Skip to contentSkip to conversation

Migration: Remove copying of guest ram pages

PATCH Superseded Orit Wasserman 5 participants 37 messages Mar 21, 2013

v3 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

v3 cover Migration: Remove copying of guest ram pages

In migration all data is copied to a static buffer in QEMUFile, this hurts our network bandwidth and CPU usage especially with large guests. We switched to iovec for storing different buffers to send (even a byte field is considered as a buffer) and use writev to send the iovec.
writev was chosen (as apposed to sendmsg) because it supprts non socket fds.

Guest memory pages are not copied by calling a new function qemu_put_buffer_no_copy.
The page header data and device state data are still copied into the static buffer. This data consists of a lot of bytes and integer fields and the static buffer is used to store it during batching.
Another improvement is changing qemu_putbe64/32/16 to create a single buffer instead of several byte sized buffer.

git repository: git://github.com/oritwas/qemu.git sendv_v2

Change from v2:
Always send data for the iovec even if writev_buffer is not implemented. Coalesce adjacent iovecs to create one big buffer from small adjacent buffer.

Changes from v1:
Use iov_send for socket.
Make writev_buffer optional and if it is not implemented use put_buffer

Future work: Make number of iovec changeable

Orit Wasserman (9):
Add QemuFileWritevBuffer QemuFileOps
Add socket_writev_buffer function
Update bytes_xfer in qemu_put_byte
Store the data to send also in iovec
Use writev ops if available
More optimized qemu_put_be64/32/16
Add qemu_put_buffer_no_copy
Use qemu_put_buffer_no_copy for guest memory pages coalesce adjacent iovecs

 arch_init.c                   |   2 +-
 include/migration/qemu-file.h |  12 ++++
 savevm.c                      | 127 ++++++++++++++++++++++++++++++++++--------
 3 files changed, 116 insertions(+), 25 deletions(-)
Signature · 1 line
1.7.11.7
#1 lore raw

PATCH v3 1/9 Add QemuFileWritevBuffer QemuFileOps

This will allow us to write an iovec

---
 include/migration/qemu-file.h | 7 +++++++
 1 file changed, 7 insertions(+)
include/migration/qemu-file.h +7 -0
index df81261..8d3da9b 100644
@@ -51,11 +51,18 @@ typedef int (QEMUFileCloseFunc)(void *opaque);
  */
 typedef int (QEMUFileGetFD)(void *opaque);
 
+/*
+ * This function writes an iovec to file.
+ */
+typedef int (QEMUFileWritevBufferFunc)(void *opaque, struct iovec *iov,
+                                       int iovcnt);
+
 typedef struct QEMUFileOps {
     QEMUFilePutBufferFunc *put_buffer;
     QEMUFileGetBufferFunc *get_buffer;
     QEMUFileCloseFunc *close;
     QEMUFileGetFD *get_fd;
+    QEMUFileWritevBufferFunc *writev_buffer;
 } QEMUFileOps;
 
 QEMUFile *qemu_fopen_ops(void *opaque, const QEMUFileOps *ops);
Signature · 1 line
1.7.11.7
#3 lore raw

PATCH v3 2/9 Add socket_writev_buffer function

In reply to Orit Wasserman
---
 savevm.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
savevm.c +14 -0
index 35c8d1e..baa45ae 100644
@@ -39,6 +39,7 @@
 #include "qmp-commands.h"
 #include "trace.h"
 #include "qemu/bitops.h"
+#include "qemu/iov.h"
 
 #define SELF_ANNOUNCE_ROUNDS 5
 
@@ -171,6 +172,18 @@ static void coroutine_fn yield_until_fd_readable(int fd)
     qemu_coroutine_yield();
 }
 
+static int socket_writev_buffer(void *opaque, struct iovec *iov, int iovcnt)
+{
+    QEMUFileSocket *s = opaque;
+    ssize_t len;
+
+    len = iov_send(s->fd, iov, iovcnt, 0, iov_size(iov, iovcnt));
+    if (len < iov_size(iov, iovcnt)) {
+        len = -socket_error();
+    }
+    return len;
+}
+
 static int socket_get_fd(void *opaque)
 {
     QEMUFileSocket *s = opaque;
@@ -387,6 +400,7 @@ static const QEMUFileOps socket_read_ops = {
 static const QEMUFileOps socket_write_ops = {
     .get_fd =     socket_get_fd,
     .put_buffer = socket_put_buffer,
+    .writev_buffer = socket_writev_buffer,
     .close =      socket_close
 };
 
Signature · 1 line
1.7.11.7
#4 lore raw

Re: [Qemu-devel] [PATCH v3 2/9] Add socket_writev_buffer function

Orit Wasserman <[email protected]> wrote:
D> Signed-off-by: Orit Wasserman <[email protected]>

Quoted 17 lines
---
 savevm.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
savevm.c +1 -0
index 35c8d1e..baa45ae 100644
@@ -39,6 +39,7 @@
 #include "qmp-commands.h"
 #include "trace.h"
 #include "qemu/bitops.h"
+#include "qemu/iov.h"
 
 #define SELF_ANNOUNCE_ROUNDS 5
 

savevm.c static void coroutine_fn yield_until_fd_readable(int fd) · line 181

Quoted 10 lines
savevm.c +7 -0
@@ -171,6 +172,18 @@ static void coroutine_fn yield_until_fd_readable(int fd)
     qemu_coroutine_yield();
 }
 
+static int socket_writev_buffer(void *opaque, struct iovec *iov, int iovcnt)
+{
+    QEMUFileSocket *s = opaque;
+    ssize_t len;
+
+    len = iov_send(s->fd, iov, iovcnt, 0, iov_size(iov, iovcnt));
+    if (len < iov_size(iov, iovcnt)) {

Not the more expensive operation ever, but we can factor
iov_size(iov, iovcnt)
value into one variable.

