date:20130823

Re: [Qemu-devel] [PATCH 13/18] arch_init: adjust ram_save_setup() for migrate_is_localhost

2013-08-23 Thread Lei Li


On 08/21/2013 06:48 PM, Paolo Bonzini wrote:

Il 21/08/2013 09:18, Lei Li ha scritto:

Send all the ram blocks hooked by save_page, which will copy
ram page and MADV_DONTNEED the page just copied.

You should implement this entirely in the hook.

It will be a little less efficient because of the dirty bitmap overhead,
but you should aim at having *zero* changes in arch_init.c and migration.c.


Yes, the reason I modify the migration_thread() to have new process that send 
all
the ram pages in adjusted qemu_savevm_state_begin stage and send device states 
in
qemu_savevm_device_state stage for localhost migration is to avoid the bitmap 
thing,
which is a little less efficient just like you mentioned above.

The performance assurance is very important to this feature, our goal is 100ms
of downtime for a 1TB guest.



Paolo


Signed-off-by: Lei Li li...@linux.vnet.ibm.com
---
  arch_init.c |   19 +--
  1 files changed, 13 insertions(+), 6 deletions(-)

diff --git a/arch_init.c b/arch_init.c
index 434a4ca..cbbb4db 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -474,7 +474,7 @@ static int ram_save_block(QEMUFile *f, bool last_stage)
  /* In doubt sent page as normal */
  bytes_sent = -1;
  ret = ram_control_save_page(f, block-offset,
-   offset, TARGET_PAGE_SIZE, bytes_sent);
+offset, TARGET_PAGE_SIZE, bytes_sent);
  
  if (ret != RAM_SAVE_CONTROL_NOT_SUPP) {

  if (ret != RAM_SAVE_CONTROL_DELAYED) {
@@ -613,11 +613,13 @@ static int ram_save_setup(QEMUFile *f, void *opaque)
  RAMBlock *block;
  int64_t ram_pages = last_ram_offset()  TARGET_PAGE_BITS;
  
-migration_bitmap = bitmap_new(ram_pages);

-bitmap_set(migration_bitmap, 0, ram_pages);
-migration_dirty_pages = ram_pages;
-mig_throttle_on = false;
-dirty_rate_high_cnt = 0;
+if (!migrate_is_localhost()) {
+migration_bitmap = bitmap_new(ram_pages);
+bitmap_set(migration_bitmap, 0, ram_pages);
+migration_dirty_pages = ram_pages;
+mig_throttle_on = false;
+dirty_rate_high_cnt = 0;
+}
  
  if (migrate_use_xbzrle()) {

  XBZRLE.cache = cache_init(migrate_xbzrle_cache_size() /
@@ -641,6 +643,11 @@ static int ram_save_setup(QEMUFile *f, void *opaque)
  migration_bitmap_sync();
  qemu_mutex_unlock_iothread();
  
+if (migrate_is_localhost()) {

+ram_save_blocks(f);
+return 0;
+}
+
  qemu_put_be64(f, ram_bytes_total() | RAM_SAVE_FLAG_MEM_SIZE);
  
  QTAILQ_FOREACH(block, ram_list.blocks, next) {







--
Lei

Re: [Qemu-devel] [PATCH 06/18] bugfix: wrong error set by ram_control_load_hook()

2013-08-23 Thread Lei Li


On 08/23/2013 01:34 PM, Paolo Bonzini wrote:

On 08/21/2013 06:40 PM, Paolo Bonzini wrote:

Il 21/08/2013 09:18, Lei Li ha scritto:

It should set negative error value if there has been an error.

Signed-off-by: Lei Li li...@linux.vnet.ibm.com
---
   savevm.c |2 +-
   1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/savevm.c b/savevm.c
index 1522d95..f10e031 100644
--- a/savevm.c
+++ b/savevm.c
@@ -649,7 +649,7 @@ void ram_control_after_iterate(QEMUFile *f, uint64_t
flags)
   
   void ram_control_load_hook(QEMUFile *f, uint64_t flags)

   {
-int ret = 0;
+int ret = -1;

-EINVAL

OK, thanks.

Can you extract the patches that I reviewed positively, plus this
one, and send them separately?


Sure, I will send it later. :)



Thanks!

Paolo




--
Lei

Re: [Qemu-devel] [PATCH] block: Fix race in gluster_finish_aiocb

2013-08-23 Thread Bharata B Rao

On Wed, Aug 21, 2013 at 05:40:11PM +0200, Paolo Bonzini wrote:
 
 We could just use a bottom half, too.  Add a bottom half to acb,
 schedule it in gluster_finish_aiocb, delete it in the bottom half's own
 callback.

I tried this approach (the patch at the end of this mail), but see this
occasional errors, doesn't happen always. Any clues on what am I missing here ?

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffcbfff700 (LWP 11301)]
0x5597156b in event_notifier_set (e=0xa4)
at util/event_notifier-posix.c:97
97  ret = write(e-wfd, value, sizeof(value));
(gdb) bt
#0  0x5597156b in event_notifier_set (e=0xa4)
at util/event_notifier-posix.c:97
#1  0x555dd3cd in aio_notify (ctx=0x0) at async.c:233
#2  0x555dd00f in qemu_bh_schedule (bh=0x7fff38e0) at async.c:128
#3  0x556002da in gluster_finish_aiocb (fd=0x562ae5d0, ret=4096, 
arg=0x7fffd00419c0) at block/gluster.c:409
#4  0x75e70cdc in glfs_io_async_cbk (ret=4096, frame=0x0, data=
0x56443ee0) at glfs-fops.c:567
#5  0x75c2843e in synctask_wrap (old_task=0x56365940)
at syncop.c:133
#6  0x73b03370 in ?? () from /lib64/libc.so.6
#7  0x in ?? ()
(gdb) up
#1  0x555dd3cd in aio_notify (ctx=0x0) at async.c:233
233 event_notifier_set(ctx-notifier);
(gdb) up
#2  0x555dd00f in qemu_bh_schedule (bh=0x7fff38e0) at async.c:128
128 aio_notify(bh-ctx);
(gdb) p *bh
$1 = {ctx = 0x0, cb = 0x555ffdcd qemu_gluster_aio_bh, opaque = 
0x7fffd00419c0, next = 0x56345e70, scheduled = false, idle = false, 
  deleted = true}


gluster: Use BH mechanism for AIO completion

Replace the existing pipe based AIO completion handing by BH based method.

Signed-off-by: Bharata B Rao bhar...@linux.vnet.ibm.com
---
 block/gluster.c |   69 ++-
 1 file changed, 8 insertions(+), 61 deletions(-)

diff --git a/block/gluster.c b/block/gluster.c
index 46f36f8..598b335 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -30,7 +30,6 @@ typedef struct GlusterAIOCB {
 
 typedef struct BDRVGlusterState {
 struct glfs *glfs;
-int fds[2];
 struct glfs_fd *fd;
 int event_reader_pos;
 GlusterAIOCB *event_acb;
@@ -231,12 +230,13 @@ out:
 return NULL;
 }
 
-static void qemu_gluster_complete_aio(GlusterAIOCB *acb, BDRVGlusterState *s)
+static void qemu_gluster_aio_bh(void *opaque)
 {
 int ret;
+GlusterAIOCB *acb = opaque;
 bool *finished = acb-finished;
 BlockDriverCompletionFunc *cb = acb-common.cb;
-void *opaque = acb-common.opaque;
+void *cb_opaque = acb-common.opaque;
 
 if (!acb-ret || acb-ret == acb-size) {
 ret = 0; /* Success */
@@ -246,33 +246,15 @@ static void qemu_gluster_complete_aio(GlusterAIOCB *acb, 
BDRVGlusterState *s)
 ret = -EIO; /* Partial read/write - fail it */
 }
 
+qemu_bh_delete(acb-bh);
 qemu_aio_release(acb);
-cb(opaque, ret);
+
+cb(cb_opaque, ret);
 if (finished) {
 *finished = true;
 }
 }
 
-static void qemu_gluster_aio_event_reader(void *opaque)
-{
-BDRVGlusterState *s = opaque;
-ssize_t ret;
-
-do {
-char *p = (char *)s-event_acb;
-
-ret = read(s-fds[GLUSTER_FD_READ], p + s-event_reader_pos,
-   sizeof(s-event_acb) - s-event_reader_pos);
-if (ret  0) {
-s-event_reader_pos += ret;
-if (s-event_reader_pos == sizeof(s-event_acb)) {
-s-event_reader_pos = 0;
-qemu_gluster_complete_aio(s-event_acb, s);
-}
-}
-} while (ret  0  errno == EINTR);
-}
-
 /* TODO Convert to fine grained options */
 static QemuOptsList runtime_opts = {
 .name = gluster,
@@ -329,17 +311,7 @@ static int qemu_gluster_open(BlockDriverState *bs,  QDict 
*options,
 s-fd = glfs_open(s-glfs, gconf-image, open_flags);
 if (!s-fd) {
 ret = -errno;
-goto out;
-}
-
-ret = qemu_pipe(s-fds);
-if (ret  0) {
-ret = -errno;
-goto out;
 }
-fcntl(s-fds[GLUSTER_FD_READ], F_SETFL, O_NONBLOCK);
-qemu_aio_set_fd_handler(s-fds[GLUSTER_FD_READ],
-qemu_gluster_aio_event_reader, NULL, s);
 
 out:
 qemu_opts_del(opts);
@@ -417,31 +389,10 @@ static const AIOCBInfo gluster_aiocb_info = {
 static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
 {
 GlusterAIOCB *acb = (GlusterAIOCB *)arg;
-BlockDriverState *bs = acb-common.bs;
-BDRVGlusterState *s = bs-opaque;
-int retval;
 
 acb-ret = ret;
-retval = qemu_write_full(s-fds[GLUSTER_FD_WRITE], acb, sizeof(acb));
-if (retval != sizeof(acb)) {
-/*
- * Gluster AIO callback thread failed to notify the waiting
- * QEMU thread about IO completion.
- *
- * Complete this IO request and make the disk inaccessible for
- * subsequent reads and

Re: [Qemu-devel] VNC key presses not correct

2013-08-23 Thread Erik Rull

On August 23, 2013 at 7:42 AM Markus Armbruster arm...@redhat.com wrote:

Anthony Liguori anth...@codemonkey.ws writes:

On Aug 22, 2013 4:55 PM, Erik Rull erik.r...@rdsoftware.de wrote:

Markus Armbruster wrote:

Erik Rull erik.r...@rdsoftware.de writes:

Markus Armbruster wrote:

Erik Rull erik.r...@rdsoftware.de writes:

Hi all,

I'm struggling with the QEMU VNC on qemu-kvm-1.2.0 a bit, the following
two
things are not working properly:

Have you tried to reproduce on a current version?

Hello Markus,

yes, I was able to reproduce this with the released qemu-1.6.0 from
the qemu-wiki website. Character mistakes are the same when opening a
Notepad on Windows XP on the guest.

Thanks!

Maybe nobody cares because english is the layout used nearly everywhere
:-)

I nobody cared, we wouldn't ship it :)

Cc'ing Gerd, who knows more about VNC than I do.

And it is not related to the host operating system, I used a Debian
6.0 and a self built linux - both with the same effect.
And it is not related to the VNC client, I tried several clients on
several operating systems, all with the same result.

Do you see this effect with a non-english input layout on your side, too?
When using SDL/direct input everything is great.

Your initial report has details on some keypresses. Please also tell us
your complete QEMU command line, and your complete VNC client command
line(s).

Sure, here it is:
./qemu-system-x86_64 -drive file=../../../qemu-img/windows.img,cache=off
-readconfig ../docs/ich9-ehci-uhci.cfg -device usb-tablet,bus=ehci.0 -boot
c -no-acpi -monitor telnet::5100,server,nowait -vnc :1 -m 1024 --enable-kvm
-cpu host

You need to specify a keymap with -k

Yes, because...

Client is e.g. Ultra VNC 32 bit under Windows XP V 1.0.8.0 = Just enter
the IP::Port and connect. Or RealVNC in the same way.
Client OS has german keyboard and german keyboard layout - on the QEMU
guest side it doesn't matter which input language is set.
Guest is Windows XP SP3 32 bit.

... you use a VNC client that doesn't understand the extended key event
extension. The extension enables 100% faithful key interpretation, -k
can only approximate it.

https://www.berrange.com/posts/2010/07/04/more-than-you-or-i-ever-wanted-to-know-about-virtual-keyboard-handling/

Thanks, that's it! Updated to the latest VNC client version - works (btw.
TigerVNC crashes when sending Ctrl+Alt+Del). One thing has to be done due to the
default english keyboard layout: switch the input language on the client side to
english - then all keypresses are interpreted correctly. This is acceptable due
to the multilanguage user group of the guest and less effort than rebooting the
guest when a user with a different keyboard layout logs in :-)

Best regards,

Erik

[Qemu-devel] [PATCH] migration: Fix debug print type

2013-08-23 Thread Christoffer Dall

The printf args are uint64_t and with -Werr QEMU doesn't compile with
migration debugging turned on unless this is fixed.  Fix it.

Signed-off-by: Christoffer Dall christoffer.d...@linaro.org
---
 migration.c |2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/migration.c b/migration.c
index 1402fa7..0d8706f 100644
--- a/migration.c
+++ b/migration.c
@@ -566,7 +566,7 @@ static void *migration_thread(void *opaque)
 if (!qemu_file_rate_limit(s-file)) {
 DPRINTF(iterate\n);
 pending_size = qemu_savevm_state_pending(s-file, max_size);
-DPRINTF(pending size %lu max %lu\n, pending_size, max_size);
+DPRINTF(pending size %llu max %llu\n, pending_size, max_size);
 if (pending_size  pending_size = max_size) {
 qemu_savevm_state_iterate(s-file);
 } else {
-- 
1.7.10.4

Re: [Qemu-devel] [PATCH] migration: Fix debug print type

2013-08-23 Thread Stefan Weil

See my comment below.

Am 23.08.2013 09:06, schrieb Christoffer Dall:
 The printf args are uint64_t and with -Werr QEMU doesn't compile with
 migration debugging turned on unless this is fixed.  Fix it.

 Signed-off-by: Christoffer Dall christoffer.d...@linaro.org
 ---
  migration.c |2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

 diff --git a/migration.c b/migration.c
 index 1402fa7..0d8706f 100644
 --- a/migration.c
 +++ b/migration.c
 @@ -566,7 +566,7 @@ static void *migration_thread(void *opaque)
  if (!qemu_file_rate_limit(s-file)) {
  DPRINTF(iterate\n);
  pending_size = qemu_savevm_state_pending(s-file, max_size);
 -DPRINTF(pending size %lu max %lu\n, pending_size, max_size);
 +DPRINTF(pending size %llu max %llu\n, pending_size, max_size);

Please use PRIu64 for uint64_t.

  if (pending_size  pending_size = max_size) {
  qemu_savevm_state_iterate(s-file);
  } else {

1 2 3 >

1 - 100 of 203 matches

Mail list logo