* [GIT PULL nf] IPVS Fixes for v3.18
From: Simon Horman @ 2014-10-28 1:05 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
Hi Pablo,
please consider this fix for v3.18.
It fixes a null-pointer dereference that may occur when logging
errors.
This problem was introduced by 4a4739d56b0 ("ipvs: Pull out
crosses_local_route_boundary logic") in v3.17-rc5. As such I would
also like it considered for 3.17-stable.
The following changes since commit 7965ee93719921ea5978f331da653dfa2d7b99f5:
netfilter: nft_compat: fix wrong target lookup in nft_target_select_ops() (2014-10-27 22:17:46 +0100)
are available in the git repository at:
https://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs.git tags/ipvs-fixes-for-v3.18
for you to fetch changes up to 3d53666b40007b55204ee8890618da79a20c9940:
ipvs: Avoid null-pointer deref in debug code (2014-10-28 09:48:31 +0900)
----------------------------------------------------------------
Alex Gartrell (1):
ipvs: Avoid null-pointer deref in debug code
net/netfilter/ipvs/ip_vs_xmit.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
^ permalink raw reply
* [PATCH nf] ipvs: Avoid null-pointer deref in debug code
From: Simon Horman @ 2014-10-28 1:05 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Alex Gartrell, Simon Horman
In-Reply-To: <1414458334-22479-1-git-send-email-horms@verge.net.au>
From: Alex Gartrell <agartrell@fb.com>
Use daddr instead of reaching into dest.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Gartrell <agartrell@fb.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_xmit.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 91f17c1..437a366 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -316,7 +316,7 @@ __ip_vs_get_out_rt(int skb_af, struct sk_buff *skb, struct ip_vs_dest *dest,
if (unlikely(crosses_local_route_boundary(skb_af, skb, rt_mode,
local))) {
IP_VS_DBG_RL("We are crossing local and non-local addresses"
- " daddr=%pI4\n", &dest->addr.ip);
+ " daddr=%pI4\n", &daddr);
goto err_put;
}
@@ -458,7 +458,7 @@ __ip_vs_get_out_rt_v6(int skb_af, struct sk_buff *skb, struct ip_vs_dest *dest,
if (unlikely(crosses_local_route_boundary(skb_af, skb, rt_mode,
local))) {
IP_VS_DBG_RL("We are crossing local and non-local addresses"
- " daddr=%pI6\n", &dest->addr.in6);
+ " daddr=%pI6\n", daddr);
goto err_put;
}
--
2.1.1
^ permalink raw reply related
* [PATCH nf-next] ipvs: remove unnecessary assignment in __ip_vs_get_out_rt
From: Simon Horman @ 2014-10-28 0:59 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Alex Gartrell, Simon Horman
In-Reply-To: <1414457960-20864-1-git-send-email-horms@verge.net.au>
From: Alex Gartrell <agartrell@fb.com>
It is a precondition of the function that daddr be equal to dest->addr.ip
if dest is non-NULL, so this additional assignment is just confusing for
stupid engineers like me.
Signed-off-by: Alex Gartrell <agartrell@fb.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
---
net/netfilter/ipvs/ip_vs_xmit.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 91f17c1..5efa597 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -293,7 +293,6 @@ __ip_vs_get_out_rt(int skb_af, struct sk_buff *skb, struct ip_vs_dest *dest,
&dest->addr.ip, &dest_dst->dst_saddr.ip,
atomic_read(&rt->dst.__refcnt));
}
- daddr = dest->addr.ip;
if (ret_saddr)
*ret_saddr = dest_dst->dst_saddr.ip;
} else {
--
2.1.1
^ permalink raw reply related
* [GIT PULL nf-next] IPVS Updates for v3.19
From: Simon Horman @ 2014-10-28 0:59 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Simon Horman
Hi Pablo,
please consider these IPVS updates for v3.19.
The single patch in this series fixes some minor fallout from adding
support IPv6 real servers in IPv4 virtual-services and vice versa.
It should not have any run-time affect other than perhaps saving a few cycles.
The following changes since commit 61ed53deb1c6a4386d8710dbbfcee8779c381931:
Merge tag 'ntb-3.18' of git://github.com/jonmason/ntb (2014-10-19 12:58:22 -0700)
are available in the git repository at:
https://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs-next.git tags/ipvs-for-v3.19
for you to fetch changes up to d7701089118d23bfed03bad0a6b5cc5115990c9e:
ipvs: remove unnecessary assignment in __ip_vs_get_out_rt (2014-10-28 09:50:06 +0900)
----------------------------------------------------------------
Alex Gartrell (1):
ipvs: remove unnecessary assignment in __ip_vs_get_out_rt
net/netfilter/ipvs/ip_vs_xmit.c | 1 -
1 file changed, 1 deletion(-)
^ permalink raw reply
* Re: [PATCH RFC 4/4] virtio_net: bigger header when VERSION_1 is set
From: Rusty Russell @ 2014-10-28 0:28 UTC (permalink / raw)
To: Michael S. Tsirkin, linux-kernel; +Cc: netdev, virtualization
In-Reply-To: <1414099656-28090-4-git-send-email-mst@redhat.com>
"Michael S. Tsirkin" <mst@redhat.com> writes:
> With VERSION_1 virtio_net uses same header size
> whether mergeable buffers are enabled or not.
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
These two are great too, thanks:
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Cheers,
Rusty.
> ---
> drivers/net/virtio_net.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 9c6d50f..a2fe340 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -1764,7 +1764,8 @@ static int virtnet_probe(struct virtio_device *vdev)
> if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> vi->mergeable_rx_bufs = true;
>
> - if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> + if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF) ||
> + virtio_has_feature(vdev, VIRTIO_F_VERSION_1))
> vi->hdr_len = sizeof(struct virtio_net_hdr_mrg_rxbuf);
> else
> vi->hdr_len = sizeof(struct virtio_net_hdr);
> --
> MST
^ permalink raw reply
* Re: [PATCH RFC 2/4] virtio_net: get rid of virtio_net_hdr/skb_vnet_hdr
From: Rusty Russell @ 2014-10-28 0:27 UTC (permalink / raw)
To: Michael S. Tsirkin, linux-kernel; +Cc: netdev, virtualization
In-Reply-To: <1414099656-28090-2-git-send-email-mst@redhat.com>
"Michael S. Tsirkin" <mst@redhat.com> writes:
> virtio 1.0 doesn't use virtio_net_hdr anymore, and in fact, it's not
> really useful since virtio_net_hdr_mrg_rxbuf includes that as the first
> field anyway.
>
> Let's drop it, precalculate header len and store within vi instead.
>
> This way we can also remove struct skb_vnet_hdr.
Yes, this is definitely a win.
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Thanks,
Rusty.
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> drivers/net/virtio_net.c | 88 ++++++++++++++++++++++--------------------------
> 1 file changed, 40 insertions(+), 48 deletions(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 36f3dfc..a795a23 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -123,6 +123,9 @@ struct virtnet_info {
> /* Host can handle any s/g split between our header and packet data */
> bool any_header_sg;
>
> + /* Packet virtio header size */
> + u8 hdr_len;
> +
> /* Active statistics */
> struct virtnet_stats __percpu *stats;
>
> @@ -139,21 +142,14 @@ struct virtnet_info {
> struct notifier_block nb;
> };
>
> -struct skb_vnet_hdr {
> - union {
> - struct virtio_net_hdr hdr;
> - struct virtio_net_hdr_mrg_rxbuf mhdr;
> - };
> -};
> -
> struct padded_vnet_hdr {
> - struct virtio_net_hdr hdr;
> + struct virtio_net_hdr_mrg_rxbuf 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.
> + * hdr is in a separate sg buffer, and data sg buffer shares same page
> + * with this header sg. This padding makes next sg 16 byte aligned
> + * after the header.
> */
> - char padding[6];
> + char padding[4];
> };
>
> /* Converting between virtqueue no. and kernel tx/rx queue no.
