On 04/18/2012 12:40 PM, Juan Quintela wrote:
Anthony Liguori<anth...@codemonkey.ws> wrote:
On 04/11/2012 01:49 PM, Orit Wasserman wrote:
Signed-off-by: Orit Wasserman<owass...@redhat.com>
Signed-off-by: Benoit Hudzia<benoit.hud...@sap.com>
Signed-off-by: Petter Svard<pett...@cs.umu.se>
Signed-off-by: Aidan Shribman<aidan.shrib...@sap.com>
---
arch_init.c | 26 ++++++++++++++------------
1 files changed, 14 insertions(+), 12 deletions(-)
diff --git a/arch_init.c b/arch_init.c
index 2e534f1..47b9fef 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -347,6 +347,18 @@ void cache_resize(int64_t new_size)
g_free(old_page_cache);
}
+static void save_block_hdr(QEMUFile *f, RAMBlock *block, ram_addr_t offset,
+ int cont, int flag)
+{
+ qemu_put_be64(f, offset | cont | flag);
+ if (!cont) {
+ qemu_put_byte(f, strlen(block->idstr));
+ qemu_put_buffer(f, (uint8_t *)block->idstr,
+ strlen(block->idstr));
+ }
+
+}
Any time we're changing protocols/formats, we need to document it. We
need a docs/xbzrle.txt (and it should be before this patch).
Agreed.
It's not clear to me how we preserve compatibility here either.
You're potentially passing garbage to an older QEMU.
I think not. Only if you are migrating with the -x option. And then,
you get what you asked for.
I thought I had previously asked for a monitor command to negotiate
extensions for migration?
I think it is in another patch.
So I think we also need:
{ 'command': 'set-migration-capabilities',
'data': { 'enable': ['MigrationCapability'] }
Then a management tool just needs to:
caps_from_src = query-migration-capabilities(src_qmp_session)
caps_from_dst = query-migration-capabilities(dst_qmp_session)
common_caps = intersection(caps_from_src, caps_from_dst)
set-migration-capabilities(src_qmp_session, common_caps);
set-migration-capabilities(dst_qmp_session, common_caps);
Then libvirt doesn't special code to handle XBRLE or whatever the next new
migration protocol feature is. With the current patches, libvirt would need to
create a table of:
migration_features = {'xblre': '-x'}
Which would change every time we add a feature. That seems pretty undesirable
to me.
Regards,
Anthony Liguori
Later, Juan
