Hi Mark, On 2/4/25 4:20 PM, Mark Rutland wrote: > There are several problems with the way hyp code lazily saves the host's > FPSIMD/SVE state, including: > > * Host SVE being discarded unexpectedly due to inconsistent > configuration of TIF_SVE and CPACR_ELx.ZEN. This has been seen to > result in QEMU crashes where SVE is used by memmove(), as reported by > Eric Auger: > > https://issues.redhat.com/browse/RHEL-68997 I tested the above test case with the whole series. Tested-by: Eric Auger <eric.auger@xxxxxxxxxx> Thanks Eric > > * Host SVE state is discarded *after* modification by ptrace, which was an > unintentional ptrace ABI change introduced with lazy discarding of SVE state. > > * The host FPMR value can be discarded when running a non-protected VM, > where FPMR support is not exposed to a VM, and that VM uses > FPSIMD/SVE. In these cases the hyp code does not save the host's FPMR > before unbinding the host's FPSIMD/SVE/SME state, leaving a stale > value in memory. > > Avoid these by eagerly saving and "flushing" the host's FPSIMD/SVE/SME > state when loading a vCPU such that KVM does not need to save any of the > host's FPSIMD/SVE/SME state. For clarity, fpsimd_kvm_prepare() is > removed and the necessary call to fpsimd_save_and_flush_cpu_state() is > placed in kvm_arch_vcpu_load_fp(). As 'fpsimd_state' and 'fpmr_ptr' > should not be used, they are set to NULL; all uses of these will be > removed in subsequent patches. > > Historical problems go back at least as far as v5.17, e.g. erroneous > assumptions about TIF_SVE being clear in commit: > > 8383741ab2e773a9 ("KVM: arm64: Get rid of host SVE tracking/saving") > > ... and so this eager save+flush probably needs to be backported to ALL > stable trees. > > Fixes: 93ae6b01bafee8fa ("KVM: arm64: Discard any SVE state when entering KVM guests") > Fixes: 8c845e2731041f0f ("arm64/sve: Leave SVE enabled on syscall if we don't context switch") > Fixes: ef3be86021c3bdf3 ("KVM: arm64: Add save/restore support for FPMR") > Reported-by: Eric Auger <eauger@xxxxxxxxxx> > Reported-by: Wilco Dijkstra <wilco.dijkstra@xxxxxxx> > Cc: stable@xxxxxxxxxxxxxxx > Cc: Catalin Marinas <catalin.marinas@xxxxxxx> > Cc: Florian Weimer <fweimer@xxxxxxxxxx> > Cc: Fuad Tabba <tabba@xxxxxxxxxx> > Cc: Jeremy Linton <jeremy.linton@xxxxxxx> > Cc: Marc Zyngier <maz@xxxxxxxxxx> > Cc: Mark Brown <broonie@xxxxxxxxxx> > Cc: Oliver Upton <oliver.upton@xxxxxxxxx> > Cc: Paolo Bonzini <pbonzini@xxxxxxxxxx> > Cc: Will Deacon <will@xxxxxxxxxx> > Signed-off-by: Mark Rutland <mark.rutland@xxxxxxx> > --- > arch/arm64/kernel/fpsimd.c | 25 ------------------------- > arch/arm64/kvm/fpsimd.c | 35 ++++++++++------------------------- > 2 files changed, 10 insertions(+), 50 deletions(-) > > diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c > index 2b601d88762d4..8370d55f03533 100644 > --- a/arch/arm64/kernel/fpsimd.c > +++ b/arch/arm64/kernel/fpsimd.c > @@ -1694,31 +1694,6 @@ void fpsimd_signal_preserve_current_state(void) > sve_to_fpsimd(current); > } > > -/* > - * Called by KVM when entering the guest. > - */ > -void fpsimd_kvm_prepare(void) > -{ > - if (!system_supports_sve()) > - return; > - > - /* > - * KVM does not save host SVE state since we can only enter > - * the guest from a syscall so the ABI means that only the > - * non-saved SVE state needs to be saved. If we have left > - * SVE enabled for performance reasons then update the task > - * state to be FPSIMD only. > - */ > - get_cpu_fpsimd_context(); > - > - if (test_and_clear_thread_flag(TIF_SVE)) { > - sve_to_fpsimd(current); > - current->thread.fp_type = FP_STATE_FPSIMD; > - } > - > - put_cpu_fpsimd_context(); > -} > - > /* > * Associate current's FPSIMD context with this cpu > * The caller must have ownership of the cpu FPSIMD context before calling > diff --git a/arch/arm64/kvm/fpsimd.c b/arch/arm64/kvm/fpsimd.c > index 4d3d1a2eb1570..ceeb0a4893aa7 100644 > --- a/arch/arm64/kvm/fpsimd.c > +++ b/arch/arm64/kvm/fpsimd.c > @@ -54,16 +54,18 @@ void kvm_arch_vcpu_load_fp(struct kvm_vcpu *vcpu) > if (!system_supports_fpsimd()) > return; > > - fpsimd_kvm_prepare(); > - > /* > - * We will check TIF_FOREIGN_FPSTATE just before entering the > - * guest in kvm_arch_vcpu_ctxflush_fp() and override this to > - * FP_STATE_FREE if the flag set. > + * Ensure that any host FPSIMD/SVE/SME state is saved and unbound such > + * that the host kernel is responsible for restoring this state upon > + * return to userspace, and the hyp code doesn't need to save anything. > + * > + * When the host may use SME, fpsimd_save_and_flush_cpu_state() ensures > + * that PSTATE.{SM,ZA} == {0,0}. > */ > - *host_data_ptr(fp_owner) = FP_STATE_HOST_OWNED; > - *host_data_ptr(fpsimd_state) = kern_hyp_va(¤t->thread.uw.fpsimd_state); > - *host_data_ptr(fpmr_ptr) = kern_hyp_va(¤t->thread.uw.fpmr); > + fpsimd_save_and_flush_cpu_state(); > + *host_data_ptr(fp_owner) = FP_STATE_FREE; > + *host_data_ptr(fpsimd_state) = NULL; > + *host_data_ptr(fpmr_ptr) = NULL; > > host_data_clear_flag(HOST_SVE_ENABLED); > if (read_sysreg(cpacr_el1) & CPACR_EL1_ZEN_EL0EN) > @@ -73,23 +75,6 @@ void kvm_arch_vcpu_load_fp(struct kvm_vcpu *vcpu) > host_data_clear_flag(HOST_SME_ENABLED); > if (read_sysreg(cpacr_el1) & CPACR_EL1_SMEN_EL0EN) > host_data_set_flag(HOST_SME_ENABLED); > - > - /* > - * If PSTATE.SM is enabled then save any pending FP > - * state and disable PSTATE.SM. If we leave PSTATE.SM > - * enabled and the guest does not enable SME via > - * CPACR_EL1.SMEN then operations that should be valid > - * may generate SME traps from EL1 to EL1 which we > - * can't intercept and which would confuse the guest. > - * > - * Do the same for PSTATE.ZA in the case where there > - * is state in the registers which has not already > - * been saved, this is very unlikely to happen. > - */ > - if (read_sysreg_s(SYS_SVCR) & (SVCR_SM_MASK | SVCR_ZA_MASK)) { > - *host_data_ptr(fp_owner) = FP_STATE_FREE; > - fpsimd_save_and_flush_cpu_state(); > - } > } > > /*
