Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-02-06 Thread Rik van Riel
On 02/04/2013 09:02 AM, Borislav Petkov wrote: On Mon, Feb 04, 2013 at 08:50:33AM -0500, Rik van Riel wrote: We need to know whether we are actually running on top of a hypervisor, not whether we have the code compiled in to do so. Oh ok, I see. The thing is, if CONFIG_PARAVIRT_GUEST is disab

Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-02-04 Thread Borislav Petkov
On Mon, Feb 04, 2013 at 08:50:33AM -0500, Rik van Riel wrote: > We need to know whether we are actually running on top of a > hypervisor, not whether we have the code compiled in to do so. Oh ok, I see. The thing is, if CONFIG_PARAVIRT_GUEST is disabled, x86_hyper won't exist, see: http://marc.in

Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-02-04 Thread Rik van Riel
On 01/26/2013 07:47 AM, Borislav Petkov wrote: On Sat, Jan 26, 2013 at 01:00:43PM +0100, Ingo Molnar wrote: +void __init init_spinlock_delay(void) +{ + if (x86_hyper) + max_spinlock_delay = MAX_SPINLOCK_DELAY_GUEST; I realize that you took existing code and extended it, but

Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-01-28 Thread Stefano Stabellini
On Fri, 25 Jan 2013, Rik van Riel wrote: > Modern Intel and AMD CPUs will trap to the host when the guest > is spinning on a spinlock, allowing the host to schedule in > something else. > > This effectively means the host is taking care of spinlock > backoff for virtual machines. It also means tha

Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-01-26 Thread Borislav Petkov
On Sat, Jan 26, 2013 at 01:00:43PM +0100, Ingo Molnar wrote: > > +void __init init_spinlock_delay(void) > > +{ > > + if (x86_hyper) > > + max_spinlock_delay = MAX_SPINLOCK_DELAY_GUEST; > > I realize that you took existing code and extended it, but that > chunk of code looks pretty dis

Re: [PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-01-26 Thread Ingo Molnar
* Rik van Riel wrote: > static bool smp_no_nmi_ipi = false; > > /* > + * Modern Intel and AMD CPUs tell the hypervisor when a guest is > + * spinning excessively on a spinlock. The hypervisor will then > + * schedule something else, effectively taking care of the backoff > + * for us. Doing

[PATCH -v4 5/5] x86,smp: limit spinlock delay on virtual machines

2013-01-25 Thread Rik van Riel
Modern Intel and AMD CPUs will trap to the host when the guest is spinning on a spinlock, allowing the host to schedule in something else. This effectively means the host is taking care of spinlock backoff for virtual machines. It also means that doing the spinlock backoff in the guest anyway can