2009-09-24 15:59:20 +00:00
|
|
|
/* A network driver using virtio.
|
2007-10-22 01:03:37 +00:00
|
|
|
*
|
|
|
|
* Copyright 2007 Rusty Russell <rusty@rustcorp.com.au> IBM Corporation
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
|
*/
|
|
|
|
//#define DEBUG
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/etherdevice.h>
|
2008-04-18 03:21:42 +00:00
|
|
|
#include <linux/ethtool.h>
|
2007-10-22 01:03:37 +00:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/virtio.h>
|
|
|
|
#include <linux/virtio_net.h>
|
|
|
|
#include <linux/scatterlist.h>
|
2009-01-26 02:06:26 +00:00
|
|
|
#include <linux/if_vlan.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
|
|
|
#include <linux/slab.h>
|
2013-01-24 23:51:31 +00:00
|
|
|
#include <linux/cpu.h>
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2013-05-09 19:50:51 +00:00
|
|
|
static int napi_weight = NAPI_POLL_WEIGHT;
|
2007-12-16 13:19:43 +00:00
|
|
|
module_param(napi_weight, int, 0444);
|
|
|
|
|
2011-12-19 14:08:01 +00:00
|
|
|
static bool csum = true, gso = true;
|
2008-02-05 04:50:02 +00:00
|
|
|
module_param(csum, bool, 0444);
|
|
|
|
module_param(gso, bool, 0444);
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
/* FIXME: MTU in config. */
|
2009-01-26 02:06:26 +00:00
|
|
|
#define MAX_PACKET_LEN (ETH_HLEN + VLAN_HLEN + ETH_DATA_LEN)
|
2008-11-17 06:41:34 +00:00
|
|
|
#define GOOD_COPY_LEN 128
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2011-11-14 14:17:08 +00:00
|
|
|
#define VIRTNET_DRIVER_VERSION "1.0.0"
|
2009-02-04 09:02:34 +00:00
|
|
|
|
2011-06-15 06:36:29 +00:00
|
|
|
struct virtnet_stats {
|
2012-06-05 22:35:24 +00:00
|
|
|
struct u64_stats_sync tx_syncp;
|
|
|
|
struct u64_stats_sync rx_syncp;
|
2011-06-15 06:36:29 +00:00
|
|
|
u64 tx_bytes;
|
|
|
|
u64 tx_packets;
|
|
|
|
|
|
|
|
u64 rx_bytes;
|
|
|
|
u64 rx_packets;
|
|
|
|
};
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* Internal representation of a send virtqueue */
|
|
|
|
struct send_queue {
|
|
|
|
/* Virtqueue associated with this send _queue */
|
|
|
|
struct virtqueue *vq;
|
|
|
|
|
|
|
|
/* TX: fragments + linear part + virtio header */
|
|
|
|
struct scatterlist sg[MAX_SKB_FRAGS + 2];
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
/* Name of the send queue: output.$index */
|
|
|
|
char name[40];
|
2012-12-07 07:04:55 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
/* Internal representation of a receive virtqueue */
|
|
|
|
struct receive_queue {
|
|
|
|
/* Virtqueue associated with this receive_queue */
|
|
|
|
struct virtqueue *vq;
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
struct napi_struct napi;
|
|
|
|
|
|
|
|
/* Number of input buffers, and max we've ever had. */
|
|
|
|
unsigned int num, max;
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* Chain pages by the private ptr. */
|
|
|
|
struct page *pages;
|
|
|
|
|
|
|
|
/* RX: fragments + linear part + virtio header */
|
|
|
|
struct scatterlist sg[MAX_SKB_FRAGS + 2];
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
/* Name of this receive queue: input.$index */
|
|
|
|
char name[40];
|
2012-12-07 07:04:55 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
struct virtnet_info {
|
|
|
|
struct virtio_device *vdev;
|
|
|
|
struct virtqueue *cvq;
|
|
|
|
struct net_device *dev;
|
2012-12-07 07:04:56 +00:00
|
|
|
struct send_queue *sq;
|
|
|
|
struct receive_queue *rq;
|
2012-12-07 07:04:55 +00:00
|
|
|
unsigned int status;
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* Max # of queue pairs supported by the device */
|
|
|
|
u16 max_queue_pairs;
|
|
|
|
|
|
|
|
/* # of queue pairs currently used by the driver */
|
|
|
|
u16 curr_queue_pairs;
|
|
|
|
|
2008-04-18 03:24:27 +00:00
|
|
|
/* I like... big packets and I cannot lie! */
|
|
|
|
bool big_packets;
|
|
|
|
|
2008-11-17 06:41:34 +00:00
|
|
|
/* Host will merge rx buffers for big packets (shake it! shake it!) */
|
|
|
|
bool mergeable_rx_bufs;
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* Has control virtqueue */
|
|
|
|
bool has_cvq;
|
|
|
|
|
2013-07-25 00:50:23 +00:00
|
|
|
/* Host can handle any s/g split between our header and packet data */
|
|
|
|
bool any_header_sg;
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
/* enable config space updates */
|
|
|
|
bool config_enable;
|
|
|
|
|
2011-06-15 06:36:29 +00:00
|
|
|
/* Active statistics */
|
|
|
|
struct virtnet_stats __percpu *stats;
|
|
|
|
|
2009-08-26 19:22:32 +00:00
|
|
|
/* Work struct for refilling if we run low on memory. */
|
|
|
|
struct delayed_work refill;
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
/* Work struct for config space updates */
|
|
|
|
struct work_struct config_work;
|
|
|
|
|
|
|
|
/* Lock for config space updates */
|
|
|
|
struct mutex config_lock;
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
/* Does the affinity hint is set for virtqueues? */
|
|
|
|
bool affinity_hint_set;
|
2013-01-24 23:51:29 +00:00
|
|
|
|
|
|
|
/* Per-cpu variable to show the mapping from CPU to virtqueue */
|
|
|
|
int __percpu *vq_index;
|
2013-01-24 23:51:31 +00:00
|
|
|
|
|
|
|
/* CPU hot plug notifier */
|
|
|
|
struct notifier_block nb;
|
2007-10-22 01:03:37 +00:00
|
|
|
};
|
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
struct skb_vnet_hdr {
|
|
|
|
union {
|
|
|
|
struct virtio_net_hdr hdr;
|
|
|
|
struct virtio_net_hdr_mrg_rxbuf mhdr;
|
|
|
|
};
|
|
|
|
};
|
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
struct padded_vnet_hdr {
|
|
|
|
struct virtio_net_hdr hdr;
|
|
|
|
/*
|
|
|
|
* virtio_net_hdr should be in a separated sg buffer because of a
|
|
|
|
* QEMU bug, and data sg buffer shares same page with this header sg.
|
|
|
|
* This padding makes next sg 16 byte aligned after virtio_net_hdr.
|
|
|
|
*/
|
|
|
|
char padding[6];
|
|
|
|
};
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* Converting between virtqueue no. and kernel tx/rx queue no.
|
|
|
|
* 0:rx0 1:tx0 2:rx1 3:tx1 ... 2N:rxN 2N+1:txN 2N+2:cvq
|
|
|
|
*/
|
|
|
|
static int vq2txq(struct virtqueue *vq)
|
|
|
|
{
|
2013-03-21 14:17:34 +00:00
|
|
|
return (vq->index - 1) / 2;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int txq2vq(int txq)
|
|
|
|
{
|
|
|
|
return txq * 2 + 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int vq2rxq(struct virtqueue *vq)
|
|
|
|
{
|
2013-03-21 14:17:34 +00:00
|
|
|
return vq->index / 2;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int rxq2vq(int rxq)
|
|
|
|
{
|
|
|
|
return rxq * 2;
|
|
|
|
}
|
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2009-09-24 15:59:19 +00:00
|
|
|
return (struct skb_vnet_hdr *)skb->cb;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
/*
|
|
|
|
* private is used to chain pages for big packets, put the whole
|
|
|
|
* most recent used list in the beginning for reuse
|
|
|
|
*/
|
2012-12-07 07:04:55 +00:00
|
|
|
static void give_pages(struct receive_queue *rq, struct page *page)
|
2008-11-17 06:39:18 +00:00
|
|
|
{
|
2010-01-29 03:20:04 +00:00
|
|
|
struct page *end;
|
2008-11-17 06:39:18 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* Find end of list, sew whole thing into vi->rq.pages. */
|
2010-01-29 03:20:04 +00:00
|
|
|
for (end = page; end->private; end = (struct page *)end->private);
|
2012-12-07 07:04:55 +00:00
|
|
|
end->private = (unsigned long)rq->pages;
|
|
|
|
rq->pages = page;
|
2008-11-17 06:39:18 +00:00
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static struct page *get_a_page(struct receive_queue *rq, gfp_t gfp_mask)
|
2008-07-25 17:06:01 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct page *p = rq->pages;
|
2008-07-25 17:06:01 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (p) {
|
2012-12-07 07:04:55 +00:00
|
|
|
rq->pages = (struct page *)p->private;
|
2010-01-29 03:20:04 +00:00
|
|
|
/* clear private here, it is used to chain pages */
|
|
|
|
p->private = 0;
|
|
|
|
} else
|
2008-07-25 17:06:01 +00:00
|
|
|
p = alloc_page(gfp_mask);
|
|
|
|
return p;
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static void skb_xmit_done(struct virtqueue *vq)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = vq->vdev->priv;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2008-02-05 04:50:07 +00:00
|
|
|
/* Suppress further interrupts. */
|
2012-12-07 07:04:55 +00:00
|
|
|
virtqueue_disable_cb(vq);
|
2008-05-26 07:48:13 +00:00
|
|
|
|
2008-06-08 10:51:55 +00:00
|
|
|
/* We were probably waiting for more output buffers. */
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_wake_subqueue(vi->dev, vq2txq(vq));
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
static void set_skb_frag(struct sk_buff *skb, struct page *page,
|
|
|
|
unsigned int offset, unsigned int *len)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2011-10-19 22:17:27 +00:00
|
|
|
int size = min((unsigned)PAGE_SIZE - offset, *len);
|
2010-01-29 03:20:04 +00:00
|
|
|
int i = skb_shinfo(skb)->nr_frags;
|
|
|
|
|
2011-10-19 22:17:27 +00:00
|
|
|
__skb_fill_page_desc(skb, i, page, offset, size);
|
2010-01-29 03:20:04 +00:00
|
|
|
|
2011-10-19 22:17:27 +00:00
|
|
|
skb->data_len += size;
|
|
|
|
skb->len += size;
|
2011-10-19 23:14:46 +00:00
|
|
|
skb->truesize += PAGE_SIZE;
|
2010-01-29 03:20:04 +00:00
|
|
|
skb_shinfo(skb)->nr_frags++;
|
2013-02-11 09:27:41 +00:00
|
|
|
skb_shinfo(skb)->tx_flags |= SKBTX_SHARED_FRAG;
|
2011-10-19 22:17:27 +00:00
|
|
|
*len -= size;
|
2010-01-29 03:20:04 +00:00
|
|
|
}
|
2008-06-08 10:49:00 +00:00
|
|
|
|
2012-01-04 12:52:32 +00:00
|
|
|
/* Called from bottom half context */
|
2012-12-07 07:04:55 +00:00
|
|
|
static struct sk_buff *page_to_skb(struct receive_queue *rq,
|
2010-01-29 03:20:04 +00:00
|
|
|
struct page *page, unsigned int len)
|
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = rq->vq->vdev->priv;
|
2010-01-29 03:20:04 +00:00
|
|
|
struct sk_buff *skb;
|
|
|
|
struct skb_vnet_hdr *hdr;
|
|
|
|
unsigned int copy, hdr_len, offset;
|
|
|
|
char *p;
|
2008-07-25 17:06:01 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
p = page_address(page);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
/* copy small packet so we can reuse these pages for small data */
|
|
|
|
skb = netdev_alloc_skb_ip_align(vi->dev, GOOD_COPY_LEN);
|
|
|
|
if (unlikely(!skb))
|
|
|
|
return NULL;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
hdr = skb_vnet_hdr(skb);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (vi->mergeable_rx_bufs) {
|
|
|
|
hdr_len = sizeof hdr->mhdr;
|
|
|
|
offset = hdr_len;
|
|
|
|
} else {
|
|
|
|
hdr_len = sizeof hdr->hdr;
|
|
|
|
offset = sizeof(struct padded_vnet_hdr);
|
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
memcpy(hdr, p, hdr_len);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
len -= hdr_len;
|
|
|
|
p += offset;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
copy = len;
|
|
|
|
if (copy > skb_tailroom(skb))
|
|
|
|
copy = skb_tailroom(skb);
|
|
|
|
memcpy(skb_put(skb, copy), p, copy);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
len -= copy;
|
|
|
|
offset += copy;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2011-09-28 04:40:54 +00:00
|
|
|
/*
|
|
|
|
* Verify that we can indeed put this data into a skb.
|
|
|
|
* This is here to handle cases when the device erroneously
|
|
|
|
* tries to receive more than is possible. This is usually
|
|
|
|
* the case of a broken device.
|
|
|
|
*/
|
|
|
|
if (unlikely(len > MAX_SKB_FRAGS * PAGE_SIZE)) {
|
2012-11-08 17:47:28 +00:00
|
|
|
net_dbg_ratelimited("%s: too much data\n", skb->dev->name);
|
2011-09-28 04:40:54 +00:00
|
|
|
dev_kfree_skb(skb);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
while (len) {
|
|
|
|
set_skb_frag(skb, page, offset, &len);
|
|
|
|
page = (struct page *)page->private;
|
|
|
|
offset = 0;
|
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (page)
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, page);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
return skb;
|
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static int receive_mergeable(struct receive_queue *rq, struct sk_buff *skb)
|
2010-01-29 03:20:04 +00:00
|
|
|
{
|
|
|
|
struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
|
|
|
|
struct page *page;
|
|
|
|
int num_buf, i, len;
|
|
|
|
|
|
|
|
num_buf = hdr->mhdr.num_buffers;
|
|
|
|
while (--num_buf) {
|
|
|
|
i = skb_shinfo(skb)->nr_frags;
|
|
|
|
if (i >= MAX_SKB_FRAGS) {
|
|
|
|
pr_debug("%s: packet too long\n", skb->dev->name);
|
|
|
|
skb->dev->stats.rx_length_errors++;
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2012-12-07 07:04:55 +00:00
|
|
|
page = virtqueue_get_buf(rq->vq, &len);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (!page) {
|
|
|
|
pr_debug("%s: rx error: %d buffers missing\n",
|
|
|
|
skb->dev->name, hdr->mhdr.num_buffers);
|
|
|
|
skb->dev->stats.rx_length_errors++;
|
|
|
|
return -EINVAL;
|
2008-11-17 06:41:34 +00:00
|
|
|
}
|
2011-06-15 06:36:29 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (len > PAGE_SIZE)
|
|
|
|
len = PAGE_SIZE;
|
|
|
|
|
|
|
|
set_skb_frag(skb, page, 0, &len);
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
--rq->num;
|
2010-01-29 03:20:04 +00:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static void receive_buf(struct receive_queue *rq, void *buf, unsigned int len)
|
2010-01-29 03:20:04 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = rq->vq->vdev->priv;
|
|
|
|
struct net_device *dev = vi->dev;
|
2012-02-13 06:53:41 +00:00
|
|
|
struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
|
2010-01-29 03:20:04 +00:00
|
|
|
struct sk_buff *skb;
|
|
|
|
struct page *page;
|
|
|
|
struct skb_vnet_hdr *hdr;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (unlikely(len < sizeof(struct virtio_net_hdr) + ETH_HLEN)) {
|
|
|
|
pr_debug("%s: short packet %i\n", dev->name, len);
|
|
|
|
dev->stats.rx_length_errors++;
|
|
|
|
if (vi->mergeable_rx_bufs || vi->big_packets)
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, buf);
|
2010-01-29 03:20:04 +00:00
|
|
|
else
|
|
|
|
dev_kfree_skb(buf);
|
|
|
|
return;
|
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
if (!vi->mergeable_rx_bufs && !vi->big_packets) {
|
|
|
|
skb = buf;
|
|
|
|
len -= sizeof(struct virtio_net_hdr);
|
|
|
|
skb_trim(skb, len);
|
|
|
|
} else {
|
|
|
|
page = buf;
|
2012-12-07 07:04:55 +00:00
|
|
|
skb = page_to_skb(rq, page, len);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (unlikely(!skb)) {
|
2008-11-17 06:41:34 +00:00
|
|
|
dev->stats.rx_dropped++;
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, page);
|
2010-01-29 03:20:04 +00:00
|
|
|
return;
|
2008-11-17 06:41:34 +00:00
|
|
|
}
|
2010-01-29 03:20:04 +00:00
|
|
|
if (vi->mergeable_rx_bufs)
|
2012-12-07 07:04:55 +00:00
|
|
|
if (receive_mergeable(rq, skb)) {
|
2010-01-29 03:20:04 +00:00
|
|
|
dev_kfree_skb(skb);
|
|
|
|
return;
|
|
|
|
}
|
2008-04-18 03:24:27 +00:00
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
hdr = skb_vnet_hdr(skb);
|
2011-06-15 06:36:29 +00:00
|
|
|
|
2012-06-05 22:35:24 +00:00
|
|
|
u64_stats_update_begin(&stats->rx_syncp);
|
2011-06-15 06:36:29 +00:00
|
|
|
stats->rx_bytes += skb->len;
|
|
|
|
stats->rx_packets++;
|
2012-06-05 22:35:24 +00:00
|
|
|
u64_stats_update_end(&stats->rx_syncp);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
if (hdr->hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
|
2007-10-22 01:03:37 +00:00
|
|
|
pr_debug("Needs csum!\n");
|
2009-09-24 15:59:19 +00:00
|
|
|
if (!skb_partial_csum_set(skb,
|
|
|
|
hdr->hdr.csum_start,
|
|
|
|
hdr->hdr.csum_offset))
|
2007-10-22 01:03:37 +00:00
|
|
|
goto frame_err;
|
2011-06-10 00:56:17 +00:00
|
|
|
} else if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID) {
|
|
|
|
skb->ip_summed = CHECKSUM_UNNECESSARY;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2008-06-08 10:49:00 +00:00
|
|
|
skb->protocol = eth_type_trans(skb, dev);
|
|
|
|
pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
|
|
|
|
ntohs(skb->protocol), skb->len, skb->pkt_type);
|
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
if (hdr->hdr.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
|
2007-10-22 01:03:37 +00:00
|
|
|
pr_debug("GSO!\n");
|
2009-09-24 15:59:19 +00:00
|
|
|
switch (hdr->hdr.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
|
2007-10-22 01:03:37 +00:00
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV4:
|
2013-02-11 09:27:41 +00:00
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
|
2007-10-22 01:03:37 +00:00
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_UDP:
|
2013-02-11 09:27:41 +00:00
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
|
2007-10-22 01:03:37 +00:00
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV6:
|
2013-02-11 09:27:41 +00:00
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
|
2007-10-22 01:03:37 +00:00
|
|
|
break;
|
|
|
|
default:
|
2012-11-08 17:47:28 +00:00
|
|
|
net_warn_ratelimited("%s: bad gso type %u.\n",
|
|
|
|
dev->name, hdr->hdr.gso_type);
|
2007-10-22 01:03:37 +00:00
|
|
|
goto frame_err;
|
|
|
|
}
|
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
if (hdr->hdr.gso_type & VIRTIO_NET_HDR_GSO_ECN)
|
2013-02-11 09:27:41 +00:00
|
|
|
skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
|
2008-02-05 04:50:02 +00:00
|
|
|
|
2009-09-24 15:59:19 +00:00
|
|
|
skb_shinfo(skb)->gso_size = hdr->hdr.gso_size;
|
2007-10-22 01:03:37 +00:00
|
|
|
if (skb_shinfo(skb)->gso_size == 0) {
|
2012-11-08 17:47:28 +00:00
|
|
|
net_warn_ratelimited("%s: zero gso size.\n", dev->name);
|
2007-10-22 01:03:37 +00:00
|
|
|
goto frame_err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Header must be checked, and gso_segs computed. */
|
|
|
|
skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
|
|
|
|
skb_shinfo(skb)->gso_segs = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
netif_receive_skb(skb);
|
|
|
|
return;
|
|
|
|
|
|
|
|
frame_err:
|
|
|
|
dev->stats.rx_frame_errors++;
|
|
|
|
dev_kfree_skb(skb);
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static int add_recvbuf_small(struct receive_queue *rq, gfp_t gfp)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = rq->vq->vdev->priv;
|
2007-10-22 01:03:37 +00:00
|
|
|
struct sk_buff *skb;
|
2010-01-29 03:20:04 +00:00
|
|
|
struct skb_vnet_hdr *hdr;
|
|
|
|
int err;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2012-01-04 12:52:32 +00:00
|
|
|
skb = __netdev_alloc_skb_ip_align(vi->dev, MAX_PACKET_LEN, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (unlikely(!skb))
|
|
|
|
return -ENOMEM;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
skb_put(skb, MAX_PACKET_LEN);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
hdr = skb_vnet_hdr(skb);
|
2012-12-07 07:04:55 +00:00
|
|
|
sg_set_buf(rq->sg, &hdr->hdr, sizeof hdr->hdr);
|
2008-04-18 03:24:27 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
skb_to_sgvec(skb, rq->sg + 1, 0, skb->len);
|
2008-04-18 03:24:27 +00:00
|
|
|
|
2013-03-20 05:14:28 +00:00
|
|
|
err = virtqueue_add_inbuf(rq->vq, rq->sg, 2, skb, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (err < 0)
|
|
|
|
dev_kfree_skb(skb);
|
2008-04-18 03:24:27 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
return err;
|
|
|
|
}
|
2008-04-18 03:24:27 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static int add_recvbuf_big(struct receive_queue *rq, gfp_t gfp)
|
2010-01-29 03:20:04 +00:00
|
|
|
{
|
|
|
|
struct page *first, *list = NULL;
|
|
|
|
char *p;
|
|
|
|
int i, err, offset;
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* page in rq->sg[MAX_SKB_FRAGS + 1] is list tail */
|
2010-01-29 03:20:04 +00:00
|
|
|
for (i = MAX_SKB_FRAGS + 1; i > 1; --i) {
|
2012-12-07 07:04:55 +00:00
|
|
|
first = get_a_page(rq, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (!first) {
|
|
|
|
if (list)
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, list);
|
2010-01-29 03:20:04 +00:00
|
|
|
return -ENOMEM;
|
2008-04-18 03:24:27 +00:00
|
|
|
}
|
2012-12-07 07:04:55 +00:00
|
|
|
sg_set_buf(&rq->sg[i], page_address(first), PAGE_SIZE);
|
2008-04-18 03:24:27 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
/* chain new page in list head to match sg */
|
|
|
|
first->private = (unsigned long)list;
|
|
|
|
list = first;
|
|
|
|
}
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
first = get_a_page(rq, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (!first) {
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, list);
|
2010-01-29 03:20:04 +00:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
p = page_address(first);
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* rq->sg[0], rq->sg[1] share the same page */
|
|
|
|
/* a separated rq->sg[0] for virtio_net_hdr only due to QEMU bug */
|
|
|
|
sg_set_buf(&rq->sg[0], p, sizeof(struct virtio_net_hdr));
|
2010-01-29 03:20:04 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
/* rq->sg[1] for data packet, from offset */
|
2010-01-29 03:20:04 +00:00
|
|
|
offset = sizeof(struct padded_vnet_hdr);
|
2012-12-07 07:04:55 +00:00
|
|
|
sg_set_buf(&rq->sg[1], p + offset, PAGE_SIZE - offset);
|
2010-01-29 03:20:04 +00:00
|
|
|
|
|
|
|
/* chain first in list head */
|
|
|
|
first->private = (unsigned long)list;
|
2013-03-20 05:14:28 +00:00
|
|
|
err = virtqueue_add_inbuf(rq->vq, rq->sg, MAX_SKB_FRAGS + 2,
|
|
|
|
first, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (err < 0)
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, first);
|
2010-01-29 03:20:04 +00:00
|
|
|
|
|
|
|
return err;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static int add_recvbuf_mergeable(struct receive_queue *rq, gfp_t gfp)
|
2008-11-17 06:41:34 +00:00
|
|
|
{
|
2010-01-29 03:20:04 +00:00
|
|
|
struct page *page;
|
2008-11-17 06:41:34 +00:00
|
|
|
int err;
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
page = get_a_page(rq, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (!page)
|
|
|
|
return -ENOMEM;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
sg_init_one(rq->sg, page_address(page), PAGE_SIZE);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2013-03-20 05:14:28 +00:00
|
|
|
err = virtqueue_add_inbuf(rq->vq, rq->sg, 1, page, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
if (err < 0)
|
2012-12-07 07:04:55 +00:00
|
|
|
give_pages(rq, page);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
return err;
|
|
|
|
}
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2011-12-29 00:42:38 +00:00
|
|
|
/*
|
|
|
|
* Returns false if we couldn't fill entirely (OOM).
|
|
|
|
*
|
|
|
|
* Normally run in the receive path, but can also be run from ndo_open
|
|
|
|
* before we're receiving packets, or from refill_work which is
|
|
|
|
* careful to disable receiving (using napi_disable).
