On Fri, Nov 18, 2016 at 08:17:25PM +0530, Aneesh Kumar K.V wrote:
> Paul Mackerras <paulus@xxxxxxxxxx> writes:
>
> > On POWER9, the msgsnd instruction is able to send interrupts to
> > other cores, as well as other threads on the local core. Since
> > msgsnd is generally simpler and faster than sending an IPI via the
> > XICS, we use msgsnd for all IPIs sent by KVM on POWER9.
> >
> > Signed-off-by: Paul Mackerras <paulus@xxxxxxxxxx>
> > ---
> > arch/powerpc/kvm/book3s_hv.c | 11 ++++++++++-
> > arch/powerpc/kvm/book3s_hv_builtin.c | 10 ++++++++--
> > 2 files changed, 18 insertions(+), 3 deletions(-)
> >
[...]
> > diff --git a/arch/powerpc/kvm/book3s_hv_builtin.c b/arch/powerpc/kvm/book3s_hv_builtin.c
> > index 0c84d6b..37ed045 100644
> > --- a/arch/powerpc/kvm/book3s_hv_builtin.c
> > +++ b/arch/powerpc/kvm/book3s_hv_builtin.c
> > @@ -205,12 +205,18 @@ static inline void rm_writeb(unsigned long paddr, u8 val)
> > void kvmhv_rm_send_ipi(int cpu)
> > {
> > unsigned long xics_phys;
> > + unsigned long msg = PPC_DBELL_TYPE(PPC_DBELL_SERVER);
> >
> > - /* On POWER8 for IPIs to threads in the same core, use msgsnd */
> > + /* On POWER9 we can use msgsnd for any destination cpu. */
> > + if (cpu_has_feature(CPU_FTR_ARCH_300)) {
> > + msg |= get_hard_smp_processor_id(cpu);
> > + __asm__ __volatile__ (PPC_MSGSND(%0) : : "r" (msg));
> > + return;
>
> Do we need a "sync" there before msgsnd ?
The comment just above this function says:
/*
* Send an interrupt or message to another CPU.
* This can only be called in real mode.
* The caller needs to include any barrier needed to order writes
* to memory vs. the IPI/message.
*/
so no. In fact all of its callers do smp_mb() before calling it.
(And no we don't want to move the smp_mb() into kvmhv_rm_send_ipi();
see kvmhv_interrupt_vcore() for why.)
Paul.
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
