Reporting a related case on a different Lenovo model with the same
Phoenix/gfx1103 GPU family and same 6.8.x kernel track — flagging
because amdgpu.runpm=0 has NOT fully resolved my case the way it did
here, which may indicate a distinct-but-related issue worth tracking
separately.

System Information
Model: Lenovo ThinkPad T14s Gen 4, Type 21F9CTO1WW

CPU: AMD Ryzen 7 PRO 7840U w/ Radeon 780M Graphics (Phoenix, gfx1103_r1)

BIOS: R2EET49W (1.30), 03/18/2026

OS: Ubuntu 22.04.5 LTS (Xubuntu/XFCE 4.16.0)

Kernel: 6.8.0-136-generic (HWE track — this is the ceiling for 22.04; no
6.17/linux-oem-24.04 path exists on this release to test the upgrade
route documented in comment #6)

linux-firmware: 20220329.git681281e4-0ubuntu3.42

Storage: Samsung MZVL21T0HDLU-00BLL (1TB NVMe), firmware 6L2QGXD7

Why I'm flagging this as possibly distinct
Comment #6 confirms amdgpu.runpm=0 gives fully reliable, deterministic boots on 
the IdeaPad Slim 3 (Phoenix/Radeon 740M) — the freeze reportedly does not recur 
once the parameter is set, across both 6.8.0-100 and 6.17.0-22.

On my hardware (Phoenix/Radeon 780M), amdgpu.runpm=0 has been active
continuously since 2026-08-13, and the boot freeze still recurs
intermittently — roughly once every 4–11 cold boots — rather than being
fully eliminated. This is a different pattern than the deterministic
freeze-every-time-without-the-parameter behavior described here, which
makes me wonder whether this is genuinely the same root cause, or a
related-but-separate amdgpu/Phoenix boot-path issue that happens to
share the same workaround as a partial mitigation.

Symptom
Blank screen with blinking cursor top-left, immediately after selecting the 
boot entry at GRUB, before any kernel/userspace logging begins. Fully 
unresponsive, requires hard power-off. Not reproducible on warm reboots — only 
cold boots.

Steps taken (chronological)
BIOS updated 0.1.24 → 0.1.30 — no change in frequency.

Added nvme_core.default_ps_max_latency_us=5500 — hang recurred
afterward.

Tested GRUB_TERMINAL=console to rule out GRUB gfxterm/amdgpu handoff —
hang recurred, since reverted (ruled out).

Added amdgpu.runpm=0 (based on this bug) on 2026-08-13 — hang has
recurred at least twice since (2026-08-20, 2026-08-24), so the
workaround that fully resolved comment #6's case has only partially
helped, if at all, in mine.

Currently also testing amdgpu.dcdebugmask=0x10 (disables PSR) alongside
runpm=0, added 2026-08-25 — result pending.

Current kernel command line
text
quiet nvme_core.default_ps_max_latency_us=5500 amdgpu.runpm=0 
amdgpu.dcdebugmask=0x10 usbcore.autosuspend=-1 i8042.nomux=1 i8042.reset 
i2c_hid.polling_mode=1
Log evidence
No journal entries are ever produced for the failed boots — journalctl -b -1 -p 
err..emerg, journalctl -b -1 -u lightdm, and Xorg.0.log all show clean output 
on the subsequent successful boot, confirming the freeze happens before 
systemd-journald initializes (GRUB-to-kernel handoff or very early 
kernel/initramfs stage).

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2141725

Title:
  Boot freezes unless amdgpu.runpm=0 on Lenovo IdeaPad Slim 3 (Phoenix /
  Radeon 740M)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2141725/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to