|
|
|
|
*/
|
2012-12-07 07:04:55 +00:00
|
|
|
static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
|
2010-01-29 03:20:04 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = rq->vq->vdev->priv;
|
2010-01-29 03:20:04 +00:00
|
|
|
int err;
|
2010-07-02 16:32:55 +00:00
|
|
|
bool oom;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-01-29 03:20:04 +00:00
|
|
|
do {
|
|
|
|
if (vi->mergeable_rx_bufs)
|
2012-12-07 07:04:55 +00:00
|
|
|
err = add_recvbuf_mergeable(rq, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
else if (vi->big_packets)
|
2012-12-07 07:04:55 +00:00
|
|
|
err = add_recvbuf_big(rq, gfp);
|
2010-01-29 03:20:04 +00:00
|
|
|
else
|
2012-12-07 07:04:55 +00:00
|
|
|
err = add_recvbuf_small(rq, gfp);
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2010-07-02 16:32:55 +00:00
|
|
|
oom = err == -ENOMEM;
|
2012-10-16 13:26:14 +00:00
|
|
|
if (err)
|
2008-11-17 06:41:34 +00:00
|
|
|
break;
|
2012-12-07 07:04:55 +00:00
|
|
|
++rq->num;
|
2012-12-20 16:37:04 +00:00
|
|
|
} while (rq->vq->num_free);
|
2012-12-07 07:04:55 +00:00
|
|
|
if (unlikely(rq->num > rq->max))
|
|
|
|
rq->max = rq->num;
|
|
|
|
virtqueue_kick(rq->vq);
|
2009-08-26 19:22:32 +00:00
|
|
|
return !oom;
|
2008-11-17 06:41:34 +00:00
|
|
|
}
|
|
|
|
|
2008-02-05 04:49:57 +00:00
|
|
|
static void skb_recv_done(struct virtqueue *rvq)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = rvq->vdev->priv;
|
2012-12-07 07:04:56 +00:00
|
|
|
struct receive_queue *rq = &vi->rq[vq2rxq(rvq)];
|
2012-12-07 07:04:55 +00:00
|
|
|
|
2008-02-05 04:49:57 +00:00
|
|
|
/* Schedule NAPI, Suppress further interrupts if successful. */
|
2012-12-07 07:04:55 +00:00
|
|
|
if (napi_schedule_prep(&rq->napi)) {
|
2010-04-12 13:19:04 +00:00
|
|
|
virtqueue_disable_cb(rvq);
|
2012-12-07 07:04:55 +00:00
|
|
|
__napi_schedule(&rq->napi);
|
2008-02-05 04:49:57 +00:00
|
|
|
}
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static void virtnet_napi_enable(struct receive_queue *rq)
|
2011-02-10 19:03:31 +00:00
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
napi_enable(&rq->napi);
|
2011-02-10 19:03:31 +00:00
|
|
|
|
|
|
|
/* If all buffers were filled by other side before we napi_enabled, we
|
|
|
|
* won't get another interrupt, so process any outstanding packets
|
|
|
|
* now. virtnet_poll wants re-enable the queue, so we disable here.
|
|
|
|
* We synchronize against interrupts via NAPI_STATE_SCHED */
|
2012-12-07 07:04:55 +00:00
|
|
|
if (napi_schedule_prep(&rq->napi)) {
|
|
|
|
virtqueue_disable_cb(rq->vq);
|
2012-05-16 07:57:12 +00:00
|
|
|
local_bh_disable();
|
2012-12-07 07:04:55 +00:00
|
|
|
__napi_schedule(&rq->napi);
|
2012-05-16 07:57:12 +00:00
|
|
|
local_bh_enable();
|
2011-02-10 19:03:31 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-08-26 19:22:32 +00:00
|
|
|
static void refill_work(struct work_struct *work)
|
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi =
|
|
|
|
container_of(work, struct virtnet_info, refill.work);
|
2009-08-26 19:22:32 +00:00
|
|
|
bool still_empty;
|
2012-12-07 07:04:56 +00:00
|
|
|
int i;
|
|
|
|
|
2013-04-29 02:30:08 +00:00
|
|
|
for (i = 0; i < vi->curr_queue_pairs; i++) {
|
2012-12-07 07:04:56 +00:00
|
|
|
struct receive_queue *rq = &vi->rq[i];
|
2009-08-26 19:22:32 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
napi_disable(&rq->napi);
|
|
|
|
still_empty = !try_fill_recv(rq, GFP_KERNEL);
|
|
|
|
virtnet_napi_enable(rq);
|
2009-08-26 19:22:32 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* In theory, this can happen: if we don't get any buffers in
|
|
|
|
* we will *never* try to fill again.
|
|
|
|
*/
|
|
|
|
if (still_empty)
|
|
|
|
schedule_delayed_work(&vi->refill, HZ/2);
|
|
|
|
}
|
2009-08-26 19:22:32 +00:00
|
|
|
}
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
static int virtnet_poll(struct napi_struct *napi, int budget)
|
|
|
|
{
|
2012-12-07 07:04:55 +00:00
|
|
|
struct receive_queue *rq =
|
|
|
|
container_of(napi, struct receive_queue, napi);
|
|
|
|
struct virtnet_info *vi = rq->vq->vdev->priv;
|
2010-01-29 03:20:04 +00:00
|
|
|
void *buf;
|
virtio_net: fix race in RX VQ processing
virtio net called virtqueue_enable_cq on RX path after napi_complete, so
with NAPI_STATE_SCHED clear - outside the implicit napi lock.
This violates the requirement to synchronize virtqueue_enable_cq wrt
virtqueue_add_buf. In particular, used event can move backwards,
causing us to lose interrupts.
In a debug build, this can trigger panic within START_USE.
Jason Wang reports that he can trigger the races artificially,
by adding udelay() in virtqueue_enable_cb() after virtio_mb().
However, we must call napi_complete to clear NAPI_STATE_SCHED before
polling the virtqueue for used buffers, otherwise napi_schedule_prep in
a callback will fail, causing us to lose RX events.
To fix, call virtqueue_enable_cb_prepare with NAPI_STATE_SCHED
set (under napi lock), later call virtqueue_poll with
NAPI_STATE_SCHED clear (outside the lock).
Reported-by: Jason Wang <jasowang@redhat.com>
Tested-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-07-09 05:13:04 +00:00
|
|
|
unsigned int r, len, received = 0;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
again:
|
|
|
|
while (received < budget &&
|
2012-12-07 07:04:55 +00:00
|
|
|
(buf = virtqueue_get_buf(rq->vq, &len)) != NULL) {
|
|
|
|
receive_buf(rq, buf, len);
|
|
|
|
--rq->num;
|
2007-10-22 01:03:37 +00:00
|
|
|
received++;
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
if (rq->num < rq->max / 2) {
|
|
|
|
if (!try_fill_recv(rq, GFP_ATOMIC))
|
2012-08-20 21:51:24 +00:00
|
|
|
schedule_delayed_work(&vi->refill, 0);
|
2009-08-26 19:22:32 +00:00
|
|
|
}
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2007-11-19 16:20:43 +00:00
|
|
|
/* Out of packets? */
|
|
|
|
if (received < budget) {
|
virtio_net: fix race in RX VQ processing
virtio net called virtqueue_enable_cq on RX path after napi_complete, so
with NAPI_STATE_SCHED clear - outside the implicit napi lock.
This violates the requirement to synchronize virtqueue_enable_cq wrt
virtqueue_add_buf. In particular, used event can move backwards,
causing us to lose interrupts.
In a debug build, this can trigger panic within START_USE.
Jason Wang reports that he can trigger the races artificially,
by adding udelay() in virtqueue_enable_cb() after virtio_mb().
However, we must call napi_complete to clear NAPI_STATE_SCHED before
polling the virtqueue for used buffers, otherwise napi_schedule_prep in
a callback will fail, causing us to lose RX events.
To fix, call virtqueue_enable_cb_prepare with NAPI_STATE_SCHED
set (under napi lock), later call virtqueue_poll with
NAPI_STATE_SCHED clear (outside the lock).
Reported-by: Jason Wang <jasowang@redhat.com>
Tested-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-07-09 05:13:04 +00:00
|
|
|
r = virtqueue_enable_cb_prepare(rq->vq);
|
2009-01-20 00:43:59 +00:00
|
|
|
napi_complete(napi);
|
virtio_net: fix race in RX VQ processing
virtio net called virtqueue_enable_cq on RX path after napi_complete, so
with NAPI_STATE_SCHED clear - outside the implicit napi lock.
This violates the requirement to synchronize virtqueue_enable_cq wrt
virtqueue_add_buf. In particular, used event can move backwards,
causing us to lose interrupts.
In a debug build, this can trigger panic within START_USE.
Jason Wang reports that he can trigger the races artificially,
by adding udelay() in virtqueue_enable_cb() after virtio_mb().
However, we must call napi_complete to clear NAPI_STATE_SCHED before
polling the virtqueue for used buffers, otherwise napi_schedule_prep in
a callback will fail, causing us to lose RX events.
To fix, call virtqueue_enable_cb_prepare with NAPI_STATE_SCHED
set (under napi lock), later call virtqueue_poll with
NAPI_STATE_SCHED clear (outside the lock).
Reported-by: Jason Wang <jasowang@redhat.com>
Tested-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-07-09 05:13:04 +00:00
|
|
|
if (unlikely(virtqueue_poll(rq->vq, r)) &&
|
2009-12-03 07:58:21 +00:00
|
|
|
napi_schedule_prep(napi)) {
|
2012-12-07 07:04:55 +00:00
|
|
|
virtqueue_disable_cb(rq->vq);
|
2009-01-20 00:43:59 +00:00
|
|
|
__napi_schedule(napi);
|
2007-10-22 01:03:37 +00:00
|
|
|
goto again;
|
virtio: fix race in enable_cb
There is a race in virtio_net, dealing with disabling/enabling the callback.
I saw the following oops:
kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:218!
illegal operation: 0001 [#1] SMP
Modules linked in: sunrpc dm_mod
CPU: 2 Not tainted 2.6.25-rc1zlive-host-10623-gd358142-dirty #99
Process swapper (pid: 0, task: 000000000f85a610, ksp: 000000000f873c60)
Krnl PSW : 0404300180000000 00000000002b81a6 (vring_disable_cb+0x16/0x20)
R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
Krnl GPRS: 0000000000000001 0000000000000001 0000000010005800 0000000000000001
000000000f3a0900 000000000f85a610 0000000000000000 0000000000000000
0000000000000000 000000000f870000 0000000000000000 0000000000001237
000000000f3a0920 000000000010ff74 00000000002846f6 000000000fa0bcd8
Krnl Code: 00000000002b819a: a7110001 tmll %r1,1
00000000002b819e: a7840004 brc 8,2b81a6
00000000002b81a2: a7f40001 brc 15,2b81a4
>00000000002b81a6: a51b0001 oill %r1,1
00000000002b81aa: 40102000 sth %r1,0(%r2)
00000000002b81ae: 07fe bcr 15,%r14
00000000002b81b0: eb7ff0380024 stmg %r7,%r15,56(%r15)
00000000002b81b6: a7f13e00 tmll %r15,15872
Call Trace:
([<000000000fa0bcd0>] 0xfa0bcd0)
[<00000000002b8350>] vring_interrupt+0x5c/0x6c
[<000000000010ab08>] do_extint+0xb8/0xf0
[<0000000000110716>] ext_no_vtime+0x16/0x1a
[<0000000000107e72>] cpu_idle+0x1c2/0x1e0
The problem can be triggered with a high amount of host->guest traffic.
I think its the following race:
poll says netif_rx_complete
poll calls enable_cb
enable_cb opens the interrupt mask
a new packet comes, an interrupt is triggered----\
enable_cb sees that there is more work |
enable_cb disables the interrupt |
. V
. interrupt is delivered
. skb_recv_done does atomic napi test, ok
some waiting disable_cb is called->check fails->bang!
.
poll would do napi check
poll would do disable_cb
The fix is to let enable_cb not disable the interrupt again, but expect the
caller to do the cleanup if it returns false. In that case, the interrupt is
only disabled, if the napi test_set_bit was successful.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (cleaned up doco)
2008-03-14 13:17:05 +00:00
|
|
|
}
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return received;
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
static int virtnet_open(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
int i;
|
|
|
|
|
2013-05-21 20:03:58 +00:00
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
if (i < vi->curr_queue_pairs)
|
|
|
|
/* Make sure we have some buffers: if oom use wq. */
|
|
|
|
if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
|
|
|
|
schedule_delayed_work(&vi->refill, 0);
|
2012-12-07 07:04:56 +00:00
|
|
|
virtnet_napi_enable(&vi->rq[i]);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-12-20 16:37:04 +00:00
|
|
|
static void free_old_xmit_skbs(struct send_queue *sq)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
2012-10-16 13:26:14 +00:00
|
|
|
unsigned int len;
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = sq->vq->vdev->priv;
|
2012-02-13 06:53:41 +00:00
|
|
|
struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
while ((skb = virtqueue_get_buf(sq->vq, &len)) != NULL) {
|
2007-10-22 01:03:37 +00:00
|
|
|
pr_debug("Sent skb %p\n", skb);
|
2011-06-15 06:36:29 +00:00
|
|
|
|
2012-06-05 22:35:24 +00:00
|
|
|
u64_stats_update_begin(&stats->tx_syncp);
|
2011-06-15 06:36:29 +00:00
|
|
|
stats->tx_bytes += skb->len;
|
|
|
|
stats->tx_packets++;
|
2012-06-05 22:35:24 +00:00
|
|
|
u64_stats_update_end(&stats->tx_syncp);
|
2011-06-15 06:36:29 +00:00
|
|
|
|
2009-10-14 14:36:43 +00:00
|
|
|
dev_kfree_skb_any(skb);
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2013-07-25 00:50:23 +00:00
|
|
|
struct skb_vnet_hdr *hdr;
|
2007-10-22 01:03:37 +00:00
|
|
|
const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;
|
2012-12-07 07:04:55 +00:00
|
|
|
struct virtnet_info *vi = sq->vq->vdev->priv;
|
2012-10-16 13:26:14 +00:00
|
|
|
unsigned num_sg;
|
2013-07-25 00:50:23 +00:00
|
|
|
unsigned hdr_len;
|
|
|
|
bool can_push;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2008-10-27 22:59:26 +00:00
|
|
|
pr_debug("%s: xmit %p %pM\n", vi->dev->name, skb, dest);
|
2013-07-25 00:50:23 +00:00
|
|
|
if (vi->mergeable_rx_bufs)
|
|
|
|
hdr_len = sizeof hdr->mhdr;
|
|
|
|
else
|
|
|
|
hdr_len = sizeof hdr->hdr;
|
|
|
|
|
|
|
|
can_push = vi->any_header_sg &&
|
|
|
|
!((unsigned long)skb->data & (__alignof__(*hdr) - 1)) &&
|
|
|
|
!skb_header_cloned(skb) && skb_headroom(skb) >= hdr_len;
|
|
|
|
/* Even if we can, don't push here yet as this would skew
|
|
|
|
* csum_start offset below. */
|
|
|
|
if (can_push)
|
|
|
|
hdr = (struct skb_vnet_hdr *)(skb->data - hdr_len);
|
|
|
|
else
|
|
|
|
hdr = skb_vnet_hdr(skb);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
if (skb->ip_summed == CHECKSUM_PARTIAL) {
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
|
2010-12-14 15:24:08 +00:00
|
|
|
hdr->hdr.csum_start = skb_checksum_start_offset(skb);
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.csum_offset = skb->csum_offset;
|
2007-10-22 01:03:37 +00:00
|
|
|
} else {
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.flags = 0;
|
|
|
|
hdr->hdr.csum_offset = hdr->hdr.csum_start = 0;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if (skb_is_gso(skb)) {
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.hdr_len = skb_headlen(skb);
|
|
|
|
hdr->hdr.gso_size = skb_shinfo(skb)->gso_size;
|
2008-02-05 04:50:02 +00:00
|
|
|
if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
|
2007-10-22 01:03:37 +00:00
|
|
|
else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
|
2007-10-22 01:03:37 +00:00
|
|
|
else if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_UDP;
|
2007-10-22 01:03:37 +00:00
|
|
|
else
|
|
|
|
BUG();
|
2008-02-05 04:50:02 +00:00
|
|
|
if (skb_shinfo(skb)->gso_type & SKB_GSO_TCP_ECN)
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
|
2007-10-22 01:03:37 +00:00
|
|
|
} else {
|
2009-09-24 15:59:19 +00:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_NONE;
|
|
|
|
hdr->hdr.gso_size = hdr->hdr.hdr_len = 0;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2008-11-17 06:41:34 +00:00
|
|
|
if (vi->mergeable_rx_bufs)
|
2013-07-25 00:50:23 +00:00
|
|
|
hdr->mhdr.num_buffers = 0;
|
2008-11-17 06:41:34 +00:00
|
|
|
|
2013-07-25 00:50:23 +00:00
|
|
|
if (can_push) {
|
|
|
|
__skb_push(skb, hdr_len);
|
|
|
|
num_sg = skb_to_sgvec(skb, sq->sg, 0, skb->len);
|
|
|
|
/* Pull header back to avoid skew in tx bytes calculations. */
|
|
|
|
__skb_pull(skb, hdr_len);
|
|
|
|
} else {
|
|
|
|
sg_set_buf(sq->sg, hdr, hdr_len);
|
|
|
|
num_sg = skb_to_sgvec(skb, sq->sg + 1, 0, skb->len) + 1;
|
|
|
|
}
|
2013-03-20 05:14:28 +00:00
|
|
|
return virtqueue_add_outbuf(sq->vq, sq->sg, num_sg, skb, GFP_ATOMIC);
|
2008-05-26 07:48:13 +00:00
|
|
|
}
|
|
|
|
|
2009-08-31 19:50:51 +00:00
|
|
|
static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
|
2008-05-03 02:50:46 +00:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2012-12-07 07:04:56 +00:00
|
|
|
int qnum = skb_get_queue_mapping(skb);
|
|
|
|
struct send_queue *sq = &vi->sq[qnum];
|
2012-10-16 13:26:14 +00:00
|
|
|
int err;
|
2008-02-05 04:50:07 +00:00
|
|
|
|
|
|
|
/* Free up any pending old buffers before queueing new ones. */
|
2012-12-07 07:04:55 +00:00
|
|
|
free_old_xmit_skbs(sq);
|
2008-05-03 02:50:46 +00:00
|
|
|
|
2009-10-28 11:03:38 +00:00
|
|
|
/* Try to transmit */
|
2012-12-20 16:37:04 +00:00
|
|
|
err = xmit_skb(sq, skb);
|
2009-09-24 15:59:20 +00:00
|
|
|
|
2012-10-16 13:26:14 +00:00
|
|
|
/* This should not happen! */
|
2012-10-16 13:26:15 +00:00
|
|
|
if (unlikely(err)) {
|
2012-10-16 13:26:14 +00:00
|
|
|
dev->stats.tx_fifo_errors++;
|
|
|
|
if (net_ratelimit())
|
|
|
|
dev_warn(&dev->dev,
|
2012-12-20 16:37:04 +00:00
|
|
|
"Unexpected TXQ (%d) queue failure: %d\n", qnum, err);
|
2010-07-02 16:34:01 +00:00
|
|
|
dev->stats.tx_dropped++;
|
|
|
|
kfree_skb(skb);
|
|
|
|
return NETDEV_TX_OK;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
2012-12-07 07:04:55 +00:00
|
|
|
virtqueue_kick(sq->vq);
|
2009-10-28 11:03:38 +00:00
|
|
|
|
2009-09-24 15:59:20 +00:00
|
|
|
/* Don't wait up for transmitted skbs to be freed. */
|
|
|
|
skb_orphan(skb);
|
|
|
|
nf_reset(skb);
|
|
|
|
|
|
|
|
/* Apparently nice girls don't return TX_BUSY; stop the queue
|
|
|
|
* before it gets out of hand. Naturally, this wastes entries. */
|
2012-12-20 16:37:04 +00:00
|
|
|
if (sq->vq->num_free < 2+MAX_SKB_FRAGS) {
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_stop_subqueue(dev, qnum);
|
2012-12-07 07:04:55 +00:00
|
|
|
if (unlikely(!virtqueue_enable_cb_delayed(sq->vq))) {
|
2009-09-24 15:59:20 +00:00
|
|
|
/* More just got used, free them then recheck. */
|
2012-12-20 16:37:04 +00:00
|
|
|
free_old_xmit_skbs(sq);
|
|
|
|
if (sq->vq->num_free >= 2+MAX_SKB_FRAGS) {
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_start_subqueue(dev, qnum);
|
2012-12-07 07:04:55 +00:00
|
|
|
virtqueue_disable_cb(sq->vq);
|
2009-09-24 15:59:20 +00:00
|
|
|
}
|
|
|
|
}
|
2008-05-03 02:50:46 +00:00
|
|
|
}
|
2009-09-24 15:59:20 +00:00
|
|
|
|
|
|
|
return NETDEV_TX_OK;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2013-01-21 01:17:21 +00:00
|
|
|
/*
|
|
|
|
* Send command via the control virtqueue and check status. Commands
|
|
|
|
* supported by the hypervisor, as indicated by feature bits, should
|
|
|
|
* never fail unless improperly formated.
|
|
|
|
*/
|
|
|
|
static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
|
2013-03-20 05:14:28 +00:00
|
|
|
struct scatterlist *out,
|
|
|
|
struct scatterlist *in)
|
2013-01-21 01:17:21 +00:00
|
|
|
{
|
2013-03-20 05:14:28 +00:00
|
|
|
struct scatterlist *sgs[4], hdr, stat;
|
2013-01-21 01:17:21 +00:00
|
|
|
struct virtio_net_ctrl_hdr ctrl;
|
|
|
|
virtio_net_ctrl_ack status = ~0;
|
2013-03-20 05:14:28 +00:00
|
|
|
unsigned out_num = 0, in_num = 0, tmp;
|
2013-01-21 01:17:21 +00:00
|
|
|
|
|
|
|
/* Caller should know better */
|
2013-03-20 05:14:28 +00:00
|
|
|
BUG_ON(!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ));
|
2013-01-21 01:17:21 +00:00
|
|
|
|
|
|
|
ctrl.class = class;
|
|
|
|
ctrl.cmd = cmd;
|
2013-03-20 05:14:28 +00:00
|
|
|
/* Add header */
|
|
|
|
sg_init_one(&hdr, &ctrl, sizeof(ctrl));
|
|
|
|
sgs[out_num++] = &hdr;
|
2013-01-21 01:17:21 +00:00
|
|
|
|
2013-03-20 05:14:28 +00:00
|
|
|
if (out)
|
|
|
|
sgs[out_num++] = out;
|
|
|
|
if (in)
|
|
|
|
sgs[out_num + in_num++] = in;
|
2013-01-21 01:17:21 +00:00
|
|
|
|
2013-03-20 05:14:28 +00:00
|
|
|
/* Add return status. */
|
|
|
|
sg_init_one(&stat, &status, sizeof(status));
|
|
|
|
sgs[out_num + in_num++] = &stat;
|
2013-01-21 01:17:21 +00:00
|
|
|
|
2013-03-20 05:14:28 +00:00
|
|
|
BUG_ON(out_num + in_num > ARRAY_SIZE(sgs));
|
|
|
|
BUG_ON(virtqueue_add_sgs(vi->cvq, sgs, out_num, in_num, vi, GFP_ATOMIC)
|
|
|
|
< 0);
|
2013-01-21 01:17:21 +00:00
|
|
|
|
|
|
|
virtqueue_kick(vi->cvq);
|
|
|
|
|
|
|
|
/* Spin for a response, the kick causes an ioport write, trapping
|
|
|
|
* into the hypervisor, so the request should be handled immediately.
