drm/atomic-helper: Re-order CRTC and Bridge ops
Subsystems (3)
v13 cover drm/atomic-helper: Re-order CRTC and Bridge ops
Hello all,
This series re-orders the sequences in which the drm CRTC and the drm Bridge get enabled and disabled with respect to each other.
The bridge pre_enable calls have been shifted before the crtc_enable and the bridge post_disable calls have been shifted after the crtc_disable.
This has been done as per the definition of bridge pre_enable. "The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".
Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are pre_enabled.
The original sequence. for display pipe enable looks like:
crtc_enable
bridge[n]_pre_enable
...
bridge[1]_pre_enable
encoder_enable
bridge[1]_enable
...
bridge[n]_enableThe sequence of enable after this patch-set will look like:
bridge[n]_pre_enable
...
bridge[1]_pre_enable
crtc_enable
encoder_enable
bridge[1]_enable
...
bridge[n]_enableFor the disable sequence, this is what the original looks like:
bridge[n]_disable
...
bridge[1]_disable
encoder_disable
bridge[1]_post_disable
...
bridge[n]_post_disable
crtc_disableThis is what the disable sequence will be, after this series of patches:
bridge[n]_disable
...
bridge[1]_disable
encoder_disable
crtc_disable
bridge[1]_post_disable
...
bridge[n]_post_disableThis series further updates the bridge API definitions to accurately reflect the updated scenario.
This series is a subset of its v11[0] which had 14 patches in the revision. 9 of those 14 patches (which were specific to the cdns-dsi bridge driver) were merged[1].
Regards
Aradhya
---
References:
[0]: Revision v11 of this series.
https://lore.kernel.org/all/[email protected]/
[1]: Patches 1 through 9 getting merged.
https://lore.kernel.org/all/[email protected]/
---
Change Log:
- Changes in v13:
- Style changes in patch 2/4. (Thomas)
- Squash patch v12:4/5 into v12:3/5, which is now v13:3/4. (Thomas)
- Add R-b tags from Thomas Zimmermann in patches 1-3.
- Rebase onto latest drm-misc-next.
- Changes in v12:
- Drop patches 1 through 9 since they have been merged.
- Rebase onto newer drm-misc-next.
- Re-word the patch 3/4, ("drm/bridge: Update the bridge enable/disable doc") to make it more readable.
- Changes in v11:
- Add patch v11:13/14 ("drm/bridge: Update the bridge enable/disable doc"), that updates the documentation about the order of the various bridge enable/disable hooks being called wrt the CRTC and encoder hooks.
- Rebase on drm-misc-next instead of linux-next.
As part of rebase, accommodate the following change:
- Change patch v10:08/13 ("drm/bridge: cdns-dsi: Support atomic bridge APIs") to v11:08/13 ("drm/bridge: cdns-dsi: Add input format negotiation"), since Maxime has already updated the bridge hooks to their atomic versions in commit 68c98e227a96 ("drm/bridge: cdns-csi:
Switch to atomic helpers").
My new patch now only adds the format negotiation hook for the cdns-dsi. (Note: Since the new patch is now only a subset of the old one, without any change in logic, I decided to carry forward the R-b and T-b tags.)
- Add Alexander Sverdlin's T-b in patches 10, 11, 12.
- Changes in v10:
- Rebase on latest linux-next (next-20250226).
- As part of rebase, update the patches to accommodate a couple of widespread changes in DRM Framework -
- All the ("drm/atomic-helper: Change parameter name of ***") commits.
- All the ("drm/bridge: Pass full state to ***") commits.
(These updates are only trivial substitutions.)
- Add Tomi Valkeinen's T-b tags in all the patches.
- Changes in v9:
- Fix the oops in 11/13 - where the encoder_bridge_enable _was_ pre_enabling the bridges instead of enabling.
- Add the following tags:
- Dmitry Baryshkov's R-b in patches 2, 10, 11, and A-b in patch 12.
- Jayesh Choudhary's R-b in patch 12.
- Tomi Valkeinen's R-b in patches 2, 10, 11, 12.
- Changes in v8:
- Move the phy de-initialization to bridge post_disable() instead of bridge disable() in patch-3.
- Copy the private bridge state (dsi_cfg), in addition to the bridge_state, in patch-9.
- Split patch v7:11/12 into three patches, v8:{10,11,12}/13, to separate out different refactorings into different patches, and improve bisectability.
- Move patch v7:02/12 down to v8:06/12, to keep the initial patches for fixes only.
- Drop patch v7:04/12 as it doesn't become relevant until patch v7:12/12.
- Add R-b tags of Dmitry Baryshkov in patch-9 and patch-3, and of Tomi Valkeinen in patch-9.
- Changes in v7:
- phy_init()/exit() were called from the PM path in v6. Change it back to the bridge enable/disable path in patch-3, so that the phy_init() can go back to being called after D-Phy reset assert.
- Reword commit text in patch-5 to explain the need of the fix.
- Drop the stray code in patch-10.
- Add R-b tag of Dmitry Baryshkov in patch-6.
- Changes in v6:
- Reword patch 3 to better explain the fixes around phy de-init.
- Fix the Lane ready timeout condition in patch 7.
- Fix the dsi _bridge_atomic_check() implementation by adding a new bridge state structure in patch 10.
- Rework and combine patches v5:11/13 and v5:12/13 to v6:11/12.
- Generate the patches of these series using the "patience" algorithm.
Note: All patches, except v6:11/12, *do not* differ from their default (greedy) algorithm variants.
For patch 11, the patience algorithm significantly improves the readability.
- Rename and move the Bridge enable/disable enums from public to private in patch 11.
- Add R-b tags of Tomi Valkeinen in patch 6, and Dmitry Baryshkov in patch 2.
- Changes in v5:
- Fix subject and description in patch 1/13.
- Add patch to check the return value of phy_mipi_dphy_get_default_config() (patch: 6/13).
- Change the Clk and Data Lane ready timeout from forever to 5s.
- Print an error instead of calling WARN_ON_ONCE in patch 7/13.
- Drop patch v4-07/11: "drm/bridge: cdns-dsi: Reset the DCS write FIFO". There has been some inconsistencies found with this patch upon further testing. This patch was being used to enable a DSI panel based on ILITEK ILI9881C bridge. This will be debugged separately.
- Add patch to move the DSI mode check from _atomic_enable() to _atomic_check() (patch: 10/13).
- Split patch v4-10/11 into 2 patches - 11/13 and 12/13.
Patch 11/13 separates out the Encoder-Bridge operations into a helper function *without* changing the logic. Patch 12/13 then changes the order of the encoder-bridge operations as was intended in the original patch.
- Add detailed comment for patch 13/13.
- Add Tomi Valkeinen's R-b in patches 1, 2, 4, 5, 7, 8, 9, 13.
- Changes in v4:
- Add new patch, "drm/bridge: cdns-dsi: Move to devm_drm_of_get_bridge()", to update to an auto-managed way of finding next bridge in the chain.
- Drop patch "drm/bridge: cdns-dsi: Fix the phy_initialized variable" and add "drm/bridge: cdns-dsi: Fix Phy _init() and _exit()" that properly de-initializes the Phy and maintains the initialization state.
- Reword patch "drm/bridge: cdns-dsi: Reset the DCS write FIFO" to explain the HW concerns better.
- Add R-b tag from Dmitry Baryshkov for patches 1/11 and 8/11.
- Changes in v3:
- Reword the commit message for patch "drm/bridge: cdns-dsi: Fix OF node pointer".
- Add a new helper API to figure out DSI host input pixel format in patch "drm/mipi-dsi: Add helper to find input format".
- Use a common function for bridge pre-enable and enable, and bridge disable and post-disable, to avoid code duplication.
