From patchwork Fri Sep 27 07:00:38 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: =?utf-8?b?SsO8cmdlbiBHcm/Dnw==?= X-Patchwork-Id: 11163953 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 0BC1A112B for ; Fri, 27 Sep 2019 07:03:45 +0000 (UTC) Received: from lists.xenproject.org (lists.xenproject.org [192.237.175.120]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id D9C6C207FF for ; Fri, 27 Sep 2019 07:03:44 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org D9C6C207FF Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1iDkGc-0004XY-5S; Fri, 27 Sep 2019 07:02:14 +0000 Received: from us1-rack-iad1.inumbo.com ([172.99.69.81]) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1iDkGa-0004V7-RV for xen-devel@lists.xenproject.org; Fri, 27 Sep 2019 07:02:12 +0000 X-Inumbo-ID: 93f9d360-e0f4-11e9-bf31-bc764e2007e4 Received: from mx1.suse.de (unknown [195.135.220.15]) by localhost (Halon) with ESMTPS id 93f9d360-e0f4-11e9-bf31-bc764e2007e4; Fri, 27 Sep 2019 07:01:07 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx1.suse.de (Postfix) with ESMTP id C52D8B181; Fri, 27 Sep 2019 07:01:04 +0000 (UTC) From: Juergen Gross To: xen-devel@lists.xenproject.org Date: Fri, 27 Sep 2019 09:00:38 +0200 Message-Id: <20190927070050.12405-35-jgross@suse.com> X-Mailer: git-send-email 2.16.4 In-Reply-To: <20190927070050.12405-1-jgross@suse.com> References: <20190927070050.12405-1-jgross@suse.com> Subject: [Xen-devel] [PATCH v4 34/46] xen/sched: add fall back to idle vcpu when scheduling unit X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Juergen Gross , Stefano Stabellini , Wei Liu , Konrad Rzeszutek Wilk , George Dunlap , Andrew Cooper , Ian Jackson , Tim Deegan , Julien Grall , Jan Beulich , Dario Faggioli , Volodymyr Babchuk , =?utf-8?q?Roger_Pau_Monn?= =?utf-8?q?=C3=A9?= MIME-Version: 1.0 Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" When scheduling an unit with multiple vcpus there is no guarantee all vcpus are available (e.g. above maxvcpus or vcpu offline). Fall back to idle vcpu of the current cpu in that case. This requires to store the correct schedule_unit pointer in the idle vcpu as long as it used as fallback vcpu. In order to modify the runstates of the correct vcpus when switching schedule units merge sched_unit_runstate_change() into sched_switch_units() and loop over the affected physical cpus instead of the unit's vcpus. This in turn requires an access function to the current variable of other cpus. Today context_saved() is called in case previous and next vcpus differ when doing a context switch. With an idle vcpu being capable to be a substitute for an offline vcpu this is problematic when switching to an idle scheduling unit. An idle previous vcpu leaves us in doubt which schedule unit was active previously, so save the previous unit pointer in the per-schedule resource area. If it is NULL the unit has not changed and we don't have to set the previous unit to be not running. When running an idle vcpu in a non-idle scheduling unit use a specific guest idle loop not performing any non-softirq tasklets and livepatching in order to avoid populating the cpu caches with memory used by other domains (as far as possible). Softirqs are considered to be save. In order to avoid livepatching when going to