|
|
|
|
*/
|
|
|
|
while (!virtqueue_get_buf(vi->cvq, &tmp))
|
|
|
|
cpu_relax();
|
|
|
|
|
|
|
|
return status == VIRTIO_NET_OK;
|
|
|
|
}
|
|
|
|
|
2009-02-05 00:36:34 +00:00
|
|
|
static int virtnet_set_mac_address(struct net_device *dev, void *p)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct virtio_device *vdev = vi->vdev;
|
2012-06-29 05:10:06 +00:00
|
|
|
int ret;
|
2013-01-21 01:17:23 +00:00
|
|
|
struct sockaddr *addr = p;
|
|
|
|
struct scatterlist sg;
|
2009-02-05 00:36:34 +00:00
|
|
|
|
2013-01-21 01:17:23 +00:00
|
|
|
ret = eth_prepare_mac_addr_change(dev, p);
|
2012-06-29 05:10:06 +00:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
2009-02-05 00:36:34 +00:00
|
|
|
|
2013-01-21 01:17:23 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_MAC_ADDR)) {
|
|
|
|
sg_init_one(&sg, addr->sa_data, dev->addr_len);
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_MAC,
|
|
|
|
VIRTIO_NET_CTRL_MAC_ADDR_SET,
|
2013-03-20 05:14:28 +00:00
|
|
|
&sg, NULL)) {
|
2013-01-21 01:17:23 +00:00
|
|
|
dev_warn(&vdev->dev,
|
|
|
|
"Failed to set mac address by vq command.\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
} else if (virtio_has_feature(vdev, VIRTIO_NET_F_MAC)) {
|
2009-04-04 23:40:19 +00:00
|
|
|
vdev->config->set(vdev, offsetof(struct virtio_net_config, mac),
|
2013-01-21 01:17:23 +00:00
|
|
|
addr->sa_data, dev->addr_len);
|
|
|
|
}
|
|
|
|
|
|
|
|
eth_commit_mac_addr_change(dev, p);
|
2009-02-05 00:36:34 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-06-15 06:36:29 +00:00
|
|
|
static struct rtnl_link_stats64 *virtnet_stats(struct net_device *dev,
|
|
|
|
struct rtnl_link_stats64 *tot)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
int cpu;
|
|
|
|
unsigned int start;
|
|
|
|
|
|
|
|
for_each_possible_cpu(cpu) {
|
2012-02-13 06:53:41 +00:00
|
|
|
struct virtnet_stats *stats = per_cpu_ptr(vi->stats, cpu);
|
2011-06-15 06:36:29 +00:00
|
|
|
u64 tpackets, tbytes, rpackets, rbytes;
|
|
|
|
|
|
|
|
do {
|
2012-07-21 06:30:50 +00:00
|
|
|
start = u64_stats_fetch_begin_bh(&stats->tx_syncp);
|
2011-06-15 06:36:29 +00:00
|
|
|
tpackets = stats->tx_packets;
|
|
|
|
tbytes = stats->tx_bytes;
|
2012-07-21 06:30:50 +00:00
|
|
|
} while (u64_stats_fetch_retry_bh(&stats->tx_syncp, start));
|
2012-06-05 22:35:24 +00:00
|
|
|
|
|
|
|
do {
|
2012-07-21 06:30:50 +00:00
|
|
|
start = u64_stats_fetch_begin_bh(&stats->rx_syncp);
|
2011-06-15 06:36:29 +00:00
|
|
|
rpackets = stats->rx_packets;
|
|
|
|
rbytes = stats->rx_bytes;
|
2012-07-21 06:30:50 +00:00
|
|
|
} while (u64_stats_fetch_retry_bh(&stats->rx_syncp, start));
|
2011-06-15 06:36:29 +00:00
|
|
|
|
|
|
|
tot->rx_packets += rpackets;
|
|
|
|
tot->tx_packets += tpackets;
|
|
|
|
tot->rx_bytes += rbytes;
|
|
|
|
tot->tx_bytes += tbytes;
|
|
|
|
}
|
|
|
|
|
|
|
|
tot->tx_dropped = dev->stats.tx_dropped;
|
2011-11-21 09:28:17 +00:00
|
|
|
tot->tx_fifo_errors = dev->stats.tx_fifo_errors;
|
2011-06-15 06:36:29 +00:00
|
|
|
tot->rx_dropped = dev->stats.rx_dropped;
|
|
|
|
tot->rx_length_errors = dev->stats.rx_length_errors;
|
|
|
|
tot->rx_frame_errors = dev->stats.rx_frame_errors;
|
|
|
|
|
|
|
|
return tot;
|
|
|
|
}
|
|
|
|
|
2008-02-29 10:54:50 +00:00
|
|
|
#ifdef CONFIG_NET_POLL_CONTROLLER
|
|
|
|
static void virtnet_netpoll(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2012-12-07 07:04:56 +00:00
|
|
|
int i;
|
2008-02-29 10:54:50 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
for (i = 0; i < vi->curr_queue_pairs; i++)
|
|
|
|
napi_schedule(&vi->rq[i].napi);
|
2008-02-29 10:54:50 +00:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
static void virtnet_ack_link_announce(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
rtnl_lock();
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_ANNOUNCE,
|
2013-03-20 05:14:28 +00:00
|
|
|
VIRTIO_NET_CTRL_ANNOUNCE_ACK, NULL, NULL))
|
2012-04-11 20:43:52 +00:00
|
|
|
dev_warn(&vi->dev->dev, "Failed to ack link announce.\n");
|
|
|
|
rtnl_unlock();
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
static int virtnet_set_queues(struct virtnet_info *vi, u16 queue_pairs)
|
|
|
|
{
|
|
|
|
struct scatterlist sg;
|
|
|
|
struct virtio_net_ctrl_mq s;
|
|
|
|
struct net_device *dev = vi->dev;
|
|
|
|
|
|
|
|
if (!vi->has_cvq || !virtio_has_feature(vi->vdev, VIRTIO_NET_F_MQ))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
s.virtqueue_pairs = queue_pairs;
|
|
|
|
sg_init_one(&sg, &s, sizeof(s));
|
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_MQ,
|
2013-03-20 05:14:28 +00:00
|
|
|
VIRTIO_NET_CTRL_MQ_VQ_PAIRS_SET, &sg, NULL)) {
|
2012-12-07 07:04:56 +00:00
|
|
|
dev_warn(&dev->dev, "Fail to set num of queue pairs to %d\n",
|
|
|
|
queue_pairs);
|
|
|
|
return -EINVAL;
|
2013-04-29 02:30:08 +00:00
|
|
|
} else {
|
2012-12-07 07:04:56 +00:00
|
|
|
vi->curr_queue_pairs = queue_pairs;
|
2013-10-15 03:18:59 +00:00
|
|
|
/* virtnet_open() will refill when device is going to up. */
|
|
|
|
if (dev->flags & IFF_UP)
|
|
|
|
schedule_delayed_work(&vi->refill, 0);
|
2013-04-29 02:30:08 +00:00
|
|
|
}
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
static int virtnet_close(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2012-12-07 07:04:56 +00:00
|
|
|
int i;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2011-12-29 00:42:38 +00:00
|
|
|
/* Make sure refill_work doesn't re-enable napi! */
|
|
|
|
cancel_delayed_work_sync(&vi->refill);
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++)
|
|
|
|
napi_disable(&vi->rq[i].napi);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-02-04 09:02:40 +00:00
|
|
|
static void virtnet_set_rx_mode(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2009-02-04 09:02:45 +00:00
|
|
|
struct scatterlist sg[2];
|
2009-02-04 09:02:40 +00:00
|
|
|
u8 promisc, allmulti;
|
2009-02-04 09:02:45 +00:00
|
|
|
struct virtio_net_ctrl_mac *mac_data;
|
2009-05-22 23:22:17 +00:00
|
|
|
struct netdev_hw_addr *ha;
|
2010-01-25 21:36:10 +00:00
|
|
|
int uc_count;
|
2010-02-08 04:30:35 +00:00
|
|
|
int mc_count;
|
2009-02-04 09:02:45 +00:00
|
|
|
void *buf;
|
|
|
|
int i;
|
2009-02-04 09:02:40 +00:00
|
|
|
|
|
|
|
/* We can't dynamicaly set ndo_set_rx_mode, so return gracefully */
|
|
|
|
if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_RX))
|
|
|
|
return;
|
|
|
|
|
2009-02-04 09:02:45 +00:00
|
|
|
promisc = ((dev->flags & IFF_PROMISC) != 0);
|
|
|
|
allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
|
2009-02-04 09:02:40 +00:00
|
|
|
|
2009-05-01 17:27:56 +00:00
|
|
|
sg_init_one(sg, &promisc, sizeof(promisc));
|
2009-02-04 09:02:40 +00:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
|
|
|
|
VIRTIO_NET_CTRL_RX_PROMISC,
|
2013-03-20 05:14:28 +00:00
|
|
|
sg, NULL))
|
2009-02-04 09:02:40 +00:00
|
|
|
dev_warn(&dev->dev, "Failed to %sable promisc mode.\n",
|
|
|
|
promisc ? "en" : "dis");
|
|
|
|
|
2009-05-01 17:27:56 +00:00
|
|
|
sg_init_one(sg, &allmulti, sizeof(allmulti));
|
2009-02-04 09:02:40 +00:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
|
|
|
|
VIRTIO_NET_CTRL_RX_ALLMULTI,
|
2013-03-20 05:14:28 +00:00
|
|
|
sg, NULL))
|
2009-02-04 09:02:40 +00:00
|
|
|
dev_warn(&dev->dev, "Failed to %sable allmulti mode.\n",
|
|
|
|
allmulti ? "en" : "dis");
|
2009-02-04 09:02:45 +00:00
|
|
|
|
2010-01-25 21:36:10 +00:00
|
|
|
uc_count = netdev_uc_count(dev);
|
2010-02-08 04:30:35 +00:00
|
|
|
mc_count = netdev_mc_count(dev);
|
2009-02-04 09:02:45 +00:00
|
|
|
/* MAC filter - use one buffer for both lists */
|
2010-02-08 04:30:35 +00:00
|
|
|
buf = kzalloc(((uc_count + mc_count) * ETH_ALEN) +
|
|
|
|
(2 * sizeof(mac_data->entries)), GFP_ATOMIC);
|
|
|
|
mac_data = buf;
|
2013-02-03 17:28:15 +00:00
|
|
|
if (!buf)
|
2009-02-04 09:02:45 +00:00
|
|
|
return;
|
|
|
|
|
2009-05-01 17:27:56 +00:00
|
|
|
sg_init_table(sg, 2);
|
|
|
|
|
2009-02-04 09:02:45 +00:00
|
|
|
/* Store the unicast list and count in the front of the buffer */
|
2010-01-25 21:36:10 +00:00
|
|
|
mac_data->entries = uc_count;
|
2009-05-22 23:22:17 +00:00
|
|
|
i = 0;
|
2010-01-25 21:36:10 +00:00
|
|
|
netdev_for_each_uc_addr(ha, dev)
|
2009-05-22 23:22:17 +00:00
|
|
|
memcpy(&mac_data->macs[i++][0], ha->addr, ETH_ALEN);
|
2009-02-04 09:02:45 +00:00
|
|
|
|
|
|
|
sg_set_buf(&sg[0], mac_data,
|
2010-01-25 21:36:10 +00:00
|
|
|
sizeof(mac_data->entries) + (uc_count * ETH_ALEN));
|
2009-02-04 09:02:45 +00:00
|
|
|
|
|
|
|
/* multicast list and count fill the end */
|
2010-01-25 21:36:10 +00:00
|
|
|
mac_data = (void *)&mac_data->macs[uc_count][0];
|
2009-02-04 09:02:45 +00:00
|
|
|
|
2010-02-08 04:30:35 +00:00
|
|
|
mac_data->entries = mc_count;
|
2010-02-23 23:17:07 +00:00
|
|
|
i = 0;
|
2010-04-01 21:22:57 +00:00
|
|
|
netdev_for_each_mc_addr(ha, dev)
|
|
|
|
memcpy(&mac_data->macs[i++][0], ha->addr, ETH_ALEN);
|
2009-02-04 09:02:45 +00:00
|
|
|
|
|
|
|
sg_set_buf(&sg[1], mac_data,
|
2010-02-08 04:30:35 +00:00
|
|
|
sizeof(mac_data->entries) + (mc_count * ETH_ALEN));
|
2009-02-04 09:02:45 +00:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_MAC,
|
|
|
|
VIRTIO_NET_CTRL_MAC_TABLE_SET,
|
2013-03-20 05:14:28 +00:00
|
|
|
sg, NULL))
|
2009-02-04 09:02:45 +00:00
|
|
|
dev_warn(&dev->dev, "Failed to set MAC fitler table.\n");
|
|
|
|
|
|
|
|
kfree(buf);
|
2009-02-04 09:02:40 +00:00
|
|
|
}
|
|
|
|
|
2013-04-19 02:04:28 +00:00
|
|
|
static int virtnet_vlan_rx_add_vid(struct net_device *dev,
|
|
|
|
__be16 proto, u16 vid)
|
2009-02-04 09:02:50 +00:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct scatterlist sg;
|
|
|
|
|
2009-05-01 17:27:56 +00:00
|
|
|
sg_init_one(&sg, &vid, sizeof(vid));
|
2009-02-04 09:02:50 +00:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
|
2013-03-20 05:14:28 +00:00
|
|
|
VIRTIO_NET_CTRL_VLAN_ADD, &sg, NULL))
|
2009-02-04 09:02:50 +00:00
|
|
|
dev_warn(&dev->dev, "Failed to add VLAN ID %d.\n", vid);
|
2011-12-09 00:52:37 +00:00
|
|
|
return 0;
|
2009-02-04 09:02:50 +00:00
|
|
|
}
|
|
|
|
|
2013-04-19 02:04:28 +00:00
|
|
|
static int virtnet_vlan_rx_kill_vid(struct net_device *dev,
|
|
|
|
__be16 proto, u16 vid)
|
2009-02-04 09:02:50 +00:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct scatterlist sg;
|
|
|
|
|
2009-05-01 17:27:56 +00:00
|
|
|
sg_init_one(&sg, &vid, sizeof(vid));
|
2009-02-04 09:02:50 +00:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
|
2013-03-20 05:14:28 +00:00
|
|
|
VIRTIO_NET_CTRL_VLAN_DEL, &sg, NULL))
|
2009-02-04 09:02:50 +00:00
|
|
|
dev_warn(&dev->dev, "Failed to kill VLAN ID %d.\n", vid);
|
2011-12-09 00:52:37 +00:00
|
|
|
return 0;
|
2009-02-04 09:02:50 +00:00
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
static void virtnet_clean_affinity(struct virtnet_info *vi, long hcpu)
|
2012-12-07 07:04:56 +00:00
|
|
|
{
|
|
|
|
int i;
|
2013-01-24 23:51:29 +00:00
|
|
|
int cpu;
|
2012-12-07 07:04:56 +00:00
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
if (vi->affinity_hint_set) {
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
2013-01-24 23:51:29 +00:00
|
|
|
virtqueue_set_affinity(vi->rq[i].vq, -1);
|
|
|
|
virtqueue_set_affinity(vi->sq[i].vq, -1);
|
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
vi->affinity_hint_set = false;
|
|
|
|
}
|
|
|
|
|
|
|
|
i = 0;
|
|
|
|
for_each_online_cpu(cpu) {
|
|
|
|
if (cpu == hcpu) {
|
|
|
|
*per_cpu_ptr(vi->vq_index, cpu) = -1;
|
|
|
|
} else {
|
2013-01-24 23:51:29 +00:00
|
|
|
*per_cpu_ptr(vi->vq_index, cpu) =
|
|
|
|
++i % vi->curr_queue_pairs;
|
2013-01-24 23:51:30 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2013-01-24 23:51:29 +00:00
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
static void virtnet_set_affinity(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
int cpu;
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
/* In multiqueue mode, when the number of cpu is equal to the number of
|
|
|
|
* queue pairs, we let the queue pairs to be private to one cpu by
|
|
|
|
* setting the affinity hint to eliminate the contention.
