From nobody Mon Feb 9 17:27:06 2026 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1637857798245701.4799420055426; Thu, 25 Nov 2021 08:29:58 -0800 (PST) Received: from localhost ([::1]:44602 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1mqHdE-0001Bi-Tl for importer@patchew.org; Thu, 25 Nov 2021 11:29:56 -0500 Received: from eggs.gnu.org ([209.51.188.92]:58714) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1mqHbZ-0007g1-US for qemu-devel@nongnu.org; Thu, 25 Nov 2021 11:28:13 -0500 Received: from prt-mail.chinatelecom.cn ([42.123.76.219]:41521 helo=chinatelecom.cn) by eggs.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1mqHbV-0005y4-Ne for qemu-devel@nongnu.org; Thu, 25 Nov 2021 11:28:13 -0500 Received: from clientip-171.223.98.21 (unknown [172.18.0.188]) by chinatelecom.cn (HERMES) with SMTP id 7988A280097; Fri, 26 Nov 2021 00:28:01 +0800 (CST) Received: from ([172.18.0.188]) by app0023 with ESMTP id e304611bb4004ebfbed4155877a4ac60 for qemu-devel@nongnu.org; Fri, 26 Nov 2021 00:28:08 CST HMM_SOURCE_IP: 172.18.0.188:42636.659795330 HMM_ATTACHE_NUM: 0000 HMM_SOURCE_TYPE: SMTP X-189-SAVE-TO-SEND: +huangy81@chinatelecom.cn X-Transaction-ID: e304611bb4004ebfbed4155877a4ac60 X-Real-From: huangy81@chinatelecom.cn X-Receive-IP: 172.18.0.188 X-MEDUSA-Status: 0 From: huangy81@chinatelecom.cn To: qemu-devel Subject: [PATCH v6 2/3] cpu-throttle: implement vCPU throttle Date: Fri, 26 Nov 2021 00:27:51 +0800 Message-Id: <2d9259a1e7cf45352a264890bb9cc3154568ec65.1637857372.git.huangy81@chinatelecom.cn> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: References: In-Reply-To: References: MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=42.123.76.219; envelope-from=huangy81@chinatelecom.cn; helo=chinatelecom.cn X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Juan Quintela , Hyman , David Hildenbrand , Richard Henderson , Markus ArmBruster , Peter Xu , "Dr. David Alan Gilbert" , Paolo Bonzini , =?UTF-8?q?Philippe=20Mathieu-Daud=C3=A9?= Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZM-MESSAGEID: 1637857802204100002 From: Hyman Huang(=E9=BB=84=E5=8B=87) Impose dirty restraint on vCPU by kicking it and sleep as the auto-converge does during migration, but just kick the specified vCPU instead, not all vCPUs of vm. Start a thread to track the dirtylimit status and adjust the throttle pencentage dynamically depend on current and quota dirtyrate. Introduce the util function in the header for dirty page limit implementation. Signed-off-by: Hyman Huang(=E9=BB=84=E5=8B=87) --- include/sysemu/cpu-throttle.h | 23 +++ softmmu/cpu-throttle.c | 316 ++++++++++++++++++++++++++++++++++++++= ++++ softmmu/trace-events | 5 + 3 files changed, 344 insertions(+) diff --git a/include/sysemu/cpu-throttle.h b/include/sysemu/cpu-throttle.h index d65bdef..726c1ce 100644 --- a/include/sysemu/cpu-throttle.h +++ b/include/sysemu/cpu-throttle.h @@ -65,4 +65,27 @@ bool cpu_throttle_active(void); */ int cpu_throttle_get_percentage(void); =20 +/** + * dirtylimit_state_init: + * + * initialize golobal state for dirtylimit + */ +void dirtylimit_state_init(int max_cpus); + +/** + * dirtylimit_vcpu: + * + * impose dirtylimit on vcpu util reaching the quota dirtyrate + */ +void dirtylimit_vcpu(int cpu_index, + uint64_t quota); +/** + * dirtylimit_cancel_vcpu: + * + * cancel dirtylimit for the specified vcpu + * + * Returns: the number of running threads for dirtylimit + */ +int dirtylimit_cancel_vcpu(int cpu_index); + #endif /* SYSEMU_CPU_THROTTLE_H */ diff --git a/softmmu/cpu-throttle.c b/softmmu/cpu-throttle.c index 8c2144a..2d45388 100644 --- a/softmmu/cpu-throttle.c +++ b/softmmu/cpu-throttle.c @@ -29,6 +29,8 @@ #include "qemu/main-loop.h" #include "sysemu/cpus.h" #include "sysemu/cpu-throttle.h" +#include "sysemu/dirtylimit.h" +#include "trace.h" =20 /* vcpu throttling controls */ static QEMUTimer *throttle_timer; @@ -38,6 +40,320 @@ static unsigned int throttle_percentage; #define CPU_THROTTLE_PCT_MAX 99 #define CPU_THROTTLE_TIMESLICE_NS 10000000 =20 +#define DIRTYLIMIT_TOLERANCE_RANGE 15 /* 15MB/s */ + +#define DIRTYLIMIT_THROTTLE_HEAVY_WATERMARK 75 +#define DIRTYLIMIT_THROTTLE_SLIGHT_WATERMARK 90 + +#define DIRTYLIMIT_THROTTLE_HEAVY_STEP_SIZE 5 +#define DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE 2 + +typedef enum { + RESTRAIN_KEEP, + RESTRAIN_RATIO, + RESTRAIN_HEAVY, + RESTRAIN_SLIGHT, +} RestrainPolicy; + +typedef struct DirtyLimitState { + int cpu_index; + bool enabled; + uint64_t quota; /* quota dirtyrate MB/s */ + QemuThread thread; + char *name; /* thread name */ +} DirtyLimitState; + +struct { + DirtyLimitState *states; + int max_cpus; + unsigned long *bmap; /* running thread bitmap */ + unsigned long nr; +} *dirtylimit_state; + +static inline bool dirtylimit_enabled(int cpu_index) +{ + return qatomic_read(&dirtylimit_state->states[cpu_index].enabled); +} + +static inline void dirtylimit_set_quota(int cpu_index, uint64_t quota) +{ + qatomic_set(&dirtylimit_state->states[cpu_index].quota, quota); +} + +static inline uint64_t dirtylimit_quota(int cpu_index) +{ + return qatomic_read(&dirtylimit_state->states[cpu_index].quota); +} + +static int64_t dirtylimit_current(int cpu_index) +{ + return dirtylimit_calc_current(cpu_index); +} + +static void dirtylimit_vcpu_thread(CPUState *cpu, run_on_cpu_data data) +{ + double pct; + double throttle_ratio; + int64_t sleeptime_ns, endtime_ns; + int *percentage =3D (int *)data.host_ptr; + + pct =3D (double)(*percentage) / 100; + throttle_ratio =3D pct / (1 - pct); + /* Add 1ns to fix double's rounding error (like 0.9999999...) */ + sleeptime_ns =3D (int64_t)(throttle_ratio * CPU_THROTTLE_TIMESLICE_NS = + 1); + endtime_ns =3D qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + sleeptime_ns; + while (sleeptime_ns > 0 && !cpu->stop) { + if (sleeptime_ns > SCALE_MS) { + qemu_cond_timedwait_iothread(cpu->halt_cond, + sleeptime_ns / SCALE_MS); + } else { + qemu_mutex_unlock_iothread(); + g_usleep(sleeptime_ns / SCALE_US); + qemu_mutex_lock_iothread(); + } + sleeptime_ns =3D endtime_ns - qemu_clock_get_ns(QEMU_CLOCK_REALTIM= E); + } + qatomic_set(&cpu->throttle_thread_scheduled, 0); + + free(percentage); +} + +static void dirtylimit_check(int cpu_index, + int percentage) +{ + CPUState *cpu; + int64_t sleeptime_ns, starttime_ms, currenttime_ms; + int *pct_parameter; + double pct; + + pct =3D (double) percentage / 100; + + starttime_ms =3D qemu_clock_get_ms(QEMU_CLOCK_REALTIME); + + while (true) { + CPU_FOREACH(cpu) { + if ((cpu_index =3D=3D cpu->cpu_index) && + (!qatomic_xchg(&cpu->throttle_thread_scheduled, 1))) { + pct_parameter =3D malloc(sizeof(*pct_parameter)); + *pct_parameter =3D percentage; + async_run_on_cpu(cpu, dirtylimit_vcpu_thread, + RUN_ON_CPU_HOST_PTR(pct_parameter)); + break; + } + } + + sleeptime_ns =3D CPU_THROTTLE_TIMESLICE_NS / (1 - pct); + g_usleep(sleeptime_ns / SCALE_US); + + currenttime_ms =3D qemu_clock_get_ms(QEMU_CLOCK_REALTIME); + if (unlikely((currenttime_ms - starttime_ms) > + (DIRTYLIMIT_CALC_PERIOD_TIME_S * 1000))) { + break; + } + } +} + +static uint64_t dirtylimit_init_pct(uint64_t quota, + uint64_t current) +{ + uint64_t limit_pct =3D 0; + + if (quota >=3D current || (current =3D=3D 0) || + ((current - quota) <=3D DIRTYLIMIT_TOLERANCE_RANGE)) { + limit_pct =3D 0; + } else { + limit_pct =3D (current - quota) * 100 / current; + + limit_pct =3D MIN(limit_pct, + DIRTYLIMIT_THROTTLE_HEAVY_WATERMARK); + } + + return limit_pct; +} + +static RestrainPolicy dirtylimit_policy(unsigned int last_pct, + uint64_t quota, + uint64_t current) +{ + uint64_t max, min; + + max =3D MAX(quota, current); + min =3D MIN(quota, current); + if ((max - min) <=3D DIRTYLIMIT_TOLERANCE_RANGE) { + return RESTRAIN_KEEP; + } + if (last_pct < DIRTYLIMIT_THROTTLE_HEAVY_WATERMARK) { + /* last percentage locates in [0, 75)*/ + return RESTRAIN_RATIO; + } else if (last_pct < DIRTYLIMIT_THROTTLE_SLIGHT_WATERMARK) { + /* last percentage locates in [75, 90)*/ + return RESTRAIN_HEAVY; + } else { + /* last percentage locates in [90, 99]*/ + return RESTRAIN_SLIGHT; + } +} + +static uint64_t dirtylimit_pct(unsigned int last_pct, + uint64_t quota, + uint64_t current) +{ + uint64_t limit_pct =3D 0; + RestrainPolicy policy; + bool mitigate =3D (quota > current) ? true : false; + + if (mitigate && ((current =3D=3D 0) || + (last_pct <=3D DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE))) { + return 0; + } + + policy =3D dirtylimit_policy(last_pct, quota, current); + switch (policy) { + case RESTRAIN_SLIGHT: + /* [90, 99] */ + if (mitigate) { + limit_pct =3D + last_pct - DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE; + } else { + limit_pct =3D + last_pct + DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE; + + limit_pct =3D MIN(limit_pct, CPU_THROTTLE_PCT_MAX); + } + break; + case RESTRAIN_HEAVY: + /* [75, 90) */ + if (mitigate) { + limit_pct =3D + last_pct - DIRTYLIMIT_THROTTLE_HEAVY_STEP_SIZE; + } else { + limit_pct =3D + last_pct + DIRTYLIMIT_THROTTLE_HEAVY_STEP_SIZE; + + limit_pct =3D MIN(limit_pct, + DIRTYLIMIT_THROTTLE_SLIGHT_WATERMARK); + } + break; + case RESTRAIN_RATIO: + /* [0, 75) */ + if (mitigate) { + if (last_pct <=3D (((quota - current) * 100 / quota))) { + limit_pct =3D 0; + } else { + limit_pct =3D last_pct - + ((quota - current) * 100 / quota); + limit_pct =3D MAX(limit_pct, CPU_THROTTLE_PCT_MIN); + } + } else { + limit_pct =3D last_pct + + ((current - quota) * 100 / current); + + limit_pct =3D MIN(limit_pct, + DIRTYLIMIT_THROTTLE_HEAVY_WATERMARK); + } + break; + case RESTRAIN_KEEP: + default: + limit_pct =3D last_pct; + break; + } + + return limit_pct; +} + +static void *dirtylimit_thread(void *opaque) +{ + int