> @@ -179,9 +175,9 @@ static int rxq2vq(int rxq)
> return rxq * 2;
> }
>
> -static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
> +static inline struct virtio_net_hdr_mrg_rxbuf *skb_vnet_hdr(struct sk_buff *skb)
> {
> - return (struct skb_vnet_hdr *)skb->cb;
> + return (struct virtio_net_hdr_mrg_rxbuf *)skb->cb;
> }
>
> /*
> @@ -247,7 +243,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> unsigned int len, unsigned int truesize)
> {
> struct sk_buff *skb;
> - struct skb_vnet_hdr *hdr;
> + struct virtio_net_hdr_mrg_rxbuf *hdr;
> unsigned int copy, hdr_len, hdr_padded_len;
> char *p;
>
> @@ -260,13 +256,11 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
>
> hdr = skb_vnet_hdr(skb);
>
> - if (vi->mergeable_rx_bufs) {
> - hdr_len = sizeof hdr->mhdr;
> - hdr_padded_len = sizeof hdr->mhdr;
> - } else {
> - hdr_len = sizeof hdr->hdr;
> + hdr_len = vi->hdr_len;
> + if (vi->mergeable_rx_bufs)
> + hdr_padded_len = sizeof *hdr;
> + else
> hdr_padded_len = sizeof(struct padded_vnet_hdr);
> - }
>
> memcpy(hdr, p, hdr_len);
>
> @@ -317,11 +311,11 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> return skb;
> }
>
> -static struct sk_buff *receive_small(void *buf, unsigned int len)
> +static struct sk_buff *receive_small(struct virtnet_info *vi, void *buf, unsigned int len)
> {
> struct sk_buff * skb = buf;
>
> - len -= sizeof(struct virtio_net_hdr);
> + len -= vi->hdr_len;
> skb_trim(skb, len);
>
> return skb;
> @@ -354,8 +348,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
> unsigned int len)
> {
> void *buf = mergeable_ctx_to_buf_address(ctx);
> - struct skb_vnet_hdr *hdr = buf;
> - u16 num_buf = virtio16_to_cpu(rq->vq->vdev, hdr->mhdr.num_buffers);
> + struct virtio_net_hdr_mrg_rxbuf *hdr = buf;
> + u16 num_buf = virtio16_to_cpu(vi->vdev, hdr->num_buffers);
> struct page *page = virt_to_head_page(buf);
> int offset = buf - page_address(page);
> unsigned int truesize = max(len, mergeable_ctx_to_buf_truesize(ctx));
> @@ -373,8 +367,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
> if (unlikely(!ctx)) {
> pr_debug("%s: rx error: %d buffers out of %d missing\n",
> dev->name, num_buf,
> - virtio16_to_cpu(rq->vq->vdev,
> - hdr->mhdr.num_buffers));
> + virtio16_to_cpu(vi->vdev,
> + hdr->num_buffers));
> dev->stats.rx_length_errors++;
> goto err_buf;
> }
> @@ -441,7 +435,7 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> struct net_device *dev = vi->dev;
> struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
> struct sk_buff *skb;
> - struct skb_vnet_hdr *hdr;
> + struct virtio_net_hdr_mrg_rxbuf *hdr;
>
> if (unlikely(len < sizeof(struct virtio_net_hdr) + ETH_HLEN)) {
> pr_debug("%s: short packet %i\n", dev->name, len);
> @@ -536,7 +530,7 @@ static int add_recvbuf_small(struct virtnet_info *vi, struct receive_queue *rq,
> gfp_t gfp)
> {
> struct sk_buff *skb;
> - struct skb_vnet_hdr *hdr;
> + struct virtio_net_hdr_mrg_rxbuf *hdr;
> int err;
>
> skb = __netdev_alloc_skb_ip_align(vi->dev, GOOD_PACKET_LEN, gfp);
> @@ -547,7 +541,7 @@ static int add_recvbuf_small(struct virtnet_info *vi, struct receive_queue *rq,
>
> hdr = skb_vnet_hdr(skb);
> sg_init_table(rq->sg, MAX_SKB_FRAGS + 2);
> - sg_set_buf(rq->sg, &hdr->hdr, sizeof hdr->hdr);
> + sg_set_buf(rq->sg, hdr, vi->hdr_len);
> skb_to_sgvec(skb, rq->sg + 1, 0, skb->len);
>
> err = virtqueue_add_inbuf(rq->vq, rq->sg, 2, skb, gfp);
> @@ -557,7 +551,8 @@ static int add_recvbuf_small(struct virtnet_info *vi, struct receive_queue *rq,
> return err;
> }
>
> -static int add_recvbuf_big(struct receive_queue *rq, gfp_t gfp)
> +static int add_recvbuf_big(struct virtnet_info *vi, struct receive_queue *rq,
> + gfp_t gfp)
> {
> struct page *first, *list = NULL;
> char *p;
> @@ -588,8 +583,8 @@ static int add_recvbuf_big(struct receive_queue *rq, gfp_t gfp)
> p = page_address(first);
>
> /* 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));
> + /* a separated rq->sg[0] for header - required in case !any_header_sg */
> + sg_set_buf(&rq->sg[0], p, vi->hdr_len);
>
> /* rq->sg[1] for data packet, from offset */
> offset = sizeof(struct padded_vnet_hdr);
> @@ -668,7 +663,7 @@ static bool try_fill_recv(struct virtnet_info *vi, struct receive_queue *rq,
> if (vi->mergeable_rx_bufs)
> err = add_recvbuf_mergeable(rq, gfp);
> else if (vi->big_packets)
> - err = add_recvbuf_big(rq, gfp);
> + err = add_recvbuf_big(vi, rq, gfp);
> else
> err = add_recvbuf_small(vi, rq, gfp);
>
> @@ -848,18 +843,14 @@ static void free_old_xmit_skbs(struct send_queue *sq)
>
> static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
> {
> - struct skb_vnet_hdr *hdr;
> + struct virtio_net_hdr_mrg_rxbuf *hdr;
> const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;
> struct virtnet_info *vi = sq->vq->vdev->priv;
> unsigned num_sg;
> - unsigned hdr_len;
> + unsigned hdr_len = vi->hdr_len;
> bool can_push;
>
> pr_debug("%s: xmit %p %pM\n", vi->dev->name, skb, dest);
> - 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)) &&
> @@ -867,7 +858,7 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
> /* 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);
> + hdr = (struct virtio_net_hdr_mrg_rxbuf *)(skb->data - hdr_len);
> else
> hdr = skb_vnet_hdr(skb);
>
> @@ -902,7 +893,7 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
> }
>
> if (vi->mergeable_rx_bufs)
> - hdr->mhdr.num_buffers = 0;
> + hdr->num_buffers = 0;
>
> sg_init_table(sq->sg, MAX_SKB_FRAGS + 2);
> if (can_push) {
> @@ -1773,18 +1764,19 @@ static int virtnet_probe(struct virtio_device *vdev)
> if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> vi->mergeable_rx_bufs = true;
>
> + if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
> + vi->hdr_len = sizeof(struct virtio_net_hdr_mrg_rxbuf);
> + else
> + vi->hdr_len = sizeof(struct virtio_net_hdr);
> +
> if (virtio_has_feature(vdev, VIRTIO_F_ANY_LAYOUT))
> vi->any_header_sg = true;
>
> if (virtio_has_feature(vdev, VIRTIO_NET_F_CTRL_VQ))
> vi->has_cvq = true;
>
> - if (vi->any_header_sg) {
> - if (vi->mergeable_rx_bufs)
> - dev->needed_headroom = sizeof(struct virtio_net_hdr_mrg_rxbuf);
> - else
> - dev->needed_headroom = sizeof(struct virtio_net_hdr);
> - }
> + if (vi->any_header_sg)
> + dev->needed_headroom = vi->hdr_len;
>
> /* Use single tx/rx queue pair as default */
> vi->curr_queue_pairs = 1;
> --
> MST
^ permalink raw reply
* Re: [PATCH RFC 1/4] virtio_net: pass vi around
From: Rusty Russell @ 2014-10-28 0:27 UTC (permalink / raw)
To: Michael S. Tsirkin, linux-kernel; +Cc: netdev, virtualization
In-Reply-To: <1414099656-28090-1-git-send-email-mst@redhat.com>
"Michael S. Tsirkin" <mst@redhat.com> writes:
> Too many places poke at [rs]q->vq->vdev->priv just to get
> the the vi structure. Let's just pass the pointer around: seems
> cleaner, and might even be faster.
Agreed, it's neater.
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Thanks,
Rusty.
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> drivers/net/virtio_net.c | 36 +++++++++++++++++++-----------------
> 1 file changed, 19 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 57cbc7d..36f3dfc 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -241,11 +241,11 @@ static unsigned long mergeable_buf_to_ctx(void *buf, unsigned int truesize)
> }
>
> /* Called from bottom half context */
> -static struct sk_buff *page_to_skb(struct receive_queue *rq,
> +static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> + struct receive_queue *rq,
> struct page *page, unsigned int offset,
> unsigned int len, unsigned int truesize)
> {
> - struct virtnet_info *vi = rq->vq->vdev->priv;
> struct sk_buff *skb;
> struct skb_vnet_hdr *hdr;
> unsigned int copy, hdr_len, hdr_padded_len;
> @@ -328,12 +328,13 @@ static struct sk_buff *receive_small(void *buf, unsigned int len)
> }
>
> static struct sk_buff *receive_big(struct net_device *dev,
> + struct virtnet_info *vi,
> struct receive_queue *rq,
> void *buf,
> unsigned int len)
> {
> struct page *page = buf;
> - struct sk_buff *skb = page_to_skb(rq, page, 0, len, PAGE_SIZE);
> + struct sk_buff *skb = page_to_skb(vi, rq, page, 0, len, PAGE_SIZE);
>
> if (unlikely(!skb))
> goto err;
> @@ -359,7 +360,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
> int offset = buf - page_address(page);
> unsigned int truesize = max(len, mergeable_ctx_to_buf_truesize(ctx));
>
> - struct sk_buff *head_skb = page_to_skb(rq, page, offset, len, truesize);
> + struct sk_buff *head_skb = page_to_skb(vi, rq, page, offset, len,
> + truesize);
> struct sk_buff *curr_skb = head_skb;
>
> if (unlikely(!curr_skb))
> @@ -433,9 +435,9 @@ err_buf:
> return NULL;
> }
>
> -static void receive_buf(struct receive_queue *rq, void *buf, unsigned int len)
> +static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> + void *buf, unsigned int len)
> {
> - struct virtnet_info *vi = rq->vq->vdev->priv;
> struct net_device *dev = vi->dev;
> struct virtnet_stats *stats = this_cpu_ptr(vi->stats);
> struct sk_buff *skb;
> @@ -459,9 +461,9 @@ static void receive_buf(struct receive_queue *rq, void *buf, unsigned int len)
> if (vi->mergeable_rx_bufs)
> skb = receive_mergeable(dev, vi, rq, (unsigned long)buf, len);
> else if (vi->big_packets)
> - skb = receive_big(dev, rq, buf, len);
> + skb = receive_big(dev, vi, rq, buf, len);
> else
> - skb = receive_small(buf, len);
> + skb = receive_small(vi, buf, len);
>
> if (unlikely(!skb))
> return;
> @@ -530,9 +532,9 @@ frame_err:
> dev_kfree_skb(skb);
> }
>
> -static int add_recvbuf_small(struct receive_queue *rq, gfp_t gfp)
> +static int add_recvbuf_small(struct virtnet_info *vi, struct receive_queue *rq,
> + gfp_t gfp)
> {
> - struct virtnet_info *vi = rq->vq->vdev->priv;
> struct sk_buff *skb;
> struct skb_vnet_hdr *hdr;
> int err;
> @@ -655,9 +657,9 @@ static int add_recvbuf_mergeable(struct receive_queue *rq, gfp_t gfp)
> * before we're receiving packets, or from refill_work which is
> * careful to disable receiving (using napi_disable).
> */
> -static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
> +static bool try_fill_recv(struct virtnet_info *vi, struct receive_queue *rq,
> + gfp_t gfp)
> {
> - struct virtnet_info *vi = rq->vq->vdev->priv;
> int err;
> bool oom;
>
> @@ -668,7 +670,7 @@ static bool try_fill_recv(struct receive_queue *rq, gfp_t gfp)
> else if (vi->big_packets)
> err = add_recvbuf_big(rq, gfp);
> else
> - err = add_recvbuf_small(rq, gfp);
> + err = add_recvbuf_small(vi, rq, gfp);
>
> oom = err == -ENOMEM;
> if (err)
> @@ -717,7 +719,7 @@ static void refill_work(struct work_struct *work)
> struct receive_queue *rq = &vi->rq[i];
>
> napi_disable(&rq->napi);
> - still_empty = !try_fill_recv(rq, GFP_KERNEL);
> + still_empty = !try_fill_recv(vi, rq, GFP_KERNEL);
> virtnet_napi_enable(rq);
>
> /* In theory, this can happen: if we don't get any buffers in
> @@ -736,12 +738,12 @@ static int virtnet_receive(struct receive_queue *rq, int budget)
>
> while (received < budget &&
> (buf = virtqueue_get_buf(rq->vq, &len)) != NULL) {
> - receive_buf(rq, buf, len);
> + receive_buf(vi, rq, buf, len);
> received++;
> }
>
> if (rq->vq->num_free > virtqueue_get_vring_size(rq->vq) / 2) {
> - if (!try_fill_recv(rq, GFP_ATOMIC))
> + if (!try_fill_recv(vi, rq, GFP_ATOMIC))
> schedule_delayed_work(&vi->refill, 0);
> }
>
> @@ -817,7 +819,7 @@ static int virtnet_open(struct net_device *dev)
> 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))
> + if (!try_fill_recv(vi, &vi->rq[i], GFP_KERNEL))
> schedule_delayed_work(&vi->refill, 0);
> virtnet_napi_enable(&vi->rq[i]);
> }
> --
> MST
^ permalink raw reply
* Re: [PATCH] ovs: Turn vports with dependencies into separate modules
From: Pravin Shelar @ 2014-10-28 0:27 UTC (permalink / raw)
To: Thomas Graf; +Cc: dev@openvswitch.org, netdev
In-Reply-To: <20141027214722.GA2783@casper.infradead.org>
On Mon, Oct 27, 2014 at 2:47 PM, Thomas Graf <tgraf@suug.ch> wrote:
> On 10/27/14 at 10:14am, Pravin Shelar wrote:
>> On Fri, Oct 24, 2014 at 2:57 PM, Thomas Graf <tgraf@suug.ch> wrote:
>> > I was refering to how many other kernel APIs have been designed, a
>> > registration API allowing a vport to be implemented exclusively in the
>> > scope of a single file tends to be cleaner than having to touch multiple
>> > files and maintaining an init list.
>> >
>> This has never been issue in openvswitch. Plus we do not need loadable
>> vport module to fix this issue.
>>
>> > It also allows for OVS to be built into vmlinuz while vports can
>> > remain as modules even if vxlan itself is built as a module.
>> >
>>
>> What is problem with current OVS built into kernel?
>
> What I mean specifically is the following dependency logic which will
> no longer be required:
>
> depends on NET_IPGRE_DEMUX && !(OPENVSWITCH=y && NET_IPGRE_DEMUX=m)
>
> The patch also brings additional flexibility to users of
> distributions. Distros typically ship something like an allmodconfig
> so a user can either run openvswitch.ko with all encaps compiled in
> or not run openvswitch.ko. With vports as module, a user can blacklist
> a certain encap type.
>
> Another advantage is obviously that users can run additional vport
> types on top of their distribution kernels.
>
> Is there anything specific that you are concerned with in regard
> to this proposed change?
OVS vport code is not alot and making it plugable module does not save
much space. Even with this patch user can not load any vport type
since we still need to define the type in kernel interface and add the
support in userspace netdev layer. Therefore this patch adds
complexity without much gain.
^ permalink raw reply
* Re: [PATCH net] bpf: split eBPF out of NET
From: Alexei Starovoitov @ 2014-10-28 0:18 UTC (permalink / raw)
To: David Miller
Cc: Geert Uytterhoeven, Josh Triplett, Ingo Molnar, Steven Rostedt,
Hannes Frederic Sowa, Eric Dumazet, Daniel Borkmann,
Network Development, LKML
In-Reply-To: <20141027.191043.246099210901442100.davem@davemloft.net>
On Mon, Oct 27, 2014 at 4:10 PM, David Miller <davem@davemloft.net> wrote:
> From: Alexei Starovoitov <ast@plumgrid.com>
> Date: Thu, 23 Oct 2014 18:41:08 -0700
>
>> introduce two configs:
>> - hidden CONFIG_BPF to select eBPF interpreter that classic socket filters
>> depend on
>> - visible CONFIG_BPF_SYSCALL (default off) that tracing and sockets can use
>>
>> that solves several problems:
>> - tracing and others that wish to use eBPF don't need to depend on NET.