|
|
|
|
*/
|
2013-01-24 23:51:30 +00:00
|
|
|
if (vi->curr_queue_pairs == 1 ||
|
|
|
|
vi->max_queue_pairs != num_online_cpus()) {
|
|
|
|
virtnet_clean_affinity(vi, -1);
|
|
|
|
return;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
i = 0;
|
|
|
|
for_each_online_cpu(cpu) {
|
2012-12-07 07:04:56 +00:00
|
|
|
virtqueue_set_affinity(vi->rq[i].vq, cpu);
|
|
|
|
virtqueue_set_affinity(vi->sq[i].vq, cpu);
|
2013-01-24 23:51:30 +00:00
|
|
|
*per_cpu_ptr(vi->vq_index, cpu) = i;
|
|
|
|
i++;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
vi->affinity_hint_set = true;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:31 +00:00
|
|
|
static int virtnet_cpu_callback(struct notifier_block *nfb,
|
|
|
|
unsigned long action, void *hcpu)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = container_of(nfb, struct virtnet_info, nb);
|
|
|
|
|
2013-10-15 03:18:58 +00:00
|
|
|
mutex_lock(&vi->config_lock);
|
|
|
|
|
|
|
|
if (!vi->config_enable)
|
|
|
|
goto done;
|
|
|
|
|
2013-01-24 23:51:31 +00:00
|
|
|
switch(action & ~CPU_TASKS_FROZEN) {
|
|
|
|
case CPU_ONLINE:
|
|
|
|
case CPU_DOWN_FAILED:
|
|
|
|
case CPU_DEAD:
|
|
|
|
virtnet_set_affinity(vi);
|
|
|
|
break;
|
|
|
|
case CPU_DOWN_PREPARE:
|
|
|
|
virtnet_clean_affinity(vi, (long)hcpu);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2013-10-15 03:18:58 +00:00
|
|
|
|
|
|
|
done:
|
|
|
|
mutex_unlock(&vi->config_lock);
|
2013-01-24 23:51:31 +00:00
|
|
|
return NOTIFY_OK;
|
2012-12-07 07:04:56 +00:00
|
|
|
}
|
|
|
|
|
2011-10-19 08:10:59 +00:00
|
|
|
static void virtnet_get_ringparam(struct net_device *dev,
|
|
|
|
struct ethtool_ringparam *ring)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
ring->rx_max_pending = virtqueue_get_vring_size(vi->rq[0].vq);
|
|
|
|
ring->tx_max_pending = virtqueue_get_vring_size(vi->sq[0].vq);
|
2011-10-19 08:10:59 +00:00
|
|
|
ring->rx_pending = ring->rx_max_pending;
|
|
|
|
ring->tx_pending = ring->tx_max_pending;
|
|
|
|
}
|
|
|
|
|
2011-11-14 14:17:08 +00:00
|
|
|
|
|
|
|
static void virtnet_get_drvinfo(struct net_device *dev,
|
|
|
|
struct ethtool_drvinfo *info)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct virtio_device *vdev = vi->vdev;
|
|
|
|
|
|
|
|
strlcpy(info->driver, KBUILD_MODNAME, sizeof(info->driver));
|
|
|
|
strlcpy(info->version, VIRTNET_DRIVER_VERSION, sizeof(info->version));
|
|
|
|
strlcpy(info->bus_info, virtio_bus_name(vdev), sizeof(info->bus_info));
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:57 +00:00
|
|
|
/* TODO: Eliminate OOO packets during switching */
|
|
|
|
static int virtnet_set_channels(struct net_device *dev,
|
|
|
|
struct ethtool_channels *channels)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
u16 queue_pairs = channels->combined_count;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
/* We don't support separate rx/tx channels.
|
|
|
|
* We don't allow setting 'other' channels.
|
|
|
|
*/
|
|
|
|
if (channels->rx_count || channels->tx_count || channels->other_count)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (queue_pairs > vi->max_queue_pairs)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2013-01-24 23:51:29 +00:00
|
|
|
get_online_cpus();
|
2012-12-07 07:04:57 +00:00
|
|
|
err = virtnet_set_queues(vi, queue_pairs);
|
|
|
|
if (!err) {
|
|
|
|
netif_set_real_num_tx_queues(dev, queue_pairs);
|
|
|
|
netif_set_real_num_rx_queues(dev, queue_pairs);
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
virtnet_set_affinity(vi);
|
2012-12-07 07:04:57 +00:00
|
|
|
}
|
2013-01-24 23:51:29 +00:00
|
|
|
put_online_cpus();
|
2012-12-07 07:04:57 +00:00
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void virtnet_get_channels(struct net_device *dev,
|
|
|
|
struct ethtool_channels *channels)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
|
|
|
channels->combined_count = vi->curr_queue_pairs;
|
|
|
|
channels->max_combined = vi->max_queue_pairs;
|
|
|
|
channels->max_other = 0;
|
|
|
|
channels->rx_count = 0;
|
|
|
|
channels->tx_count = 0;
|
|
|
|
channels->other_count = 0;
|
|
|
|
}
|
|
|
|
|
2009-09-02 08:03:33 +00:00
|
|
|
static const struct ethtool_ops virtnet_ethtool_ops = {
|
2011-11-14 14:17:08 +00:00
|
|
|
.get_drvinfo = virtnet_get_drvinfo,
|
2009-01-20 01:09:49 +00:00
|
|
|
.get_link = ethtool_op_get_link,
|
2011-10-19 08:10:59 +00:00
|
|
|
.get_ringparam = virtnet_get_ringparam,
|
2012-12-07 07:04:57 +00:00
|
|
|
.set_channels = virtnet_set_channels,
|
|
|
|
.get_channels = virtnet_get_channels,
|
2008-04-18 03:21:42 +00:00
|
|
|
};
|
|
|
|
|
2008-11-26 13:58:11 +00:00
|
|
|
#define MIN_MTU 68
|
|
|
|
#define MAX_MTU 65535
|
|
|
|
|
|
|
|
static int virtnet_change_mtu(struct net_device *dev, int new_mtu)
|
|
|
|
{
|
|
|
|
if (new_mtu < MIN_MTU || new_mtu > MAX_MTU)
|
|
|
|
return -EINVAL;
|
|
|
|
dev->mtu = new_mtu;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* To avoid contending a lock hold by a vcpu who would exit to host, select the
|
|
|
|
* txq based on the processor id.
|
|
|
|
*/
|
|
|
|
static u16 virtnet_select_queue(struct net_device *dev, struct sk_buff *skb)
|
|
|
|
{
|
2013-01-24 23:51:29 +00:00
|
|
|
int txq;
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
|
|
|
if (skb_rx_queue_recorded(skb)) {
|
|
|
|
txq = skb_get_rx_queue(skb);
|
|
|
|
} else {
|
|
|
|
txq = *__this_cpu_ptr(vi->vq_index);
|
|
|
|
if (txq == -1)
|
|
|
|
txq = 0;
|
|
|
|
}
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
while (unlikely(txq >= dev->real_num_tx_queues))
|
|
|
|
txq -= dev->real_num_tx_queues;
|
|
|
|
|
|
|
|
return txq;
|
|
|
|
}
|
|
|
|
|
2009-01-06 18:44:22 +00:00
|
|
|
static const struct net_device_ops virtnet_netdev = {
|
|
|
|
.ndo_open = virtnet_open,
|
|
|
|
.ndo_stop = virtnet_close,
|
|
|
|
.ndo_start_xmit = start_xmit,
|
|
|
|
.ndo_validate_addr = eth_validate_addr,
|
2009-02-05 00:36:34 +00:00
|
|
|
.ndo_set_mac_address = virtnet_set_mac_address,
|
2009-02-04 09:02:40 +00:00
|
|
|
.ndo_set_rx_mode = virtnet_set_rx_mode,
|
2009-01-06 18:44:22 +00:00
|
|
|
.ndo_change_mtu = virtnet_change_mtu,
|
2011-06-15 06:36:29 +00:00
|
|
|
.ndo_get_stats64 = virtnet_stats,
|
2009-05-01 17:31:10 +00:00
|
|
|
.ndo_vlan_rx_add_vid = virtnet_vlan_rx_add_vid,
|
|
|
|
.ndo_vlan_rx_kill_vid = virtnet_vlan_rx_kill_vid,
|
2012-12-07 07:04:56 +00:00
|
|
|
.ndo_select_queue = virtnet_select_queue,
|
2009-01-06 18:44:22 +00:00
|
|
|
#ifdef CONFIG_NET_POLL_CONTROLLER
|
|
|
|
.ndo_poll_controller = virtnet_netpoll,
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
static void virtnet_config_changed_work(struct work_struct *work)
|
2009-01-20 01:09:49 +00:00
|
|
|
{
|
2012-04-11 20:43:52 +00:00
|
|
|
struct virtnet_info *vi =
|
|
|
|
container_of(work, struct virtnet_info, config_work);
|
2009-01-20 01:09:49 +00:00
|
|
|
u16 v;
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
mutex_lock(&vi->config_lock);
|
|
|
|
if (!vi->config_enable)
|
|
|
|
goto done;
|
|
|
|
|
2011-08-14 14:52:33 +00:00
|
|
|
if (virtio_config_val(vi->vdev, VIRTIO_NET_F_STATUS,
|
2009-01-20 01:09:49 +00:00
|
|
|
offsetof(struct virtio_net_config, status),
|
2011-08-14 14:52:33 +00:00
|
|
|
&v) < 0)
|
2012-04-11 20:43:52 +00:00
|
|
|
goto done;
|
|
|
|
|
|
|
|
if (v & VIRTIO_NET_S_ANNOUNCE) {
|
2012-08-09 22:14:56 +00:00
|
|
|
netdev_notify_peers(vi->dev);
|
2012-04-11 20:43:52 +00:00
|
|
|
virtnet_ack_link_announce(vi);
|
|
|
|
}
|
2009-01-20 01:09:49 +00:00
|
|
|
|
|
|
|
/* Ignore unknown (future) status bits */
|
|
|
|
v &= VIRTIO_NET_S_LINK_UP;
|
|
|
|
|
|
|
|
if (vi->status == v)
|
2012-04-11 20:43:52 +00:00
|
|
|
goto done;
|
2009-01-20 01:09:49 +00:00
|
|
|
|
|
|
|
vi->status = v;
|
|
|
|
|
|
|
|
if (vi->status & VIRTIO_NET_S_LINK_UP) {
|
|
|
|
netif_carrier_on(vi->dev);
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_tx_wake_all_queues(vi->dev);
|
2009-01-20 01:09:49 +00:00
|
|
|
} else {
|
|
|
|
netif_carrier_off(vi->dev);
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_tx_stop_all_queues(vi->dev);
|
2009-01-20 01:09:49 +00:00
|
|
|
}
|
2012-04-11 20:43:52 +00:00
|
|
|
done:
|
|
|
|
mutex_unlock(&vi->config_lock);
|
2009-01-20 01:09:49 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void virtnet_config_changed(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = vdev->priv;
|
|
|
|
|
2012-08-20 21:51:24 +00:00
|
|
|
schedule_work(&vi->config_work);
|
2009-01-20 01:09:49 +00:00
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
static void virtnet_free_queues(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
kfree(vi->rq);
|
|
|
|
kfree(vi->sq);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void free_receive_bufs(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
while (vi->rq[i].pages)
|
|
|
|
__free_pages(get_a_page(&vi->rq[i], GFP_KERNEL), 0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void free_unused_bufs(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
void *buf;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
struct virtqueue *vq = vi->sq[i].vq;
|
|
|
|
while ((buf = virtqueue_detach_unused_buf(vq)) != NULL)
|
|
|
|
dev_kfree_skb(buf);
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
struct virtqueue *vq = vi->rq[i].vq;
|
|
|
|
|
|
|
|
while ((buf = virtqueue_detach_unused_buf(vq)) != NULL) {
|
|
|
|
if (vi->mergeable_rx_bufs || vi->big_packets)
|
|
|
|
give_pages(&vi->rq[i], buf);
|
|
|
|
else
|
|
|
|
dev_kfree_skb(buf);
|
|
|
|
--vi->rq[i].num;
|
|
|
|
}
|
|
|
|
BUG_ON(vi->rq[i].num != 0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
static void virtnet_del_vqs(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
struct virtio_device *vdev = vi->vdev;
|
|
|
|
|
2013-01-24 23:51:30 +00:00
|
|
|
virtnet_clean_affinity(vi, -1);
|
2012-12-07 07:04:56 +00:00
|
|
|
|
2012-12-07 07:04:55 +00:00
|
|
|
vdev->config->del_vqs(vdev);
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
virtnet_free_queues(vi);
|
2012-12-07 07:04:55 +00:00
|
|
|
}
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
static int virtnet_find_vqs(struct virtnet_info *vi)
|
2011-12-22 11:28:31 +00:00
|
|
|
{
|
2012-12-07 07:04:56 +00:00
|
|
|
vq_callback_t **callbacks;
|
|
|
|
struct virtqueue **vqs;
|
|
|
|
int ret = -ENOMEM;
|
|
|
|
int i, total_vqs;
|
|
|
|
const char **names;
|
|
|
|
|
|
|
|
/* We expect 1 RX virtqueue followed by 1 TX virtqueue, followed by
|
|
|
|
* possible N-1 RX/TX queue pairs used in multiqueue mode, followed by
|
|
|
|
* possible control vq.
