Re: [PATCH 0/3] x86/KVM/VMX: Introduce and use try_cmpxchg64()

[Sean Christopherson]

On Tue, Dec 15, 2020, Uros Bizjak wrote:
> This patch series introduces try_cmpxchg64() atomic locking function.
> 
> try_cmpxchg64() provides the same interface for 64 bit and 32 bit targets,
> emits CMPXCHGQ for 64 bit targets and CMPXCHG8B for 32 bit targets,
> and provides appropriate fallbacks when CMPXCHG8B is unavailable.
> 
> try_cmpxchg64() reuses flags from CMPXCHGQ/CMPXCHG8B instructions and
> avoids unneeded CMP for 64 bit targets or XOR/XOR/OR sequence for
> 32 bit targets.
> 
> Cc: Will Deacon <will@xxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Cc: Boqun Feng <boqun.feng@xxxxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> Cc: Borislav Petkov <bp@xxxxxxxxx>
> Cc: "H. Peter Anvin" <hpa@xxxxxxxxx>
> Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> Cc: Sean Christopherson <seanjc@xxxxxxxxxx>
> Cc: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> Cc: Wanpeng Li <wanpengli@xxxxxxxxxxx>
> Cc: Jim Mattson <jmattson@xxxxxxxxxx>
> Cc: Joerg Roedel <joro@xxxxxxxxxx>
> 
> Uros Bizjak (3):
>   asm-generic/atomic: Add try_cmpxchg64() instrumentation
>   locking/atomic/x86: Introduce arch_try_cmpxchg64()
>   KVM/VMX: Use try_cmpxchg64() in posted_intr.c

For anyone else trying to apply this, it depends on v5.11-rc1 (commit
29f006fdefe6, "asm-generic/atomic: Add try_cmpxchg() fallbacks"), which hasn't
yet been merged into Paolo's tree.

>  arch/x86/include/asm/cmpxchg_32.h         | 62 +++++++++++++++++++----
>  arch/x86/include/asm/cmpxchg_64.h         |  6 +++
>  arch/x86/kvm/vmx/posted_intr.c            |  9 ++--
>  include/asm-generic/atomic-instrumented.h | 46 ++++++++++++++++-
>  scripts/atomic/gen-atomic-instrumented.sh |  2 +-
>  5 files changed, 108 insertions(+), 17 deletions(-)
> 
> -- 
> 2.26.2
>