Opened 3 years ago
Closed 3 years ago
#17655 closed bug (fixed)
intel_extreme, screen display downscaled by 3/4
Reported by: | skirst | Owned by: | rudolfc |
---|---|---|---|
Priority: | normal | Milestone: | R1/beta4 |
Component: | Drivers/Graphics/intel_extreme/kabylake | Version: | R1/Development |
Keywords: | Cc: | ||
Blocked By: | Blocking: | #17654 | |
Platform: | All |
Description
With hrev55942, on-metal install on an intel Kaby-Lake system, the display has begun being presented at what seems to be 3/4 native resolution. The lower right portion of the display padded with black. Mouse interaction works fine, within the confines of the the display, and the system seems to work fine otherwise. Rolling back to hrev55926 or using failsafe display from the boot loader absolves the issue. I have attached syslogs from each system state.
Attachments (6)
Change History (28)
by , 3 years ago
Attachment: | syslog-ok-hrev55926 added |
---|
by , 3 years ago
Attachment: | syslog-bug-hrev55942 added |
---|
comment:1 by , 3 years ago
Component: | - General → Drivers/Graphics/intel_extreme/kabylake |
---|---|
Owner: | changed from | to
Platform: | x86-64 → All |
comment:2 by , 3 years ago
Blocking: | 17654 added |
---|
comment:3 by , 3 years ago
by , 3 years ago
Attachment: | PXL_20220315_084453505.jpg added |
---|
comment:6 by , 3 years ago
the picture shows 1366x768 (comes from the VBT). You should be able to adjust to 1920x1080 in the screen preferences.
comment:7 by , 3 years ago
i tried that to no avail. setting other resolutions has a similar effect.
comment:8 by , 3 years ago
forgive me, new to this. another thing i've found that despite the screen being scaled, if i use flameshot to take a screenshot, the resulting image is at native resolution, minus 1 pixel in each dimension.
comment:9 by , 3 years ago
Hi, that's OK of course ;-) korli just created a new piece of code for fetching the native modeline. as soon as that sits in the driver maybe you can try again to see what happens? I/we will keep you posted :-)
by , 3 years ago
Attachment: | syslog-hrev55954 added |
---|
comment:11 by , 3 years ago
The problem still persists with hrev55954, unfortunately. I've attached an updated syslog.
comment:13 by , 3 years ago
Indeed the problem seems to be resolved now. Thanks! You can close this ticket.
by , 3 years ago
Attachment: | syslog-hrev56004 added |
---|
comment:16 by , 3 years ago
at the moment it only picks the EDID from the first dp aux channel. Rudolf, any thoughts?
comment:17 by , 3 years ago
I did not have time yet unfortunately to look at the code.but looking at this syslog it seems my connected code has been overruled. If instead only on dp connections the aux would be checked chances would be the right info would be at the right connection/ screen. At least this would be my route instead of for instance checking the bios/rom like stuff for routes, if possible. Since it seems closer to a quick result where also multiple screens are connected. I would assume for now only one screen uses dp protocol.
As said, I did not look closely yet, so I could be talking nonsense here..
comment:18 by , 3 years ago
Please check with hrev56024 or newer. Obviously assigning 3 pipes to 4 connected ports won't work nicely.
comment:19 by , 3 years ago
Please check with hrev56088 or newer, and provide a new syslog. the error "No pipes left to assign to port" should be gone.
by , 3 years ago
Attachment: | syslog-hrev56092 added |
---|
comment:21 by , 3 years ago
Yes. Native resolution is working, as well as brightness control. Thanks!
comment:22 by , 3 years ago
Milestone: | Unscheduled → R1/beta4 |
---|---|
Resolution: | → fixed |
Status: | new → closed |
Nice!
Which laptop is this?