Most likely it will be fixed very soon indeed. (If someone reports it 
)
Personally I just enjoy figuring out the root cause and digging into those things.
It’s somewhat fun for me and I enjoy it (and ideally also learn something). 
2 Likes
It’s not working on my Ryzen 5500U and also on the desktop Ryzen 3800X which both worked before. I will wait til next kernel updates. My Ryzen 4700U never did work.
Neither did on my Ryzen 4800U 
1 Like
Has the mitigation for this been already implemented in 5.18.11 ?
This is what inxi -aC shows regarding the vulnerabilities:
Vulnerabilities:
Type: itlb_multihit status: Not affected
Type: l1tf status: Not affected
Type: mds status: Not affected
Type: meltdown status: Not affected
Type: mmio_stale_data status: Not affected
Type: spec_store_bypass
mitigation: Speculative Store Bypass disabled via prctl
Type: spectre_v1
mitigation: usercopy/swapgs barriers and __user pointer sanitization
Type: spectre_v2 mitigation: Retpolines, IBPB: conditional, IBRS_FW,
STIBP: conditional, RSB filling
Type: srbds status: Not affected
Type: tsx_async_abort status: Not affected
same here
also in this case I do simple have no error or anything shown any indication on why it does not work…
Actually no. I was jumping between the mainline and latest trees and just saw retbleed here and there 
Sorry. Thanks for noticing.
1 Like
i dont think that parameter is a thing
OP is still working on my end
yes you need shared mem for desktop zen3
AMD Ryzen™ 7 Mobile Processors
so also mobile one and working … (like the one from @fbodymechanic )
2 Likes
Working on my latest machine with 5.18.11…
just updated to .12 and is still not there for Desktop CPU
[16:37:17] joekamprad :: i3-private ➜ ~ » uname -a
Linux i3-private 5.18.12-AMD #1 SMP PREEMPT_DYNAMIC Fri Jul 15 10:46:59 CEST 2022 x86_64 GNU/Linux
i see this in journal:
Jul 15 16:31:26 i3-private kernel: amd_uncore: 4 amd_df counters detected
Jul 15 16:31:26 i3-private kernel: amd_uncore: 6 amd_l3 counters detected
Jul 15 16:31:26 i3-private kernel: perf/amd_iommu: Detected AMD IOMMU #0 (2 banks, 4 counters/bank).
Jul 15 16:31:26 i3-private systemd-modules-load[426]: Module 'amd_pstate' is built in
looks like maybe loading with “/etc/modules-load.d/amd-pstate.conf” may no longer be needed? or are you running a custom kernel that may have it built in?
yes … but should not harm loading it … i use linux-amd kernel where it is set to =y …
https://lore.kernel.org/all/3559249.JlDtxWtqDm@natalenko.name/
validation @jonathon
2 Likes
that would likely be why it still works for me, im on 5.18.10 lol
1 Like
Same here though 5800H laptop but also get acpi cpufreq since .11 and now also with .12 (Linux-amd kernel) .10 everything was working fine. I’ve seen this also remarked already in a Fedora forum thread so we’re not alone in this.
cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_driver gives me: acpi-cpufreq
/etc/modules-load.d/amd-pstate.conf I added it again (it disappeared after updating to .11 and again after going to .12) and rebooted but still acpi-cpufreq stays in place once rebooted. (amd-pstate.conf file did stay in place once rebooted)
executing watch -n1 “grep "^[c]pu MHz" /proc/cpuinfo” gives me 1200Mhz on all threads while 400 was set over cpupower frequency-set -u 400 (400 is the lowest for the 5800H). Setting 400 again has no effect whatsoever.
Putting some stress on the cpu doesn’t raise the core frequencies over +/- 1500MHz when checking with watch -n1 “grep "^[c]pu MHz" /proc/cpuinfo”
Really strange, i guess 2 options, rollback to .10 or wait for .13 and hope it’s solved in .13
1 Like
stress test result with s-tui shows cpu doesn’t go higher than 1.2Ghz
will give you the same on hardware limits… like only showing 1.2GHz as lowest and nothing lower…
1 Like
Seems fix is on the way: https://bugzilla.kernel.org/show_bug.cgi?id=216248
2 Likes
5.18.13 will have the needed fix !!
https://lwn.net/ml/linux-kernel/20220721182818.743726259%40linuxfoundation.org/
4 Likes
5.18.13 will soon be in the testing area, and a bit later as official.