Quoted 15 lines
savevm.c +6 -0
+        len = -socket_error();
+    }
+    return len;
+}
+
 static int socket_get_fd(void *opaque)
 {
     QEMUFileSocket *s = opaque;
@@ -387,6 +400,7 @@ static const QEMUFileOps socket_read_ops = {
 static const QEMUFileOps socket_write_ops = {
     .get_fd =     socket_get_fd,
     .put_buffer = socket_put_buffer,
+    .writev_buffer = socket_writev_buffer,
     .close =      socket_close
 };
#5 lore raw

PATCH v3 3/9 Update bytes_xfer in qemu_put_byte

In reply to Orit Wasserman
---
 savevm.c | 2 ++
 1 file changed, 2 insertions(+)
savevm.c +2 -0
index baa45ae..686c8c8 100644
@@ -647,6 +647,8 @@ void qemu_put_byte(QEMUFile *f, int v)
 
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
+    f->bytes_xfer += 1;
+
     if (f->buf_index >= IO_BUF_SIZE) {
         qemu_fflush(f);
     }
Signature · 1 line
1.7.11.7
#7 lore raw

Re: [Qemu-devel] [PATCH v3 3/9] Update bytes_xfer in qemu_put_byte

In reply to Orit Wasserman
Quoted 10 lines · Orit Wasserman
---
 savevm.c | 2 ++
 1 file changed, 2 insertions(+)
savevm.c +0 -0
index baa45ae..686c8c8 100644

savevm.c void qemu_put_byte(QEMUFile *f, int v) · line 650

Quoted 4 lines
savevm.c +1 -0
@@ -647,6 +647,8 @@ void qemu_put_byte(QEMUFile *f, int v)
 
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
+    f->bytes_xfer += 1;

Why ' += 1' instead of the shorter '++'?

Signature · 2 lines
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org
#8 lore raw

Re: [Qemu-devel] [PATCH v3 3/9] Update bytes_xfer in qemu_put_byte

Quoted 12 lines · Eric Blake
Quoted 11 lines · Orit Wasserman
---
 savevm.c | 2 ++
 1 file changed, 2 insertions(+)
savevm.c +0 -0
index baa45ae..686c8c8 100644
@@ -647,6 +647,8 @@ void qemu_put_byte(QEMUFile *f, int v)
 
Quoted 6 lines

savevm.c void qemu_put_byte(QEMUFile *f, int v) · line 650

Quoted 3 lines
savevm.c +1 -0
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
+    f->bytes_xfer += 1;

Why ' += 1' instead of the shorter '++'?

It compiles to the same code but I can change it :)

#9 lore raw

PATCH v3 4/9 Store the data to send also in iovec

In reply to Orit Wasserman

All data is still copied into the static buffer.

---
 savevm.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
savevm.c +11 -2
index 686c8c8..ab81dd3 100644
@@ -114,6 +114,7 @@ void qemu_announce_self(void)
 /* savevm/loadvm support */
 
 #define IO_BUF_SIZE 32768
+#define MAX_IOV_SIZE MIN(IOV_MAX, 64)
 
 struct QEMUFile {
     const QEMUFileOps *ops;
@@ -129,6 +130,9 @@ struct QEMUFile {
     int buf_size; /* 0 when writing */
     uint8_t buf[IO_BUF_SIZE];
 
+    struct iovec iov[MAX_IOV_SIZE];
+    unsigned int iovcnt;
+
     int last_error;
 };
 
@@ -527,6 +531,7 @@ static void qemu_fflush(QEMUFile *f)
             f->pos += f->buf_index;
         }
         f->buf_index = 0;
+        f->iovcnt = 0;
     }
     if (ret < 0) {
         qemu_file_set_error(f, ret);
@@ -619,12 +624,14 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size)
         if (l > size)
             l = size;
         memcpy(f->buf + f->buf_index, buf, l);
+        f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+        f->iov[f->iovcnt++].iov_len = l;
         f->is_write = 1;
         f->buf_index += l;
         f->bytes_xfer += l;
         buf += l;
         size -= l;
-        if (f->buf_index >= IO_BUF_SIZE) {
+        if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
             qemu_fflush(f);
             if (qemu_file_get_error(f)) {
                 break;
@@ -648,8 +655,10 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
+    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+    f->iov[f->iovcnt++].iov_len = 1;
 
-    if (f->buf_index >= IO_BUF_SIZE) {
+    if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
     }
 }
Signature · 1 line
1.7.11.7
Reviewed-by on v3 4/9
#10 lore raw

Re: [Qemu-devel] [PATCH v3 4/9] Store the data to send also in iovec

Quoted 38 lines · Orit Wasserman

All data is still copied into the static buffer.