- Add T-b tag from Dominik Haller in patch 5/10. (Missed to add it in v2).
- Add R-b tag from Dmitry Baryshkov for patch 8/10.
- Changes in v2:
- Drop patch "drm/tidss: Add CRTC mode_fixup"
- Split patch "drm/bridge: cdns-dsi: Fix minor bugs" into 4 separate ones
- Drop support for early_enable/late_disable APIs and instead re-order the pre_enable / post_disable APIs to be called before / after crtc_enable / crtc_disable.
- Drop support for early_enable/late_disable in cdns-dsi and use pre_enable/post_disable APIs instead to do bridge enable/disable.
Previous versions:
v1: https://lore.kernel.org/all/[email protected]/ v2: https://lore.kernel.org/all/[email protected]/ v3: https://lore.kernel.org/all/[email protected]/ v4: https://lore.kernel.org/all/[email protected]/ v5: https://lore.kernel.org/all/[email protected]/ v6: https://lore.kernel.org/all/[email protected]/ v7: https://lore.kernel.org/all/[email protected]/ v8: https://lore.kernel.org/all/[email protected]/ v9: https://lore.kernel.org/all/[email protected]/ v10: https://lore.kernel.org/all/[email protected]/ v11: https://lore.kernel.org/all/[email protected]/ v12: https://lore.kernel.org/all/[email protected]/
---
Aradhya Bhatia (4):
drm/atomic-helper: Refactor crtc & encoder-bridge op loops into
separate functions
drm/atomic-helper: Separate out bridge pre_enable/post_disable from
enable/disable
drm/atomic-helper: Re-order bridge chain pre-enable and post-disable drm/bridge: cdns-dsi: Use pre_enable/post_disable to enable/disable
.../gpu/drm/bridge/cadence/cdns-dsi-core.c | 64 +++-- drivers/gpu/drm/drm_atomic_helper.c | 160 +++++++++-- include/drm/drm_bridge.h | 249 +++++++++++++----- 3 files changed, 355 insertions(+), 118 deletions(-)
base-commit: 0b3d99425891e3c4a87259afb88fbd1168dc7707
Signature · 1 line
2.34.1
PATCH v13 1/4 drm/atomic-helper: Refactor crtc & encoder-bridge op loops into separate functions
From: Aradhya Bhatia <[email protected]>
The way any singular display pipeline, in need of a modeset, gets enabled is as follows -
crtc enable
(all) bridge pre-enable
encoder enable
(all) bridge enable
- and the disable sequence is exactly the reverse of this.
The crtc operations occur by looping over the old and new crtc states, while the encoder and bridge operations occur together, by looping over the connector states of the display pipelines.
Refactor these operations - crtc enable/disable, and encoder & bridge (pre/post) enable/disable - into separate functions each, to make way for the re-ordering of the enable/disable sequences.
This patch doesn't alter the sequence of crtc/encoder/bridge operations in any way, but helps to cleanly pave the way for the next two patches, by maintaining logical bisectability.
--- drivers/gpu/drm/drm_atomic_helper.c | 69 ++++++++++++++++++++--------- 1 file changed, 49 insertions(+), 20 deletions(-)
drivers/gpu/drm/drm_atomic_helper.c +49 -20
@@ -1160,11 +1160,10 @@ crtc_needs_disable(struct drm_crtc_state *old_state, }static void -disable_outputs (struct drm_device *dev,struct drm_atomic_state *state) +encoder_bridge_disable (struct drm_device *dev,struct drm_atomic_state *state) {struct drm_connector *connector;struct drm_connector_state *old_conn_state, *new_conn_state; -struct drm_crtc *crtc;struct drm_crtc_state *old_crtc_state, *new_crtc_state;int i;
@@ -1227,6 +1226,14 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *state)drm_atomic_bridge_chain_post_disable (bridge, state); } +} + +static void +crtc_disable (struct drm_device *dev,struct drm_atomic_state *state) +{ +struct drm_crtc *crtc; +struct drm_crtc_state *old_crtc_state, *new_crtc_state; +int i;for_each_oldnew_crtc_in_state (state, crtc, old_crtc_state, new_crtc_state, i) {const struct drm_crtc_helper_funcs *funcs;
@@ -1274,6 +1281,14 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *state) } } +static void +disable_outputs (struct drm_device *dev,struct drm_atomic_state *state) +{ +encoder_bridge_disable (dev, state); + +crtc_disable (dev, state); +} +/** * drm_atomic_helper_update_legacy_modeset_state - update legacy modeset state * @dev: DRM device
@@ -1483,28 +1498,12 @@ static void drm_atomic_helper_commit_writebacks(struct drm_device *dev, } } -/** -* drm_atomic_helper_commit_modeset_enables - modeset commit to enable outputs -* @dev: DRM device -* @state: atomic state object being committed -* -* This function enables all the outputs with the new configuration which had to -* be turned off for the update. -* -* For compatibility with legacy CRTC helpers this should be called after -* drm_atomic_helper_commit_planes(), which is what the default commit function -* does. But drivers with different needs can group the modeset commits together -* and do the plane commits at the end. This is useful for drivers doing runtime -* PM since planes updates then only happen when the CRTC is actually enabled. -*/ -void drm_atomic_helper_commit_modeset_enables (struct drm_device *dev, -struct drm_atomic_state *state) +static void +crtc_enable (struct drm_device *dev,struct drm_atomic_state *state) {struct drm_crtc *crtc;struct drm_crtc_state *old_crtc_state;struct drm_crtc_state *new_crtc_state; -struct drm_connector *connector; -struct drm_connector_state *new_conn_state;int i;for_each_oldnew_crtc_in_state (state, crtc, old_crtc_state, new_crtc_state, i) {
@@ -1528,6 +1527,14 @@ void drm_atomic_helper_commit_modeset_enables(struct drm_device *dev, funcs->commit (crtc); } } +} + +static void +encoder_bridge_enable (struct drm_device *dev,struct drm_atomic_state *state) +{ +struct drm_connector *connector; +struct drm_connector_state *new_conn_state; +int i;for_each_new_connector_in_state (state, connector, new_conn_state, i) {const struct drm_encoder_helper_funcs *funcs;
@@ -1565,6 +1572,28 @@ void drm_atomic_helper_commit_modeset_enables(struct drm_device *dev,drm_atomic_bridge_chain_enable (bridge, state); } +} + +/** +* drm_atomic_helper_commit_modeset_enables - modeset commit to enable outputs +* @dev: DRM device +* @state: atomic state object being committed +* +* This function enables all the outputs with the new configuration which had to +* be turned off for the update. +* +* For compatibility with legacy CRTC helpers this should be called after +* drm_atomic_helper_commit_planes(), which is what the default commit function +* does. But drivers with different needs can group the modeset commits together +* and do the plane commits at the end. This is useful for drivers doing runtime +* PM since planes updates then only happen when the CRTC is actually enabled. +*/ +void drm_atomic_helper_commit_modeset_enables (struct drm_device *dev, +struct drm_atomic_state *state) +{ +crtc_enable (dev, state); + +encoder_bridge_enable (dev, state);drm_atomic_helper_commit_writebacks (dev, state); }
Signature · 1 line
2.34.1
PATCH v13 2/4 drm/atomic-helper: Separate out bridge pre_enable/post_disable from enable/disable
In reply to Aradhya BhatiaThe encoder-bridge ops occur by looping over the new connector states of the display pipelines. The enable sequence runs as follows -
- pre_enable(bridge),
- enable(encoder),
- enable(bridge),
while the disable sequnce runs as follows -
- disable(bridge),
- disable(encoder),
- post_disable(bridge).
Separate out the pre_enable(bridge), and the post_disable(bridge) operations into separate functions each.