|
|
|
|
*/
|
|
|
|
total_vqs = vi->max_queue_pairs * 2 +
|
|
|
|
virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ);
|
|
|
|
|
|
|
|
/* Allocate space for find_vqs parameters */
|
|
|
|
vqs = kzalloc(total_vqs * sizeof(*vqs), GFP_KERNEL);
|
|
|
|
if (!vqs)
|
|
|
|
goto err_vq;
|
|
|
|
callbacks = kmalloc(total_vqs * sizeof(*callbacks), GFP_KERNEL);
|
|
|
|
if (!callbacks)
|
|
|
|
goto err_callback;
|
|
|
|
names = kmalloc(total_vqs * sizeof(*names), GFP_KERNEL);
|
|
|
|
if (!names)
|
|
|
|
goto err_names;
|
|
|
|
|
|
|
|
/* Parameters for control virtqueue, if any */
|
|
|
|
if (vi->has_cvq) {
|
|
|
|
callbacks[total_vqs - 1] = NULL;
|
|
|
|
names[total_vqs - 1] = "control";
|
|
|
|
}
|
2011-12-22 11:28:31 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
/* Allocate/initialize parameters for send/receive virtqueues */
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
callbacks[rxq2vq(i)] = skb_recv_done;
|
|
|
|
callbacks[txq2vq(i)] = skb_xmit_done;
|
|
|
|
sprintf(vi->rq[i].name, "input.%d", i);
|
|
|
|
sprintf(vi->sq[i].name, "output.%d", i);
|
|
|
|
names[rxq2vq(i)] = vi->rq[i].name;
|
|
|
|
names[txq2vq(i)] = vi->sq[i].name;
|
|
|
|
}
|
2011-12-22 11:28:31 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
ret = vi->vdev->config->find_vqs(vi->vdev, total_vqs, vqs, callbacks,
|
|
|
|
names);
|
|
|
|
if (ret)
|
|
|
|
goto err_find;
|
2011-12-22 11:28:31 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
if (vi->has_cvq) {
|
|
|
|
vi->cvq = vqs[total_vqs - 1];
|
2011-12-22 11:28:31 +00:00
|
|
|
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN))
|
2013-04-19 02:04:27 +00:00
|
|
|
vi->dev->features |= NETIF_F_HW_VLAN_CTAG_FILTER;
|
2011-12-22 11:28:31 +00:00
|
|
|
}
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
vi->rq[i].vq = vqs[rxq2vq(i)];
|
|
|
|
vi->sq[i].vq = vqs[txq2vq(i)];
|
|
|
|
}
|
|
|
|
|
|
|
|
kfree(names);
|
|
|
|
kfree(callbacks);
|
|
|
|
kfree(vqs);
|
|
|
|
|
2011-12-22 11:28:31 +00:00
|
|
|
return 0;
|
2012-12-07 07:04:56 +00:00
|
|
|
|
|
|
|
err_find:
|
|
|
|
kfree(names);
|
|
|
|
err_names:
|
|
|
|
kfree(callbacks);
|
|
|
|
err_callback:
|
|
|
|
kfree(vqs);
|
|
|
|
err_vq:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int virtnet_alloc_queues(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
vi->sq = kzalloc(sizeof(*vi->sq) * vi->max_queue_pairs, GFP_KERNEL);
|
|
|
|
if (!vi->sq)
|
|
|
|
goto err_sq;
|
|
|
|
vi->rq = kzalloc(sizeof(*vi->rq) * vi->max_queue_pairs, GFP_KERNEL);
|
2012-12-10 02:24:08 +00:00
|
|
|
if (!vi->rq)
|
2012-12-07 07:04:56 +00:00
|
|
|
goto err_rq;
|
|
|
|
|
|
|
|
INIT_DELAYED_WORK(&vi->refill, refill_work);
|
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
vi->rq[i].pages = NULL;
|
|
|
|
netif_napi_add(vi->dev, &vi->rq[i].napi, virtnet_poll,
|
|
|
|
napi_weight);
|
|
|
|
|
|
|
|
sg_init_table(vi->rq[i].sg, ARRAY_SIZE(vi->rq[i].sg));
|
|
|
|
sg_init_table(vi->sq[i].sg, ARRAY_SIZE(vi->sq[i].sg));
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_rq:
|
|
|
|
kfree(vi->sq);
|
|
|
|
err_sq:
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int init_vqs(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Allocate send & receive queues */
|
|
|
|
ret = virtnet_alloc_queues(vi);
|
|
|
|
if (ret)
|
|
|
|
goto err;
|
|
|
|
|
|
|
|
ret = virtnet_find_vqs(vi);
|
|
|
|
if (ret)
|
|
|
|
goto err_free;
|
|
|
|
|
2013-01-24 23:51:29 +00:00
|
|
|
get_online_cpus();
|
2013-01-24 23:51:30 +00:00
|
|
|
virtnet_set_affinity(vi);
|
2013-01-24 23:51:29 +00:00
|
|
|
put_online_cpus();
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_free:
|
|
|
|
virtnet_free_queues(vi);
|
|
|
|
err:
|
|
|
|
return ret;
|
2011-12-22 11:28:31 +00:00
|
|
|
}
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
static int virtnet_probe(struct virtio_device *vdev)
|
|
|
|
{
|
2012-12-07 07:04:56 +00:00
|
|
|
int i, err;
|
2007-10-22 01:03:37 +00:00
|
|
|
struct net_device *dev;
|
|
|
|
struct virtnet_info *vi;
|
2012-12-07 07:04:56 +00:00
|
|
|
u16 max_queue_pairs;
|
|
|
|
|
|
|
|
/* Find if host supports multiqueue virtio_net device */
|
|
|
|
err = virtio_config_val(vdev, VIRTIO_NET_F_MQ,
|
|
|
|
offsetof(struct virtio_net_config,
|
|
|
|
max_virtqueue_pairs), &max_queue_pairs);
|
|
|
|
|
|
|
|
/* We need at least 2 queue's */
|
|
|
|
if (err || max_queue_pairs < VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MIN ||
|
|
|
|
max_queue_pairs > VIRTIO_NET_CTRL_MQ_VQ_PAIRS_MAX ||
|
|
|
|
!virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
|
|
|
|
max_queue_pairs = 1;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
/* Allocate ourselves a network device with room for our info */
|
2012-12-07 07:04:56 +00:00
|
|
|
dev = alloc_etherdev_mq(sizeof(struct virtnet_info), max_queue_pairs);
|
2007-10-22 01:03:37 +00:00
|
|
|
if (!dev)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
/* Set up network device as normal. */
|
2012-06-29 05:10:06 +00:00
|
|
|
dev->priv_flags |= IFF_UNICAST_FLT | IFF_LIVE_ADDR_CHANGE;
|
2009-01-06 18:44:22 +00:00
|
|
|
dev->netdev_ops = &virtnet_netdev;
|
2007-10-22 01:03:37 +00:00
|
|
|
dev->features = NETIF_F_HIGHDMA;
|
2011-06-15 06:36:29 +00:00
|
|
|
|
2008-04-18 03:21:42 +00:00
|
|
|
SET_ETHTOOL_OPS(dev, &virtnet_ethtool_ops);
|
2007-10-22 01:03:37 +00:00
|
|
|
SET_NETDEV_DEV(dev, &vdev->dev);
|
|
|
|
|
|
|
|
/* Do we support "hardware" checksums? */
|
2011-03-31 01:01:35 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_CSUM)) {
|
2007-10-22 01:03:37 +00:00
|
|
|
/* This opens up the world of extra features. */
|
2011-03-31 01:01:35 +00:00
|
|
|
dev->hw_features |= NETIF_F_HW_CSUM|NETIF_F_SG|NETIF_F_FRAGLIST;
|
|
|
|
if (csum)
|
|
|
|
dev->features |= NETIF_F_HW_CSUM|NETIF_F_SG|NETIF_F_FRAGLIST;
|
|
|
|
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_GSO)) {
|
|
|
|
dev->hw_features |= NETIF_F_TSO | NETIF_F_UFO
|
2008-02-05 04:50:02 +00:00
|
|
|
| NETIF_F_TSO_ECN | NETIF_F_TSO6;
|
|
|
|
}
|
2008-05-03 02:50:46 +00:00
|
|
|
/* Individual feature bits: what can host handle? */
|
2011-03-31 01:01:35 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO4))
|
|
|
|
dev->hw_features |= NETIF_F_TSO;
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO6))
|
|
|
|
dev->hw_features |= NETIF_F_TSO6;
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_HOST_ECN))
|
|
|
|
dev->hw_features |= NETIF_F_TSO_ECN;
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_HOST_UFO))
|
|
|
|
dev->hw_features |= NETIF_F_UFO;
|
|
|
|
|
|
|
|
if (gso)
|
|
|
|
dev->features |= dev->hw_features & (NETIF_F_ALL_TSO|NETIF_F_UFO);
|
|
|
|
/* (!csum && gso) case will be fixed by register_netdev() */
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
2013-08-27 15:09:02 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_CSUM))
|
|
|
|
dev->features |= NETIF_F_RXCSUM;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2013-04-10 23:32:21 +00:00
|
|
|
dev->vlan_features = dev->features;
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
/* Configuration may specify what MAC to use. Otherwise random. */
|
2011-08-14 14:52:33 +00:00
|
|
|
if (virtio_config_val_len(vdev, VIRTIO_NET_F_MAC,
|
2008-02-05 04:49:56 +00:00
|
|
|
offsetof(struct virtio_net_config, mac),
|
2011-08-14 14:52:33 +00:00
|
|
|
dev->dev_addr, dev->addr_len) < 0)
|
2012-02-15 06:45:39 +00:00
|
|
|
eth_hw_addr_random(dev);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
/* Set up our device-specific information */
|
|
|
|
vi = netdev_priv(dev);
|
|
|
|
vi->dev = dev;
|
|
|
|
vi->vdev = vdev;
|
2008-02-18 09:02:51 +00:00
|
|
|
vdev->priv = vi;
|
2011-06-15 06:36:29 +00:00
|
|
|
vi->stats = alloc_percpu(struct virtnet_stats);
|
|
|
|
err = -ENOMEM;
|
|
|
|
if (vi->stats == NULL)
|
|
|
|
goto free;
|
|
|
|
|
2013-01-24 23:51:29 +00:00
|
|
|
vi->vq_index = alloc_percpu(int);
|
|
|
|
if (vi->vq_index == NULL)
|
|
|
|
goto free_stats;
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
mutex_init(&vi->config_lock);
|
|
|
|
vi->config_enable = true;
|
|
|
|
INIT_WORK(&vi->config_work, virtnet_config_changed_work);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2008-04-18 03:24:27 +00:00
|
|
|
/* If we can receive ANY GSO packets, we must allocate large ones. */
|
2009-12-03 07:58:21 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
|
|
|
|
virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO6) ||
|
|
|
|
virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_ECN))
|
2008-04-18 03:24:27 +00:00
|
|
|
vi->big_packets = true;
|
|
|
|
|
2008-11-17 06:41:34 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
|
|
|
|
vi->mergeable_rx_bufs = true;
|
|
|
|
|
2013-07-25 00:50:23 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_F_ANY_LAYOUT))
|
|
|
|
vi->any_header_sg = true;
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
|
|
|
|
vi->has_cvq = true;
|
|
|
|
|
|
|
|
/* Use single tx/rx queue pair as default */
|
|
|
|
vi->curr_queue_pairs = 1;
|
|
|
|
vi->max_queue_pairs = max_queue_pairs;
|
|
|
|
|
|
|
|
/* Allocate/initialize the rx/tx queues, and invoke find_vqs */
|
2011-12-22 11:28:31 +00:00
|
|
|
err = init_vqs(vi);
|
2009-06-13 04:16:36 +00:00
|
|
|
if (err)
|
2013-01-24 23:51:29 +00:00
|
|
|
goto free_index;
|
2007-10-22 01:03:37 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
netif_set_real_num_tx_queues(dev, 1);
|
|
|
|
netif_set_real_num_rx_queues(dev, 1);
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
err = register_netdev(dev);
|
|
|
|
if (err) {
|
|
|
|
pr_debug("virtio_net: registering device failed\n");
|
2009-06-13 04:16:36 +00:00
|
|
|
goto free_vqs;
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
2008-02-05 04:50:02 +00:00
|
|
|
|
|
|
|
/* Last of all, set up some receive buffers. */
|
2013-04-29 02:30:08 +00:00
|
|
|
for (i = 0; i < vi->curr_queue_pairs; i++) {
|
2012-12-07 07:04:56 +00:00
|
|
|
try_fill_recv(&vi->rq[i], GFP_KERNEL);
|
|
|
|
|
|
|
|
/* If we didn't even get one input buffer, we're useless. */
|
|
|
|
if (vi->rq[i].num == 0) {
|
|
|
|
free_unused_bufs(vi);
|
|
|
|
err = -ENOMEM;
|
|
|
|
goto free_recv_bufs;
|
|
|
|
}
|
2008-02-05 04:50:02 +00:00
|
|
|
}
|
|
|
|
|
2013-01-24 23:51:31 +00:00
|
|
|
vi->nb.notifier_call = &virtnet_cpu_callback;
|
|
|
|
err = register_hotcpu_notifier(&vi->nb);
|
|
|
|
if (err) {
|
|
|
|
pr_debug("virtio_net: registering cpu notifier failed\n");
|
|
|
|
goto free_recv_bufs;
|
|
|
|
}
|
|
|
|
|
2010-11-10 14:45:41 +00:00
|
|
|
/* Assume link up if device can't report link status,
|
|
|
|
otherwise get link status from config. */
|
|
|
|
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_STATUS)) {
|
|
|
|
netif_carrier_off(dev);
|
2012-08-20 21:51:24 +00:00
|
|
|
schedule_work(&vi->config_work);
|
2010-11-10 14:45:41 +00:00
|
|
|
} else {
|
|
|
|
vi->status = VIRTIO_NET_S_LINK_UP;
|
|
|
|
netif_carrier_on(dev);
|
|
|
|
}
|
2009-01-20 01:09:49 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
pr_debug("virtnet: registered device %s with %d RX and TX vq's\n",
|
|
|
|
dev->name, max_queue_pairs);
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
return 0;
|
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
free_recv_bufs:
|
|
|
|
free_receive_bufs(vi);
|
2008-02-05 04:50:02 +00:00
|
|
|
unregister_netdev(dev);
|
2009-06-13 04:16:36 +00:00
|
|
|
free_vqs:
|
2012-12-07 07:04:56 +00:00
|
|
|
cancel_delayed_work_sync(&vi->refill);
|
2012-12-07 07:04:55 +00:00
|
|
|
virtnet_del_vqs(vi);
|
2013-01-24 23:51:29 +00:00
|
|
|
free_index:
|
|
|
|
free_percpu(vi->vq_index);
|
2011-06-15 06:36:29 +00:00
|
|
|
free_stats:
|
|
|
|
free_percpu(vi->stats);
|
2007-10-22 01:03:37 +00:00
|
|
|
free:
|
|
|
|
free_netdev(dev);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-12-22 11:28:32 +00:00
|
|
|
static void remove_vq_common(struct virtnet_info *vi)
|
2007-10-22 01:03:37 +00:00
|
|
|
{
|
2011-12-22 11:28:32 +00:00
|
|
|
vi->vdev->config->reset(vi->vdev);
|
2010-02-08 14:14:42 +00:00
|
|
|
|
|
|
|
/* Free unused buffers in both send and recv, if any. */
|
2010-01-29 03:20:04 +00:00
|
|
|
free_unused_bufs(vi);
|
2008-07-25 17:06:01 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
free_receive_bufs(vi);
|
2009-06-13 04:16:36 +00:00
|
|
|
|
2012-12-07 07:04:56 +00:00
|
|
|
virtnet_del_vqs(vi);
|
2011-12-22 11:28:32 +00:00
|
|
|
}
|
|
|
|
|
2012-12-03 14:24:15 +00:00
|
|
|
static void virtnet_remove(struct virtio_device *vdev)
|
2011-12-22 11:28:32 +00:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = vdev->priv;
|
|
|
|
|
2013-01-24 23:51:31 +00:00
|
|
|
unregister_hotcpu_notifier(&vi->nb);
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
/* Prevent config work handler from accessing the device. */
|
|
|
|
mutex_lock(&vi->config_lock);
|
|
|
|
vi->config_enable = false;
|
|
|
|
mutex_unlock(&vi->config_lock);
|
|
|
|
|
2011-12-22 11:28:32 +00:00
|
|
|
unregister_netdev(vi->dev);
|
|
|
|
|
|
|
|
remove_vq_common(vi);
|
2008-07-25 17:06:01 +00:00
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
flush_work(&vi->config_work);
|
|
|
|
|
2013-01-24 23:51:29 +00:00
|
|
|
free_percpu(vi->vq_index);
|
2011-07-20 03:56:02 +00:00
|
|
|
free_percpu(vi->stats);
|
2007-11-19 16:20:42 +00:00
|
|
|
free_netdev(vi->dev);
|
2007-10-22 01:03:37 +00:00
|
|
|
}
|
|
|
|
|
2011-12-22 11:28:33 +00:00
|
|
|
#ifdef CONFIG_PM
|
|
|
|
static int virtnet_freeze(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = vdev->priv;
|
2012-12-07 07:04:56 +00:00
|
|
|
int i;
|
2011-12-22 11:28:33 +00:00
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
/* Prevent config work handler from accessing the device */
|
|
|
|
mutex_lock(&vi->config_lock);
|
|
|
|
vi->config_enable = false;
|
|
|
|
mutex_unlock(&vi->config_lock);
|
|
|
|
|
2011-12-22 11:28:33 +00:00
|
|
|
netif_device_detach(vi->dev);
|
|
|
|
cancel_delayed_work_sync(&vi->refill);
|
|
|
|
|
|
|
|
if (netif_running(vi->dev))
|
2012-12-07 07:04:56 +00:00
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++) {
|
|
|
|
napi_disable(&vi->rq[i].napi);
|
|
|
|
netif_napi_del(&vi->rq[i].napi);
|
|
|
|
}
|
2011-12-22 11:28:33 +00:00
|
|
|
|
|
|
|
remove_vq_common(vi);
|
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
flush_work(&vi->config_work);
|
|
|
|
|
2011-12-22 11:28:33 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int virtnet_restore(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = vdev->priv;
|
2012-12-07 07:04:56 +00:00
|
|
|
int err, i;
|
2011-12-22 11:28:33 +00:00
|
|
|
|
|
|
|
err = init_vqs(vi);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
if (netif_running(vi->dev))
|
2012-12-07 07:04:56 +00:00
|
|
|
for (i = 0; i < vi->max_queue_pairs; i++)
|
|
|
|
virtnet_napi_enable(&vi->rq[i]);
|
2011-12-22 11:28:33 +00:00
|
|
|
|
|
|
|
netif_device_attach(vi->dev);
|
|
|
|
|
2013-04-29 02:30:08 +00:00
|
|
|
for (i = 0; i < vi->curr_queue_pairs; i++)
|
2012-12-07 07:04:56 +00:00
|
|
|
if (!try_fill_recv(&vi->rq[i], GFP_KERNEL))
|
|
|
|
schedule_delayed_work(&vi->refill, 0);
|
2011-12-22 11:28:33 +00:00
|
|
|
|
2012-04-11 20:43:52 +00:00
|
|
|
mutex_lock(&vi->config_lock);
|
|
|
|
vi->config_enable = true;
|
|
|
|
mutex_unlock(&vi->config_lock);
|
|
|
|
|
2013-10-15 03:18:59 +00:00
|
|
|
rtnl_lock();
|
2012-12-07 07:04:56 +00:00
|
|
|
virtnet_set_queues(vi, vi->curr_queue_pairs);
|
2013-10-15 03:18:59 +00:00
|
|
|
rtnl_unlock();
|
2012-12-07 07:04:56 +00:00
|
|
|
|
2011-12-22 11:28:33 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2007-10-22 01:03:37 +00:00
|
|
|
static struct virtio_device_id id_table[] = {
|
|
|
|
{ VIRTIO_ID_NET, VIRTIO_DEV_ANY_ID },
|
|
|
|
{ 0 },
|
|
|
|
};
|
|
|
|
|
2008-05-03 02:50:50 +00:00
|
|
|
static unsigned int features[] = {
|
2008-07-08 07:10:42 +00:00
|
|
|
VIRTIO_NET_F_CSUM, VIRTIO_NET_F_GUEST_CSUM,
|
|
|
|
VIRTIO_NET_F_GSO, VIRTIO_NET_F_MAC,
|
2008-05-03 02:50:50 +00:00
|
|
|
VIRTIO_NET_F_HOST_TSO4, VIRTIO_NET_F_HOST_UFO, VIRTIO_NET_F_HOST_TSO6,
|
2008-04-18 03:24:27 +00:00
|
|
|
VIRTIO_NET_F_HOST_ECN, VIRTIO_NET_F_GUEST_TSO4, VIRTIO_NET_F_GUEST_TSO6,
|
2009-07-14 14:21:02 +00:00
|
|
|
VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
|
2009-02-04 09:02:34 +00:00
|
|
|
VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
|
2009-02-04 09:02:50 +00:00
|
|
|
VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
|
2012-12-07 07:04:56 +00:00
|
|
|
VIRTIO_NET_F_GUEST_ANNOUNCE, VIRTIO_NET_F_MQ,
|
2013-01-21 01:17:23 +00:00
|
|
|
VIRTIO_NET_F_CTRL_MAC_ADDR,
|
2013-07-25 00:50:23 +00:00
|
|
|
VIRTIO_F_ANY_LAYOUT,
|
2008-05-03 02:50:50 +00:00
|
|
|
};
|
|
|
|
|
2009-11-05 09:32:44 +00:00
|
|
|
static struct virtio_driver virtio_net_driver = {
|
2008-05-03 02:50:50 +00:00
|
|
|
.feature_table = features,
|
|
|
|
.feature_table_size = ARRAY_SIZE(features),
|
2007-10-22 01:03:37 +00:00
|
|
|
.driver.name = KBUILD_MODNAME,
|
|
|
|
.driver.owner = THIS_MODULE,
|
|
|
|
.id_table = id_table,
|
|
|
|
.probe = virtnet_probe,
|
2012-12-03 14:24:15 +00:00
|
|
|
.remove = virtnet_remove,
|
2009-01-20 01:09:49 +00:00
|
|
|
.config_changed = virtnet_config_changed,
|
2011-12-22 11:28:33 +00:00
|
|
|
#ifdef CONFIG_PM
|
|
|
|
.freeze = virtnet_freeze,
|
|
|
|
.restore = virtnet_restore,
|
|
|
|
#endif
|
2007-10-22 01:03:37 +00:00
|
|
|
};
|
|
|
|
|
2013-02-13 06:29:28 +00:00
|
|
|
module_virtio_driver(virtio_net_driver);
|
2007-10-22 01:03:37 +00:00
|
|
|
|
|
|
|
MODULE_DEVICE_TABLE(virtio, id_table);
|
|
|
|
MODULE_DESCRIPTION("Virtio network driver");
|
|
|
|
MODULE_LICENSE("GPL");
|