From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>
Subject: [Qemu-devel] [PULL v2 47/50] vhost: Huge page align and merge
Date: Tue, 20 Mar 2018 05:18:00 +0200 [thread overview]
Message-ID: <1521515720-612046-48-git-send-email-mst@redhat.com> (raw)
In-Reply-To: <1521515720-612046-1-git-send-email-mst@redhat.com>
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Align RAMBlocks to page size alignment, and adjust the merging code
to deal with partial overlap due to that alignment.
This is needed for postcopy so that we can place/fetch whole hugepages
when under userfault.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/virtio/vhost.c | 66 ++++++++++++++++++++++++++++++++++++++++++--------
hw/virtio/trace-events | 3 ++-
2 files changed, 58 insertions(+), 11 deletions(-)
diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
index d8d0ef9..250f886 100644
--- a/hw/virtio/vhost.c
+++ b/hw/virtio/vhost.c
@@ -522,10 +522,28 @@ static void vhost_region_add_section(struct vhost_dev *dev,
uint64_t mrs_gpa = section->offset_within_address_space;
uintptr_t mrs_host = (uintptr_t)memory_region_get_ram_ptr(section->mr) +
section->offset_within_region;
+ RAMBlock *mrs_rb = section->mr->ram_block;
+ size_t mrs_page = qemu_ram_pagesize(mrs_rb);
trace_vhost_region_add_section(section->mr->name, mrs_gpa, mrs_size,
mrs_host);
+ /* Round the section to it's page size */
+ /* First align the start down to a page boundary */
+ uint64_t alignage = mrs_host & (mrs_page - 1);
+ if (alignage) {
+ mrs_host -= alignage;
+ mrs_size += alignage;
+ mrs_gpa -= alignage;
+ }
+ /* Now align the size up to a page boundary */
+ alignage = mrs_size & (mrs_page - 1);
+ if (alignage) {
+ mrs_size += mrs_page - alignage;
+ }
+ trace_vhost_region_add_section_aligned(section->mr->name, mrs_gpa, mrs_size,
+ mrs_host);
+
if (dev->n_tmp_sections) {
/* Since we already have at least one section, lets see if
* this extends it; since we're scanning in order, we only
@@ -542,18 +560,46 @@ static void vhost_region_add_section(struct vhost_dev *dev,
prev_sec->offset_within_region;
uint64_t prev_host_end = range_get_last(prev_host_start, prev_size);
- if (prev_gpa_end + 1 == mrs_gpa &&
- prev_host_end + 1 == mrs_host &&
- section->mr == prev_sec->mr &&
- (!dev->vhost_ops->vhost_backend_can_merge ||
- dev->vhost_ops->vhost_backend_can_merge(dev,
+ if (mrs_gpa <= (prev_gpa_end + 1)) {
+ /* OK, looks like overlapping/intersecting - it's possible that
+ * the rounding to page sizes has made them overlap, but they should
+ * match up in the same RAMBlock if they do.
+ */
+ if (mrs_gpa < prev_gpa_start) {
+ error_report("%s:Section rounded to %"PRIx64
+ " prior to previous %"PRIx64,
+ __func__, mrs_gpa, prev_gpa_start);
+ /* A way to cleanly fail here would be better */
+ return;
+ }
+ /* Offset from the start of the previous GPA to this GPA */
+ size_t offset = mrs_gpa - prev_gpa_start;
+
+ if (prev_host_start + offset == mrs_host &&
+ section->mr == prev_sec->mr &&
+ (!dev->vhost_ops->vhost_backend_can_merge ||
+ dev->vhost_ops->vhost_backend_can_merge(dev,
mrs_host, mrs_size,
prev_host_start, prev_size))) {
- /* The two sections abut */
- need_add = false;
- prev_sec->size = int128_add(prev_sec->size, section->size);
- trace_vhost_region_add_section_abut(section->mr->name,
- mrs_size + prev_size);
+ uint64_t max_end = MAX(prev_host_end, mrs_host + mrs_size);
+ need_add = false;
+ prev_sec->offset_within_address_space =
+ MIN(prev_gpa_start, mrs_gpa);
+ prev_sec->offset_within_region =
+ MIN(prev_host_start, mrs_host) -
+ (uintptr_t)memory_region_get_ram_ptr(prev_sec->mr);
+ prev_sec->size = int128_make64(max_end - MIN(prev_host_start,
+ mrs_host));
+ trace_vhost_region_add_section_merge(section->mr->name,
+ int128_get64(prev_sec->size),
+ prev_sec->offset_within_address_space,
+ prev_sec->offset_within_region);
+ } else {
+ error_report("%s: Overlapping but not coherent sections "
+ "at %"PRIx64,
+ __func__, mrs_gpa);
+ return;
+ }
}
}
diff --git a/hw/virtio/trace-events b/hw/virtio/trace-events
index 857c495..1422ff0 100644
--- a/hw/virtio/trace-events
+++ b/hw/virtio/trace-events
@@ -3,7 +3,8 @@
# hw/virtio/vhost.c
vhost_commit(bool started, bool changed) "Started: %d Changed: %d"
vhost_region_add_section(const char *name, uint64_t gpa, uint64_t size, uint64_t host) "%s: 0x%"PRIx64"+0x%"PRIx64" @ 0x%"PRIx64
-vhost_region_add_section_abut(const char *name, uint64_t new_size) "%s: 0x%"PRIx64
+vhost_region_add_section_merge(const char *name, uint64_t new_size, uint64_t gpa, uint64_t owr) "%s: size: 0x%"PRIx64 " gpa: 0x%"PRIx64 " owr: 0x%"PRIx64
+vhost_region_add_section_aligned(const char *name, uint64_t gpa, uint64_t size, uint64_t host) "%s: 0x%"PRIx64"+0x%"PRIx64" @ 0x%"PRIx64
vhost_section(const char *name, int r) "%s:%d"
# hw/virtio/vhost-user.c
--
MST
next prev parent reply other threads:[~2018-03-20 3:18 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-20 3:16 [Qemu-devel] [PULL v2 00/50] virtio, vhost, pci, pc: features, cleanups Michael S. Tsirkin
2018-03-20 3:16 ` [Qemu-devel] [PULL v2 01/50] scripts/update-linux-headers: add ethtool.h and update to 4.16.0-rc4 Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 03/50] virtio-net: add linkspeed and duplex settings to virtio-net Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 02/50] virtio-net: use 64-bit values for feature flags Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 04/50] acpi: remove unused acpi-dsdt.aml Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 05/50] pc: replace pm object initialization with one-liner in acpi_get_pm_info() Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 07/50] acpi: add build_append_gas() helper for Generic Address Structure Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 06/50] acpi: reuse AcpiGenericAddress instead of Acpi20GenericAddress Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 08/50] acpi: move ACPI_PORT_SMI_CMD define to header it belongs to Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 09/50] pc: acpi: isolate FADT specific data into AcpiFadtData structure Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 10/50] pc: acpi: use build_append_foo() API to construct FADT Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 11/50] acpi: move build_fadt() from i386 specific to generic ACPI source Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 12/50] virt_arm: acpi: reuse common build_fadt() Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 13/50] tests: acpi: don't read all fields in test_acpi_fadt_table() Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 15/50] hw/pci: remove obsolete PCIDevice->init() Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 14/50] standard-headers: update virtio_net.h Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 16/50] pc-dimm: make qmp_pc_dimm_device_list() sort devices by address Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 17/50] qmp: distinguish PC-DIMM and NVDIMM in MemoryDeviceInfoList Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 18/50] hw/acpi-build: build SRAT memory affinity structures for DIMM devices Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 19/50] tests/bios-tables-test: add test cases for DIMM proximity Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 20/50] test/acpi-test-data: add ACPI tables for dimmpxm test Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 21/50] Makefile: add target to print generated files Michael S. Tsirkin
2018-04-13 7:21 ` Markus Armbruster
2018-04-13 10:04 ` Marc-André Lureau
2018-04-13 12:51 ` Michael S. Tsirkin
2018-05-04 5:44 ` Markus Armbruster
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 22/50] migrate: Update ram_block_discard_range for shared Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 23/50] qemu_ram_block_host_offset Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 24/50] postcopy: use UFFDIO_ZEROPAGE only when available Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 25/50] postcopy: Add notifier chain Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 26/50] postcopy: Add vhost-user flag for postcopy and check it Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 27/50] vhost-user: Add 'VHOST_USER_POSTCOPY_ADVISE' message Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 28/50] libvhost-user: Support sending fds back to qemu Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 29/50] libvhost-user: Open userfaultfd Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 30/50] postcopy: Allow registering of fd handler Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 31/50] vhost+postcopy: Register shared ufd with postcopy Michael S. Tsirkin
2018-04-27 16:12 ` Peter Maydell
2018-05-02 10:58 ` Dr. David Alan Gilbert
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 32/50] vhost+postcopy: Transmit 'listen' to slave Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 33/50] postcopy+vhost-user: Split set_mem_table for postcopy Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 34/50] migration/ram: ramblock_recv_bitmap_test_byte_offset Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 35/50] libvhost-user+postcopy: Register new regions with the ufd Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 37/50] vhost+postcopy: Stash RAMBlock and offset Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 36/50] vhost+postcopy: Send address back to qemu Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 38/50] vhost+postcopy: Helper to send requests to source for shared pages Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 39/50] vhost+postcopy: Resolve client address Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 40/50] postcopy: helper for waking shared Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 41/50] postcopy: postcopy_notify_shared_wake Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 44/50] libvhost-user: mprotect & madvises for postcopy Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 42/50] vhost+postcopy: Add vhost waker Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 43/50] vhost+postcopy: Call wakeups Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 46/50] vhost+postcopy: Wire up POSTCOPY_END notify Michael S. Tsirkin
2018-03-20 3:17 ` [Qemu-devel] [PULL v2 45/50] vhost-user: Add VHOST_USER_POSTCOPY_END message Michael S. Tsirkin
2018-03-20 3:18 ` Michael S. Tsirkin [this message]
2018-03-20 3:18 ` [Qemu-devel] [PULL v2 48/50] postcopy: Allow shared memory Michael S. Tsirkin
2018-03-20 3:18 ` [Qemu-devel] [PULL v2 49/50] libvhost-user: Claim support for postcopy Michael S. Tsirkin
2018-03-20 3:18 ` [Qemu-devel] [PULL v2 50/50] postcopy shared docs Michael S. Tsirkin
2018-03-20 14:18 ` [Qemu-devel] [PULL v2 00/50] virtio, vhost, pci, pc: features, cleanups Peter Maydell
2018-03-20 14:37 ` Michael S. Tsirkin
2018-03-20 15:05 ` Michael S. Tsirkin
2018-03-20 15:41 ` Peter Maydell
2018-03-20 15:51 ` Michael S. Tsirkin
2018-03-20 15:54 ` Peter Maydell
2018-03-20 16:02 ` Michael S. Tsirkin
2018-03-20 17:18 ` Peter Maydell
2018-03-20 17:22 ` Michael S. Tsirkin
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=1521515720-612046-48-git-send-email-mst@redhat.com \
--to=mst@redhat.com \
--cc=dgilbert@redhat.com \
--cc=peter.maydell@linaro.org \
--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).