Observed on two AMD Ryzen systems, one desktop with vendor_id : AuthenticAMD cpu family : 23 model : 96 model name : AMD Ryzen 7 PRO 4750G with Radeon Graphics and one laptop with vendor_id : AuthenticAMD cpu family : 23 model : 104 model name : AMD Ryzen 5 5500U with Radeon Graphics On selecting suspend to RAM, the display is powered off, but the power LED remains lit, suggesting the problem lies with suspend, not resume. The only way to recover is a long press on the power button to force the machine off. The last few messages in the system journal are systemd[1]: Reached target sleep.target. wpa_supplicant[1257]: p2p-dev-wlp5s0: CTRL-EVENT-DSCP-POLICY clear_all wpa_supplicant[1257]: p2p-dev-wlp5s0: CTRL-EVENT-DSCP-POLICY clear_all wpa_supplicant[1257]: nl80211: deinit ifname=p2p-dev-wlp5s0 disabled_11b_rates=0 systemd[1]: Starting systemd-suspend.service... systemd-sleep[2206]: Entering sleep state 'suspend'... kernel: PM: suspend entry (deep) Installing and using kernel-desktop-6.1.45-1.mga8-1-1.mga8 from Mageia 8 backports fixes the bug. And there was no such problem in Mageia 8 using the 5.x series kernels.
Thank you for the report. Can we take it that this relates to recent kernels (please say which one gave you the problem; I currently see 6.4.9), since at least 6.1.45 works. Assigning to the kernel maintainer.
Assignee: bugsquad => kernel
Amd is working on fixing up suspend/resume issues on Amd platforms, but no ETA for it landing in Linus tree yet... so far they are up to revision 16 on their patchsets ...
Is this something to note for errata? Or just silently accept there are many such problems nowadays? (or else we would have a long list of obscure resume issues...)
CC: (none) => fri
This is fixed in kernel-desktop-6.5.11-5.mga9
Resolution: (none) => FIXEDStatus: NEW => RESOLVED