Hello to everyone. I'm a new net-bsd user. I'm learning and trying this OS. The first step is to configure correctly the resolution of my screen. On my PC I have 3 graphic cards : 1) 000:02:0: Intel UHD Graphics 630 (VGA display, revision 0x02) 2) 001:00:0: NVIDIA product 1e04 (VGA display, revision 0xa1) : (NVIDIA RTX 2080ti) 3) 002:00:0: NVIDIA GeForce GTX 1060 3GB (VGA display, revision 0xa1) Good. On the BIOS I have configured the Intel UHD Graphics 630 as primary GPU device. In addition I have installed XFCE4. The problem that I have is that I can't increase the default resolution that have been chosen automatically,that's 1024x768. Probably I haven't installed the correct driver,I don't know. Below u can see all the hardware specs of my PC : netbsd-marietto# pcictl pci0 list 000:00:0: Intel Xeon E (S, Desktop) Host Bridge, DRAM (host bridge, revision 0x0d) 000:01:0: Intel Core 6G PCIe x16 (PCI bridge, revision 0x0d) 000:01:1: Intel Core 6G PCIe x8 (PCI bridge, revision 0x0d) 000:02:0: Intel UHD Graphics 630 (VGA display, revision 0x02) 000:18:0: Intel 300 Series Thermal (miscellaneous DASP, revision 0x10) 000:20:0: Intel 300 Series USB 3.1 xHCI (USB serial bus, xHCI, revision 0x10) 000:20:2: Intel 300 Series Shared SRAM (RAM memory, revision 0x10) 000:22:0: Intel 300 Series MEI (miscellaneous communications, revision 0x10) 000:23:0: Intel 300 Series SATA (AHCI) desktop (SATA mass storage, AHCI 1.0, revision 0x10) 000:27:0: Intel 300 Series PCIe (PCI bridge, revision 0xf0) 000:28:0: Intel 300 Series PCIe (PCI bridge, revision 0xf0) 000:28:5: Intel 300 Series PCIe (PCI bridge, revision 0xf0) 000:29:0: Intel 300 Series PCIe (PCI bridge, revision 0xf0) 000:31:0: Intel Z390 LPC (ISA bridge, revision 0x10) 000:31:3: Intel 300 Series cAVS (mixed mode multimedia, revision 0x10) 000:31:4: Intel 300 Series SMBus (SMBus serial bus, revision 0x10) 000:31:5: Intel 300 Series SPI (FLASH) (miscellaneous serial bus, revision 0x10) 000:31:6: Intel I219-V Ethernet Connection (ethernet network, revision 0x10) 001:00:0: NVIDIA product 1e04 (VGA display, revision 0xa1) 001:00:1: NVIDIA product 10f7 (mixed mode multimedia, revision 0xa1) 001:00:2: NVIDIA product 1ad6 (USB serial bus, xHCI, revision 0xa1) 001:00:3: NVIDIA product 1ad7 (miscellaneous serial bus, revision 0xa1) 002:00:0: NVIDIA GeForce GTX 1060 3GB (VGA display, revision 0xa1) 002:00:1: NVIDIA product 10f1 (mixed mode multimedia, revision 0xa1) 003:00:0: Micron/Crucial Technology product 5403 (Flash mass storage, NVMe, revision 0x03) 005:00:0: Renesas Technologies uPD720201 USB 3.0 Host Controller (USB serial bus, xHCI, revision 0x03) netbsd-marietto# dmesg [ 15.871239] wsdisplay0: screen 1 added (default, vt100 emulation) [ 15.871239] wsdisplay0: screen 2 added (default, vt100 emulation) [ 15.871239] wsdisplay0: screen 3 added (default, vt100 emulation) [ 15.871239] wsdisplay0: screen 4 added (default, vt100 emulation) I've installed this server Xorg : modular-xorg-server-1.20.11 Modular X11 server from modular X.org At the moment I've installed these drivers with the command pkgin install : (please be patient if I made some mistake,I'm experimenting / learning) **intel-vaapi-driver-2.4.0 VA-API user mode driver for Intel GEN Graphics family xf86-video-intel-2.99.917.20200515 Modular Xorg Intel video driver** Probably one some of them shouldn't be installed ? And I'm not using any xorg.conf file because the one that I have generated with the command : Xorg -configure does not work at all,so I have removed it. I mean the one that u see below : https://pastebin.ubuntu.com/p/9bSCXjnZZm/ Below u can read the Xorg log file : https://pastebin.ubuntu.com/p/9PnxYmkqxH/ It's all. I would like to understand how to configure correctly my Intel graphic integrated chipset. Very thanks.

