Opened 2 years ago

Last modified 2 years ago

#18084 new bug

[intel_extreme] Black screen (lit)

Reported by: thebuck Owned by: pulkomandy
Priority: normal Milestone: Unscheduled
Component: Drivers/Graphics/intel_extreme/broadwell Version: R1/beta4
Keywords: Cc:
Blocked By: Blocking:
Platform: All

Description (last modified by thebuck)

R1beta4tc0, EFI boot, x86_64.

After all icons have lit up (native resolution), briefly shows the same picture with wrong stride (not sure if it is THAT picture, of course), then black, lit.

Intel HD Graphics 5300 (rev 09), PCI-ID 8086:161e (from lspci on Linux)

Internal panel: 1920x1080, capable of 60 and 120 Hz (with Linux)

Micro-HDMI (D type) socket present, but I have no cable for it.

Attachments (4)

syslog.redacted.txt (25.1 KB ) - added by thebuck 2 years ago.
syslog, only intel_extreme-blocks.
card0-eDP-1.bin (128 bytes ) - added by thebuck 2 years ago.
Linux /sys/class/drm/card0-eDP-1/edid
i915_frequency_info.txt (990 bytes ) - added by thebuck 2 years ago.
Linux /sys/kernel/debug/dri/0/i915_frequency_info
syslog-hrev56605-5830-1.redacted.txt (24.6 KB ) - added by thebuck 2 years ago.
No change in behaviour :(

Download all attachments as: .zip

Change History (13)

by thebuck, 2 years ago

Attachment: syslog.redacted.txt added

syslog, only intel_extreme-blocks.

comment:1 by thebuck, 2 years ago

or Haswell? Both names appear in the log...

comment:2 by thebuck, 2 years ago

Graphics used to work on this device (beta3, and some nightlies until halfway to beta4 at least), but I do not remember whether that was running on framebuffer automatically.

So this may be a regression.

comment:3 by korli, 2 years ago

Broadwell is correct. No edid is found, only the resolution in the VBt is found. The driver isn't really able to switch to a non native resolution it seems. So it's supposed to be connected with eDP.

What's the exact PCI id for the graphics de vice?

comment:4 by thebuck, 2 years ago

Description: modified (diff)

PCI-ID and connector information added.

comment:5 by korli, 2 years ago

Thanks. It seems the EDID transmission somehow works, but the received data isn't checksummed OK. Have you the EDID available for instance from Linux?

by thebuck, 2 years ago

Attachment: card0-eDP-1.bin added

Linux /sys/class/drm/card0-eDP-1/edid

comment:6 by thebuck, 2 years ago

In that drm folder there is also card0-HDMI-A-1 and card0-HDMI-A-2, but there is no dock or anything that might warrant two HDMI ports...

comment:7 by thebuck, 2 years ago

/sys/class/drm/card0/device/graphics/fb0/stride = 7680

Wow. 9600 pixels wide virtual screen space

comment:8 by korli, 2 years ago

Could you attach the contents of /sys/kernel/debug/dri/0/i915_frequency_info ?

by thebuck, 2 years ago

Attachment: i915_frequency_info.txt added

Linux /sys/kernel/debug/dri/0/i915_frequency_info

by thebuck, 2 years ago

No change in behaviour :(

Note: See TracTickets for help on using tickets.