This patch keeps the sequence same for any singular disaplay pipe, but changes the sequence across multiple display pipelines.
This patch is meant to be an interim patch, to cleanly pave the way for the sequence re-ordering patch, and maintain bisectability in the process.
--- drivers/gpu/drm/drm_atomic_helper.c | 91 ++++++++++++++++++++++++++++- 1 file changed, 88 insertions(+), 3 deletions(-)
drivers/gpu/drm/drm_atomic_helper.c +88 -3
@@ -1223,8 +1223,6 @@ encoder_bridge_disable(struct drm_device *dev, struct drm_atomic_state *state)else if (funcs->dpms) funcs->dpms (encoder,DRM_MODE_DPMS_OFF ); } - -drm_atomic_bridge_chain_post_disable (bridge, state); } }
@@ -1281,11 +1279,65 @@ crtc_disable(struct drm_device *dev, struct drm_atomic_state *state) } } +static void +encoder_bridge_post_disable (struct drm_device *dev,struct drm_atomic_state *state) +{ +struct drm_connector *connector; +struct drm_connector_state *old_conn_state, *new_conn_state; +struct drm_crtc_state *old_crtc_state, *new_crtc_state; +int i; + +for_each_oldnew_connector_in_state (state, connector, old_conn_state, new_conn_state, i) { +struct drm_encoder *encoder; +struct drm_bridge *bridge; + +/* +* Shut down everything that's in the changeset and currently +* still on. So need to check the old, saved state. +*/ +if (!old_conn_state->crtc) +continue ; + + old_crtc_state =drm_atomic_get_old_crtc_state (state, old_conn_state->crtc); + +if (new_conn_state->crtc) + new_crtc_state =drm_atomic_get_new_crtc_state (state, + new_conn_state->crtc); +else + new_crtc_state =NULL ; + +if (!crtc_needs_disable (old_crtc_state, new_crtc_state) || + !drm_atomic_crtc_needs_modeset (old_conn_state->crtc->state)) +continue ; + + encoder = old_conn_state->best_encoder; + +/* +* We shouldn't get this far if we didn't previously have +* an encoder.. but WARN_ON() rather than explode. +*/ +if (WARN_ON (!encoder)) +continue ; + +drm_dbg_atomic (dev,"post-disabling bridges [ENCODER:%d:%s]\n" , + encoder->base.id, encoder->name); + +/* +* Each encoder has at most one connector (since we always steal +* it away), so we won't call disable hooks twice. +*/ + bridge =drm_bridge_chain_get_first_bridge (encoder); +drm_atomic_bridge_chain_post_disable (bridge, state); + } +} +static void disable_outputs (struct drm_device *dev,struct drm_atomic_state *state) {encoder_bridge_disable (dev, state); +encoder_bridge_post_disable (dev, state); +crtc_disable (dev, state); }
@@ -1498,6 +1550,38 @@ static void drm_atomic_helper_commit_writebacks(struct drm_device *dev, } } +static void +encoder_bridge_pre_enable (struct drm_device *dev,struct drm_atomic_state *state) +{ +struct drm_connector *connector; +struct drm_connector_state *new_conn_state; +int i; + +for_each_new_connector_in_state (state, connector, new_conn_state, i) { +struct drm_encoder *encoder; +struct drm_bridge *bridge; + +if (!new_conn_state->best_encoder) +continue ; + +if (!new_conn_state->crtc->state->active || + !drm_atomic_crtc_needs_modeset (new_conn_state->crtc->state)) +continue ; + + encoder = new_conn_state->best_encoder; + +drm_dbg_atomic (dev,"pre-enabling bridges [ENCODER:%d:%s]\n" , + encoder->base.id, encoder->name); + +/* +* Each encoder has at most one connector (since we always steal +* it away), so we won't call enable hooks twice. +*/ + bridge =drm_bridge_chain_get_first_bridge (encoder); +drm_atomic_bridge_chain_pre_enable (bridge, state); + } +} +static void crtc_enable (struct drm_device *dev,struct drm_atomic_state *state) {
@@ -1559,7 +1643,6 @@ encoder_bridge_enable(struct drm_device *dev, struct drm_atomic_state *state)* it away), so we won't call enable hooks twice. */ bridge =drm_bridge_chain_get_first_bridge (encoder); -drm_atomic_bridge_chain_pre_enable (bridge, state);if (funcs) {if (funcs->atomic_enable)
@@ -1593,6 +1676,8 @@ void drm_atomic_helper_commit_modeset_enables(struct drm_device *dev, {crtc_enable (dev, state); +encoder_bridge_pre_enable (dev, state); +encoder_bridge_enable (dev, state);drm_atomic_helper_commit_writebacks (dev, state);
Signature · 1 line
2.34.1
PATCH v13 3/4 drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
In reply to Aradhya BhatiaFrom: Aradhya Bhatia <[email protected]>
Move the bridge pre_enable call before crtc enable, and the bridge post_disable call after the crtc disable.
The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enable
And, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disable
The definition of bridge pre_enable hook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".
Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are pre_enabled. Fix that by re-ordering the sequence of bridge pre_enable and bridge post_disable.
While at it, update the drm bridge API documentation as well.
--- drivers/gpu/drm/drm_atomic_helper.c | 8 +- include/drm/drm_bridge.h | 249 ++++++++++++++++++++-------- 2 files changed, 187 insertions(+), 70 deletions(-)
drivers/gpu/drm/drm_atomic_helper.c +4 -4
@@ -1336,9 +1336,9 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *state) {encoder_bridge_disable (dev, state); -encoder_bridge_post_disable (dev, state); -crtc_disable (dev, state); + +encoder_bridge_post_disable (dev, state); }/**
@@ -1674,10 +1674,10 @@ encoder_bridge_enable(struct drm_device *dev, struct drm_atomic_state *state)void drm_atomic_helper_commit_modeset_enables (struct drm_device *dev,struct drm_atomic_state *state) { -crtc_enable (dev, state); -encoder_bridge_pre_enable (dev, state); +crtc_enable (dev, state); +encoder_bridge_enable (dev, state);drm_atomic_helper_commit_writebacks (dev, state);
include/drm/drm_bridge.h +183 -66
321 lines not on this page · raw message
Signature · 1 line
2.34.1
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Hi,
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>
This patch landed in today's linux-next as commit c9b1150a68d9 ("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM with Exynos_DP sub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.
