Hi,
Sorry to post usage questions here, but I didn't find a better place.
I used libdrm for my wayland based compositor, it generally works but I've been experiencing hiccups here and there. I have a few questions. Most of them are generic and one of them is specific.
1. Is there any way to debug when the functions like drmModeSetCrtc, drmModePageFlip and drmModeAtomicCommit returns non zero, besides getting the errno?
2. Do I need to attach a 0 fb_id and 0 crtc_id if I intend to deactivate the connector? It seems I don't need to do it in legacy drm code, and I am required to do it using atomic functions.
3. I tried what I described above for returning to tty, however I got -ENIVAL on return(maybe it is because I didn't wait for vblank).
4. Do I need to wait for vblank for modesetting? I know it will be an error if I tried to do drmModePageFlip before the last one is handled but what about drmModeSetCrtc? Do I need to wait for a vblank for the last pageflip?
Is it actually valid to set a 0 crtc_id for connector and 0 fb_id for crtc?
Sorry I didn't place the actual code here, my application is already rather complex and a lot of abstraction on top of it.
Best regards, Sichem
On 2021-06-05 6:06 a.m., Sichem Zhou wrote:
- Is there any way to debug when the functions like drmModeSetCrtc,
drmModePageFlip and drmModeAtomicCommit returns non zero, besides getting the errno?
Enable appropriate debugging bits in /sys/module/drm/parameters/debug, reproduce problem, (disable debugging bits again), check dmesg.
- Do I need to wait for vblank for modesetting? I know it will be an
error if I tried to do drmModePageFlip before the last one is handled but what about drmModeSetCrtc? Do I need to wait for a vblank for the last pageflip?
Right now, if you don't wait for any pending page flip to complete before drmModeSetCrtc, the two may race, and the CRTC may end up scanning out from the FB set by the page flip.
Maybe this race should be prevented in the kernel though.
There's no need to wait for vblank.
I have to defer to others for your other questions.
Hi Michel,
Thanks for your answer, I just enabled the debug and captured a drm debug log from dmesg, but I don't seem to find anything that looks like an error, Is there anything specific I should be looking for?
Sorry for attaching my log here, here is the last drmModeAtomicCommit where I had -einval.
[ 56.005133] [drm:drm_atomic_state_init [drm]] Allocated atomic state 00000000c20a40d6 [ 56.005218] [drm:drm_mode_object_get [drm]] OBJ ID: 97 (3) [ 56.005297] [drm:drm_mode_object_get [drm]] OBJ ID: 97 (4) [ 56.005373] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:31:plane 1A] 0000000040b2d9eb state to 00000000c20a40d6 [ 56.005457] [drm:drm_mode_object_get [drm]] OBJ ID: 122 (1) [ 56.005534] [drm:drm_atomic_get_crtc_state [drm]] Added [CRTC:51:pipe A] 0000000073659c18 state to 00000000c20a40d6 [ 56.005611] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:39:plane 2A] 0000000071f640c7 state to 00000000c20a40d6 [ 56.005687] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:39:plane 2A] state 0000000071f640c7 [ 56.005765] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:47:cursor A] 000000000ebe824b state to 00000000c20a40d6 [ 56.005839] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:47:cursor A] state 000000000ebe824b [ 56.005911] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:52:plane 1B] 00000000b6f083af state to 00000000c20a40d6 [ 56.005983] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:60:plane 2B] 00000000243f0105 state to 00000000c20a40d6 [ 56.006055] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:60:plane 2B] state 00000000243f0105 [ 56.006124] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:68:cursor B] 00000000c385344b state to 00000000c20a40d6 [ 56.006196] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:68:cursor B] state 00000000c385344b [ 56.006264] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:73:plane 1C] 000000006229658a state to 00000000c20a40d6 [ 56.006336] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:81:plane 2C] 0000000065fc3015 state to 00000000c20a40d6 [ 56.006407] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:81:plane 2C] state 0000000065fc3015 [ 