linux-toradex.git/drivers/gpu/drm/mediatek, branch v6.9-rc4

Merge tag 'mediatek-drm-next-6.9' of https://git.kernel.org/pub/scm/linux/kernel/git/chunkuang.hu/linux into drm-next

2024-03-01T09:14:33+00:00

Mediatek DRM Next for Linux 6.9

1. Add display driver for MT8188 VDOSYS1
2. DSI driver cleanups
3. Filter modes according to hardware capability
4. Fix a null pointer crash in mtk_drm_crtc_finish_page_flip

Signed-off-by: Dave Airlie 

From: Chun-Kuang Hu 
Link: https://patchwork.freedesktop.org/patch/msgid/20240229162143.28957-1-chunkuang.hu@kernel.org

drm/mediatek: Fix a null pointer crash in mtk_drm_crtc_finish_page_flip

2024-02-29T16:15:28+00:00

It's possible that mtk_crtc->event is NULL in
mtk_drm_crtc_finish_page_flip().

pending_needs_vblank value is set by mtk_crtc->event, but in
mtk_drm_crtc_atomic_flush(), it's is not guarded by the same
lock in mtk_drm_finish_page_flip(), thus a race condition happens.

Consider the following case:

CPU1                              CPU2
step 1:
mtk_drm_crtc_atomic_begin()
mtk_crtc->event is not null,
                                  step 1:
                                  mtk_drm_crtc_atomic_flush:
                                  mtk_drm_crtc_update_config(
                                      !!mtk_crtc->event)
step 2:
mtk_crtc_ddp_irq ->
mtk_drm_finish_page_flip:
lock
mtk_crtc->event set to null,
pending_needs_vblank set to false
unlock
                                  pending_needs_vblank set to true,

                                  step 2:
                                  mtk_crtc_ddp_irq ->
                                  mtk_drm_finish_page_flip called again,
                                  pending_needs_vblank is still true
                                  //null pointer

Instead of guarding the entire mtk_drm_crtc_atomic_flush(), it's more
efficient to just check if mtk_crtc->event is null before use.

Fixes: 119f5173628a ("drm/mediatek: Add DRM Driver for Mediatek SoC MT8173.")
Signed-off-by: Hsin-Yi Wang 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240223212404.3709690-1-hsinyi@chromium.org/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: Filter modes according to hardware capability

2024-02-27T00:02:55+00:00

We found a stability issue on MT8188 when connecting an external monitor
in 2560x1440@144Hz mode. Checked with the designer, there is a function
called "prefetch" which is working during VBP (triggered by VSYNC).
If the duration of VBP is too short, the throughput requirement could
increase more than 3 times and lead to stability issues.

The mode settings that VDOSYS supports are mainly affected by clock
rate and throughput, display driver should filter these settings
according to the SoC's limitation to avoid unstable conditions.

Since currently the mode filter is only available on MT8195 and MT8188
and they share the same compatible name, the reference number (8250)
is hard coded instead of in the driver data.

Signed-off-by: Hsiao Chien Sung 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240220093711.20546-2-shawn.sung@mediatek.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Use mipi_dsi_pixel_format_to_bpp() helper function

2024-02-19T14:49:51+00:00

Instead of open coding, use the mipi_dsi_pixel_format_to_bpp() helper
function from drm_mipi_dsi.h in mtk_dsi_poweron() and for validation
in mtk_dsi_bridge_mode_valid().

Note that this function changes the behavior of this driver: previously,
in case of unknown formats, it would (wrongly) assume that it should
account for a 24-bits format - now it will return an error and refuse
to set clocks and/or enable the DSI.

This is done because setting the wrong data rate will only produce a
garbage output that the display will misinterpret both because this
driver doesn't actually provide any extra-spec format support and/or
because the data rate (hence, the HS clock) will be wrong.

Reviewed-by: Alexandre Mergnat 
Signed-off-by: AngeloGioacchino Del Regno 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-10-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Compress of_device_id entries and add sentinel

2024-02-19T14:48:22+00:00

All entries fit in 82 columns, which is acceptable: compress all of
the mtk_dsi_of_match[] entries to a single line for each.

While at it, also add the usual sentinel comment to the last entry.

Reviewed-by: Alexandre Mergnat 
Signed-off-by: AngeloGioacchino Del Regno 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-9-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Simplify with dev_err_probe and remove gotos

2024-02-19T14:46:46+00:00

Most of the functions that are called in the probe callback are
devm managed, or all but mipi_dsi_host_register(): simplify the probe
function's error paths with dev_err_probe() and remove the lonely
instance of `goto err_unregister_host` by just directly calling the
mipi_dsi_host_unregister() function in the devm_request_irq() error
path, allowing to also remove the same label.

Reviewed-by: Alexandre Mergnat 
Signed-off-by: AngeloGioacchino Del Regno 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-8-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Register DSI host after acquiring clocks and PHY

2024-02-19T14:45:06+00:00

Registering the dsi host with its ops before getting dsi->regs is
simply wrong: even though there's nothing (for now) asynchronously
calling those ops before the end of the probe function, installing
ops that are using iospace(s) and clocks before even initializing
those is too fragile.

Register the DSI host after getting clocks, iospace and PHY.
This wil also allow to simplify the error paths in a later commit.

Reviewed-by: Alexandre Mergnat 
Signed-off-by: AngeloGioacchino Del Regno 
Reviewed-by: CK Hu 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-7-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Replace open-coded instance of HZ_PER_MHZ

2024-02-19T14:43:30+00:00

In mtk_dsi_phy_timconfig(), we're dividing the `data_rate` variable,
expressed in Hz to retrieve a value in MHz: instead of open-coding,
use the HZ_PER_MHZ definition, available in linux/units.h.

Reviewed-by: Alexandre Mergnat 
Reviewed-by: CK Hu 
Signed-off-by: AngeloGioacchino Del Regno 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-6-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Use bitfield macros where useful

2024-02-19T14:42:01+00:00

Instead of open coding bitshifting for various register fields,
use the bitfield macro FIELD_PREP(): this allows to enhance the
human readability, decrease likeliness of mistakes (and register
field overflowing) and also to simplify the code.
The latter is especially seen in mtk_dsi_rxtx_control(), where
it was possible to change a switch to a short for loop and to
also remove the need to check for maximum DSI lanes == 4 thanks
to the FIELD_PREP macro masking the value.

While at it, also add the missing DA_HS_SYNC bitmask, used in
mtk_dsi_phy_timconfig().

Reviewed-by: Alexandre Mergnat 
Reviewed-by: CK Hu 
Signed-off-by: AngeloGioacchino Del Regno 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-5-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu

drm/mediatek: dsi: Cleanup functions mtk_dsi_ps_control{_vact}()

2024-02-19T14:36:19+00:00

Function mtk_dsi_ps_control() is a subset of mtk_dsi_ps_control_vact():
merge the two in one mtk_dsi_ps_control() function by adding one
function parameter `config_vact` which, when true, writes the VACT
related registers.

Reviewed-by: Fei Shao 
Reviewed-by: Alexandre Mergnat 
Reviewed-by: CK Hu 
Signed-off-by: AngeloGioacchino Del Regno 
Link: https://patchwork.kernel.org/project/dri-devel/patch/20240215085316.56835-4-angelogioacchino.delregno@collabora.com/
Signed-off-by: Chun-Kuang Hu