---
 savevm.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
savevm.c +5 -0
index 686c8c8..ab81dd3 100644
@@ -114,6 +114,7 @@ void qemu_announce_self(void)
 /* savevm/loadvm support */
 
 #define IO_BUF_SIZE 32768
+#define MAX_IOV_SIZE MIN(IOV_MAX, 64)
 
 struct QEMUFile {
     const QEMUFileOps *ops;
@@ -129,6 +130,9 @@ struct QEMUFile {
     int buf_size; /* 0 when writing */
     uint8_t buf[IO_BUF_SIZE];
 
+    struct iovec iov[MAX_IOV_SIZE];
+    unsigned int iovcnt;
+
     int last_error;
 };
 
@@ -527,6 +531,7 @@ static void qemu_fflush(QEMUFile *f)
             f->pos += f->buf_index;
         }
         f->buf_index = 0;
+        f->iovcnt = 0;
     }
     if (ret < 0) {
         qemu_file_set_error(f, ret);

savevm.c void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size) · line 631

Quoted 8 lines
savevm.c +2 -0
@@ -619,12 +624,14 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size)
         if (l > size)
             l = size;
         memcpy(f->buf + f->buf_index, buf, l);
+        f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+        f->iov[f->iovcnt++].iov_len = l;
         f->is_write = 1;
         f->buf_index += l;
         f->bytes_xfer += l;

After reading the patch several times, we need to do a s/l/len/ on this function, 1 and l are too similar, and I was reading the wrong value :p

Quoted 9 lines
savevm.c +1 -1
         buf += l;
         size -= l;
-        if (f->buf_index >= IO_BUF_SIZE) {
+        if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
             qemu_fflush(f);
             if (qemu_file_get_error(f)) {
                 break;
@@ -648,8 +655,10 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;

savevm.c void qemu_put_byte(QEMUFile *f, int v) · line 664

Quoted 10 lines
savevm.c +3 -1
     f->is_write = 1;
     f->bytes_xfer += 1;
+    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+    f->iov[f->iovcnt++].iov_len = 1;
 
-    if (f->buf_index >= IO_BUF_SIZE) {
+    if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
     }
 }
#11 lore raw

PATCH v3 5/9 Use writev ops if available

In reply to Orit Wasserman

Update qemu_fflush and stdio_close to use writev ops if they are available Use the buffers stored in the iovec.

---
 savevm.c | 30 +++++++++++++++++++++++-------
 1 file changed, 23 insertions(+), 7 deletions(-)
savevm.c +23 -7
index ab81dd3..f69dce3 100644
@@ -292,7 +292,7 @@ static int stdio_fclose(void *opaque)
     QEMUFileStdio *s = opaque;
     int ret = 0;
 
-    if (s->file->ops->put_buffer) {
+    if (s->file->ops->put_buffer || s->file->ops->writev_buffer) {
         int fd = fileno(s->stdio_file);
         struct stat st;
 
@@ -515,24 +515,40 @@ static void qemu_file_set_error(QEMUFile *f, int ret)
     }
 }
 
-/** Flushes QEMUFile buffer
+/**
+ * Flushes QEMUFile buffer
  *
+ * If there is writev_buffer QEMUFileOps it uses it otherwise uses
+ * put_buffer ops.
  */
 static void qemu_fflush(QEMUFile *f)
 {
     int ret = 0;
+    int i = 0;
 
-    if (!f->ops->put_buffer) {
+    if (!f->ops->writev_buffer && !f->ops->put_buffer) {
         return;
     }
-    if (f->is_write && f->buf_index > 0) {
-        ret = f->ops->put_buffer(f->opaque, f->buf, f->pos, f->buf_index);
-        if (ret >= 0) {
-            f->pos += f->buf_index;
+
+    if (f->is_write && f->iovcnt > 0) {
+        if (f->ops->writev_buffer) {
+            ret = f->ops->writev_buffer(f->opaque, f->iov, f->iovcnt);
+            if (ret >= 0) {
+                f->pos += ret;
+            }
+        } else {
+            for (i = 0; i < f->iovcnt && ret >= 0; i++) {
+                ret = f->ops->put_buffer(f->opaque, f->iov[i].iov_base, f->pos,
+                                         f->iov[i].iov_len);
+                if (ret >= 0) {
+                    f->pos += ret;
+                }
+            }
         }
         f->buf_index = 0;
         f->iovcnt = 0;
     }
+
     if (ret < 0) {
         qemu_file_set_error(f, ret);
     }
Signature · 1 line
1.7.11.7
Reviewed-by on v3 5/9
#12 lore raw

Re: [Qemu-devel] [PATCH v3 5/9] Use writev ops if available

Quoted 4 lines · Orit Wasserman

Update qemu_fflush and stdio_close to use writev ops if they are available Use the buffers stored in the iovec.

I think that qemu_fflush() should only be called for QEMUFiles opened in write mode, but that is independent of this series.

#13 lore raw

PATCH v3 6/9 More optimized qemu_put_be64/32/16

In reply to Orit Wasserman

qemu_put_be functions used qemu_put_byte this caused lots of 1 bytes buffers in the iovec.
we move to use cpu_put_be64/32/16wu and put a single buffer per call.

---
 savevm.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
savevm.c +12 -8
index f69dce3..83aa9e7 100644
@@ -795,22 +795,26 @@ void qemu_file_reset_rate_limit(QEMUFile *f)
 
 void qemu_put_be16(QEMUFile *f, unsigned int v)
 {
-    qemu_put_byte(f, v >> 8);
-    qemu_put_byte(f, v);
+    uint16_t p;
+    cpu_to_be16wu(&p, v);
+
+    qemu_put_buffer(f, (uint8_t *)&p, sizeof(p));
 }
 
 void qemu_put_be32(QEMUFile *f, unsigned int v)
 {
-    qemu_put_byte(f, v >> 24);
-    qemu_put_byte(f, v >> 16);
-    qemu_put_byte(f, v >> 8);
-    qemu_put_byte(f, v);
+    uint32_t p;
+    cpu_to_be32wu(&p, v);
+
+    qemu_put_buffer(f, (uint8_t *)&p, sizeof(p));
 }
 
 void qemu_put_be64(QEMUFile *f, uint64_t v)
 {
-    qemu_put_be32(f, v >> 32);
-    qemu_put_be32(f, v);
+    uint64_t p;
+    cpu_to_be64wu(&p, v);
+
+    qemu_put_buffer(f, (uint8_t *)&p, sizeof(p));
 }
 
 unsigned int qemu_get_be16(QEMUFile *f)
Signature · 1 line
1.7.11.7
#15 lore raw

PATCH v3 7/9 Add qemu_put_buffer_no_copy

In reply to Orit Wasserman

This allow us to add a buffer to the iovec to send without copying it into the static buffer.

---
 include/migration/qemu-file.h |  5 +++++
 savevm.c                      | 37 ++++++++++++++++++++++++++++---------
 2 files changed, 33 insertions(+), 9 deletions(-)
include/migration/qemu-file.h +5 -0
index 8d3da9b..5168be2 100644
@@ -75,6 +75,11 @@ int qemu_fclose(QEMUFile *f);
 int64_t qemu_ftell(QEMUFile *f);
 void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size);
 void qemu_put_byte(QEMUFile *f, int v);
+/*
+ * put_buffer without copying the buffer.
+ * The buffer should be available till it is sent.
+ */
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size);
 
 static inline void qemu_put_ubyte(QEMUFile *f, unsigned int v)
 {
savevm.c +28 -9
index 83aa9e7..fbfb9e3 100644
@@ -621,6 +621,30 @@ int qemu_fclose(QEMUFile *f)
     return ret;
 }
 
+
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
+{
+    if (f->last_error) {
+        return;
+    }
+
+    if (f->is_write == 0 && f->buf_index > 0) {
+        fprintf(stderr,
+                "Attempted to write to buffer while read buffer is not empty\n");
+        abort();
+    }
+
+    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt++].iov_len = size;
+
+    f->is_write = 1;
+    f->bytes_xfer += size;
+
+    if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
+        qemu_fflush(f);
+    }
+}
+
 void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size)
 {
     int l;
@@ -640,19 +664,14 @@ void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size)
         if (l > size)
             l = size;
         memcpy(f->buf + f->buf_index, buf, l);
-        f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
-        f->iov[f->iovcnt++].iov_len = l;
         f->is_write = 1;
         f->buf_index += l;
-        f->bytes_xfer += l;
+        qemu_put_buffer_no_copy(f, f->buf + (f->buf_index - l), l);
+        if (qemu_file_get_error(f)) {
+            break;
+        }
         buf += l;
         size -= l;
-        if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
-            qemu_fflush(f);
-            if (qemu_file_get_error(f)) {
-                break;
-            }
-        }
     }
 }
 
Signature · 1 line
1.7.11.7
#16 lore raw

Re: [Qemu-devel] [PATCH v3 7/9] Add qemu_put_buffer_no_copy

Quoted 35 lines · Orit Wasserman

This allow us to add a buffer to the iovec to send without copying it into the static buffer.

---
 include/migration/qemu-file.h |  5 +++++
 savevm.c                      | 37 ++++++++++++++++++++++++++++---------
 2 files changed, 33 insertions(+), 9 deletions(-)
include/migration/qemu-file.h +5 -0
index 8d3da9b..5168be2 100644
@@ -75,6 +75,11 @@ int qemu_fclose(QEMUFile *f);
 int64_t qemu_ftell(QEMUFile *f);
 void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size);
 void qemu_put_byte(QEMUFile *f, int v);
+/*
+ * put_buffer without copying the buffer.
+ * The buffer should be available till it is sent.
+ */
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size);
 
 static inline void qemu_put_ubyte(QEMUFile *f, unsigned int v)
 {
savevm.c +2 -0
index 83aa9e7..fbfb9e3 100644
@@ -621,6 +621,30 @@ int qemu_fclose(QEMUFile *f)
     return ret;
 }
 
+
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)

savevm.c int qemu_fclose(QEMUFile *f) · line 635

Quoted 10 lines
savevm.c +10 -0
+{
+    if (f->last_error) {
+        return;
+    }
+
+    if (f->is_write == 0 && f->buf_index > 0) {
+        fprintf(stderr,
+                "Attempted to write to buffer while read buffer is not empty\n");
+        abort();
+    }

I don't understand this test at all (yes, I know that the test already existed).

We shouldn't never arrived qemu_put_buffer() with a QEMUFile with f->is_write == 0.

Change it for one assert()?

savevm.c int qemu_fclose(QEMUFile *f) · line 638

Quoted 2 lines
savevm.c +2 -0
+    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt++].iov_len = size;

This is clearly wrong, or I have completely missunderstood it (I will give a 50% to each posiblitiy).

Here we should be using "buf" and "size" passed as paramenters, f->buf and f->buf_index shouldn't be used, no?

savevm.c int qemu_fclose(QEMUFile *f) · line 639

Quoted 2 lines
savevm.c +2 -0
+
+    f->is_write = 1;

is_write is completely redundant, and should just be a:

f->ops->put_buffer test (or now with writev). We only set it up when there is anything to be written?

But again, this is independent of this series.

#17 lore raw

Re: [Qemu-devel] [PATCH v3 7/9] Add qemu_put_buffer_no_copy

Quoted 55 lines · Juan Quintela
Quoted 35 lines · Orit Wasserman

This allow us to add a buffer to the iovec to send without copying it into the static buffer.

---
 include/migration/qemu-file.h |  5 +++++
 savevm.c                      | 37 ++++++++++++++++++++++++++++---------
 2 files changed, 33 insertions(+), 9 deletions(-)
include/migration/qemu-file.h +5 -0
index 8d3da9b..5168be2 100644
@@ -75,6 +75,11 @@ int qemu_fclose(QEMUFile *f);
 int64_t qemu_ftell(QEMUFile *f);
 void qemu_put_buffer(QEMUFile *f, const uint8_t *buf, int size);
 void qemu_put_byte(QEMUFile *f, int v);
+/*
+ * put_buffer without copying the buffer.
+ * The buffer should be available till it is sent.
+ */
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size);
 
 static inline void qemu_put_ubyte(QEMUFile *f, unsigned int v)
 {
savevm.c +2 -0
index 83aa9e7..fbfb9e3 100644
@@ -621,6 +621,30 @@ int qemu_fclose(QEMUFile *f)
     return ret;
 }
 
+
+void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)

savevm.c int qemu_fclose(QEMUFile *f) · line 635

Quoted 10 lines
savevm.c +10 -0
+{
+    if (f->last_error) {
+        return;
+    }
+
+    if (f->is_write == 0 && f->buf_index > 0) {
+        fprintf(stderr,
+                "Attempted to write to buffer while read buffer is not empty\n");
+        abort();
+    }

I don't understand this test at all (yes, I know that the test already existed).

We shouldn't never arrived qemu_put_buffer() with a QEMUFile with f->is_write == 0.

Change it for one assert()?

Quoted 8 lines

savevm.c int qemu_fclose(QEMUFile *f) · line 638

Quoted 2 lines
savevm.c +2 -0
+    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt++].iov_len = size;

This is clearly wrong, or I have completely missunderstood it (I will give a 50% to each posiblitiy).

Here we should be using "buf" and "size" passed as paramenters, f->buf and f->buf_index shouldn't be used, no?

you are right, it somehow moved to the next patch :( I will send a fixed v4 after you finish review the series ..

Quoted 11 lines

savevm.c int qemu_fclose(QEMUFile *f) · line 639

Quoted 3 lines
savevm.c +2 -0
+
+    f->is_write = 1;

is_write is completely redundant, and should just be a:

f->ops->put_buffer test (or now with writev). We only set it up when there is anything to be written?

But again, this is independent of this series.

I agree, maybe we need a cleanup series and remove it. It will certainly simplify this code

Orit

#18 lore raw

PATCH v3 8/9 Use qemu_put_buffer_no_copy for guest memory pages

In reply to Orit Wasserman

This will remove an unneeded copy of guest memory pages. For the page header and device state we still copy the data to the static buffer the other option is to allocate the memory on demand which is more expensive.

---
 arch_init.c | 2 +-
 savevm.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch_init.c +1 -1
index 98e2bc6..27b53eb 100644
@@ -481,7 +481,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
             /* XBZRLE overflow or normal page */
             if (bytes_sent == -1) {
                 bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
-                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
+                qemu_put_buffer_no_copy(f, p, TARGET_PAGE_SIZE);
                 bytes_sent += TARGET_PAGE_SIZE;
                 acct_info.norm_pages++;
             }
savevm.c +1 -1
index fbfb9e3..50e8fb2 100644
@@ -634,7 +634,7 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
     f->iov[f->iovcnt++].iov_len = size;
 
     f->is_write = 1;
Signature · 1 line
1.7.11.7
#19 lore raw

Re: [Qemu-devel] [PATCH v3 8/9] Use qemu_put_buffer_no_copy for guest memory pages

Quoted 6 lines · Orit Wasserman

This will remove an unneeded copy of guest memory pages. For the page header and device state we still copy the data to the static buffer the other option is to allocate the memory on demand which is more expensive.

Okay so with this, the _nocopy can be later rewritten to do vmsplice to save another copy? Cool.

Quoted 33 lines
---
 arch_init.c | 2 +-
 savevm.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch_init.c +1 -1
index 98e2bc6..27b53eb 100644
@@ -481,7 +481,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
             /* XBZRLE overflow or normal page */
             if (bytes_sent == -1) {
                 bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
-                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
+                qemu_put_buffer_no_copy(f, p, TARGET_PAGE_SIZE);
                 bytes_sent += TARGET_PAGE_SIZE;
                 acct_info.norm_pages++;
             }
savevm.c +1 -1
index fbfb9e3..50e8fb2 100644
@@ -634,7 +634,7 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
     f->iov[f->iovcnt++].iov_len = size;
 
     f->is_write = 1;
Signature · 1 line
1.7.11.7
#20 lore raw

Re: [Qemu-devel] [PATCH v3 8/9] Use qemu_put_buffer_no_copy for guest memory pages

In reply to Orit Wasserman
Quoted 16 lines · Orit Wasserman

This will remove an unneeded copy of guest memory pages. For the page header and device state we still copy the data to the static buffer the other option is to allocate the memory on demand which is more expensive.

---
 arch_init.c | 2 +-
 savevm.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch_init.c +0 -0
index 98e2bc6..27b53eb 100644

arch_init.c static int ram_save_block(QEMUFile *f, bool last_stage) · line 487

Quoted 8 lines
arch_init.c +1 -1
@@ -481,7 +481,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
             /* XBZRLE overflow or normal page */
             if (bytes_sent == -1) {
                 bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
-                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
+                qemu_put_buffer_no_copy(f, p, TARGET_PAGE_SIZE);
                 bytes_sent += TARGET_PAGE_SIZE;
                 acct_info.norm_pages++;
             }

Once here, shouldn't we also change:

block-migration.c::blk_send()

    qemu_put_buffer(f, blk->buf, BLOCK_SIZE);

to nocopy?

Again, this can be an additional patch.

savevm.c void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size) · line 638

Quoted 11 lines
savevm.c +1 -1
index fbfb9e3..50e8fb2 100644
@@ -634,7 +634,7 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
     f->iov[f->iovcnt++].iov_len = size;

This bit should be in the previous patch, the error that I pointed?

Later, Juan.

#21 lore raw

Re: [Qemu-devel] [PATCH v3 8/9] Use qemu_put_buffer_no_copy for guest memory pages

Quoted 28 lines · Juan Quintela
Quoted 16 lines · Orit Wasserman

This will remove an unneeded copy of guest memory pages. For the page header and device state we still copy the data to the static buffer the other option is to allocate the memory on demand which is more expensive.

---
 arch_init.c | 2 +-
 savevm.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch_init.c +0 -0
index 98e2bc6..27b53eb 100644

arch_init.c static int ram_save_block(QEMUFile *f, bool last_stage) · line 487

Quoted 8 lines
arch_init.c +1 -1
@@ -481,7 +481,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
             /* XBZRLE overflow or normal page */
             if (bytes_sent == -1) {
                 bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
-                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
+                qemu_put_buffer_no_copy(f, p, TARGET_PAGE_SIZE);
                 bytes_sent += TARGET_PAGE_SIZE;
                 acct_info.norm_pages++;
             }

Once here, shouldn't we also change:

Quoted 5 lines

block-migration.c::blk_send()

    qemu_put_buffer(f, blk->buf, BLOCK_SIZE);

to nocopy?

Sadly no, I looked at the code and I saw the buffer is freed immediately after call blk_send :(. look at mig_save_device_dirty.

Orit

Quoted 20 lines

Again, this can be an additional patch.

savevm.c void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size) · line 638

Quoted 11 lines
savevm.c +1 -1
index fbfb9e3..50e8fb2 100644
@@ -634,7 +634,7 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
     f->iov[f->iovcnt++].iov_len = size;

This bit should be in the previous patch, the error that I pointed?

Later, Juan.

#22 lore raw

Re: [Qemu-devel] [PATCH v3 8/9] Use qemu_put_buffer_no_copy for guest memory pages

Quoted 31 lines · Orit Wasserman
Quoted 30 lines · Juan Quintela
Quoted 16 lines · Orit Wasserman

This will remove an unneeded copy of guest memory pages. For the page header and device state we still copy the data to the static buffer the other option is to allocate the memory on demand which is more expensive.

---
 arch_init.c | 2 +-
 savevm.c    | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch_init.c +0 -0
index 98e2bc6..27b53eb 100644

arch_init.c static int ram_save_block(QEMUFile *f, bool last_stage) · line 487

Quoted 8 lines
arch_init.c +1 -1
@@ -481,7 +481,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
             /* XBZRLE overflow or normal page */
             if (bytes_sent == -1) {
                 bytes_sent = save_block_hdr(f, block, offset, cont, RAM_SAVE_FLAG_PAGE);
-                qemu_put_buffer(f, p, TARGET_PAGE_SIZE);
+                qemu_put_buffer_no_copy(f, p, TARGET_PAGE_SIZE);
                 bytes_sent += TARGET_PAGE_SIZE;
                 acct_info.norm_pages++;
             }

Once here, shouldn't we also change:

block-migration.c::blk_send()

Quoted 7 lines
Quoted 3 lines
    qemu_put_buffer(f, blk->buf, BLOCK_SIZE);

to nocopy?

Sadly no, I looked at the code and I saw the buffer is freed immediately after call blk_send :(. look at mig_save_device_dirty.

Orit

Hmm if _nocopy also implies it is not safe to free the buffer immediately, then I'd rather put this in a name e.g. _async instead of _nocopy.

Quoted 20 lines
Quoted 20 lines

Again, this can be an additional patch.

savevm.c void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size) · line 638

Quoted 11 lines
savevm.c +1 -1
index fbfb9e3..50e8fb2 100644
@@ -634,7 +634,7 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = f->buf + f->buf_index;
+    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
     f->iov[f->iovcnt++].iov_len = size;

This bit should be in the previous patch, the error that I pointed?

Later, Juan.

#23 lore raw

PATCH v3 9/9 coalesce adjacent iovecs

In reply to Orit Wasserman

This way we send one big buffer instead of many small ones

---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#24 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

Quoted 45 lines
---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#25 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

It happens in the last phase when we send the device state that consists of a lot bytes and int field that are written using qemu_put_byte/be16/...

Quoted 46 lines
Quoted 46 lines
---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#26 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 11 lines · Orit Wasserman
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

It happens in the last phase when we send the device state that consists of a lot bytes and int field that are written using qemu_put_byte/be16/...

Confused I thought device_state does not use _nocopy? My idea of using vmsplice relies exactly on this:
we can not splice device state ...

Quoted 46 lines
Quoted 46 lines
Quoted 46 lines
---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#27 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 9 lines · Michael S. Tsirkin
Quoted 8 lines · Orit Wasserman
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

Quoted 7 lines
Quoted 3 lines

It happens in the last phase when we send the device state that consists of a lot bytes and int field that are written using qemu_put_byte/be16/...

Confused I thought device_state does not use _nocopy? My idea of using vmsplice relies exactly on this:
we can not splice device state ...

qemu_put_buffer calls qemu_put_buffer_no_copy (you know code reuse) So I guess we will need to add some flag if we want to use vmsplice or not or split no_copy into external and internal. It won't be a big change

Quoted 47 lines
Quoted 47 lines
Quoted 47 lines
Quoted 47 lines

---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#28 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 14 lines · Orit Wasserman
Quoted 13 lines · Michael S. Tsirkin
Quoted 11 lines · Orit Wasserman
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

It happens in the last phase when we send the device state that consists of a lot bytes and int field that are written using qemu_put_byte/be16/...

Quoted 7 lines
Quoted 3 lines

Confused I thought device_state does not use _nocopy? My idea of using vmsplice relies exactly on this:
we can not splice device state ...

qemu_put_buffer calls qemu_put_buffer_no_copy (you know code reuse) So I guess we will need to add some flag if we want to use vmsplice or not or split no_copy into external and internal. It won't be a big change

Fair enough. Have time to look into this? Data copies is like 15% CPU, more with swap ...

Quoted 47 lines
Quoted 47 lines
Quoted 47 lines
Quoted 47 lines
Quoted 47 lines

---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +17 -4
index 50e8fb2..13a533b 100644
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }
 
     f->is_write = 1;
     f->bytes_xfer += size;
@@ -690,8 +696,15 @@ void qemu_put_byte(QEMUFile *f, int v)
     f->buf[f->buf_index++] = v;
     f->is_write = 1;
     f->bytes_xfer += 1;
-    f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
-    f->iov[f->iovcnt++].iov_len = 1;
+
+    /* check for adjoint buffer and colace them */
+    if (f->iovcnt > 0 && f->buf + (f->buf_index - 1)  ==
+        f->iov[f->iovcnt - 1].iov_base + f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += 1;
+    } else {
+        f->iov[f->iovcnt].iov_base = f->buf + (f->buf_index - 1);
+        f->iov[f->iovcnt++].iov_len = 1;
+    }
 
     if (f->buf_index >= IO_BUF_SIZE || f->iovcnt >= MAX_IOV_SIZE) {
         qemu_fflush(f);
Signature · 1 line
1.7.11.7
#29 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

In reply to Michael S. Tsirkin
Quoted 9 lines · Michael S. Tsirkin
Quoted 8 lines · Orit Wasserman
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

Quoted 8 lines
Quoted 4 lines

It happens in the last phase when we send the device state that consists of a lot
bytes and int field that are written using qemu_put_byte/be16/...

Confused I thought device_state does not use _nocopy? My idea of using vmsplice relies exactly on this:
we can not splice device state ...

As it is today, I am not sure that we can use vmsplice() because we are sending:

<header>
<page>
<header>
<page>
<header>
<page>

We can optimize at some pount to write a bigger/different header and sent a bunch of pages together, but just now we don't have that code.

Later, Juan.

#30 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 31 lines · Juan Quintela
Quoted 9 lines · Michael S. Tsirkin
Quoted 8 lines · Orit Wasserman
Quoted 6 lines · Michael S. Tsirkin
Quoted 3 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

Why does this happen BTW?

Quoted 8 lines
Quoted 4 lines

It happens in the last phase when we send the device state that consists of a lot
bytes and int field that are written using qemu_put_byte/be16/...

Confused I thought device_state does not use _nocopy? My idea of using vmsplice relies exactly on this:
we can not splice device state ...

As it is today, I am not sure that we can use vmsplice() because we are sending:

<header>
<page>
<header>
<page>
<header>
<page>

Quoted 4 lines

We can optimize at some pount to write a bigger/different header and sent a bunch of pages together, but just now we don't have that code.

Later, Juan.

Sending the page can do vmsplice, can't it? Multipage is likely a good idea anyway, e.g. RDMA wants to do this too.

#31 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 7 lines · Michael S. Tsirkin
Quoted 7 lines

<header>
<page>
<header>
<page>
<header>
<page>

Quoted 8 lines
Quoted 4 lines

We can optimize at some pount to write a bigger/different header and sent a bunch of pages together, but just now we don't have that code.

Later, Juan.

Sending the page can do vmsplice, can't it? Multipage is likely a good idea anyway, e.g. RDMA wants to do this too.

RDMA requires de pages lock into memory, no?

Later, Juan.

#32 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

Quoted 14 lines · Juan Quintela
Quoted 12 lines · Michael S. Tsirkin
Quoted 11 lines

<header>
<page>
<header>
<page>
<header>
<page>

We can optimize at some pount to write a bigger/different header and sent a bunch of pages together, but just now we don't have that code.

Later, Juan.

Quoted 7 lines
Quoted 3 lines

Sending the page can do vmsplice, can't it? Multipage is likely a good idea anyway, e.g. RDMA wants to do this too.

RDMA requires de pages lock into memory, no?

Later, Juan.

That's a separate issue, but the point is it is asynchronous. It can be made to behave exactly the same implementing this API:

- send async: lock, send
- flush: poll for completion, unlock

#33 lore raw

Re: [Qemu-devel] [PATCH v3 9/9] coalesce adjacent iovecs

In reply to Orit Wasserman
Quoted 12 lines · Orit Wasserman

This way we send one big buffer instead of many small ones

---
 savevm.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
savevm.c +0 -0
index 50e8fb2..13a533b 100644

savevm.c void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size) · line 639

Quoted 8 lines
savevm.c +1 -2
@@ -634,8 +634,14 @@ void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size)
         abort();
     }
 
-    f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
-    f->iov[f->iovcnt++].iov_len = size;
 
 
+    /* check for adjoint buffer and colace them */

s/colace/coalesce/

savevm.c void qemu_put_buffer_no_copy(QEMUFile *f, const uint8_t *buf, int size) · line 644

Quoted 7 lines
savevm.c +7 -0
+    if (f->iovcnt > 0 && buf == f->iov[f->iovcnt - 1].iov_base +
+        f->iov[f->iovcnt - 1].iov_len) {
+        f->iov[f->iovcnt - 1].iov_len += size;
+    } else {
+        f->iov[f->iovcnt].iov_base = (uint8_t *)buf;
+        f->iov[f->iovcnt++].iov_len = size;
+    }

Can we create a function for this?
Would help make sure that we don't forget to "fix" both places when we have changes.

Later, Juan.

#34 lore raw
In reply to Orit Wasserman

Il 21/03/2013 17:05, Orit Wasserman ha scritto:

Quoted 16 lines

In migration all data is copied to a static buffer in QEMUFile, this hurts our network bandwidth and CPU usage especially with large guests. We switched to iovec for storing different buffers to send (even a byte field is considered as a buffer) and use writev to send the iovec.
writev was chosen (as apposed to sendmsg) because it supprts non socket fds.

Guest memory pages are not copied by calling a new function qemu_put_buffer_no_copy.
The page header data and device state data are still copied into the static buffer. This data consists of a lot of bytes and integer fields and the static buffer is used to store it during batching.
Another improvement is changing qemu_putbe64/32/16 to create a single buffer instead of several byte sized buffer.

git repository: git://github.com/oritwas/qemu.git sendv_v2

Quoted 7 lines

Change from v2:
Always send data for the iovec even if writev_buffer is not implemented. Coalesce adjacent iovecs to create one big buffer from small adjacent buffer.

Changes from v1:
Use iov_send for socket.
Make writev_buffer optional and if it is not implemented use put_buffer

I didn't review it too closely, so I don't feel like giving my Reviewed-by. Still, I must say it is a very nice and simple solution. Kudos!

Just one question: patch 6 is not strictly needed anymore now that you have coalescing, no? Should Juan leave it out?

Paolo

#35 lore raw
Quoted 25 lines · Paolo Bonzini

Il 21/03/2013 17:05, Orit Wasserman ha scritto:

Quoted 16 lines

In migration all data is copied to a static buffer in QEMUFile, this hurts our network bandwidth and CPU usage especially with large guests. We switched to iovec for storing different buffers to send (even a byte field is considered as a buffer) and use writev to send the iovec.
writev was chosen (as apposed to sendmsg) because it supprts non socket fds.

Guest memory pages are not copied by calling a new function qemu_put_buffer_no_copy.
The page header data and device state data are still copied into the static buffer. This data consists of a lot of bytes and integer fields and the static buffer is used to store it during batching.
Another improvement is changing qemu_putbe64/32/16 to create a single buffer instead of several byte sized buffer.

git repository: git://github.com/oritwas/qemu.git sendv_v2

Quoted 7 lines

Change from v2:
Always send data for the iovec even if writev_buffer is not implemented. Coalesce adjacent iovecs to create one big buffer from small adjacent buffer.

Changes from v1:
Use iov_send for socket.
Make writev_buffer optional and if it is not implemented use put_buffer

Quoted 6 lines

I didn't review it too closely, so I don't feel like giving my Reviewed-by. Still, I must say it is a very nice and simple solution. Kudos!

Just one question: patch 6 is not strictly needed anymore now that you have coalescing, no? Should Juan leave it out?

true, Juan had a small fix so I'm sending v4 I will remove it.

Quoted 3 lines

Paolo

#36 lore raw
In reply to Orit Wasserman
Quoted 14 lines · Orit Wasserman

In migration all data is copied to a static buffer in QEMUFile, this hurts our network bandwidth and CPU usage especially with large guests. We switched to iovec for storing different buffers to send (even a byte field is considered as a buffer) and use writev to send the iovec.
writev was chosen (as apposed to sendmsg) because it supprts non socket fds.

Guest memory pages are not copied by calling a new function qemu_put_buffer_no_copy.
The page header data and device state data are still copied into the static buffer. This data consists of a lot of bytes and integer fields and the static buffer is used to store it during batching.
Another improvement is changing qemu_putbe64/32/16 to create a single buffer instead of several byte sized buffer.

Quoted 1 line

git repository: git://github.com/oritwas/qemu.git sendv_v2

it is still sendv_v2 or sendv_v3 O:-)

Quoted 21 lines

Change from v2:
Always send data for the iovec even if writev_buffer is not implemented. Coalesce adjacent iovecs to create one big buffer from small adjacent buffer.

Changes from v1:
Use iov_send for socket.
Make writev_buffer optional and if it is not implemented use put_buffer

Future work: Make number of iovec changeable

Orit Wasserman (9):
Add QemuFileWritevBuffer QemuFileOps
Add socket_writev_buffer function
Update bytes_xfer in qemu_put_byte
Store the data to send also in iovec
Use writev ops if available
More optimized qemu_put_be64/32/16
Add qemu_put_buffer_no_copy
Use qemu_put_buffer_no_copy for guest memory pages coalesce adjacent iovecs

Quoted 4 lines
 arch_init.c                   |   2 +-
 include/migration/qemu-file.h |  12 ++++
 savevm.c                      | 127 ++++++++++++++++++++++++++++++++++--------
 3 files changed, 116 insertions(+), 25 deletions(-)

Really nice series. Just a couple of small nits.

Later, Juan.