>> They can use BPF_SYSCALL to allow loading from userspace or select BPF
>> to use it directly from kernel in NET-less configs.
>> - in 3.18 programs cannot be attached to events yet, so don't force it on
>> - when the rest of eBPF infra is there in 3.19+, it's still useful to
>> switch it off to minimize kernel size
>>
>> Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
>> ---
>>
>> bloat-o-meter on x64 shows:
>> add/remove: 0/60 grow/shrink: 0/2 up/down: 0/-15601 (-15601)
>>
>> tested with many different config combinations. Hopefully didn't miss anything.
>
> Applied with two changes:
>
> 1) boolean --> bool
> 2) Moved bloat-o-meter and testing information into commit message.
>
> Thanks.
Thank you for taking care of it!
^ permalink raw reply
* Re: [Bug 86851] New: Reproducible panic on heavy UDP traffic
From: Eric Dumazet @ 2014-10-28 0:16 UTC (permalink / raw)
To: Nikolay Aleksandrov
Cc: Patrick McLean, Florian Westphal, Stephen Hemminger, netdev
In-Reply-To: <544ECFFA.8080402@redhat.com>
On Tue, 2014-10-28 at 00:06 +0100, Nikolay Aleksandrov wrote:
> Great! Thanks for testing.
> As I said earlier we have a valid case that can hit the WARN_ON in
> inet_evict_frag().
> Anyhow, Eric would you mind posting the patch officially ?
> If you'd like me to remove the WARN_ON() in a separate one just let me
> know, otherwise feel free to remove it in the fix for the race.
Please Nikolay take ownership of this patch, I am busy on other stuff at
the moment, thanks !
^ permalink raw reply
* [PATCH v2] net: ethernet: realtek: atp: checkpatch errors and warnings corrected
From: Roberto Medina @ 2014-10-27 23:51 UTC (permalink / raw)
To: netdev; +Cc: linux-kernel, Roberto Medina
From: Roberto Medina <robertoxmed@gmail.com>
Several warnings and errors of coding style rules corrected.
Compile tested.
Signed-off-by: Roberto Medina <robertoxmed@gmail.com>
---
drivers/net/ethernet/realtek/atp.h | 246 +++++++++++++++++++------------------
1 file changed, 127 insertions(+), 119 deletions(-)
diff --git a/drivers/net/ethernet/realtek/atp.h b/drivers/net/ethernet/realtek/atp.h
index 040b137..32497f0 100644
--- a/drivers/net/ethernet/realtek/atp.h
+++ b/drivers/net/ethernet/realtek/atp.h
@@ -6,10 +6,10 @@
/* The header prepended to received packets. */
struct rx_header {
- ushort pad; /* Pad. */
- ushort rx_count;
- ushort rx_status; /* Unknown bit assignments :-<. */
- ushort cur_addr; /* Apparently the current buffer address(?) */
+ ushort pad; /* Pad. */
+ ushort rx_count;
+ ushort rx_status; /* Unknown bit assignments :-<. */
+ ushort cur_addr; /* Apparently the current buffer address(?) */
};
#define PAR_DATA 0
@@ -29,22 +29,25 @@ struct rx_header {
#define RdAddr 0xC0
#define HNib 0x10
-enum page0_regs
-{
- /* The first six registers hold the ethernet physical station address. */
- PAR0 = 0, PAR1 = 1, PAR2 = 2, PAR3 = 3, PAR4 = 4, PAR5 = 5,
- TxCNT0 = 6, TxCNT1 = 7, /* The transmit byte count. */
- TxSTAT = 8, RxSTAT = 9, /* Tx and Rx status. */
- ISR = 10, IMR = 11, /* Interrupt status and mask. */
- CMR1 = 12, /* Command register 1. */
- CMR2 = 13, /* Command register 2. */
- MODSEL = 14, /* Mode select register. */
- MAR = 14, /* Memory address register (?). */
- CMR2_h = 0x1d, };
-
-enum eepage_regs
-{ PROM_CMD = 6, PROM_DATA = 7 }; /* Note that PROM_CMD is in the "high" bits. */
+enum page0_regs {
+ /* The first six registers hold
+ * the ethernet physical station address.
+ */
+ PAR0 = 0, PAR1 = 1, PAR2 = 2, PAR3 = 3, PAR4 = 4, PAR5 = 5,
+ TxCNT0 = 6, TxCNT1 = 7, /* The transmit byte count. */
+ TxSTAT = 8, RxSTAT = 9, /* Tx and Rx status. */
+ ISR = 10, IMR = 11, /* Interrupt status and mask. */
+ CMR1 = 12, /* Command register 1. */
+ CMR2 = 13, /* Command register 2. */
+ MODSEL = 14, /* Mode select register. */
+ MAR = 14, /* Memory address register (?). */
+ CMR2_h = 0x1d,
+};
+enum eepage_regs {
+ PROM_CMD = 6,
+ PROM_DATA = 7 /* Note that PROM_CMD is in the "high" bits. */
+};
#define ISR_TxOK 0x01
#define ISR_RxOK 0x04
@@ -72,141 +75,146 @@ enum eepage_regs
#define CMR2h_Normal 2 /* Accept physical and broadcast address. */
#define CMR2h_PROMISC 3 /* Promiscuous mode. */
-/* An inline function used below: it differs from inb() by explicitly return an unsigned
- char, saving a truncation. */
+/* An inline function used below: it differs from inb() by explicitly
+ * return an unsigned char, saving a truncation.
+ */
static inline unsigned char inbyte(unsigned short port)
{
- unsigned char _v;
- __asm__ __volatile__ ("inb %w1,%b0" :"=a" (_v):"d" (port));
- return _v;
+ unsigned char _v;
+
+ __asm__ __volatile__ ("inb %w1,%b0" : "=a" (_v) : "d" (port));
+ return _v;
}
/* Read register OFFSET.
- This command should always be terminated with read_end(). */
+ * This command should always be terminated with read_end().
