[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH v10 10/14, resend] rdma: introduce capability x-rdma
From: |
mrhines |
Subject: |
[Qemu-devel] [PATCH v10 10/14, resend] rdma: introduce capability x-rdma-pin-all |
Date: |
Mon, 24 Jun 2013 09:46:26 -0400 |
From: "Michael R. Hines" <address@hidden>
This capability allows you to disable dynamic chunk registration
for better throughput on high-performance links.
For example, using an 8GB RAM virtual machine with all 8GB of memory in
active use and the VM itself is completely idle using a 40 gbps infiniband link:
1. x-rdma-pin-all disabled total time: approximately 7.5 seconds @ 9.5 Gbps
2. x-rdma-pin-all enabled total time: approximately 4 seconds @ 26 Gbps
These numbers would of course scale up to whatever size virtual machine
you have to migrate using RDMA.
Enabling this feature does *not* have any measurable affect on
migration *downtime*. This is because, without this feature, all of the
memory will have already been registered already in advance during
the bulk round and does not need to be re-registered during the successive
iteration rounds.
Reviewed-by: Paolo Bonzini <address@hidden>
Reviewed-by: Chegu Vinod <address@hidden>
Reviewed-by: Eric Blake <address@hidden>
Tested-by: Chegu Vinod <address@hidden>
Tested-by: Michael R. Hines <address@hidden>
Signed-off-by: Michael R. Hines <address@hidden>
---
include/migration/migration.h | 2 ++
migration.c | 9 +++++++++
qapi-schema.json | 7 ++++++-
3 files changed, 17 insertions(+), 1 deletion(-)
diff --git a/include/migration/migration.h b/include/migration/migration.h
index d60c64c..2e6001d 100644
--- a/include/migration/migration.h
+++ b/include/migration/migration.h
@@ -125,6 +125,8 @@ void migrate_add_blocker(Error *reason);
*/
void migrate_del_blocker(Error *reason);
+bool migrate_rdma_pin_all(void);
+
int xbzrle_encode_buffer(uint8_t *old_buf, uint8_t *new_buf, int slen,
uint8_t *dst, int dlen);
int xbzrle_decode_buffer(uint8_t *src, int slen, uint8_t *dst, int dlen);
diff --git a/migration.c b/migration.c
index 441a3b2..a704d48 100644
--- a/migration.c
+++ b/migration.c
@@ -476,6 +476,15 @@ void qmp_migrate_set_downtime(double value, Error **errp)
max_downtime = (uint64_t)value;
}
+bool migrate_rdma_pin_all(void)
+{
+ MigrationState *s;
+
+ s = migrate_get_current();
+
+ return s->enabled_capabilities[MIGRATION_CAPABILITY_X_RDMA_PIN_ALL];
+}
+
int migrate_use_xbzrle(void)
{
MigrationState *s;
diff --git a/qapi-schema.json b/qapi-schema.json
index b239adb..4e9cd6d 100644
--- a/qapi-schema.json
+++ b/qapi-schema.json
@@ -608,10 +608,15 @@
# This feature allows us to minimize migration traffic for certain
work
# loads, by sending compressed difference of the pages
#
+# @x-rdma-pin-all: Controls whether or not the entire VM memory footprint is
+# mlock()'d on demand or all at once. Refer to docs/rdma.txt for
usage.
+# Disabled by default. Experimental: may (or may not) be renamed after
+# further testing is complete. (since 1.6)
+#
# Since: 1.2
##
{ 'enum': 'MigrationCapability',
- 'data': ['xbzrle'] }
+ 'data': ['xbzrle', 'x-rdma-pin-all'] }
##
# @MigrationCapabilityStatus
--
1.7.10.4
- [Qemu-devel] [PATCH v10 08/14, resend] rdma: introduce qemu_ram_foreach_block(), (continued)
[Qemu-devel] [PATCH v10 11/14, resend] rdma: core logic, mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 10/14, resend] rdma: introduce capability x-rdma-pin-all,
mrhines <=
[Qemu-devel] [PATCH v10 03/14, resend] rdma: export yield_until_fd_readable(), mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 04/14, resend] rdma: export throughput w/ MigrationStats QMP, mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 06/14, resend] rdma: export qemu_fflush(), mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 13/14, resend] rdma: fix mlock() freezes and accounting, mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 09/14, resend] rdma: new QEMUFileOps hooks, mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 05/14, resend] rdma: introduce qemu_file_mode_is_not_valid(), mrhines, 2013/06/24
[Qemu-devel] [PATCH v10 02/14, resend] rdma: introduce qemu_update_position(), mrhines, 2013/06/24