Improving the Intel Extreme Driver (Was: Graphics on Dell laptop in Vesa mode only)

Sorry for the delay: but this chipset looks to have DDI ports indeed, so that could work. The PLL’s might work differently though.
All in all I expect this to be not a 1:1 fit, but it should be doable if it can be tested.

@Kapix: what do you think about this chipset seeing your research in types? I looked at apollo lake’s Volume 7: Display PDF document. All the others on the ‘01’ site (2016 Intel(R) Processors Based on Apollo Lake Platform (Broxton Graphics) | 01.org) seem to not give more details. Are you able to find more precise and complete register level docs for instance?

Thanks!

2 Likes

In the meantime it became clear that another small update is needed in the driver, it has to let app-server know it does have EDID info in case of laptops, which it does not do yet for internal panels.

So this will be fixed soon too :relaxed:

8 Likes

Registers manual for Apollo Lake is linked in the main page.

It seems clock programming sequence (not just PLL calculation) is different for Gen9LP.

Note:

  • Intel Iris Xe MAX Graphics - PCI ID: 4905
  • Intel Iris Xe Pod - PCI ID: 4906
  • Intel Server GPU SG-18M - PCI ID: 4907
  • Intel Iris Xe Graphics - PCI ID: 4908
  • Intel UHD Discrete Graphics - PCI ID: 4909

Dell XPS 13 (2160p UHD screen) is Haiku-compatible.

  • Rocket Lake (desktop) - Gen 12
  • Tiger Lake (mobile) - Gen 12
1 Like

Ah, yes I see now. Missed that before, thanks! Anyhow, more work to get it running then…

Hi, do you have that laptop? In that case can you find the memory controller’s ID? This card is not yet in the driver, though kallisti5 also has a XPS13 (1800p) which -is- detected (But I don’t know it’s IDs yet).

1 Like

Yes. I sent you notes on it.

It’s a TigerLake, a different beast :slight_smile:

1 Like

@extrowerk could you check this bug report of yours? Thanks.

I am checking, but which one do you refer to?
Probably this one, right? https://dev.haiku-os.org/ticket/15396
Will check it.

1 Like

device Display controller (VGA compatible controller, VGA controller) [3|0|0]
vendor 8086: Intel Corporation
device 3e92: CometLake-S GT2 [UHD Graphics 630]

screenshot1215

HDMI cable
The screen brightness is not regulated.

2 Likes

It looks like the pci.ids entry isn’t correct. I’ve submitted a new name for it with the Linux sources as reference.

3 Likes

Very encouraging, this is potentially my graphics also. Time to retry (when I can find spare time)!! Thanks @kim1963 !!

2 Likes

Does the driver (Skylake and it’s a HP PRODESK 600 G2) work with DP? DP did work with the Vesa drive (didn’t thought it would) but now it’s blank (Updated today)

https://dev.haiku-os.org/ticket/17437

1 Like

Should be possible yes. But you really need to provide at least a syslog in that ticket :grin:

2 Likes

I assume you mean it should be flagged desktop instead of mobile, correct? That slider for intensity should not be there I think…

No, I was talking about:

which is now corrected at PCI Devices

PXL_20211203_205732157
:frowning: Don’t start

Bug report :slight_smile:

1 Like

Hello Andrea,

Is this a boot attempt on your system in this ticket?
https://dev.haiku-os.org/ticket/17201

If so, maybe we should re-open it and update it with the latest info. Also please do try to get a syslog from that system. If that won’t work because you cannot make it fully boot, I guess you could do a try on a USB stick, and then after the failed boot, mount that stick on another system running Haiku and grab the syslog from that stick that way?

Interesting: for instance:

  • no display is detected (which would only make the booticons resolution work at most)
  • the adpa, etc. register locations are not implemented in your hardware from the looks of it: did not see that behaviour before…

(I sure hope I’m not babling to much :wink: )