xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

tuntap: limit head length of skb allocated

Browse Source 
We currently use hdr_len as a hint of head length which is advertised by
guest. But when guest advertise a very big value, it can lead to an 64K+
allocating of kmalloc() which has a very high possibility of failure when host
memory is fragmented or under heavy stress. The huge hdr_len also reduce the
effect of zerocopy or even disable if a gso skb is linearized in guest.

To solves those issues, this patch introduces an upper limit (PAGE_SIZE) of the
head, which guarantees an order 0 allocation each time.

Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Jason Wang
2013-11-13 14:00:39 +08:00
committed by David S. Miller
parent 6115c11fe1
commit 96f8d9ecf2
1 changed files with 9 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
drivers/net/tun.c
View File

@@ -981,6 +981,7 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile,
struct sk_buff *skb; struct sk_buff *skb;
size_t len = total_len, align = NET_SKB_PAD, linear; size_t len = total_len, align = NET_SKB_PAD, linear;
struct virtio_net_hdr gso = { 0 }; struct virtio_net_hdr gso = { 0 };
int good_linear;
int offset = 0; int offset = 0;
int copylen; int copylen;
bool zerocopy = false; bool zerocopy = false;
@@ -1021,12 +1022,16 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile,
return -EINVAL; return -EINVAL;
} }
good_linear = SKB_MAX_HEAD(align);
if (msg_control) { if (msg_control) {
/* There are 256 bytes to be copied in skb, so there is /* There are 256 bytes to be copied in skb, so there is
* enough room for skb expand head in case it is used. * enough room for skb expand head in case it is used.
* The rest of the buffer is mapped from userspace. * The rest of the buffer is mapped from userspace.
*/ */
copylen = gso.hdr_len ? gso.hdr_len : GOODCOPY_LEN; copylen = gso.hdr_len ? gso.hdr_len : GOODCOPY_LEN;
if (copylen > good_linear)
copylen = good_linear;
linear = copylen; linear = copylen;
if (iov_pages(iv, offset + copylen, count) <= MAX_SKB_FRAGS) if (iov_pages(iv, offset + copylen, count) <= MAX_SKB_FRAGS)
zerocopy = true; zerocopy = true;
@@ -1034,7 +1039,10 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile,
if (!zerocopy) { if (!zerocopy) {
copylen = len; copylen = len;
linear = gso.hdr_len; if (gso.hdr_len > good_linear)
linear = good_linear;
else
linear = gso.hdr_len;
} }
skb = tun_alloc_skb(tfile, align, copylen, linear, noblock); skb = tun_alloc_skb(tfile, align, copylen, linear, noblock);