[SOLVED] Slackware64-current kernel 6.6.7 and i915 module crash and blank screen

kingbeowulf · 12-16-2023, 07:32 PM

I have a 2006vintage Lenovo Thinkpad T510 I use for portable computing sometimes and Slackware64-current testing. It's been running just fine on Current for the past year. The recent upgrade to kernel-6.6.7, however, results in some sort of i915 crash and no viable framebuffer. I tried both generic and huge version. System is useable via ssh, just no display. I've reverted to 6.1.66 that works for the intel GPU. No issues with AMD GPUs (so far!).

laptop is i7-620M integrated iGPU i915

Code:

smod |grep i915
i915                 3915776  7
drm_buddy              20480  1 i915
ttm                    98304  1 i915
drm_display_helper    212992  1 i915
cec                    77824  1 i915
drm_kms_helper        249856  2 drm_display_helper,i915
drm                   831488  7 drm_kms_helper,drm_display_helper,drm_buddy,thinkpad_acpi,i915,ttm
intel_gtt              24576  2 intel_agp,i915
i2c_algo_bit           12288  1 i915
i2c_core              122880  12 videodev,drm_kms_helper,i2c_algo_bit,at24,drm_display_helper,i2c_smbus,thinkpad_acpi,i2c_i801,i915,psmouse,regmap_i2c,drm
video                  65536  2 thinkpad_acpi,i915

Code:

00:02.0 VGA compatible controller: Intel Corporation Core Processor Integrated Graphics Controller (rev 02) (prog-if 00 [VGA controller])
	Subsystem: Lenovo Core Processor Integrated Graphics Controller
	Flags: bus master, fast devsel, latency 0, IRQ 27
	Memory at f2000000 (64-bit, non-prefetchable) [size=4M]
	Memory at d0000000 (64-bit, prefetchable) [size=256M]
	I/O ports at 1800 [size=8]
	Expansion ROM at 000c0000 [virtual] [disabled] [size=128K]
	Capabilities: <access denied>
	Kernel driver in use: i915
	Kernel modules: i915

syslog looks normal

Code:

Dec 16 14:38:00 arwen kernel: pci 0000:00:00.0: Intel HD Graphics Chipset
Dec 16 14:38:00 arwen kernel: pci 0000:00:00.0: detected gtt size: 2097152K total, 262144K mappable
Dec 16 14:38:00 arwen kernel: pci 0000:00:00.0: detected 32768K stolen memory
Dec 16 14:38:00 arwen kernel: i915 0000:00:02.0: vgaarb: deactivate vga console
Dec 16 14:38:00 arwen kernel: Console: switching to colour dummy device 80x25
Dec 16 14:38:00 arwen kernel: i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=io+mem:owns=io+mem
Dec 16 14:38:00 arwen kernel: input: SynPS/2 Synaptics TouchPad as /devices/platform/i8042/serio1/input/input5
Dec 16 14:38:00 arwen kernel: i915 0000:00:02.0: [drm] Skipping intel_backlight registration
Dec 16 14:38:00 arwen kernel: [drm] Initialized i915 1.6.0 20201103 for 0000:00:02.0 on minor 0
Dec 16 14:38:00 arwen kernel: ACPI: video: Video Device [VID] (multi-head: yes  rom: no  post: no)
Dec 16 14:38:00 arwen kernel: input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input8
Dec 16 14:38:00 arwen kernel: acpi device:01: registered as cooling_device5
Dec 16 14:38:00 arwen kernel: snd_hda_intel 0000:00:1b.0: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
Dec 16 14:38:00 arwen kernel: cfg80211: Loading compiled-in X.509 certificates for regulatory database
Dec 16 14:38:00 arwen kernel: Loaded X.509 cert 'sforshee: 00b28ddf47aef9cea7'
Dec 16 14:38:00 arwen kernel: videodev: Linux video capture interface: v2.00
Dec 16 14:38:00 arwen kernel: fbcon: i915drmfb (fb0) is primary device
Dec 16 14:38:00 arwen kernel: Console: switching to colour frame buffer device 170x48
Dec 16 14:38:00 arwen kernel: i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device

messages shows an error:

Code:

Dec 16 13:12:17 arwen kernel: ------------[ cut here ]------------
Dec 16 13:12:17 arwen kernel: i915 0000:00:02.0: Could not determine valid watermarks for inherited state
Dec 16 13:12:17 arwen kernel: WARNING: CPU: 1 PID: 513 at drivers/gpu/drm/i915/display/i9xx_wm.c:3497 ilk_wm_sanitize+0x2be/0x2e0 [i915]
Dec 16 13:12:17 arwen kernel: Modules linked in: kvm at24 regmap_i2c think_lmi mei_pxp irqbypass i915(+) wmi_bmof firmware_attributes_class mxm_wmi crct10dif_pclmul crc32_pclmul snd_hda_intel snd_intel_dspcfg drm_buddy snd_intel_sdw_acpi polyval_generic ttm snd_hda_codec ghash_clmulni_intel thinkpad_acpi drm_display_helper sha512_ssse3 snd_hda_core cec sha256_ssse3 psmouse nvram sdhci_pci sha1_ssse3 rc_core snd_hwdep ledtrig_audio platform_profile cqhci drm_kms_helper intel_cstate rfkill acpi_cpufreq serio_raw sdhci i2c_i801 snd_pcm intel_agp firewire_ohci mmc_core drm snd_timer i2c_algo_bit firewire_core i2c_smbus intel_gtt snd tpm_tis video mei_me e1000e wmi intel_ips soundcore mei i2c_core agpgart tpm_tis_core ehci_pci lpc_ich ehci_hcd mfd_core evdev loop
Dec 16 13:12:17 arwen kernel: CPU: 1 PID: 513 Comm: udevd Not tainted 6.6.7 #1
Dec 16 13:12:17 arwen kernel: Hardware name: LENOVO 4349AQ2/4349AQ2, BIOS 6MET72WW (1.32 ) 08/24/2010
Dec 16 13:12:17 arwen kernel: RIP: 0010:ilk_wm_sanitize+0x2be/0x2e0 [i915]
Dec 16 13:12:17 arwen kernel: Code: 0b e9 61 ff ff ff 48 8b 7b 08 48 8b 5f 50 48 85 db 74 2b e8 a4 f9 36 cb 48 89 da 48 c7 c7 00 ec b7 c0 48 89 c6 e8 42 e7 73 ca <0f> 0b e9 12 ff ff ff 48 8b 1f e9 79 ff ff ff 48 8b 1f eb a3 48 8b
Dec 16 13:12:17 arwen kernel: RSP: 0018:ffffba8440f9fa80 EFLAGS: 00010282
Dec 16 13:12:17 arwen kernel: RAX: 0000000000000000 RBX: ffff902b80e57880 RCX: 0000000000000000
Dec 16 13:12:17 arwen kernel: RDX: 0000000000000002 RSI: 00000000ffffdfff RDI: 00000000ffffffff
Dec 16 13:12:17 arwen kernel: RBP: ffffba8440f9fa80 R08: 0000000000000000 R09: ffffba8440f9f920
Dec 16 13:12:17 arwen kernel: R10: 0000000000000003 R11: ffffffff8cf3df28 R12: ffff902b85b15000
Dec 16 13:12:17 arwen kernel: R13: ffff902b890e82c0 R14: ffff902b8098e0c0 R15: ffff902b8098e000
Dec 16 13:12:17 arwen kernel: FS:  00007f5fec06e740(0000) GS:ffff902cb3c80000(0000) knlGS:0000000000000000
Dec 16 13:12:17 arwen kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Dec 16 13:12:17 arwen kernel: CR2: 00007f647533ba98 CR3: 0000000103da4000 CR4: 00000000000006e0
Dec 16 13:12:17 arwen kernel: Call Trace:
Dec 16 13:12:17 arwen kernel:  <TASK>
Dec 16 13:12:17 arwen kernel:  ? ilk_wm_sanitize+0x2be/0x2e0 [i915]
