From: "Dr. David Alan Gilbert (git)" <dgilbert@redhat.com>
To: qemu-devel@nongnu.org, quintela@redhat.com, amit.shah@redhat.com
Cc: aarcange@redhat.com, pbonzini@redhat.com, liang.z.li@intel.com,
luis@cs.umu.se, bharata@linux.vnet.ibm.com
Subject: [Qemu-devel] [PATCH v8 51/54] Postcopy: Mark nohugepage before discard
Date: Tue, 29 Sep 2015 09:38:15 +0100 [thread overview]
Message-ID: <1443515898-3594-52-git-send-email-dgilbert@redhat.com> (raw)
In-Reply-To: <1443515898-3594-1-git-send-email-dgilbert@redhat.com>
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Prior to servicing userfault requests we must ensure we've not got
huge pages in the area that might include non-transferred memory,
since a hugepage could incorrectly mark the whole huge page as present.
We mark the area as non-huge page (nhp) just before we perform
discards; the discard code now tells us to discard any areas
that haven't been sent (as well as any that are redirtied);
any already formed transparent-huge-pages get fragmented
by this discard process if they cotnain any discards.
Transparent huge pages that have been entirely transferred
and don't contain any discards are not broken by this mechanism;
they stay as huge pages.
By starting postcopy after a full precopy pass, many of the pages
then stay as huge pages; this is important for maintaining performance
after the end of the migration.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
include/migration/postcopy-ram.h | 6 ++++++
migration/postcopy-ram.c | 46 +++++++++++++++++++++++++++++++++++++---
migration/savevm.c | 9 +++++++-
trace-events | 1 +
4 files changed, 58 insertions(+), 4 deletions(-)
diff --git a/include/migration/postcopy-ram.h b/include/migration/postcopy-ram.h
index 50c1ce5..60ac5b1 100644
--- a/include/migration/postcopy-ram.h
+++ b/include/migration/postcopy-ram.h
@@ -41,6 +41,12 @@ int postcopy_ram_incoming_cleanup(MigrationIncomingState *mis);
int postcopy_ram_discard_range(MigrationIncomingState *mis, uint8_t *start,
size_t length);
+/*
+ * Userfault requires us to mark RAM as NOHUGEPAGE prior to discard
+ * however leaving it until after precopy means that most of the precopy
+ * data is still THPd
+ */
+int postcopy_ram_prepare_discard(MigrationIncomingState *mis);
/*
* Called at the start of each RAMBlock by the bitmap code.
diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c
index 0b021ca..10e5a5b 100644
--- a/migration/postcopy-ram.c
+++ b/migration/postcopy-ram.c
@@ -226,12 +226,10 @@ static int cleanup_range(const char *block_name, void *host_addr,
* We turned off hugepage for the precopy stage with postcopy enabled
* we can turn it back on now.
*/
-#ifdef MADV_HUGEPAGE
- if (madvise(host_addr, length, MADV_HUGEPAGE)) {
+ if (qemu_madvise(host_addr, length, QEMU_MADV_HUGEPAGE)) {
error_report("%s HUGEPAGE: %s", __func__, strerror(errno));
return -1;
}
-#endif
/*
* We can also turn off userfault now since we should have all the
@@ -308,6 +306,43 @@ int postcopy_ram_incoming_cleanup(MigrationIncomingState *mis)
}
/*
+ * Disable huge pages on an area
+ */
+static int nhp_range(const char *block_name, void *host_addr,
+ ram_addr_t offset, ram_addr_t length, void *opaque)
+{
+ trace_postcopy_nhp_range(block_name, host_addr, offset, length);
+
+ /*
+ * Before we do discards we need to ensure those discards really
+ * do delete areas of the page, even if THP thinks a hugepage would
+ * be a good idea, so force hugepages off.
+ */
+ if (qemu_madvise(host_addr, length, QEMU_MADV_NOHUGEPAGE)) {
+ error_report("%s: NOHUGEPAGE: %s", __func__, strerror(errno));
+ return -1;
+ }
+
+ return 0;
+}
+
+/*
+ * Userfault requires us to mark RAM as NOHUGEPAGE prior to discard
+ * however leaving it until after precopy means that most of the precopy
+ * data is still THPd
+ */
+int postcopy_ram_prepare_discard(MigrationIncomingState *mis)
+{
+ if (qemu_ram_foreach_block(nhp_range, mis)) {
+ return -1;
+ }
+
+ postcopy_state_set(POSTCOPY_INCOMING_DISCARD);
+
+ return 0;
+}
+
+/*
* Mark the given area of RAM as requiring notification to unwritten areas
* Used as a callback on qemu_ram_foreach_block.
* host_addr: Base of area to mark
@@ -581,6 +616,11 @@ int postcopy_ram_discard_range(MigrationIncomingState *mis, uint8_t *start,
assert(0);
}
+int postcopy_ram_prepare_discard(MigrationIncomingState *mis)
+{
+ assert(0);
+}
+
int postcopy_ram_enable_notify(MigrationIncomingState *mis)
{
assert(0);
diff --git a/migration/savevm.c b/migration/savevm.c
index 7c5b9d1..894e085 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1314,7 +1314,7 @@ static int loadvm_postcopy_ram_handle_discard(MigrationIncomingState *mis,
switch (ps) {
case POSTCOPY_INCOMING_ADVISE:
/* 1st discard */
- tmp = 0; /* TODO: later patch postcopy_ram_prepare_discard(mis); */
+ tmp = postcopy_ram_prepare_discard(mis);
if (tmp) {
return tmp;
}
@@ -1446,6 +1446,13 @@ static int loadvm_postcopy_handle_listen(MigrationIncomingState *mis)
error_report("CMD_POSTCOPY_LISTEN in wrong postcopy state (%d)", ps);
return -1;
}
+ if (ps == POSTCOPY_INCOMING_ADVISE) {
+ /*
+ * A rare case, we entered listen without having to do any discards,
+ * so do the setup that's normally done at the time of the 1st discard.
+ */
+ postcopy_ram_prepare_discard(mis);
+ }
/*
* Sensitise RAM - can now generate requests for blocks that don't exist
diff --git a/trace-events b/trace-events
index 42a7577..81f9ca7 100644
--- a/trace-events
+++ b/trace-events
@@ -1542,6 +1542,7 @@ postcopy_discard_send_finish(const char *ramblock, int nwords, int ncmds) "%s ma
postcopy_ram_discard_range(void *start, size_t length) "%p,+%zx"
postcopy_cleanup_range(const char *ramblock, void *host_addr, size_t offset, size_t length) "%s: %p offset=%zx length=%zx"
postcopy_init_range(const char *ramblock, void *host_addr, size_t offset, size_t length) "%s: %p offset=%zx length=%zx"
+postcopy_nhp_range(const char *ramblock, void *host_addr, size_t offset, size_t length) "%s: %p offset=%zx length=%zx"
postcopy_place_page(void *host_addr) "host=%p"
postcopy_place_page_zero(void *host_addr) "host=%p"
postcopy_ram_enable_notify(void) ""
--
2.5.0
next prev parent reply other threads:[~2015-09-29 8:40 UTC|newest]
Thread overview: 118+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-29 8:37 [Qemu-devel] [PATCH v8 00/54] Postcopy implementation Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 01/54] Add postcopy documentation Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 02/54] Provide runtime Target page information Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 03/54] Init page sizes in qtest Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 04/54] Move configuration section writing Dr. David Alan Gilbert (git)
2015-10-05 6:44 ` Amit Shah
2015-10-30 12:47 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 05/54] qemu_ram_block_from_host Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 06/54] Rename mis->file to from_src_file Dr. David Alan Gilbert (git)
2015-09-29 10:41 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 07/54] Add qemu_get_buffer_in_place to avoid copies some of the time Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 08/54] Add wrapper for setting blocking status on a QEMUFile Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 09/54] Add QEMU_MADV_NOHUGEPAGE Dr. David Alan Gilbert (git)
2015-10-28 10:35 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 10/54] migration/ram.c: Use RAMBlock rather than MemoryRegion Dr. David Alan Gilbert (git)
2015-10-28 10:36 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 11/54] ram_debug_dump_bitmap: Dump a migration bitmap as text Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 12/54] migrate_init: Call from savevm Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 13/54] Move dirty page search state into separate structure Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 14/54] ram_find_and_save_block: Split out the finding Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 15/54] Rename save_live_complete to save_live_complete_precopy Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 16/54] Return path: Open a return path on QEMUFile for sockets Dr. David Alan Gilbert (git)
2015-10-02 15:29 ` Daniel P. Berrange
2015-10-02 16:32 ` Dr. David Alan Gilbert
2015-10-02 17:03 ` Daniel P. Berrange
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 17/54] Return path: socket_writev_buffer: Block even on non-blocking fd's Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 18/54] Migration commands Dr. David Alan Gilbert (git)
2015-10-20 11:22 ` Juan Quintela
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 19/54] Return path: Control commands Dr. David Alan Gilbert (git)
2015-10-20 11:27 ` Juan Quintela
2015-10-26 11:42 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 20/54] Return path: Send responses from destination to source Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 21/54] Return path: Source handling of return path Dr. David Alan Gilbert (git)
2015-10-20 11:33 ` Juan Quintela
2015-10-26 12:06 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 22/54] Rework loadvm path for subloops Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 23/54] Add migration-capability boolean for postcopy-ram Dr. David Alan Gilbert (git)
2015-09-29 20:22 ` Eric Blake
2015-09-30 7:00 ` Amit Shah
2015-09-30 12:44 ` Eric Blake
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 24/54] Add wrappers and handlers for sending/receiving the postcopy-ram migration messages Dr. David Alan Gilbert (git)
2015-10-20 11:50 ` Juan Quintela
2015-10-26 12:22 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 25/54] MIG_CMD_PACKAGED: Send a packaged chunk of migration stream Dr. David Alan Gilbert (git)
2015-10-20 13:25 ` Juan Quintela
2015-10-26 16:21 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 26/54] Modify save_live_pending for postcopy Dr. David Alan Gilbert (git)
2015-10-28 11:03 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 27/54] postcopy: OS support test Dr. David Alan Gilbert (git)
2015-10-20 13:31 ` Juan Quintela
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 28/54] migrate_start_postcopy: Command to trigger transition to postcopy Dr. David Alan Gilbert (git)
2015-09-30 16:25 ` Eric Blake
2015-09-30 16:30 ` Dr. David Alan Gilbert
2015-10-20 13:33 ` Juan Quintela
2015-10-28 11:17 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 29/54] MIGRATION_STATUS_POSTCOPY_ACTIVE: Add new migration state Dr. David Alan Gilbert (git)
2015-10-20 13:35 ` Juan Quintela
2015-10-30 18:19 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 30/54] Avoid sending vmdescription during postcopy Dr. David Alan Gilbert (git)
2015-10-20 13:35 ` Juan Quintela
2015-10-28 11:19 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 31/54] Add qemu_savevm_state_complete_postcopy Dr. David Alan Gilbert (git)
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 32/54] Postcopy: Maintain sentmap and calculate discard Dr. David Alan Gilbert (git)
2015-10-21 11:17 ` Juan Quintela
2015-10-30 18:43 ` Dr. David Alan Gilbert
2015-11-02 17:31 ` Dr. David Alan Gilbert
2015-11-02 18:19 ` Dr. David Alan Gilbert
2015-11-02 20:14 ` Dr. David Alan Gilbert
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 33/54] postcopy: Incoming initialisation Dr. David Alan Gilbert (git)
2015-10-21 8:35 ` Juan Quintela
2015-11-03 17:59 ` Dr. David Alan Gilbert
2015-11-03 18:32 ` Juan Quintela
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 34/54] postcopy: ram_enable_notify to switch on userfault Dr. David Alan Gilbert (git)
2015-10-28 11:40 ` Amit Shah
2015-09-29 8:37 ` [Qemu-devel] [PATCH v8 35/54] Postcopy: Postcopy startup in migration thread Dr. David Alan Gilbert (git)
2015-10-21 8:57 ` Juan Quintela
2015-10-26 17:12 ` Dr. David Alan Gilbert
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 36/54] Split out end of migration code from migration_thread Dr. David Alan Gilbert (git)
2015-10-21 9:11 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 37/54] Postcopy: End of iteration Dr. David Alan Gilbert (git)
2015-10-21 9:16 ` Juan Quintela
2015-10-29 5:10 ` Amit Shah
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 38/54] Page request: Add MIG_RP_MSG_REQ_PAGES reverse command Dr. David Alan Gilbert (git)
2015-10-21 11:12 ` Juan Quintela
2015-10-26 16:58 ` Dr. David Alan Gilbert
2015-10-29 5:17 ` Amit Shah
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 39/54] Page request: Process incoming page request Dr. David Alan Gilbert (git)
2015-10-21 11:17 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 40/54] Page request: Consume pages off the post-copy queue Dr. David Alan Gilbert (git)
2015-10-26 16:32 ` Juan Quintela
2015-11-03 11:52 ` Dr. David Alan Gilbert
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 41/54] postcopy_ram.c: place_page and helpers Dr. David Alan Gilbert (git)
2015-10-28 10:28 ` Juan Quintela
2015-10-28 13:11 ` Dr. David Alan Gilbert
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 42/54] Postcopy: Use helpers to map pages during migration Dr. David Alan Gilbert (git)
2015-10-28 10:58 ` Juan Quintela
2015-10-30 12:59 ` Dr. David Alan Gilbert
2015-10-30 16:35 ` Dr. David Alan Gilbert
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 43/54] Don't sync dirty bitmaps in postcopy Dr. David Alan Gilbert (git)
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 44/54] Don't iterate on precopy-only devices during postcopy Dr. David Alan Gilbert (git)
2015-10-28 11:01 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 45/54] Host page!=target page: Cleanup bitmaps Dr. David Alan Gilbert (git)
2015-10-28 11:24 ` Juan Quintela
2015-11-03 17:32 ` Dr. David Alan Gilbert
2015-11-03 18:30 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 46/54] postcopy: Check order of received target pages Dr. David Alan Gilbert (git)
2015-10-28 11:26 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 47/54] Round up RAMBlock sizes to host page sizes Dr. David Alan Gilbert (git)
2015-10-28 11:28 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 48/54] Postcopy; Handle userfault requests Dr. David Alan Gilbert (git)
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 49/54] Start up a postcopy/listener thread ready for incoming page data Dr. David Alan Gilbert (git)
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 50/54] postcopy: Wire up loadvm_postcopy_handle_ commands Dr. David Alan Gilbert (git)
2015-09-29 8:38 ` Dr. David Alan Gilbert (git) [this message]
2015-10-28 14:02 ` [Qemu-devel] [PATCH v8 51/54] Postcopy: Mark nohugepage before discard Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 52/54] End of migration for postcopy Dr. David Alan Gilbert (git)
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 53/54] Disable mlock around incoming postcopy Dr. David Alan Gilbert (git)
2015-10-21 9:17 ` Juan Quintela
2015-09-29 8:38 ` [Qemu-devel] [PATCH v8 54/54] Inhibit ballooning during postcopy Dr. David Alan Gilbert (git)
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1443515898-3594-52-git-send-email-dgilbert@redhat.com \
--to=dgilbert@redhat.com \
--cc=aarcange@redhat.com \
--cc=amit.shah@redhat.com \
--cc=bharata@linux.vnet.ibm.com \
--cc=liang.z.li@intel.com \
--cc=luis@cs.umu.se \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).