08:31 <lucaceresoli> hi, following the discussion on https://www.spinics.net/lists/linux-omap/msg159409.html ...

08:32 <lucaceresoli> I'm currently facing the lack of /dev/fb0 after upgrading from ti/ti-linux-4.19.y to ti/ti-linux-5.10.y

08:32 <lucaceresoli> Any hints on what I should check first?

08:55 <lucaceresoli> my panel driver (not yet published) gets probed; in DT it has ports/port@0/endpoint/remote-endpoint = <&dpi_out>, basically the same as the LCD in am57xx-evm-common.dts from ti-linux-5.10.y

08:55 <lucaceresoli> Should I take am57xx-evm-common.dts from ti-linux-5.10.y as working code so I can use it as an example?

11:57 <NiksDev> lucaceresoli, Make sure following configs are enabled:

11:57 <NiksDev> CONFIG_DRM_KMS_HELPER=y

11:57 <NiksDev> CONFIG_DRM_KMS_FB_HELPER=y

11:57 <NiksDev> CONFIG_DRM_FBDEV_OVERALLOC=100

11:57 <NiksDev> CONFIG_DRM_FBDEV_EMULATION=y

11:59 <NiksDev> Once you have the basic DRM functionality working (can be verified by running modetest), you can enable the FBDEV emulation to get /dev/fb0. No special change in DTS is required

12:33 <lucaceresoli> NiksDev, thanks. All the config parameters are there, but...

12:33 <lucaceresoli> # modetest

12:33 <lucaceresoli> [...]

12:33 <lucaceresoli> no device found

12:33 <lucaceresoli> #

12:34 <lucaceresoli> dmesg | grep -E '(>>>|omapdrm|OF:)'

12:34 <lucaceresoli> (ignore the last line, bad copy paste)

12:36 <lucaceresoli> yet the omapdrm driver is instantiated, dss_probe() returns 0... I'm not sure I understood exactly which part of the omapdrm driver I should see enabled

13:45 <NiksDev> I would suggest to enable dynamic debugs and check the prints

13:46 <NiksDev> first step is to get the modetest running, fbdev gets automatically enabled after that if above configs are present

13:52 <lucaceresoli> NiksDev, OK, I'm now building with DYNAMIC_DEBUG and omapdrm as a module...

14:32 <lucaceresoli> NiksDev, with dynamic debug I'm not getting much info :(

14:32 <lucaceresoli> After:

14:33 <lucaceresoli> echo 'file drivers/gpu/drm/omapdrm/* +p' >/sys/kernel/debug/dynamic_debug/control

14:33 <lucaceresoli> echo 'module *drm* +p' >/sys/kernel/debug/dynamic_debug/control

14:33 <lucaceresoli> modprobe omapdrm

14:33 <lucaceresoli> ...all I get in dmesg is:

14:33 <lucaceresoli> [ 117.035517] dmm 4e000000.dmm: workaround for errata i878 in use

14:33 <lucaceresoli> [ 117.052333] DSS: dss_restore_context

14:33 <lucaceresoli> [ 117.046840] dmm 4e000000.dmm: initialized all PAT entries

14:33 <lucaceresoli> [ 117.052340] DSS: context restored

14:33 <lucaceresoli> [ 117.053749] DSS: dss_save_context

14:33 <lucaceresoli> [ 117.053756] DSS: context saved

14:34 <lucaceresoli> What would you expect to find in the logs?

14:34 <lucaceresoli> (I'm now building the same setup on the 4.19 branch to see what is different there)

14:44 <lucaceresoli> nonono, wait, I _do_ have more messages in dmesg, but they happen while booting, not after modprobe. They must be triggered by the LCD controller which is described in DT:

14:45 <lucaceresoli> [ 9.991038] omapdss_dss 58000000.dss: trying to bring up master

14:45 <lucaceresoli> [ 9.991047] omapdss_dss 58000000.dss: Looking for component 0

14:45 <lucaceresoli> [ 9.991054] omapdss_dss 58000000.dss: master has incomplete components

14:45 <lucaceresoli> [ 10.040190] omapdss_dispc 58001000.dispc: adding component (ops dispc_component_ops [omapdss])

14:45 <lucaceresoli> [ 10.040204] omapdss_dss 58000000.dss: Looking for component 0

14:45 <lucaceresoli> [ 10.040215] omapdss_dss 58000000.dss: Looking for component 1

14:45 <lucaceresoli> [ 10.040210] omapdss_dss 58000000.dss: found component 58001000.dispc, duplicate 0

14:45 <lucaceresoli> [ 10.040198] omapdss_dss 58000000.dss: trying to bring up master

14:45 <lucaceresoli> [ 10.040220] omapdss_dss 58000000.dss: master has incomplete components

14:45 <lucaceresoli> does it mean anything to you?

14:49 <lucaceresoli> NiksDev, looks like omapdss gets enabled, not omapdrm. I'm studying the relationship between the two.

17:26 <NishanthMenon> lucaceresoli, NiksDev works in IST timezone.. just a headsup

17:27 <NishanthMenon> lucaceresoli, also a future reference, please use https://pastebin.ubuntu.com/ => unfortunately our TI IT security folks have blocked pastebin.com and hastebin.com among others.. :( I updated the $topic to reflect that

17:34 <NishanthMenon> kristo, https://libera.irclog.whitequark.org/beagle/2021-05-28#29978567; -> exactly what we spoke about today morning.. :)

17:35 <NishanthMenon> it is a recurring theme

18:22 <lucaceresoli> 3w00DguyB

18:28 <NishanthMenon> lucaceresoli, ^^ please change paswd

18:28 <lucaceresoli> already done... my bad! thanks

18:29 <lucaceresoli> today I'm typing in the wrong place all the time :(