[PATCH v3 01/21] drm/vc4: Declare the DSI encoder as a bridge element
Boris Brezillon
boris.brezillon at collabora.com
Sun Nov 24 10:47:41 UTC 2019
On Sun, 24 Nov 2019 12:01:30 +0200
Laurent Pinchart <laurent.pinchart at ideasonboard.com> wrote:
> Hi Boris,
>
> Thank you for the patch.
>
> On Wed, Oct 23, 2019 at 05:44:52PM +0200, Boris Brezillon wrote:
> > Encoder drivers will progressively transition to the drm_bridge
> > interface in place of the drm_encoder one.
> >
> > Let's start with the VC4 driver, and use the ->pre_{enable,disable}()
> > hooks to get rid of the hack resetting encoder->bridge.next (which was
> > needed to control the encoder/bridge enable/disable sequence).
> >
> > Signed-off-by: Boris Brezillon <boris.brezillon at collabora.com>
> > ---
> > Changes in v3:
> > * Embed a drm_bridge object in vc4_dsi since drm_encoder no longer has
> > a dummy bridge
> >
> > Changes in v2:
> > * New patch (replaces "drm/vc4: Get rid of the dsi->bridge field")
> > ---
> > drivers/gpu/drm/vc4/vc4_dsi.c | 88 +++++++++++++++++++++--------------
> > 1 file changed, 52 insertions(+), 36 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/vc4/vc4_dsi.c b/drivers/gpu/drm/vc4/vc4_dsi.c
> > index c9ba83ed49b9..49f8a313e759 100644
> > --- a/drivers/gpu/drm/vc4/vc4_dsi.c
> > +++ b/drivers/gpu/drm/vc4/vc4_dsi.c
> > @@ -498,7 +498,11 @@ struct vc4_dsi {
> >
> > struct mipi_dsi_host dsi_host;
> > struct drm_encoder *encoder;
> > - struct drm_bridge *bridge;
> > +
> > + /* Embed a bridge object so we can implement bridge funcs instead of
> > + * encoder ones.
> > + */
> > + struct drm_bridge bridge;
> >
> > void __iomem *regs;
> >
> > @@ -543,6 +547,11 @@ struct vc4_dsi {
> > struct debugfs_regset32 regset;
> > };
> >
> > +static inline struct vc4_dsi *bridge_to_vc4_dsi(struct drm_bridge *bridge)
> > +{
> > + return container_of(bridge, struct vc4_dsi, bridge);
> > +}
> > +
> > #define host_to_dsi(host) container_of(host, struct vc4_dsi, dsi_host)
> >
> > static inline void
> > @@ -747,16 +756,11 @@ dsi_esc_timing(u32 ns)
> > return DIV_ROUND_UP(ns, ESC_TIME_NS);
> > }
> >
> > -static void vc4_dsi_encoder_disable(struct drm_encoder *encoder)
> > +static void vc4_dsi_bridge_post_disable(struct drm_bridge *bridge)
> > {
> > - struct vc4_dsi_encoder *vc4_encoder = to_vc4_dsi_encoder(encoder);
> > - struct vc4_dsi *dsi = vc4_encoder->dsi;
> > + struct vc4_dsi *dsi = bridge_to_vc4_dsi(bridge);
> > struct device *dev = &dsi->pdev->dev;
> >
> > - drm_bridge_disable(dsi->bridge);
> > - vc4_dsi_ulps(dsi, true);
> > - drm_bridge_post_disable(dsi->bridge);
> > -
> > clk_disable_unprepare(dsi->pll_phy_clock);
> > clk_disable_unprepare(dsi->escape_clock);
> > clk_disable_unprepare(dsi->pixel_clock);
> > @@ -764,6 +768,13 @@ static void vc4_dsi_encoder_disable(struct drm_encoder *encoder)
> > pm_runtime_put(dev);
> > }
> >
> > +static void vc4_dsi_bridge_disable(struct drm_bridge *bridge)
> > +{
> > + struct vc4_dsi *dsi = bridge_to_vc4_dsi(bridge);
> > +
> > + vc4_dsi_ulps(dsi, true);
> > +}
> > +
> > /* Extends the mode's blank intervals to handle BCM2835's integer-only
> > * DSI PLL divider.
