[PATCH v4 0/3] support dirty restraint on vCPU

[PATCH v4 0/3] support dirty restraint on vCPU

[huangy81]

From: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>

v4:
- rebase on master
- modify the following points according to the advice given by Markus
  1. move the defination into migration.json
  2. polish the comments of set-dirty-limit
  3. do the syntax check and change dirty rate to dirty page rate

Thanks for the carefule reviews made by Markus.

Please review, thanks!

v3:
- rebase on master
- modify the following points according to the advice given by Markus
  1. remove the DirtyRateQuotaVcpu and use its field as option directly
  2. add comments to show details of what dirtylimit setup do
  3. explain how to use dirtylimit in combination with existing qmp
     commands "calc-dirty-rate" and "query-dirty-rate" in documentation. 

Thanks for the carefule reviews made by Markus.

Please review, thanks!

Hyman

v2:
- rebase on master
- modify the following points according to the advices given by Juan
  1. rename dirtyrestraint to dirtylimit
  2. implement the full lifecyle function of dirtylimit_calc, include
     dirtylimit_calc and dirtylimit_calc_quit
  3. introduce 'quit' field in dirtylimit_calc_state to implement the
     dirtylimit_calc_quit
  4. remove the ready_cond and ready_mtx since it may not be suitable
  5. put the 'record_dirtypage' function code at the beggining of the 
     file 
  6. remove the unnecesary return;
- other modifications has been made after code review
  1. introduce 'bmap' and 'nr' field in dirtylimit_state to record the
     number of running thread forked by dirtylimit
  2. stop the dirtyrate calculation thread if all the dirtylimit thread
     are stopped
  3. do some renaming works
     dirtyrate calulation thread -> dirtylimit-calc 
     dirtylimit thread -> dirtylimit-{cpu_index}
     function name do_dirtyrestraint -> dirtylimit_check
     qmp command dirty-restraint -> set-drity-limit
     qmp command dirty-restraint-cancel -> cancel-dirty-limit
     header file dirtyrestraint.h -> dirtylimit.h

Please review, thanks !

thanks for the accurate and timely advices given by Juan. we really
appreciate it if corrections and suggetions about this patchset are
proposed.

Best Regards ! 

Hyman

v1:
this patchset introduce a mechanism to impose dirty restraint
on vCPU, aiming to keep the vCPU running in a certain dirtyrate
given by user. dirty restraint on vCPU maybe an alternative
method to implement convergence logic for live migration,
which could improve guest memory performance during migration
compared with traditional method in theory.

For the current live migration implementation, the convergence
logic throttles all vCPUs of the VM, which has some side effects. 
-'read processes' on vCPU will be unnecessarily penalized
- throttle increase percentage step by step, which seems
  struggling to find the optimal throttle percentage when
  dirtyrate is high. 
- hard to predict the remaining time of migration if the
  throttling percentage reachs 99%

to a certain extent, the dirty restraint machnism can fix these
effects by throttling at vCPU granularity during migration.

the implementation is rather straightforward, we calculate
vCPU dirtyrate via the Dirty Ring mechanism periodically
as the commit 0e21bf246 "implement dirty-ring dirtyrate calculation"
does, for vCPU that be specified to impose dirty restraint,
we throttle it periodically as the auto-converge does, once after
throttling, we compare the quota dirtyrate with current dirtyrate,
if current dirtyrate is not under the quota, increase the throttling
percentage until current dirtyrate is under the quota.

this patchset is the basis of implmenting a new auto-converge method
for live migration, we introduce two qmp commands for impose/cancel
the dirty restraint on specified vCPU, so it also can be an independent
api to supply the upper app such as libvirt, which can use it to
implement the convergence logic during live migration, supplemented
with the qmp 'calc-dirty-rate' command or whatever. 

we post this patchset for RFC and any corrections and suggetions about
the implementation, api, throttleing algorithm or whatever are very
appreciated!

Please review, thanks !

