Groups | Search | Server Info | Login | Register

Groups > linux.gentoo.user > #116059

[gentoo-user] Hard kernel lock with gentoo-sources-6.18.18

Show all headers | View raw

[Multipart message — attachments visible in raw view] - view raw

I upgraded my kernel from gentoo-sources-6.18.12 to gentoo-
sources-6.18.18 and noticed the kernel crashes just as SDDM is being 
launched.  The only thing I can do is press Ctrl+Alt+Del which allows a 
reboot.

On the odd occasion, typically straight after a reboot from 6.18.12, I 
can launch SDDM and I am able to login with the 6.18.18 kernel.  In 
these rare occasions I noticed this in dmesg, which shows a single CPU 
core crashing:

[   30.949469] ata3.00: configured for UDMA/133
[   30.973669] nvme nvme0: using unchecked data buffer
[   57.775566] ------------[ cut here ]------------
[   57.775568] atomic remove_fb failed with -22
[   57.775575] WARNING: CPU: 1 PID: 221 at drivers/gpu/drm/
drm_framebuffer.c:1176 drm_framebuffer_remove+0x58d/0x5a0
[   57.775582] Modules linked in: xt_hl xt_CT xt_helper nf_conntrack_ftp 
xt_set ip_set_hash_net ip_set nfnetlink ip6t_rt iptable_nat nf_nat 
xt_TCPMSS xt_LOG nf_log_syslog ip6t_REJECT nf_reject_ipv6 ipt_REJECT 
nf_reject_ipv4 ip6table_raw iptable_raw ip6table_mangle iptable_mangle 
xt_multiport xt_state xt_limit xt_conntrack nf_conntrack nf_defrag_ipv6 
nf_defrag_ipv4 ip6table_filter ip6_tables iptable_filter ip_tables 
dm_crypt algif_skcipher af_alg asus_wmi_sensors nct6775 nct6775_core 
hwmon_vid ee1004 jc42 asus_ec_sensors regmap_i2c tpm_crb tpm_tis 
tpm_tis_core wmi_bmof i2c_piix4 tpm k10temp i2c_smbus libaescfb
[   57.775635] CPU: 1 UID: 0 PID: 221 Comm: kworker/1:1 Not tainted 
6.18.18-gentoo #2 PREEMPT(voluntary) 
[   57.775638] Hardware name: ASUS System Product Name/ROG CROSSHAIR 
VIII DARK HERO, BIOS 5302 10/03/2025
[   57.775640] Workqueue: events drm_mode_rmfb_work_fn
[   57.775644] RIP: 0010:drm_framebuffer_remove+0x58d/0x5a0
[   57.775647] Code: 27 eb d6 48 8d 7c 24 18 e8 00 e7 00 00 48 8d 7c 24 
18 e8 a6 e4 00 00 ba f4 ff ff ff 89 d6 48 c7 c7 f8 3b e0 b0 e8 e3 af 6f 
ff <0f> 0b e9 f6 fc ff ff 31 ff eb 9b e8 d3 ae b5 00 0f 1f 00 90 90 90
[   57.775649] RSP: 0018:ffffcd9780be3d98 EFLAGS: 00010246
[   57.775652] RAX: 0000000000000000 RBX: ffff8aea192b4900 RCX: 
0000000000000000
[   57.775653] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 
0000000000000000
[   57.775655] RBP: ffff8aea0208e100 R08: 0000000000000000 R09: 
0000000000000000
[   57.775657] R10: 0000000000000000 R11: 0000000000000000 R12: 
ffff8aea06480010
[   57.775658] R13: 0000000000000004 R14: 00000000ffffffea R15: 
0000000000000000
[   57.775660] FS:  0000000000000000(0000) GS:ffff8af93cb10000(0000) 
knlGS:0000000000000000
[   57.775662] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   57.775663] CR2: 00007f99ba56e6dc CR3: 00000001204fe000 CR4: 
0000000000750ef0
[   57.775665] PKRU: 55555554
[   57.775667] Call Trace:
[   57.775668]  <TASK>
[   57.775671]  drm_mode_rmfb_work_fn+0x79/0xa0
[   57.775674]  process_one_work+0x18d/0x350
[   57.775679]  worker_thread+0x199/0x320
[   57.775682]  ? __pfx_worker_thread+0x10/0x10
[   57.775685]  kthread+0xf7/0x220
[   57.775688]  ? __pfx_kthread+0x10/0x10
[   57.775691]  ? __pfx_kthread+0x10/0x10
[   57.775693]  ret_from_fork+0x141/0x1c0
[   57.775697]  ? __pfx_kthread+0x10/0x10
[   57.775699]  ? __pfx_kthread+0x10/0x10
[   57.775701]  ret_from_fork_asm+0x1a/0x30
[   57.775706]  </TASK>
[   57.775707] ---[ end trace 0000000000000000 ]---
[   59.013359] block nvme0n1: No UUID available providing old NGUID

It seems to me when the efifb tries to hand over the graphics to SDDM 
the kernel falls apart.  I assume the hard lockups happen when more than 
one core crashes, but I don't know.  When the kernel locks up the SDDM 
log shows this:

[18:26:57.111] (II) HELPER: Jumping to VT 7
[18:26:57.112] (II) HELPER: VT mode didn't need to be fixed
[18:26:57.112] (II) DAEMON: Greeter session started successfully
[18:26:57.113] (II) HELPER: [PAM] Closing session
[18:26:57.113] (II) HELPER: [PAM] Ended.
[18:26:57.116] (WW) DAEMON: Auth: sddm-helper exited with 63
[18:26:57.116] (II) DAEMON: Greeter stopped. 
SDDM::Auth::HelperExitStatus(63)
[18:27:18.714] (WW) DAEMON: Signal received: SIGTERM
[18:27:18.714] (II) DAEMON: Socket server stopping...
[18:27:18.714] (II) DAEMON: Socket server stopped.

The only difference between 6.18.12 and 6.18.18 kernel config is these 
two new lines:

> CONFIG_UNWIND_USER=y
> CONFIG_HAVE_UNWIND_USER_FP=y

NOTES: 
=====
1. If I disable the display-manager service from the default runlevel 
there is no crash.  The moment I try to start the display-manager 
manually from the console the kernel crashes as described above.

2. SDDM is configured to launch using Wayland with the kwin compositor.  
It has been doing so successfully for years now  and it still does with 
kernel 6.18.12.  I have not tried (yet) to launch SDDM in X with 
6.18.18.

3. This system is running a Zen 3 CPU with a Radeon graphics card and 
the AMDGPU driver.

Have you had any similarly unpleasant experience with 6.18.18?

Any idea how I could fix this?

Back to linux.gentoo.user | Previous | Next — Next in thread | Find similar

Thread

[gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 Michael <confabulate@kintzios.com> - 2026-03-23 21:30 +0100
  Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 Michael <confabulate@kintzios.com> - 2026-03-24 00:30 +0100
    Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 Michael <confabulate@kintzios.com> - 2026-03-25 19:30 +0100
      Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 "D. Ben Knoble" <ben.knoble@gmail.com> - 2026-03-26 00:40 +0100
        Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 Michael <confabulate@kintzios.com> - 2026-03-26 09:50 +0100
          Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 "D. Ben Knoble" <ben.knoble@gmail.com> - 2026-03-26 17:00 +0100
            Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 "D. Ben Knoble" <ben.knoble@gmail.com> - 2026-03-26 18:00 +0100
          Re: [gentoo-user] Hard kernel lock with gentoo-sources-6.18.18 Michael <confabulate@kintzios.com> - 2026-04-28 20:20 +0200

csiph-web