Quoted 357 lines
--- drivers/gpu/drm/drm_atomic_helper.c | 8 +- include/drm/drm_bridge.h | 249 ++++++++++++++++++++-------- 2 files changed, 187 insertions(+), 70 deletions(-)drivers/gpu/drm/drm_atomic_helper.c +4
-4@@ -1336,9 +1336,9 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *state) {encoder_bridge_disable (dev, state); -encoder_bridge_post_disable (dev, state); -crtc_disable (dev, state); + +encoder_bridge_post_disable (dev, state); }/** @@ -1674,10 +1674,10 @@ encoder_bridge_enable(struct drm_device *dev, struct drm_atomic_state *state)void drm_atomic_helper_commit_modeset_enables (struct drm_device *dev,struct drm_atomic_state *state) { -crtc_enable (dev, state); -encoder_bridge_pre_enable (dev, state); +crtc_enable (dev, state); +encoder_bridge_enable (dev, state);drm_atomic_helper_commit_writebacks (dev, state);include/drm/drm_bridge.h +176
-66@@ -165,17 +165,33 @@ struct drm_bridge_funcs {/** * @disable: * -* This callback should disable the bridge. It is called right before -* the preceding element in the display pipe is disabled. If the -* preceding element is a bridge this means it's called before that -* bridge's @disable vfunc. If the preceding element is a &drm_encoder -* it's called right before the &drm_encoder_helper_funcs.disable, -* &drm_encoder_helper_funcs.prepare or &drm_encoder_helper_funcs.dpms -* hook. +* The @disable callback should disable the bridge. * * The bridge can assume that the display pipe (i.e. clocks and timing * signals) feeding it is still running when this callback is called. * +* +* If the preceding element is a &drm_bridge, then this is called before +* that bridge is disabled via one of: +* +* - &drm_bridge_funcs.disable +* - &drm_bridge_funcs.atomic_disable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called before the encoder is disabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_disable +* - &drm_encoder_helper_funcs.prepare +* - &drm_encoder_helper_funcs.disable +* - &drm_encoder_helper_funcs.dpms +* +* and the CRTC is disabled via one of: +* +* - &drm_crtc_helper_funcs.prepare +* - &drm_crtc_helper_funcs.atomic_disable +* - &drm_crtc_helper_funcs.disable +* - &drm_crtc_helper_funcs.dpms. +* * The @disable callback is optional. * * NOTE: @@ -188,17 +204,34 @@ struct drm_bridge_funcs {/** * @post_disable: * -* This callback should disable the bridge. It is called right after the -* preceding element in the display pipe is disabled. If the preceding -* element is a bridge this means it's called after that bridge's -* @post_disable function. If the preceding element is a &drm_encoder -* it's called right after the encoder's -* &drm_encoder_helper_funcs.disable, &drm_encoder_helper_funcs.prepare -* or &drm_encoder_helper_funcs.dpms hook. -* * The bridge must assume that the display pipe (i.e. clocks and timing -* signals) feeding it is no longer running when this callback is -* called. +* signals) feeding this bridge is no longer running when the +* @post_disable is called. +* +* This callback should perform all the actions required by the hardware +* after it has stopped receiving signals from the preceding element. +* +* If the preceding element is a &drm_bridge, then this is called after +* that bridge is post-disabled (unless marked otherwise by the +* @pre_enable_prev_first flag) via one of: +* +* - &drm_bridge_funcs.post_disable +* - &drm_bridge_funcs.atomic_post_disable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called after the encoder is disabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_disable +* - &drm_encoder_helper_funcs.prepare +* - &drm_encoder_helper_funcs.disable +* - &drm_encoder_helper_funcs.dpms +* +* and the CRTC is disabled via one of: +* +* - &drm_crtc_helper_funcs.prepare +* - &drm_crtc_helper_funcs.atomic_disable +* - &drm_crtc_helper_funcs.disable +* - &drm_crtc_helper_funcs.dpms * * The @post_disable callback is optional. * @@ -241,18 +274,30 @@ struct drm_bridge_funcs {/** * @pre_enable: * -* This callback should enable the bridge. It is called right before -* the preceding element in the display pipe is enabled. If the -* preceding element is a bridge this means it's called before that -* bridge's @pre_enable function. If the preceding element is a -* &drm_encoder it's called right before the encoder's -* &drm_encoder_helper_funcs.enable, &drm_encoder_helper_funcs.commit or -* &drm_encoder_helper_funcs.dpms hook. -* * The display pipe (i.e. clocks and timing signals) feeding this bridge -* will not yet be running when this callback is called. The bridge must -* not enable the display link feeding the next bridge in the chain (if -* there is one) when this callback is called. +* will not yet be running when the @pre_enable is called. +* +* This callback should perform all the necessary actions to prepare the +* bridge to accept signals from the preceding element. +* +* If the preceding element is a &drm_bridge, then this is called before +* that bridge is pre-enabled (unless marked otherwise by +* @pre_enable_prev_first flag) via one of: +* +* - &drm_bridge_funcs.pre_enable +* - &drm_bridge_funcs.atomic_pre_enable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called before the CRTC is enabled via one of: +* +* - &drm_crtc_helper_funcs.atomic_enable +* - &drm_crtc_helper_funcs.commit +* +* and the encoder is enabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_enable +* - &drm_encoder_helper_funcs.enable +* - &drm_encoder_helper_funcs.commit * * The @pre_enable callback is optional. * @@ -266,19 +311,31 @@ struct drm_bridge_funcs {/** * @enable: * -* This callback should enable the bridge. It is called right after -* the preceding element in the display pipe is enabled. If the -* preceding element is a bridge this means it's called after that -* bridge's @enable function. If the preceding element is a -* &drm_encoder it's called right after the encoder's -* &drm_encoder_helper_funcs.enable, &drm_encoder_helper_funcs.commit or -* &drm_encoder_helper_funcs.dpms hook. +* The @enable callback should enable the bridge. * * The bridge can assume that the display pipe (i.e. clocks and timing * signals) feeding it is running when this callback is called. This * callback must enable the display link feeding the next bridge in the * chain if there is one. * +* If the preceding element is a &drm_bridge, then this is called after +* that bridge is enabled via one of: +* +* - &drm_bridge_funcs.enable +* - &drm_bridge_funcs.atomic_enable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called after the CRTC is enabled via one of: +* +* - &drm_crtc_helper_funcs.atomic_enable +* - &drm_crtc_helper_funcs.commit +* +* and the encoder is enabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_enable +* - &drm_encoder_helper_funcs.enable +* - drm_encoder_helper_funcs.commit +* * The @enable callback is optional. * * NOTE: @@ -291,17 +348,30 @@ struct drm_bridge_funcs {/** * @atomic_pre_enable: * -* This callback should enable the bridge. It is called right before -* the preceding element in the display pipe is enabled. If the -* preceding element is a bridge this means it's called before that -* bridge's @atomic_pre_enable or @pre_enable function. If the preceding -* element is a &drm_encoder it's called right before the encoder's -* &drm_encoder_helper_funcs.atomic_enable hook. -* * The display pipe (i.e. clocks and timing signals) feeding this bridge -* will not yet be running when this callback is called. The bridge must -* not enable the display link feeding the next bridge in the chain (if -* there is one) when this callback is called. +* will not yet be running when the @atomic_pre_enable is called. +* +* This callback should perform all the necessary actions to prepare the +* bridge to accept signals from the preceding element. +* +* If the preceding element is a &drm_bridge, then this is called before +* that bridge is pre-enabled (unless marked otherwise by +* @pre_enable_prev_first flag) via one of: +* +* - &drm_bridge_funcs.pre_enable +* - &drm_bridge_funcs.atomic_pre_enable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called before the CRTC is enabled via one of: +* +* - &drm_crtc_helper_funcs.atomic_enable +* - &drm_crtc_helper_funcs.commit +* +* and the encoder is enabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_enable +* - &drm_encoder_helper_funcs.enable +* - &drm_encoder_helper_funcs.commit * * The @atomic_pre_enable callback is optional. */ @@ -311,18 +381,31 @@ struct drm_bridge_funcs {/** * @atomic_enable: * -* This callback should enable the bridge. It is called right after -* the preceding element in the display pipe is enabled. If the -* preceding element is a bridge this means it's called after that -* bridge's @atomic_enable or @enable function. If the preceding element -* is a &drm_encoder it's called right after the encoder's -* &drm_encoder_helper_funcs.atomic_enable hook. +* The @atomic_enable callback should enable the bridge. * * The bridge can assume that the display pipe (i.e. clocks and timing * signals) feeding it is running when this callback is called. This * callback must enable the display link feeding the next bridge in the * chain if there is one. * +* If the preceding element is a &drm_bridge, then this is called after +* that bridge is enabled via one of: +* +* - &drm_bridge_funcs.enable +* - &drm_bridge_funcs.atomic_enable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called