+ */
static inline unsigned char read_nibble(short port, unsigned char offset)
{
- unsigned char retval;
- outb(EOC+offset, port + PAR_DATA);
- outb(RdAddr+offset, port + PAR_DATA);
- inbyte(port + PAR_STATUS); /* Settling time delay */
- retval = inbyte(port + PAR_STATUS);
- outb(EOC+offset, port + PAR_DATA);
-
- return retval;
+ unsigned char retval;
+
+ outb(EOC+offset, port + PAR_DATA);
+ outb(RdAddr+offset, port + PAR_DATA);
+ inbyte(port + PAR_STATUS); /* Settling time delay */
+ retval = inbyte(port + PAR_STATUS);
+ outb(EOC+offset, port + PAR_DATA);
+
+ return retval;
}
/* Functions for bulk data read. The interrupt line is always disabled. */
/* Get a byte using read mode 0, reading data from the control lines. */
static inline unsigned char read_byte_mode0(short ioaddr)
{
- unsigned char low_nib;
-
- outb(Ctrl_LNibRead, ioaddr + PAR_CONTROL);
- inbyte(ioaddr + PAR_STATUS);
- low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
- outb(Ctrl_HNibRead, ioaddr + PAR_CONTROL);
- inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
- inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
- return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
+ unsigned char low_nib;
+
+ outb(Ctrl_LNibRead, ioaddr + PAR_CONTROL);
+ inbyte(ioaddr + PAR_STATUS);
+ low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
+ outb(Ctrl_HNibRead, ioaddr + PAR_CONTROL);
+ inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
+ inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
+ return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
}
/* The same as read_byte_mode0(), but does multiple inb()s for stability. */
static inline unsigned char read_byte_mode2(short ioaddr)
{
- unsigned char low_nib;
-
- outb(Ctrl_LNibRead, ioaddr + PAR_CONTROL);
- inbyte(ioaddr + PAR_STATUS);
- low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
- outb(Ctrl_HNibRead, ioaddr + PAR_CONTROL);
- inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
- return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
+ unsigned char low_nib;
+
+ outb(Ctrl_LNibRead, ioaddr + PAR_CONTROL);
+ inbyte(ioaddr + PAR_STATUS);
+ low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
+ outb(Ctrl_HNibRead, ioaddr + PAR_CONTROL);
+ inbyte(ioaddr + PAR_STATUS); /* Settling time delay -- needed! */
+ return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
}
/* Read a byte through the data register. */
static inline unsigned char read_byte_mode4(short ioaddr)
{
- unsigned char low_nib;
+ unsigned char low_nib;
- outb(RdAddr | MAR, ioaddr + PAR_DATA);
- low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
- outb(RdAddr | HNib | MAR, ioaddr + PAR_DATA);
- return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
+ outb(RdAddr | MAR, ioaddr + PAR_DATA);
+ low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
+ outb(RdAddr | HNib | MAR, ioaddr + PAR_DATA);
+ return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
}
/* Read a byte through the data register, double reading to allow settling. */
static inline unsigned char read_byte_mode6(short ioaddr)
{
- unsigned char low_nib;
-
- outb(RdAddr | MAR, ioaddr + PAR_DATA);
- inbyte(ioaddr + PAR_STATUS);
- low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
- outb(RdAddr | HNib | MAR, ioaddr + PAR_DATA);
- inbyte(ioaddr + PAR_STATUS);
- return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
+ unsigned char low_nib;
+
+ outb(RdAddr | MAR, ioaddr + PAR_DATA);
+ inbyte(ioaddr + PAR_STATUS);
+ low_nib = (inbyte(ioaddr + PAR_STATUS) >> 3) & 0x0f;
+ outb(RdAddr | HNib | MAR, ioaddr + PAR_DATA);
+ inbyte(ioaddr + PAR_STATUS);
+ return low_nib | ((inbyte(ioaddr + PAR_STATUS) << 1) & 0xf0);
}
static inline void
write_reg(short port, unsigned char reg, unsigned char value)
{
- unsigned char outval;
- outb(EOC | reg, port + PAR_DATA);
- outval = WrAddr | reg;
- outb(outval, port + PAR_DATA);
- outb(outval, port + PAR_DATA); /* Double write for PS/2. */
-
- outval &= 0xf0;
- outval |= value;
- outb(outval, port + PAR_DATA);
- outval &= 0x1f;
- outb(outval, port + PAR_DATA);
- outb(outval, port + PAR_DATA);
-
- outb(EOC | outval, port + PAR_DATA);
+ unsigned char outval;
+
+ outb(EOC | reg, port + PAR_DATA);
+ outval = WrAddr | reg;
+ outb(outval, port + PAR_DATA);
+ outb(outval, port + PAR_DATA); /* Double write for PS/2. */
+
+ outval &= 0xf0;
+ outval |= value;
+ outb(outval, port + PAR_DATA);
+ outval &= 0x1f;
+ outb(outval, port + PAR_DATA);
+ outb(outval, port + PAR_DATA);
+
+ outb(EOC | outval, port + PAR_DATA);
}
static inline void
write_reg_high(short port, unsigned char reg, unsigned char value)
{
- unsigned char outval = EOC | HNib | reg;
+ unsigned char outval = EOC | HNib | reg;
- outb(outval, port + PAR_DATA);
- outval &= WrAddr | HNib | 0x0f;
- outb(outval, port + PAR_DATA);
- outb(outval, port + PAR_DATA); /* Double write for PS/2. */
+ outb(outval, port + PAR_DATA);
+ outval &= WrAddr | HNib | 0x0f;
+ outb(outval, port + PAR_DATA);
+ outb(outval, port + PAR_DATA); /* Double write for PS/2. */
- outval = WrAddr | HNib | value;
- outb(outval, port + PAR_DATA);
- outval &= HNib | 0x0f; /* HNib | value */
- outb(outval, port + PAR_DATA);
- outb(outval, port + PAR_DATA);
+ outval = WrAddr | HNib | value;
+ outb(outval, port + PAR_DATA);
+ outval &= HNib | 0x0f; /* HNib | value */
+ outb(outval, port + PAR_DATA);
+ outb(outval, port + PAR_DATA);
- outb(EOC | HNib | outval, port + PAR_DATA);
+ outb(EOC | HNib | outval, port + PAR_DATA);
}
/* Write a byte out using nibble mode. The low nibble is written first. */
static inline void
write_reg_byte(short port, unsigned char reg, unsigned char value)
{
- unsigned char outval;
- outb(EOC | reg, port + PAR_DATA); /* Reset the address register. */
- outval = WrAddr | reg;
- outb(outval, port + PAR_DATA);
- outb(outval, port + PAR_DATA); /* Double write for PS/2. */
-
- outb((outval & 0xf0) | (value & 0x0f), port + PAR_DATA);
- outb(value & 0x0f, port + PAR_DATA);
- value >>= 4;
- outb(value, port + PAR_DATA);
- outb(0x10 | value, port + PAR_DATA);
- outb(0x10 | value, port + PAR_DATA);
-
- outb(EOC | value, port + PAR_DATA); /* Reset the address register. */
+ unsigned char outval;
+
+ outb(EOC | reg, port + PAR_DATA); /* Reset the address register. */
+ outval = WrAddr | reg;
+ outb(outval, port + PAR_DATA);
+ outb(outval, port + PAR_DATA); /* Double write for PS/2. */
+
+ outb((outval & 0xf0) | (value & 0x0f), port + PAR_DATA);
+ outb(value & 0x0f, port + PAR_DATA);
+ value >>= 4;
+ outb(value, port + PAR_DATA);
+ outb(0x10 | value, port + PAR_DATA);
+ outb(0x10 | value, port + PAR_DATA);
+
+ outb(EOC | value, port + PAR_DATA); /* Reset the address register. */
}
-/*
- * Bulk data writes to the packet buffer. The interrupt line remains enabled.
+/* Bulk data writes to the packet buffer. The interrupt line remains enabled.
* The first, faster method uses only the dataport (data modes 0, 2 & 4).
* The second (backup) method uses data and control regs (modes 1, 3 & 5).
* It should only be needed when there is skew between the individual data
@@ -214,28 +222,28 @@ write_reg_byte(short port, unsigned char reg, unsigned char value)
*/
static inline void write_byte_mode0(short ioaddr, unsigned char value)
{
- outb(value & 0x0f, ioaddr + PAR_DATA);
- outb((value>>4) | 0x10, ioaddr + PAR_DATA);
+ outb(value & 0x0f, ioaddr + PAR_DATA);
+ outb((value>>4) | 0x10, ioaddr + PAR_DATA);
}
static inline void write_byte_mode1(short ioaddr, unsigned char value)
{
- outb(value & 0x0f, ioaddr + PAR_DATA);
- outb(Ctrl_IRQEN | Ctrl_LNibWrite, ioaddr + PAR_CONTROL);
- outb((value>>4) | 0x10, ioaddr + PAR_DATA);
- outb(Ctrl_IRQEN | Ctrl_HNibWrite, ioaddr + PAR_CONTROL);
+ outb(value & 0x0f, ioaddr + PAR_DATA);
+ outb(Ctrl_IRQEN | Ctrl_LNibWrite, ioaddr + PAR_CONTROL);
+ outb((value>>4) | 0x10, ioaddr + PAR_DATA);
+ outb(Ctrl_IRQEN | Ctrl_HNibWrite, ioaddr + PAR_CONTROL);
}
/* Write 16bit VALUE to the packet buffer: the same as above just doubled. */
static inline void write_word_mode0(short ioaddr, unsigned short value)
{
- outb(value & 0x0f, ioaddr + PAR_DATA);
- value >>= 4;
- outb((value & 0x0f) | 0x10, ioaddr + PAR_DATA);
- value >>= 4;
- outb(value & 0x0f, ioaddr + PAR_DATA);
- value >>= 4;
- outb((value & 0x0f) | 0x10, ioaddr + PAR_DATA);
+ outb(value & 0x0f, ioaddr + PAR_DATA);
+ value >>= 4;
+ outb((value & 0x0f) | 0x10, ioaddr + PAR_DATA);
+ value >>= 4;
+ outb(value & 0x0f, ioaddr + PAR_DATA);
+ value >>= 4;
+ outb((value & 0x0f) | 0x10, ioaddr + PAR_DATA);
}
/* EEPROM_Ctrl bits. */
@@ -248,10 +256,10 @@ static inline void write_word_mode0(short ioaddr, unsigned short value)
/* Delay between EEPROM clock transitions. */
#define eeprom_delay(ticks) \
-do { int _i = 40; while (--_i > 0) { __SLOW_DOWN_IO; }} while (0)
+do { int _i = 40; while (--_i > 0) { __SLOW_DOWN_IO; } } while (0)
/* The EEPROM commands include the alway-set leading bit. */
#define EE_WRITE_CMD(offset) (((5 << 6) + (offset)) << 17)
-#define EE_READ(offset) (((6 << 6) + (offset)) << 17)
+#define EE_READ(offset) (((6 << 6) + (offset)) << 17)
#define EE_ERASE(offset) (((7 << 6) + (offset)) << 17)
#define EE_CMD_SIZE 27 /* The command+address+data size. */
--
2.1.2
^ permalink raw reply related
* Re: [PATCH net-next 2/2] udp: Reset flow table for flows over unconnected sockets
From: Eric Dumazet @ 2014-10-27 23:19 UTC (permalink / raw)
To: Tom Herbert; +Cc: David Miller, Linux Netdev List
In-Reply-To: <CA+mtBx_V3WT1bbXY9F731GNdDdb3+ebHwj9hRyVEFynAPYhSXg@mail.gmail.com>
On Mon, 2014-10-27 at 12:36 -0700, Tom Herbert wrote:
> Please try this patch and provide real data to support your points.