Best Regards ! 

Hyman Huang (3):
  migration/dirtyrate: implement vCPU dirtyrate calculation periodically
  cpu-throttle: implement vCPU throttle
  cpus-common: implement dirty limit on vCPU

 cpus-common.c                 |  41 ++++++
 include/exec/memory.h         |   5 +-
 include/hw/core/cpu.h         |   9 ++
 include/sysemu/cpu-throttle.h |  23 +++
 include/sysemu/dirtylimit.h   |  44 ++++++
 migration/dirtyrate.c         | 139 ++++++++++++++++--
 migration/dirtyrate.h         |   2 +
 qapi/migration.json           |  43 ++++++
 softmmu/cpu-throttle.c        | 319 ++++++++++++++++++++++++++++++++++++++++++
 softmmu/trace-events          |   5 +
 softmmu/vl.c                  |   1 +
 11 files changed, 620 insertions(+), 11 deletions(-)
 create mode 100644 include/sysemu/dirtylimit.h

-- 
1.8.3.1

[PATCH v4 1/3] migration/dirtyrate: implement vCPU dirtyrate calculation periodically

[huangy81]

From: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>

Introduce the third method GLOBAL_DIRTY_LIMIT of dirty
tracking for calculate dirtyrate periodly for dirty restraint.

Implement thread for calculate dirtyrate periodly, which will
be used for dirty restraint.

Add dirtylimit.h to introduce the util function for dirty
limit implementation.

Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
---
 include/exec/memory.h       |   5 +-
 include/sysemu/dirtylimit.h |  44 ++++++++++++++
 migration/dirtyrate.c       | 139 ++++++++++++++++++++++++++++++++++++++++----
 migration/dirtyrate.h       |   2 +
 4 files changed, 179 insertions(+), 11 deletions(-)
 create mode 100644 include/sysemu/dirtylimit.h

diff --git a/include/exec/memory.h b/include/exec/memory.h
index 20f1b27..606bec8 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -69,7 +69,10 @@ static inline void fuzz_dma_read_cb(size_t addr,
 /* Dirty tracking enabled because measuring dirty rate */
 #define GLOBAL_DIRTY_DIRTY_RATE (1U << 1)
 
-#define GLOBAL_DIRTY_MASK  (0x3)
+/* Dirty tracking enabled because dirty limit */
+#define GLOBAL_DIRTY_LIMIT      (1U << 2)
+
+#define GLOBAL_DIRTY_MASK  (0x7)
 
 extern unsigned int global_dirty_tracking;
 
diff --git a/include/sysemu/dirtylimit.h b/include/sysemu/dirtylimit.h
new file mode 100644
index 0000000..49298a2
--- /dev/null
+++ b/include/sysemu/dirtylimit.h
@@ -0,0 +1,44 @@
+/*
+ * dirty limit helper functions
+ *
+ * Copyright (c) 2021 CHINA TELECOM CO.,LTD.
+ *
+ * Authors:
+ *  Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or later.
+ * See the COPYING file in the top-level directory.
+ */
+#ifndef QEMU_DIRTYRLIMIT_H
+#define QEMU_DIRTYRLIMIT_H
+
+#define DIRTYLIMIT_CALC_PERIOD_TIME_S   15      /* 15s */
+
+/**
+ * dirtylimit_calc_current:
+ *
+ * get current dirty page rate for specified vCPU.
+ */
+int64_t dirtylimit_calc_current(int cpu_index);
+
+/**
+ * dirtylimit_calc:
+ *
+ * start dirty page rate calculation thread.
+ */
+void dirtylimit_calc(void);
+
+/**
+ * dirtylimit_calc_quit:
+ *
+ * quit dirty page rate calculation thread.
+ */
+void dirtylimit_calc_quit(void);
+
+/**
+ * dirtylimit_calc_state_init:
+ *
+ * initialize dirty page rate calculation state.
+ */
+void dirtylimit_calc_state_init(int max_cpus);
+#endif
diff --git a/migration/dirtyrate.c b/migration/dirtyrate.c
index d65e744..d370a21 100644
--- a/migration/dirtyrate.c
+++ b/migration/dirtyrate.c
@@ -27,6 +27,7 @@
 #include "qapi/qmp/qdict.h"
 #include "sysemu/kvm.h"
 #include "sysemu/runstate.h"
+#include "sysemu/dirtylimit.h"
 #include "exec/memory.h"
 