after the CRTC is enabled via one of: +* +* - &drm_crtc_helper_funcs.atomic_enable +* - &drm_crtc_helper_funcs.commit +* +* and the encoder is enabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_enable +* - &drm_encoder_helper_funcs.enable +* - drm_encoder_helper_funcs.commit +* * The @atomic_enable callback is optional. */ void (*atomic_enable)(struct drm_bridge *bridge,@@ -330,16 +413,32 @@ struct drm_bridge_funcs {/** * @atomic_disable: * -* This callback should disable the bridge. It is called right before -* the preceding element in the display pipe is disabled. If the -* preceding element is a bridge this means it's called before that -* bridge's @atomic_disable or @disable vfunc. If the preceding element -* is a &drm_encoder it's called right before the -* &drm_encoder_helper_funcs.atomic_disable hook. +* The @atomic_disable callback should disable the bridge. * * The bridge can assume that the display pipe (i.e. clocks and timing * signals) feeding it is still running when this callback is called. * +* If the preceding element is a &drm_bridge, then this is called before +* that bridge is disabled via one of: +* +* - &drm_bridge_funcs.disable +* - &drm_bridge_funcs.atomic_disable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called before the encoder is disabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_disable +* - &drm_encoder_helper_funcs.prepare +* - &drm_encoder_helper_funcs.disable +* - &drm_encoder_helper_funcs.dpms +* +* and the CRTC is disabled via one of: +* +* - &drm_crtc_helper_funcs.prepare +* - &drm_crtc_helper_funcs.atomic_disable +* - &drm_crtc_helper_funcs.disable +* - &drm_crtc_helper_funcs.dpms. +* * The @atomic_disable callback is optional. */ void (*atomic_disable)(struct drm_bridge *bridge,@@ -348,16 +447,34 @@ struct drm_bridge_funcs {/** * @atomic_post_disable: * -* This callback should disable the bridge. It is called right after the -* preceding element in the display pipe is disabled. If the preceding -* element is a bridge this means it's called after that bridge's -* @atomic_post_disable or @post_disable function. If the preceding -* element is a &drm_encoder it's called right after the encoder's -* &drm_encoder_helper_funcs.atomic_disable hook. -* * The bridge must assume that the display pipe (i.e. clocks and timing -* signals) feeding it is no longer running when this callback is -* called. +* signals) feeding this bridge is no longer running when the +* @atomic_post_disable is called. +* +* This callback should perform all the actions required by the hardware +* after it has stopped receiving signals from the preceding element. +* +* If the preceding element is a &drm_bridge, then this is called after +* that bridge is post-disabled (unless marked otherwise by the +* @pre_enable_prev_first flag) via one of: +* +* - &drm_bridge_funcs.post_disable +* - &drm_bridge_funcs.atomic_post_disable +* +* If the preceding element of the bridge is a display controller, then +* this callback is called after the encoder is disabled via one of: +* +* - &drm_encoder_helper_funcs.atomic_disable +* - &drm_encoder_helper_funcs.prepare +* - &drm_encoder_helper_funcs.disable +* - &drm_encoder_helper_funcs.dpms
include/drm/drm_bridge.h struct drm_bridge_funcs { · line 480
Quoted 10 lines
include/drm/drm_bridge.h +7
-0+* +* and the CRTC is disabled via one of: +* +* - &drm_crtc_helper_funcs.prepare +* - &drm_crtc_helper_funcs.atomic_disable +* - &drm_crtc_helper_funcs.disable +* - &drm_crtc_helper_funcs.dpms * * The @atomic_post_disable callback is optional. */
Best regards
Signature · 2 lines
Marek Szyprowski, PhD Samsung R&D Institute Poland
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Hi,
Quoted 52 lines · Marek Szyprowski
Hi,
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>
Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.
Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects a dp->lvds bridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.
Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Tomi
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Quoted 64 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.
Quoted 7 lines
Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Best regards
Signature · 2 lines
Marek Szyprowski, PhD Samsung R&D Institute Poland
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Hi,
Quoted 71 lines · Marek Szyprowski
Quoted 70 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.
Quoted 7 lines
Quoted 1 line
Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Did you get a chance to test this more? Any hints what happens will help =)
Tomi
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Quoted 72 lines · Tomi Valkeinen
Quoted 71 lines · Marek Szyprowski
Quoted 70 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.
Quoted 7 lines
Quoted 6 lines
Quoted 1 line
Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Did you get a chance to test this more? Any hints what happens will help =)
I've spent some time debugging this issue, but so far I only got something I don't really understand. This issue is somehow related with the DP clock enabling and disabling, what is being done from exynos_dp_poweron() and exynos_dp_poweroff() functions, which are called from analogix_dp_resume() and analogix_dp_suspend(). The lockup happens somewhere while registering the fbdev console, with console lock held, what makes debugging much harder.
I've did some experiments with pm runtime of the analogix_dp_core driver and increased autosuspend delay to 200 msec. This magically fixed the issue, but I still see no direct calls to analogix_dp without proper pm runtime guards.
Best regards
Signature · 2 lines
Marek Szyprowski, PhD Samsung R&D Institute Poland
Maxime Ripard
maintainerRe: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Quoted 81 lines · Marek Szyprowski
Quoted 72 lines · Tomi Valkeinen
Quoted 71 lines · Marek Szyprowski
Quoted 70 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.Quoted 7 lines
Quoted 6 lines
Quoted 1 line
Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Did you get a chance to test this more? Any hints what happens will help =)
Quoted 7 lines
I've spent some time debugging this issue, but so far I only got something I don't really understand. This issue is somehow related with the DP clock enabling and disabling, what is being done from
exynos_dp_poweron()andexynos_dp_poweroff()functions, which are called fromanalogix_dp_resume()andanalogix_dp_suspend(). The lockup happens somewhere while registering the fbdev console, with console lock held, what makes debugging much harder.
You can skip the locking part with the fb.lockless_register_fb=1 kernel parameter. It's of course not meant for anything but debugging, but it's useful :)
Maxime
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Quoted 88 lines · Maxime Ripard
Quoted 79 lines · Marek Szyprowski
Quoted 78 lines · Tomi Valkeinen
Quoted 77 lines · Marek Szyprowski
Quoted 64 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.Quoted 7 lines
Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Quoted 8 lines
Quoted 1 line
Did you get a chance to test this more? Any hints what happens will help =)
I've spent some time debugging this issue, but so far I only got something I don't really understand. This issue is somehow related with the DP clock enabling and disabling, what is being done from
exynos_dp_poweron()andexynos_dp_poweroff()functions, which are called fromanalogix_dp_resume()andanalogix_dp_suspend(). The lockup happens somewhere while registering the fbdev console, with console lock held, what makes debugging much harder.
Quoted 3 lines
You can skip the locking part with the fb.lockless_register_fb=1 kernel parameter. It's of course not meant for anything but debugging, but it's useful :)
I've finally found where is the problem! It turned out that the issue is in the exynos_drm_fimd driver, in the code for enabling the display port clock (fimd_dp_clock_enable()
function). It touches FIMD regs, but it was not guarded with FIMD's
runtime pm calls and after the $subject change it happened that exynos_dp/analogix_dp code called the clock enable/disable when FIMD driver (which implements the CRTC object) was not runtime resumed. Previously all dp clock handling was done when CRTC was enabled, thus the FIMD device was in the resumed runtime pm state.
I will send a patch fixing this soon.
Best regards
Signature · 2 lines
Marek Szyprowski, PhD Samsung R&D Institute Poland
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
Hi,
Quoted 103 lines · Marek Szyprowski
Quoted 88 lines · Maxime Ripard
Quoted 79 lines · Marek Szyprowski
Quoted 78 lines · Tomi Valkeinen
Quoted 77 lines · Marek Szyprowski
Quoted 64 lines · Tomi Valkeinen
Quoted 49 lines · Marek Szyprowski
Quoted 41 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.The sequence of enable after this patch will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableAnd, the disable sequence for the display pipeline will look like:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThe definition of bridge
pre_enablehook says that,
"The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled. Fix that by re-ordering the sequence of bridgepre_enableand bridgepost_disable.While at it, update the drm bridge API documentation as well.
