Message-ID: <wepxz7qfph5cqxa7jnxiwdya3znaljqng7flggn65bmqrkwqyj@hjklrehup452>
Date:   Wed, 13 Dec 2023 13:17:54 +0100
From:   Maxime Ripard <mripard@...nel.org>
To:     Dario Binacchi <dario.binacchi@...rulasolutions.com>
Cc:     Jagan Teki <jagan@...rulasolutions.com>,
        Dave Stevenson <dave.stevenson@...pberrypi.com>,
        Michael Nazzareno Trimarchi <michael@...rulasolutions.com>,
        linux-kernel@...r.kernel.org,
        Amarula patchwork <linux-amarula@...rulasolutions.com>,
        Andrzej Hajda <andrzej.hajda@...el.com>,
        Daniel Vetter <daniel@...ll.ch>,
        David Airlie <airlied@...il.com>,
        Frieder Schrempf <frieder.schrempf@...tron.de>,
        Jernej Skrabec <jernej.skrabec@...il.com>,
        Jonas Karlman <jonas@...boo.se>,
        Laurent Pinchart <Laurent.pinchart@...asonboard.com>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Neil Armstrong <neil.armstrong@...aro.org>,
        Robert Foss <rfoss@...nel.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        dri-devel@...ts.freedesktop.org
Subject: Re: [PATCH v4 02/10] drm/bridge: Fix a use case in the bridge
 disable logic

On Wed, Dec 13, 2023 at 12:59:05PM +0100, Dario Binacchi wrote:
> Hi Jagan and Dave,
> 
> On Wed, Dec 6, 2023 at 2:57 PM Michael Nazzareno Trimarchi
> <michael@...rulasolutions.com> wrote:
> >
> > Hi Jagan
> >
> > On Wed, Dec 6, 2023 at 2:31 PM Jagan Teki <jagan@...rulasolutions.com> wrote:
> > >
> > > Hi Dario,
> > >
> > > On Wed, Dec 6, 2023 at 6:57 PM Dario Binacchi
> > > <dario.binacchi@...rulasolutions.com> wrote:
> > > >
> > > > Hi Dave and Jagan,
> > > >
> > > > On Tue, Dec 5, 2023 at 4:39 PM Dave Stevenson
> > > > <dave.stevenson@...pberrypi.com> wrote:
> > > > >
> > > > > Hi Dario
> > > > >
> > > > > On Tue, 5 Dec 2023 at 10:54, Dario Binacchi
> > > > > <dario.binacchi@...rulasolutions.com> wrote:
> > > > > >
> > > > > > The patch fixes the code for finding the next bridge with the
> > > > > > "pre_enable_prev_first" flag set to false. In case this condition is
> > > > > > not verified, i. e. there is no subsequent bridge with the flag set to
> > > > > > false, the whole bridge list is traversed, invalidating the "next"
> > > > > > variable.
> > > > > >
> > > > > > The use of a new iteration variable (i. e. "iter") ensures that the value
> > > > > > of the "next" variable is not invalidated.
> > > > >
> > > > > We already have https://patchwork.freedesktop.org/patch/529288/ that
> > > > > has been reviewed (but not applied) to resolve this. What does this
> > > > > version do differently and why?
> > > >
> > > > My patches only affect drm_atomic_bridge_chain_post_disable(), whereas
> > > > Jagan's patch affects both
> > > > drm_atomic_bridge_chain_post_disable() and drm_atomic_bridge_chain_pre_enable().
> > > > I tested Jagan's patch on my system with success and I reviewed with
> > > > Michael Trimarchi the
> > > > drm_atomic_bridge_chain_pre_enable() fixing and we think it's okay.
> > > > We also believe that our changes to post_disable() are better, as we
> > > > set the 'next' variable only when required,
> > > > and the code is more optimized since the list_is_last() is not called
> > > > within the loop.
> > > > Would it be possible to use Jagan's patch for fixing
> > > > drm_atomic_bridge_chain_pre_enable() and mine for
> > > > fixing drm_atomic_bridge_chain_post_disable()?
> > > >
> > >
> > > Can you please share the post-disabled bridge chain list with the
> > > below example before and after your change?
> >
> > We have already git commit the description in how the patch affects
> > the post_disable. As Dario
> > reported your patch is ok even in our use case. We don't have a real
> > use case as the one you describe.
> >
> > Can we know how you test it in this use case here? Can you test our
> > patches of post_disable?
> >
> > Thanks
> > Michael
> >
> > >
> > > Example:
> > > - Panel
> > > - Bridge 1
> > > - Bridge 2 pre_enable_prev_first
> > > - Bridge 3
> > > - Bridge 4 pre_enable_prev_first
> > > - Bridge 5 pre_enable_prev_first
> > > - Bridge 6
> > > - Encoder
> > >
> > > Thanks,
> > > Jagan.
> 
> Starting from my use case:
> 
> # cat /sys/kernel/debug/dri/32e00000.lcdif/bridge_chains
> encoder[36]
> bridge[0] type: 16, ops: 0x0, OF:
> /soc@...us@...00000/dsi@...10000:fsl,imx8mn-mipi-dsim
> bridge[1] type: 16, ops: 0x8, OF:
> /soc@...us@...00000/dsi@...10000/panel@0:sharp,ls068b3sx0
> 
> I developed a pass through MIPI-DSI bridge driver to try to test your case:
> # cat /sys/kernel/debug/dri/32e00000.lcdif/bridge_chains
> encoder[36]
> bridge[0] type: 16, ops: 0x0, OF:
> /soc@...us@...00000/dsi@...10000:fsl,imx8mn-mipi-dsim
> bridge[1] type: 16, ops: 0x0, OF: /pt_mipi_dsi1:amarula,pt-mipi-dsi
> bridge[2] type: 16, ops: 0x0, OF: /pt_mipi_dsi2:amarula,pt-mipi-dsi
> bridge[3] type: 16, ops: 0x0, OF: /pt_mipi_dsi3:amarula,pt-mipi-dsi
> bridge[4] type: 16, ops: 0x0, OF: /pt_mipi_dsi4:amarula,pt-mipi-dsi
> bridge[5] type: 16, ops: 0x0, OF: /pt_mipi_dsi5:amarula,pt-mipi-dsi
> bridge[6] type: 16, ops: 0x8, OF: /pt_mipi_dsi5/panel@0:sharp,ls068b3sx02
> 
> The pre_enable_prev_first flag is set through the
> "amarula,pre_enable_prev_first" dts property I put
> in my dts.
> Your and my patches give the same results (result: OK) in both your
> use case and mine.
> But If I test my new "enlarged" use case:
> 
> - Encoder
> - bridge[0] (samsung-dsim)
> - bridge[1] pre_enable_prev_first
> - bridge[2] pre_enable_prev_first
> - bridge[3] pre_enable_prev_first
> - bridge[4] pre_enable_prev_first
> - bridge[5] pre_enable_prev_first
> - bridge[6] pre_enable_prev_first (Panel)
> 
> the result is:
> my patches: KO
> your patch: OK
> 
> So, I will remove my patches from the series.
> 
> Can the driver I implemented to test the use cases (pass through
> MIPI-DSI) be considered useful for testing these
> bridge pipelines?
> Does it make sense to send its patch?

As it is, not really, but kunit tests would be very welcome

Maxime

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)

Powered by blists - more mailing lists