Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches (rebased Stefan's pull request)

# gpg: Signature made Thu 12 Nov 2015 15:34:16 GMT using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream: (43 commits)
  block: Update copyright of the accounting code
  scsi-disk: Account for failed operations
  macio: Account for failed operations
  ide: Account for failed and invalid operations
  atapi: Account for failed and invalid operations
  xen_disk: Account for failed and invalid operations
  virtio-blk: Account for failed and invalid operations
  nvme: Account for failed and invalid operations
  iotests: Add test for the block device statistics
  block: Use QEMU_CLOCK_VIRTUAL for the accounting code in qtest mode
  qemu-io: Account for failed, invalid and flush operations
  block: New option to define the intervals for collecting I/O statistics
  block: Add average I/O queue depth to BlockDeviceTimedStats
  block: Compute minimum, maximum and average I/O latencies
  block: Allow configuring whether to account failed and invalid ops
  block: Add statistics for failed and invalid I/O operations
  block: Add idle_time_ns to BlockDeviceStats
  util: Infrastructure for computing recent averages
  block: define 'clock_type' for the accounting code
  ide: Account for write operations correctly
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

5 files changed, 207 insertions, 11 deletions

diff --git a/include/block/accounting.h b/include/block/accounting.h
index 66637cdfed..0f46cb4ec1 100644
--- a/include/block/accounting.h
+++ b/include/block/accounting.h
@@ -2,6 +2,7 @@
  * QEMU System Emulator block accounting
  *
  * Copyright (c) 2011 Christoph Hellwig
+ * Copyright (c) 2015 Igalia, S.L.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a copy
  * of this software and associated documentation files (the "Software"), to deal
@@ -25,8 +26,12 @@
 #define BLOCK_ACCOUNTING_H
 
 #include <stdint.h>
+#include <stdbool.h>
 
 #include "qemu/typedefs.h"
+#include "qemu/timed-average.h"
+
+typedef struct BlockAcctTimedStats BlockAcctTimedStats;
 
 enum BlockAcctType {
     BLOCK_ACCT_READ,
@@ -35,11 +40,23 @@ enum BlockAcctType {
     BLOCK_MAX_IOTYPE,
 };
 
+struct BlockAcctTimedStats {
+    TimedAverage latency[BLOCK_MAX_IOTYPE];
+    unsigned interval_length; /* in seconds */
+    QSLIST_ENTRY(BlockAcctTimedStats) entries;
+};
+
 typedef struct BlockAcctStats {
     uint64_t nr_bytes[BLOCK_MAX_IOTYPE];
     uint64_t nr_ops[BLOCK_MAX_IOTYPE];
+    uint64_t invalid_ops[BLOCK_MAX_IOTYPE];
+    uint64_t failed_ops[BLOCK_MAX_IOTYPE];
     uint64_t total_time_ns[BLOCK_MAX_IOTYPE];
     uint64_t merged[BLOCK_MAX_IOTYPE];
+    int64_t last_access_time_ns;
+    QSLIST_HEAD(, BlockAcctTimedStats) intervals;
+    bool account_invalid;
+    bool account_failed;
 } BlockAcctStats;
 
 typedef struct BlockAcctCookie {
@@ -48,10 +65,21 @@ typedef struct BlockAcctCookie {
     enum BlockAcctType type;
 } BlockAcctCookie;
 
+void block_acct_init(BlockAcctStats *stats, bool account_invalid,
+                     bool account_failed);
+void block_acct_cleanup(BlockAcctStats *stats);
+void block_acct_add_interval(BlockAcctStats *stats, unsigned interval_length);
+BlockAcctTimedStats *block_acct_interval_next(BlockAcctStats *stats,
+                                              BlockAcctTimedStats *s);
 void block_acct_start(BlockAcctStats *stats, BlockAcctCookie *cookie,
                       int64_t bytes, enum BlockAcctType type);
 void block_acct_done(BlockAcctStats *stats, BlockAcctCookie *cookie);
+void block_acct_failed(BlockAcctStats *stats, BlockAcctCookie *cookie);
+void block_acct_invalid(BlockAcctStats *stats, enum BlockAcctType type);
 void block_acct_merge_done(BlockAcctStats *stats, enum BlockAcctType type,
                            int num_requests);
+int64_t block_acct_idle_time_ns(BlockAcctStats *stats);
+double block_acct_queue_depth(BlockAcctTimedStats *stats,
+                              enum BlockAcctType type);
 
 #endif
diff --git a/include/block/block.h b/include/block/block.h
index 610db923d5..73edb1a79c 100644
--- a/include/block/block.h
+++ b/include/block/block.h
@@ -14,6 +14,7 @@ typedef struct BlockDriver BlockDriver;
 typedef struct BlockJob BlockJob;
 typedef struct BdrvChild BdrvChild;
 typedef struct BdrvChildRole BdrvChildRole;
+typedef struct BlockJobTxn BlockJobTxn;
 
 typedef struct BlockDriverInfo {
     /* in bytes, 0 if irrelevant */
@@ -335,10 +336,18 @@ void bdrv_aio_cancel_async(BlockAIOCB *acb);
 
 typedef struct BlockRequest {
     /* Fields to be filled by multiwrite caller */
-    int64_t sector;
-    int nb_sectors;
-    int flags;
-    QEMUIOVector *qiov;
+    union {
+        struct {
+            int64_t sector;
+            int nb_sectors;
+            int flags;
+            QEMUIOVector *qiov;
+        };
+        struct {
+            int req;
+            void *buf;
+        };
+    };
     BlockCompletionFunc *cb;
     void *opaque;
 
@@ -493,7 +502,6 @@ void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
                            int64_t cur_sector, int nr_sectors);
 void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
                              int64_t cur_sector, int nr_sectors);
-void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap);
 void bdrv_dirty_iter_init(BdrvDirtyBitmap *bitmap, struct HBitmapIter *hbi);
 void bdrv_set_dirty_iter(struct HBitmapIter *hbi, int64_t offset);
 int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap);
diff --git a/include/block/block_int.h b/include/block/block_int.h
index 603145a21d..4012e36437 100644
--- a/include/block/block_int.h
+++ b/include/block/block_int.h
@@ -60,11 +60,19 @@
 
 #define BLOCK_PROBE_BUF_SIZE        512
 
+enum BdrvTrackedRequestType {
+    BDRV_TRACKED_READ,
+    BDRV_TRACKED_WRITE,
+    BDRV_TRACKED_FLUSH,
+    BDRV_TRACKED_IOCTL,
+    BDRV_TRACKED_DISCARD,
+};
+
 typedef struct BdrvTrackedRequest {
     BlockDriverState *bs;
     int64_t offset;
     unsigned int bytes;
-    bool is_write;
+    enum BdrvTrackedRequestType type;
 
     bool serialising;
     int64_t overlap_offset;
@@ -219,7 +227,6 @@ struct BlockDriver {
     void (*bdrv_lock_medium)(BlockDriverState *bs, bool locked);
 
     /* to control generic scsi devices */
-    int (*bdrv_ioctl)(BlockDriverState *bs, unsigned long int req, void *buf);
     BlockAIOCB *(*bdrv_aio_ioctl)(BlockDriverState *bs,
         unsigned long int req, void *buf,
         BlockCompletionFunc *cb, void *opaque);
@@ -288,6 +295,12 @@ struct BlockDriver {
      */
     int (*bdrv_probe_geometry)(BlockDriverState *bs, HDGeometry *geo);
 
+    /**
+     * Drain and stop any internal sources of requests in the driver, and
+     * remain so until next I/O callback (e.g. bdrv_co_writev) is called.
+     */
+    void (*bdrv_drain)(BlockDriverState *bs);
+
     QLIST_ENTRY(BlockDriver) list;
 };
 
@@ -656,6 +669,7 @@ void mirror_start(BlockDriverState *bs, BlockDriverState *target,
  * @on_target_error: The action to take upon error writing to the target.
  * @cb: Completion function for the job.
  * @opaque: Opaque pointer value passed to @cb.
+ * @txn: Transaction that this job is part of (may be NULL).
  *
  * Start a backup operation on @bs.  Clusters in @bs are written to @target
  * until the job is cancelled or manually completed.
@@ -666,7 +680,7 @@ void backup_start(BlockDriverState *bs, BlockDriverState *target,
                   BlockdevOnError on_source_error,
                   BlockdevOnError on_target_error,
                   BlockCompletionFunc *cb, void *opaque,
-                  Error **errp);
+                  BlockJobTxn *txn, Error **errp);
 
 void blk_set_bs(BlockBackend *blk, BlockDriverState *bs);
 
@@ -680,4 +694,7 @@ void blk_dev_resize_cb(BlockBackend *blk);
 void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector, int nr_sectors);
 bool bdrv_requests_pending(BlockDriverState *bs);
 
+void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap **out);
+void bdrv_undo_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap, HBitmap *in);
+
 #endif /* BLOCK_INT_H */
diff --git a/include/block/blockjob.h b/include/block/blockjob.h
index 289b13f0c0..d84ccd8d2c 100644
--- a/include/block/blockjob.h
+++ b/include/block/blockjob.h
@@ -50,6 +50,26 @@ typedef struct BlockJobDriver {
      * manually.
      */
     void (*complete)(BlockJob *job, Error **errp);
+
+    /**
+     * If the callback is not NULL, it will be invoked when all the jobs
+     * belonging to the same transaction complete; or upon this job's
+     * completion if it is not in a transaction. Skipped if NULL.
+     *
+     * All jobs will complete with a call to either .commit() or .abort() but
+     * never both.
+     */
+    void (*commit)(BlockJob *job);
+
+    /**
+     * If the callback is not NULL, it will be invoked when any job in the
+     * same transaction fails; or upon this job's failure (due to error or
+     * cancellation) if it is not in a transaction. Skipped if NULL.
+     *
+     * All jobs will complete with a call to either .commit() or .abort() but
+     * never both.
+     */
+    void (*abort)(BlockJob *job);
 } BlockJobDriver;
 