Quoted 7 lines
Acked-by Dmitry Baryshkov <[email protected]>Reviewed-by Tomi Valkeinen <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Tested-by Tomi Valkeinen <[email protected]>Tested-by Alexander Sverdlin <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Signed-off-by Aradhya Bhatia <[email protected]>Quoted 14 lines
This patch landed in today's linux-next as commit
c9b1150a68d9("drm/atomic-helper: Re-order bridge chain pre-enable and post-disable"). In my tests I found that it breaks booting of Samsung Exynos 5420/5800 based Chromebooks (Peach-Pit and Peach-Pi). Both of them use Exynos DRM withExynos_DPsub-driver (Analogix DP) and EDP panel. Booting stops at '[drm] Initialized exynos 1.1.0 for exynos-drm on minor 0' message. On the other hand, the Samsung Exynos5250 based Snow Chromebook boots fine, but it uses dp-lvds nxp,ptn3460 bridge and lvds panel instead of edp panels. This looks like some sort of deadlock, because if I disable FBDEV emulation, those boards boots fine and I'm able to run modetest and enable the display. Also the DRM kernel logger seems to be working fine, although I didn't check the screen output yet, as I only have a remote access to those boards. I will investigate it further and let You know.Quoted 7 lines
Thanks for the report. I was trying to understand the pipeline, but I'm a bit confused. Above you say Peach-Pit uses DP and EDP panel, but if I look at
arch/arm/boot/dts/samsung/exynos5420-peach-pit.dts, it connects adp->lvdsbridge (parade,ps8625). Peach-Pi seems to connect to an eDP panel.Is the above correct? Do both Peach-Pi and Peach-Pit fail?
Yes, sorry, my fault. I much have checked the same (peach-pi) dts 2 times. Both Peach-Pi and Peach-Pit fails, while Snow works fine. All three use the same Exynos DP (based on analogix dp) driver. I will try to play a bit more with those boards in the afternoon, hopefully getting some more hints where the issue is.
Quoted 8 lines
Quoted 1 line
Did you get a chance to test this more? Any hints what happens will help =)
I've spent some time debugging this issue, but so far I only got something I don't really understand. This issue is somehow related with the DP clock enabling and disabling, what is being done from
exynos_dp_poweron()andexynos_dp_poweroff()functions, which are called fromanalogix_dp_resume()andanalogix_dp_suspend(). The lockup happens somewhere while registering the fbdev console, with console lock held, what makes debugging much harder.Quoted 3 lines
You can skip the locking part with the fb.lockless_register_fb=1 kernel parameter. It's of course not meant for anything but debugging, but it's useful :)
I've finally found where is the problem! It turned out that the issue is in the
exynos_drm_fimddriver, in the code for enabling the display port clock (fimd_dp_clock_enable()
function). It touches FIMD regs, but it was not guarded with FIMD's
runtime pm calls and after the $subject change it happened that exynos_dp/analogix_dp code called the clock enable/disable when FIMD driver (which implements the CRTC object) was not runtime resumed. Previously all dp clock handling was done when CRTC was enabled, thus the FIMD device was in the resumed runtime pm state.
Quoted 1 line
I will send a patch fixing this soon.
Alright, sounds like we don't need to revert the re-ordering patch after all =). Thanks for debugging this.
Tomi
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
In reply to Marek SzyprowskiQuoted 9 lines · Marek Szyprowski
I've finally found where is the problem! It turned out that the issue is in the
exynos_drm_fimddriver, in the code for enabling the display port clock (fimd_dp_clock_enable()
function). It touches FIMD regs, but it was not guarded with FIMD's
runtime pm calls and after the $subject change it happened that exynos_dp/analogix_dp code called the clock enable/disable when FIMD driver (which implements the CRTC object) was not runtime resumed. Previously all dp clock handling was done when CRTC was enabled, thus the FIMD device was in the resumed runtime pm state.
We have another issue the this patch: i.MX8MM (and MP) seem to have a problem with DSI, which is using the samsung-dsim.c driver. For me, the first time I run kmstest I get a black screen. The second time I get a display, but it's horizontally in the wrong place.
Perhaps you have an idea about this if you're familiar with the samsung-dsim?.
The issue seems to be that earlier the order was:
mxsfb_crtc_atomic_enable()samsung_dsim_atomic_pre_enable()samsung_dsim_atomic_enable()
now the order is:
samsung_dsim_atomic_pre_enable()mxsfb_crtc_atomic_enable()samsung_dsim_atomic_enable()
If I move the samsung_dsim_init() call from samsung_dsim_atomic_pre_enable() to samsung_dsim_atomic_enable(), it starts to work.
My guess is that the DSI init requires a pixel stream from the crtc, and now that the (pre)init is done without the pixel stream, it goes wrong. But this is purely a guess so far.
I also see that the samsung_dsim_init() call is behind "!samsung_dsim_hw_is_exynos()", so you probably don't hit this issue...
Tomi
Re: [PATCH v13 3/4] drm/atomic-helper: Re-order bridge chain pre-enable and post-disable
In reply to Aradhya BhatiaQuoted 4 lines · Aradhya Bhatia
From: Aradhya Bhatia <[email protected]>
Move the bridge
pre_enablecall before crtc enable, and the bridgepost_disablecall after the crtc disable.
This commit regresses the MCDE driver to a hard boot failure. Reverting the patch fixes the issue.
I think it has something to do with the internal DSI bridge in drivers/gpu/drm/mcde/mcde_dsi.c
Any hints to what needs to be fixed here?
Yours,
Linus Walleij
PATCH v13 4/4 drm/bridge: cdns-dsi: Use pre_enable/post_disable to enable/disable
In reply to Aradhya BhatiaFrom: Aradhya Bhatia <[email protected]>
The cdns-dsi controller requires that it be turned on completely before the input DPI's source has begun streaming[0]. Not having that, allows for a small window before cdns-dsi enable and after cdns-dsi disable where the previous entity (in this case tidss's videoport) to continue streaming DPI video signals. This small window where cdns-dsi is disabled but is still receiving signals causes the input FIFO of cdns-dsi to get corrupted. This causes the colors to shift on the output display. The colors can either shift by one color component (R->G, G->B, B->R), or by two color components (R->B, G->R, B->G).
Since tidss's videoport starts streaming via crtc enable hooks, we need cdns-dsi to be up and running before that. Now that the bridges are pre_enabled before crtc is enabled, and post_disabled after crtc is disabled, use the pre_enable and post_disable hooks to get cdns-dsi ready and running before the tidss videoport to get pass the color shift issues.