guest idle another variant of reset_stack_and_jump() not calling check_for_livepatch_work is needed. Signed-off-by: Juergen Gross Acked-by: Julien Grall Reviewed-by: Dario Faggioli --- RFC V2: - new patch (Andrew Cooper) V1: - use urgent_count to select correct idle routine (Jan Beulich) V2: - set vcpu->is_running in context_saved() - introduce reset_stack_and_jump_nolp() (Jan Beulich) - readd scrubbing (Jan Beulich, Andrew Cooper) - get_cpu_current() _NOT_ moved to include/asm-x86/current.h as the needed reference of stack_base[] results in a #include hell V3: - split context_saved() into unit_context_saved() and vcpu_context_saved() V4: - rename sd -> sr (Jan Beulich) - use unsigned int for cpu (Jan Beulich) - add comment in sched_context_switch() (Jan Beulich) - add comment before definition of get_cpu_current() (Jan Beulich) --- xen/arch/x86/domain.c | 23 ++++++ xen/common/schedule.c | 188 +++++++++++++++++++++++++++++------------- xen/include/asm-arm/current.h | 1 + xen/include/asm-x86/current.h | 19 ++++- xen/include/asm-x86/smp.h | 7 ++ xen/include/xen/sched-if.h | 4 +- xen/include/xen/sched.h | 1 + 7 files changed, 180 insertions(+), 63 deletions(-) diff --git a/xen/arch/x86/domain.c b/xen/arch/x86/domain.c index 27f99d3bcc..c8d7f491ea 100644 --- a/xen/arch/x86/domain.c +++ b/xen/arch/x86/domain.c @@ -159,6 +159,25 @@ static void idle_loop(void) } } +/* + * Idle loop for siblings in active schedule units. + * We don't do any standard idle work like tasklets or livepatching. + */ +static void guest_idle_loop(void) +{ + unsigned int cpu = smp_processor_id(); + + for ( ; ; ) + { + ASSERT(!cpu_is_offline(cpu)); + + if ( !softirq_pending(cpu) && !scrub_free_pages() && + !softirq_pending(cpu)) + sched_guest_idle(pm_idle, cpu); + do_softirq(); + } +} + void startup_cpu_idle_loop(void) { struct vcpu *v = current; @@ -172,6 +191,10 @@ void startup_cpu_idle_loop(void) static void noreturn continue_idle_domain(struct vcpu *v) { + /* Idle vcpus might be attached to non-idle units! */ + if ( !is_idle_domain(v->sched_unit->domain) ) + reset_stack_and_jump_nolp(guest_idle_loop); + reset_stack_and_jump(idle_loop); } diff --git a/xen/common/schedule.c b/xen/common/schedule.c index f79cd2a5a6..854ef39ee2 100644 --- a/xen/common/schedule.c +++ b/xen/common/schedule.c @@ -143,10 +143,21 @@ static struct scheduler sched_idle_ops = { .switch_sched = sched_idle_switch_sched, }; +static inline struct vcpu *unit2vcpu_cpu(const struct sched_unit *unit, + unsigned int cpu) +{ + unsigned int idx = unit->unit_id + per_cpu(sched_res_idx, cpu); + const struct domain *d = unit->domain; + + return (idx < d->max_vcpus) ? d->vcpu[idx] : NULL; +} + static inline struct vcpu *sched_unit2vcpu_cpu(const struct sched_unit *unit, unsigned int cpu) { - return unit->domain->vcpu[unit->unit_id + per_cpu(sched_res_idx, cpu)]; + struct vcpu *v = unit2vcpu_cpu(unit, cpu); + + return (v && v->new_state == RUNSTATE_running) ? v : idle_vcpu[cpu]; } static inline struct scheduler *dom_scheduler(const struct domain *d) @@ -266,8 +277,11 @@ static inline void vcpu_runstate_change( trace_runstate_change(v, new_state); - unit->runstate_cnt[v->runstate.state]--; - unit->runstate_cnt[new_state]++; + if ( !is_idle_vcpu(v) ) + { + unit->runstate_cnt[v->runstate.state]--; + unit->runstate_cnt[new_state]++; + } delta = new_entry_time - v->runstate.state_entry_time; if ( delta > 0 ) @@ -279,21 +293,11 @@ static inline void vcpu_runstate_change( v->runstate.state = new_state; } -static inline void sched_unit_runstate_change(struct sched_unit *unit, - bool running, s_time_t new_entry_time) +void sched_guest_idle(void (*idle) (void), unsigned int cpu) { - struct vcpu *v; - - for_each_sched_unit_vcpu ( unit, v ) - { - if ( running ) - vcpu_runstate_change(v, v->new_state, new_entry_time); - else - vcpu_runstate_change(v, - ((v->pause_flags & VPF_blocked) ? RUNSTATE_blocked : - (vcpu_runnable(v) ? RUNSTATE_runnable : RUNSTATE_offline)), - new_entry_time); - } + atomic_inc(&per_cpu(sched_urgent_count, cpu)); + idle(); + atomic_dec(&per_cpu(sched_urgent_count, cpu)); } void vcpu_runstate_get(struct vcpu *v, struct vcpu_runstate_info *runstate) @@ -543,6 +547,7 @@ int sched_init_vcpu(struct vcpu *v) if ( is_idle_domain(d) ) { get_sched_res(v->processor)->curr = unit; + get_sched_res(v->processor)->sched_unit_idle = unit; v->is_running = 1; unit->is_running = true; unit->state_entry_time = NOW(); @@ -875,7 +880,7 @@ static void sched_unit_move_locked(struct sched_unit *unit, * * sched_unit_migrate_finish() will do the work now if it can, or simply * return if it can't (because unit is still running); in that case - * sched_unit_migrate_finish() will be called by context_saved(). + * sched_unit_migrate_finish() will be called by unit_context_saved(). */ static void sched_unit_migrate_start(struct sched_unit *unit) { @@ -898,7 +903,7 @@ static void sched_unit_migrate_finish(struct sched_unit *unit) /* * If the unit is currently running, this will be handled by - * context_saved(); and in any case, if the bit is cleared, then + * unit_context_saved(); and in any case, if the bit is cleared, then * someone else has already done the work so we don't need to. */ if ( unit->is_running ) @@ -1783,33 +1788,66 @@ static void sched_switch_units(struct sched_resource *sr, struct sched_unit *next, struct sched_unit *prev, s_time_t now) { - sr->curr = next; - - TRACE_3D(TRC_SCHED_SWITCH_INFPREV, prev->domain->domain_id, prev->unit_id, - now - prev->state_entry_time); - TRACE_4D(TRC_SCHED_SWITCH_INFNEXT, next->domain->domain_id, next->unit_id, - (next->vcpu_list->runstate.state == RUNSTATE_runnable) ? - (now - next->state_entry_time) : 0, prev->next_time); + unsigned int cpu; ASSERT(unit_running(prev)); - TRACE_4D(TRC_SCHED_SWITCH, prev->domain->domain_id, prev->unit_id, - next->domain->domain_id, next->unit_id); + if ( prev != next ) + { + sr->curr = next; + sr->prev = prev; - sched_unit_runstate_change(prev, false, now); + TRACE_3D(TRC_SCHED_SWITCH_INFPREV, prev->domain->domain_id, + prev->unit_id, now - prev->state_entry_time); + TRACE_4D(TRC_SCHED_SWITCH_INFNEXT, next->domain->domain_id, + next->unit_id, + (next->vcpu_list->runstate.state == RUNSTATE_runnable) ? + (now - next->state_entry_time) : 0, prev->next_time); + TRACE_4D(TRC_SCHED_SWITCH, prev->domain->domain_id, prev->unit_id, + next->domain->domain_id, next->unit_id); - ASSERT(!unit_running(next)); - sched_unit_runstate_change(next, true, now); + ASSERT(!unit_running(next)); - /* - * NB. Don't add any trace records from here until the actual context - * switch, else lost_records