linux-kernel - Re: [PATCH] drm/rcar-du: dsi: Handle both DRM_MODE_FLAG_N.SYNC and !DRM_MODE_FLAG

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <31ce0f27-b716-4b3a-b6b6-35bcee0a33f3@mailbox.org>
Date: Tue, 2 Dec 2025 19:18:58 +0100
From: Marek Vasut <marek.vasut@...lbox.org>
To: Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc: dri-devel@...ts.freedesktop.org, David Airlie <airlied@...il.com>,
 Geert Uytterhoeven <geert+renesas@...der.be>,
 Kieran Bingham <kieran.bingham+renesas@...asonboard.com>,
 Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
 Magnus Damm <magnus.damm@...il.com>, Maxime Ripard <mripard@...nel.org>,
 Simona Vetter <simona@...ll.ch>, Thomas Zimmermann <tzimmermann@...e.de>,
 Tomi Valkeinen <tomi.valkeinen+renesas@...asonboard.com>,
 linux-kernel@...r.kernel.org, linux-renesas-soc@...r.kernel.org
Subject: Re: [PATCH] drm/rcar-du: dsi: Handle both DRM_MODE_FLAG_N.SYNC and
 !DRM_MODE_FLAG_P.SYNC

On 12/1/25 7:09 AM, Laurent Pinchart wrote:

Hello Laurent,

> On Tue, Nov 25, 2025 at 09:13:02PM +0100, Marek Vasut wrote:
>> On 11/8/25 12:23 AM, Laurent Pinchart wrote:
>>> On Sat, Nov 08, 2025 at 12:04:10AM +0100, Marek Vasut wrote:
>>>> Since commit 94fe479fae96 ("drm/rcar-du: dsi: Clean up handling of DRM mode flags")
>>>> the driver does not set TXVMVPRMSET0R_VSPOL_LOW and TXVMVPRMSET0R_HSPOL_LOW
>>>> for modes which set neither DRM_MODE_FLAG_[PN].SYNC.
>>>
>>> Could you please explain what broke ?
> 
> Sorry, I wasn't clear. I meant could you summarize the explanation in
> the commit message ?
> 
>> Consider mode->flags, V-ones for simplicity:
>>
>> Before 94fe479fae96 :
>>
>> DRM_MODE_FLAG_PVSYNC => vprmset0r |= 0
>> DRM_MODE_FLAG_NVSYNC => vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW
>> Neither DRM_MODE_FLAG_[PN]VSYNC => vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW
>>
>> After 94fe479fae96 :
>>
>> DRM_MODE_FLAG_PVSYNC => vprmset0r |= 0
>> DRM_MODE_FLAG_NVSYNC => vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW
>> Neither DRM_MODE_FLAG_[PN]VSYNC => vprmset0r |= 0 <---------- This broke
>>
>> The "Neither" case behavior is different. I did not realize that:
>>
>> DRM_MODE_FLAG_N[HV]SYNC is not equivalent !DRM_MODE_FLAG_P[HV]SYNC
>>
>> They really are not equivalent .
>>
>> [...]
>>
>>>>    	/* Configuration for Video Parameters, input is always RGB888 */
>>>>    	vprmset0r = TXVMVPRMSET0R_BPP_24;
>>>> -	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
>>>> +	if ((mode->flags & DRM_MODE_FLAG_NVSYNC) ||
>>>> +	    !(mode->flags & DRM_MODE_FLAG_PVSYNC))
>>>>    		vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW;
>>>
>>> I don't think this restores the previous behaviour. You would need to
>>> write
>>>
>>> 	if (!(mode->flags & DRM_MODE_FLAG_PVSYNC))
>>> 		vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW;
>>
>> This patch covers both the N[HV]SYNC and !P[HV]SYNC , so that should
>> restore the behavior to "Before" and explicitly be clear that N[HV]SYNC
>> and !P[HV]SYNC are not the same thing.
> 
> Before commit 94fe479fae96 we had
> 
> 	vprmset0r = (mode->flags & DRM_MODE_FLAG_PVSYNC ?
> 		     TXVMVPRMSET0R_VSPOL_HIG : TXVMVPRMSET0R_VSPOL_LOW)
> 		  | (mode->flags & DRM_MODE_FLAG_PHSYNC ?
> 		     TXVMVPRMSET0R_HSPOL_HIG : TXVMVPRMSET0R_HSPOL_LOW)
> 		  | TXVMVPRMSET0R_CSPC_RGB | TXVMVPRMSET0R_BPP_24;
> 
> Considering the vertical sync for simplicity, this gives us
> 
> NVSYNC \ PVSYNC		0		1
>   0			VSPOL_LOW	VSPOL_HIG
>   1			VSPOL_LOW	VSPOL_HIG
> 
> With this patch, the code becomes
> 
> 	/* Configuration for Video Parameters, input is always RGB888 */
> 	vprmset0r = TXVMVPRMSET0R_BPP_24;
> 	if ((mode->flags & DRM_MODE_FLAG_NVSYNC) ||
> 	    !(mode->flags & DRM_MODE_FLAG_PVSYNC))
> 		vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW;
> 	if ((mode->flags & DRM_MODE_FLAG_NHSYNC) ||
> 	    !(mode->flags & DRM_MODE_FLAG_PHSYNC))
> 		vprmset0r |= TXVMVPRMSET0R_HSPOL_LOW;
> 
> which gives us
> 
> NVSYNC \ PVSYNC		0		1
>   0			VSPOL_LOW	VSPOL_HIG
>   1			VSPOL_LOW	VSPOL_LOW
> 
> This is a different behaviour. Granted, we should never have both NVSYNC
> and PVSYNC set together (unless I'm missing something), so the
> difference in behaviour shouldn't matter. I'm fine with that if you
> explain it in the commit message, however I think that writing
> 
>   	if (!(mode->flags & DRM_MODE_FLAG_PVSYNC))
>   		vprmset0r |= TXVMVPRMSET0R_VSPOL_LOW;
>   	if (!(mode->flags & DRM_MODE_FLAG_PHSYNC))
>   		vprmset0r |= TXVMVPRMSET0R_HSPOL_LOW;
> 
> would both restore the previous behaviour in all cases, and be simpler.
I sent a V2 which addresses both, the commit message update and this 
comment.