(cc Jocelyn, dri-devel)
Am 08.06.26 um 17:30 schrieb Guilherme Bember Duarte:
I don't know what exactly happened but the update to 7.0.11 might have
something to do with it (or possibly some stupid mistake on my part).
Now I can switch to other TTYs where kmscon spawns fine (before that
there's still no display initialisation, verifiable by checking
"/sys/class/drm/card0-*/status", all return unknown), but with or
without that boot argument, TTY 1 stays dead instead of the usual logger.
Switching to TTY 1 again makes the screen hang until i switch back and
on boot I get no signal on the display until I switch over to another TTY.
On 8 June 2026 07:52:59 GMT-03:00, Thomas Zimmermann
<[email protected]> wrote:
Hi, Am 06.06.26 um 16:49 schrieb Guilherme Bember Duarte:
Hi! I replaced DRM_FBDEV_EMULATION with DRM_CLIENT_LOG
(DRM_CLIENT_DEFAULT="log") on a kernel with DRM_PANIC enabled,
using kmscon as the userspace terminal. Nothing initialises
the display after the native driver loads. The result is a
blank screen, no output at all from boot. DRM_CLIENT_LOG (boot
logs don't show up at all) and DRM_PANIC (forced a panic to
try this out) also don't work, since neither can render
without an initialised display. This can be verified by
checking "/sys/class/drm/card0-*/status" (all outputs show
"unknown") and "/sys/class/drm/card0-*/modes" (empty) after
boot. Writing "detect" to
"/sys/class/drm/card0-HDMI-A-1/status" (or any output for that
matter) resolves the issue, the connector is probed, modes are
populated, and kmscon can be started correctly. This suggests
DRM_FBDEV_EMULATION previously triggered some display
initialisation/probing during boot that is not happening with
DRM_CLIENT_LOG. I can't say if SIMPLEDRM (and SIMPLEFB) are
affected, the result in my system is a immediate blank screen
and no signal, I don't know if its because the main graphics
driver initialise very quickly during boot or if SIMPLEDRM is
also affected.
Did you boot with "console=drm_log" at the kernel command line?
Best regards Thomas
Hardware: AMD RX 6600 XT (Navi23/AMDGPU), although I believe
this affects any driver. Kernel version: 7.0.11-gentoo Config
diff: 2849a2850
# CONFIG_VGA_SWITCHEROO is not set
2874,2879c2875,2878 < CONFIG_DRM_FBDEV_EMULATION=y <
CONFIG_DRM_FBDEV_OVERALLOC=100 < #
CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM is not set < #
CONFIG_DRM_CLIENT_LOG is not set <
CONFIG_DRM_CLIENT_DEFAULT_FBDEV=y <
CONFIG_DRM_CLIENT_DEFAULT="fbdev"
------------------------------------------------------------------------
# CONFIG_DRM_FBDEV_EMULATION is not set
CONFIG_DRM_CLIENT_LOG=y CONFIG_DRM_CLIENT_DEFAULT_LOG=y
CONFIG_DRM_CLIENT_DEFAULT="log"
2978,2986d2976 < CONFIG_FB_CORE=y < # CONFIG_FB_DEVICE is not
set < CONFIG_FB_SYS_FILLRECT=y < CONFIG_FB_SYS_COPYAREA=y <
CONFIG_FB_SYS_IMAGEBLIT=y < CONFIG_FB_SYSMEM_FOPS=y <
CONFIG_FB_DEFERRED_IO=y < CONFIG_FB_SYSMEM_HELPERS=y <
CONFIG_FB_SYSMEM_HELPERS_DEFERRED=y 3021,3024d3010 <
CONFIG_FRAMEBUFFER_CONSOLE=y < #
CONFIG_FRAMEBUFFER_CONSOLE_LEGACY_ACCELERATION is not set <
CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y < #
CONFIG_FRAMEBUFFER_CONSOLE_ROTATION is not set 3027d3012 < #
CONFIG_LOGO is not set 4473d4457 < # CONFIG_FONT_SUN8x16 is
not set Guilherme B.
Guilherme B.
--
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
