* Re: [net-next 0/8][pull request] Intel Wired LAN Driver Updates
From: David Miller @ 2012-07-17 10:22 UTC (permalink / raw)
To: jeffrey.t.kirsher; +Cc: netdev, gospo, sassmann
In-Reply-To: <1342519759-25137-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Tue, 17 Jul 2012 03:09:11 -0700
> This series contains updates to ixgbevf.
>
> The following are changes since commit 282f23c6ee343126156dd41218b22ece96d747e3:
> tcp: implement RFC 5961 3.2
> and are available in the git repository at:
> git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
>
> Alexander Duyck (8):
> ixgbevf: Drop all dead or unnecessary code
> ixgbevf: Drop netdev_registered value since that is already stored in
> netdev
> ixgbevf: Make use of NETIF_F_RXCSUM instead of keeping our own flag
> ixgbevf: Drop use of eitr_low and eitr_high for hard coded values
> ixgbevf: Cleanup accounting for space needed at start of xmit_frame
> ixgbevf: Update q_vector to contain ring pointers instead of bitmaps
> ixgbevf: Move Tx clean-up into NAPI context
> ixgbevf: Use igb style interrupt masks instead of ixgbe style
Pulled, thanks Jeff.
^ permalink raw reply
* [PATCH] libertas: remove duplicated include
From: djduanjiong @ 2012-07-17 10:32 UTC (permalink / raw)
To: linville; +Cc: netdev, Duan Jiong
From: Duan Jiong <djduanjiong@gmail.com>
Signed-off-by: Duan Jiong <djduanjiong@gmail.com>
---
drivers/net/wireless/libertas/firmware.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/drivers/net/wireless/libertas/firmware.c b/drivers/net/wireless/libertas/firmware.c
index 601f207..c0f9e7e 100644
--- a/drivers/net/wireless/libertas/firmware.c
+++ b/drivers/net/wireless/libertas/firmware.c
@@ -4,9 +4,7 @@
#include <linux/sched.h>
#include <linux/firmware.h>
-#include <linux/firmware.h>
#include <linux/module.h>
-#include <linux/sched.h>
#include "dev.h"
#include "decl.h"
--
1.7.9.5
^ permalink raw reply related
* [net-next PATCH v7] net: ethernet: davinci_emac: add OF support
From: Anatolij Gustschin @ 2012-07-17 10:34 UTC (permalink / raw)
To: netdev
Cc: David Miller, Heiko Schocher, davinci-linux-open-source,
linux-arm-kernel, devicetree-discuss, Grant Likely, Sekhar Nori,
Wolfgang Denk, Anatoly Sivov
From: Heiko Schocher <hs@denx.de>
add OF support for the davinci_emac driver.
Signed-off-by: Heiko Schocher <hs@denx.de>
Acked-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Anatolij Gustschin <agust@denx.de>
Cc: netdev@vger.kernel.org
Cc: davinci-linux-open-source@linux.davincidsp.com
Cc: linux-arm-kernel@lists.infradead.org
Cc: devicetree-discuss@lists.ozlabs.org
Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Sekhar Nori <nsekhar@ti.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Anatoly Sivov <mm05@mail.ru>
Cc: David Miller <davem@davemloft.net>
---
- changes for v2:
- add comment from Anatoly Sivov
- fix typo in davinci_emac.txt
- add comment from Grant Likely:
- add prefix "ti,davinci-" to davinci specific property names
- remove version property
- use compatible name "ti,davinci-dm6460-emac"
- use devm_kzalloc()
- use of_match_ptr()
- document all new properties
- remove of_address_to_resource() and do not overwrite
resource table
- whitespace fixes
- remove hw_ram_addr as it is not used in current
board code
- no changes for v3
- changes for v4:
add comments from Nori Sekhar:
- move devictree documentation to:
Documentation/devicetree/bindings/net/davinci_emac.txt
- fix typo in it
- rename compatible property to "ti,davinci-dm6467-emac"
- remove pinmux-handle
- set version directly in pdata->version
- no changes for v5
- changes for v6:
add comment from Nori, Sekhar:
- use mac address in DT data only if no valid address is passed
through platform data
- added Acked-by from Sekhar Nori
- changes Subject line from "ARM: davinci: net:" to
"net, ethernet, davinci_emac:"
- add David Miller to Cc
- changes for v7:
- rebased to apply on current net-next tree
- slightly change the subject line and commit log
.../devicetree/bindings/net/davinci_emac.txt | 41 +++++++++
drivers/net/ethernet/ti/davinci_emac.c | 89 +++++++++++++++++++-
2 files changed, 129 insertions(+), 1 deletions(-)
create mode 100644 Documentation/devicetree/bindings/net/davinci_emac.txt
diff --git a/Documentation/devicetree/bindings/net/davinci_emac.txt b/Documentation/devicetree/bindings/net/davinci_emac.txt
new file mode 100644
index 0000000..48b259e
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/davinci_emac.txt
@@ -0,0 +1,41 @@
+* Texas Instruments Davinci EMAC
+
+This file provides information, what the device node
+for the davinci_emac interface contains.
+
+Required properties:
+- compatible: "ti,davinci-dm6467-emac";
+- reg: Offset and length of the register set for the device
+- ti,davinci-ctrl-reg-offset: offset to control register
+- ti,davinci-ctrl-mod-reg-offset: offset to control module register
+- ti,davinci-ctrl-ram-offset: offset to control module ram
+- ti,davinci-ctrl-ram-size: size of control module ram
+- ti,davinci-rmii-en: use RMII
+- ti,davinci-no-bd-ram: has the emac controller BD RAM
+- phy-handle: Contains a phandle to an Ethernet PHY.
+ if not, davinci_emac driver defaults to 100/FULL
+- interrupts: interrupt mapping for the davinci emac interrupts sources:
+ 4 sources: <Receive Threshold Interrupt
+ Receive Interrupt
+ Transmit Interrupt
+ Miscellaneous Interrupt>
+
+Optional properties:
+- local-mac-address : 6 bytes, mac address
+
+Example (enbw_cmc board):
+ eth0: emac@1e20000 {
+ compatible = "ti,davinci-dm6467-emac";
+ reg = <0x220000 0x4000>;
+ ti,davinci-ctrl-reg-offset = <0x3000>;
+ ti,davinci-ctrl-mod-reg-offset = <0x2000>;
+ ti,davinci-ctrl-ram-offset = <0>;
+ ti,davinci-ctrl-ram-size = <0x2000>;
+ local-mac-address = [ 00 00 00 00 00 00 ];
+ interrupts = <33
+ 34
+ 35
+ 36
+ >;
+ interrupt-parent = <&intc>;
+ };
diff --git a/drivers/net/ethernet/ti/davinci_emac.c b/drivers/net/ethernet/ti/davinci_emac.c
index ab0bbb7..b298ab0 100644
--- a/drivers/net/ethernet/ti/davinci_emac.c
+++ b/drivers/net/ethernet/ti/davinci_emac.c
@@ -58,6 +58,12 @@
#include <linux/io.h>
#include <linux/uaccess.h>
#include <linux/davinci_emac.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_net.h>
+
+#include <mach/mux.h>
#include <asm/irq.h>
#include <asm/page.h>
@@ -339,6 +345,9 @@ struct emac_priv {
u32 rx_addr_type;
atomic_t cur_tx;
const char *phy_id;
+#ifdef CONFIG_OF
+ struct device_node *phy_node;
+#endif
struct phy_device *phydev;
spinlock_t lock;
/*platform specific members*/
@@ -1760,6 +1769,77 @@ static const struct net_device_ops emac_netdev_ops = {
#endif
};
+#ifdef CONFIG_OF
+static struct emac_platform_data
+ *davinci_emac_of_get_pdata(struct platform_device *pdev,
+ struct emac_priv *priv)
+{
+ struct device_node *np;
+ struct emac_platform_data *pdata = NULL;
+ const u8 *mac_addr;
+ u32 data;
+ int ret;
+
+ pdata = pdev->dev.platform_data;
+ if (!pdata) {
+ pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+ if (!pdata)
+ goto nodata;
+ }
+
+ np = pdev->dev.of_node;
+ if (!np)
+ goto nodata;
+ else
+ pdata->version = EMAC_VERSION_2;
+
+ if (!is_valid_ether_addr(pdata->mac_addr)) {
+ mac_addr = of_get_mac_address(np);
+ if (mac_addr)
+ memcpy(pdata->mac_addr, mac_addr, ETH_ALEN);
+ }
+
+ ret = of_property_read_u32(np, "ti,davinci-ctrl-reg-offset", &data);
+ if (!ret)
+ pdata->ctrl_reg_offset = data;
+
+ ret = of_property_read_u32(np, "ti,davinci-ctrl-mod-reg-offset",
+ &data);
+ if (!ret)
+ pdata->ctrl_mod_reg_offset = data;
+
+ ret = of_property_read_u32(np, "ti,davinci-ctrl-ram-offset", &data);
+ if (!ret)
+ pdata->ctrl_ram_offset = data;
+
+ ret = of_property_read_u32(np, "ti,davinci-ctrl-ram-size", &data);
+ if (!ret)
+ pdata->ctrl_ram_size = data;
+
+ ret = of_property_read_u32(np, "ti,davinci-rmii-en", &data);
+ if (!ret)
+ pdata->rmii_en = data;
+
+ ret = of_property_read_u32(np, "ti,davinci-no-bd-ram", &data);
+ if (!ret)
+ pdata->no_bd_ram = data;
+
+ priv->phy_node = of_parse_phandle(np, "phy-handle", 0);
+ if (!priv->phy_node)
+ pdata->phy_id = "";
+
+ pdev->dev.platform_data = pdata;
+nodata:
+ return pdata;
+}
+#else
+static struct emac_platform_data
+ *davinci_emac_of_get_pdata(struct platform_device *pdev,
+ struct emac_priv *priv)
+{
+ return pdev->dev.platform_data;
+}
+#endif
/**
* davinci_emac_probe - EMAC device probe
* @pdev: The DaVinci EMAC device that we are removing
@@ -1802,7 +1882,7 @@ static int __devinit davinci_emac_probe(struct platform_device *pdev)
spin_lock_init(&priv->lock);
- pdata = pdev->dev.platform_data;
+ pdata = davinci_emac_of_get_pdata(pdev, priv);
if (!pdata) {
dev_err(&pdev->dev, "no platform data\n");
rc = -ENODEV;
@@ -2013,12 +2093,19 @@ static const struct dev_pm_ops davinci_emac_pm_ops = {
.resume = davinci_emac_resume,
};
+static const struct of_device_id davinci_emac_of_match[] = {
+ {.compatible = "ti,davinci-dm6467-emac", },
+ {},
+};
+MODULE_DEVICE_TABLE(of, davinci_emac_of_match);
+
/* davinci_emac_driver: EMAC platform driver structure */
static struct platform_driver davinci_emac_driver = {
.driver = {
.name = "davinci_emac",
.owner = THIS_MODULE,
.pm = &davinci_emac_pm_ops,
+ .of_match_table = of_match_ptr(davinci_emac_of_match),
},
.probe = davinci_emac_probe,
.remove = __devexit_p(davinci_emac_remove),
--
1.7.1
^ permalink raw reply related
* Re: [PATCH v6 4/7] net, ethernet, davinci_emac: add OF support
From: Anatolij Gustschin @ 2012-07-17 10:37 UTC (permalink / raw)
To: Sekhar Nori
Cc: David Miller, Heiko Schocher, davinci-linux-open-source,
linux-arm-kernel, devicetree-discuss, netdev, Grant Likely,
Wolfgang Denk, Anatoly Sivov
In-Reply-To: <50044F05.6060707@ti.com>
Hi,
On Mon, 16 Jul 2012 22:57:33 +0530
Sekhar Nori <nsekhar@ti.com> wrote:
> Hi Dave,
>
> On 7/9/2012 2:14 PM, Heiko Schocher wrote:
> > add of support for the davinci_emac driver.
