linux-kernel - Re: [PATCH RFC 0/4] drm/bridge: samsung-dsim: Stop controlling vsync display FIFO flush in panels

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: <d96f5578-4c87-4b94-b42f-7e8e54d75b04@samsung.com>
Date: Wed, 28 May 2025 08:41:27 +0200
From: Marek Szyprowski <m.szyprowski@...sung.com>
To: Philipp Zabel <p.zabel@...gutronix.de>, Inki Dae <inki.dae@...sung.com>,
	Jagan Teki <jagan@...rulasolutions.com>, Andrzej Hajda
	<andrzej.hajda@...el.com>, Neil Armstrong <neil.armstrong@...aro.org>,
	Robert Foss <rfoss@...nel.org>, Laurent Pinchart
	<Laurent.pinchart@...asonboard.com>, Jonas Karlman <jonas@...boo.se>, Jernej
	Skrabec <jernej.skrabec@...il.com>, Maarten Lankhorst
	<maarten.lankhorst@...ux.intel.com>, Maxime Ripard <mripard@...nel.org>,
	Thomas Zimmermann <tzimmermann@...e.de>, David Airlie <airlied@...il.com>,
	Simona Vetter <simona@...ll.ch>, Artur Weber <aweber.kernel@...il.com>,
	Jessica Zhang <quic_jesszhan@...cinc.com>
Cc: dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
	kernel@...gutronix.de
Subject: Re: [PATCH RFC 0/4] drm/bridge: samsung-dsim: Stop controlling
 vsync display FIFO flush in panels

On 27.05.2025 16:14, Philipp Zabel wrote:
> This series enables the vsync flush feature in the samsung-dsim driver
> unconditionally and removes the MIPI_DSI_MODE_VSYNC_FLUSH flag.
>
> Background: I've recently seen shifted display issues on two different
> i.MX8MM boards (mxsfb + samsung-dsim) with different DSI panels.
> The symptoms were horizonally shifted display contents, with a stable
> offset, in about 0.1 to 0.6 percent of modesets.
> Enabling the MIPI_DSI_MODE_VSYNC_FLUSH flag in the panels' mode_flags
> fixed the issue in both cases.
>
> The samsung-dsim driver is the only DSI bridge driver that uses this
> flag: If the flag is absent, the driver sets the DSIM_MFLUSH_VS bit in
> the DSIM_CONFIG_REG register, which disables the vsync flush feature.
> The reset value of this bit is cleared (vsync flush is default-enabled).
> According to the i.MX8MM reference manual,
>
>      "It needs that Main display FIFO should be flushed for deleting
>       garbage data."
>
> This appears to match the comment in mxsfb_reset_block() in mxsfb_kms.c:
>
>      /*
>       * It seems, you can't re-program the controller if it is still
>       * running. This may lead to shifted pictures (FIFO issue?), so
>       * first stop the controller and drain its FIFOs.
>       */
>
> Now I wonder why the bit is controlled by a flag in the panel drivers.
> Whether the display controller pushes up to a FIFO worth of garbage data
> into the DSI bridge during initialization seems to be a property of the
> display controller / DSI bridge integration (whether this is due to
> hardware or driver bugs), not a specific requirement of the panel.
> Surely no panel needs to receive a partial line of garbage data in front
> of the first frame?
>
> Instead of adding the flag to every panel connected to affected SoCs,
> the vsync flush feature could just be enabled unconditionally.
> Clearing an already-empty display FIFO should have no effect, unless
> I'm missing something? With that, the MIPI_DSI_MODE_VSYNC_FLUSH flag
> would not be used anymore and could be removed.

The Exynos5433 datasheet doesn't give us anything more about this bit:

"Auto flush of display FIFO in only video mode.
It requires that the display FIFO should be flushed for
deleting garbage data in video mode."

Your reasoning seems to be correct, it probably slipped into 
MIPI_DSI_MODE flags just because it is in the same register.

Feel free to add:

Acked-by: Marek Szyprowski <m.szyprowski@...sung.com>

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland