[Qemu-devel] [PULL 0/1] Block patches
[PULL 0/1] Block patches
1
The following changes since commit 0b5e750bea635b167eb03d86c3d9a09bbd43bc06:
1
The following changes since commit 887cba855bb6ff4775256f7968409281350b568c:
2
2
3
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging (2019-02-12 10:53:37 +0000)
3
configure: Fix cross-building for RISCV host (v5) (2023-07-11 17:56:09 +0100)
4
4
5
are available in the Git repository at:
5
are available in the Git repository at:
6
6
7
git://github.com/stefanha/qemu.git tags/block-pull-request
7
https://gitlab.com/stefanha/qemu.git tags/block-pull-request
8
8
9
for you to fetch changes up to 42824b4d16da56a50ff4027f6cd22378e0e2666e:
9
for you to fetch changes up to 75dcb4d790bbe5327169fd72b185960ca58e2fa6:
10
10
11
virtio-blk: set correct config size for the host driver (2019-02-13 16:18:17 +0800)
11
virtio-blk: fix host notifier issues during dataplane start/stop (2023-07-12 15:20:32 -0400)
12
12
13
----------------------------------------------------------------
13
----------------------------------------------------------------
14
Pull request
14
Pull request
15
15
16
Fix a virtio-blk migration regression.
17
18
----------------------------------------------------------------
16
----------------------------------------------------------------
19
17
20
Changpeng Liu (1):
18
Stefan Hajnoczi (1):
21
virtio-blk: set correct config size for the host driver
19
virtio-blk: fix host notifier issues during dataplane start/stop
22
20
23
hw/block/virtio-blk.c | 13 +++++++++----
21
hw/block/dataplane/virtio-blk.c | 67 +++++++++++++++++++--------------
24
1 file changed, 9 insertions(+), 4 deletions(-)
22
1 file changed, 38 insertions(+), 29 deletions(-)
25
23
26
--
24
--
27
2.20.1
25
2.40.1
28
29
diff view generated by jsdifflib
[Qemu-devel] [PULL 1/1] virtio-blk: set correct config size for the host driver
[PULL 1/1] virtio-blk: fix host notifier issues during dataplane start/stop
1
From: Changpeng Liu <changpeng.liu@intel.com>
1
The main loop thread can consume 100% CPU when using --device
2
virtio-blk-pci,iothread=<iothread>. ppoll() constantly returns but
3
reading virtqueue host notifiers fails with EAGAIN. The file descriptors
4
are stale and remain registered with the AioContext because of bugs in
5
the virtio-blk dataplane start/stop code.
2
6
3
Commit caa1ee43 "vhost-user-blk: add discard/write zeroes features
7
The problem is that the dataplane start/stop code involves drain
4
support" added fields to struct virtio_blk_config. This changes
8
operations, which call virtio_blk_drained_begin() and
5
the size of the config space and breaks migration from QEMU 3.1
9
virtio_blk_drained_end() at points where the host notifier is not
6
and older:
10
operational:
11
- In virtio_blk_data_plane_start(), blk_set_aio_context() drains after
12
vblk->dataplane_started has been set to true but the host notifier has
13
not been attached yet.
14
- In virtio_blk_data_plane_stop(), blk_drain() and blk_set_aio_context()
15
drain after the host notifier has already been detached but with
16
vblk->dataplane_started still set to true.
7
17
8
qemu-system-ppc64: get_pci_config_device: Bad config data: i=0x10 read: 41 device: 1 cmask: ff wmask: 80 w1cmask:0
18
I would like to simplify ->ioeventfd_start/stop() to avoid interactions
9
qemu-system-ppc64: Failed to load PCIDevice:config
19
with drain entirely, but couldn't find a way to do that. Instead, this
10
qemu-system-ppc64: Failed to load virtio-blk:virtio
20
patch accepts the fragile nature of the code and reorders it so that
11
qemu-system-ppc64: error while loading state for instance 0x0 of device 'pci@800000020000000:01.0/virtio-blk'
21
vblk->dataplane_started is false during drain operations. This way the
12
qemu-system-ppc64: load of migration failed: Invalid argument
22
virtio_blk_drained_begin() and virtio_blk_drained_end() calls don't
23
touch the host notifier. The result is that
24
virtio_blk_data_plane_start() and virtio_blk_data_plane_stop() have
25
complete control over the host notifier and stale file descriptors are
26
no longer left in the AioContext.