> >
> > Signed-off-by: Heiko Schocher <hs@denx.de>
> > Acked-by: Sekhar Nori <nsekhar@ti.com>
> > Cc: davinci-linux-open-source@linux.davincidsp.com
> > Cc: linux-arm-kernel@lists.infradead.org
> > Cc: devicetree-discuss@lists.ozlabs.org
> > Cc: netdev@vger.kernel.org
> > Cc: Grant Likely <grant.likely@secretlab.ca>
> > Cc: Sekhar Nori <nsekhar@ti.com>
> > Cc: Wolfgang Denk <wd@denx.de>
> > Cc: Anatoly Sivov <mm05@mail.ru>
> > Cc: David Miller <davem@davemloft.net>
>
> Can you please consider this patch for v3.6? I tested it on DaVinci
> AM18x EVM with and without CONFIG_OF using NFS root.
>
> This patch can be independently queued and does not have any dependencies.
unfortunately the patch didn't apply on net-next tree. I've
send a rebased patch.
Thanks,
Anatolij
^ permalink raw reply
* [PATCH net-next] tcp: implement RFC 5961 4.2
From: Eric Dumazet @ 2012-07-17 11:41 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Kiran Kumar Kella
From: Eric Dumazet <edumazet@google.com>
Implement the RFC 5691 mitigation against Blind
Reset attack using SYN bit.
Section 4.2 of RFC 5961 advises to send a Challenge ACK and drop
incoming packet, instead of resetting the session.
Add a new SNMP counter to count number of challenge acks sent
in response to SYN packets.
(netstat -s | grep TCPSYNChallenge)
Remove obsolete TCPAbortOnSyn, since we no longer abort a TCP session
because of a SYN flag.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Kiran Kumar Kella <kkiran@broadcom.com>
---
include/linux/snmp.h | 2 +-
net/ipv4/proc.c | 2 +-
net/ipv4/tcp_input.c | 32 +++++++++++++++-----------------
3 files changed, 17 insertions(+), 19 deletions(-)
diff --git a/include/linux/snmp.h b/include/linux/snmp.h
index 673e0e9..e5fcbd0 100644
--- a/include/linux/snmp.h
+++ b/include/linux/snmp.h
@@ -208,7 +208,6 @@ enum
LINUX_MIB_TCPDSACKOFOSENT, /* TCPDSACKOfoSent */
LINUX_MIB_TCPDSACKRECV, /* TCPDSACKRecv */
LINUX_MIB_TCPDSACKOFORECV, /* TCPDSACKOfoRecv */
- LINUX_MIB_TCPABORTONSYN, /* TCPAbortOnSyn */
LINUX_MIB_TCPABORTONDATA, /* TCPAbortOnData */
LINUX_MIB_TCPABORTONCLOSE, /* TCPAbortOnClose */
LINUX_MIB_TCPABORTONMEMORY, /* TCPAbortOnMemory */
@@ -238,6 +237,7 @@ enum
LINUX_MIB_TCPOFODROP, /* TCPOFODrop */
LINUX_MIB_TCPOFOMERGE, /* TCPOFOMerge */
LINUX_MIB_TCPCHALLENGEACK, /* TCPChallengeACK */
+ LINUX_MIB_TCPSYNCHALLENGE, /* TCPSYNChallenge */
__LINUX_MIB_MAX
};
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index 3e8e78f..2a5240b 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -232,7 +232,6 @@ static const struct snmp_mib snmp4_net_list[] = {
SNMP_MIB_ITEM("TCPDSACKOfoSent", LINUX_MIB_TCPDSACKOFOSENT),
SNMP_MIB_ITEM("TCPDSACKRecv", LINUX_MIB_TCPDSACKRECV),
SNMP_MIB_ITEM("TCPDSACKOfoRecv", LINUX_MIB_TCPDSACKOFORECV),
- SNMP_MIB_ITEM("TCPAbortOnSyn", LINUX_MIB_TCPABORTONSYN),
SNMP_MIB_ITEM("TCPAbortOnData", LINUX_MIB_TCPABORTONDATA),
SNMP_MIB_ITEM("TCPAbortOnClose", LINUX_MIB_TCPABORTONCLOSE),
SNMP_MIB_ITEM("TCPAbortOnMemory", LINUX_MIB_TCPABORTONMEMORY),
@@ -262,6 +261,7 @@ static const struct snmp_mib snmp4_net_list[] = {
SNMP_MIB_ITEM("TCPOFODrop", LINUX_MIB_TCPOFODROP),
SNMP_MIB_ITEM("TCPOFOMerge", LINUX_MIB_TCPOFOMERGE),
SNMP_MIB_ITEM("TCPChallengeACK", LINUX_MIB_TCPCHALLENGEACK),
+ SNMP_MIB_ITEM("TCPSYNChallenge", LINUX_MIB_TCPSYNCHALLENGE),
SNMP_MIB_SENTINEL
};
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index c841a89..8aaec55 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5270,8 +5270,8 @@ static void tcp_send_challenge_ack(struct sock *sk)
/* Does PAWS and seqno based validation of an incoming segment, flags will
* play significant role here.
*/
-static int tcp_validate_incoming(struct sock *sk, struct sk_buff *skb,
- const struct tcphdr *th, int syn_inerr)
+static bool tcp_validate_incoming(struct sock *sk, struct sk_buff *skb,
+ const struct tcphdr *th, int syn_inerr)
{
const u8 *hash_location;
struct tcp_sock *tp = tcp_sk(sk);
@@ -5323,20 +5323,22 @@ static int tcp_validate_incoming(struct sock *sk, struct sk_buff *skb,
/* step 3: check security and precedence [ignored] */
- /* step 4: Check for a SYN in window. */
- if (th->syn && !before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) {
+ /* step 4: Check for a SYN
+ * RFC 5691 4.2 : Send a challenge ack
+ */
+ if (th->syn) {
if (syn_inerr)
TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_INERRS);
- NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPABORTONSYN);
- tcp_reset(sk);
- return -1;
+ NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPSYNCHALLENGE);
+ tcp_send_challenge_ack(sk);
+ goto discard;
}
- return 1;
+ return true;
discard:
__kfree_skb(skb);
- return 0;
+ return false;
}
/*
@@ -5366,7 +5368,6 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
const struct tcphdr *th, unsigned int len)
{
struct tcp_sock *tp = tcp_sk(sk);
- int res;
if (sk->sk_rx_dst) {
struct dst_entry *dst = sk->sk_rx_dst;
@@ -5555,9 +5556,8 @@ slow_path:
* Standard slow path.
*/
- res = tcp_validate_incoming(sk, skb, th, 1);
- if (res <= 0)
- return -res;
+ if (!tcp_validate_incoming(sk, skb, th, 1))
+ return 0;
step5:
if (th->ack && tcp_ack(sk, skb, FLAG_SLOWPATH) < 0)
@@ -5877,7 +5877,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
struct tcp_sock *tp = tcp_sk(sk);
struct inet_connection_sock *icsk = inet_csk(sk);
int queued = 0;
- int res;
tp->rx_opt.saw_tstamp = 0;
@@ -5932,9 +5931,8 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
return 0;
}
- res = tcp_validate_incoming(sk, skb, th, 0);
- if (res <= 0)
- return -res;
+ if (!tcp_validate_incoming(sk, skb, th, 0))
+ return 0;
/* step 5: check the ACK field */
if (th->ack) {
^ permalink raw reply related
* [PATCH] [RFC] tcp: TSQ - do not always throttle.
From: Krishna Kumar @ 2012-07-17 12:03 UTC (permalink / raw)
To: davem, eric.dumazet; +Cc: netdev, Krishna Kumar
Do not throttle if sysctl_tcp_limit_output_bytes==0.
Maybe it is better to throttle earlier in the loop, after
calling tcp_init_tso_segs().
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
---
tcp_output.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff -ruNp org/net/ipv4/tcp_output.c new/net/ipv4/tcp_output.c
--- org/net/ipv4/tcp_output.c 2012-07-17 09:56:12.000000000 +0530
+++ new/net/ipv4/tcp_output.c 2012-07-17 13:02:12.476111697 +0530
@@ -1948,7 +1948,8 @@ static bool tcp_write_xmit(struct sock *
/* TSQ : sk_wmem_alloc accounts skb truesize,
* including skb overhead. But thats OK.
*/
- if (atomic_read(&sk->sk_wmem_alloc) >= sysctl_tcp_limit_output_bytes) {
+ if (sysctl_tcp_limit_output_bytes > 0 &&
+ atomic_read(&sk->sk_wmem_alloc) >= sysctl_tcp_limit_output_bytes) {
set_bit(TSQ_THROTTLED, &tp->tsq_flags);
break;
}
^ permalink raw reply
* [PATCH] skbuff: Use correct allocation in skb_copy_ubufs
From: Krishna Kumar @ 2012-07-17 12:05 UTC (permalink / raw)
To: davem; +Cc: xma, netdev, Krishna Kumar
Use correct allocation flags during copy of user space fragments
to the kernel. Also "improve" couple of for loops.
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
---
skbuff.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff -ruNp org/net/core/skbuff.c new/net/core/skbuff.c
--- org/net/core/skbuff.c 2012-07-17 09:56:12.000000000 +0530
+++ new/net/core/skbuff.c 2012-07-17 11:05:43.715853844 +0530
@@ -751,7 +751,7 @@ int skb_copy_ubufs(struct sk_buff *skb,
u8 *vaddr;
skb_frag_t *f = &skb_shinfo(skb)->frags[i];
- page = alloc_page(GFP_ATOMIC);
+ page = alloc_page(gfp_mask);
if (!page) {
while (head) {
struct page *next = (struct page *)head->private;
@@ -769,15 +769,15 @@ int skb_copy_ubufs(struct sk_buff *skb,
}
/* skb frags release userspace buffers */
- for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
+ for (i = 0; i < num_frags; i++)
skb_frag_unref(skb, i);
uarg->callback(uarg);
/* skb frags point to kernel buffers */
- for (i = skb_shinfo(skb)->nr_frags; i > 0; i--) {
- __skb_fill_page_desc(skb, i-1, head, 0,
- skb_shinfo(skb)->frags[i - 1].size);
+ for (i = num_frags - 1; i >= 0; i--) {
+ __skb_fill_page_desc(skb, i, head, 0,
+ skb_shinfo(skb)->frags[i].size);
head = (struct page *)head->private;
}
^ permalink raw reply
* Re: [PATCH v2] sctp: Fix list corruption resulting from freeing an association on a list
From: Neil Horman @ 2012-07-17 12:25 UTC (permalink / raw)
To: David Miller; +Cc: netdev, davej, vyasevich, sri, linux-sctp
In-Reply-To: <20120716.223250.2238626170464909220.davem@davemloft.net>
On Mon, Jul 16, 2012 at 10:32:50PM -0700, David Miller wrote:
> From: Neil Horman <nhorman@tuxdriver.com>
> Date: Mon, 16 Jul 2012 15:13:51 -0400
>
> > A few days ago Dave Jones reported this oops:
> ...
> > It appears from his analysis and some staring at the code that this is likely
> > occuring because an association is getting freed while still on the
> > sctp_assoc_hashtable. As a result, we get a gpf when traversing the hashtable
> > while a freed node corrupts part of the list.
> >
> > Nominally I would think that an mibalanced refcount was responsible for this,
> > but I can't seem to find any obvious imbalance. What I did note however was
> > that the two places where we create an association using
> > sctp_primitive_ASSOCIATE (__sctp_connect and sctp_sendmsg), have failure paths
> > which free a newly created association after calling sctp_primitive_ASSOCIATE.
> > sctp_primitive_ASSOCIATE brings us into the sctp_sf_do_prm_asoc path, which
> > issues a SCTP_CMD_NEW_ASOC side effect, which in turn adds a new association to
> > the aforementioned hash table. the sctp command interpreter that process side
> > effects has not way to unwind previously processed commands, so freeing the
> > association from the __sctp_connect or sctp_sendmsg error path would lead to a
> > freed association remaining on this hash table.
> >
> > I've fixed this but modifying sctp_[un]hash_established to use hlist_del_init,
> > which allows us to proerly use hlist_unhashed to check if the node is on a
> > hashlist safely during a delete. That in turn alows us to safely call
> > sctp_unhash_established in the __sctp_connect and sctp_sendmsg error paths
> > before freeing them, regardles of what the associations state is on the hash
> > list.
> >
> > I noted, while I was doing this, that the __sctp_unhash_endpoint was using
> > hlist_unhsashed in a simmilar fashion, but never nullified any removed nodes
> > pointers to make that function work properly, so I fixed that up in a simmilar
> > fashion.
> >
> > I attempted to test this using a virtual guest running the SCTP_RR test from
> > netperf in a loop while running the trinity fuzzer, both in a loop. I wasn't
> > able to recreate the problem prior to this fix, nor was I able to trigger the
> > failure after (neither of which I suppose is suprising). Given the trace above
> > however, I think its likely that this is what we hit.
> >
> > Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> > Reported-by: davej@redhat.com
>
> Looks great, applied and queued up for -stable, thanks Neil.
>
Thanks Dave!
Neil
^ permalink raw reply
* RE: [PATCH] mlx4_en: map entire pages to increase throughput
From: David Laight @ 2012-07-17 12:42 UTC (permalink / raw)
To: David Miller, rick.jones2
Cc: cascardo, netdev, yevgenyp, ogerlitz, amirv, brking, leitao,
klebers
In-Reply-To: <20120716.222903.367603216293954363.davem@davemloft.net>
> > That seems rather extraordinarily low - Power7 is supposed to be a
> > rather high performance CPU. The last time I noticed O(3Gbit/s) on
> > 10G for bulk transfer was before the advent of LRO/GRO - that was in
> > the x86 space though. Is mapping really that expensive with Power7?
>
> Unfortunately, IOMMU mappings are incredibly expensive. I see effects
> like this on Sparc too.
Would there be any mileage in permanently allocating IOMMU
virtual address to the ring entries, then 'just' assigning
the correct physical address during rx/tx setup?
A long time ago it used to be much faster on sparc systems
to receive into a permanently mapped buffer area and then
do a maximally aligned copy into the actual rx buffer.
David
^ permalink raw reply
* Re: [PATCH] mlx4_en: map entire pages to increase throughput
From: David Miller @ 2012-07-17 12:50 UTC (permalink / raw)
To: David.Laight
Cc: rick.jones2, cascardo, netdev, yevgenyp, ogerlitz, amirv, brking,
leitao, klebers
In-Reply-To: <AE90C24D6B3A694183C094C60CF0A2F6026B6F8B@saturn3.aculab.com>
From: "David Laight" <David.Laight@ACULAB.COM>
Date: Tue, 17 Jul 2012 13:42:04 +0100
> Would there be any mileage in permanently allocating IOMMU
> virtual address to the ring entries, then 'just' assigning
> the correct physical address during rx/tx setup?
There is a not a one to one mapping between these two entities,
in particular on the transmit side.
A transmit packet can have multiple segments, some of which are
larger than one IOMMU page.
^ permalink raw reply
* Re: [PATCH] [RFC] tcp: TSQ - do not always throttle.
From: Eric Dumazet @ 2012-07-17 13:10 UTC (permalink / raw)
To: Krishna Kumar; +Cc: davem, netdev
In-Reply-To: <20120717120358.16611.98190.sendpatchset@localhost.localdomain>
On Tue, 2012-07-17 at 17:33 +0530, Krishna Kumar wrote:
> Do not throttle if sysctl_tcp_limit_output_bytes==0.
>
> Maybe it is better to throttle earlier in the loop, after
> calling tcp_init_tso_segs().
>
I wonder why, and why you put this question in a changelog instead of
outside of it...
Idea was to avoid setting TSQ_THROTTLED if we break out the loop.
About disabling TSQ, my initial intent was to instead use a negative
sysctl_tcp_limit_output_bytes value.
Thats why I have in tcp_transmit_skb() :
skb->destructor = (sysctl_tcp_limit_output_bytes > 0) ?
tcp_wfree : sock_wfree;
So I suggest you change the tcp_write_xmit(() test to a single unsigned
compare :
if (atomic_read(&sk->sk_wmem_alloc) >=
(unsigned) sysctl_tcp_limit_output_bytes) {
Also use :
skb->destructor = (sysctl_tcp_limit_output_bytes >= 0) ?
tcp_wfree : sock_wfree;
and document the 'negative value disables TSQ' in
Documentation/networking/ip-sysctl.txt
^ permalink raw reply
* [PATCH 0/5] Long term PMTU/redirect storage in ipv4.
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
These patches implement the final mechanism necessary to really allow
us to go without the route cache in ipv4.
We need a place to have long-term storage of PMTU/redirect information
which is independent of the routes themselves, yet does not get us
back into a situation where we have to write to metrics or anything
like that.
For this we use an "next-hop exception" table in the FIB nexthops.
Currently it is a simple linked list and uses a single global lock
for synchronization, but that can be easily adjusted as-needed.
The one thing I desperately want to avoid is having to create clone
routes in the FIB trie for this purpose, because that is very
expensive. However, I'm willing to entertain such an idea later
if this current scheme proves to have downsides that the FIB trie
variant would not have.
In order to accomodate this any such scheme, we need to be able to
produce a full flow key at PMTU/redirect time. That required an
adjustment of the interface call-sites used to propagate these events.
For a PMTU/redirect with a fully specified socket, we pass that socket
and use it to produce the flow key.
Otherwise we use a passed in SKB to formulate the key. There are two
cases that need to be distinguished, ICMP message processing (in which
case the IP header is at skb->data) and output packet processing
(mostly tunnels, and in all such cases the IP header is at ip_hdr(skb)).
We also have to make the code able to handle the case where the dst
itself passed into the dst_ops->{update_pmtu,redirect} method is
invalidated. This matters for calls from sockets that have cached
that route. We provide a inet{,6} helper function for this purpose,
and edit SCTP specially since it caches routes at the transport rather
than socket level.
Signed-off-by: David S. Miller <davem@davemloft.net>
^ permalink raw reply
* [PATCH 1/5] ipv4: Add helper inet_csk_update_pmtu().
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
This abstracts away the call to dst_ops->update_pmtu() so that we can
transparently handle the fact that, in the future, the dst itself can
be invalidated by the PMTU update (when we have non-host routes cached
in sockets).
So we try to rebuild the socket cached route after the method
invocation if necessary.
This isn't used by SCTP because it needs to cache dsts per-transport,
and thus will need it's own local version of this helper.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/inet_connection_sock.h | 2 ++
net/dccp/ipv4.c | 11 ++-------
net/ipv4/inet_connection_sock.c | 46 ++++++++++++++++++++++++++++++++++++
net/ipv4/tcp_ipv4.c | 11 ++-------
4 files changed, 52 insertions(+), 18 deletions(-)
diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h
index 291e7ce..2cf44b4 100644
--- a/include/net/inet_connection_sock.h
+++ b/include/net/inet_connection_sock.h
@@ -337,4 +337,6 @@ extern int inet_csk_compat_getsockopt(struct sock *sk, int level, int optname,
char __user *optval, int __user *optlen);
extern int inet_csk_compat_setsockopt(struct sock *sk, int level, int optname,
char __user *optval, unsigned int optlen);
+
+extern struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu);
#endif /* _INET_CONNECTION_SOCK_H */
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 129ed8f..683902f 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -161,17 +161,10 @@ static inline void dccp_do_pmtu_discovery(struct sock *sk,
if (sk->sk_state == DCCP_LISTEN)
return;
- /* We don't check in the destentry if pmtu discovery is forbidden
- * on this route. We just assume that no packet_to_big packets
- * are send back when pmtu discovery is not active.
- * There is a small race when the user changes this flag in the
- * route, but I think that's acceptable.
- */
- if ((dst = __sk_dst_check(sk, 0)) == NULL)
+ dst = inet_csk_update_pmtu(sk, mtu);
+ if (!dst)
return;
- dst->ops->update_pmtu(dst, mtu);
-
/* Something is about to be wrong... Remember soft error
* for the case, if this connection will not able to recover.
*/
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 76825be..200d218 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -803,3 +803,49 @@ int inet_csk_compat_setsockopt(struct sock *sk, int level, int optname,
}
EXPORT_SYMBOL_GPL(inet_csk_compat_setsockopt);
#endif
+
+static struct dst_entry *inet_csk_rebuild_route(struct sock *sk, struct flowi *fl)
+{
+ struct inet_sock *inet = inet_sk(sk);
+ struct ip_options_rcu *inet_opt;
+ __be32 daddr = inet->inet_daddr;
+ struct flowi4 *fl4;
+ struct rtable *rt;
+
+ rcu_read_lock();
+ inet_opt = rcu_dereference(inet->inet_opt);
+ if (inet_opt && inet_opt->opt.srr)
+ daddr = inet_opt->opt.faddr;
+ fl4 = &fl->u.ip4;
+ rt = ip_route_output_ports(sock_net(sk), fl4, sk, daddr,
+ inet->inet_saddr, inet->inet_dport,
+ inet->inet_sport, sk->sk_protocol,
+ RT_CONN_FLAGS(sk), sk->sk_bound_dev_if);
+ if (IS_ERR(rt))
+ rt = NULL;
+ if (rt)
+ sk_setup_caps(sk, &rt->dst);
+ rcu_read_unlock();
+
+ return &rt->dst;
+}
+
+struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu)
+{
+ struct dst_entry *dst = __sk_dst_check(sk, 0);
+ struct inet_sock *inet = inet_sk(sk);
+
+ if (!dst) {
+ dst = inet_csk_rebuild_route(sk, &inet->cork.fl);
+ if (!dst)
+ goto out;
+ }
+ dst->ops->update_pmtu(dst, mtu);
+
+ dst = __sk_dst_check(sk, 0);
+ if (!dst)
+ dst = inet_csk_rebuild_route(sk, &inet->cork.fl);
+out:
+ return dst;
+}
+EXPORT_SYMBOL_GPL(inet_csk_update_pmtu);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 7a0062c..b8e7e05 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -289,17 +289,10 @@ static void do_pmtu_discovery(struct sock *sk, const struct iphdr *iph, u32 mtu)
if (sk->sk_state == TCP_LISTEN)
return;
- /* We don't check in the destentry if pmtu discovery is forbidden
- * on this route. We just assume that no packet_to_big packets
- * are send back when pmtu discovery is not active.
- * There is a small race when the user changes this flag in the
- * route, but I think that's acceptable.
- */
- if ((dst = __sk_dst_check(sk, 0)) == NULL)
+ dst = inet_csk_update_pmtu(sk, mtu);
+ if (!dst)
return;
- dst->ops->update_pmtu(dst, mtu);
-
/* Something is about to be wrong... Remember soft error
* for the case, if this connection will not able to recover.
*/
--
1.7.10.4
^ permalink raw reply related
* [PATCH 2/5] ipv6: Add helper inet6_csk_update_pmtu().
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
This is the ipv6 version of inet_csk_update_pmtu().
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/inet6_connection_sock.h | 2 ++
net/dccp/ipv6.c | 35 +++----------------------
net/ipv6/inet6_connection_sock.c | 49 +++++++++++++++++++++++++----------
net/ipv6/tcp_ipv6.c | 37 +++-----------------------
4 files changed, 45 insertions(+), 78 deletions(-)
diff --git a/include/net/inet6_connection_sock.h b/include/net/inet6_connection_sock.h
index df2a857..04642c9 100644
--- a/include/net/inet6_connection_sock.h
+++ b/include/net/inet6_connection_sock.h
@@ -43,4 +43,6 @@ extern void inet6_csk_reqsk_queue_hash_add(struct sock *sk,
extern void inet6_csk_addr2sockaddr(struct sock *sk, struct sockaddr *uaddr);
extern int inet6_csk_xmit(struct sk_buff *skb, struct flowi *fl);
+
+extern struct dst_entry *inet6_csk_update_pmtu(struct sock *sk, u32 mtu);
#endif /* _INET6_CONNECTION_SOCK_H */
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index 090c080..3ee0342 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -145,39 +145,12 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
if ((1 << sk->sk_state) & (DCCPF_LISTEN | DCCPF_CLOSED))
goto out;
- /* icmp should have updated the destination cache entry */
- dst = __sk_dst_check(sk, np->dst_cookie);
- if (dst == NULL) {
- struct inet_sock *inet = inet_sk(sk);
- struct flowi6 fl6;
-
- /* BUGGG_FUTURE: Again, it is not clear how
- to handle rthdr case. Ignore this complexity
- for now.
- */
- memset(&fl6, 0, sizeof(fl6));
- fl6.flowi6_proto = IPPROTO_DCCP;
- fl6.daddr = np->daddr;
- fl6.saddr = np->saddr;
- fl6.flowi6_oif = sk->sk_bound_dev_if;
- fl6.fl6_dport = inet->inet_dport;
- fl6.fl6_sport = inet->inet_sport;
- security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
-
- dst = ip6_dst_lookup_flow(sk, &fl6, NULL, false);
- if (IS_ERR(dst)) {
- sk->sk_err_soft = -PTR_ERR(dst);
- goto out;
- }
- } else
- dst_hold(dst);
-
- dst->ops->update_pmtu(dst, ntohl(info));
+ dst = inet6_csk_update_pmtu(sk, ntohl(info));
+ if (!dst)
+ goto out;
- if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
+ if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst))
dccp_sync_mss(sk, dst_mtu(dst));
- } /* else let the usual retransmit timer handle it */
- dst_release(dst);
goto out;
}
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index bceb144..62539a4 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -203,15 +203,13 @@ struct dst_entry *__inet6_csk_dst_check(struct sock *sk, u32 cookie)
return dst;
}
-int inet6_csk_xmit(struct sk_buff *skb, struct flowi *fl_unused)
+static struct dst_entry *inet6_csk_route_socket(struct sock *sk)
{
- struct sock *sk = skb->sk;
struct inet_sock *inet = inet_sk(sk);
struct ipv6_pinfo *np = inet6_sk(sk);
- struct flowi6 fl6;
- struct dst_entry *dst;
struct in6_addr *final_p, final;
- int res;
+ struct dst_entry *dst;
+ struct flowi6 fl6;
memset(&fl6, 0, sizeof(fl6));
fl6.flowi6_proto = sk->sk_protocol;
@@ -228,18 +226,29 @@ int inet6_csk_xmit(struct sk_buff *skb, struct flowi *fl_unused)
final_p = fl6_update_dst(&fl6, np->opt, &final);
dst = __inet6_csk_dst_check(sk, np->dst_cookie);
-
- if (dst == NULL) {
+ if (!dst) {
dst = ip6_dst_lookup_flow(sk, &fl6, final_p, false);
- if (IS_ERR(dst)) {
- sk->sk_err_soft = -PTR_ERR(dst);
- sk->sk_route_caps = 0;
- kfree_skb(skb);
- return PTR_ERR(dst);
- }
+ if (!IS_ERR(dst))
+ __inet6_csk_dst_store(sk, dst, NULL, NULL);
+ }
+ return dst;
+}
- __inet6_csk_dst_store(sk, dst, NULL, NULL);
+int inet6_csk_xmit(struct sk_buff *skb, struct flowi *fl_unused)
+{
+ struct sock *sk = skb->sk;
+ struct ipv6_pinfo *np = inet6_sk(sk);
+ struct flowi6 fl6;
+ struct dst_entry *dst;
+ int res;
+
+ dst = inet6_csk_route_socket(sk);
+ if (IS_ERR(dst)) {
+ sk->sk_err_soft = -PTR_ERR(dst);
+ sk->sk_route_caps = 0;
+ kfree_skb(skb);
+ return PTR_ERR(dst);
}
rcu_read_lock();
@@ -253,3 +262,15 @@ int inet6_csk_xmit(struct sk_buff *skb, struct flowi *fl_unused)
return res;
}
EXPORT_SYMBOL_GPL(inet6_csk_xmit);
+
+struct dst_entry *inet6_csk_update_pmtu(struct sock *sk, u32 mtu)
+{
+ struct dst_entry *dst = inet6_csk_route_socket(sk);
+
+ if (IS_ERR(dst))
+ return NULL;
+ dst->ops->update_pmtu(dst, mtu);
+
+ return inet6_csk_route_socket(sk);
+}
+EXPORT_SYMBOL_GPL(inet6_csk_update_pmtu);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 3071f37..ecdf241 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -378,43 +378,14 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
if ((1 << sk->sk_state) & (TCPF_LISTEN | TCPF_CLOSE))
goto out;
- /* icmp should have updated the destination cache entry */
- dst = __sk_dst_check(sk, np->dst_cookie);
-
- if (dst == NULL) {
- struct inet_sock *inet = inet_sk(sk);
- struct flowi6 fl6;
-
- /* BUGGG_FUTURE: Again, it is not clear how
- to handle rthdr case. Ignore this complexity
- for now.
- */
- memset(&fl6, 0, sizeof(fl6));
- fl6.flowi6_proto = IPPROTO_TCP;
- fl6.daddr = np->daddr;
- fl6.saddr = np->saddr;
- fl6.flowi6_oif = sk->sk_bound_dev_if;
- fl6.flowi6_mark = sk->sk_mark;
- fl6.fl6_dport = inet->inet_dport;
- fl6.fl6_sport = inet->inet_sport;
- security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
-
- dst = ip6_dst_lookup_flow(sk, &fl6, NULL, false);
- if (IS_ERR(dst)) {
- sk->sk_err_soft = -PTR_ERR(dst);
- goto out;
- }
-
- } else
- dst_hold(dst);
-
- dst->ops->update_pmtu(dst, ntohl(info));
+ dst = inet6_csk_update_pmtu(sk, ntohl(info));
+ if (!dst)
+ goto out;
if (inet_csk(sk)->icsk_pmtu_cookie > dst_mtu(dst)) {
tcp_sync_mss(sk, dst_mtu(dst));
tcp_simple_retransmit(sk);
- } /* else let the usual retransmit timer handle it */
- dst_release(dst);
+ }
goto out;
}
--
1.7.10.4
^ permalink raw reply related
* [PATCH 3/5] sctp: Adjust PMTU updates to accomodate route invalidation.
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
This adjusts the call to dst_ops->update_pmtu() so that we can
transparently handle the fact that, in the future, the dst itself can
be invalidated by the PMTU update (when we have non-host routes cached
in sockets).
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/sctp/sctp.h | 4 ++--
include/net/sctp/structs.h | 4 ++--
net/sctp/associola.c | 4 ++--
net/sctp/input.c | 4 ++--
net/sctp/output.c | 2 +-
net/sctp/socket.c | 6 +++---
net/sctp/transport.c | 12 ++++++++++--
7 files changed, 22 insertions(+), 14 deletions(-)
diff --git a/include/net/sctp/sctp.h b/include/net/sctp/sctp.h
index 1f2735d..ff49964 100644
--- a/include/net/sctp/sctp.h
+++ b/include/net/sctp/sctp.h
@@ -519,10 +519,10 @@ static inline int sctp_frag_point(const struct sctp_association *asoc, int pmtu)
return frag;
}
-static inline void sctp_assoc_pending_pmtu(struct sctp_association *asoc)
+static inline void sctp_assoc_pending_pmtu(struct sock *sk, struct sctp_association *asoc)
{
- sctp_assoc_sync_pmtu(asoc);
+ sctp_assoc_sync_pmtu(sk, asoc);
asoc->pmtu_pending = 0;
}
diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index fecdf31..536e439 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -1091,7 +1091,7 @@ void sctp_transport_burst_limited(struct sctp_transport *);
void sctp_transport_burst_reset(struct sctp_transport *);
unsigned long sctp_transport_timeout(struct sctp_transport *);
void sctp_transport_reset(struct sctp_transport *);
-void sctp_transport_update_pmtu(struct sctp_transport *, u32);
+void sctp_transport_update_pmtu(struct sock *, struct sctp_transport *, u32);
void sctp_transport_immediate_rtx(struct sctp_transport *);
@@ -2003,7 +2003,7 @@ void sctp_assoc_update(struct sctp_association *old,
__u32 sctp_association_get_next_tsn(struct sctp_association *);
-void sctp_assoc_sync_pmtu(struct sctp_association *);
+void sctp_assoc_sync_pmtu(struct sock *, struct sctp_association *);
void sctp_assoc_rwnd_increase(struct sctp_association *, unsigned int);
void sctp_assoc_rwnd_decrease(struct sctp_association *, unsigned int);
void sctp_assoc_set_primary(struct sctp_association *,
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index b16517e..8cf348e 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -1360,7 +1360,7 @@ struct sctp_transport *sctp_assoc_choose_alter_transport(
/* Update the association's pmtu and frag_point by going through all the
* transports. This routine is called when a transport's PMTU has changed.
*/
-void sctp_assoc_sync_pmtu(struct sctp_association *asoc)
+void sctp_assoc_sync_pmtu(struct sock *sk, struct sctp_association *asoc)
{
struct sctp_transport *t;
__u32 pmtu = 0;
@@ -1372,7 +1372,7 @@ void sctp_assoc_sync_pmtu(struct sctp_association *asoc)
list_for_each_entry(t, &asoc->peer.transport_addr_list,
transports) {
if (t->pmtu_pending && t->dst) {
- sctp_transport_update_pmtu(t, dst_mtu(t->dst));
+ sctp_transport_update_pmtu(sk, t, dst_mtu(t->dst));
t->pmtu_pending = 0;
}
if (!pmtu || (t->pathmtu < pmtu))
diff --git a/net/sctp/input.c b/net/sctp/input.c
index f050d45..a67bc31 100644
--- a/net/sctp/input.c
+++ b/net/sctp/input.c
@@ -408,10 +408,10 @@ void sctp_icmp_frag_needed(struct sock *sk, struct sctp_association *asoc,
if (t->param_flags & SPP_PMTUD_ENABLE) {
/* Update transports view of the MTU */
- sctp_transport_update_pmtu(t, pmtu);
+ sctp_transport_update_pmtu(sk, t, pmtu);
/* Update association pmtu. */
- sctp_assoc_sync_pmtu(asoc);
+ sctp_assoc_sync_pmtu(sk, asoc);
}
/* Retransmit with the new pmtu setting.
diff --git a/net/sctp/output.c b/net/sctp/output.c
index 539f35d..838e18b 100644
--- a/net/sctp/output.c
+++ b/net/sctp/output.c
@@ -410,7 +410,7 @@ int sctp_packet_transmit(struct sctp_packet *packet)
if (!sctp_transport_dst_check(tp)) {
sctp_transport_route(tp, NULL, sctp_sk(sk));
if (asoc && (asoc->param_flags & SPP_PMTUD_ENABLE)) {
- sctp_assoc_sync_pmtu(asoc);
+ sctp_assoc_sync_pmtu(sk, asoc);
}
}
dst = dst_clone(tp->dst);
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index b3b8a8d..74bd3c4 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -1853,7 +1853,7 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
}
if (asoc->pmtu_pending)
- sctp_assoc_pending_pmtu(asoc);
+ sctp_assoc_pending_pmtu(sk, asoc);
/* If fragmentation is disabled and the message length exceeds the
* association fragmentation point, return EMSGSIZE. The I-D
@@ -2365,7 +2365,7 @@ static int sctp_apply_peer_addr_params(struct sctp_paddrparams *params,
if ((params->spp_flags & SPP_PMTUD_DISABLE) && params->spp_pathmtu) {
if (trans) {
trans->pathmtu = params->spp_pathmtu;
- sctp_assoc_sync_pmtu(asoc);
+ sctp_assoc_sync_pmtu(sctp_opt2sk(sp), asoc);
} else if (asoc) {
asoc->pathmtu = params->spp_pathmtu;
sctp_frag_point(asoc, params->spp_pathmtu);
@@ -2382,7 +2382,7 @@ static int sctp_apply_peer_addr_params(struct sctp_paddrparams *params,
(trans->param_flags & ~SPP_PMTUD) | pmtud_change;
if (update) {
sctp_transport_pmtu(trans, sctp_opt2sk(sp));
- sctp_assoc_sync_pmtu(asoc);
+ sctp_assoc_sync_pmtu(sctp_opt2sk(sp), asoc);
}
} else if (asoc) {
asoc->param_flags =
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index 1dcceb6..e69e1a2 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -228,7 +228,7 @@ void sctp_transport_pmtu(struct sctp_transport *transport, struct sock *sk)
transport->pathmtu = SCTP_DEFAULT_MAXSEGMENT;
}
-void sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu)
+void sctp_transport_update_pmtu(struct sock *sk, struct sctp_transport *t, u32 pmtu)
{
struct dst_entry *dst;
@@ -245,8 +245,16 @@ void sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu)
}
dst = sctp_transport_dst_check(t);
- if (dst)
+ if (!dst)
+ t->af_specific->get_dst(t, &t->saddr, &t->fl, sk);
+
+ if (dst) {
dst->ops->update_pmtu(dst, pmtu);
+
+ dst = sctp_transport_dst_check(t);
+ if (!dst)
+ t->af_specific->get_dst(t, &t->saddr, &t->fl, sk);
+ }
}
/* Caches the dst entry and source address for a transport's destination
--
1.7.10.4
^ permalink raw reply related
* [PATCH 4/5] net: Pass optional SKB and SK arguments to dst_ops->{update_pmtu,redirect}()
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
This will be used so that we can compose a full flow key.
Even though we have a route in this context, we need more. In the
future the routes will be without destination address, source address,
etc. keying. One ipv4 route will cover entire subnets, etc.
In this environment we have to have a way to possess persistent storage
for redirects and PMTU information. This persistent storage will exist
in the FIB tables, and that's why we'll need to be able to rebuild a
full lookup flow key here. Using that flow key will do a fib_lookup()
and create/update the persistent entry.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
drivers/infiniband/ulp/ipoib/ipoib_cm.c | 2 +-
include/net/dst_ops.h | 6 ++++--
net/bridge/br_netfilter.c | 6 ++++--
net/dccp/ipv4.c | 2 +-
net/dccp/ipv6.c | 2 +-
net/decnet/dn_route.c | 12 ++++++++----
net/ipv4/inet_connection_sock.c | 2 +-
net/ipv4/ip_gre.c | 2 +-
net/ipv4/ipip.c | 2 +-
net/ipv4/route.c | 21 +++++++++++++--------
net/ipv4/tcp_ipv4.c | 2 +-
net/ipv4/xfrm4_policy.c | 10 ++++++----
net/ipv6/inet6_connection_sock.c | 2 +-
net/ipv6/ip6_tunnel.c | 6 +++---
net/ipv6/route.c | 21 +++++++++++++--------
net/ipv6/sit.c | 2 +-
net/ipv6/tcp_ipv6.c | 2 +-
net/ipv6/xfrm6_policy.c | 10 ++++++----
net/netfilter/ipvs/ip_vs_xmit.c | 4 ++--
net/sctp/input.c | 2 +-
net/sctp/transport.c | 2 +-
21 files changed, 71 insertions(+), 49 deletions(-)
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_cm.c b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
index 014504d..1ca7322 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_cm.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_cm.c
@@ -1397,7 +1397,7 @@ void ipoib_cm_skb_too_long(struct net_device *dev, struct sk_buff *skb,
int e = skb_queue_empty(&priv->cm.skb_queue);
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
skb_queue_tail(&priv->cm.skb_queue, skb);
if (e)
diff --git a/include/net/dst_ops.h b/include/net/dst_ops.h
index 085931f..d079fc6 100644
--- a/include/net/dst_ops.h
+++ b/include/net/dst_ops.h
@@ -24,8 +24,10 @@ struct dst_ops {
struct net_device *dev, int how);
struct dst_entry * (*negative_advice)(struct dst_entry *);
void (*link_failure)(struct sk_buff *);
- void (*update_pmtu)(struct dst_entry *dst, u32 mtu);
- void (*redirect)(struct dst_entry *dst, struct sk_buff *skb);
+ void (*update_pmtu)(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu);
+ void (*redirect)(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb);
int (*local_out)(struct sk_buff *skb);
struct neighbour * (*neigh_lookup)(const struct dst_entry *dst,
struct sk_buff *skb,
diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c
index 81f76c4..68e8f36 100644
--- a/net/bridge/br_netfilter.c
+++ b/net/bridge/br_netfilter.c
@@ -111,11 +111,13 @@ static inline __be16 pppoe_proto(const struct sk_buff *skb)
pppoe_proto(skb) == htons(PPP_IPV6) && \
brnf_filter_pppoe_tagged)
-static void fake_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void fake_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
}
-static void fake_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void fake_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
}
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 683902f..ab4f44c 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -193,7 +193,7 @@ static void dccp_do_redirect(struct sk_buff *skb, struct sock *sk)
struct dst_entry *dst = __sk_dst_check(sk, 0);
if (dst)
- dst->ops->redirect(dst, skb);
+ dst->ops->redirect(dst, sk, skb);
}
/*
diff --git a/net/dccp/ipv6.c b/net/dccp/ipv6.c
index 3ee0342..56840b2 100644
--- a/net/dccp/ipv6.c
+++ b/net/dccp/ipv6.c
@@ -134,7 +134,7 @@ static void dccp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
struct dst_entry *dst = __sk_dst_check(sk, np->dst_cookie);
if (dst)
- dst->ops->redirect(dst, skb);
+ dst->ops->redirect(dst, sk, skb);
}
if (type == ICMPV6_PKT_TOOBIG) {
diff --git a/net/decnet/dn_route.c b/net/decnet/dn_route.c
index e9c4e2e..47de90d 100644
--- a/net/decnet/dn_route.c
+++ b/net/decnet/dn_route.c
@@ -117,8 +117,10 @@ static void dn_dst_destroy(struct dst_entry *);
static void dn_dst_ifdown(struct dst_entry *, struct net_device *dev, int how);
static struct dst_entry *dn_dst_negative_advice(struct dst_entry *);
static void dn_dst_link_failure(struct sk_buff *);
-static void dn_dst_update_pmtu(struct dst_entry *dst, u32 mtu);
-static void dn_dst_redirect(struct dst_entry *dst, struct sk_buff *skb);
+static void dn_dst_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb , u32 mtu);
+static void dn_dst_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb);
static struct neighbour *dn_dst_neigh_lookup(const struct dst_entry *dst,
struct sk_buff *skb,
const void *daddr);
@@ -266,7 +268,8 @@ static int dn_dst_gc(struct dst_ops *ops)
* We update both the mtu and the advertised mss (i.e. the segment size we
* advertise to the other end).
*/
-static void dn_dst_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void dn_dst_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
struct dn_route *rt = (struct dn_route *) dst;
struct neighbour *n = rt->n;
@@ -294,7 +297,8 @@ static void dn_dst_update_pmtu(struct dst_entry *dst, u32 mtu)
}
}
-static void dn_dst_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void dn_dst_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
}
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 200d218..3ea4652 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -840,7 +840,7 @@ struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu)
if (!dst)
goto out;
}
- dst->ops->update_pmtu(dst, mtu);
+ dst->ops->update_pmtu(dst, sk, NULL, mtu);
dst = __sk_dst_check(sk, 0);
if (!dst)
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 0c31235..42c44b1 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -833,7 +833,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
mtu = skb_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
if (skb->protocol == htons(ETH_P_IP)) {
df |= (old_iph->frag_off&htons(IP_DF));
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index c2d0e6d..2c2c35b 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -519,7 +519,7 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
}
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
if ((old_iph->frag_off & htons(IP_DF)) &&
mtu < ntohs(old_iph->tot_len)) {
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index aad2181..b35d3bf 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -148,8 +148,10 @@ static unsigned int ipv4_mtu(const struct dst_entry *dst);
static void ipv4_dst_destroy(struct dst_entry *dst);
static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst);
static void ipv4_link_failure(struct sk_buff *skb);
-static void ip_rt_update_pmtu(struct dst_entry *dst, u32 mtu);
-static void ip_do_redirect(struct dst_entry *dst, struct sk_buff *skb);
+static void ip_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu);
+static void ip_do_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb);
static int rt_garbage_collect(struct dst_ops *ops);
static void ipv4_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
@@ -1273,7 +1275,7 @@ static void rt_del(unsigned int hash, struct rtable *rt)
spin_unlock_bh(rt_hash_lock_addr(hash));
}
-static void ip_do_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buff *skb)
{
__be32 new_gw = icmp_hdr(skb)->un.gateway;
__be32 old_gw = ip_hdr(skb)->saddr;
@@ -1506,7 +1508,8 @@ out: kfree_skb(skb);
return 0;
}
-static void ip_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void ip_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
struct rtable *rt = (struct rtable *) dst;
@@ -1531,7 +1534,7 @@ void ipv4_update_pmtu(struct sk_buff *skb, struct net *net, u32 mtu,
iph->daddr, iph->saddr, 0, 0);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
- ip_rt_update_pmtu(&rt->dst, mtu);
+ ip_rt_update_pmtu(&rt->dst, NULL, skb, mtu);
ip_rt_put(rt);
}
}
@@ -1559,7 +1562,7 @@ void ipv4_redirect(struct sk_buff *skb, struct net *net,
protocol, flow_flags, iph->daddr, iph->saddr, 0, 0);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
- ip_do_redirect(&rt->dst, skb);
+ ip_do_redirect(&rt->dst, NULL, skb);
ip_rt_put(rt);
}
}
@@ -2587,11 +2590,13 @@ static unsigned int ipv4_blackhole_mtu(const struct dst_entry *dst)
return mtu ? : dst->dev->mtu;
}
-static void ipv4_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void ipv4_rt_blackhole_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
}
-static void ipv4_rt_blackhole_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void ipv4_rt_blackhole_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
}
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index b8e7e05..d9caf5c 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -319,7 +319,7 @@ static void do_redirect(struct sk_buff *skb, struct sock *sk)
struct dst_entry *dst = __sk_dst_check(sk, 0);
if (dst)
- dst->ops->redirect(dst, skb);
+ dst->ops->redirect(dst, sk, skb);
}
/*
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 737131c..fcf7678 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -194,20 +194,22 @@ static inline int xfrm4_garbage_collect(struct dst_ops *ops)
return (dst_entries_get_slow(ops) > ops->gc_thresh * 2);
}
-static void xfrm4_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void xfrm4_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
struct dst_entry *path = xdst->route;
- path->ops->update_pmtu(path, mtu);
+ path->ops->update_pmtu(path, sk, skb, mtu);
}
-static void xfrm4_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void xfrm4_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
struct dst_entry *path = xdst->route;
- path->ops->redirect(path, skb);
+ path->ops->redirect(path, sk, skb);
}
static void xfrm4_dst_destroy(struct dst_entry *dst)
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index 62539a4..4a0c4d2 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -269,7 +269,7 @@ struct dst_entry *inet6_csk_update_pmtu(struct sock *sk, u32 mtu)
if (IS_ERR(dst))
return NULL;
- dst->ops->update_pmtu(dst, mtu);
+ dst->ops->update_pmtu(dst, sk, NULL, mtu);
return inet6_csk_route_socket(sk);
}
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 61d1065..db32846 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -609,10 +609,10 @@ ip4ip6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
if (rel_info > dst_mtu(skb_dst(skb2)))
goto out;