marietto It's probably because your GPU is too new. NetBSD 9.x supports Intel GPUs up to and including Kaby Lake, with DRM implementation equal to Linux 4.4. Here's the link for 9.0 Announcement which mentions the DRM version. EDIT: UHD 630 is Coffee Lake and supported in Linux since around 4.13.

marietto The Xorg server is using the NVIDIA RTX 2080ti card: [ 1202.633] (--) PCI: (0@0:2:0) 8086:3e98:1458:d000 rev 2, Mem @ 0x98000000/16777216, 0x40000000/536870912, I/O @ 0x00005000/64 [ 1202.633] (--) PCI:*(1@1:0:0) 10de:1e04:19da:2503 rev 161, Mem @ 0x96000000/16777216, 0x60000000/268435456, 0x94000000/33554432, I/O @ 0x00004000/128, BIOS @ 0x????????/524288 [ 1202.633] (--) PCI: (2@2:0:0) 10de:1c02:19da:2438 rev 161, Mem @ 0x92000000/16777216, 0x80000000/268435456, 0x90000000/33554432, I/O @ 0x00003000/128, BIOS @ 0x????????/524288 if(window.hljsLoader && !document.currentScript.parentNode.hasAttribute('data-s9e-livepreview-onupdate')) { window.hljsLoader.highlightBlocks(document.currentScript.parentNode); } You can see the * on the 2nd line. However, the nvidia driver fails to initialize the card: [ 1202.649] (EE) [drm] Failed to open DRM device for pci:0001:01:00.0: -19 [ 1202.661] (EE) [drm] Failed to open DRM device for pci:0001:01:00.0: -19 [ 1202.661] (EE) open /dev/dri/card0: No such file or directory [ 1202.661] (WW) Falling back to old probe method for modesetting [ 1202.661] (EE) open /dev/dri/card0: No such file or directory if(window.hljsLoader && !document.currentScript.parentNode.hasAttribute('data-s9e-livepreview-onupdate')) { window.hljsLoader.highlightBlocks(document.currentScript.parentNode); } so, Xorg falls back to the safe option for your HW which is the wsfb driver: [ 1202.661] (II) wsfb(0): Creating default Display subsection in Screen section "Builtin Default wsfb Screen 0" for depth/fbbpp 24/32 [ 1202.661] (==) wsfb(0): Depth 24, (==) framebuffer bpp 32 if(window.hljsLoader && !document.currentScript.parentNode.hasAttribute('data-s9e-livepreview-onupdate')) { window.hljsLoader.highlightBlocks(document.currentScript.parentNode); } and that , since you probably haven't configured it to use a better resolution, defaults to 1024x768. Try: Uninstall modular-xorg and xf86-video-intel --Just use what's in the base install. Re-create the DRM devices. As root: cd /dev sh ./MAKEDEV drm0 drm1 drm2 if(window.hljsLoader && !document.currentScript.parentNode.hasAttribute('data-s9e-livepreview-onupdate')) { window.hljsLoader.highlightBlocks(document.currentScript.parentNode); } Edit /boot.cfg and disable the nouveau drivers by adding this line: userconf=disable nouveau* if(window.hljsLoader && !document.currentScript.parentNode.hasAttribute('data-s9e-livepreview-onupdate')) { window.hljsLoader.highlightBlocks(document.currentScript.parentNode); } Reboot. Post /var/run/dmesg.boot and current Xorg log file if X starts.

I already said this on reddit, but you shouldn't use modular-xorg with the official binary packages. They're built for native X only. Coffee Lake is indeed too new for the DRM/KMS stack in 9, but you should still be able to get the correct resolution with gop+wsfb.

rvp Hello. thanks for your explanation. I would like to know why NetBSD chosen to use the NVIDIA graphic card instead of the Intel UHD 630 as primary choice. I have attached the screen to the HDMI port of the UHD 630,not on the NVIDIA card. And on the BIOS,I have chosen the UHD 630 as primary,not the NVIDIA card. I want to use the UHD 630 as primary graphic card and I want to make the NVIDIA card available to be passed thru with qemu + nvmm. About this : do u know if the pass thru of my graphic card can work on NetBSD ?

marietto I would like to know why NetBSD chosen to use the NVIDIA graphic card instead of the Intel UHD 630 as primary choice. No idea--will have to look at your /var/run/dmesg.boot . Try loading the intel driver explicitly. (Add a BusID to further nail it down.) marietto do u know if the pass thru of my graphic card can work on NetBSD ? No idea. Sorry.

rvp like this ? /etc/X11/xorg.conf Section "Device" Identifier "Card0" Driver "intel" EndSection

marietto /etc/X11/xorg.conf Location is wrong. Move the wsfb.conf file out of the way and put that fragment, above, in the same dir.

[unknown] I haven't any wsfb.conf file in any place of my disk. can u be more clear about what I should do ? thanks.

I tried the solution below but it didn't work : netbsd-marietto# pkgin install xf86-video-wsfb calculating dependencies...done. 2 packages to install: xf86-video-wsfb-0.4.0nb8 modular-xorg-server-1.20.11 0 to refresh, 0 to upgrade, 2 to install 16K to download, 7861K to install proceed ? [Y/n] y cd /etc/X11/xorg.conf.d nano wsfb.conf Section "Device" Identifier "Card0" Driver "intel" EndSection

Did you actually disable nouveau and/or i915drmkms on your boot.cfg as suggested above by @rvp ?

Trying to install the driver for my main GPU,the Intel UHD Graphics 630.(i915)

rvp

marietto I've also did the following :

That works also 😀

OK, now that we have the logs, things can be explained:

[     9.576848] kern info: [drm] failed to retrieve link info, disabling eDP
[    11.136843] i915drmkms0: info: [drm] Cannot find any crtc or sizes
[    13.286834] i915drmkms0: info: [drm] Cannot find any crtc or sizes

This is why the screen turns off. The DRM driver can't find any monitors or the EDID info. returned by the monitor is wrong or not read properly. You have the DRM_MAX_RESOLUTION stuff uncommented in the kernel config., right? (You should also report this on Github.)

Now we have this situation:

nouveau disabled; i915drmkms enabled, but, not finding any monitors.
Xorg runs, modesetting driver loaded (nouveau_drv also loaded, don't know why), but can't run because no monitors found. nouveau_drv won't run because the nouveau DRM driver is disabled.

Can you add this to the X config file (Section "Device") and try to run X again. Post the Xorg log file:

    BusID    "pci:0:2:0"     # use the Intel card
    # BusID    "pci:1:0:0"     # use default Nvidia card
    # BusID    "pci:2:0:0"     # use other Nvidia card

Then, disable the Intel card in /boot.cfg, uncomment one of the Nvidia BusIDs, reboot and run startx again. Post both logs for this.

marietto

Is there a way to check if DRM_MAX_RESOLUTION is configured correctly ?

ok. now inside the folder /etc/X11/xorg.conf.d I have a file called "xorg-modesetting.conf" with this content inside :

Section "Device"
Identifier "Card0"
Driver "modesetting"
BusID "pci:0:2:0"
EndSection

xorg.1.log : https://pastebin.ubuntu.com/p/mYWP9nfShm/
dmesg : https://pastebin.ubuntu.com/p/432SPGkRR5/

rvp

marietto Is there a way to check if DRM_MAX_RESOLUTION is configured correctly ?

As long as its uncommented, it should be OK.

marietto ok. now inside the folder /etc/X11/xorg.conf.d I have a file called "xorg-modesetting.conf" with this content inside :

Section "Device"
Identifier "Card0"
Driver "modesetting"
BusID "pci:0:2:0"
EndSection

Even with BusID "pci:0:2:0" added, things are the same with the Intel DRM driver:

[     1.000000] [   377.000000] nouveau* disabled
...
[     1.015531] i915drmkms0 at pci0 dev 2 function 0: Intel UHD Graphics 630 (rev. 0x02)
...
[    10.221872] kern.module.path=/stand/amd64/9.99.82/modules
[    10.231872] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[    10.231872] kern info: [drm] Driver supports precise vblank timestamp query.
[    10.231872] kern info: [drm] Failed to find VBIOS tables (VBT)
[    10.231872] allocated pic msi9 type edge pin 0 level 6 to cpu1 slot 2 idt entry 98
[    10.231872] i915drmkms0: interrupting at msi9 vec 0 (i915drmkms0)
[    10.231872] kern info: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
[    10.501871] kern info: [drm] failed to retrieve link info, disabling eDP
[    10.521872] kern info: [drm] Initialized i915 1.6.0 20200114 for i915drmkms0 on minor 0
[    12.071867] i915drmkms0: info: [drm] Cannot find any crtc or sizes
[    12.081866] kern info: [drm] DRM_I915_DEBUG enabled
[    12.081866] kern info: [drm] DRM_I915_DEBUG_GEM enabled
[    14.231859] i915drmkms0: info: [drm] Cannot find any crtc or sizes
...

No intelfb0 framebuffer is created, and so we have in the Xorg.log file:

[   320.124] (EE) open /dev/dri/card0: No such file or directory
[   320.124] (WW) Falling back to old probe method for modesetting
[   320.124] (EE) open /dev/dri/card0: No such file or directory
[   320.124] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card support
[   320.124] (EE) Screen 0 deleted because of no matching config section.
[   320.125] (II) UnloadModule: "modesetting"
[   320.125] (EE) Device(s) detected, but none match those in the config file.

X is still loading the nouveau video driver for some reason...
Post these logs on Github.

For the Nvidia card:

marietto experiment n. 2,with this xorg.conf file 😀

Section "Device"
Identifier "Card0"
Driver "modesetting"
BusID "pci:1:0:0"
EndSection

The nouveau DRM driver didn't configure either of your Nvidia cards, so only genfb is running on all 3:

[     1.000000] [   373.000000] i915drmkms* disabled
...
[     1.012984] genfb0 at pci1 dev 0 function 0: NVIDIA product 1e04 (rev. 0xa1)
[     1.012984] genfb0: framebuffer at 0x40000000, size 1920x1080, depth 32, stride 7680
[     1.012984] genfb0: shadow framebuffer enabled, size 8100 KB
[     1.012984] wsdisplay0 at genfb0 kbdmux 1: console (default, vt100 emulation)
[     1.012984] wsmux1: connecting to wsdisplay0
[     1.012984] drm at genfb0 not configured
...
[     1.012984] genfb1 at pci2 dev 0 function 0: NVIDIA GeForce GTX 1060 3GB (rev. 0xa1)
...
[     1.012984] genfb2 at pci0 dev 2 function 0: Intel UHD Graphics 630 (rev. 0x02)

And, in Xorg.log:

[    53.577] (EE) open /dev/dri/card0: Operation not supported by device
[    53.577] (WW) Falling back to old probe method for modesetting
[    53.577] (EE) open /dev/dri/card0: Operation not supported by device
[    53.577] (EE) No devices detected.

(in this case,for some reason,the screen didn't turn off,but X does not start)

The screen didn't turn off because you're running on genfb (no Nvidia DRM drivers have taken over!); and X hasn't started because the modesetting driver needs DRM kernel drivers. It won't run on top of genfb.

For the Intel test, the Intel card was set as primary in BIOS, right?
And, no KVM (keyboard, video, mouse) switches between the card and monitor, right? Direct connection?

Fortune cookie: You have bad luck. Maybe luck will change later.

JuvenalUrbino

marietto Is there a way to check if DRM_MAX_RESOLUTION is configured correctly ?

# config -x /netbsd | grep DRM_MAX_RESOLUTION

marietto

experiment n. 2,with this xorg.conf file 😀

Section "Device"
Identifier "Card0"
Driver "modesetting"
BusID "pci:1:0:0"
EndSection

(in this case,for some reason,the screen didn't turn off,but X does not start)

Xorg.0.log : https://pastebin.ubuntu.com/p/jKBz4PmSRY/
dmesg.log : https://pastebin.ubuntu.com/p/DMmXQDH2n3/

marietto

rvp 😀

yes,I have the Intel set as primary card in BIOS. I don't use KVM switches. I created this bug ticket :

https://gitlab.freedesktop.org/drm/intel/-/issues/4279

rvp

marietto I created this bug ticket

That actually is a good summary, but, since this is a NetBSD DRM issue, you should file it on Github.

Can you make a dmesg log after booting the kernel with -v -x? Power off; then interrupt the bootloader; at the prompt type boot -v -x. Connect direct to the monitor's HDMI port. Xorg log not needed.

marietto

options DRM_MAX_RESOLUTION_HORIZONTAL=1920
options DRM_MAX_RESOLUTION_VERTICAL=1080

marietto

I made another experiment. I've attached the hdmi cable from the intel graphic card to the VGA port of my monitor,instead of attach it to the hdmi port. I've used a converter from hdmi to vga. It means vga on the monitor and hdmi on the intel graphic card. The monitor turned off anyway. I've saved the log files.

dmesg_log_intel : https://pastebin.ubuntu.com/p/KJCsthZ87T/
xorg.log_intel : https://pastebin.ubuntu.com/p/whcWDMctvv/

marietto

Im not sure that it is only a netbsd issue,because I'm using a firmware that comes from intel called i915/kbl_dmc_ver1_04.bin. the problem could be inside this file. Do u agree ? Anyway,give me the Github page where you think that I should report the problem.

----> Can you make a dmesg log after booting the kernel with -v -x? Power off; then interrupt the bootloader; at the prompt type boot -v -x

ok. will do.

rvp

marietto Im not sure that it is only a netbsd issue,because I'm using a firmware that comes from intel called i915/kbl_dmc_ver1_04.bin. the problem could be inside this file. Do u agree ?

We're not seeing any firmware-related errors...

marietto Anyway,give me the Github page where you think that I should report the problem.

Just add it to your existing issue there.

Anyway, we can see if your monitor is returning proper EDID data. SSH into your machine, then compile and run (as root) the program below, then email the binary output file to me (same username as here) lumaca SDF punto ORG_anizzazione_

~~If the program says that it cannot open /dev/ttyE0, then retry it on genfb (ie. just disable both Intel and Nouveau DRM drivers in /boot.cfg)~~
Note: the program doesn't work on genfb

cc -o wsedid wsedid.c                   # compile
./wsedid > /tmp/hdmi_edid.bin           # run
/dev/ttyE0 EDID data_size = 128         # printed by program

