kernel/linux.git/drivers/gpu/drm/msm/dsi, branch linux-5.3.y

drm/msm/dsi: Fix return value check for clk_get_parent

2019-10-11T16:36:14+00:00

commit 5fb9b797d5ccf311ae4aba69e86080d47668b5f7 upstream. clk_get_parent returns an error pointer upon failure, not NULL. So the checks as they exist won't catch a failure. This patch changes the checks and the return values to properly handle an error pointer. Fixes: c4d8cfe516dc ("drm/msm/dsi: add implementation for helper functions") Cc: Sibi Sankar Cc: Sean Paul Cc: Rob Clark Cc: # v4.19+ Signed-off-by: Sean Paul Signed-off-by: Rob Clark Signed-off-by: Greg Kroah-Hartman

Merge tag 'drm-msm-next-2019-06-25' of https://gitlab.freedesktop.org/drm/msm into drm-next

2019-06-28T00:16:40+00:00

+ usual progress on cleanups + dsi vs EPROBE_DEFER fixes + msm8998 (snapdragon 835 support) + a540 gpu support (mesa support already landed) + dsi, dsi-phy support + mdp5 and dpu interconnect (bus/memory scaling) support + initial prep work for per-context pagetables (at least the parts that don't have external dependencies like iommu/arm-smmu) There is one more patch for fixing DSI cmd mode panels (part of a set of patches to get things working on nexus5), but it would be conflicty with 1cff7440a86e04a613665803b42034 in drm-next without rebasing or back-merge, and since it doesn't conflict with anything in msm-next, I think it best if Sean merges that through drm-mix-fixes instead. (In other news, I've been making some progress w/ getting efifb working properly on sdm850 laptop without horrible hacks, and drm/msm + clk stuff not totally falling over when bootloader enables display and things are already running when driver probes.. but not quite ready yet, hopefully we can post some of that for 5.4.. should help for both the sdm835 and sdm850 laptops.) Signed-off-by: Dave Airlie From: Rob Clark Link: https://patchwork.freedesktop.org/patch/msgid/CAF6AEGsj3N4XzDLSDoa+4RHZ9wXObYmhcep0M3LjnRg48BeLvg@mail.gmail.com

drm/msm/dsi: Move setup_encoder to modeset_init

2019-06-20T18:42:19+00:00

Now that the panel probe/setup is in the modeset path, we can call dsi_manager_setup_encoder() in a common place for both internal and external bridge setups. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-10-sean@poorly.run

drm/msm/dsi: Move dsi panel init into modeset init path

2019-06-20T18:42:17+00:00

Since deferred probe from the modeset init path now works, we can move the panel initialization from detect() into connector init. This avoids doing work in detect() and hopefully will result in a more deterministic boot sequence between devices with a dsi panel, and those with an external bridge. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-9-sean@poorly.run

drm/msm/dsi: Use the new setup_encoder function in attach_dsi_device

2019-06-20T18:42:15+00:00

Now that we have a function to call set_encoder_mode() for us, use it. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-8-sean@poorly.run

drm/msm/dsi: Simplify the logic in msm_dsi_manager_panel_init()

2019-06-20T18:42:13+00:00

This patch moves things around a bit to be a little more readable and pulls out the set_encoder_mode() call into its own function for later use. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-7-sean@poorly.run

drm/msm/dsi: Pull out panel init code into function

2019-06-20T18:42:11+00:00

Pull all of the panel init code out of detect() and put it in its own function. This will be useful in future patches where it's moved from detect(). Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-6-sean@poorly.run

drm/msm/dsi: Don't store dsi host mode_flags in msm_dsi

2019-06-20T18:42:09+00:00

It's a bit dangerous to store the flags in msm_dsi since there's no way to tell when they're populated. Fortunately the only place that uses them is the same place that fills them. So just use a local variable and delete the struct member. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-5-sean@poorly.run

drm/msm/dsi: Split mode_flags out of msm_dsi_host_get_panel()

2019-06-20T18:42:06+00:00

We use the flags in more places than just get_panel, so split them out into a separate function. Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-4-sean@poorly.run

drm/msm/dsi_pll_10nm: Remove impossible check

2019-06-20T18:42:03+00:00

While I'm in here, cut this out, pdev can't be NULL Reviewed-by: Rob Clark Signed-off-by: Sean Paul Link: https://patchwork.freedesktop.org/patch/msgid/20190617201301.133275-2-sean@poorly.run