Opened 5 years ago
Last modified 2 years ago
#15665 assigned bug
intel mobile 945GSE wait_for_retrace does not work (teapot shows black screen)
Reported by: | bruno | Owned by: | rudolfc |
---|---|---|---|
Priority: | low | Milestone: | Unscheduled |
Component: | Drivers/Graphics/intel_extreme/9xx | Version: | R1/Development |
Keywords: | Cc: | ||
Blocked By: | Blocking: | ||
Platform: | All |
Description (last modified by )
hrev53794 32bit...
boots fine to native resolution
1024 x 600
changing resolution with lower screensize works...
bigger ones will result in black screen...
changing workspace will not work... have to force reboot
Attachments (8)
Change History (28)
by , 5 years ago
Attachment: | screenshot1.png added |
---|
by , 5 years ago
by , 5 years ago
Attachment: | media_addon_server-441-debug-29-01-2020-20-38-04.report added |
---|
by , 5 years ago
comment:1 by , 5 years ago
Component: | - General → Drivers/Graphics/intel_extreme |
---|---|
Description: | modified (diff) |
Owner: | changed from | to
Status: | new → assigned |
comment:2 by , 5 years ago
comment:3 by , 5 years ago
What would this have to dowith hardware cursor? The teapot uses wait_for_retrace to synchronize itself, so if that doesn't work it justwaits forever.
comment:4 by , 5 years ago
Like you said it might have nothing to do with it. I was just guessing by noticing that that the revolving haiku letters one work (no black screen) and it does not have user interaction (no mouse input). I agree that both may be totally coded differently internally. Speaking of that demo, it would be more fun if the user could spin the letter just like the teapot :).
comment:6 by , 5 years ago
Hi, I'm confused because your ticket title mentions teapot, but description mentions other problems.
For the workspace switching, you can fix that by re-setting the same (working) resolution for all workspaces.
For teapot, the problem is not related to OpenGL. It is because the application uses wait_for_retrace to synchronize itself with the display. The way to detect the retrace has changed several times on different cards and yours is not looking at the right place, yet.
comment:7 by , 5 years ago
Description: | modified (diff) |
---|
ah sorry for confusing you here... I can confirm: Switching workspaces works with all resolutions down the native one! Also trying to use higher screen size will result in black/white screen and automatically change to last resolution without problems. So it is not a bug!
left problem is only: 1: Black teapot screen! 2: Brightness slider has no effect!
The card is working and recognized so far no vesa needed.
comment:8 by , 5 years ago
Is this an Asus eeePC? In that case it is known that the brightness doesn't work. Asus did not connect the Intel video card brightness to the LCD and used something else. You can set the brightness using the hardware shortcuts, only. This will require a specific driver just for this machine if we want brightness to work.
I pushed a fix for the teapot (can't check if it works) and I'm working on one for limiting the video resolutions to avoid the ones leading to black screen.
comment:9 by , 5 years ago
with hrev53864 32bit Teapot still will not show off... anything else works! add new syslog... The teapot app dont close! have to kill
by , 5 years ago
Attachment: | syslog10022020 added |
---|
comment:10 by , 5 years ago
Component: | Drivers/Graphics/intel_extreme → Drivers/Graphics/intel_extreme/9xx |
---|---|
Priority: | normal → low |
Summary: | intel mobile 945GSE black screen with Teapot → intel mobile 945GSE wait_for_retrace does not work (teapot shows black screen) |
by , 5 years ago
Attachment: | syslog_hrev54226 added |
---|
by , 4 years ago
comment:12 by , 4 years ago
Would the test v4 intel_extreme driver maybe change something here? https://discuss.haiku-os.org/t/improving-the-intel-extreme-driver-was-graphics-on-dell-laptop-in-vesa-mode-only/10583/134
comment:13 by , 4 years ago
Owner: | changed from | to
---|
Hi, I took the liberty to assign it to me because I'm working on it currently. Bruno, could you upload the syslog here from your latest attempt with hrev55189 in which the 800x600 mode works in dualhead clone mode (laptop output correct but not centered, instead on left side of the panel)
Thanks!
comment:14 by , 3 years ago
Hi, for your info, currently there's a patch being worked on to imrpove the vertical retrace behaviour on a lot of cards: https://review.haiku-os.org/c/haiku/+/4760
So I expect it will soon improve :-)
comment:15 by , 3 years ago
Hello, Could you try a current nightly for the GLTeapot trouble? If all is right it works now as the patch has been submitted some time ago. Thank you!
comment:16 by , 3 years ago
Sorry to report it is still not working... I removed the experimental accelerant in Non system folder. Still not working! What do you need to solve the problem?!
comment:17 by , 3 years ago
Hmm, that's unfortunate. I'll see if it works on all my systems first then, upto now only checked the Skylake which works now for this..
I'll report back here when I'm done testing the teapots.. :)
Thank you!
comment:18 by , 3 years ago
Hello again: In the meantime I tested all my systems, also I am asking users: GLTeapot is spinning at the expected speed sofar everywhere..
Could you try hrev55928 or later to see how it behaves, as an update to the situation? Are you able to get to the desktop? can you set various resolutions? does GLTeapot spin at 60Hz?
I think I need a recap so to speak, and also the driver has been expanded again..
Thank you!
comment:19 by , 3 years ago
hrev.55954 32bit haiku still same results:
no teapot to see... black window PulkoMandy wrote once it is not difficult to make the teapot visible...
boots nice to 1024 x 600 intern laptop screen...
boots nice to 1920 x 1080 external Monitor/TV (still with garbage on bottom screen with 32bit)
various screen resolutions I get only with extern Monitor, but it works
by , 3 years ago
Attachment: | syslog 18 March added |
---|
0x27ae pci on my acer apire one (also 945GSE) also don't have GL teapot (the haiku revolving letter logo work so i guess the hardware cursor is in play in the problem?).
0x2592 (915 chip) on my toshiba laptop work ok with open GL teapot.
both on hrev53797