/**
 * wsedid.c: Print EDID info. using wsdisplay(4) on NetBSD/OpenBSD.
 */
#include <dev/wscons/wsconsio.h>
#include <sys/ioctl.h>
#include <sys/utsname.h>
#include <err.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

int
main(int argc, char* argv[])
{
	struct utsname un;
	struct wsdisplayio_edid_info ei;
	char ebuf[4096];
	char* dev, *os;
	int fd, rc = EXIT_FAILURE;

	if ((uname(&un)) < 0)
		err(rc, "uname failed");
	os = un.sysname;
	if (argc == 2)
		dev = argv[1];
	else {
		if (strcmp(os, "NetBSD") == 0)
			dev = "/dev/ttyE0";
		else if (strcmp(os, "OpenBSD") == 0)
			dev = "/dev/ttyC0";
		else
			errx(rc, "%s: OS is not supported", os);
	}
	ei.edid_data = ebuf;
	ei.buffer_size = sizeof ebuf;
	if ((fd = open(dev, O_RDONLY)) < 0)
		err(rc, "%s: open failed", dev);
	if (ioctl(fd, WSDISPLAYIO_GET_EDID, &ei) < 0)
		err(rc, "ioctl(WSDISPLAYIO_GET_EDID) failed");
	fprintf(stderr, "%s EDID data_size = %u\n", dev, ei.data_size);
	fwrite(ebuf, 1, ei.data_size, stdout);

	rc = EXIT_SUCCESS;
	return rc;
}

marietto

https://pastebin.ubuntu.com/p/rBwpMMjtZ6/

marietto

netbsd-marietto# cc -o wsedid wsedid.c

netbsd-marietto# ./wsedid > /tmp/hdmi_edid.bin

wsedid: ioctl(WSDISPLAYIO_GET_EDID) failed: Operation not supported by device

I tried to send you the file produced,but I can't do that because it's empty,its 0 byte and gmail won't send it.

rvp

marietto I can't do that because it's empty,its 0 byte

Don't send me a 0-byte file, mate. What'll I do with it?

~~Retry with genfb~~

EDIT: the program won't work on genfb

marietto

hahahahaha I know. But I didn't realize that it was 0 byte until I tried to send it 4 email 😃
ok. I did that. I've got the same error as before. Probably my monitor does not support EDID,right ?

rvp

marietto I've got the same error as before.

EDIT: Note: program won't work on genfb

Hmm. Probably the wrong device, then. Try:

./wsedid /dev/constty > /tmp/hdmi_edid.intel.bin    # with intel DRM
./wsedid /dev/constty > /tmp/hdmi_edid.genfb.bin    # with genfb (won't work--skip)

Mail both files. We'll see if there's a difference. (And, if /dev/constty doesn't work, try /dev/console)

marietto

what's the scenario within I should produce those files ? choose from below the right comtinabion 😀

1) with Intel and Nouveau DRM drivers disabled in boot.cfg
2) with Nouveau disabled in boot.cfg
3) with intel disabled in boot.cfg
4) with intel and nouveau enabled in boot.cfg

rvp

marietto what's the scenario within I should produce those files ?

Don't bother. This is a kernel issue. I get the same error with my old integrated Intel card where the console display is OK (but X doesn't work as I mentioned above).

If you want to get the EDID data, then you'll have to put back the GENERIC kernel and run the program on that. (Though I don't see, now, what use that would be as the 5.6 DRM drivers are still very much a WIP... I had hoped the issue was your monitor sending slightly non-standard EDID data, and that we could work around it by applying some known workarounds.)

marietto

I tried with intel and nouveau driver disabled....also tried with nouveau disabled and intel enabled. nothing to do.

marietto

I changed the monitor. I'm using a very old monitor,Samsung syncMaster 740N. It is a VGA only monitor. So,the cable is VGA behind and it became HDMI on the intel graphic card port,thanks to a hardware signal converter. It works like the AOC. Same errors and it turns off with the intel driver enabled. So,it does not depend on my monitor. I suspect that only some specific monitors will work. But it's hard to determine in advance which monitors are ok.

« Previous Page Next Page »