[0]: See section 12.6.5.7.3 "Start-up Procedure" in J721E SoC TRM
TRM Link: http://www.ti.com/lit/pdf/spruil1
--- .../gpu/drm/bridge/cadence/cdns-dsi-core.c | 64 ++++++++++--------- 1 file changed, 35 insertions(+), 29 deletions(-)
drivers/gpu/drm/bridge/cadence/cdns-dsi-core.c +35 -29
@@ -670,13 +670,28 @@ cdns_dsi_bridge_mode_valid(struct drm_bridge *bridge,return MODE_OK ; } -static void cdns_dsi_bridge_atomic_disable (struct drm_bridge *bridge, -struct drm_atomic_state *state) +static void cdns_dsi_bridge_atomic_post_disable (struct drm_bridge *bridge, +struct drm_atomic_state *state) {struct cdns_dsi_input *input =bridge_to_cdns_dsi_input (bridge);struct cdns_dsi *dsi =input_to_dsi (input);u32 val; +/* +* The cdns-dsi controller needs to be disabled after it's DPI source +* has stopped streaming. If this is not followed, there is a brief +* window before DPI source is disabled and after cdns-dsi controller +* has been disabled where the DPI stream is still on, but the cdns-dsi +* controller is not ready anymore to accept the incoming signals. This +* is one of the reasons why a shift in pixel colors is observed on +* displays that have cdns-dsi as one of the bridges. +* +* To mitigate this, disable this bridge from the bridge post_disable() +* hook, instead of the bridge _disable() hook. The bridge post_disable() +* hook gets called after the CRTC disable, where often many DPI sources +* disable their streams. +*/ + val =readl (dsi->regs +MCTL_MAIN_DATA_CTL ); val &= ~(IF_VID_SELECT_MASK |IF_VID_MODE |VID_EN |HOST_EOT_GEN |DISP_EOT_GEN );
@@ -688,15 +703,6 @@ static void cdns_dsi_bridge_atomic_disable(struct drm_bridge *bridge,if (dsi->platform_ops && dsi->platform_ops->disable) dsi->platform_ops->disable (dsi); -pm_runtime_put (dsi->base.dev); -} - -static void cdns_dsi_bridge_atomic_post_disable (struct drm_bridge *bridge, -struct drm_atomic_state *state) -{ -struct cdns_dsi_input *input =bridge_to_cdns_dsi_input (bridge); -struct cdns_dsi *dsi =input_to_dsi (input); - dsi->phy_initialized = false; dsi->link_initialized = false;phy_power_off (dsi->dphy);
@@ -774,8 +780,8 @@ static void cdns_dsi_init_link(struct cdns_dsi *dsi) dsi->link_initialized = true; } -static void cdns_dsi_bridge_atomic_enable (struct drm_bridge *bridge, -struct drm_atomic_state *state) +static void cdns_dsi_bridge_atomic_pre_enable (struct drm_bridge *bridge, +struct drm_atomic_state *state) {struct cdns_dsi_input *input =bridge_to_cdns_dsi_input (bridge);struct cdns_dsi *dsi =input_to_dsi (input);
@@ -792,6 +798,21 @@ static void cdns_dsi_bridge_atomic_enable(struct drm_bridge *bridge,u32 tmp, reg_wakeup, div, status;int nlanes; +/* +* The cdns-dsi controller needs to be enabled before it's DPI source +* has begun streaming. If this is not followed, there is a brief window +* after DPI source enable and before cdns-dsi controller enable where +* the DPI stream is on, but the cdns-dsi controller is not ready to +* accept the incoming signals. This is one of the reasons why a shift +* in pixel colors is observed on displays that have cdns-dsi as one of +* the bridges. +* +* To mitigate this, enable this bridge from the bridge pre_enable() +* hook, instead of the bridge _enable() hook. The bridge pre_enable() +* hook gets called before the CRTC enable, where often many DPI sources +* enable their streams. +*/ +if (WARN_ON (pm_runtime_get_sync (dsi->base.dev) <0 ))return ;
@@ -811,8 +832,8 @@ static void cdns_dsi_bridge_atomic_enable(struct drm_bridge *bridge, mode = &crtc_state->adjusted_mode; nlanes = output->dev->lanes; -cdns_dsi_hs_init (dsi);cdns_dsi_init_link (dsi); +cdns_dsi_hs_init (dsi);/* * Now that the DSI Link and DSI Phy are initialized,
@@ -941,19 +962,6 @@ static void cdns_dsi_bridge_atomic_enable(struct drm_bridge *bridge,writel (tmp, dsi->regs +MCTL_MAIN_EN ); } -static void cdns_dsi_bridge_atomic_pre_enable (struct drm_bridge *bridge, -struct drm_atomic_state *state) -{ -struct cdns_dsi_input *input =bridge_to_cdns_dsi_input (bridge); -struct cdns_dsi *dsi =input_to_dsi (input); - -if (WARN_ON (pm_runtime_get_sync (dsi->base.dev) <0 )) -return ; - -cdns_dsi_init_link (dsi); -cdns_dsi_hs_init (dsi); -} - static u32 *cdns_dsi_bridge_get_input_bus_fmts(struct drm_bridge *bridge, struct drm_bridge_state *bridge_state, struct drm_crtc_state *crtc_state,
@@ -1048,9 +1056,7 @@ cdns_dsi_bridge_atomic_reset(struct drm_bridge *bridge)static const struct drm_bridge_funcs cdns_dsi_bridge_funcs = { .attach = cdns_dsi_bridge_attach, .mode_valid = cdns_dsi_bridge_mode_valid, - .atomic_disable = cdns_dsi_bridge_atomic_disable, .atomic_pre_enable = cdns_dsi_bridge_atomic_pre_enable, - .atomic_enable = cdns_dsi_bridge_atomic_enable, .atomic_post_disable = cdns_dsi_bridge_atomic_post_disable, .atomic_check = cdns_dsi_bridge_atomic_check, .atomic_reset = cdns_dsi_bridge_atomic_reset,
Signature · 1 line
2.34.1
Hello Aradhya,
Thanks for the patch series.
Quoted 239 lines · Aradhya Bhatia
Hello all,
This series re-orders the sequences in which the drm CRTC and the drm Bridge get enabled and disabled with respect to each other.
The bridge
pre_enablecalls have been shifted before thecrtc_enableand the bridgepost_disablecalls have been shifted after thecrtc_disable.This has been done as per the definition of bridge
pre_enable. "The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".Since CRTC is also a source feeding the bridge, it should not be enabled before the bridges in the pipeline are
pre_enabled.The original sequence. for display pipe enable looks like:
crtc_enable bridge[n]_pre_enable ... bridge[1]_pre_enable encoder_enable bridge[1]_enable ... bridge[n]_enableThe sequence of enable after this patch-set will look like:
bridge[n]_pre_enable ... bridge[1]_pre_enable crtc_enable encoder_enable bridge[1]_enable ... bridge[n]_enableFor the disable sequence, this is what the original looks like:
bridge[n]_disable ... bridge[1]_disable encoder_disable bridge[1]_post_disable ... bridge[n]_post_disable crtc_disableThis is what the disable sequence will be, after this series of patches:
bridge[n]_disable ... bridge[1]_disable encoder_disable crtc_disable bridge[1]_post_disable ... bridge[n]_post_disableThis series further updates the bridge API definitions to accurately reflect the updated scenario.
This series is a subset of its v11[0] which had 14 patches in the revision. 9 of those 14 patches (which were specific to the cdns-dsi bridge driver) were merged[1].
Regards
Aradhya---
References:
[0]: Revision v11 of this series.
https://lore.kernel.org/all/[email protected]/[1]: Patches 1 through 9 getting merged.
https://lore.kernel.org/all/[email protected]/---
Change Log:- Changes in v13:
- Style changes in patch 2/4. (Thomas)
- Squash patch v12:4/5 into v12:3/5, which is now v13:3/4. (Thomas)
- Add R-b tags from Thomas Zimmermann in patches 1-3.
- Rebase onto latest drm-misc-next.- Changes in v12:
- Drop patches 1 through 9 since they have been merged.
- Rebase onto newer drm-misc-next.
- Re-word the patch 3/4, ("drm/bridge: Update the bridge enable/disable doc") to make it more readable.- Changes in v11:
- Add patch v11:13/14 ("drm/bridge: Update the bridge enable/disable doc"), that updates the documentation about the order of the various bridge enable/disable hooks being called wrt the CRTC and encoder hooks.
- Rebase on drm-misc-next instead of linux-next.