resume will not work properly. - */ + /* + * NB. Don't add any trace records from here until the actual context + * switch, else lost_records resume will not work properly. + */ + + ASSERT(!next->is_running); + next->is_running = true; + next->state_entry_time = now; + + if ( is_idle_unit(prev) ) + { + prev->runstate_cnt[RUNSTATE_running] = 0; + prev->runstate_cnt[RUNSTATE_runnable] = sched_granularity; + } + if ( is_idle_unit(next) ) + { + next->runstate_cnt[RUNSTATE_running] = sched_granularity; + next->runstate_cnt[RUNSTATE_runnable] = 0; + } + } + + for_each_cpu ( cpu, sr->cpus ) + { + struct vcpu *vprev = get_cpu_current(cpu); + struct vcpu *vnext = sched_unit2vcpu_cpu(next, cpu); + + if ( vprev != vnext || vprev->runstate.state != vnext->new_state ) + { + vcpu_runstate_change(vprev, + ((vprev->pause_flags & VPF_blocked) ? RUNSTATE_blocked : + (vcpu_runnable(vprev) ? RUNSTATE_runnable : RUNSTATE_offline)), + now); + vcpu_runstate_change(vnext, vnext->new_state, now); + } + + vnext->is_running = 1; - ASSERT(!next->is_running); - next->vcpu_list->is_running = 1; - next->is_running = true; - next->state_entry_time = now; + if ( is_idle_vcpu(vnext) ) + vnext->sched_unit = next; + } } static bool sched_tasklet_check_cpu(unsigned int cpu) @@ -1865,29 +1903,39 @@ static struct sched_unit *do_schedule(struct sched_unit *prev, s_time_t now, if ( prev->next_time >= 0 ) /* -ve means no limit */ set_timer(&sr->s_timer, now + prev->next_time); - if ( likely(prev != next) ) - sched_switch_units(sr, next, prev, now); + sched_switch_units(sr, next, prev, now); return next; } -static void context_saved(struct vcpu *prev) +static void vcpu_context_saved(struct vcpu *vprev, struct vcpu *vnext) { - struct sched_unit *unit = prev->sched_unit; - /* Clear running flag /after/ writing context to memory. */ smp_wmb(); - prev->is_running = 0; + if ( vprev != vnext ) + vprev->is_running = 0; +} + +static void unit_context_saved(struct sched_resource *sr) +{ + struct sched_unit *unit = sr->prev; + + if ( !unit ) + return; + unit->is_running = false; unit->state_entry_time = NOW(); + sr->prev = NULL; /* Check for migration request /after/ clearing running flag. */ smp_mb(); - sched_context_saved(vcpu_scheduler(prev), unit); + sched_context_saved(unit_scheduler(unit), unit); - sched_unit_migrate_finish(unit); + /* Idle never migrates and idle vcpus might belong to other units. */ + if ( !is_idle_unit(unit) ) + sched_unit_migrate_finish(unit); } /* @@ -1897,35 +1945,44 @@ static void context_saved(struct vcpu *prev) * The counter will be 0 in case no rendezvous is needed. For the rendezvous * case it is initialised to the number of cpus to rendezvous plus 1. Each * member entering decrements the counter. The last one will decrement it to - * 1 and perform the final needed action in that case (call of context_saved() - * if vcpu was switched), and then set the counter to zero. The other members + * 1 and perform the final needed action in that case (call of + * unit_context_saved()), and then set the counter to zero. The other members * will wait until the counter becomes zero until they proceed. */ void sched_context_switched(struct vcpu *vprev, struct vcpu *vnext) { struct sched_unit *next = vnext->sched_unit; + struct sched_resource *sr = get_sched_res(smp_processor_id()); if ( atomic_read(&next->rendezvous_out_cnt) ) { int cnt = atomic_dec_return(&next->rendezvous_out_cnt); - /* Call context_saved() before releasing other waiters. */ + vcpu_context_saved(vprev, vnext); + + /* Call unit_context_saved() before releasing other waiters. */ if ( cnt == 1 ) { - if ( vprev != vnext ) - context_saved(vprev); + unit_context_saved(sr); atomic_set(&next->rendezvous_out_cnt, 0); } else while ( atomic_read(&next->rendezvous_out_cnt) ) cpu_relax(); } - else if ( vprev != vnext && sched_granularity == 1 ) - context_saved(vprev); + else + { + vcpu_context_saved(vprev, vnext); + if ( sched_granularity == 1 ) + unit_context_saved(sr); + } + + if ( is_idle_vcpu(vprev) && vprev != vnext ) + vprev->sched_unit = sr->sched_unit_idle; } static void sched_context_switch(struct vcpu *vprev, struct vcpu *vnext, - s_time_t now) + bool reset_idle_unit, s_time_t now) { if ( unlikely(vprev == vnext) ) { @@ -1934,6 +1991,17 @@ static void sched_context_switch(struct vcpu *vprev, struct vcpu *vnext, now - vprev->runstate.state_entry_time, vprev->sched_unit->next_time); sched_context_switched(vprev, vnext); + + /* + * We are switching from a non-idle to an idle unit. + * A vcpu of the idle unit might have been running before due to + * the guest vcpu being blocked. We must adjust the unit of the idle + * vcpu which might have been set to the guest's one. + */ + if ( reset_idle_unit ) + vnext->sched_unit = + get_sched_res(smp_processor_id())->sched_unit_idle; + trace_continue_running(vnext); return continue_running(vprev); } @@ -1992,7 +2060,7 @@ static struct sched_unit *sched_wait_rendezvous_in(struct sched_unit *prev, pcpu_schedule_unlock_irq(*lock, cpu); raise_softirq(SCHED_SLAVE_SOFTIRQ); - sched_context_switch(vprev, vprev, now); + sched_context_switch(vprev, vprev, false, now); return NULL; /* ARM only. */ } @@ -2035,7 +2103,8 @@ static void sched_slave(void) pcpu_schedule_unlock_irq(lock, cpu); - sched_context_switch(vprev, sched_unit2vcpu_cpu(next, cpu), now); + sched_context_switch(vprev, sched_unit2vcpu_cpu(next, cpu), + is_idle_unit(next) && !is_idle_unit(prev), now); } /* @@ -2097,7 +2166,8 @@ static void schedule(void) pcpu_schedule_unlock_irq(lock, cpu); vnext = sched_unit2vcpu_cpu(next, cpu); - sched_context_switch(vprev, vnext, now); + sched_context_switch(vprev, vnext, + !is_idle_unit(prev) && is_idle_unit(next), now); } /* The scheduler timer: force a run through the scheduler */ @@ -2168,6 +2238,7 @@ static int cpu_schedule_up(unsigned int cpu) */ sr->curr = idle_vcpu[cpu]->sched_unit; + sr->sched_unit_idle = idle_vcpu[cpu]->sched_unit; sr->sched_priv = NULL; @@ -2337,6 +2408,7 @@ void __init scheduler_init(void) if ( vcpu_create(idle_domain, 0) == NULL ) BUG(); get_sched_res(0)->curr = idle_vcpu[0]->sched_unit; + get_sched_res(0)->sched_unit_idle = idle_vcpu[0]->sched_unit; } /* diff --git a/xen/include/asm-arm/current.h b/xen/include/asm-arm/current.h index 1653e89d30..88beb4645a 100644 --- a/xen/include/asm-arm/current.h +++ b/xen/include/asm-arm/current.h @@ -18,6 +18,7 @@ DECLARE_PER_CPU(struct vcpu *, curr_vcpu); #define current (this_cpu(curr_vcpu)) #define set_current(vcpu) do { current = (vcpu); } while (0) +#define get_cpu_current(cpu) (per_cpu(curr_vcpu, cpu)) /* Per-VCPU state that lives