From: Yuanhan Liu <yuanhan.liu@linux.intel.com>
To: dev@dpdk.org
Subject: [PATCH 3/3] vhost: fix vq realloc at numa_realloc
Date: Fri, 18 Dec 2015 15:04:07 +0800 [thread overview]
Message-ID: <1450422247-6814-3-git-send-email-yuanhan.liu@linux.intel.com> (raw)
In-Reply-To: <1450422247-6814-1-git-send-email-yuanhan.liu@linux.intel.com>
vq is allocated on pairs, hence we should do pair reallocation
at numa_realloc() as well, otherwise an error like following
occurs while do numa reallocation:
VHOST_CONFIG: reallocate vq from 0 to 1 node
PANIC in rte_free():
Fatal error: Invalid memory
The reason we don't catch it is because numa_realloc() will
not take effect when RTE_LIBRTE_VHOST_NUMA is not enabled,
which is the default case.
Fixes: e049ca6d10e0 ("vhost-user: prepare multiple queue setup")
Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
---
lib/librte_vhost/virtio-net.c | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/lib/librte_vhost/virtio-net.c b/lib/librte_vhost/virtio-net.c
index 31ca4f7..a962667 100644
--- a/lib/librte_vhost/virtio-net.c
+++ b/lib/librte_vhost/virtio-net.c
@@ -445,6 +445,13 @@ numa_realloc(struct virtio_net *dev, int index)
struct vhost_virtqueue *old_vq, *vq;
int ret;
+ /*
+ * vq is allocated on pairs, we should try to do realloc
+ * on first queue of one queue pair only.
+ */
+ if (index % VIRTIO_QNUM != 0)
+ return dev;
+
old_dev = dev;
vq = old_vq = dev->virtqueue[index];
@@ -461,11 +468,12 @@ numa_realloc(struct virtio_net *dev, int index)
if (oldnode != newnode) {
RTE_LOG(INFO, VHOST_CONFIG,
"reallocate vq from %d to %d node\n", oldnode, newnode);
- vq = rte_malloc_socket(NULL, sizeof(*vq), 0, newnode);
+ vq = rte_malloc_socket(NULL, sizeof(*vq) * VIRTIO_QNUM, 0,
+ newnode);
if (!vq)
return dev;
- memcpy(vq, old_vq, sizeof(*vq));
+ memcpy(vq, old_vq, sizeof(*vq) * VIRTIO_QNUM);
rte_free(old_vq);
}
@@ -491,6 +499,7 @@ numa_realloc(struct virtio_net *dev, int index)
out:
dev->virtqueue[index] = vq;
+ dev->virtqueue[index + 1] = vq + 1;
vhost_devices[dev->device_fh] = dev;
return dev;
--
1.9.0
next prev parent reply other threads:[~2015-12-18 7:03 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-18 7:04 [PATCH 1/3] vhost: get rid of linked list dev Yuanhan Liu
2015-12-18 7:04 ` [PATCH 2/3] vhost: simplify numa_realloc Yuanhan Liu
2015-12-22 6:46 ` Xie, Huawei
2015-12-22 6:52 ` Yuanhan Liu
2015-12-18 7:04 ` Yuanhan Liu [this message]
2015-12-22 6:56 ` [PATCH 3/3] vhost: fix vq realloc at numa_realloc Xie, Huawei
2015-12-22 3:33 ` [PATCH 1/3] vhost: get rid of linked list dev Xie, Huawei
2015-12-22 6:21 ` Xie, Huawei
2015-12-22 7:28 ` [PATCH v2 " Yuanhan Liu
2015-12-22 7:28 ` [PATCH v2 2/3] vhost: simplify numa_realloc Yuanhan Liu
2015-12-22 14:40 ` Xie, Huawei
2015-12-22 7:28 ` [PATCH v2 3/3] vhost: fix vq realloc at numa_realloc Yuanhan Liu
2016-03-07 13:49 ` Loftus, Ciara
2016-03-08 11:54 ` Yuanhan Liu
2016-03-10 4:19 ` [PATCH v3 0/3] vhost: virtio-net.c cleanups and fixes Yuanhan Liu
2016-03-10 4:19 ` [PATCH v3 1/3] vhost: get rid of linked list dev Yuanhan Liu
2016-03-10 4:20 ` [PATCH v3 2/3] vhost: simplify numa_realloc Yuanhan Liu
2016-03-10 4:20 ` [PATCH v3 3/3] vhost: fix vq realloc at numa_realloc Yuanhan Liu
2016-03-11 15:35 ` [PATCH v3 0/3] vhost: virtio-net.c cleanups and fixes Thomas Monjalon
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=1450422247-6814-3-git-send-email-yuanhan.liu@linux.intel.com \
--to=yuanhan.liu@linux.intel.com \
--cc=dev@dpdk.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.