13
27
14
Since virtio-blk doesn't support the "discard" and "write zeroes"
28
This patch fixes the 100% CPU consumption in the main loop thread and
15
features, it shouldn't even expose the associated fields in the
29
correctly moves host notifier processing to the IOThread.
16
config space actually. Just include all fields up to num_queues to
17
match QEMU 3.1 and older.
18
30
19
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
31
Fixes: 1665d9326fd2 ("virtio-blk: implement BlockDevOps->drained_begin()")
20
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
32
Reported-by: Lukáš Doktor <ldoktor@redhat.com>
21
Message-id: 1550022537-27565-1-git-send-email-changpeng.liu@intel.com
33
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
22
Message-Id: <1550022537-27565-1-git-send-email-changpeng.liu@intel.com>
34
Tested-by: Lukas Doktor <ldoktor@redhat.com>
35
Message-id: 20230704151527.193586-1-stefanha@redhat.com
23
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
36
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
24
---
37
---
25
hw/block/virtio-blk.c | 13 +++++++++----
38
hw/block/dataplane/virtio-blk.c | 67 +++++++++++++++++++--------------
26
1 file changed, 9 insertions(+), 4 deletions(-)
39
1 file changed, 38 insertions(+), 29 deletions(-)
27
40
28
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
41
diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c
29
index XXXXXXX..XXXXXXX 100644
42
index XXXXXXX..XXXXXXX 100644
30
--- a/hw/block/virtio-blk.c
43
--- a/hw/block/dataplane/virtio-blk.c
31
+++ b/hw/block/virtio-blk.c
44
+++ b/hw/block/dataplane/virtio-blk.c
32
@@ -XXX,XX +XXX,XX @@
45
@@ -XXX,XX +XXX,XX @@ int virtio_blk_data_plane_start(VirtIODevice *vdev)
33
#include "hw/virtio/virtio-bus.h"
46
34
#include "hw/virtio/virtio-access.h"
47
memory_region_transaction_commit();
35
48
36
+/* We don't support discard yet, hide associated config fields. */
49
- /*
37
+#define VIRTIO_BLK_CFG_SIZE offsetof(struct virtio_blk_config, \
50
- * These fields are visible to the IOThread so we rely on implicit barriers
38
+ max_discard_sectors)
51
- * in aio_context_acquire() on the write side and aio_notify_accept() on
52
- * the read side.
53
- */
54
- s->starting = false;
55
- vblk->dataplane_started = true;
56
trace_virtio_blk_data_plane_start(s);
57
58
old_context = blk_get_aio_context(s->conf->conf.blk);
59
@@ -XXX,XX +XXX,XX @@ int virtio_blk_data_plane_start(VirtIODevice *vdev)
60
event_notifier_set(virtio_queue_get_host_notifier(vq));
61
}
62
63
+ /*
64
+ * These fields must be visible to the IOThread when it processes the
65
+ * virtqueue, otherwise it will think dataplane has not started yet.
66
+ *
67
+ * Make sure ->dataplane_started is false when blk_set_aio_context() is
68
+ * called above so that draining does not cause the host notifier to be
69
+ * detached/attached prematurely.
