From: Juan Quintela <quintela@redhat.com>
To: qemu-devel@nongnu.org
Cc: amit.shah@redhat.com, dgilbert@redhat.com
Subject: [Qemu-devel] [PULL 43/57] postcopy_ram.c: place_page and helpers
Date: Tue, 10 Nov 2015 15:25:32 +0100 [thread overview]
Message-ID: <1447165546-27784-44-git-send-email-quintela@redhat.com> (raw)
In-Reply-To: <1447165546-27784-1-git-send-email-quintela@redhat.com>
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
postcopy_place_page (etc) provide a way for postcopy to place a page
into guests memory atomically (using the copy ioctl on the ufd).
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
---
include/migration/migration.h | 1 +
include/migration/postcopy-ram.h | 21 ++++++++
migration/postcopy-ram.c | 100 +++++++++++++++++++++++++++++++++++++++
trace-events | 2 +
4 files changed, 124 insertions(+)
diff --git a/include/migration/migration.h b/include/migration/migration.h
index 1491bf3..a48471e 100644
--- a/include/migration/migration.h
+++ b/include/migration/migration.h
@@ -96,6 +96,7 @@ struct MigrationIncomingState {
int userfault_fd;
QEMUFile *to_src_file;
QemuMutex rp_mutex; /* We send replies from multiple threads */
+ void *postcopy_tmp_page;
/* See savevm.c */
LoadStateEntry_Head loadvm_handlers;
diff --git a/include/migration/postcopy-ram.h b/include/migration/postcopy-ram.h
index b10c03d..d7c292f 100644
--- a/include/migration/postcopy-ram.h
+++ b/include/migration/postcopy-ram.h
@@ -69,4 +69,25 @@ void postcopy_discard_send_range(MigrationState *ms, PostcopyDiscardState *pds,
void postcopy_discard_send_finish(MigrationState *ms,
PostcopyDiscardState *pds);
+/*
+ * Place a page (from) at (host) efficiently
+ * There are restrictions on how 'from' must be mapped, in general best
+ * to use other postcopy_ routines to allocate.
+ * returns 0 on success
+ */
+int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from);
+
+/*
+ * Place a zero page at (host) atomically
+ * returns 0 on success
+ */
+int postcopy_place_page_zero(MigrationIncomingState *mis, void *host);
+
+/*
+ * Allocate a page of memory that can be mapped at a later point in time
+ * using postcopy_place_page
+ * Returns: Pointer to allocated page
+ */
+void *postcopy_get_tmp_page(MigrationIncomingState *mis);
+
#endif
diff --git a/migration/postcopy-ram.c b/migration/postcopy-ram.c
index 3110b2a..58492c0 100644
--- a/migration/postcopy-ram.c
+++ b/migration/postcopy-ram.c
@@ -272,6 +272,10 @@ int postcopy_ram_incoming_cleanup(MigrationIncomingState *mis)
return -1;
}
+ if (mis->postcopy_tmp_page) {
+ munmap(mis->postcopy_tmp_page, getpagesize());
+ mis->postcopy_tmp_page = NULL;
+ }
return 0;
}
@@ -338,6 +342,83 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis)
return 0;
}
+/*
+ * Place a host page (from) at (host) atomically
+ * returns 0 on success
+ */
+int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from)
+{
+ struct uffdio_copy copy_struct;
+
+ copy_struct.dst = (uint64_t)(uintptr_t)host;
+ copy_struct.src = (uint64_t)(uintptr_t)from;
+ copy_struct.len = getpagesize();
+ copy_struct.mode = 0;
+
+ /* copy also acks to the kernel waking the stalled thread up
+ * TODO: We can inhibit that ack and only do it if it was requested
+ * which would be slightly cheaper, but we'd have to be careful
+ * of the order of updating our page state.
+ */
+ if (ioctl(mis->userfault_fd, UFFDIO_COPY, ©_struct)) {
+ int e = errno;
+ error_report("%s: %s copy host: %p from: %p",
+ __func__, strerror(e), host, from);
+
+ return -e;
+ }
+
+ trace_postcopy_place_page(host);
+ return 0;
+}
+
+/*
+ * Place a zero page at (host) atomically
+ * returns 0 on success
+ */
+int postcopy_place_page_zero(MigrationIncomingState *mis, void *host)
+{
+ struct uffdio_zeropage zero_struct;
+
+ zero_struct.range.start = (uint64_t)(uintptr_t)host;
+ zero_struct.range.len = getpagesize();
+ zero_struct.mode = 0;
+
+ if (ioctl(mis->userfault_fd, UFFDIO_ZEROPAGE, &zero_struct)) {
+ int e = errno;
+ error_report("%s: %s zero host: %p",
+ __func__, strerror(e), host);
+
+ return -e;
+ }
+
+ trace_postcopy_place_page_zero(host);
+ return 0;
+}
+
+/*
+ * Returns a target page of memory that can be mapped at a later point in time
+ * using postcopy_place_page
+ * The same address is used repeatedly, postcopy_place_page just takes the
+ * backing page away.
+ * Returns: Pointer to allocated page
+ *
+ */
+void *postcopy_get_tmp_page(MigrationIncomingState *mis)
+{
+ if (!mis->postcopy_tmp_page) {
+ mis->postcopy_tmp_page = mmap(NULL, getpagesize(),
+ PROT_READ | PROT_WRITE, MAP_PRIVATE |
+ MAP_ANONYMOUS, -1, 0);
+ if (!mis->postcopy_tmp_page) {
+ error_report("%s: %s", __func__, strerror(errno));
+ return NULL;
+ }
+ }
+
+ return mis->postcopy_tmp_page;
+}
+
#else
/* No target OS support, stubs just fail */
bool postcopy_ram_supported_by_host(void)
@@ -370,6 +451,25 @@ int postcopy_ram_enable_notify(MigrationIncomingState *mis)
assert(0);
return -1;
}
+
+int postcopy_place_page(MigrationIncomingState *mis, void *host, void *from)
+{
+ assert(0);
+ return -1;
+}
+
+int postcopy_place_page_zero(MigrationIncomingState *mis, void *host)
+{
+ assert(0);
+ return -1;
+}
+
+void *postcopy_get_tmp_page(MigrationIncomingState *mis)
+{
+ assert(0);
+ return NULL;
+}
+
#endif
/* ------------------------------------------------------------------------- */
diff --git a/trace-events b/trace-events
index 3df3656..c493f5d 100644
--- a/trace-events
+++ b/trace-events
@@ -1551,6 +1551,8 @@ postcopy_discard_send_range(const char *ramblock, unsigned long start, unsigned
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_place_page(void *host_addr) "host=%p"
+postcopy_place_page_zero(void *host_addr) "host=%p"
# kvm-all.c
kvm_ioctl(int type, void *arg) "type 0x%x, arg %p"
--
2.5.0
next prev parent reply other threads:[~2015-11-10 14:27 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-11-10 14:24 [Qemu-devel] [PULL 00/57] Migration pull (take 2) Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 01/57] Add postcopy documentation Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 02/57] Provide runtime Target page information Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 03/57] Move configuration section writing Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 04/57] Move page_size_init earlier Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 05/57] qemu_ram_block_from_host Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 06/57] qemu_ram_block_by_name Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 07/57] Rename mis->file to from_src_file Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 08/57] Add qemu_get_buffer_in_place to avoid copies some of the time Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 09/57] Add wrapper for setting blocking status on a QEMUFile Juan Quintela
2015-11-10 14:24 ` [Qemu-devel] [PULL 10/57] Add QEMU_MADV_NOHUGEPAGE Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 11/57] ram_debug_dump_bitmap: Dump a migration bitmap as text Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 12/57] ram_load: Factor out host_from_stream_offset call and check Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 13/57] migrate_init: Call from savevm Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 14/57] Rename save_live_complete to save_live_complete_precopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 15/57] Add Linux userfaultfd.h header Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 16/57] Return path: Open a return path on QEMUFile for sockets Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 17/57] Return path: socket_writev_buffer: Block even on non-blocking fd's Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 18/57] Migration commands Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 19/57] Return path: Control commands Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 20/57] Return path: Send responses from destination to source Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 21/57] migration_is_setup_or_active Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 22/57] Return path: Source handling of return path Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 23/57] Rework loadvm path for subloops Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 24/57] Add migration-capability boolean for postcopy-ram Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 25/57] Add wrappers and handlers for sending/receiving the postcopy-ram migration messages Juan Quintela