at the top of the stack */ struct cpu_info { diff --git a/xen/include/asm-x86/current.h b/xen/include/asm-x86/current.h index f3508c3c08..0b47485337 100644 --- a/xen/include/asm-x86/current.h +++ b/xen/include/asm-x86/current.h @@ -77,6 +77,11 @@ struct cpu_info { /* get_stack_bottom() must be 16-byte aligned */ }; +static inline struct cpu_info *get_cpu_info_from_stack(unsigned long sp) +{ + return (struct cpu_info *)((sp | (STACK_SIZE - 1)) + 1) - 1; +} + static inline struct cpu_info *get_cpu_info(void) { #ifdef __clang__ @@ -87,7 +92,7 @@ static inline struct cpu_info *get_cpu_info(void) register unsigned long sp asm("rsp"); #endif - return (struct cpu_info *)((sp | (STACK_SIZE - 1)) + 1) - 1; + return get_cpu_info_from_stack(sp); } #define get_current() (get_cpu_info()->current_vcpu) @@ -124,16 +129,22 @@ unsigned long get_stack_dump_bottom (unsigned long sp); # define CHECK_FOR_LIVEPATCH_WORK "" #endif -#define reset_stack_and_jump(__fn) \ +#define switch_stack_and_jump(fn, instr) \ ({ \ __asm__ __volatile__ ( \ "mov %0,%%"__OP"sp;" \ - CHECK_FOR_LIVEPATCH_WORK \ + instr \ "jmp %c1" \ - : : "r" (guest_cpu_user_regs()), "i" (__fn) : "memory" ); \ + : : "r" (guest_cpu_user_regs()), "i" (fn) : "memory" ); \ unreachable(); \ }) +#define reset_stack_and_jump(fn) \ + switch_stack_and_jump(fn, CHECK_FOR_LIVEPATCH_WORK) + +#define reset_stack_and_jump_nolp(fn) \ + switch_stack_and_jump(fn, "") + /* * Which VCPU's state is currently running on each CPU? * This is not necesasrily the same as 'current' as a CPU may be diff --git a/xen/include/asm-x86/smp.h b/xen/include/asm-x86/smp.h index 61446d0efd..dbeed2fd41 100644 --- a/xen/include/asm-x86/smp.h +++ b/xen/include/asm-x86/smp.h @@ -77,6 +77,13 @@ void set_nr_sockets(void); /* Representing HT and core siblings in each socket. */ extern cpumask_t **socket_cpumask; +/* + * To be used only while no context switch can occur on the cpu, i.e. + * by certain scheduling code only. + */ +#define get_cpu_current(cpu) \ + (get_cpu_info_from_stack((unsigned long)stack_base[cpu])->current_vcpu) + #endif /* !__ASSEMBLY__ */ #endif diff --git a/xen/include/xen/sched-if.h b/xen/include/xen/sched-if.h index 1b296b150f..41a1083a08 100644 --- a/xen/include/xen/sched-if.h +++ b/xen/include/xen/sched-if.h @@ -39,6 +39,8 @@ struct sched_resource { spinlock_t *schedule_lock, _lock; struct sched_unit *curr; + struct sched_unit *sched_unit_idle; + struct sched_unit *prev; void *sched_priv; struct timer s_timer; /* scheduling timer */ @@ -194,7 +196,7 @@ static inline void sched_clear_pause_flags_atomic(struct sched_unit *unit, static inline struct sched_unit *sched_idle_unit(unsigned int cpu) { - return idle_vcpu[cpu]->sched_unit; + return get_sched_res(cpu)->sched_unit_idle; } static inline unsigned int sched_get_resource_cpu(unsigned int cpu) diff --git a/xen/include/xen/sched.h b/xen/include/xen/sched.h index 12f00cd78d..ce4329db72 100644 --- a/xen/include/xen/sched.h +++ b/xen/include/xen/sched.h @@ -929,6 +929,7 @@ void restore_vcpu_affinity(struct domain *d); void vcpu_runstate_get(struct vcpu *v, struct vcpu_runstate_info *runstate); uint64_t get_cpu_idle_time(unsigned int cpu); +void sched_guest_idle(void (*idle) (void), unsigned int cpu); /* * Used by idle loop to decide whether there is work to do: