hwpmc: fix false callchain assertion on the PMC_UR ring
ClosedPublic
Actions

Authored by afscoelho_gmail.com on Jul 31 2026, 2:21 PM.

Details

Reviewers

ali_mashtizadeh.com
mhorne
janeri_gmail.com
raghavendra.kt_amd.com
alfredo
gnn
anascime_amd.com
davichazbh_gmail.com

Group Reviewers

pmc

Commits

rG84b241f1cc12: hwpmc: fix false callchain assertion on the PMC_UR ring
rGaf3929c5152b: hwpmc: fix false callchain assertion on the PMC_UR ring

Summary

pmc_capture_user_callchain() asserts that TDP_CALLCHAIN is set on the
current thread, but PMC_UR samples never set that flag -- only PMC_HR
and PMC_SR do. That makes the assertion always fail for PMC_UR,
panicking INVARIANTS kernels as soon as pmcstat -U is used.

Skip the assertion for PMC_UR. No functional change on kernels built
without INVARIANTS.

Signed-off-by: Andre Silva <andasilv@amd.com>
Sponsored by: AMD

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

afscoelho_gmail.com created this revision.Jul 31 2026, 2:21 PM

afscoelho_gmail.com held this revision as a draft.

Herald added a subscriber: imp. · View Herald TranscriptJul 31 2026, 2:21 PM

afscoelho_gmail.com added a parent revision: D58571: hwpmc: fix false runcount assertion in user callchain capture.Jul 31 2026, 2:21 PM

Harbormaster completed remote builds in B75314: Diff 183156.Jul 31 2026, 2:22 PM

afscoelho_gmail.com published this revision for review.Jul 31 2026, 2:31 PM

afscoelho_gmail.com mentioned this in D58256: hwpmc/amd: add PerfMonV2 global-control path.Jul 31 2026, 2:37 PM

These two diffs should be merged but does it fix the -U flag bug? I don’t see why we need to keep the two path I had a patch to unify them but I hit a different assert.

In D58572#1344314, @ali_mashtizadeh.com wrote:

These two diffs should be merged but does it fix the -U flag bug? I don’t see why we need to keep the two path I had a patch to unify them but I hit a different assert.

Yes, we can merge, no problem. I just kept them separate because I thought the two assertion bugs were clearer as two small diffs. D58572 fixes the -U panic, while D58571 fixes the other runcount assertion. With both applied, -U survived my stress test.