2015-11-14 19:11 ` Stefan Weil
2015-11-16 10:07 ` Dr. David Alan Gilbert
2015-11-16 13:14 ` Stefan Weil
2015-11-16 14:20 ` Dr. David Alan Gilbert
2015-11-10 14:25 ` [Qemu-devel] [PULL 26/57] MIG_CMD_PACKAGED: Send a packaged chunk of migration stream Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 27/57] Modify save_live_pending for postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 28/57] postcopy: OS support test Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 29/57] migrate_start_postcopy: Command to trigger transition to postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 30/57] migration_completion: Take current state Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 31/57] MIGRATION_STATUS_POSTCOPY_ACTIVE: Add new migration state Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 32/57] Avoid sending vmdescription during postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 33/57] Add qemu_savevm_state_complete_postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 34/57] Postcopy: Maintain unsentmap Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 35/57] migration_completion: Take current state Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 36/57] postcopy: Incoming initialisation Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 37/57] postcopy: ram_enable_notify to switch on userfault Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 38/57] Postcopy: Postcopy startup in migration thread Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 39/57] Postcopy: End of iteration Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 40/57] Page request: Add MIG_RP_MSG_REQ_PAGES reverse command Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 41/57] Page request: Process incoming page request Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 42/57] Page request: Consume pages off the post-copy queue Juan Quintela
2015-11-12 11:53 ` Peter Maydell
2015-11-12 12:04 ` Dr. David Alan Gilbert
2015-11-12 12:15 ` Peter Maydell
2015-11-12 12:23 ` Dr. David Alan Gilbert
2015-11-12 12:57 ` Peter Maydell
2015-11-12 13:08 ` Dr. David Alan Gilbert
2015-11-12 13:18 ` Peter Maydell
2015-11-12 13:53 ` Peter Maydell
2015-11-12 14:20 ` Dr. David Alan Gilbert
2015-11-12 15:25 ` Juan Quintela
2015-11-12 15:57 ` Dr. David Alan Gilbert
2015-11-12 14:22 ` Juan Quintela
2015-11-12 13:36 ` Markus Armbruster
2015-11-16 10:10 ` Andreas Färber
2015-11-10 14:25 ` Juan Quintela [this message]
2015-11-10 14:25 ` [Qemu-devel] [PULL 44/57] Postcopy: Use helpers to map pages during migration Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 45/57] postcopy: Check order of received target pages Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 46/57] Don't sync dirty bitmaps in postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 47/57] Don't iterate on precopy-only devices during postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 48/57] Host page!=target page: Cleanup bitmaps Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 49/57] Round up RAMBlock sizes to host page sizes Juan Quintela
2015-12-30 0:26 ` Peter Crosthwaite
2016-01-04 9:48 ` Dr. David Alan Gilbert
2016-01-06 21:27 ` Paolo Bonzini
2015-11-10 14:25 ` [Qemu-devel] [PULL 50/57] Postcopy; Handle userfault requests Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 51/57] Start up a postcopy/listener thread ready for incoming page data Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 52/57] postcopy: Wire up loadvm_postcopy_handle_ commands Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 53/57] Postcopy: Mark nohugepage before discard Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 54/57] End of migration for postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 55/57] Disable mlock around incoming postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 56/57] Inhibit ballooning during postcopy Juan Quintela
2015-11-10 14:25 ` [Qemu-devel] [PULL 57/57] migration: qemu_savevm_state_cleanup becomes mandatory operation Juan Quintela
2015-11-10 22:21 ` [Qemu-devel] [PULL 00/57] Migration pull (take 2) Peter Maydell
-- strict thread matches above, loose matches on Subject: below --
2015-11-09 17:28 [Qemu-devel] [PULL 00/57] Migration pull Juan Quintela
2015-11-09 17:28 ` [Qemu-devel] [PULL 43/57] postcopy_ram.c: place_page and helpers Juan Quintela
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=1447165546-27784-44-git-send-email-quintela@redhat.com \
--to=quintela@redhat.com \
--cc=amit.shah@redhat.com \
--cc=dgilbert@redhat.com \
--cc=qemu-devel@nongnu.org \
/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).