[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH] remove blocked time accounting from xen "clockchip"
On Fri, Jan 20, 2012 at 09:57:04AM +0000, Jan Beulich wrote: > >>> On 19.01.12 at 20:42, Konrad Rzeszutek Wilk <konrad@xxxxxxxxxx> wrote: > > I finally got some time to look at them and I think they are these ones: > > > > git log --oneline > > e03b644fe68b1c6401465b02724d261538dba10f..3c404b578fab699c4708279938078d9404b > > 255a4 > > 3c404b5 KVM guest: Add a pv_ops stub for steal time > > c9aaa89 KVM: Steal time implementation > > 9ddabbe KVM: KVM Steal time guest/host interface > > 4b6b35f KVM: Add constant to represent KVM MSRs enabled bit in guest/host > > interface > > > > What is interesting is that they end up inserting a bunch of: > > > > > > + if (steal_account_process_tick()) > > + return; > > + > > > > in irqtime_account_process_tick and in account_process_tick. > > And this (particularly the "return" part of it) is what I have a hard > time to understand: How can it be correct to not do any of the > other accounting? After all, the function calls only > account_steal_time(), but its certainly going to be common that > part of the time was stolen, and part was spent executing. > > Further, it's being called only from the process tick accounting > functions, but clearly part of idle or interrupt time can also be > stolen. I took a stab at trying to implement this using this API and basing it on top Laszlo patch.. But I've doubts about it too and I haven't run any benchmarks past booting a guest. commit b1acd2adad821fd87da6941c38f0dbaddd37dc6b Author: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx> Date: Mon Jan 23 14:21:20 2012 -0500 xen/time: Use the pv_ops steal_time. In the past we were using do_stolen_account() to update the stolen time and this was done when by calling account_steal_ticks(). The do_stolen_account() was called from xen_timer_interrupt(). The xen_timer_interrupt() would be called periodically (or frequently) depending on what the timer decided. We would piggy-back on the timer IRQ to update the steal time and idle time (fixed by 'remove blocked time accounting from xen "clockchip"' patch). Instead of piggy-backing on the timer interrupt lets do it via the provided API pv-ops time calls - the steal_clock. Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx> diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c index 5c4a7f8..59cdd1b 100644 --- a/arch/x86/xen/time.c +++ b/arch/x86/xen/time.c @@ -71,14 +71,14 @@ static u64 get64(const u64 *p) /* * Runstate accounting */ -static void get_runstate_snapshot(struct vcpu_runstate_info *res) +static void get_runstate_snapshot(struct vcpu_runstate_info *res, int cpu) { u64 state_time; struct vcpu_runstate_info *state; BUG_ON(preemptible()); - state = &__get_cpu_var(xen_runstate); + state = &per_cpu(xen_runstate, cpu); /* * The runstate info is always updated by the hypervisor on @@ -110,18 +110,18 @@ void xen_setup_runstate_info(int cpu) BUG(); } -static void do_stolen_accounting(void) +static u64 xen_steal_clock(int cpu) { struct vcpu_runstate_info state; struct vcpu_runstate_info *snap; s64 runnable, offline, stolen; - cputime_t ticks; + u64 ticks; - get_runstate_snapshot(&state); + get_runstate_snapshot(&state, cpu); WARN_ON(state.state != RUNSTATE_running); - snap = &__get_cpu_var(xen_runstate_snapshot); + snap = &per_cpu(xen_runstate_snapshot, cpu); /* work out how much time the VCPU has not been runn*ing* */ runnable = state.time[RUNSTATE_runnable] - snap->time[RUNSTATE_runnable]; @@ -138,7 +138,8 @@ static void do_stolen_accounting(void) ticks = iter_div_u64_rem(stolen, NS_PER_TICK, &stolen); __this_cpu_write(xen_residual_stolen, stolen); - account_steal_ticks(ticks); + + return ticks; } /* Get the TSC speed from Xen */ @@ -377,8 +378,6 @@ static irqreturn_t xen_timer_interrupt(int irq, void *dev_id) ret = IRQ_HANDLED; } - do_stolen_accounting(); - return ret; } @@ -439,6 +438,7 @@ void xen_timer_resume(void) static const struct pv_time_ops xen_time_ops __initconst = { .sched_clock = xen_clocksource_read, + .steal_clock = xen_steal_clock, }; static void __init xen_time_init(void) @@ -464,6 +464,9 @@ static void __init xen_time_init(void) xen_setup_runstate_info(cpu); xen_setup_timer(cpu); xen_setup_cpu_clockevents(); + + jump_label_inc(¶virt_steal_enabled); + jump_label_inc(¶virt_steal_rq_enabled); } void __init xen_init_time_ops(void) > > Jan _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |