From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Peter Maydell" <peter.maydell@linaro.org>,
"Jason Wang" <jasowang@redhat.com>,
"Eugenio Pérez" <eperezma@redhat.com>
Subject: [PULL 01/35] vhost-vdpa: Do not send empty IOTLB update batches
Date: Sat, 4 Sep 2021 17:35:56 -0400 [thread overview]
Message-ID: <20210904213506.486886-2-mst@redhat.com> (raw)
In-Reply-To: <20210904213506.486886-1-mst@redhat.com>
From: Eugenio Pérez <eperezma@redhat.com>
With the introduction of the batch hinting, meaningless batches can be
created with no IOTLB updates if the memory region was skipped by
vhost_vdpa_listener_skipped_section. This is the case of host notifiers
memory regions, device un/realize, and others. This causes the vdpa
device to receive dma mapping settings with no changes, a possibly
expensive operation for nothing.
To avoid that, VHOST_IOTLB_BATCH_BEGIN hint is delayed until we have a
meaningful (not skipped section) mapping or unmapping operation, and
VHOST_IOTLB_BATCH_END is not written unless at least one of _UPDATE /
_INVALIDATE has been issued.
v3:
* Use a bool instead of a counter avoiding potential number wrapping
* Fix bad check on _commit
* Move VHOST_BACKEND_F_IOTLB_BATCH check to
vhost_vdpa_iotlb_batch_begin_once
v2 (from RFC):
* Rename misleading name
* Abstract start batching function for listener_add/del
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20210812140933.226288-1-eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
include/hw/virtio/vhost-vdpa.h | 1 +
hw/virtio/vhost-vdpa.c | 35 ++++++++++++++++++++++------------
2 files changed, 24 insertions(+), 12 deletions(-)
diff --git a/include/hw/virtio/vhost-vdpa.h b/include/hw/virtio/vhost-vdpa.h
index 9188226d8b..a8963da2d9 100644
--- a/include/hw/virtio/vhost-vdpa.h
+++ b/include/hw/virtio/vhost-vdpa.h
@@ -22,6 +22,7 @@ typedef struct VhostVDPAHostNotifier {
typedef struct vhost_vdpa {
int device_fd;
uint32_t msg_type;
+ bool iotlb_batch_begin_sent;
MemoryListener listener;
struct vhost_dev *dev;
VhostVDPAHostNotifier notifier[VIRTIO_QUEUE_MAX];
diff --git a/hw/virtio/vhost-vdpa.c b/hw/virtio/vhost-vdpa.c
index 4fa414feea..ca1227e5dc 100644
--- a/hw/virtio/vhost-vdpa.c
+++ b/hw/virtio/vhost-vdpa.c
@@ -89,19 +89,13 @@ static int vhost_vdpa_dma_unmap(struct vhost_vdpa *v, hwaddr iova,
return ret;
}
-static void vhost_vdpa_listener_begin(MemoryListener *listener)
+static void vhost_vdpa_listener_begin_batch(struct vhost_vdpa *v)
{
- struct vhost_vdpa *v = container_of(listener, struct vhost_vdpa, listener);
- struct vhost_dev *dev = v->dev;
- struct vhost_msg_v2 msg = {};
int fd = v->device_fd;
-
- if (!(dev->backend_cap & (0x1ULL << VHOST_BACKEND_F_IOTLB_BATCH))) {
- return;
- }
-
- msg.type = v->msg_type;
- msg.iotlb.type = VHOST_IOTLB_BATCH_BEGIN;
+ struct vhost_msg_v2 msg = {
+ .type = v->msg_type,
+ .iotlb.type = VHOST_IOTLB_BATCH_BEGIN,
+ };
if (write(fd, &msg, sizeof(msg)) != sizeof(msg)) {
error_report("failed to write, fd=%d, errno=%d (%s)",
@@ -109,6 +103,16 @@ static void vhost_vdpa_listener_begin(MemoryListener *listener)
}
}
+static void vhost_vdpa_iotlb_batch_begin_once(struct vhost_vdpa *v)
+{
+ if (v->dev->backend_cap & (0x1ULL << VHOST_BACKEND_F_IOTLB_BATCH) &&
+ !v->iotlb_batch_begin_sent) {
+ vhost_vdpa_listener_begin_batch(v);
+ }
+
+ v->iotlb_batch_begin_sent = true;
+}
+
static void vhost_vdpa_listener_commit(MemoryListener *listener)
{
struct vhost_vdpa *v = container_of(listener, struct vhost_vdpa, listener);
@@ -120,6 +124,10 @@ static void vhost_vdpa_listener_commit(MemoryListener *listener)
return;
}
+ if (!v->iotlb_batch_begin_sent) {
+ return;
+ }
+
msg.type = v->msg_type;
msg.iotlb.type = VHOST_IOTLB_BATCH_END;
@@ -127,6 +135,8 @@ static void vhost_vdpa_listener_commit(MemoryListener *listener)
error_report("failed to write, fd=%d, errno=%d (%s)",
fd, errno, strerror(errno));
}
+
+ v->iotlb_batch_begin_sent = false;
}
static void vhost_vdpa_listener_region_add(MemoryListener *listener,
@@ -170,6 +180,7 @@ static void vhost_vdpa_listener_region_add(MemoryListener *listener,
llsize = int128_sub(llend, int128_make64(iova));
+ vhost_vdpa_iotlb_batch_begin_once(v);
ret = vhost_vdpa_dma_map(v, iova, int128_get64(llsize),
vaddr, section->readonly);
if (ret) {
@@ -221,6 +232,7 @@ static void vhost_vdpa_listener_region_del(MemoryListener *listener,
llsize = int128_sub(llend, int128_make64(iova));
+ vhost_vdpa_iotlb_batch_begin_once(v);
ret = vhost_vdpa_dma_unmap(v, iova, int128_get64(llsize));
if (ret) {
error_report("vhost_vdpa dma unmap error!");
@@ -234,7 +246,6 @@ static void vhost_vdpa_listener_region_del(MemoryListener *listener,
* depends on the addnop().
*/
static const MemoryListener vhost_vdpa_memory_listener = {
- .begin = vhost_vdpa_listener_begin,
.commit = vhost_vdpa_listener_commit,
.region_add = vhost_vdpa_listener_region_add,
.region_del = vhost_vdpa_listener_region_del,
--
MST
next prev parent reply other threads:[~2021-09-04 21:37 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-04 21:35 [PULL 00/35] pc,pci,virtio: fixes, cleanups Michael S. Tsirkin
2021-09-04 21:35 ` Michael S. Tsirkin [this message]
2021-09-04 21:35 ` [PULL 02/35] hw/virtio: Fix leak of host-notifier memory-region Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 03/35] vhost: make SET_VRING_ADDR, SET_FEATURES send replies Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 04/35] hw/acpi: define PIIX4 acpi pci hotplug property strings at a single place Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 05/35] q35: catch invalid cpu hotplug configuration Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 06/35] hw/acpi: refactor acpi hp modules so that targets can just use what they need Michael S. Tsirkin
2021-09-06 9:58 ` Philippe Mathieu-Daudé
2021-09-06 10:03 ` Ani Sinha
2021-09-06 10:24 ` Philippe Mathieu-Daudé
2021-09-06 10:49 ` Ani Sinha
2021-09-07 5:55 ` Ani Sinha
2021-09-07 6:13 ` Philippe Mathieu-Daudé
2021-09-07 6:34 ` Ani Sinha
2021-09-07 9:49 ` Ani Sinha
2022-07-19 16:12 ` Peter Maydell
2022-07-19 16:21 ` Peter Maydell
2022-07-20 18:37 ` Ani Sinha
2022-07-20 21:34 ` Peter Maydell
2022-07-20 22:13 ` Ani Sinha
2022-07-21 10:51 ` BB
2022-07-21 12:35 ` Dr. David Alan Gilbert
2022-07-25 17:57 ` Ani Sinha
2021-09-04 21:36 ` [PULL 07/35] hw/virtio: move vhost_set_backend_type() to vhost.c Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 08/35] vhost-user: add missing space in error message Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 09/35] acpi: Delete broken ACPI_GED_X86 macro Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 10/35] Use PCI_HOST_BRIDGE macro Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 11/35] virtio-balloon: don't start free page hinting if postcopy is possible Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 12/35] virtio-balloon: free page hinting cleanups Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 13/35] virtio-bus: introduce iommu_enabled() Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 14/35] virtio-pci: implement iommu_enabled() Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 15/35] vhost: correctly detect the enabling IOMMU Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 16/35] hw/i386/acpi-build: Get NUMA information from struct NumaState Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 17/35] hw/pci: remove all references to find_i440fx function Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 18/35] hw/acpi: use existing references to pci device struct within functions Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 19/35] MAINTAINERS: Added myself as a reviewer for acpi/smbios subsystem Michael S. Tsirkin
2021-09-04 21:36 ` [PULL 20/35] hw/virtio: Document virtio_queue_packed_empty_rcu is called within RCU Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 21/35] hw/virtio: Remove NULL check in virtio_free_region_cache() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 22/35] hw/virtio: Add flatview update in vhost_user_cleanup() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 23/35] tests/vhost-user-bridge.c: Sanity check socket path length Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 24/35] tests/vhost-user-bridge.c: Fix typo in help message Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 25/35] vhost-vdpa: remove unused variable "acked_features" Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 26/35] vhost-vdpa: correctly return err in vhost_vdpa_set_backend_cap() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 27/35] vhost_net: remove the meaningless assignment in vhost_net_start_one() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 28/35] vhost: use unsigned int for nvqs Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 29/35] vhost_net: do not assume nvqs is always 2 Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 30/35] vhost-vdpa: remove the unnecessary check in vhost_vdpa_add() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 31/35] vhost-vdpa: don't cleanup twice " Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 32/35] vhost-vdpa: fix leaking of vhost_net " Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 33/35] vhost-vdpa: tweak the error label " Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 34/35] vhost-vdpa: fix the wrong assertion in vhost_vdpa_init() Michael S. Tsirkin
2021-09-04 21:37 ` [PULL 35/35] vhost-vdpa: remove the unncessary queue_index assignment Michael S. Tsirkin
2021-09-06 9:41 ` [PULL 00/35] pc,pci,virtio: fixes, cleanups Peter Maydell
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=20210904213506.486886-2-mst@redhat.com \
--to=mst@redhat.com \
--cc=eperezma@redhat.com \
--cc=jasowang@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).