cpu_index =3D *(int *)opaque; + uint64_t quota_dirtyrate, current_dirtyrate; + unsigned int last_pct =3D 0; + unsigned int pct =3D 0; + + rcu_register_thread(); + + quota_dirtyrate =3D dirtylimit_quota(cpu_index); + current_dirtyrate =3D dirtylimit_current(cpu_index); + + pct =3D dirtylimit_init_pct(quota_dirtyrate, current_dirtyrate); + + do { + trace_dirtylimit_impose(cpu_index, + quota_dirtyrate, current_dirtyrate, pct); + + last_pct =3D pct; + if (pct =3D=3D 0) { + sleep(DIRTYLIMIT_CALC_PERIOD_TIME_S); + } else { + dirtylimit_check(cpu_index, pct); + } + + quota_dirtyrate =3D dirtylimit_quota(cpu_index); + current_dirtyrate =3D dirtylimit_current(cpu_index); + + pct =3D dirtylimit_pct(last_pct, quota_dirtyrate, current_dirtyrat= e); + } while (dirtylimit_enabled(cpu_index)); + + rcu_unregister_thread(); + + return NULL; +} + +int dirtylimit_cancel_vcpu(int cpu_index) +{ + int i; + int nr_threads =3D 0; + + qatomic_set(&dirtylimit_state->states[cpu_index].enabled, 0); + bitmap_test_and_clear_atomic(dirtylimit_state->bmap, cpu_index, 1); + + for (i =3D 0; i < dirtylimit_state->nr; i++) { + unsigned long temp =3D dirtylimit_state->bmap[i]; + nr_threads +=3D ctpopl(temp); + } + + return nr_threads; +} + +void dirtylimit_vcpu(int cpu_index, + uint64_t quota) +{ + trace_dirtylimit_vcpu(cpu_index, quota); + + dirtylimit_set_quota(cpu_index, quota); + + if (unlikely(!dirtylimit_enabled(cpu_index))) { + qatomic_set(&dirtylimit_state->states[cpu_index].enabled, 1); + dirtylimit_state->states[cpu_index].name =3D + g_strdup_printf("dirtylimit-%d", cpu_index); + qemu_thread_create(&dirtylimit_state->states[cpu_index].thread, + dirtylimit_state->states[cpu_index].name, + dirtylimit_thread, + (void *)&dirtylimit_state->states[cpu_index].cpu_index, + QEMU_THREAD_DETACHED); + bitmap_set_atomic(dirtylimit_state->bmap, cpu_index, 1); + } +} + +void dirtylimit_state_init(int max_cpus) +{ + int i; + + dirtylimit_state =3D g_malloc0(sizeof(*dirtylimit_state)); + + dirtylimit_state->states =3D + g_malloc0(sizeof(DirtyLimitState) * max_cpus); + + for (i =3D 0; i < max_cpus; i++) { + dirtylimit_state->states[i].cpu_index =3D i; + } + + dirtylimit_state->max_cpus =3D max_cpus; + dirtylimit_state->bmap =3D bitmap_new(max_cpus); + bitmap_clear(dirtylimit_state->bmap, 0, max_cpus); + dirtylimit_state->nr =3D BITS_TO_LONGS(max_cpus); + + trace_dirtylimit_state_init(max_cpus); +} + static void cpu_throttle_thread(CPUState *cpu, run_on_cpu_data opaque) { double pct; diff --git a/softmmu/trace-events b/softmmu/trace-events index 9c88887..a7c9c04 100644 --- a/softmmu/trace-events +++ b/softmmu/trace-events @@ -31,3 +31,8 @@ runstate_set(int current_state, const char *current_state= _str, int new_state, co system_wakeup_request(int reason) "reason=3D%d" qemu_system_shutdown_request(int reason) "reason=3D%d" qemu_system_powerdown_request(void) "" + +#cpu-throttle.c +dirtylimit_state_init(int max_cpus) "dirtylimit state init: max cpus %d" +dirtylimit_impose(int cpu_index, uint64_t quota, uint64_t current, int pct= ) "CPU[%d] impose dirtylimit: quota %" PRIu64 ", current %" PRIu64 ", perce= ntage %d" +dirtylimit_vcpu(int cpu_index, uint64_t quota) "CPU[%d] set quota dirtylim= it %"PRIu64 --=20 1.8.3.1