56.006475] [drm:drm_atomic_get_plane_state [drm]] Added [PLANE:89:cursor C] 00000000c5c2644d state to 00000000c20a40d6 [ 56.006546] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:89:cursor C] state 00000000c5c2644d [ 56.006618] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [FB:97] for [PLANE:31:plane 1A] state 0000000040b2d9eb [ 56.006686] [drm:drm_mode_object_get [drm]] OBJ ID: 97 (5) [ 56.006763] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 97 (6) [ 56.006858] [drm:drm_atomic_add_affected_connectors [drm]] Adding all current connectors for [CRTC:51:pipe A] to 00000000c20a40d6 [ 56.006942] [drm:drm_mode_object_get [drm]] OBJ ID: 95 (4) [ 56.007019] [drm:drm_mode_object_get [drm]] OBJ ID: 95 (5) [ 56.007094] [drm:drm_atomic_get_connector_state [drm]] Added [CONNECTOR:95:eDP-1] 0000000066710aab state to 00000000c20a40d6 [ 56.007172] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 95 (5) [ 56.007250] i915 0000:00:02.0: [drm:drm_atomic_set_crtc_for_connector [drm]] Link [CONNECTOR:95:eDP-1] state 0000000066710aab to [NOCRTC] [ 56.007321] [drm:drm_mode_object_get [drm]] OBJ ID: 95 (4) [ 56.007397] i915 0000:00:02.0: [drm:drm_atomic_set_crtc_for_connector [drm]] Link [CONNECTOR:95:eDP-1] state 0000000066710aab to [CRTC:51:pipe A] [ 56.007471] [drm:drm_atomic_get_crtc_state [drm]] Added [CRTC:72:pipe B] 0000000065ce9351 state to 00000000c20a40d6 [ 56.007548] i915 0000:00:02.0: [drm:drm_atomic_set_mode_for_crtc [drm]] Set [NOMODE] for [CRTC:72:pipe B] state 0000000065ce9351 [ 56.007618] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:52:plane 1B] state 00000000b6f083af [ 56.007684] [drm:drm_atomic_add_affected_connectors [drm]] Adding all current connectors for [CRTC:72:pipe B] to 00000000c20a40d6 [ 56.007765] [drm:drm_atomic_get_crtc_state [drm]] Added [CRTC:93:pipe C] 0000000067f37a81 state to 00000000c20a40d6 [ 56.007838] i915 0000:00:02.0: [drm:drm_atomic_set_mode_for_crtc [drm]] Set [NOMODE] for [CRTC:93:pipe C] state 0000000067f37a81 [ 56.007907] i915 0000:00:02.0: [drm:drm_atomic_set_fb_for_plane [drm]] Set [NOFB] for [PLANE:73:plane 1C] state 000000006229658a [ 56.007975] [drm:drm_atomic_add_affected_connectors [drm]] Adding all current connectors for [CRTC:93:pipe C] to 00000000c20a40d6 [ 56.008051] [drm:drm_atomic_check_only [drm]] checking 00000000c20a40d6 [ 56.008133] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] Updating routing for [CONNECTOR:95:eDP-1] [ 56.008184] [drm:drm_atomic_helper_check_modeset [drm_kms_helper]] [CONNECTOR:95:eDP-1] keeps [ENCODER:94:DDI A/PHY A], now on [CRTC:51:pipe A] [ 56.008232] [drm:drm_atomic_add_encoder_bridges [drm]] Adding all bridges for [encoder:94:DDI A/PHY A] to 00000000c20a40d6 [ 56.008311] [drm:drm_atomic_add_encoder_bridges [drm]] Adding all bridges for [encoder:94:DDI A/PHY A] to 00000000c20a40d6 [ 56.008390] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 97 (5) [ 56.008469] [drm:drm_mode_object_get [drm]] OBJ ID: 97 (4) [ 56.008554] i915 0000:00:02.0: [drm:intel_plane_atomic_calc_changes [i915]] [CRTC:51:pipe A] with [PLANE:31:plane 1A] visible 1 -> 1, off 0, on 0, ms 0 [ 56.008824] i915 0000:00:02.0: [drm:intel_atomic_get_global_obj_state [i915]] Added new global object 000000008a949845 state 00000000a680ff36 to 00000000c20a40d6 [ 56.009059] i915 0000:00:02.0: [drm:intel_atomic_get_global_obj_state [i915]] Added new global object 00000000db701695 state 0000000099e760e4 to 00000000c20a40d6 [ 56.009292] [drm:drm_atomic_commit [drm]] committing 00000000c20a40d6 [ 56.009430] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,1029)@ 55.633486 -> 55.627611 [e 0 us, 0 rep] [ 56.009519] i915 0000:00:02.0: [drm:drm_vblank_restore [drm]] missed 38 vblanks in 263172022 ns, frame duration=6943097 ns, hw_diff=38 [ 56.009599] i915 0000:00:02.0: [drm:drm_vblank_enable [drm]] enabling vblank on crtc 0, ret: 0 [ 56.009675] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,1071)@ 55.633731 -> 55.627616 [e 0 us, 0 rep] [ 56.009755] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-131)@ 55.633811 -> 55.634559 [e 0 us, 0 rep] [ 56.009831] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7575, diff=1, hw=8436 hw_last=8435 [ 56.009917] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-103)@ 55.633970 -> 55.634558 [e 0 us, 0 rep] [ 56.010025] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7576, diff=0, hw=8436 hw_last=8436 [ 56.010112] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-68)@ 55.634168 -> 55.634556 [e 0 us, 0 rep] [ 56.010191] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7576, diff=0, hw=8436 hw_last=8436 [ 56.016771] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-118)@ 55.640809 -> 55.641483 [e 0 us, 0 rep] [ 56.016798] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7576, diff=1, hw=8437 hw_last=8436 [ 56.016814] i915 0000:00:02.0: [drm:drm_handle_vblank [drm]] vblank event on 7577, current 7577 [ 56.016907] [drm:drm_atomic_state_default_clear [drm]] Clearing atomic state 00000000c20a40d6 [ 56.016927] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 95 (5) [ 56.016945] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 95 (4) [ 56.016964] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 122 (2) [ 56.016981] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 97 (5) [ 56.016997] [drm:drm_mode_object_put.part.0 [drm]] OBJ ID: 97 (4) [ 56.017015] [drm:__drm_atomic_state_free [drm]] Freeing atomic state 00000000c20a40d6 [ 56.017034] [drm:drm_release_noglobal [drm]] driver lastclose completed [ 56.023639] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-131)@ 55.647660 -> 55.648408 [e 0 us, 0 rep] [ 56.023668] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7577, diff=1, hw=8438 hw_last=8437 [ 56.023688] i915 0000:00:02.0: [drm:vblank_disable_fn [drm]] disabling vblank on crtc 0 [ 56.023706] i915 0000:00:02.0: [drm:drm_crtc_vblank_helper_get_vblank_timestamp_internal [drm]] crtc 0 : v p(0,-119)@ 55.647729 -> 55.648408 [e 0 us, 0 rep] [ 56.023725] i915 0000:00:02.0: [drm:drm_update_vblank_count [drm]] updating vblank count on crtc 0: current=7578, diff=0, hw=8438 hw_last=8438 [ 57.924428] i915 0000:00:02.0: [drm:intel_pps_vdd_off_sync_unlocked [i915]] Turning [ENCODER:94:DDI A/PHY A] VDD off [ 57.924748] i915 0000:00:02.0: [drm:intel_pps_vdd_off_sync_unlocked [i915]] PP_STATUS: 0x80000008 PP_CONTROL: 0x00000067 [ 57.924956] i915 0000:00:02.0: [drm:__intel_display_power_put_domain [i915]] disabling DC off [ 57.925163] i915 0000:00:02.0: [drm:skl_enable_dc6 [i915]] Enabling DC6 [ 57.925361] i915 0000:00:02.0: [drm:gen9_set_dc_state [i915]] Setting DC state from 00 to 02
Best regards, Sichem
On Mon, Jun 7, 2021 at 4:32 PM Michel Dänzer michel@daenzer.net wrote:
On 2021-06-05 6:06 a.m., Sichem Zhou wrote:
- Is there any way to debug when the functions like drmModeSetCrtc,
drmModePageFlip and drmModeAtomicCommit returns non zero, besides getting the errno?
Enable appropriate debugging bits in /sys/module/drm/parameters/debug, reproduce problem, (disable debugging bits again), check dmesg.
- Do I need to wait for vblank for modesetting? I know it will be an
error if I tried to do drmModePageFlip before the last one is handled but what about drmModeSetCrtc? Do I need to wait for a vblank for the last pageflip?
Right now, if you don't wait for any pending page flip to complete before drmModeSetCrtc, the two may race, and the CRTC may end up scanning out from the FB set by the page flip.
Maybe this race should be prevented in the kernel though.
There's no need to wait for vblank.
I have to defer to others for your other questions.
-- Earthling Michel Dänzer | https://redhat.com Libre software enthusiast | Mesa and X developer
- Do I need to wait for vblank for modesetting? I know it will be an
error if I tried to do drmModePageFlip before the last one is handled but what about drmModeSetCrtc? Do I need to wait for a vblank for the last pageflip?
Right now, if you don't wait for any pending page flip to complete before drmModeSetCrtc, the two may race, and the CRTC may end up scanning out from the FB set by the page flip.
Maybe this race should be prevented in the kernel though.
Maybe this could happen on non-atomic drivers, but on atomic drivers there's no race, the setcrtc will wait for the previous pageflip to complete first. Ofc if you call both from 2 threads at the same time then either could win, but if you do one after the other there's no ordering issues with atomic drivers. -Daniel
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Michel Dänzer -
Sichem Zhou