 /*
@@ -46,6 +47,134 @@ static struct DirtyRateStat DirtyStat;
 static DirtyRateMeasureMode dirtyrate_mode =
                 DIRTY_RATE_MEASURE_MODE_PAGE_SAMPLING;
 
+#define DIRTYLIMIT_CALC_TIME_MS         1000    /* 1000ms */
+
+struct {
+    DirtyRatesData data;
+    int64_t period;
+    bool quit;
+} *dirtylimit_calc_state;
+
+static void dirtylimit_global_dirty_log_start(void)
+{
+    qemu_mutex_lock_iothread();
+    memory_global_dirty_log_start(GLOBAL_DIRTY_LIMIT);
+    qemu_mutex_unlock_iothread();
+}
+
+static void dirtylimit_global_dirty_log_stop(void)
+{
+    qemu_mutex_lock_iothread();
+    memory_global_dirty_log_sync();
+    memory_global_dirty_log_stop(GLOBAL_DIRTY_LIMIT);
+    qemu_mutex_unlock_iothread();
+}
+
+static inline void record_dirtypages(DirtyPageRecord *dirty_pages,
+                                     CPUState *cpu, bool start)
+{
+    if (start) {
+        dirty_pages[cpu->cpu_index].start_pages = cpu->dirty_pages;
+    } else {
+        dirty_pages[cpu->cpu_index].end_pages = cpu->dirty_pages;
+    }
+}
+
+static void dirtylimit_calc_func(void)
+{
+    CPUState *cpu;
+    DirtyPageRecord *dirty_pages;
+    int64_t start_time, end_time, calc_time;
+    DirtyRateVcpu rate;
+    int i = 0;
+
+    dirty_pages = g_malloc0(sizeof(*dirty_pages) *
+        dirtylimit_calc_state->data.nvcpu);
+
+    dirtylimit_global_dirty_log_start();
+
+    CPU_FOREACH(cpu) {
+        record_dirtypages(dirty_pages, cpu, true);
+    }
+
+    start_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
+    g_usleep(DIRTYLIMIT_CALC_TIME_MS * 1000);
+    end_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
+    calc_time = end_time - start_time;
+
+    dirtylimit_global_dirty_log_stop();
+
+    CPU_FOREACH(cpu) {
+        record_dirtypages(dirty_pages, cpu, false);
+    }
+
+    for (i = 0; i < dirtylimit_calc_state->data.nvcpu; i++) {
+        uint64_t increased_dirty_pages =
+            dirty_pages[i].end_pages - dirty_pages[i].start_pages;
+        uint64_t memory_size_MB =
+            (increased_dirty_pages * TARGET_PAGE_SIZE) >> 20;
+        int64_t dirtyrate = (memory_size_MB * 1000) / calc_time;
+
+        rate.id = i;
+        rate.dirty_rate  = dirtyrate;
+        dirtylimit_calc_state->data.rates[i] = rate;
+
+        trace_dirtyrate_do_calculate_vcpu(i,
+            dirtylimit_calc_state->data.rates[i].dirty_rate);
+    }
+}
+
+static void *dirtylimit_calc_thread(void *opaque)
+{
+    rcu_register_thread();
+
+    while (!qatomic_read(&dirtylimit_calc_state->quit)) {
+        dirtylimit_calc_func();
+        sleep(dirtylimit_calc_state->period);
+    }
+
+    rcu_unregister_thread();
+    return NULL;
+}
+
+int64_t dirtylimit_calc_current(int cpu_index)
+{
+    DirtyRateVcpu *rates = dirtylimit_calc_state->data.rates;
+
+    return qatomic_read(&rates[cpu_index].dirty_rate);
+}
+
+void dirtylimit_calc(void)
+{
+    if (unlikely(qatomic_read(&dirtylimit_calc_state->quit))) {
+        qatomic_set(&dirtylimit_calc_state->quit, 0);
+        QemuThread thread;
+        qemu_thread_create(&thread, "dirtylimit-calc",
+            dirtylimit_calc_thread,
+            NULL, QEMU_THREAD_DETACHED);
+    }
+}
+
+void dirtylimit_calc_quit(void)
+{
+    qatomic_set(&dirtylimit_calc_state->quit, 1);
+}
+
+void dirtylimit_calc_state_init(int max_cpus)
+{
+    dirtylimit_calc_state =
+        g_malloc0(sizeof(*dirtylimit_calc_state));
+
+    dirtylimit_calc_state->data.nvcpu = max_cpus;
+    dirtylimit_calc_state->data.rates =
+        g_malloc0(sizeof(DirtyRateVcpu) * max_cpus);
+
+    dirtylimit_calc_state->period =
+        DIRTYLIMIT_CALC_PERIOD_TIME_S;
+
+    dirtylimit_calc_state->quit = true;
+}
+
 static int64_t set_sample_page_period(int64_t msec, int64_t initial_time)
 {
     int64_t current_time;
@@ -396,16 +525,6 @@ static bool compare_page_hash_info(struct RamblockDirtyInfo *info,
     return true;
 }
 
-static inline void record_dirtypages(DirtyPageRecord *dirty_pages,
-                                     CPUState *cpu, bool start)
-{
-    if (start) {
-        dirty_pages[cpu->cpu_index].start_pages = cpu->dirty_pages;
-    } else {
-        dirty_pages[cpu->cpu_index].end_pages = cpu->dirty_pages;
-    }
-}
-
 static void dirtyrate_global_dirty_log_start(void)
 {
     qemu_mutex_lock_iothread();
diff --git a/migration/dirtyrate.h b/migration/dirtyrate.h
index 69d4c5b..e96acdc 100644
--- a/migration/dirtyrate.h
+++ b/migration/dirtyrate.h
@@ -70,6 +70,8 @@ typedef struct VcpuStat {
     DirtyRateVcpu *rates; /* array of dirty rate for each vcpu */
 } VcpuStat;
 
+typedef struct VcpuStat DirtyRatesData;
+
 /*
  * Store calculation statistics for each measure.
  */
-- 
1.8.3.1

[PATCH v4 2/3] cpu-throttle: implement vCPU throttle

[huangy81]

From: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>

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 dirtylimit
implementation.

Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
---
 include/sysemu/cpu-throttle.h |  23 +++
 softmmu/cpu-throttle.c        | 319 ++++++++++++++++++++++++++++++++++++++++++
 softmmu/trace-events          |   5 +
 3 files changed, 347 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);
 
+/**
+ * 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..a5e67c9 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"
 
 /* vcpu throttling controls */
 static QEMUTimer *throttle_timer;
@@ -38,6 +40,323 @@ static unsigned int throttle_percentage;
 #define CPU_THROTTLE_PCT_MAX 99
 #define CPU_THROTTLE_TIMESLICE_NS 10000000
 
+#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 = (int *)data.host_ptr;
+
+    pct = (double)(*percentage) / 100;
+    throttle_ratio = pct / (1 - pct);
+    /* Add 1ns to fix double's rounding error (like 0.9999999...) */
+    sleeptime_ns = (int64_t)(throttle_ratio * CPU_THROTTLE_TIMESLICE_NS + 1);
+    endtime_ns = 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 = endtime_ns - qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
+    }
+    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 = (double) percentage / 100;
+
+    starttime_ms = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
+
+    while (true) {
+        CPU_FOREACH(cpu) {
+            if ((cpu_index == cpu->cpu_index) &&
+                (!qatomic_xchg(&cpu->throttle_thread_scheduled, 1))) {
+                pct_parameter = malloc(sizeof(*pct_parameter));
+                *pct_parameter = percentage;
+                async_run_on_cpu(cpu, dirtylimit_vcpu_thread,
+                                 RUN_ON_CPU_HOST_PTR(pct_parameter));
+                break;
+            }
+        }
+
+        sleeptime_ns = CPU_THROTTLE_TIMESLICE_NS / (1 - pct);
+        g_usleep(sleeptime_ns / SCALE_US);
+
+        currenttime_ms = 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 = 0;
+
+    if (quota >= current || (current == 0) ||
+        ((current - quota) <= DIRTYLIMIT_TOLERANCE_RANGE)) {
+        limit_pct = 0;
+    } else {
+        limit_pct = (current - quota) * 100 / current;
+
+        limit_pct = 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 = MAX(quota, current);
+    min = MIN(quota, current);
+    if ((max - min) <= 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 = 0;
+    RestrainPolicy policy;
+    bool mitigate = (quota > current) ? true : false;
+
+    if (mitigate && ((current == 0) ||
+        (last_pct <= DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE))) {
+        return 0;
+    }
+
+    policy = dirtylimit_policy(last_pct, quota, current);
+    switch (policy) {
+    case RESTRAIN_SLIGHT:
+        /* [90, 99] */
+        if (mitigate) {
+            limit_pct =
+                last_pct - DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE;
+        } else {
+            limit_pct =
+                last_pct + DIRTYLIMIT_THROTTLE_SLIGHT_STEP_SIZE;
+
+            limit_pct = MIN(limit_pct, CPU_THROTTLE_PCT_MAX);
+        }
+       break;
+    case RESTRAIN_HEAVY:
+        /* [75, 90) */
+        if (mitigate) {
+            limit_pct =
+                last_pct - DIRTYLIMIT_THROTTLE_HEAVY_STEP_SIZE;
+        } else {
+            limit_pct =
+                last_pct + DIRTYLIMIT_THROTTLE_HEAVY_STEP_SIZE;
+
+            limit_pct = MIN(limit_pct,
+                DIRTYLIMIT_THROTTLE_SLIGHT_WATERMARK);
+        }
+       break;
+    case RESTRAIN_RATIO:
+        /* [0, 75) */
+        if (mitigate) {
+            if (last_pct <= (((quota - current) * 100 / quota) / 2)) {
+                limit_pct = 0;
+            } else {
+                limit_pct = last_pct -
+                    ((quota - current) * 100 / quota) / 2;
+                limit_pct = MAX(limit_pct, CPU_THROTTLE_PCT_MIN);
+            }
+        } else {
+            /*
+             * increase linearly with dirtyrate
+             * but tune a little by divide it by 2
+             */
+            limit_pct = last_pct +
+                ((current - quota) * 100 / current) / 2;
+
+            limit_pct = MIN(limit_pct,
+                DIRTYLIMIT_THROTTLE_HEAVY_WATERMARK);
+        }
+       break;
+    case RESTRAIN_KEEP:
+    default:
+       limit_pct = last_pct;
+       break;
+    }
+
+    return limit_pct;
+}
+
+static void *dirtylimit_thread(void *opaque)
+{
+    int cpu_index = *(int *)opaque;
+    uint64_t quota_dirtyrate, current_dirtyrate;
+    unsigned int last_pct = 0;
+    unsigned int pct = 0;
+
+    rcu_register_thread();
+
+    quota_dirtyrate = dirtylimit_quota(cpu_index);
+    current_dirtyrate = dirtylimit_current(cpu_index);
+
+    pct = dirtylimit_init_pct(quota_dirtyrate, current_dirtyrate);
+
+    do {
+        trace_dirtylimit_impose(cpu_index,
+            quota_dirtyrate, current_dirtyrate, pct);
+        if (pct == 0) {
+            sleep(DIRTYLIMIT_CALC_PERIOD_TIME_S);
+        } else {
+            last_pct = pct;
+            dirtylimit_check(cpu_index, pct);
+        }
+
+        quota_dirtyrate = dirtylimit_quota(cpu_index);
+        current_dirtyrate = dirtylimit_current(cpu_index);
+
+        pct = dirtylimit_pct(last_pct, quota_dirtyrate, current_dirtyrate);
+    } while (dirtylimit_enabled(cpu_index));
+
+    rcu_unregister_thread();
+
+    return NULL;
+}
+
+int dirtylimit_cancel_vcpu(int cpu_index)
+{
+    int i;
+    int nr_threads = 0;
+
+    qatomic_set(&dirtylimit_state->states[cpu_index].enabled, 0);
+    bitmap_test_and_clear_atomic(dirtylimit_state->bmap, cpu_index, 1);
+
+    for (i = 0; i < dirtylimit_state->nr; i++) {
+        unsigned long temp = dirtylimit_state->bmap[i];
+        nr_threads += 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 =
+            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 = g_malloc0(sizeof(*dirtylimit_state));
+
+    dirtylimit_state->states =
+            g_malloc0(sizeof(DirtyLimitState) * max_cpus);
+
+    for (i = 0; i < max_cpus; i++) {
+        dirtylimit_state->states[i].cpu_index = i;
+    }
+
+    dirtylimit_state->max_cpus = max_cpus;
+    dirtylimit_state->bmap = bitmap_new(max_cpus);
+    bitmap_clear(dirtylimit_state->bmap, 0, max_cpus);
+    dirtylimit_state->nr = 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=%d"
 qemu_system_shutdown_request(int reason) "reason=%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 ", percentage %d"
+dirtylimit_vcpu(int cpu_index, uint64_t quota) "CPU[%d] set quota dirtylimit %"PRIu64
-- 
1.8.3.1

[PATCH v4 3/3] cpus-common: implement dirty limit on vCPU

[huangy81]

From: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>

Implement dirtyrate calculation periodically basing on
dirty-ring and throttle vCPU until it reachs the quota
dirtyrate given by user.

Introduce qmp commands set-dirty-limit/cancel-dirty-limit to
set/cancel dirty limit on vCPU.

Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
---
 cpus-common.c         | 41 +++++++++++++++++++++++++++++++++++++++++
 include/hw/core/cpu.h |  9 +++++++++
 qapi/migration.json   | 43 +++++++++++++++++++++++++++++++++++++++++++
 softmmu/vl.c          |  1 +
 4 files changed, 94 insertions(+)

diff --git a/cpus-common.c b/cpus-common.c
index 6e73d3e..43b0078 100644
--- a/cpus-common.c
+++ b/cpus-common.c
@@ -23,6 +23,11 @@
 #include "hw/core/cpu.h"
 #include "sysemu/cpus.h"
 #include "qemu/lockable.h"
+#include "sysemu/dirtylimit.h"
+#include "sysemu/cpu-throttle.h"
+#include "sysemu/kvm.h"
+#include "qapi/error.h"
+#include "qapi/qapi-commands-migration.h"
 
 static QemuMutex qemu_cpu_list_lock;
 static QemuCond exclusive_cond;
@@ -352,3 +357,39 @@ void process_queued_cpu_work(CPUState *cpu)
     qemu_mutex_unlock(&cpu->work_mutex);
     qemu_cond_broadcast(&qemu_work_cond);
 }
+
+void qmp_set_dirty_limit(int64_t idx,
+                         uint64_t dirtyrate,
+                         Error **errp)
+{
+    if (!kvm_dirty_ring_enabled()) {
+        error_setg(errp, "dirty ring not enable, needed by dirty restraint!");
+        return;
+    }
+
+    dirtylimit_calc();
+    dirtylimit_vcpu(idx, dirtyrate);
+}
+
+void qmp_cancel_dirty_limit(int64_t idx,
+                            Error **errp)
+{
+    if (!kvm_dirty_ring_enabled()) {
+        error_setg(errp, "dirty ring not enable, needed by dirty restraint!");
+        return;
+    }
+
+    if (unlikely(!dirtylimit_cancel_vcpu(idx))) {
+        dirtylimit_calc_quit();
+    }
+}
+
+void dirtylimit_setup(int max_cpus)
+{
+    if (!kvm_dirty_ring_enabled()) {
+        return;
+    }
+
+    dirtylimit_calc_state_init(max_cpus);
+    dirtylimit_state_init(max_cpus);
+}
diff --git a/include/hw/core/cpu.h b/include/hw/core/cpu.h
index e948e81..11df012 100644
--- a/include/hw/core/cpu.h
+++ b/include/hw/core/cpu.h
@@ -881,6 +881,15 @@ void end_exclusive(void);
  */
 void qemu_init_vcpu(CPUState *cpu);
 
+/**
+ * dirtylimit_setup:
+ *
+ * Initializes the global state of dirtylimit calculation and
+ * dirtylimit itself. This is prepared for vCPU dirtylimit which
+ * could be triggered during vm lifecycle.
+ */
+void dirtylimit_setup(int max_cpus);
+
 #define SSTEP_ENABLE  0x1  /* Enable simulated HW single stepping */
 #define SSTEP_NOIRQ   0x2  /* Do not use IRQ while single stepping */
 #define SSTEP_NOTIMER 0x4  /* Do not Timers while single stepping */
diff --git a/qapi/migration.json b/qapi/migration.json
index bbfd48c..42b260e 100644
--- a/qapi/migration.json
+++ b/qapi/migration.json
@@ -1850,6 +1850,49 @@
 { 'command': 'query-dirty-rate', 'returns': 'DirtyRateInfo' }
 
 ##
+# @set-dirty-limit:
+#
+# Set the upper limit of dirty page rate for the interested vCPU.
+#
+# This command could be used to cap the vCPU memory load, which is also
+# refered as "dirty page rate". Users can use set-dirty-limit unconditionally,
+# but if one want to know which vCPU is in high memory load and which vCPU
+# should be limited, "calc-dirty-rate" with "dirty-ring" mode maybe an
+# availiable method.
+#
+# @idx: vCPU index to set dirtylimit.
+#
+# @dirtyrate: upper limit for the specified vCPU's dirty page rate (MB/s)
+#
+# Since: 6.3
+#
+# Example:
+#   {"execute": "set-dirty-limit"}
+#    "arguments": { "idx": 0,
+#                   "dirtyrate": 200 } }
+#
+##
+{ 'command': 'set-dirty-limit',
+  'data': { 'idx': 'int', 'dirtyrate': 'uint64' } }
+
+##
+# @cancel-dirty-limit:
+#
+# Cancel the dirtylimit for the vCPU which has been set with set-dirty-limit.
+#
+# @idx: vCPU index to canceled the dirtylimit
+#
+# Since: 6.3
+#
+# Example:
+#   {"execute": "cancel-dirty-limit"}
+#    "arguments": { "idx": 0 } }
+#
+##
+{ 'command': 'cancel-dirty-limit',
+  'data': { 'idx': 'int' } }
+
+##
 # @snapshot-save:
 #
 # Save a VM snapshot
diff --git a/softmmu/vl.c b/softmmu/vl.c
index 620a1f1..0f83ce3 100644
--- a/softmmu/vl.c
+++ b/softmmu/vl.c
@@ -3777,5 +3777,6 @@ void qemu_init(int argc, char **argv, char **envp)
     qemu_init_displays();
     accel_setup_post(current_machine);
     os_setup_post();
+    dirtylimit_setup(current_machine->smp.max_cpus);
     resume_mux_open();
 }
-- 
1.8.3.1