> > *
> > @@ -777,12 +788,11 @@ static void vc4_dsi_encoder_disable(struct drm_encoder *encoder)
> > * higher-than-expected clock rate to the panel, but that's what the
> > * firmware does too.
> > */
> > -static bool vc4_dsi_encoder_mode_fixup(struct drm_encoder *encoder,
> > - const struct drm_display_mode *mode,
> > - struct drm_display_mode *adjusted_mode)
> > +static bool vc4_dsi_bridge_mode_fixup(struct drm_bridge *bridge,
> > + const struct drm_display_mode *mode,
> > + struct drm_display_mode *adjusted_mode)
> > {
> > - struct vc4_dsi_encoder *vc4_encoder = to_vc4_dsi_encoder(encoder);
> > - struct vc4_dsi *dsi = vc4_encoder->dsi;
> > + struct vc4_dsi *dsi = bridge_to_vc4_dsi(bridge);
> > struct clk *phy_parent = clk_get_parent(dsi->pll_phy_clock);
> > unsigned long parent_rate = clk_get_rate(phy_parent);
> > unsigned long pixel_clock_hz = mode->clock * 1000;
> > @@ -816,11 +826,11 @@ static bool vc4_dsi_encoder_mode_fixup(struct drm_encoder *encoder,
> > return true;
> > }
> >
> > -static void vc4_dsi_encoder_enable(struct drm_encoder *encoder)
> > +static void vc4_dsi_bridge_pre_enable(struct drm_bridge *bridge)
> > {
> > + struct drm_encoder *encoder = bridge->encoder;
> > struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
> > - struct vc4_dsi_encoder *vc4_encoder = to_vc4_dsi_encoder(encoder);
> > - struct vc4_dsi *dsi = vc4_encoder->dsi;
> > + struct vc4_dsi *dsi = bridge_to_vc4_dsi(bridge);
> > struct device *dev = &dsi->pdev->dev;
> > bool debug_dump_regs = false;
> > unsigned long hs_clock;
> > @@ -1054,8 +1064,12 @@ static void vc4_dsi_encoder_enable(struct drm_encoder *encoder)
> > }
> >
> > vc4_dsi_ulps(dsi, false);
> > +}
> >
> > - drm_bridge_pre_enable(dsi->bridge);
>
> If I'm not mistaken this switches the order of the DSI's encoder
> pre-enable and the next bridge's pre-enable. I think it's true for
> post-disable too. It may not be a problem, but have this been tested ?
No, it hasn't been tested (I don't have a Rpi with a DSI panel at
hand), and now that you mention it, I don't think it will work (I had
forgotten drm_bridge_pre_enable() iterates the bridge chain in reverse
order).
Well, it will work if the DSI encoder is connected to the RPi DSI panel
since ->prepare() doesn't do anything, but I see that some DSI panel
drivers send DSI commands in their ->prepare() method (BTW, we should
really document the fact that panel drivers can send DSI commands in
the ->prepare() hook).
The only way this can be fixed is by putting the
vc4_dsi_pre_enable/post_disable() code in runtime PM resume/suspend
hooks and let the vc4_dsi_host_transfer() call pm_runtime_get_sync(),
pm_runtime_put() every time a msg is sent.
>
> > +static void vc4_dsi_bridge_enable(struct drm_bridge *bridge)
> > +{
> > + struct vc4_dsi *dsi = bridge_to_vc4_dsi(bridge);
> > + bool debug_dump_regs = false;
> >
> > if (dsi->mode_flags & MIPI_DSI_MODE_VIDEO) {
> > DSI_PORT_WRITE(DISP0_CTRL,
> > @@ -1072,8 +1086,6 @@ static void vc4_dsi_encoder_enable(struct drm_encoder *encoder)
> > DSI_DISP0_ENABLE);
> > }
> >
> > - drm_bridge_enable(dsi->bridge);
> > -
> > if (debug_dump_regs) {
> > struct drm_printer p = drm_info_printer(&dsi->pdev->dev);
> > dev_info(&dsi->pdev->dev, "DSI regs after:\n");
> > @@ -1290,10 +1302,12 @@ static const struct mipi_dsi_host_ops vc4_dsi_host_ops = {
> > .transfer = vc4_dsi_host_transfer,
> > };
> >
> > -static const struct drm_encoder_helper_funcs vc4_dsi_encoder_helper_funcs = {
> > - .disable = vc4_dsi_encoder_disable,
> > - .enable = vc4_dsi_encoder_enable,
> > - .mode_fixup = vc4_dsi_encoder_mode_fixup,
> > +static const struct drm_bridge_funcs vc4_dsi_bridge_funcs = {
> > + .pre_enable = vc4_dsi_bridge_pre_enable,
> > + .enable = vc4_dsi_bridge_enable,
> > + .disable = vc4_dsi_bridge_disable,
> > + .post_disable = vc4_dsi_bridge_post_disable,
> > + .mode_fixup = vc4_dsi_bridge_mode_fixup,
> > };
> >
> > static const struct of_device_id vc4_dsi_dt_match[] = {
> > @@ -1445,6 +1459,7 @@ static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
> > struct vc4_dev *vc4 = to_vc4_dev(drm);
> > struct vc4_dsi *dsi = dev_get_drvdata(dev);
> > struct vc4_dsi_encoder *vc4_dsi_encoder;
> > + struct drm_bridge *next_bridge;
> > struct drm_panel *panel;
> > const struct of_device_id *match;
> > dma_cap_mask_t dma_mask;
> > @@ -1561,7 +1576,7 @@ static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
> > }
> >
> > ret = drm_of_find_panel_or_bridge(dev->of_node, 0, 0,
> > - &panel, &dsi->bridge);
> > + &panel, &next_bridge);
> > if (ret) {
> > /* If the bridge or panel pointed by dev->of_node is not
> > * enabled, just return 0 here so that we don't prevent the DRM
> > @@ -1576,10 +1591,10 @@ static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
> > }
> >
> > if (panel) {
> > - dsi->bridge = devm_drm_panel_bridge_add_typed(dev, panel,
> > + next_bridge = devm_drm_panel_bridge_add_typed(dev, panel,
> > DRM_MODE_CONNECTOR_DSI);
> > - if (IS_ERR(dsi->bridge))
> > - return PTR_ERR(dsi->bridge);
> > + if (IS_ERR(next_bridge))
> > + return PTR_ERR(next_bridge);
> > }
> >
> > /* The esc clock rate is supposed to always be 100Mhz. */
> > @@ -1598,19 +1613,20 @@ static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
> >
> > drm_encoder_init(drm, dsi->encoder, &vc4_dsi_encoder_funcs,
> > DRM_MODE_ENCODER_DSI, NULL);
> > - drm_encoder_helper_add(dsi->encoder, &vc4_dsi_encoder_helper_funcs);
> >
> > - ret = drm_bridge_attach(dsi->encoder, dsi->bridge, NULL);
> > + /* Declare ourself as the first bridge element. */
> > + dsi->bridge.funcs = &vc4_dsi_bridge_funcs;
> > + ret = drm_bridge_attach(dsi->encoder, &dsi->bridge, NULL);
> > + if (ret) {
> > + dev_err(dev, "bridge attach failed: %d\n", ret);
> > + return ret;
> > + }
> > +
> > + ret = drm_bridge_attach(dsi->encoder, next_bridge, &dsi->bridge);
> > if (ret) {
> > dev_err(dev, "bridge attach failed: %d\n", ret);
> > return ret;
> > }
>
> This is usually done in the bridge attach operation. As we're in control
> we can attach the next bridge here, but I think the driver would look
> more standard if you moved the second attach call to this bridge's
> attach operation.
I agree.
>
> With this fixed, and if the driver has been tested and the
> enable/disable order change doesn't cause issues,
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
>
> > - /* Disable the atomic helper calls into the bridge. We
> > - * manually call the bridge pre_enable / enable / etc. calls
> > - * from our driver, since we need to sequence them within the
> > - * encoder's enable/disable paths.
> > - */
> > - dsi->encoder->bridge = NULL;
> >
> > if (dsi->port == 0)
> > vc4_debugfs_add_regset32(drm, "dsi0_regs", &dsi->regset);
> > @@ -1629,7 +1645,7 @@ static void vc4_dsi_unbind(struct device *dev, struct device *master,
> > struct vc4_dev *vc4 = to_vc4_dev(drm);
> > struct vc4_dsi *dsi = dev_get_drvdata(dev);
> >
> > - if (dsi->bridge)
> > + if (dsi->bridge.next)
> > pm_runtime_disable(dev);
> >
> > vc4_dsi_encoder_destroy(dsi->encoder);
>
More information about the dri-devel
mailing list