Dec 16 13:12:17 arwen kernel:  ? __warn+0x7d/0x130
Dec 16 13:12:17 arwen kernel:  ? ilk_wm_sanitize+0x2be/0x2e0 [i915]
Dec 16 13:12:17 arwen kernel:  ? report_bug+0x18d/0x1c0
Dec 16 13:12:17 arwen kernel:  ? handle_bug+0x3a/0x70
Dec 16 13:12:17 arwen kernel:  ? exc_invalid_op+0x13/0x60
Dec 16 13:12:17 arwen kernel:  ? asm_exc_invalid_op+0x16/0x20
Dec 16 13:12:17 arwen kernel:  ? ilk_wm_sanitize+0x2be/0x2e0 [i915]
Dec 16 13:12:17 arwen kernel:  intel_display_driver_probe_nogem+0x1f3/0x210 [i915]
Dec 16 13:12:17 arwen kernel:  i915_driver_probe+0x6ce/0xb80 [i915]
Dec 16 13:12:17 arwen kernel:  local_pci_probe+0x41/0x90
Dec 16 13:12:17 arwen kernel:  pci_device_probe+0xbd/0x220
Dec 16 13:12:17 arwen kernel:  really_probe+0x166/0x300
Dec 16 13:12:17 arwen kernel:  ? __pfx___driver_attach+0x10/0x10
Dec 16 13:12:17 arwen kernel:  __driver_probe_device+0x73/0x120
Dec 16 13:12:17 arwen kernel:  driver_probe_device+0x1f/0xe0
Dec 16 13:12:17 arwen kernel:  __driver_attach+0x84/0x180
Dec 16 13:12:17 arwen kernel:  bus_for_each_dev+0x84/0xd0
Dec 16 13:12:17 arwen kernel:  bus_add_driver+0xe8/0x1f0
Dec 16 13:12:17 arwen kernel:  driver_register+0x55/0x100
Dec 16 13:12:17 arwen kernel:  i915_init+0x21/0xc0 [i915]
Dec 16 13:12:17 arwen kernel:  ? __pfx_i915_init+0x10/0x10 [i915]
Dec 16 13:12:17 arwen kernel:  do_one_initcall+0x59/0x310
Dec 16 13:12:17 arwen kernel:  do_init_module+0x60/0x230
Dec 16 13:12:17 arwen kernel:  init_module_from_file+0x86/0xc0
Dec 16 13:12:17 arwen kernel:  idempotent_init_module+0x120/0x2b0
Dec 16 13:12:17 arwen kernel:  __x64_sys_finit_module+0x52/0x90
Dec 16 13:12:17 arwen kernel:  do_syscall_64+0x3b/0x90
Dec 16 13:12:17 arwen kernel:  entry_SYSCALL_64_after_hwframe+0x6e/0xd8
Dec 16 13:12:17 arwen kernel: RIP: 0033:0x7f5febf1bc79
Dec 16 13:12:17 arwen kernel: Code: ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 6f 11 13 00 f7 d8 64 89 01 48
Dec 16 13:12:17 arwen kernel: RSP: 002b:00007ffd797ece18 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
Dec 16 13:12:17 arwen kernel: RAX: ffffffffffffffda RBX: 0000000001e04f90 RCX: 00007f5febf1bc79
Dec 16 13:12:17 arwen kernel: RDX: 0000000000000000 RSI: 00007f5fec0b6355 RDI: 0000000000000013
Dec 16 13:12:17 arwen kernel: RBP: 0000000000000000 R08: 0000000001e24bc0 R09: 0000000000000000
Dec 16 13:12:17 arwen kernel: R10: 0000000000000050 R11: 0000000000000246 R12: 00007f5fec0b6355
Dec 16 13:12:17 arwen kernel: R13: 0000000000020000 R14: 0000000001e15230 R15: 000000000aba9500
Dec 16 13:12:17 arwen kernel:  </TASK>
Dec 16 13:12:17 arwen kernel: ---[ end trace 0000000000000000 ]---

Kernel-6.1.66 does not have the above error, but does have a call trace regarding intel_modeset_init warning/error that does not stop normal frame buffer or X.org operation.

There is some chatter on fedora, redhat, *buntoo etc on similar problems with kernel 6.5.x and 6.6.x.

Posting here to see if any other old timers with antique laptops are also having intel GPU issues.

RandomTroll · 12-16-2023, 10:59 PM

I run 6.6.7, Slackware, i915 video, have no problem. I build my own kernel.

zeebra · 12-17-2023, 06:36 AM

Quote:

Originally Posted by kingbeowulf

There is some chatter on fedora, redhat, *buntoo etc on similar problems with kernel 6.5.x and 6.6.x.

Posting here to see if any other old timers with antique laptops are also having intel GPU issues.

Edit.Unrelated..

Perhaps there is some change in the Kernel (series) that we're unaware of (Perhaps study Linus release statements for Kernel 6.5 and 6.6 on Kernel mailing lists).
I assume you installed a new firmware package as well.
You use any boot parameters with that?

Anyways, your trace points to udevd being the problem in this case.. And/or check line 3497 in drivers/gpu/drm/i915/display/i9xx_wm.c:3497, for further details.

Might be somewhat related, some regression back in 2018, causing a similar issue with modprobe.
https://bugs.freedesktop.org/show_bug.cgi?id=105992

cwizardone · 12-17-2023, 08:24 AM

Quote:

Slackware64-current kernel 6.6.7 and i915 module crash and blank screen

Something like this happened recently with the 6.6.7 kernel and the nvidia-NVIDIA-Linux-x86_64-545.29.06.run (new features) driver on this desktop. The screen blinked and the screen went blank (black). I could move the cursor, but nothing else worked, including the reset button. Finally, it became necessary to pull the plug.
With the 6.6.6 and 6.7-rc2 kernels there were spontaneous reboots, so I've switched to the NVIDIA-Linux-x86_64-535.129.03.run (production) driver. So far, so good. Perhaps, the problem could be with the recently updated mesa or xorg packages?
Just a thought.

marav · 12-17-2023, 08:51 AM

Mesa sometimes (constantly?) drops old GPUs support
I don't think Slackware-current is the best choice for old hardware

But, you could give this a try:
https://www.linuxquestions.org/quest...ml#post6409053

kingbeowulf · 12-17-2023, 12:03 PM

@marav, I don't think this is a mesa issue, since screen stops during boot. system is up, just headless with no intel frame buffer and hence no video on the laptop panel.

@RandomTroll, how old is your intel GPU? Recent or vintage like mine?

@cwizardone, The nvidia issue is was different. Slackware loads and runs as expected, via ssh remote headless, video issue starts when i915 and drm loads during boot, way before login and before startx. I can even log in (blind) and try to 'startx'. It tries, finds monitor params, modes, EDID, but is unable to set a modeline.

As I said in the OP, fully updated Slackware64-Current with kernel-6.6.7, no intel video, and rolling back ONLY the kernel to 6.1.66 restores full function including X.Org and Xfce. No boot parameters. It's not some old bug from 2018 since 14.x through 15.0 and Current up to now have worked just fine. No initrd. Same modules loaded. hardware is detected.

LILO is simple:

Code:

# Start LILO global section
boot = /dev/sda
lba32

# This option loads the kernel and initrd much faster:
#compact

# Boot BMP Image.
# Bitmap in BMP format: 640x480x8
  bitmap = /boot/slack.bmp
# Menu colors (foreground, background, shadow, highlighted
# foreground, highlighted background, highlighted shadow):
  bmp-colors = 255,0,255,0,255,0
# Location of the option table: location x, location y, number of
# columns, lines per column (max 15), "spill" (this is how many
# entries must be in the first column before the next begins to
# be used. We don't specify it here, as there's just one column.
  bmp-table = 60,6,1,16
# Timer location x, timer location y, foreground color,
# background color, shadow color.
  bmp-timer = 65,27,0,255

# Append any additional kernel parameters:
append=""
prompt
timeout = 50

# VESA framebuffer console @ 1024x768x256
vga = 773
# Normal VGA console
#vga = normal

# Linux bootable partition config begins
image = /boot/vmlinuz-huge-6.1.66
  root = /dev/sda1
  label = Slack61
  read-only

image = /boot/vmlinuz-huge-6.6.7
  root = /dev/sda1
  label = Slackware64
  read-only
# Linux bootable partition config ends

I'll take a look later at 6.6.7 drivers/gpu/drm/i915/display/i9xx_wm.c:3497 and compare it to the one in 6.1.66.

zeebra · 12-17-2023, 12:49 PM

Quote:

Originally Posted by kingbeowulf

no intel video, and rolling back ONLY the kernel to 6.1.66 restores full function including X.Org and Xfce.

Which should make you wonder why it does that.. It seems some issue was introduced in newer Kernels somehow. Which should lead to the analyzes of the trace, and that particular line which points to drm issues (according to that line 3497), and it specifically mentions udevd as the reason. Unlike that old similar bug which mentioned modprobe as the reason. But it's a warning, so..

Perhaps see if there is something relevant here:
https://lkml.org/lkml/2023/10/30/11
And there should be a similar one for 6.5
https://lkml.org/lkml/2023/8/27/348

This line: "Could not determine valid watermarks for inherited state" probably is related if you trace that line 3497 to something relevant.

Meanwhile you could test with "nomodeset" as a Kernel parameter.

RandomTroll · 12-18-2023, 05:28 AM

Quote:

Originally Posted by kingbeowulf

@RandomTroll, how old is your intel GPU? Recent or vintage like mine?

Code:

intel_gpu_top

returns

Quote:

Intel Haswell (Gen7)

The computer is 10 years old.

kingbeowulf · 12-19-2023, 07:26 PM

Well, this is weird. Since this lappy INtell GPU seems to be still supported in the kernel, and only slightly older than RandomTroll, and I use it to test stuff, I decided to reinstall a fresh Slackware64-currnt as of 17-Dec updates. Blank screen, but otherwise operational (headless access via ssh) Tried "nomodeset i915_modeset=0" and I get a nice old timey 1024x768 frame buffer and X (Xfce) works. No GPU acceleration or other resolution options.

But wait, there's more!

I fire up the latest Slackware64-live-current (kernel 6.6.7) on a USB and *POOF* everything works. Pretty KDE/Plasma. Correct resolution. xrandr reports choices for LVDS-1 and the presence of the other VGA and DP ports. Weird.

So now I have to figure out why liveslak-current works but a regular clean, full install with lilo (MBR) boot does not. The Game's AFOOT!

zeebra · 12-20-2023, 01:31 AM

Hmm.. So, the driver and firmware of 6.6.7 works (given the right conditions at least).

What happens if you boot into the installation WITH the Slackware installation disk (as described initially when booting disk)? (skip lilo)

mudherm · 12-21-2023, 07:20 PM

Sadly, upgrading to Slackware-Current (6.6.8 Kernel) didn't change anything.

I have a Sandybridge CPU.

henca · 12-22-2023, 12:32 AM

Quote:

Originally Posted by kingbeowulf

So now I have to figure out why liveslak-current works but a regular clean, full install with lilo (MBR) boot does not.

I haven't tried liveslak-current myself, but could it be that it boots with UEFI rather than the good old legacy way? UEFI handles graphics differently.

regards Henrik

franzen · 12-23-2023, 05:00 AM

Quote:

Originally Posted by kingbeowulf

I fire up the latest Slackware64-live-current (kernel 6.6.7) on a USB and *POOF* everything works.

Is there a different output if you do

Code:

grep "acceleration enabled" /var/log/Xorg.0.log

on the live-current
and on your installed -current?

kingbeowulf · 12-23-2023, 06:43 PM

This laptop is pre-UEFI, so uses old time BIOS boot. The slackware64-currt install boots fine since huge.s does set "nomodeset". There is kms.s which replicates the problem. liveslak USB boots via syslinux and I haven't looked into those boot parameters yet,

I did discover, that once laptop is booted, pressing Fn-F4 key combo to hibernate (or closing the lid, which I stumbled upon), and the pressing power button to wake up, resets the LCD screen and all is normal. Screen power management? Is this dummy selecting on of the unused video ports (VGA, DisplayPort)?

'acpi=off' is no good. At least now I know where to go looking.

@franzen, re-read my earlier posts. This blank screen during/after boot occurs BEFORE running X.org with 'startx'

kgha · 12-24-2023, 01:19 AM

No display issues with a fully upgraded -current (kernel 6.6.8) on my old (15 yrs) Asus eee 900. Inxi gives the following gpu info:

Code:

intel mobile 915GM/GMS/910GML Express Graphic

A lot of reasons why the eee900 is not the optimal hardware for the latest -current, but outdated graphics not among them.