70
+ */
71
+ s->starting = false;
72
+ vblk->dataplane_started = true;
73
+ smp_wmb(); /* paired with aio_notify_accept() on the read side */
39
+
74
+
40
static void virtio_blk_init_request(VirtIOBlock *s, VirtQueue *vq,
75
/* Get this show started by hooking up our callbacks */
41
VirtIOBlockReq *req)
76
if (!blk_in_drain(s->conf->conf.blk)) {
42
{
77
aio_context_acquire(s->ctx);
43
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_update_config(VirtIODevice *vdev, uint8_t *config)
78
@@ -XXX,XX +XXX,XX @@ int virtio_blk_data_plane_start(VirtIODevice *vdev)
44
blkcfg.alignment_offset = 0;
79
fail_guest_notifiers:
45
blkcfg.wce = blk_enable_write_cache(s->blk);
80
vblk->dataplane_disabled = true;
46
virtio_stw_p(vdev, &blkcfg.num_queues, s->conf.num_queues);
81
s->starting = false;
47
- memcpy(config, &blkcfg, sizeof(struct virtio_blk_config));
82
- vblk->dataplane_started = true;
48
+ memcpy(config, &blkcfg, VIRTIO_BLK_CFG_SIZE);
83
return -ENOSYS;
49
+ QEMU_BUILD_BUG_ON(VIRTIO_BLK_CFG_SIZE > sizeof(blkcfg));
50
}
84
}
51
85
52
static void virtio_blk_set_config(VirtIODevice *vdev, const uint8_t *config)
86
@@ -XXX,XX +XXX,XX @@ void virtio_blk_data_plane_stop(VirtIODevice *vdev)
53
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_set_config(VirtIODevice *vdev, const uint8_t *config)
87
aio_wait_bh_oneshot(s->ctx, virtio_blk_data_plane_stop_bh, s);
54
VirtIOBlock *s = VIRTIO_BLK(vdev);
55
struct virtio_blk_config blkcfg;
56
57
- memcpy(&blkcfg, config, sizeof(blkcfg));
58
+ memcpy(&blkcfg, config, VIRTIO_BLK_CFG_SIZE);
59
+ QEMU_BUILD_BUG_ON(VIRTIO_BLK_CFG_SIZE > sizeof(blkcfg));
60
61
aio_context_acquire(blk_get_aio_context(s->blk));
62
blk_set_enable_write_cache(s->blk, blkcfg.wce != 0);
63
@@ -XXX,XX +XXX,XX @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp)
64
return;
65
}
88
}
66
89
67
- virtio_init(vdev, "virtio-blk", VIRTIO_ID_BLOCK,
90
+ /*
68
- sizeof(struct virtio_blk_config));
91
+ * Batch all the host notifiers in a single transaction to avoid
69
+ virtio_init(vdev, "virtio-blk", VIRTIO_ID_BLOCK, VIRTIO_BLK_CFG_SIZE);
92
+ * quadratic time complexity in address_space_update_ioeventfds().
70
93
+ */
71
s->blk = conf->conf.blk;
94
+ memory_region_transaction_begin();
72
s->rq = NULL;
95
+
96
+ for (i = 0; i < nvqs; i++) {
97
+ virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), i, false);
98
+ }
99
+
100
+ /*
101
+ * The transaction expects the ioeventfds to be open when it
102
+ * commits. Do it now, before the cleanup loop.
103
+ */
104
+ memory_region_transaction_commit();
105
+
106
+ for (i = 0; i < nvqs; i++) {
107
+ virtio_bus_cleanup_host_notifier(VIRTIO_BUS(qbus), i);
108
+ }
109
+
110
+ /*
111
+ * Set ->dataplane_started to false before draining so that host notifiers
112
+ * are not detached/attached anymore.
113
+ */
114
+ vblk->dataplane_started = false;
115
+
116
aio_context_acquire(s->ctx);
117
118
/* Wait for virtio_blk_dma_restart_bh() and in flight I/O to complete */
119
@@ -XXX,XX +XXX,XX @@ void virtio_blk_data_plane_stop(VirtIODevice *vdev)
120
121
aio_context_release(s->ctx);
122
123
- /*
124
- * Batch all the host notifiers in a single transaction to avoid
125
- * quadratic time complexity in address_space_update_ioeventfds().
126
- */
127
- memory_region_transaction_begin();
128
-
129
- for (i = 0; i < nvqs; i++) {
130
- virtio_bus_set_host_notifier(VIRTIO_BUS(qbus), i, false);
131
- }
132
-
133
- /*
134
- * The transaction expects the ioeventfds to be open when it
135
- * commits. Do it now, before the cleanup loop.
136
- */
137
- memory_region_transaction_commit();
138
-
139
- for (i = 0; i < nvqs; i++) {
140
- virtio_bus_cleanup_host_notifier(VIRTIO_BUS(qbus), i);
141
- }
142
-
143
qemu_bh_cancel(s->bh);
144
notify_guest_bh(s); /* final chance to notify guest */
145
146
/* Clean up guest notifier (irq) */
147
k->set_guest_notifiers(qbus->parent, nvqs, false);
148
149
- vblk->dataplane_started = false;
150
s->stopping = false;
151
}
73
--
152
--
74
2.20.1
153
2.40.1
75
154
76
155
diff view generated by jsdifflib

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.