- skb_dst(skb2)->ops->update_pmtu(skb_dst(skb2), rel_info);
+ skb_dst(skb2)->ops->update_pmtu(skb_dst(skb2), NULL, skb2, rel_info);
}
if (rel_type == ICMP_REDIRECT)
- skb_dst(skb2)->ops->redirect(skb_dst(skb2), skb2);
+ skb_dst(skb2)->ops->redirect(skb_dst(skb2), NULL, skb2);
icmp_send(skb2, rel_type, rel_code, htonl(rel_info));
@@ -952,7 +952,7 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
if (mtu < IPV6_MIN_MTU)
mtu = IPV6_MIN_MTU;
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
if (skb->len > mtu) {
*pmtu = mtu;
err = -EMSGSIZE;
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 2a4c8d4..31af1ed 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -78,8 +78,10 @@ static int ip6_dst_gc(struct dst_ops *ops);
static int ip6_pkt_discard(struct sk_buff *skb);
static int ip6_pkt_discard_out(struct sk_buff *skb);
static void ip6_link_failure(struct sk_buff *skb);
-static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu);
-static void rt6_do_redirect(struct dst_entry *dst, struct sk_buff *skb);
+static void ip6_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu);
+static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb);
#ifdef CONFIG_IPV6_ROUTE_INFO
static struct rt6_info *rt6_add_route_info(struct net *net,
@@ -187,11 +189,13 @@ static unsigned int ip6_blackhole_mtu(const struct dst_entry *dst)
return mtu ? : dst->dev->mtu;
}
-static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void ip6_rt_blackhole_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
}
-static void ip6_rt_blackhole_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void ip6_rt_blackhole_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
}
@@ -1071,7 +1075,8 @@ static void ip6_link_failure(struct sk_buff *skb)
}
}
-static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void ip6_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
struct rt6_info *rt6 = (struct rt6_info*)dst;
@@ -1108,7 +1113,7 @@ void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu,
dst = ip6_route_output(net, NULL, &fl6);
if (!dst->error)
- ip6_rt_update_pmtu(dst, ntohl(mtu));
+ ip6_rt_update_pmtu(dst, NULL, skb, ntohl(mtu));
dst_release(dst);
}
EXPORT_SYMBOL_GPL(ip6_update_pmtu);
@@ -1136,7 +1141,7 @@ void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark)
dst = ip6_route_output(net, NULL, &fl6);
if (!dst->error)
- rt6_do_redirect(dst, skb);
+ rt6_do_redirect(dst, NULL, skb);
dst_release(dst);
}
EXPORT_SYMBOL_GPL(ip6_redirect);
@@ -1639,7 +1644,7 @@ static int ip6_route_del(struct fib6_config *cfg)
return err;
}
-static void rt6_do_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buff *skb)
{
struct net *net = dev_net(skb->dev);
struct netevent_redirect netevent;
diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
index fbf1622..3bd1bfc 100644
--- a/net/ipv6/sit.c
+++ b/net/ipv6/sit.c
@@ -807,7 +807,7 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
}
if (tunnel->parms.iph.daddr && skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
if (skb->len > mtu) {
icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index ecdf241..c9dabdd 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -367,7 +367,7 @@ static void tcp_v6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
struct dst_entry *dst = __sk_dst_check(sk, np->dst_cookie);
if (dst)
- dst->ops->redirect(dst,skb);
+ dst->ops->redirect(dst, sk, skb);
}
if (type == ICMPV6_PKT_TOOBIG) {
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index f5a9cb8..ef39812 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -207,20 +207,22 @@ static inline int xfrm6_garbage_collect(struct dst_ops *ops)
return dst_entries_get_fast(ops) > ops->gc_thresh * 2;
}
-static void xfrm6_update_pmtu(struct dst_entry *dst, u32 mtu)
+static void xfrm6_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
{
struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
struct dst_entry *path = xdst->route;
- path->ops->update_pmtu(path, mtu);
+ path->ops->update_pmtu(path, sk, skb, mtu);
}
-static void xfrm6_redirect(struct dst_entry *dst, struct sk_buff *skb)
+static void xfrm6_redirect(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb)
{
struct xfrm_dst *xdst = (struct xfrm_dst *)dst;
struct dst_entry *path = xdst->route;
- path->ops->redirect(path, skb);
+ path->ops->redirect(path, sk, skb);
}
static void xfrm6_dst_destroy(struct dst_entry *dst)
diff --git a/net/netfilter/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 71d6ecb..65b616a 100644
--- a/net/netfilter/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c
@@ -797,7 +797,7 @@ ip_vs_tunnel_xmit(struct sk_buff *skb, struct ip_vs_conn *cp,
goto tx_error_put;
}
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
df |= (old_iph->frag_off & htons(IP_DF));
@@ -913,7 +913,7 @@ ip_vs_tunnel_xmit_v6(struct sk_buff *skb, struct ip_vs_conn *cp,
goto tx_error_put;
}
if (skb_dst(skb))
- skb_dst(skb)->ops->update_pmtu(skb_dst(skb), mtu);
+ skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, skb, mtu);
if (mtu < ntohs(old_iph->payload_len) + sizeof(struct ipv6hdr) &&
!skb_is_gso(skb)) {
diff --git a/net/sctp/input.c b/net/sctp/input.c
index a67bc31..c201b26 100644
--- a/net/sctp/input.c
+++ b/net/sctp/input.c
@@ -432,7 +432,7 @@ void sctp_icmp_redirect(struct sock *sk, struct sctp_transport *t,
return;
dst = sctp_transport_dst_check(t);
if (dst)
- dst->ops->redirect(dst, skb);
+ dst->ops->redirect(dst, sk, skb);
}
/*
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index e69e1a2..a6b7ee9 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -249,7 +249,7 @@ void sctp_transport_update_pmtu(struct sock *sk, struct sctp_transport *t, u32 p
t->af_specific->get_dst(t, &t->saddr, &t->fl, sk);
if (dst) {
- dst->ops->update_pmtu(dst, pmtu);
+ dst->ops->update_pmtu(dst, sk, NULL, pmtu);
dst = sctp_transport_dst_check(t);
if (!dst)
--
1.7.10.4
^ permalink raw reply related
* [PATCH 5/5] ipv4: Add FIB nexthop exceptions.
From: David Miller @ 2012-07-17 13:14 UTC (permalink / raw)
To: netdev
In a regime where we have subnetted route entries, we need a way to
store persistent storage about destination specific learned values
such as redirects and PMTU values.
This is implemented here via nexthop exceptions.
The initial implementation is a simple linked list, and can be
expanded to a hash table when it is shown to be justified.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/ip_fib.h | 9 ++
net/ipv4/fib_semantics.c | 15 ++++
net/ipv4/route.c | 216 +++++++++++++++++++++++++++++++++++++++-------
3 files changed, 209 insertions(+), 31 deletions(-)
diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index 5697ace..b6b400f 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -46,6 +46,14 @@ struct fib_config {
struct fib_info;
+struct fib_nh_exception {
+ struct hlist_node fnhe_node;
+ __be32 fnhe_daddr;
+ u32 fnhe_pmtu;
+ u32 fnhe_gw;
+ unsigned long fnhe_expires;
+};
+
struct fib_nh {
struct net_device *nh_dev;
struct hlist_node nh_hash;
@@ -63,6 +71,7 @@ struct fib_nh {
__be32 nh_gw;
__be32 nh_saddr;
int nh_saddr_genid;
+ struct hlist_head nh_exceptions;
};
/*
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index d71bfbd..d266096 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -140,6 +140,18 @@ const struct fib_prop fib_props[RTN_MAX + 1] = {
},
};
+static void free_nh_exceptions(struct fib_nh *nh)
+{
+ struct hlist_head *head = &nh->nh_exceptions;
+ struct hlist_node *node, *tmp;
+ struct fib_nh_exception *fnhe;
+
+ hlist_for_each_entry_safe(fnhe, node, tmp, head, fnhe_node) {
+ hlist_del(node);
+ kfree(fnhe);
+ }
+}
+
/* Release a nexthop info record */
static void free_fib_info_rcu(struct rcu_head *head)
{
@@ -148,6 +160,8 @@ static void free_fib_info_rcu(struct rcu_head *head)
change_nexthops(fi) {
if (nexthop_nh->nh_dev)
dev_put(nexthop_nh->nh_dev);
+ if (!hlist_empty(&nexthop_nh->nh_exceptions))
+ free_nh_exceptions(nexthop_nh);
} endfor_nexthops(fi);
release_net(fi->fib_net);
@@ -777,6 +791,7 @@ struct fib_info *fib_create_info(struct fib_config *cfg)
fi->fib_nhs = nhs;
change_nexthops(fi) {
nexthop_nh->nh_parent = fi;
+ INIT_HLIST_HEAD(&nexthop_nh->nh_exceptions);
} endfor_nexthops(fi)
if (cfg->fc_mx) {
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index b35d3bf..c27ca8f4 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1275,14 +1275,93 @@ static void rt_del(unsigned int hash, struct rtable *rt)
spin_unlock_bh(rt_hash_lock_addr(hash));
}
-static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buff *skb)
+static void __build_flow_key(struct flowi4 *fl4, struct sock *sk,
+ const struct iphdr *iph,
+ int oif, u8 tos,
+ u8 prot, u32 mark, int flow_flags)
+{
+ if (sk) {
+ const struct inet_sock *inet = inet_sk(sk);
+
+ oif = sk->sk_bound_dev_if;
+ mark = sk->sk_mark;
+ tos = RT_CONN_FLAGS(sk);
+ prot = inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol;
+ }
+ flowi4_init_output(fl4, oif, mark, tos,
+ RT_SCOPE_UNIVERSE, prot,
+ flow_flags,
+ iph->daddr, iph->saddr, 0, 0);
+}
+
+static void build_skb_flow_key(struct flowi4 *fl4, struct sk_buff *skb, struct sock *sk)
+{
+ const struct iphdr *iph = ip_hdr(skb);
+ int oif = skb->dev->ifindex;
+ u8 tos = RT_TOS(iph->tos);
+ u8 prot = iph->protocol;
+ u32 mark = skb->mark;
+
+ __build_flow_key(fl4, sk, iph, oif, tos, prot, mark, 0);
+}
+
+static void build_sk_flow_key(struct flowi4 *fl4, struct sock *sk)
+{
+ const struct inet_sock *inet = inet_sk(sk);
+ struct ip_options_rcu *inet_opt;
+ __be32 daddr = inet->inet_daddr;
+
+ rcu_read_lock();
+ inet_opt = rcu_dereference(inet->inet_opt);
+ if (inet_opt && inet_opt->opt.srr)
+ daddr = inet_opt->opt.faddr;
+ flowi4_init_output(fl4, sk->sk_bound_dev_if, sk->sk_mark,
+ RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
+ inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol,
+ inet_sk_flowi_flags(sk),
+ daddr, inet->inet_saddr, 0, 0);
+ rcu_read_unlock();
+}
+
+static void ip_rt_build_flow_key(struct flowi4 *fl4, struct sock *sk,
+ struct sk_buff *skb)
+{
+ if (skb)
+ build_skb_flow_key(fl4, skb, sk);
+ else
+ build_sk_flow_key(fl4, sk);
+}
+
+static DEFINE_SPINLOCK(fnhe_lock);
+
+static struct fib_nh_exception *find_or_create_fnhe(struct fib_nh *nh, __be32 daddr)
+{
+ struct hlist_head *head = &nh->nh_exceptions;
+ struct fib_nh_exception *fnhe;
+ struct hlist_node *node;
+
+ hlist_for_each_entry(fnhe, node, head, fnhe_node) {
+ if (fnhe->fnhe_daddr == daddr)
+ return fnhe;
+ }
+
+ fnhe = kzalloc(sizeof(*fnhe), GFP_ATOMIC);
+ if (!fnhe)
+ return NULL;
+
+ fnhe->fnhe_daddr = daddr;
+ hlist_add_head(&fnhe->fnhe_node, head);
+ return fnhe;
+}
+
+static void __ip_do_redirect(struct rtable *rt, struct sk_buff *skb, struct flowi4 *fl4)
{
__be32 new_gw = icmp_hdr(skb)->un.gateway;
__be32 old_gw = ip_hdr(skb)->saddr;
struct net_device *dev = skb->dev;
struct in_device *in_dev;
+ struct fib_result res;
struct neighbour *n;
- struct rtable *rt;
struct net *net;
switch (icmp_hdr(skb)->code & 7) {
@@ -1296,7 +1375,6 @@ static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buf
return;
}
- rt = (struct rtable *) dst;
if (rt->rt_gateway != old_gw)
return;
@@ -1320,11 +1398,21 @@ static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buf
goto reject_redirect;
}
- n = ipv4_neigh_lookup(dst, NULL, &new_gw);
+ n = ipv4_neigh_lookup(&rt->dst, NULL, &new_gw);
if (n) {
if (!(n->nud_state & NUD_VALID)) {
neigh_event_send(n, NULL);
} else {
+ if (fib_lookup(net, fl4, &res) == 0) {
+ struct fib_nh *nh = &FIB_RES_NH(res);
+ struct fib_nh_exception *fnhe;
+
+ spin_lock_bh(&fnhe_lock);
+ fnhe = find_or_create_fnhe(nh, fl4->daddr);
+ if (fnhe)
+ fnhe->fnhe_gw = new_gw;
+ spin_unlock_bh(&fnhe_lock);
+ }
rt->rt_gateway = new_gw;
rt->rt_flags |= RTCF_REDIRECTED;
call_netevent_notifiers(NETEVENT_NEIGH_UPDATE, n);
@@ -1349,6 +1437,17 @@ reject_redirect:
;
}
+static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buff *skb)
+{
+ struct rtable *rt;
+ struct flowi4 fl4;
+
+ rt = (struct rtable *) dst;
+
+ ip_rt_build_flow_key(&fl4, sk, skb);
+ __ip_do_redirect(rt, skb, &fl4);
+}
+
static struct dst_entry *ipv4_negative_advice(struct dst_entry *dst)
{
struct rtable *rt = (struct rtable *)dst;
@@ -1508,33 +1607,51 @@ out: kfree_skb(skb);
return 0;
}
-static void ip_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
- struct sk_buff *skb, u32 mtu)
+static void __ip_rt_update_pmtu(struct rtable *rt, struct flowi4 *fl4, u32 mtu)
{
- struct rtable *rt = (struct rtable *) dst;
-
- dst_confirm(dst);
+ struct fib_result res;
if (mtu < ip_rt_min_pmtu)
mtu = ip_rt_min_pmtu;
+ if (fib_lookup(dev_net(rt->dst.dev), fl4, &res) == 0) {
+ struct fib_nh *nh = &FIB_RES_NH(res);
+ struct fib_nh_exception *fnhe;
+
+ spin_lock_bh(&fnhe_lock);
+ fnhe = find_or_create_fnhe(nh, fl4->daddr);
+ if (fnhe) {
+ fnhe->fnhe_pmtu = mtu;
+ fnhe->fnhe_expires = jiffies + ip_rt_mtu_expires;
+ }
+ spin_unlock_bh(&fnhe_lock);
+ }
rt->rt_pmtu = mtu;
dst_set_expires(&rt->dst, ip_rt_mtu_expires);
}
+static void ip_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
+ struct sk_buff *skb, u32 mtu)
+{
+ struct rtable *rt = (struct rtable *) dst;
+ struct flowi4 fl4;
+
+ ip_rt_build_flow_key(&fl4, sk, skb);
+ __ip_rt_update_pmtu(rt, &fl4, mtu);
+}
+
void ipv4_update_pmtu(struct sk_buff *skb, struct net *net, u32 mtu,
int oif, u32 mark, u8 protocol, int flow_flags)
{
- const struct iphdr *iph = (const struct iphdr *)skb->data;
+ const struct iphdr *iph = (const struct iphdr *) skb->data;
struct flowi4 fl4;
struct rtable *rt;
- flowi4_init_output(&fl4, oif, mark, RT_TOS(iph->tos), RT_SCOPE_UNIVERSE,
- protocol, flow_flags,
- iph->daddr, iph->saddr, 0, 0);
+ __build_flow_key(&fl4, NULL, iph, oif,
+ RT_TOS(iph->tos), protocol, mark, flow_flags);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
- ip_rt_update_pmtu(&rt->dst, NULL, skb, mtu);
+ __ip_rt_update_pmtu(rt, &fl4, mtu);
ip_rt_put(rt);
}
}
@@ -1542,27 +1659,31 @@ EXPORT_SYMBOL_GPL(ipv4_update_pmtu);
void ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu)
{
- const struct inet_sock *inet = inet_sk(sk);
+ const struct iphdr *iph = (const struct iphdr *) skb->data;
+ struct flowi4 fl4;
+ struct rtable *rt;
- return ipv4_update_pmtu(skb, sock_net(sk), mtu,
- sk->sk_bound_dev_if, sk->sk_mark,
- inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol,
- inet_sk_flowi_flags(sk));
+ __build_flow_key(&fl4, sk, iph, 0, 0, 0, 0, 0);
+ rt = __ip_route_output_key(sock_net(sk), &fl4);
+ if (!IS_ERR(rt)) {
+ __ip_rt_update_pmtu(rt, &fl4, mtu);
+ ip_rt_put(rt);
+ }
}
EXPORT_SYMBOL_GPL(ipv4_sk_update_pmtu);
void ipv4_redirect(struct sk_buff *skb, struct net *net,
int oif, u32 mark, u8 protocol, int flow_flags)
{
- const struct iphdr *iph = (const struct iphdr *)skb->data;
+ const struct iphdr *iph = (const struct iphdr *) skb->data;
struct flowi4 fl4;
struct rtable *rt;
- flowi4_init_output(&fl4, oif, mark, RT_TOS(iph->tos), RT_SCOPE_UNIVERSE,
- protocol, flow_flags, iph->daddr, iph->saddr, 0, 0);
+ __build_flow_key(&fl4, NULL, iph, oif,
+ RT_TOS(iph->tos), protocol, mark, flow_flags);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
- ip_do_redirect(&rt->dst, NULL, skb);
+ __ip_do_redirect(rt, skb, &fl4);
ip_rt_put(rt);
}
}
@@ -1570,12 +1691,16 @@ EXPORT_SYMBOL_GPL(ipv4_redirect);
void ipv4_sk_redirect(struct sk_buff *skb, struct sock *sk)
{
- const struct inet_sock *inet = inet_sk(sk);
+ const struct iphdr *iph = (const struct iphdr *) skb->data;
+ struct flowi4 fl4;
+ struct rtable *rt;
- return ipv4_redirect(skb, sock_net(sk), sk->sk_bound_dev_if,
- sk->sk_mark,
- inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol,
- inet_sk_flowi_flags(sk));
+ __build_flow_key(&fl4, sk, iph, 0, 0, 0, 0, 0);
+ rt = __ip_route_output_key(sock_net(sk), &fl4);
+ if (!IS_ERR(rt)) {
+ __ip_do_redirect(rt, skb, &fl4);
+ ip_rt_put(rt);
+ }
}
EXPORT_SYMBOL_GPL(ipv4_sk_redirect);
@@ -1722,14 +1847,43 @@ static void rt_init_metrics(struct rtable *rt, const struct flowi4 *fl4,
dst_init_metrics(&rt->dst, fi->fib_metrics, true);
}
+static void rt_bind_exception(struct rtable *rt, struct fib_nh *nh, __be32 daddr)
+{
+ struct hlist_head *head = &nh->nh_exceptions;
+ struct fib_nh_exception *fnhe;
+ struct hlist_node *node;
+
+ spin_lock_bh(&fnhe_lock);
+ hlist_for_each_entry(fnhe, node, head, fnhe_node) {
+ if (fnhe->fnhe_daddr == daddr) {
+ if (fnhe->fnhe_pmtu) {
+ unsigned long expires = fnhe->fnhe_expires;
+ unsigned long diff = jiffies - expires;
+
+ if (time_before(jiffies, expires)) {
+ rt->rt_pmtu = fnhe->fnhe_pmtu;
+ dst_set_expires(&rt->dst, diff);
+ }
+ }
+ if (fnhe->fnhe_gw)
+ rt->rt_gateway = fnhe->fnhe_gw;
+ break;
+ }
+ }
+ spin_unlock_bh(&fnhe_lock);
+}
+
static void rt_set_nexthop(struct rtable *rt, const struct flowi4 *fl4,
const struct fib_result *res,
struct fib_info *fi, u16 type, u32 itag)
{
if (fi) {
- if (FIB_RES_GW(*res) &&
- FIB_RES_NH(*res).nh_scope == RT_SCOPE_LINK)
- rt->rt_gateway = FIB_RES_GW(*res);
+ struct fib_nh *nh = &FIB_RES_NH(*res);
+
+ if (nh->nh_gw && nh->nh_scope == RT_SCOPE_LINK)
+ rt->rt_gateway = nh->nh_gw;
+ if (unlikely(!hlist_empty(&nh->nh_exceptions)))
+ rt_bind_exception(rt, nh, fl4->daddr);
rt_init_metrics(rt, fl4, fi);
#ifdef CONFIG_IP_ROUTE_CLASSID
rt->dst.tclassid = FIB_RES_NH(*res).nh_tclassid;
--
1.7.10.4
^ permalink raw reply related
* RE: [PATCH] mlx4_en: map entire pages to increase throughput
From: David Laight @ 2012-07-17 13:36 UTC (permalink / raw)
To: David Miller
Cc: rick.jones2, cascardo, netdev, yevgenyp, ogerlitz, amirv, brking,
leitao, klebers
In-Reply-To: <20120717.055005.1912765690890797652.davem@davemloft.net>
> > Would there be any mileage in permanently allocating IOMMU
> > virtual address to the ring entries, then 'just' assigning
> > the correct physical address during rx/tx setup?
>
> There is a not a one to one mapping between these two entities,
> in particular on the transmit side.
>
> A transmit packet can have multiple segments, some of which are
> larger than one IOMMU page.
A SMOP :-) TX is probably easier than RX.
Each tx segment will already go into a separate ring entry,
page boundaries could do the same.
The driver will already have to cope with 'too many segments'
(I remember being passed a full sized frame made of a list
of 1-byte message blocks...)
Or allocate enough sequential IOMMU pages for the longest
tx segment for every ring entry - after all that is already
the 'worst case' allocation!
David
^ permalink raw reply
* Re: [PATCH] mlx4_en: map entire pages to increase throughput
From: David Miller @ 2012-07-17 13:46 UTC (permalink / raw)
To: David.Laight
Cc: rick.jones2, cascardo, netdev, yevgenyp, ogerlitz, amirv, brking,
leitao, klebers
In-Reply-To: <AE90C24D6B3A694183C094C60CF0A2F6026B6F8D@saturn3.aculab.com>
From: "David Laight" <David.Laight@ACULAB.COM>
Date: Tue, 17 Jul 2012 14:36:11 +0100
> The driver will already have to cope with 'too many segments'
> (I remember being passed a full sized frame made of a list
> of 1-byte message blocks...)
Baring driver hardware bug workarounds, no it does not have to cope
with that. The code is extremely simple now.
All the driver has to do is assume that a new TX packet can never
consume more than MAX_SKB_FRAGS.
Therefore it simply stops the queue if less than MAX_SKB_FRAGS
segments remain after queueing a transmit.
Your suggestion will significantly complicate driver TX paths.
If you're going to suggest a solution, it has to be completely
general enough to work in the current state of affairs, and
your idea absolutely does not.
^ permalink raw reply
* Re: [GIT PULL nf] IPVS
From: Simon Horman @ 2012-07-17 13:50 UTC (permalink / raw)
To: Pablo Neira Ayuso
Cc: lvs-devel, netdev, netfilter-devel, Wensong Zhang,
Julian Anastasov, Hans Schillstrom, Jesper Dangaard Brouer
In-Reply-To: <20120717101406.GC3812@1984>
On Tue, Jul 17, 2012 at 12:14:06PM +0200, Pablo Neira Ayuso wrote:
> On Wed, Jul 11, 2012 at 09:19:20AM +0900, Simon Horman wrote:
> >
> > Hi Pablo,
> >
> > this pull request consists of three bug fixes for IPVS.
> > Please consider for inclusion in 3.5 and stable.
> >
> > The bug fix from Julian, "ipvs: fix oops in ip_vs_dst_event on rmmod"
> > fixes a regression introduced in 3.4 and thus I believe it is
> > only relevant to 3.5 and 3.4-stable.
> >
> > The other two fixes appear to have been present since at least 2.6.37
> > (there were a lot of changes to IPVS around that time).
>
> I have passed the two of these patches to David. The one for the FTP
> needs a consistent description.
>
> It's fairly late in the development cycle (-rc7), but these are small.
> Let's see if David is still in time to accept them. Otherwise, they go
> to net-next and we will ask for -stable submission.
Thanks, it seems that David was in an accepting mood.
^ permalink raw reply
* Re: [PATCH] mlx4_en: map entire pages to increase throughput
From: Eric Dumazet @ 2012-07-17 13:50 UTC (permalink / raw)
To: David Miller
Cc: David.Laight, rick.jones2, cascardo, netdev, yevgenyp, ogerlitz,
amirv, brking, leitao, klebers
In-Reply-To: <20120717.055005.1912765690890797652.davem@davemloft.net>
On Tue, 2012-07-17 at 05:50 -0700, David Miller wrote:
> From: "David Laight" <David.Laight@ACULAB.COM>
> Date: Tue, 17 Jul 2012 13:42:04 +0100
>
> > Would there be any mileage in permanently allocating IOMMU
> > virtual address to the ring entries, then 'just' assigning
> > the correct physical address during rx/tx setup?
>
> There is a not a one to one mapping between these two entities,
> in particular on the transmit side.
>
> A transmit packet can have multiple segments, some of which are
> larger than one IOMMU page.
And on rx side, permanently allocating IOMMU would need to copy all
incoming frames to newly allocated memory.
Annot this IOMMU performance problem can be solved on its side,
instead of having to shuffle things in all drivers ?
^ permalink raw reply
* Re: [PATCH 5/5] ipv4: Add FIB nexthop exceptions.
From: Eric Dumazet @ 2012-07-17 14:00 UTC (permalink / raw)
To: David Miller; +Cc: netdev
In-Reply-To: <20120717.061435.1733209287175819043.davem@davemloft.net>
On Tue, 2012-07-17 at 06:14 -0700, David Miller wrote:
> In a regime where we have subnetted route entries, we need a way to
> store persistent storage about destination specific learned values
> such as redirects and PMTU values.
>
> This is implemented here via nexthop exceptions.
>
> The initial implementation is a simple linked list, and can be
> expanded to a hash table when it is shown to be justified.
Say a typical host uses a single default route, I am trying to convince
myself it can really use a simple linked list ?
Arent PMTU entries added by messages coming from untrusted sources ?
^ permalink raw reply
* Re: [PATCH] [RFC] tcp: TSQ - do not always throttle.
From: Krishna Kumar2 @ 2012-07-17 14:08 UTC (permalink / raw)
To: Eric Dumazet; +Cc: davem, netdev
In-Reply-To: <1342530654.2626.563.camel@edumazet-glaptop>
Eric Dumazet <eric.dumazet@gmail.com> wrote on 07/17/2012 06:40:54 PM:
> > Do not throttle if sysctl_tcp_limit_output_bytes==0.
> >
> > Maybe it is better to throttle earlier in the loop, after
> > calling tcp_init_tso_segs().
> >
>
> I wonder why, and why you put this question in a changelog instead of
> outside of it...
>
> Idea was to avoid setting TSQ_THROTTLED if we break out the loop.
The reason I mentioned it (in the wrong place) is because
I thought this is a likely case and the checks before that
might all pass only to get throttled. Some of the checks
are quite lengthy.
> About disabling TSQ, my initial intent was to instead use a negative
> sysctl_tcp_limit_output_bytes value.
>
> Thats why I have in tcp_transmit_skb() :
>
> skb->destructor = (sysctl_tcp_limit_output_bytes > 0) ?
> tcp_wfree : sock_wfree;
>
> So I suggest you change the tcp_write_xmit(() test to a single unsigned
> compare :
>
> if (atomic_read(&sk->sk_wmem_alloc) >=
> (unsigned) sysctl_tcp_limit_output_bytes) {
>
> Also use :
>
> skb->destructor = (sysctl_tcp_limit_output_bytes >= 0) ?
> tcp_wfree : sock_wfree;
>
> and document the 'negative value disables TSQ' in
> Documentation/networking/ip-sysctl.txt
Sure, will post with this change.
thanks,
- KK
^ permalink raw reply
* RE: [PATCH 1/4] pch_gbe: Fix the checksum fill to the error location
From: Andy Cress @ 2012-07-17 14:20 UTC (permalink / raw)
To: Eric Dumazet; +Cc: netdev, Zhong Hongbo
In-Reply-To: <1342510387.2626.174.camel@edumazet-glaptop>
Eric,
This is intriguing, and the data copy also would explain why this transmit path is slow, and is susceptible to transmit timeouts.
I want to apply and test your proposed patch, but I'll have to do that next week.
Andy
-----Original Message-----
From: Eric Dumazet [mailto:eric.dumazet@gmail.com]
Sent: Tuesday, July 17, 2012 3:33 AM
To: Andy Cress
Cc: netdev@vger.kernel.org; Zhong Hongbo
Subject: Re: [PATCH 1/4] pch_gbe: Fix the checksum fill to the error location
On Tue, 2012-07-17 at 09:09 +0200, Eric Dumazet wrote:
> Hmm... I fail to understand why you care about NIC doing checksums,
> while pch_gbe_tx_queue() make a _copy_ of each outgoing
> packets.
>
> There _must_ be a way to avoid most of these copies (ie not touching
> payload), only mess with the header to insert these 2 nul bytes ?
>
> /* [Header:14][payload] ---> [Header:14][paddong:2][payload] */
>
> So at device setup : dev->needed_headroom = 2;
>
> and in xmit,
>
> if (skb_headroom(skb) < 2) {
> struct sk_buff *skb_new;
>
> skb_new = skb_realloc_headroom(skb, 2);
> if (!skb_new) { handle error }
> consume_skb(skb);
> skb = skb_new;
> }
> ptr = skb_push(skb, 2);
> memmove(ptr, ptr + 2, ETH_HLEN);
> ptr[ETH_HLEN] = 0;
> ptr[ETH_HLEN + 1] = 0;
>
>
Something like the following (untested) patch
drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c | 55 +++++-----
1 file changed, 29 insertions(+), 26 deletions(-)
diff --git a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
index b100656..2d3d982 100644
--- a/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
+++ b/drivers/net/ethernet/oki-semi/pch_gbe/pch_gbe_main.c
@@ -1163,7 +1163,7 @@ static void pch_gbe_tx_queue(struct pch_gbe_adapter *adapter,
struct pch_gbe_hw *hw = &adapter->hw;
struct pch_gbe_tx_desc *tx_desc;
struct pch_gbe_buffer *buffer_info;
- struct sk_buff *tmp_skb;
+ char *ptr;
unsigned int frame_ctrl;
unsigned int ring_num;
@@ -1221,18 +1221,27 @@ static void pch_gbe_tx_queue(struct pch_gbe_adapter *adapter,
buffer_info = &tx_ring->buffer_info[ring_num];
- tmp_skb = buffer_info->skb;
+ if (skb_headroom(skb) < 2) {
+ struct sk_buff *skb_new;
+
+ skb_new = skb_realloc_headroom(skb, 2);
+ if (!skb_new) {
+ tx_ring->next_to_use = ring_num;
+ dev_kfree_skb_any(skb);
+ return;
+ }
+ consume_skb(skb);
+ skb = skb_new;
+ }
/* [Header:14][payload] ---> [Header:14][paddong:2][payload] */
- memcpy(tmp_skb->data, skb->data, ETH_HLEN);
- tmp_skb->data[ETH_HLEN] = 0x00;
- tmp_skb->data[ETH_HLEN + 1] = 0x00;
- tmp_skb->len = skb->len;
- memcpy(&tmp_skb->data[ETH_HLEN + 2], &skb->data[ETH_HLEN],
- (skb->len - ETH_HLEN));
+ ptr = skb_push(skb, 2);
+ memmove(ptr, ptr + 2, ETH_HLEN);
+ ptr[ETH_HLEN] = 0x00;
+ ptr[ETH_HLEN + 1] = 0x00;
/*-- Set Buffer information --*/
- buffer_info->length = tmp_skb->len;
- buffer_info->dma = dma_map_single(&adapter->pdev->dev, tmp_skb->data,
+ buffer_info->length = skb->len;
+ buffer_info->dma = dma_map_single(&adapter->pdev->dev, skb->data,
buffer_info->length,
DMA_TO_DEVICE);
if (dma_mapping_error(&adapter->pdev->dev, buffer_info->dma)) {
@@ -1240,18 +1249,20 @@ static void pch_gbe_tx_queue(struct pch_gbe_adapter *adapter,
buffer_info->dma = 0;
buffer_info->time_stamp = 0;
tx_ring->next_to_use = ring_num;
+ dev_kfree_skb_any(skb);
return;
}
buffer_info->mapped = true;
buffer_info->time_stamp = jiffies;
+ buffer_info->skb = skb;
/*-- Set Tx descriptor --*/
tx_desc = PCH_GBE_TX_DESC(*tx_ring, ring_num);
- tx_desc->buffer_addr = (buffer_info->dma);
- tx_desc->length = (tmp_skb->len);
- tx_desc->tx_words_eob = ((tmp_skb->len + 3));
+ tx_desc->buffer_addr = buffer_info->dma;
+ tx_desc->length = skb->len;
+ tx_desc->tx_words_eob = skb->len + 3;
tx_desc->tx_frame_ctrl = (frame_ctrl);
- tx_desc->gbec_status = (DSC_INIT16);
+ tx_desc->gbec_status = DSC_INIT16;
if (unlikely(++ring_num == tx_ring->count))
ring_num = 0;
@@ -1265,7 +1276,6 @@ static void pch_gbe_tx_queue(struct pch_gbe_adapter *adapter,
pch_tx_timestamp(adapter, skb);
#endif
- dev_kfree_skb_any(skb);
}
/**
@@ -1543,19 +1553,12 @@ static void pch_gbe_alloc_tx_buffers(struct pch_gbe_adapter *adapter,
struct pch_gbe_tx_ring *tx_ring)
{
struct pch_gbe_buffer *buffer_info;
- struct sk_buff *skb;
unsigned int i;
- unsigned int bufsz;
struct pch_gbe_tx_desc *tx_desc;
- bufsz =
- adapter->hw.mac.max_frame_size + PCH_GBE_DMA_ALIGN + NET_IP_ALIGN;
-
for (i = 0; i < tx_ring->count; i++) {
buffer_info = &tx_ring->buffer_info[i];
- skb = netdev_alloc_skb(adapter->netdev, bufsz);
- skb_reserve(skb, PCH_GBE_DMA_ALIGN);
- buffer_info->skb = skb;
+ buffer_info->skb = NULL;
tx_desc = PCH_GBE_TX_DESC(*tx_ring, i);
tx_desc->gbec_status = (DSC_INIT16);
}
@@ -1622,9 +1625,9 @@ pch_gbe_clean_tx(struct pch_gbe_adapter *adapter,
buffer_info->length, DMA_TO_DEVICE);
buffer_info->mapped = false;
}
- if (buffer_info->skb) {
- pr_debug("trim buffer_info->skb : %d\n", i);
- skb_trim(buffer_info->skb, 0);
+ if (skb) {
+ dev_kfree_skb_any(skb);
+ buffer_info->skb = NULL;
}
tx_desc->gbec_status = DSC_INIT16;
if (unlikely(++i == tx_ring->count))
^ permalink raw reply related
* Re: [PATCH 5/5] ipv4: Add FIB nexthop exceptions.
From: David Miller @ 2012-07-17 14:25 UTC (permalink / raw)
To: eric.dumazet; +Cc: netdev
In-Reply-To: <1342533605.2626.680.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 17 Jul 2012 16:00:05 +0200
> On Tue, 2012-07-17 at 06:14 -0700, David Miller wrote:
>> In a regime where we have subnetted route entries, we need a way to
>> store persistent storage about destination specific learned values
>> such as redirects and PMTU values.
>>
>> This is implemented here via nexthop exceptions.
>>
>> The initial implementation is a simple linked list, and can be
>> expanded to a hash table when it is shown to be justified.
>
> Say a typical host uses a single default route, I am trying to convince
> myself it can really use a simple linked list ?
>
> Arent PMTU entries added by messages coming from untrusted sources ?
They are trusted when we validate them at the socket layer, at least
as is done for TCP.
I totally agree that we'll need to adjust the list into something more
sophisticated, but that's an implementation detail rather than
something that requires the actual infrastructure to be redone.
^ 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