As part of rebase, accommodate the following change:
- Change patch v10:08/13 ("drm/bridge: cdns-dsi: Support atomic bridge APIs") to v11:08/13 ("drm/bridge: cdns-dsi: Add input format negotiation"), since Maxime has already updated the bridge hooks to their atomic versions in commit68c98e227a96("drm/bridge: cdns-csi:
Switch to atomic helpers").
My new patch now only adds the format negotiation hook for the cdns-dsi. (Note: Since the new patch is now only a subset of the old one, without any change in logic, I decided to carry forward the R-b and T-b tags.)
- Add Alexander Sverdlin's T-b in patches 10, 11, 12.- Changes in v10:
- Rebase on latest linux-next (next-20250226).
- As part of rebase, update the patches to accommodate a couple of widespread changes in DRM Framework -
- All the ("drm/atomic-helper: Change parameter name of ***") commits.
- All the ("drm/bridge: Pass full state to ***") commits.
(These updates are only trivial substitutions.)
- Add Tomi Valkeinen's T-b tags in all the patches.- Changes in v9:
- Fix the oops in 11/13 - where theencoder_bridge_enable_was_pre_enablingthe bridges instead of enabling.
- Add the following tags:
- Dmitry Baryshkov's R-b in patches 2, 10, 11, and A-b in patch 12.
- Jayesh Choudhary's R-b in patch 12.
- Tomi Valkeinen's R-b in patches 2, 10, 11, 12.- Changes in v8:
- Move the phy de-initialization to bridgepost_disable()instead of bridgedisable()in patch-3.
- Copy the private bridge state (dsi_cfg), in addition to thebridge_state, in patch-9.
- Split patch v7:11/12 into three patches, v8:{10,11,12}/13, to separate out different refactorings into different patches, and improve bisectability.
- Move patch v7:02/12 down to v8:06/12, to keep the initial patches for fixes only.
- Drop patch v7:04/12 as it doesn't become relevant until patch v7:12/12.
- Add R-b tags of Dmitry Baryshkov in patch-9 and patch-3, and of Tomi Valkeinen in patch-9.- Changes in v7:
- phy_init()/exit() were called from the PM path in v6. Change it back to the bridge enable/disable path in patch-3, so that thephy_init()can go back to being called after D-Phy reset assert.
- Reword commit text in patch-5 to explain the need of the fix.
- Drop the stray code in patch-10.
- Add R-b tag of Dmitry Baryshkov in patch-6.- Changes in v6:
- Reword patch 3 to better explain the fixes around phy de-init.
- Fix the Lane ready timeout condition in patch 7.
- Fix the dsi_bridge_atomic_check()implementation by adding a new bridge state structure in patch 10.
- Rework and combine patches v5:11/13 and v5:12/13 to v6:11/12.
- Generate the patches of these series using the "patience" algorithm.
Note: All patches, except v6:11/12, *do not* differ from their default (greedy) algorithm variants.
For patch 11, the patience algorithm significantly improves the readability.
- Rename and move the Bridge enable/disable enums from public to private in patch 11.
- Add R-b tags of Tomi Valkeinen in patch 6, and Dmitry Baryshkov in patch 2.- Changes in v5:
- Fix subject and description in patch 1/13.
- Add patch to check the return value ofphy_mipi_dphy_get_default_config()(patch: 6/13).
- Change the Clk and Data Lane ready timeout from forever to 5s.
- Print an error instead of callingWARN_ON_ONCEin patch 7/13.
- Drop patch v4-07/11: "drm/bridge: cdns-dsi: Reset the DCS write FIFO". There has been some inconsistencies found with this patch upon further testing. This patch was being used to enable a DSI panel based on ILITEK ILI9881C bridge. This will be debugged separately.
- Add patch to move the DSI mode check from_atomic_enable()to_atomic_check()(patch: 10/13).
- Split patch v4-10/11 into 2 patches - 11/13 and 12/13.
Patch 11/13 separates out the Encoder-Bridge operations into a helper function *without* changing the logic. Patch 12/13 then changes the order of the encoder-bridge operations as was intended in the original patch.
- Add detailed comment for patch 13/13.
- Add Tomi Valkeinen's R-b in patches 1, 2, 4, 5, 7, 8, 9, 13.- Changes in v4:
- Add new patch, "drm/bridge: cdns-dsi: Move todevm_drm_of_get_bridge()", to update to an auto-managed way of finding next bridge in the chain.
- Drop patch "drm/bridge: cdns-dsi: Fix thephy_initializedvariable" and add "drm/bridge: cdns-dsi: Fix Phy_init()and_exit()" that properly de-initializes the Phy and maintains the initialization state.
- Reword patch "drm/bridge: cdns-dsi: Reset the DCS write FIFO" to explain the HW concerns better.
- Add R-b tag from Dmitry Baryshkov for patches 1/11 and 8/11.- Changes in v3:
- Reword the commit message for patch "drm/bridge: cdns-dsi: Fix OF node pointer".
- Add a new helper API to figure out DSI host input pixel format in patch "drm/mipi-dsi: Add helper to find input format".
- Use a common function for bridge pre-enable and enable, and bridge disable and post-disable, to avoid code duplication.
- Add T-b tag from Dominik Haller in patch 5/10. (Missed to add it in v2).
- Add R-b tag from Dmitry Baryshkov for patch 8/10.- Changes in v2:
- Drop patch "drm/tidss: Add CRTCmode_fixup"
- Split patch "drm/bridge: cdns-dsi: Fix minor bugs" into 4 separate ones
- Drop support for early_enable/late_disable APIs and instead re-order thepre_enable/post_disableAPIs to be called before / aftercrtc_enable/crtc_disable.
- Drop support for early_enable/late_disable in cdns-dsi and use pre_enable/post_disable APIs instead to do bridge enable/disable.Previous versions:
v1: https://lore.kernel.org/all/[email protected]/ v2: https://lore.kernel.org/all/[email protected]/ v3: https://lore.kernel.org/all/[email protected]/ v4: https://lore.kernel.org/all/[email protected]/ v5: https://lore.kernel.org/all/[email protected]/ v6: https://lore.kernel.org/all/[email protected]/ v7: https://lore.kernel.org/all/[email protected]/ v8: https://lore.kernel.org/all/[email protected]/ v9: https://lore.kernel.org/all/[email protected]/ v10: https://lore.kernel.org/all/[email protected]/ v11: https://lore.kernel.org/all/[email protected]/ v12: https://lore.kernel.org/all/[email protected]/---
Quoted 8 lines
Aradhya Bhatia (4):
drm/atomic-helper: Refactor crtc & encoder-bridge op loops into
separate functions
drm/atomic-helper: Separate out bridge pre_enable/post_disable from
enable/disable
drm/atomic-helper: Re-order bridge chain pre-enable and post-disable drm/bridge: cdns-dsi: Use pre_enable/post_disable to enable/disable
For the series,
Regards
Devarsh
Quoted 7 lines
.../gpu/drm/bridge/cadence/cdns-dsi-core.c | 64 +++-- drivers/gpu/drm/drm_atomic_helper.c | 160 +++++++++-- include/drm/drm_bridge.h | 249 +++++++++++++----- 3 files changed, 355 insertions(+), 118 deletions(-)base-commit:
0b3d99425891e3c4a87259afb88fbd1168dc7707
Quoted 11 lines · Aradhya Bhatia
This series re-orders the sequences in which the drm CRTC and the drm Bridge get enabled and disabled with respect to each other.
The bridge
pre_enablecalls have been shifted before thecrtc_enableand the bridgepost_disablecalls have been shifted after thecrtc_disable.This has been done as per the definition of bridge
pre_enable. "The display pipe (i.e. clocks and timing signals) feeding this bridge will not yet be running when this callback is called".[...]
Applied to misc/kernel.git (drm-misc-next).
Thanks!
Maxime