Opened 3 years ago

Closed 10 months ago

Last modified 4 months ago

#13889 closed bug (duplicate)

No video when using intel_extreme driver

Reported by: ThomasWinwood Owned by: kallisti5
Priority: normal Milestone:
Component: Drivers/Graphics/intel_extreme Version: R1/Development
Keywords: Cc:
Blocked By: #14368 Blocking:
Platform: x86


I have a Compaq Evo D510 SFF on which I was running a build of Haiku from a few months back. I decided to upgrade to the latest build, only to be greeted by the video output switching off after the boot process finished.

I brought this up on IRC and followed the suggestion to enable the failsafe video driver, which at least gets me a working system, though at the lower resolution of 1280x1024 rather than the 1680x1050 I enjoyed from the previous build.

I have attached a syslog (which I hope is) from a run using the intel_extreme driver and the output of listdev and listimage | grep drivers as recommended on the reporting-bugs wiki page.

Attachments (4)

syslog.txt (78.2 KB ) - added by ThomasWinwood 3 years ago.
listdev.txt (1.5 KB ) - added by ThomasWinwood 3 years ago.
listimage.txt (1.4 KB ) - added by ThomasWinwood 3 years ago.
i845-simulation.png (296.5 KB ) - added by kallisti5 2 years ago.
simulation i845G

Download all attachments as: .zip

Change History (15)

by ThomasWinwood, 3 years ago

Attachment: syslog.txt added

comment:1 by ThomasWinwood, 3 years ago

patch: 01

by ThomasWinwood, 3 years ago

Attachment: listdev.txt added

by ThomasWinwood, 3 years ago

Attachment: listimage.txt added

comment:2 by ThomasWinwood, 3 years ago

Additional information: I did a little walking back through previous nightlies, and the problem has existed since at least hrev51678. I didn't think to make a note of what the working build was, though I doubt it matters.

comment:3 by diver, 3 years ago

As a workaround you can blacklist intel_extreme and intel_extreme.accelerant, extract them from a working build and put it to non-packaged.

comment:4 by ThomasWinwood, 3 years ago

Okay, I had some time on my hands so I did some bisecting (to the granularity of the nightly builds available on the Haiku website) and determined that the bug was introduced somewhere between hrev49662 and hrev50357. Any more detailed testing will have to involve someone feeding me custom builds, maybe on IRC or something.

comment:5 by pulkomandy, 3 years ago

patch: 10

comment:6 by al_xv, 2 years ago

I also have a Compaq Evo D510 and have the same issue. I didn't have any problems with R1A4.

comment:7 by kallisti5, 2 years ago

Looks like the issue for Thomas started after the big rewrite in hrev50232.

Here is the core problem per the logs:

intel_extreme: compute_dpll_9xx: best MHz: 1.15743e-12 (error: 107.964)

Something is happening during the PLL calculations.

I added a procedure to run PLL simulations in this commit:

Let me punch the numbers and see what happens.

by kallisti5, 2 years ago

Attachment: i845-simulation.png added

simulation i845G

comment:8 by pulkomandy, 22 months ago

Milestone: UnscheduledR1/beta2

comment:9 by waddlesplash, 16 months ago

Please retest after hrev53040.

comment:10 by pulkomandy, 10 months ago

Blocked By: 14368 added
Resolution: duplicate
Status: newclosed

Duplicate of #14368, please test with the changes listed there in my last comment.

comment:11 by nielx, 4 months ago

Milestone: R1/beta2

Remove milestone for tickets with status = closed and resolution != fixed

Note: See TracTickets for help on using tickets.