Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 28.04.17 at 08:45, wrote: > On 04/28/2017 09:25 AM, Jan Beulich wrote: > On 27.04.17 at 19:31, wrote: >>> On 27/04/17 09:52, Jan Beulich wrote: >>> On 27.04.17 at 10:34, wrote: > On 27/04/2017

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 04/28/2017 09:25 AM, Jan Beulich wrote: On 27.04.17 at 19:31, wrote: >> On 27/04/17 09:52, Jan Beulich wrote: >> On 27.04.17 at 10:34, wrote: On 27/04/2017 09:01, Jan Beulich wrote: On 27.04.17 at 09:22,

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 27.04.17 at 19:31, wrote: > On 27/04/17 09:52, Jan Beulich wrote: > On 27.04.17 at 10:34, wrote: >>> On 27/04/2017 09:01, Jan Beulich wrote: >>> On 27.04.17 at 09:22, wrote: > ---

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 27/04/17 09:52, Jan Beulich wrote: On 27.04.17 at 10:34, wrote: >> On 27/04/2017 09:01, Jan Beulich wrote: >> On 27.04.17 at 09:22, wrote: The introspection agent can reply to a vm_event faster than vmx_vmexit_handler()

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 04/27/17 12:00, Jan Beulich wrote: On 27.04.17 at 10:34, wrote: >> On 04/27/17 11:18, Jan Beulich wrote: >> On 27.04.17 at 10:11, wrote: On 04/27/17 11:01, Jan Beulich wrote: On 27.04.17 at 09:22,

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 27.04.17 at 10:34, wrote: > On 04/27/17 11:18, Jan Beulich wrote: > On 27.04.17 at 10:11, wrote: >>> On 04/27/17 11:01, Jan Beulich wrote: >>> On 27.04.17 at 09:22, wrote: > ---

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 27.04.17 at 10:34, wrote: > On 27/04/2017 09:01, Jan Beulich wrote: > On 27.04.17 at 09:22, wrote: >>> The introspection agent can reply to a vm_event faster than >>> vmx_vmexit_handler() can complete in some cases, where it is

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 04/27/17 11:18, Jan Beulich wrote: On 27.04.17 at 10:11, wrote: >> On 04/27/17 11:01, Jan Beulich wrote: >> On 27.04.17 at 09:22, wrote: --- a/xen/arch/x86/hvm/hvm.c +++ b/xen/arch/x86/hvm/hvm.c @@ -473,6 +473,39 @@

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 27/04/2017 09:01, Jan Beulich wrote: On 27.04.17 at 09:22, wrote: >> The introspection agent can reply to a vm_event faster than >> vmx_vmexit_handler() can complete in some cases, where it is then >> not safe for vm_event_set_registers() to modify

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 27.04.17 at 10:11, wrote: > On 04/27/17 11:01, Jan Beulich wrote: > On 27.04.17 at 09:22, wrote: >>> --- a/xen/arch/x86/hvm/hvm.c >>> +++ b/xen/arch/x86/hvm/hvm.c >>> @@ -473,6 +473,39 @@ static bool hvm_get_pending_event(struct

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 04/27/17 11:01, Jan Beulich wrote: On 27.04.17 at 09:22, wrote: >> The introspection agent can reply to a vm_event faster than >> vmx_vmexit_handler() can complete in some cases, where it is then >> not safe for vm_event_set_registers() to modify

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

>>> On 27.04.17 at 09:22, wrote: > The introspection agent can reply to a vm_event faster than > vmx_vmexit_handler() can complete in some cases, where it is then > not safe for vm_event_set_registers() to modify v->arch.user_regs. This needs more explanation: I cannot

Re: [Xen-devel] [PATCH] x86/vm_event: fix race between vmx_vmexit_handler() and vm_event_resume()

On 27/04/2017 08:22, Razvan Cojocaru wrote: > The introspection agent can reply to a vm_event faster than > vmx_vmexit_handler() can complete in some cases, where it is then > not safe for vm_event_set_registers() to modify v->arch.user_regs. > This patch ensures that vm_event_resume() code only