Paolo Bonzini <pbonz...@redhat.com> writes:

> CPU threads write exit_request as a "note to self" that they need to
> go out to a slow path.  This write happens out of the BQL and can be
> a data race with another threads' cpu_exit(); use atomic accesses
> consistently.
>
> Signed-off-by: Paolo Bonzini <pbonz...@redhat.com>
> ---
>  accel/tcg/tcg-accel-ops-mttcg.c   | 2 +-
>  accel/tcg/tcg-accel-ops-rr.c      | 4 ++--
>  hw/ppc/spapr_hcall.c              | 6 +++---
>  target/i386/kvm/kvm.c             | 2 +-
>  target/i386/nvmm/nvmm-accel-ops.c | 2 +-
>  target/i386/nvmm/nvmm-all.c       | 2 +-
>  target/i386/whpx/whpx-all.c       | 6 +++---
>  7 files changed, 12 insertions(+), 12 deletions(-)

Could you please update include/hw/core/cpu.h with a doc patch to
describe exit_request and note it should be read/written with qatomic
primitives please.

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro

Reply via email to