>
Yep. This is not good, I confirm my fear.
Google servers are shifting to serve both TCP & UDP traffic (QUIC
protocol), with an increasing UDP load.
Millions of packets per second per host, from millions of different
sources...
And your patch voids the RFS table, adds another cache miss in fast path
for UDP rx path which is already too expensive.
> If a TCP connection is hot it will continually refresh the table for
> that connection, if connection becomes idle it only takes one received
> packet to restore the CPU. The only time there could be a persistent
> problem is if collision rate is high (which probably means table is
> too small).
RFS already has a low hit/miss rate, this patch does not help neither
UDP or TCP.
Ideally, RFS should be enabled on a protocol base, not an agnostic u32
flow hash.
Whatever strategy you implement, as long as different protocols share a
common hash table, it wont be perfect for mixed workloads.
Fundamental problem is that when an UDP packet comes, its not possible
to know if its a 'flow' or 'not', unless we perform an expensive lookup,
and then RPS/RFS cost becomes prohibitive.
While for TCP, the current RFS cache miss is good enough, because about
all packets are for connected flows. We eventually have bad steering for
<not yet established> flows where the stack performs poorly anyway.
^ permalink raw reply
* Re: [PATCH net-next v3 0/5] cleanup on resource check
From: David Miller @ 2014-10-27 23:16 UTC (permalink / raw)
To: varkabhadram; +Cc: netdev, sergei.shtylyov, varkab
In-Reply-To: <1414116730-4590-1-git-send-email-varkab@cdac.in>
From: Varka Bhadram <varkabhadram@gmail.com>
Date: Fri, 24 Oct 2014 07:42:05 +0530
> This series removes the duplication of sanity check for
> platform_get_resource() return resource. It will be checked
> with devm_ioremap_resource()
>
> changes since v2:
> - Merge #1 and #2 patches into single patch
> - remove the comment
>
> changes since v1:
> - remove NULL dereference on resource_size()
Series applied, thanks.
^ permalink raw reply
* Re: [PATCH net] bpf: split eBPF out of NET
From: David Miller @ 2014-10-27 23:10 UTC (permalink / raw)
To: ast
Cc: geert, josh, mingo, rostedt, hannes, edumazet, dborkman, netdev,
linux-kernel
In-Reply-To: <1414114868-28228-1-git-send-email-ast@plumgrid.com>
From: Alexei Starovoitov <ast@plumgrid.com>
Date: Thu, 23 Oct 2014 18:41:08 -0700
> introduce two configs:
> - hidden CONFIG_BPF to select eBPF interpreter that classic socket filters
> depend on
> - visible CONFIG_BPF_SYSCALL (default off) that tracing and sockets can use
>
> that solves several problems:
> - tracing and others that wish to use eBPF don't need to depend on NET.
> They can use BPF_SYSCALL to allow loading from userspace or select BPF
> to use it directly from kernel in NET-less configs.
> - in 3.18 programs cannot be attached to events yet, so don't force it on
> - when the rest of eBPF infra is there in 3.19+, it's still useful to
> switch it off to minimize kernel size
>
> Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
> ---
>
> bloat-o-meter on x64 shows:
> add/remove: 0/60 grow/shrink: 0/2 up/down: 0/-15601 (-15601)
>
> tested with many different config combinations. Hopefully didn't miss anything.
Applied with two changes:
1) boolean --> bool
2) Moved bloat-o-meter and testing information into commit message.
Thanks.
^ permalink raw reply
* Re: [Bug 86851] New: Reproducible panic on heavy UDP traffic
From: Nikolay Aleksandrov @ 2014-10-27 23:06 UTC (permalink / raw)
To: Patrick McLean; +Cc: Eric Dumazet, Florian Westphal, Stephen Hemminger, netdev
In-Reply-To: <20141027155938.28248b5e@gentoo.org>
On 10/27/2014 11:59 PM, Patrick McLean wrote:
> On Mon, 27 Oct 2014 09:48:15 +0100
> Nikolay Aleksandrov <nikolay@redhat.com> wrote:
>
>> On 10/27/2014 01:47 AM, Eric Dumazet wrote:
>>> On Mon, 2014-10-27 at 00:28 +0100, Nikolay Aleksandrov wrote:
>>>
>>>>
>>>> Thanks for CCing me.
>>>> I'll dig in the code tomorrow but my first thought when I saw this
>>>> was could it be possible that we have a race condition between
>>>> ip_frag_queue() and inet_frag_evict(), more precisely between the
>>>> ipq_kill() calls from ip_frag_queue and inet_frag_evict since the
>>>> frag could be found before we have entered the evictor which then
>>>> can add it to its expire list but the ipq_kill() from
>>>> ip_frag_queue() can do a list_del after we release the chain lock
>>>> in the evictor so we may end up like this ?
>>>
>>> Yes, either we use hlist_del_init() but loose poison aid, or test if
>>> frag was evicted :
>>>
>>> Not sure about refcount.
>>>
>>> diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
>>> index 9eb89f3f0ee4..894ec30c5896 100644
>>> --- a/net/ipv4/inet_fragment.c
>>> +++ b/net/ipv4/inet_fragment.c
>>> @@ -285,7 +285,8 @@ static inline void fq_unlink(struct
>>> inet_frag_queue *fq, struct inet_frags *f) struct inet_frag_bucket
>>> *hb;
>>> hb = get_frag_bucket_locked(fq, f);
>>> - hlist_del(&fq->list);
>>> + if (!(fq->flags & INET_FRAG_EVICTED))
>>> + hlist_del(&fq->list);
>>> spin_unlock(&hb->chain_lock);
>>> }
>>>
>>>
>>>
>>
>> Exactly, I was thinking about a similar fix since the evict flag is
>> only set with the chain lock. IMO the refcount should be fine.
>> CCing the reporter.
>> Patrick could you please try Eric's patch ?
>>
>
> It no longer panics with that patch, but it does produce a large amount
> of warnings, here is an example of what I am getting. I will attach the
> full log to the bug.
>
Great! Thanks for testing.
As I said earlier we have a valid case that can hit the WARN_ON in
inet_evict_frag().
Anyhow, Eric would you mind posting the patch officially ?
If you'd like me to remove the WARN_ON() in a separate one just let me
know, otherwise feel free to remove it in the fix for the race.
Cheers,
Nik
^ permalink raw reply
* Re: [PATCH] net: ethernet: realtek: atp: checkpatch errors and warnings corrected
From: Roberto Medina @ 2014-10-27 23:04 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-kernel
In-Reply-To: <20141027.185355.1412632964802440855.davem@davemloft.net>
On 10/27/2014 11:53 PM, David Miller wrote:
> From: David Miller <davem@davemloft.net>
> Date: Mon, 27 Oct 2014 18:52:11 -0400 (EDT)
>
>> From: Roberto Medina <robertoxmed@gmail.com>
>> Date: Thu, 23 Oct 2014 19:10:00 +0200
>>
>>> From: Roberto Medina <robertoxmed@gmail.com>
>>>
>>> Several warnings and errors of coding style errors corrected.
