qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v3] fsdev: add IO throttle support to fsdev devi


From: Pradeep Jagadeesh
Subject: Re: [Qemu-devel] [PATCH v3] fsdev: add IO throttle support to fsdev devices
Date: Thu, 22 Sep 2016 13:46:48 +0200
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.3.0

Hi Greg,

On Wed, 21 Sep 2016 12:22:25 +0200
Pradeep Jagadeesh <address@hidden> wrote:

Hi Greg,

Thanks for having a look at patchset.
See the replies below.

On Fri, 16 Sep 2016 04:33:36 -0400
Pradeep Jagadeesh <address@hidden> wrote:

Uses throttling APIs to limit I/O bandwidth and number of operations on the
fsdev devices.

Signed-off-by: Pradeep Jagadeesh <address@hidden>
---

Hi Pradeep,

Please find some comments below.

 fsdev/Makefile.objs         |   1 +
 fsdev/file-op-9p.h          |   3 +
 fsdev/qemu-fsdev-opts.c     |  76 ++++++++++++++++++
 fsdev/qemu-fsdev-throttle.c | 191 ++++++++++++++++++++++++++++++++++++++++++++
 fsdev/qemu-fsdev-throttle.h |  42 ++++++++++

Is the qemu- prefix really needed ?
Just followed the previous file naming convention,
like qemu-fsdev-opts.c

 hw/9pfs/9p-local.c          |  11 ++-
 hw/9pfs/9p.c                |   7 ++
 hw/9pfs/cofile.c            |   3 +
 8 files changed, 332 insertions(+), 2 deletions(-)
 create mode 100644 fsdev/qemu-fsdev-throttle.c
 create mode 100644 fsdev/qemu-fsdev-throttle.h

This adds the support for the 9p-local driver.
For now this functionality can be enabled only through qemu cli options.
QMP interface and support to other drivers need further extensions.
To make it simple for other drivers, the throttle code has been put in
separate files.

v1 -> v2:

-Fixed FsContext redeclaration issue
-Removed couple of function declarations from 9p-throttle.h
-Fixed some of the .help messages

v2 -> v3:

-Addressed follwing comments by Claudio Fontana
 -Removed redundant memset calls in fsdev_throttle_configure_iolimits function
 -Checking throttle structure validity before initializing other structures
  in fsdev_throttle_configure_iolimits

-Addressed following comments by Greg Kurz
 -Moved the code from 9pfs directory to fsdev directory, because the throttling
  is for the fsdev devices.Renamed the files and functions to fsdev_ from 9pfs_
 -Renamed throttling cli options to throttling.*, as in QMP cli options
 -Removed some of the unwanted .h files from qemu-fsdev-throttle.[ch]
 -Using throttle_enabled() function to set the thottle enabled flag for fsdev.

diff --git a/fsdev/Makefile.objs b/fsdev/Makefile.objs
index 1b120a4..2c6da2d 100644
--- a/fsdev/Makefile.objs
+++ b/fsdev/Makefile.objs
@@ -7,6 +7,7 @@ common-obj-y = qemu-fsdev-dummy.o
 endif
 common-obj-y += qemu-fsdev-opts.o

+common-obj-y += qemu-fsdev-throttle.o
 # Toplevel always builds this; targets without virtio will put it in
 # common-obj-y
 common-obj-$(CONFIG_ALL) += qemu-fsdev-dummy.o
diff --git a/fsdev/file-op-9p.h b/fsdev/file-op-9p.h
index 6db9fea..33fe822 100644
--- a/fsdev/file-op-9p.h
+++ b/fsdev/file-op-9p.h
@@ -17,6 +17,7 @@
 #include <dirent.h>
 #include <utime.h>
 #include <sys/vfs.h>
+#include "qemu-fsdev-throttle.h"

 #define SM_LOCAL_MODE_BITS    0600
 #define SM_LOCAL_DIR_MODE_BITS    0700
@@ -74,6 +75,7 @@ typedef struct FsDriverEntry {
     char *path;
     int export_flags;
     FileOperations *ops;
+    FsThrottle fst;
 } FsDriverEntry;

 typedef struct FsContext
@@ -83,6 +85,7 @@ typedef struct FsContext
     int export_flags;
     struct xattr_operations **xops;
     struct extended_ops exops;
+    FsThrottle *fst;
     /* fs driver specific data */
     void *private;
 } FsContext;
diff --git a/fsdev/qemu-fsdev-opts.c b/fsdev/qemu-fsdev-opts.c
index 1dd8c7a..395d497 100644
--- a/fsdev/qemu-fsdev-opts.c
+++ b/fsdev/qemu-fsdev-opts.c
@@ -37,6 +37,82 @@ static QemuOptsList qemu_fsdev_opts = {
         }, {
             .name = "sock_fd",
             .type = QEMU_OPT_NUMBER,
+        }, {
+            .name = "throttling.iops-total",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit total I/O operations per second",
+        },{
+            .name = "throttling.iops-read",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit read operations per second",
+        },{
+            .name = "throttling.iops-write",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit write operations per second",
+        },{
+            .name = "throttling.bps-total",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit total bytes per second",
+        },{
+            .name = "throttling.bps-read",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit read bytes per second",
+        },{
+            .name = "throttling.bps-write",
+            .type = QEMU_OPT_NUMBER,
+            .help = "limit write bytes per second",
+        },{
+            .name = "throttling.iops-total-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "I/O operations burst",
+        },{
+            .name = "throttling.iops-read-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "I/O operations read burst",
+        },{
+            .name = "throttling.iops-write-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "I/O operations write burst",
+        },{
+            .name = "throttling.bps-total-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "total bytes burst",
+        },{
+            .name = "throttling.bps-read-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "total bytes read burst",
+        },{
+            .name = "throttling.bps-write-max",
+            .type = QEMU_OPT_NUMBER,
+            .help = "total bytes write burst",
+        },{
+            .name = "throttling.iops-total-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the iops-total-max burst period, in seconds",
+        },{
+            .name = "throttling.iops-read-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the iops-read-max burst period, in seconds",
+        },{
+            .name = "throttling.iops-write-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the iops-write-max burst period, in seconds",
+        },{
+            .name = "throttling.bps-total-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the bps-total-max burst period, in seconds",
+        },{
+            .name = "throttling.bps-read-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the bps-read-max burst period, in seconds",
+        },{
+            .name = "throttling.bps-write-max-length",
+            .type = QEMU_OPT_NUMBER,
+            .help = "length of the bps-write-max burst period, in seconds",
+        },{
+            .name = "throttling.iops-size",
+            .type = QEMU_OPT_NUMBER,
+            .help = "when limiting by iops max size of an I/O in bytes",
         },

         { /*End of list */ }
diff --git a/fsdev/qemu-fsdev-throttle.c b/fsdev/qemu-fsdev-throttle.c
new file mode 100644
index 0000000..0adccd1
--- /dev/null
+++ b/fsdev/qemu-fsdev-throttle.c
@@ -0,0 +1,191 @@
+/*
+ * Fsdev Throttle
+ *
+ * Copyright (C) 2016 Huawei Technologies Duesseldorf GmbH
+ *
+ * Author: Pradeep Jagadeesh <address@hidden>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or
+ * (at your option) any later version.
+ *
+ * See the COPYING file in the top-level directory for details.
+ *
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "qapi/error.h"
+#include "qemu-fsdev-throttle.h"
+
+static bool fsdev_throttle_check_for_wait(FsThrottle *fst, bool is_write)
+{
+    if (fst->any_timer_armed[is_write]) {
+        return true;
+    } else {
+        return throttle_schedule_timer(&fst->ts, &fst->tt, is_write);
+    }
+}
+
+static void fsdev_throttle_schedule_next_request(FsThrottle *fst, bool 
is_write)
+{
+    bool must_wait = fsdev_throttle_check_for_wait(fst, is_write);

Let's add a newline here.
Done!

+    if (!fst->pending_reqs[is_write]) {
+        return;
+    }
+    if (!must_wait) {
+        if (qemu_in_coroutine() &&
+            qemu_co_queue_next(&fst->throttled_reqs[is_write])) {
+            ;
+       } else {
+           int64_t now = qemu_clock_get_ns(fst->tt.clock_type);
+           timer_mod(fst->tt.timers[is_write], now + 1);
+           fst->any_timer_armed[is_write] = true;
+       }
+   }
+}
+
+static void fsdev_throttle_timer_cb(FsThrottle *fst, bool is_write)
+{
+    bool empty_queue;

Same here.
Done!

+    qemu_mutex_lock(&fst->lock);
+    fst->any_timer_armed[is_write] = false;
+    qemu_mutex_unlock(&fst->lock);
+    empty_queue = !qemu_co_enter_next(&fst->throttled_reqs[is_write]);
+    if (empty_queue) {
+        qemu_mutex_lock(&fst->lock);
+        fsdev_throttle_schedule_next_request(fst, is_write);
+        qemu_mutex_unlock(&fst->lock);
+    }
+}
+
+bool fsdev_throttle_enabled(FsThrottle *fst)
+{
+    return fst->enabled;
+}
+
+static void fsdev_iolimits_enable(FsThrottle *fst)
+{
+    fst->enabled = true;
+}
+
+static void fsdev_iolimits_disable(FsThrottle *fst)
+{
+    fst->enabled = false;
+}
+

fsdev_throttle_configure_iolimits() is the only function that will ever
write to fst->enabled. The helpers seem to indicate the opposite. It is
better to open code in fsdev_throttle_configure_iolimits().
You mean just use  fst->enabled = true; inside the
fsdev_throttle_configure_iolimits().?
If that is the case, I thought these helpers are useful in future when
we have qmp interfaces to enable,disable the throttle dynamically/run-time.

Hmm... in this case, something looks wrong.

If we start with a valid throttling config, then disable it with QMP and then
unrealize the device, we won't do the cleanup... fst->enabled looks like
a premature optimization which causes confusion actually.

I'd prefer fsdev to behave like blockdev, which doesn't have an enabled flag.

Please drop it completely. :)
Done!


+static void fsdev_throttle_read_timer_cb(void *opaque)
+{
+    fsdev_throttle_timer_cb(opaque, false);
+}
+
+static void fsdev_throttle_write_timer_cb(void *opaque)
+{
+    fsdev_throttle_timer_cb(opaque, true);
+}
+
+static void fsdev_parse_io_limit_opts(QemuOpts *opts, FsThrottle *fst)
+{
+    fst->cfg.buckets[THROTTLE_BPS_TOTAL].avg =
+          qemu_opt_get_number(opts, "throttling.bps-total", 0);
+    fst->cfg.buckets[THROTTLE_BPS_READ].avg  =
+          qemu_opt_get_number(opts, "throttling.bps-read", 0);
+    fst->cfg.buckets[THROTTLE_BPS_WRITE].avg =
+          qemu_opt_get_number(opts, "throttling.bps-write", 0);
+    fst->cfg.buckets[THROTTLE_OPS_TOTAL].avg =
+          qemu_opt_get_number(opts, "throttling.iops-total", 0);
+    fst->cfg.buckets[THROTTLE_OPS_READ].avg =
+          qemu_opt_get_number(opts, "throttling.iops-read", 0);
+    fst->cfg.buckets[THROTTLE_OPS_WRITE].avg =
+          qemu_opt_get_number(opts, "throttling.iops-write", 0);
+
+    fst->cfg.buckets[THROTTLE_BPS_TOTAL].max =
+          qemu_opt_get_number(opts, "throttling.bps-total-max", 0);
+    fst->cfg.buckets[THROTTLE_BPS_READ].max  =
+          qemu_opt_get_number(opts, "throttling.bps-read-max", 0);
+    fst->cfg.buckets[THROTTLE_BPS_WRITE].max =
+          qemu_opt_get_number(opts, "throttling.bps-write-max", 0);
+    fst->cfg.buckets[THROTTLE_OPS_TOTAL].max =
+          qemu_opt_get_number(opts, "throttling.iops-total-max", 0);
+    fst->cfg.buckets[THROTTLE_OPS_READ].max =
+          qemu_opt_get_number(opts, "throttling.iops-read-max", 0);
+    fst->cfg.buckets[THROTTLE_OPS_WRITE].max =
+          qemu_opt_get_number(opts, "throttling.iops-write-max", 0);
+
+    fst->cfg.buckets[THROTTLE_BPS_TOTAL].burst_length =
+          qemu_opt_get_number(opts, "throttling.bps-total-max-length", 1);
+    fst->cfg.buckets[THROTTLE_BPS_READ].burst_length  =
+          qemu_opt_get_number(opts, "throttling.bps-read-max-length", 1);
+    fst->cfg.buckets[THROTTLE_BPS_WRITE].burst_length =
+          qemu_opt_get_number(opts, "throttling.bps-write-max-length", 1);
+    fst->cfg.buckets[THROTTLE_OPS_TOTAL].burst_length =
+          qemu_opt_get_number(opts, "throttling.iops-total-max-length", 1);
+    fst->cfg.buckets[THROTTLE_OPS_READ].burst_length =
+          qemu_opt_get_number(opts, "throttling.iops-read-max-length", 1);
+    fst->cfg.buckets[THROTTLE_OPS_WRITE].burst_length =
+          qemu_opt_get_number(opts, "throttling.iops-write-max-length", 1);
+    fst->cfg.op_size =
+          qemu_opt_get_number(opts, "throttling.iops-size", 0);
+}
+
+int fsdev_throttle_configure_iolimits(QemuOpts *opts, FsThrottle *fst)
+{
+    Error *err = NULL;
+
+    /* Parse and set populate the cfg structure */
+    fsdev_parse_io_limit_opts(opts, fst);
+
+    if (!throttle_is_valid(&fst->cfg, &err)) {
+        error_report("Throttle configuration is not valid");

throttle_is_valid() already populates err with a meaningful message and
you should use it:

        error_reportf_err(err, "Throttle configuration is not valid: ");

OK,fixed.
+        return -1;
+    }
+    if (throttle_enabled(&fst->cfg)) {
+        fst->aioctx = qemu_get_aio_context();
+        if (!fst->aioctx) {
+            error_report("Failed to create AIO Context");
+            return -1;
+        }

qemu_get_aio_context() returns qemu_aio_context which is initialized in
qemu_init_main_loop(). It is guarantied to be non NULL and if it is, we'd
better dump core.

OK, Is there any specific dump api that exists in qemu?

QEMU uses g_assert() for elaborate checks but you really don't need to check
the return value of qemu_get_aio_context(). Just grep through the code to
be convinced :)
I did change it to g_assert((fst->aioctx = qemu_get_aio_context()));
is that ok?

+        throttle_init(&fst->ts);
+        throttle_timers_init(&fst->tt,
+                             fst->aioctx,
+                             QEMU_CLOCK_REALTIME,
+                             fsdev_throttle_read_timer_cb,
+                             fsdev_throttle_write_timer_cb,
+                             fst);
+        throttle_config(&fst->ts, &fst->tt, &fst->cfg);
+        qemu_co_queue_init(&fst->throttled_reqs[0]);
+        qemu_co_queue_init(&fst->throttled_reqs[1]);
+        fst->pending_reqs[0] = 0;
+        fst->pending_reqs[1] = 0;
+        fst->any_timer_armed[0] = false;
+        fst->any_timer_armed[1] = false;
+        qemu_mutex_init(&fst->lock);
+        fsdev_iolimits_enable(fst);
+   } else {
+        fsdev_iolimits_disable(fst);
+   }
+   return 0;
+}
+
+void fsdev_throttle_request(FsThrottle *fst, bool is_write, ssize_t bytes)
+{
+    if (fsdev_throttle_enabled(fst)) {
+        qemu_mutex_lock(&fst->lock);
+        bool must_wait = fsdev_throttle_check_for_wait(fst, is_write);

Even if the QEMU code is supposed to follow C99, I think it is prettier to
declare variables at the beginning of the code block.

           bool must_wait;
OK,fixed.

           qemu_mutex_lock(&fst->lock);
           must_wait = fsdev_throttle_check_for_wait(fst, is_write);

+        if (must_wait || fst->pending_reqs[is_write]) {
+            fst->pending_reqs[is_write]++;
+            qemu_mutex_unlock(&fst->lock);
+            qemu_co_queue_wait(&fst->throttled_reqs[is_write]);
+            qemu_mutex_lock(&fst->lock);
+            fst->pending_reqs[is_write]--;
+       }
+       throttle_account(&fst->ts, is_write, bytes);
+       fsdev_throttle_schedule_next_request(fst, is_write);
+       qemu_mutex_unlock(&fst->lock);
+    }
+}
+
+void fsdev_throttle_cleanup(FsThrottle *fst)
+{
+    throttle_timers_destroy(&fst->tt);
+    qemu_mutex_destroy(&fst->lock);
+}
diff --git a/fsdev/qemu-fsdev-throttle.h b/fsdev/qemu-fsdev-throttle.h
new file mode 100644
index 0000000..73dd367
--- /dev/null
+++ b/fsdev/qemu-fsdev-throttle.h
@@ -0,0 +1,42 @@
+/*
+ * Fsdev Throttle
+ *
+ * Copyright (C) 2016 Huawei Technologies Duesseldorf GmbH
+ *
+ * Author: Pradeep Jagadeesh <address@hidden>
+ *
+ * This work is licensed under the terms of the GNU GPL, version 2 or
+ * (at your option) any later version.
+ *
+ * See the COPYING file in the top-level directory for details.
+ *
+ */
+
+#ifndef _FSDEV_THROTTLE_H
+#define _FSDEV_THROTTLE_H
+
+#include "block/aio.h"
+#include "qemu/main-loop.h"
+#include "qemu/coroutine.h"
+#include "qemu/throttle.h"
+
+typedef struct FsThrottle {
+    ThrottleState ts;
+    ThrottleTimers tt;
+    AioContext   *aioctx;
+    ThrottleConfig cfg;
+    bool enabled;
+    CoQueue      throttled_reqs[2];
+    unsigned     pending_reqs[2];
+    bool any_timer_armed[2];
+    QemuMutex lock;
+} FsThrottle;
+
+bool fsdev_throttle_enabled(FsThrottle *);
+
+int fsdev_throttle_configure_iolimits(QemuOpts *, FsThrottle *);
+
+void fsdev_throttle_request(FsThrottle *, bool , ssize_t);
+
+void fsdev_throttle_cleanup(FsThrottle *);
+#endif /* _FSDEV_THROTTLE_H */
diff --git a/hw/9pfs/9p-local.c b/hw/9pfs/9p-local.c
index 3f271fc..b1009fc 100644
--- a/hw/9pfs/9p-local.c
+++ b/hw/9pfs/9p-local.c
@@ -436,8 +436,8 @@ static ssize_t local_pwritev(FsContext *ctx, 
V9fsFidOpenState *fs,
                              const struct iovec *iov,
                              int iovcnt, off_t offset)
 {
-    ssize_t ret
-;
+    ssize_t ret;
+
 #ifdef CONFIG_PREADV
     ret = pwritev(fs->fd, iov, iovcnt, offset);
 #else
@@ -1213,6 +1213,8 @@ static int local_parse_opts(QemuOpts *opts, struct 
FsDriverEntry *fse)
     const char *sec_model = qemu_opt_get(opts, "security_model");
     const char *path = qemu_opt_get(opts, "path");

+    FsThrottle *fst = &fse->fst;
+
     if (!sec_model) {
         error_report("Security model not specified, local fs needs security 
model");
         error_printf("valid options are:"
@@ -1240,6 +1242,11 @@ static int local_parse_opts(QemuOpts *opts, struct 
FsDriverEntry *fse)
         error_report("fsdev: No path specified");
         return -1;
     }
+
+    if (fsdev_throttle_configure_iolimits(opts, fst) < 0) {
+        return -1;
+    }
+
     fse->path = g_strdup(path);

     return 0;
diff --git a/hw/9pfs/9p.c b/hw/9pfs/9p.c
index dfe293d..3748ba9 100644
--- a/hw/9pfs/9p.c
+++ b/hw/9pfs/9p.c
@@ -3490,6 +3490,10 @@ int v9fs_device_realize_common(V9fsState *s, Error 
**errp)
         error_setg(errp, "share path %s is not a directory", fse->path);
         goto out;
     }
+
+    /* Throttle structure initialization */

I'm still thinking this comment paraphrases the code :)
I missed it!:)

+    s->ctx.fst = &fse->fst;
+
     v9fs_path_free(&path);

     rc = 0;
@@ -3504,6 +3508,9 @@ out:

 void v9fs_device_unrealize_common(V9fsState *s, Error **errp)
 {
+    if (fsdev_throttle_enabled(s->ctx.fst)) {
+        fsdev_throttle_cleanup(s->ctx.fst);
+    }
     g_free(s->ctx.fs_root);
     g_free(s->tag);
 }
diff --git a/hw/9pfs/cofile.c b/hw/9pfs/cofile.c
index 10343c0..65897d5 100644
--- a/hw/9pfs/cofile.c
+++ b/hw/9pfs/cofile.c
@@ -16,6 +16,7 @@
 #include "qemu/thread.h"
 #include "qemu/coroutine.h"
 #include "coth.h"
+#include "fsdev/qemu-fsdev-throttle.h"

 int v9fs_co_st_gen(V9fsPDU *pdu, V9fsPath *path, mode_t st_mode,
                    V9fsStatDotl *v9stat)
@@ -242,6 +243,7 @@ int v9fs_co_pwritev(V9fsPDU *pdu, V9fsFidState *fidp,
     int err;
     V9fsState *s = pdu->s;

+    fsdev_throttle_request(s->ctx.fst, true, iov->iov_len);

I think it would make more sense to move this after the cancellation point.
Done!

     if (v9fs_request_cancelled(pdu)) {
         return -EINTR;
     }
@@ -261,6 +263,7 @@ int v9fs_co_preadv(V9fsPDU *pdu, V9fsFidState *fidp,
     int err;
     V9fsState *s = pdu->s;

+    fsdev_throttle_request(s->ctx.fst, false, iov->iov_len);

Same here.

Done!
     if (v9fs_request_cancelled(pdu)) {
         return -EINTR;
     }

Cheers.

--
Greg

Regards,
Pradeep






reply via email to

[Prev in Thread] Current Thread [Next in Thread]