]> Git Repo - linux.git/commit
x86/perf: Use RET0 as default for guest_get_msrs to handle "no PMU" case
authorSean Christopherson <[email protected]>
Tue, 9 Mar 2021 17:10:19 +0000 (09:10 -0800)
committerPeter Zijlstra <[email protected]>
Wed, 10 Mar 2021 15:45:09 +0000 (16:45 +0100)
commitc8e2fe13d1d1f3a02842b7b909d4e4846a4b6a2c
tree61e5e05f3c0fabb39dd8db94478c402632f9cd8a
parentafbef30149587ad46f4780b1e0cc5e219745ce90
x86/perf: Use RET0 as default for guest_get_msrs to handle "no PMU" case

Initialize x86_pmu.guest_get_msrs to return 0/NULL to handle the "nop"
case.  Patching in perf_guest_get_msrs_nop() during setup does not work
if there is no PMU, as setup bails before updating the static calls,
leaving x86_pmu.guest_get_msrs NULL and thus a complete nop.  Ultimately,
this causes VMX abort on VM-Exit due to KVM putting random garbage from
the stack into the MSR load list.

Add a comment in KVM to note that nr_msrs is valid if and only if the
return value is non-NULL.

Fixes: abd562df94d1 ("x86/perf: Use static_call for x86_pmu.guest_get_msrs")
Reported-by: Dmitry Vyukov <[email protected]>
Reported-by: [email protected]
Suggested-by: Peter Zijlstra <[email protected]>
Signed-off-by: Sean Christopherson <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
arch/x86/events/core.c
arch/x86/kvm/vmx/vmx.c
This page took 0.051867 seconds and 4 git commands to generate.