>>>
>>> Signed-off-by: Roberto Medina <robertoxmed@gmail.com>
>>
>> Applied, thanks.
>
> Wait, are you serious, you didn't even compile test this change?
>
> For real?
>
> In file included from drivers/net/ethernet/realtek/atp.c:146:0:
> drivers/net/ethernet/realtek/atp.h:79:15: error: expected ‘;’, identifier or ‘(’ before ‘unsigned’
> static inline unsigned char inbyte(unsigned short port)
> ^
> drivers/net/ethernet/realtek/atp.h:79:15: error: ‘inline’ in empty declaration
>
Wow I'm really sorry for that, I did compile my changes, or at least
that's what I thought when I compiled the directory.
Will submit another patch. Sorry for the inconvenience again.
> Don't even submit changes meant for inclusion if you're not even
> willing to build test them.
>
> And I'm very serious about this.
>
^ permalink raw reply
* Re: [bug] sunhme unable to receive IPv4 unicasts
From: David Miller @ 2014-10-27 23:03 UTC (permalink / raw)
To: WIMPy; +Cc: netdev
In-Reply-To: <op.xn7d94v644i3fq@lx3.fl.yeti.dk>
From: "Birger Harzenetter" <WIMPy@yeti.dk>
Date: Fri, 24 Oct 2014 00:04:42 +0200
> from https://bugzilla.kernel.org/show_bug.cgi?id=86731
>
> Any kind of transmission is fine.
> Reception of IPv4 broadcast works
> Reception of non-IP works (pppoe tested)
> But IPv4 unicasts don't show up at all.
> IPv6 not tested
>
> Additional tests:
> I tried to set promiscous mode, but still no trace of unicasts seen with
> tcpdump.
>
> Last known working version: 3.15.10
Unfortunately there have been zero functional changes to this driver
since v3.15 :-/
^ permalink raw reply
* Re: [PATCH] bridge: Add support for IEEE 802.11 Proxy ARP
From: David Miller @ 2014-10-27 23:02 UTC (permalink / raw)
To: kyeyoonp; +Cc: jouni, netdev
In-Reply-To: <1414100957-8288-1-git-send-email-kyeyoonp@qca.qualcomm.com>
From: Kyeyoon Park <kyeyoonp@qca.qualcomm.com>
Date: Thu, 23 Oct 2014 14:49:17 -0700
> From: Kyeyoon Park <kyeyoonp@codeaurora.org>
>
> This feature is defined in IEEE Std 802.11-2012, 10.23.13. It allows
> the AP devices to keep track of the hardware-address-to-IP-address
> mapping of the mobile devices within the WLAN network.
>
> The AP will learn this mapping via observing DHCP, ARP, and NS/NA
> frames. When a request for such information is made (i.e. ARP request,
> Neighbor Solicitation), the AP will respond on behalf of the
> associated mobile device. In the process of doing so, the AP will drop
> the multicast request frame that was intended to go out to the wireless
> medium.
>
> It was recommended at the LKS workshop to do this implementation in
> the bridge layer. vxlan.c is already doing something very similar.
> The DHCP snooping code will be added to the userspace application
> (hostapd) per the recommendation.
>
> This RFC commit is only for IPv4. A similar approach in the bridge
> layer will be taken for IPv6 as well.
>
> Signed-off-by: Kyeyoon Park <kyeyoonp@codeaurora.org>
Looks good to me, applied, thanks.
^ permalink raw reply
* Re: [PATCH net 0/2] cxgb4 : DCBx fixes for apps/host lldp agents
From: David Miller @ 2014-10-27 23:00 UTC (permalink / raw)
To: anish; +Cc: netdev, hariprasad, leedom
In-Reply-To: <1414100251-15702-1-git-send-email-anish@chelsio.com>
From: Anish Bhatt <anish@chelsio.com>
Date: Thu, 23 Oct 2014 14:37:29 -0700
> This patchset contains some minor fixes for cxgb4 DCBx code. Chiefly, cxgb4
> was not cleaning up any apps added to kernel app table when link was lost.
> Disabling DCBx in firmware would automatically set DCBx state to host-managed
> and enabled, we now wait for an explicit enable call from an lldp agent instead
>
> First patch was originally sent to net-next, but considering it applies to
> correcting behaviour of code already in net, I think it qualifies as a bug fix.
> -Anish
Series applied, thanks Anish.
^ permalink raw reply
* Re: [Bug 86851] New: Reproducible panic on heavy UDP traffic
From: Patrick McLean @ 2014-10-27 22:59 UTC (permalink / raw)
To: Nikolay Aleksandrov
Cc: Eric Dumazet, Florian Westphal, Stephen Hemminger, netdev
In-Reply-To: <544E06CF.30709@redhat.com>
On Mon, 27 Oct 2014 09:48:15 +0100
Nikolay Aleksandrov <nikolay@redhat.com> wrote:
> On 10/27/2014 01:47 AM, Eric Dumazet wrote:
> > On Mon, 2014-10-27 at 00:28 +0100, Nikolay Aleksandrov wrote:
> >
> >>
> >> Thanks for CCing me.
> >> I'll dig in the code tomorrow but my first thought when I saw this
> >> was could it be possible that we have a race condition between
> >> ip_frag_queue() and inet_frag_evict(), more precisely between the
> >> ipq_kill() calls from ip_frag_queue and inet_frag_evict since the
> >> frag could be found before we have entered the evictor which then
> >> can add it to its expire list but the ipq_kill() from
> >> ip_frag_queue() can do a list_del after we release the chain lock
> >> in the evictor so we may end up like this ?
> >
> > Yes, either we use hlist_del_init() but loose poison aid, or test if
> > frag was evicted :
> >
> > Not sure about refcount.
> >
> > diff --git a/net/ipv4/inet_fragment.c b/net/ipv4/inet_fragment.c
> > index 9eb89f3f0ee4..894ec30c5896 100644
> > --- a/net/ipv4/inet_fragment.c
> > +++ b/net/ipv4/inet_fragment.c
> > @@ -285,7 +285,8 @@ static inline void fq_unlink(struct
> > inet_frag_queue *fq, struct inet_frags *f) struct inet_frag_bucket
> > *hb;
> > hb = get_frag_bucket_locked(fq, f);
> > - hlist_del(&fq->list);
> > + if (!(fq->flags & INET_FRAG_EVICTED))
> > + hlist_del(&fq->list);
> > spin_unlock(&hb->chain_lock);
> > }
> >
> >
> >
>
> Exactly, I was thinking about a similar fix since the evict flag is
> only set with the chain lock. IMO the refcount should be fine.
> CCing the reporter.
> Patrick could you please try Eric's patch ?
>
It no longer panics with that patch, but it does produce a large amount
of warnings, here is an example of what I am getting. I will attach the
full log to the bug.
