[SOLVED] Blank Screen on Console Login and Startx - GeForce GTX 1650 - Driver 440.82

Hello,

Linux Dist: Latest Arch Linux

During Boot:

If kernel param ‘nvidia-drm.modeset=1’ is set then immediately after login prompt appears screen goes blank.

Normal boot without any modeset param just gives standard login but low resolution typical of no modeset happening.

An error i2c error always appears before the login prompt is active. This error is present in the attached nvidia-bug-report file.

After ‘startx’ is run from VT1:

The screen goes blank, give it 30 seconds and I can switch to VT2 then back to VT1 at which point the screen comes back to life.

Other things tried:

Using Nouveau works correctly on boot and startx (switching to using this is not a viable solution for me)

Thank you! :)

Scott

nvidia-bug-report.log (763.2 KB)

The edid of the monitor is broken

Extracted contents:
header:          30 30 20 66 66 20 66 66
serial number:   20 66 66 20 66 66 20 66 66 20
version:         66 66
basic params:    20 30 30 20 20
chroma info:     30 39 20 64 31 20 30 63 20 37
established:     39 20 34
standard:        35 20 35 34 20 30 30 20 30 30 0a 32 37 20 31 32
descriptor 1:    20 30 31 20 30 33 20 38 30 20 32 66 20 31 61 20 37 38
descriptor 2:    20 20 32 65 20 33 35 20 38 35 20 61 36 20 35 36 20 34
descriptor 3:    38 20 39 61 20 32 34 0a 31 32 20 35 30 20 35 34 20 61
descriptor 4:    35 20 36 62 20 62 61 20 37 31 20 30 30 20 20 38 31 20
extensions:      30
checksum:        30

No header found
Manufacturer: HCF Model 2066 Serial Number 1713399398
EDID version: 102.102
Analog display, Input voltage level: 0.714/0.286 V
Sync: 
Maximum image size: 48 cm x 48 cm
Gamma: 1.32
DPMS levels: Off
Monochrome or grayscale display
Established timings supported:
  640x480@60Hz
  640x480@67Hz
  640x480@72Hz
  800x600@60Hz
  832x624@75Hz
Standard timings supported:
  672x672@92Hz
  672x672@112Hz
  504x504@108Hz
  632x632@92Hz
  632x632@108Hz
  328x328@110Hz
  688x688@92Hz
  640x640@110Hz
Detailed mode: Clock 123.200 MHz, 1568 mm x 305 mm
                817 1121 1665  849 hborder 32
                819  838  872 2899 vborder 55
               -hsync -vsync field sequential L/R
Detailed mode: Clock 82.240 MHz, 822 mm x 1312 mm
                562  874 1439  663 hborder 54
                563  565  581  616 vborder 32
               -hsync +vsync digital composite field sequential L/R
Detailed mode: Clock 82.480 MHz, 816 mm x 1312 mm
                569  618 1436  666 hborder 52
                 50   68   84 2662 vborder 32
               -hsync -vsync analog composite side by side interleaved
Detailed mode: Clock 82.450 MHz, 560 mm x 32 mm
                566  621 1438  664 hborder 56
                610  612  612  707 vborder 49
               -hsync -vsync analog composite field sequential L/R
Has 48 extension blocks
Checksum: 0x30 (should be 0xa3)
EDID block does not conform at all!
	Block has broken checksum
	Bad year of manufacture
	Bad week of manufacture

Completely weird values, must have some byte within toggled. So what gets used by which driver is subject to luck. Nouveau is always ‘the poor guy’, it lives on what it can get to so works on otherwise broken hardware.
I also don’t now if a fixed edid could be loaded on kernel command line with the nvidia proprietary driver.

Thank you generix! Connecting a different monitor fixed the problem for running X11 :)

I am still getting the i2c timeout error on boot but it doesn’t appear to be affecting anything at the moment.

Thanks again,

Scott

The i2c timeout error is common on Turing gen cards without an USB-C slot. The driver times out because it’s not there. Nothing to worry about.

That is great news. Thank you so much generix, you are awesome!