 /**
@@ -130,6 +150,21 @@ struct BlockJob {
 
     /** The opaque value that is passed to the completion function.  */
     void *opaque;
+
+    /** Reference count of the block job */
+    int refcnt;
+
+    /* True if this job has reported completion by calling block_job_completed.
+     */
+    bool completed;
+
+    /* ret code passed to block_job_completed.
+     */
+    int ret;
+
+    /** Non-NULL if this job is part of a transaction */
+    BlockJobTxn *txn;
+    QLIST_ENTRY(BlockJob) txn_list;
 };
 
 /**
@@ -174,12 +209,21 @@ void block_job_sleep_ns(BlockJob *job, QEMUClockType type, int64_t ns);
 void block_job_yield(BlockJob *job);
 
 /**
- * block_job_release:
+ * block_job_ref:
  * @bs: The block device.
  *
- * Release job resources when an error occurred or job completed.
+ * Grab a reference to the block job. Should be paired with block_job_unref.
  */
-void block_job_release(BlockDriverState *bs);
+void block_job_ref(BlockJob *job);
+
+/**
+ * block_job_unref:
+ * @bs: The block device.
+ *
+ * Release reference to the block job and release resources if it is the last
+ * reference.
+ */
+void block_job_unref(BlockJob *job);
 
 /**
  * block_job_completed:
@@ -364,4 +408,39 @@ void block_job_defer_to_main_loop(BlockJob *job,
                                   BlockJobDeferToMainLoopFn *fn,
                                   void *opaque);
 
+/**
+ * block_job_txn_new:
+ *
+ * Allocate and return a new block job transaction.  Jobs can be added to the
+ * transaction using block_job_txn_add_job().
+ *
+ * The transaction is automatically freed when the last job completes or is
+ * cancelled.
+ *
+ * All jobs in the transaction either complete successfully or fail/cancel as a
+ * group.  Jobs wait for each other before completing.  Cancelling one job
+ * cancels all jobs in the transaction.
+ */
+BlockJobTxn *block_job_txn_new(void);
+
+/**
+ * block_job_txn_unref:
+ *
+ * Release a reference that was previously acquired with block_job_txn_add_job
+ * or block_job_txn_new. If it's the last reference to the object, it will be
+ * freed.
+ */
+void block_job_txn_unref(BlockJobTxn *txn);
+
+/**
+ * block_job_txn_add_job:
+ * @txn: The transaction (may be NULL)
+ * @job: Job to add to the transaction
+ *
+ * Add @job to the transaction.  The @job must not already be in a transaction.
+ * The caller must call either block_job_txn_unref() or block_job_completed()
+ * to release the reference that is automatically grabbed here.
+ */
+void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job);
+
 #endif
diff --git a/include/qemu/timed-average.h b/include/qemu/timed-average.h
new file mode 100644
index 0000000000..364bf88f70
--- /dev/null
+++ b/include/qemu/timed-average.h
@@ -0,0 +1,64 @@
+/*
+ * QEMU timed average computation
+ *
+ * Copyright (C) Nodalink, EURL. 2014
+ * Copyright (C) Igalia, S.L. 2015
+ *
+ * Authors:
+ *   Benoît Canet <benoit.canet@nodalink.com>
+ *   Alberto Garcia <berto@igalia.com>
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 2 of the License, or
+ * (at your option) version 3 or any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef TIMED_AVERAGE_H
+#define TIMED_AVERAGE_H
+
+#include <stdint.h>
+
+#include "qemu/timer.h"
+
+typedef struct TimedAverageWindow TimedAverageWindow;
+typedef struct TimedAverage TimedAverage;
+
+/* All fields of both structures are private */
+
+struct TimedAverageWindow {
+    uint64_t      min;             /* minimum value accounted in the window */
+    uint64_t      max;             /* maximum value accounted in the window */
+    uint64_t      sum;             /* sum of all values */
+    uint64_t      count;           /* number of values */
+    int64_t       expiration;      /* the end of the current window in ns */
+};
+
+struct TimedAverage {
+    uint64_t           period;     /* period in nanoseconds */
+    TimedAverageWindow windows[2]; /* two overlapping windows of with
+                                    * an offset of period / 2 between them */
+    unsigned           current;    /* the current window index: it's also the
+                                    * oldest window index */
+    QEMUClockType      clock_type; /* the clock used */
+};
+
+void timed_average_init(TimedAverage *ta, QEMUClockType clock_type,
+                        uint64_t period);
+
+void timed_average_account(TimedAverage *ta, uint64_t value);
+
+uint64_t timed_average_min(TimedAverage *ta);
+uint64_t timed_average_avg(TimedAverage *ta);
+uint64_t timed_average_max(TimedAverage *ta);
+uint64_t timed_average_sum(TimedAverage *ta, uint64_t *elapsed);
+
+#endif