> [ 205.042923] ------------[ cut here ]------------
> [ 205.042933] WARNING: CPU: 4 PID: 615 at net/ipv4/inet_fragment.c:149 inet_evict_bucket+0x172/0x180()
> [ 205.042934] Modules linked in: nfs fscache nfsd auth_rpcgss nfs_acl lockd grace sunrpc 8021q garp mrp bonding x86_pkg_temp_thermal joydev sb_edac edac_core ioatdma tpm_tis ext4 mbcache jbd2 igb ixgbe i2c_algo_bit raid1 mdio crc32c_intel megaraid_sas dca
> [ 205.042953] CPU: 4 PID: 615 Comm: kworker/4:2 Not tainted 3.18.0-rc2-base-7+ #3
> [ 205.042955] Hardware name: Intel Corporation S2600GZ/S2600GZ, BIOS SE5C600.86B.02.03.0003.041920141333 04/19/2014
> [ 205.042957] Workqueue: events inet_frag_worker
> [ 205.042958] 0000000000000000 0000000000000009 ffffffff81624cd2 0000000000000000
> [ 205.042960] ffffffff81117b7d ffff8817c83a4740 0000000000000000 ffffffff81aa6820
> [ 205.042962] ffff8817ce073d70 ffff8817c83a4738 ffffffff81597cb2 ffffffff81aa8e28
> [ 205.042964] Call Trace:
> [ 205.042969] [<ffffffff81624cd2>] ? dump_stack+0x41/0x51
> [ 205.042973] [<ffffffff81117b7d>] ? warn_slowpath_common+0x6d/0x90
> [ 205.042975] [<ffffffff81597cb2>] ? inet_evict_bucket+0x172/0x180
> [ 205.042976] [<ffffffff81597d22>] ? inet_frag_worker+0x62/0x210
> [ 205.042979] [<ffffffff8112c312>] ? process_one_work+0x132/0x360
> [ 205.042981] [<ffffffff8112ca23>] ? worker_thread+0x113/0x590
> [ 205.042983] [<ffffffff8112c910>] ? rescuer_thread+0x3d0/0x3d0
> [ 205.042986] [<ffffffff8113123c>] ? kthread+0xbc/0xe0
> [ 205.042991] [<ffffffff81040000>] ? xen_teardown_timer+0x10/0x70
> [ 205.042993] [<ffffffff81131180>] ? kthread_create_on_node+0x170/0x170
> [ 205.042996] [<ffffffff8162a9fc>] ? ret_from_fork+0x7c/0xb0
> [ 205.042998] [<ffffffff81131180>] ? kthread_create_on_node+0x170/0x170
> [ 205.043000] ---[ end trace ed2bb7d412e082bc ]---
> [ 205.752744] ------------[ cut here ]------------
> [ 205.752752] WARNING: CPU: 2 PID: 610 at net/ipv4/inet_fragment.c:149 inet_evict_bucket+0x172/0x180()
> [ 205.752754] Modules linked in: nfs fscache nfsd auth_rpcgss nfs_acl lockd grace sunrpc 8021q garp mrp bonding x86_pkg_temp_thermal joydev sb_edac edac_core ioatdma tpm_tis ext4 mbcache jbd2 igb ixgbe i2c_algo_bit raid1 mdio crc32c_intel megaraid_sas dca
> [ 205.752773] CPU: 2 PID: 610 Comm: kworker/2:2 Tainted: G W 3.18.0-rc2-base-7+ #3
> [ 205.752774] Hardware name: Intel Corporation S2600GZ/S2600GZ, BIOS SE5C600.86B.02.03.0003.041920141333 04/19/2014
> [ 205.752777] Workqueue: events inet_frag_worker
> [ 205.752779] 0000000000000000 0000000000000009 ffffffff81624cd2 0000000000000000
> [ 205.752780] ffffffff81117b7d ffff882fc473c740 0000000000000000 ffffffff81aa6820
> [ 205.752782] ffff8817ce7afd70 ffff882fc473c738 ffffffff81597cb2 ffffffff81aa87a8
> [ 205.752784] Call Trace:
> [ 205.752790] [<ffffffff81624cd2>] ? dump_stack+0x41/0x51
> [ 205.752793] [<ffffffff81117b7d>] ? warn_slowpath_common+0x6d/0x90
> [ 205.752795] [<ffffffff81597cb2>] ? inet_evict_bucket+0x172/0x180
> [ 205.752797] [<ffffffff81597d22>] ? inet_frag_worker+0x62/0x210
> [ 205.752799] [<ffffffff8112c312>] ? process_one_work+0x132/0x360
> [ 205.752801] [<ffffffff8112ca23>] ? worker_thread+0x113/0x590
> [ 205.752803] [<ffffffff8112c910>] ? rescuer_thread+0x3d0/0x3d0
> [ 205.752806] [<ffffffff8113123c>] ? kthread+0xbc/0xe0
> [ 205.752810] [<ffffffff81040000>] ? xen_teardown_timer+0x10/0x70
> [ 205.752812] [<ffffffff81131180>] ? kthread_create_on_node+0x170/0x170
> [ 205.752815] [<ffffffff8162a9fc>] ? ret_from_fork+0x7c/0xb0
> [ 205.752818] [<ffffffff81131180>] ? kthread_create_on_node+0x170/0x170
> [ 205.752820] ---[ end trace ed2bb7d412e082bd ]---
> [ 206.737865] ------------[ cut here ]------------
^ permalink raw reply
* Re: [net 1/2] sctp: add transport state in /proc/net/sctp/remaddr
From: David Miller @ 2014-10-27 22:55 UTC (permalink / raw)
To: michele; +Cc: linux-sctp, vyasevich, nhorman, netdev, dborkman
In-Reply-To: <1414093721-14921-1-git-send-email-michele@acksyn.org>
From: Michele Baldessari <michele@acksyn.org>
Date: Thu, 23 Oct 2014 21:48:40 +0200
> It is often quite helpful to be able to know the state of a transport
> outside of the application itself (for troubleshooting purposes or for
> monitoring purposes). Add it under /proc/net/sctp/remaddr.
>
> Signed-off-by: Michele Baldessari <michele@acksyn.org>
You can't change the layout of procfs files, applications parse
these files and any modification can potentially break such tools.
Secondly, even if this change were acceptable, targetting this
change at anything other than the net-next tree is not appropriate
because it is a new feature.
^ permalink raw reply
* Re: [PATCH] net: ethernet: realtek: atp: checkpatch errors and warnings corrected
From: David Miller @ 2014-10-27 22:53 UTC (permalink / raw)
To: robertoxmed; +Cc: netdev, linux-kernel
In-Reply-To: <20141027.185211.1433818002293610983.davem@davemloft.net>
From: David Miller <davem@davemloft.net>
Date: Mon, 27 Oct 2014 18:52:11 -0400 (EDT)
> From: Roberto Medina <robertoxmed@gmail.com>
> Date: Thu, 23 Oct 2014 19:10:00 +0200
>
>> From: Roberto Medina <robertoxmed@gmail.com>
>>
>> Several warnings and errors of coding style errors corrected.
>>
>> Signed-off-by: Roberto Medina <robertoxmed@gmail.com>
>
> Applied, thanks.
Wait, are you serious, you didn't even compile test this change?
For real?
In file included from drivers/net/ethernet/realtek/atp.c:146:0:
drivers/net/ethernet/realtek/atp.h:79:15: error: expected ‘;’, identifier or ‘(’ before ‘unsigned’
static inline unsigned char inbyte(unsigned short port)
^
drivers/net/ethernet/realtek/atp.h:79:15: error: ‘inline’ in empty declaration
Don't even submit changes meant for inclusion if you're not even
willing to build test them.
And I'm very serious about this.
^ permalink raw reply
* Re: [PATCH] net: ethernet: realtek: atp: checkpatch errors and warnings corrected
From: David Miller @ 2014-10-27 22:52 UTC (permalink / raw)
To: robertoxmed; +Cc: netdev, linux-kernel
In-Reply-To: <1414084200-5377-1-git-send-email-robertoxmed@gmail.com>
From: Roberto Medina <robertoxmed@gmail.com>
Date: Thu, 23 Oct 2014 19:10:00 +0200
> From: Roberto Medina <robertoxmed@gmail.com>
>
> Several warnings and errors of coding style errors corrected.
>
> Signed-off-by: Roberto Medina <robertoxmed@gmail.com>
Applied, thanks.
^ permalink raw reply
* Re: [PATCH 0/8] Netfilter fixes for net
From: David Miller @ 2014-10-27 22:49 UTC (permalink / raw)
To: pablo; +Cc: netfilter-devel, netdev
In-Reply-To: <1414445887-5108-1-git-send-email-pablo@netfilter.org>
From: Pablo Neira Ayuso <pablo@netfilter.org>
Date: Mon, 27 Oct 2014 22:37:59 +0100
> The following patchset contains Netfilter fixes for your net tree,
> they are:
...
> You can pull these changes from:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf.git
Pulled, thanks a lot Pablo!
^ permalink raw reply
* Re: [PATCH v3] ipv6: notify userspace when we added or changed an ipv6 token
From: Daniel Borkmann @ 2014-10-27 22:25 UTC (permalink / raw)
To: Lubomir Rintel; +Cc: netdev, David S. Miller, Hannes Frederic Sowa
In-Reply-To: <1414427956-20056-1-git-send-email-lkundrak@v3.sk>
On 10/27/2014 05:39 PM, Lubomir Rintel wrote:
> NetworkManager might want to know that it changed when the router advertisement
> arrives.
>
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
> Cc: Daniel Borkmann <dborkman@redhat.com>
Looks better, thanks!
Acked-by: Daniel Borkmann <dborkman@redhat.com>
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox