Opened 7 years ago

Last modified 5 years ago

#8626 new enhancement

radeon_hd needs display port support

Reported by: kallisti5 Owned by: kallisti5
Priority: normal Milestone: R1
Component: Drivers/Graphics/radeon_hd Version: R1/Development
Keywords: nutmeg, travis, displayport Cc:
Blocked By: #8612 Blocking: #8485, #10335
Has a Patch: no Platform: All

Description (last modified by kallisti5)

While radeon_hd is aware of display port devices (and can even speak to them somewhat via the aux channel), the display port link training code needs finished. This is essential to supporting newer chipsets in laptop configurations with internal DisplayPort encoders such as travis and nutmeg.

TRAVIS is a DP -> lvds encoder
NUTMEG is a DP -> VGA encoder

Attachments (4)

syslog.GCC4Hyb.44523 (209.3 KB ) - added by Bouncer 7 years ago.
syslog.44523.with.DP (201.6 KB ) - added by Bouncer 7 years ago.
Dell.Precision-M4600-Radeon.HD-6700M-2013-04-01-Haiku45432.log (176.6 KB ) - added by dknoto 7 years ago.
Dell.Precision-M4600-Radeon.HD-6700M-2013-04-01-X11.log (47.6 KB ) - added by dknoto 7 years ago.

Download all attachments as: .zip

Change History (17)

comment:1 by kallisti5, 7 years ago

Description: modified (diff)

comment:2 by kallisti5, 7 years ago

Blocking: 8485 added

comment:3 by kallisti5, 7 years ago

Description: modified (diff)

comment:4 by kallisti5, 7 years ago

#8826 duplicate issue was closed.

comment:5 by kallisti5, 7 years ago

hrev44491+ may offer some relief for DisplayPort folks. Testers needed!

comment:6 by Bouncer, 7 years ago

Have just tested with hrev44523 (GCC4 Hyb). Display Port not working for HD6770. Attempting to attach a syslog.

by Bouncer, 7 years ago

Attachment: syslog.GCC4Hyb.44523 added

comment:7 by Bouncer, 7 years ago

Sorry - somehow - I submitted the wrong syslog. It shows the HD6770 using a DVI connector.

This second attachement is the correct one for the DisplayPort. eg: syslog.44523.with.DP

by Bouncer, 7 years ago

Attachment: syslog.44523.with.DP added

comment:8 by kallisti5, 7 years ago

KERN: radeon_hd: detect_displays: ERROR: 0 attached monitors were found on display connectors. Injecting first connector as a last resort.

Definitely your first issue.. I tested the DisplayPort on my 7970 this evening and got the same error. Let me see what I can do about the EDID on DisplayPort devices.

comment:9 by kallisti5, 7 years ago

Blocked By: 8612 added

adding blocked by #8612.

It's unlikely that the VESA edid matches the DisplayPort endpoint edid.

As the accelerant mode_set hook doesn't pass the crtc id, the hook functions are blissfully unaware of the needed information to properly train each display port device.

comment:10 by kallisti5, 7 years ago

The largest thing left to fix here is the lack of multi-head support in the AppServer.

The core problem is that currently we use the vesa edid mode for all monitors as the appserver doesn't pass which mode to set on which display. However the edid modes from vesa don't reflect *any* DisplayPort edid info as the DisplayPort monitors have to be probed.

We could set the "preferred/native mode" on each display, but this would mean the end user would get 0 choice on what displaymode to set where.

in reply to:  10 comment:11 by dknoto, 7 years ago

I tested hrev45432-gcc4-hybrid on Dell Precision M4600 with Radeon HD 6700M. Result of this test you can see on this short movie: http://dl.dropbox.com/u/92245884/20130331_173905.mp4 I attached syslog from Haiku and to comparing the log file of X11 from Linux Fedora 17.

comment:12 by kallisti5, 6 years ago

Blocking: 10335 added

comment:13 by kallisti5, 5 years ago

closed https://dev.haiku-os.org/ticket/9894 as duplicate. A6-4400M APU

Note: See TracTickets for help on using tickets.