>
>> 
>> For panels that do not follow Display Port specifications mask off invalid
>> values for DP_TRAINING_AUX_RD_INTERVAL. Specification lists acceptable
>> values 0-4 all other values are reserved and bit 7 of DPCD 0x0000e
>> describes another feature. Currently the code uses all of DPCD 0x0000e and
>> can cause max wait for 1024 ms instead of 16 ms as specified table 2-158.
>> This address is read for both clock recovery and channel equalization.
>> 
>> Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
>> ---
>>  drivers/gpu/drm/drm_dp_helper.c | 4 ++--
>>  include/drm/drm_dp_helper.h     | 1 +
>>  2 files changed, 3 insertions(+), 2 deletions(-)
>> 
>> diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c
>> index adf79be..a7e9b75 100644
>> --- a/drivers/gpu/drm/drm_dp_helper.c
>> +++ b/drivers/gpu/drm/drm_dp_helper.c
>> @@ -122,7 +122,7 @@ void drm_dp_link_train_clock_recovery_delay(const u8 dpcd[DP_RECEIVER_CAP_SIZE])
>>  	if (dpcd[DP_TRAINING_AUX_RD_INTERVAL] == 0)
>>  		udelay(100);
>>  	else
>> -		mdelay(dpcd[DP_TRAINING_AUX_RD_INTERVAL] * 4);
>> +		mdelay((dpcd[DP_TRAINING_AUX_RD_INTERVAL] & DP_TRAINING_AUX_RD_MASK) * 4);
>>  }
>>  EXPORT_SYMBOL(drm_dp_link_train_clock_recovery_delay);
>>  
>> @@ -130,7 +130,7 @@ void drm_dp_link_train_channel_eq_delay(const u8 dpcd[DP_RECEIVER_CAP_SIZE]) {
>>  	if (dpcd[DP_TRAINING_AUX_RD_INTERVAL] == 0)
>>  		udelay(400);
>>  	else
>> -		mdelay(dpcd[DP_TRAINING_AUX_RD_INTERVAL] * 4);
>> +		mdelay((dpcd[DP_TRAINING_AUX_RD_INTERVAL] & DP_TRAINING_AUX_RD_MASK) * 4);
>>  }
>>  EXPORT_SYMBOL(drm_dp_link_train_channel_eq_delay);
>>  
>> diff --git a/include/drm/drm_dp_helper.h b/include/drm/drm_dp_helper.h
>> index da58a42..77ba003 100644
>> --- a/include/drm/drm_dp_helper.h
>> +++ b/include/drm/drm_dp_helper.h
>> @@ -118,6 +118,7 @@
>>  # define DP_DPCD_DISPLAY_CONTROL_CAPABLE     (1 << 3) /* edp v1.2 or higher */
>>  
>>  #define DP_TRAINING_AUX_RD_INTERVAL         0x00e   /* XXX 1.2? */
>> +# define DP_TRAINING_AUX_RD_MASK            0x7     /* 1.4 */
>
> I confirmed it is already part of 1.2 and 1.3
> So probably a good idea to already to s/XXX 1.2?/1.2/
> but this is optional and up to you.
>
> With s/1.4/1.2 feel free to add:
>
> Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

Please send v2 Cc: intel-gfx so we get i915 CI on this. It doesn't
automatically happen on dri-devel yet.

BR,
Jani.

>
> (well... in a hope that no other bad panel uses 5,6 or 7,
> that are also reserved values. But I believe these cases shouldn't
> be as bad as this one you faced here anyways)
>
>>  
>>  #define DP_ADAPTER_CAP			    0x00f   /* 1.2 */
>>  # define DP_FORCE_LOAD_SENSE_CAP	    (1 << 0)
>> -- 
>> 2.7.4
>> 
>> _______________________________________________
>> dri-devel mailing list
>> dri-devel@lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/dri-devel
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel