* Re: [PATCH] ipv6: fix incorrect ipsec transport mode fragment
From: Gao feng @ 2012-05-26 9:00 UTC (permalink / raw)
To: Steffen Klassert; +Cc: netdev, davem, lw
In-Reply-To: <20120514130528.GA24733@secunet.com>
Hi Steffen
于 2012年05月14日 21:05, Steffen Klassert 写道:
> On Mon, May 14, 2012 at 11:21:00AM +0800, Gao feng wrote:
>> Since commit 299b0767(ipv6: Fix IPsec slowpath fragmentation problem)
>> the fragment of ipsec transport mode packets is incorrect.
>> because tunnel mode needs IPsec headers and trailer for all fragments,
>> while on transport mode it is sufficient to add the headers to the
>> first fragment and the trailer to the last.
>
> I mentioned this in an other thread some time ago,
> this is due to commit ad0081e43a
> "ipv6: Fragment locally generated tunnel-mode IPSec6 packets as needed"
> changed tunnel mode to do fragmentation before the transformation
> while transport mode still does fragmentation after transformation.
> Now, tunnel mode needs IPsec headers and trailer for all fragments,
> while on transport mode it is sufficient to add the headers to the
> first fragment and the trailer to the last.
>
>>
>> so modify mtu and maxfraglen base on ipsec mode and if fragment is first
>> or last.
>
> There might be other opinions, but I don't like to see this IPsec mode
> dependent stuff hacked into the generic ipv6 output path.
>
> Basically we have two cases. One where we have to add rt->dst.header_len
> to the first fragment and rt->dst.trailer_len to the last fragment,
> and the other where we have to add both to all fragments. So perhaps we
> could isolate this code and create two functions, one for each case.
>
>
I thought this problem carefully,I think the important and troubled thing is
how to deal with transport mode.
we have to use different mtu and maxfraglen for checking if the prev_skb has
extra data or has free room. so mtu_prev and maxfraglen_prev have to be used.
And I also think it's not very well to create two function for the two cases,
it will create a lot of redundant codes.
I will add a dst_entry flag DST_XFRM_TUNNEL to avoid ipsec mode dependent stuff
hacked into the generic ipv6_output path, and send patch v2.
^ permalink raw reply
* Re: Fwd: Re: WARNING: at net/sched/sch_generic.c:256 dev_watchdog+0x277/0x280()
From: Francois Romieu @ 2012-05-26 10:11 UTC (permalink / raw)
To: Alex Villacís Lasso; +Cc: netdev
In-Reply-To: <4FBE533E.6030107@palosanto.com>
[-- Attachment #1: Type: text/plain, Size: 366 bytes --]
Alex Villacís Lasso <a_villacis@palosanto.com> :
[...]
> According to my bisection, the first bad commit is 036dafa28da1e2565a8529de2ae663c37b7a0060 . Here is my bisection log:
There would be something wrong with bql support.
Can you apply the attached patches on top of 3.4 or later then
send the kernel log output whent the bug happens ?
Thanks.
--
Ueimor
[-- Attachment #2: 0001-r8169-avoid-clearing-the-end-of-Tx-descriptor-ring-m.patch --]
[-- Type: text/plain, Size: 1040 bytes --]
>From 01e33a3f6c46db63ba90f1db3b99eaf62c7071d1 Mon Sep 17 00:00:00 2001
Message-Id: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Wed, 23 May 2012 22:18:35 +0200
Subject: [PATCH 1/6] r8169: avoid clearing the end of Tx descriptor ring
marker bit.
X-Organisation: Land of Sunshine Inc.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/r8169.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 00b4f56..01f3367 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -5345,7 +5345,7 @@ static void rtl8169_unmap_tx_skb(struct device *d, struct ring_info *tx_skb,
dma_unmap_single(d, le64_to_cpu(desc->addr), len, DMA_TO_DEVICE);
- desc->opts1 = 0x00;
+ desc->opts1 &= cpu_to_le32(RingEnd);
desc->opts2 = 0x00;
desc->addr = 0x00;
tx_skb->len = 0;
--
1.7.7.6
[-- Attachment #3: 0002-r8169-TxPoll-hack-rework.patch --]
[-- Type: text/plain, Size: 4285 bytes --]
>From 7ef658faa62dc1211cabd9a6291e0921c12562bf Mon Sep 17 00:00:00 2001
Message-Id: <7ef658faa62dc1211cabd9a6291e0921c12562bf.1338026442.git.romieu@fr.zoreil.com>
In-Reply-To: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
References: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Wed, 23 May 2012 23:21:13 +0200
Subject: [PATCH 2/6] r8169: TxPoll hack rework.
X-Organisation: Land of Sunshine Inc.
I don't want to try and convince myself that it is completely safe to
issue a TxPoll request from the NAPI handler right in the middle of
a start_xmit, whence the tx_lock.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/r8169.c | 59 +++++++++++++++++++--------------
1 files changed, 34 insertions(+), 25 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 01f3367..655b293 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -5544,19 +5544,20 @@ static netdev_tx_t rtl8169_start_xmit(struct sk_buff *skb,
status = opts[0] | len | (RingEnd * !((entry + 1) % NUM_TX_DESC));
txd->opts1 = cpu_to_le32(status);
+ smp_wmb();
+
tp->cur_tx += frags + 1;
- wmb();
+ smp_wmb();
RTL_W8(TxPoll, NPQ);
mmiowb();
if (!TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS)) {
- /* Avoid wrongly optimistic queue wake-up: rtl_tx thread must
- * not miss a ring update when it notices a stopped queue.
- */
- smp_wmb();
+ /* rtl_tx thread must not miss a ring update when it notices
+ * a stopped queue. The TxPoll hack requires the smp_wmb
+ * above so we can go ahead. */
netif_stop_queue(dev);
/* Sync with rtl_tx:
* - publish queue status and cur_tx ring index (write barrier)
@@ -5640,22 +5641,36 @@ struct rtl_txc {
static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
{
struct rtl8169_stats *tx_stats = &tp->tx_stats;
- unsigned int dirty_tx, tx_left;
+ unsigned int dirty_tx, cur_tx;
struct rtl_txc txc = { 0, 0 };
dirty_tx = tp->dirty_tx;
- smp_rmb();
- tx_left = tp->cur_tx - dirty_tx;
-
- while (tx_left > 0) {
+xmit_race:
+ for (cur_tx = tp->cur_tx; dirty_tx != cur_tx; dirty_tx++) {
unsigned int entry = dirty_tx % NUM_TX_DESC;
struct ring_info *tx_skb = tp->tx_skb + entry;
u32 status;
- rmb();
status = le32_to_cpu(tp->TxDescArray[entry].opts1);
- if (status & DescOwn)
+
+ /* 8168 (only ?) hack: TxPoll requests are lost when the Tx
+ * packets are too close. Let's kick an extra TxPoll request
+ * when a burst of start_xmit activity is detected (if it is
+ * not detected, it is slow enough).
+ * The NPQ bit is cleared automatically by the chipset.
+ * The code assumes that the chipset is sane enough to clear
+ * it at a sensible time.*/
+ if (unlikely(status & DescOwn)) {
+ void __iomem *ioaddr = tp->mmio_addr;
+
+ if (!(RTL_R8(TxPoll) & NPQ)) {
+ netif_tx_lock(dev);
+ RTL_W8(TxPoll, NPQ);
+ netif_tx_unlock(dev);
+ goto done;
+ }
break;
+ }
rtl8169_unmap_tx_skb(&tp->pci_dev->dev, tx_skb,
tp->TxDescArray + entry);
@@ -5667,10 +5682,15 @@ static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
dev_kfree_skb(skb);
tx_skb->skb = NULL;
}
- dirty_tx++;
- tx_left--;
}
+ /* Rationale: if chipset stopped DMAing, enforce TxPoll write either
+ * here or in start_xmit. If chipset is still DMAing, this code
+ * will be run later anyway. */
+ smp_mb();
+ if (cur_tx != tp->cur_tx)
+ goto xmit_race;
+done:
u64_stats_update_begin(&tx_stats->syncp);
tx_stats->packets += txc.packets;
tx_stats->bytes += txc.bytes;
@@ -5692,17 +5712,6 @@ static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS)) {
netif_wake_queue(dev);
}
- /*
- * 8168 hack: TxPoll requests are lost when the Tx packets are
- * too close. Let's kick an extra TxPoll request when a burst
- * of start_xmit activity is detected (if it is not detected,
- * it is slow enough). -- FR
- */
- if (tp->cur_tx != dirty_tx) {
- void __iomem *ioaddr = tp->mmio_addr;
-
- RTL_W8(TxPoll, NPQ);
- }
}
}
--
1.7.7.6
[-- Attachment #4: 0003-r8169-ring-debug-helper.patch --]
[-- Type: text/plain, Size: 2048 bytes --]
>From c8fa99f5afe595348dd578c450681c056abb7bcc Mon Sep 17 00:00:00 2001
Message-Id: <c8fa99f5afe595348dd578c450681c056abb7bcc.1338026442.git.romieu@fr.zoreil.com>
In-Reply-To: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
References: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Fri, 25 May 2012 11:29:26 +0200
Subject: [PATCH 3/6] r8169: ring debug helper.
X-Organisation: Land of Sunshine Inc.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/r8169.c | 26 ++++++++++++++++++++++++++
1 files changed, 26 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 655b293..3feb374 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -5382,6 +5382,30 @@ static void rtl8169_tx_clear(struct rtl8169_private *tp)
netdev_reset_queue(tp->dev);
}
+static void rtl_print_ring(struct rtl8169_private *tp)
+{
+ struct net_device *dev = tp->dev;
+ int i;
+
+ netif_info(tp, drv, dev, "Rx ring: dirty=%08x current=%08x\n",
+ tp->dirty_rx, tp->cur_rx);
+ for (i = 0; i < NUM_RX_DESC; i++) {
+ struct RxDesc *desc = tp->RxDescArray + i;
+
+ netif_info(tp, drv, dev, "[%03d] %08x %08x %0llx\n", i,
+ desc->opts1, desc->opts2, (u64)desc->addr);
+ }
+
+ netif_info(tp, drv, dev, "Tx ring: dirty=%08x current=%08x\n",
+ tp->dirty_tx, tp->cur_tx);
+ for (i = 0; i < NUM_TX_DESC; i++) {
+ struct TxDesc *desc = tp->TxDescArray + i;
+
+ netif_info(tp, drv, dev, "[%03d] %08x %08x %0llx\n", i,
+ desc->opts1, desc->opts2, (u64)desc->addr);
+ }
+}
+
static void rtl_reset_work(struct rtl8169_private *tp)
{
struct net_device *dev = tp->dev;
@@ -5391,6 +5415,8 @@ static void rtl_reset_work(struct rtl8169_private *tp)
netif_stop_queue(dev);
synchronize_sched();
+ rtl_print_ring(tp);
+
rtl8169_hw_reset(tp);
for (i = 0; i < NUM_RX_DESC; i++)
--
1.7.7.6
[-- Attachment #5: 0004-r8169-add-extra-Tx-descriptor-ring-room-for-debug.patch --]
[-- Type: text/plain, Size: 6235 bytes --]
>From e7fa457acf502c73f4581cd975a37ace410da734 Mon Sep 17 00:00:00 2001
Message-Id: <e7fa457acf502c73f4581cd975a37ace410da734.1338026442.git.romieu@fr.zoreil.com>
In-Reply-To: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
References: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Fri, 25 May 2012 12:37:02 +0200
Subject: [PATCH 4/6] r8169: add extra Tx descriptor ring room for debug.
X-Organisation: Land of Sunshine Inc.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/r8169.c | 100 +++++++++++++++++++++++++++++-----
1 files changed, 85 insertions(+), 15 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 3feb374..4812515 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -82,9 +82,11 @@ static const int multicast_filter_limit = 32;
#define R8169_REGS_SIZE 256
#define R8169_NAPI_WEIGHT 64
#define NUM_TX_DESC 64 /* Number of Tx descriptor registers */
+#define NUM_TX_DESC_DBG 16 /* Tx descriptor ring canary */
#define NUM_RX_DESC 256 /* Number of Rx descriptor registers */
#define RX_BUF_SIZE 1536 /* Rx Buffer size */
-#define R8169_TX_RING_BYTES (NUM_TX_DESC * sizeof(struct TxDesc))
+#define R8169_TX_RING_BYTES \
+ ((NUM_TX_DESC + NUM_TX_DESC_DBG) * sizeof(struct TxDesc))
#define R8169_RX_RING_BYTES (NUM_RX_DESC * sizeof(struct RxDesc))
#define RTL8169_TX_TIMEOUT (6*HZ)
@@ -774,6 +776,11 @@ struct rtl8169_private {
} phy_action;
} *rtl_fw;
#define RTL_FIRMWARE_UNKNOWN ERR_PTR(-EAGAIN)
+
+ struct {
+ struct sk_buff *skb;
+ dma_addr_t mapping;
+ } guard;
};
MODULE_AUTHOR("Realtek and the Linux r8169 crew <netdev@vger.kernel.org>");
@@ -5326,28 +5333,90 @@ err_out:
return -ENOMEM;
}
-static int rtl8169_init_ring(struct net_device *dev)
+static int rtl_init_ring_tx(struct rtl8169_private *tp)
+{
+ struct device *d = &tp->pci_dev->dev;
+ struct sk_buff *skb;
+ dma_addr_t mapping;
+ int i;
+
+ memset(tp->tx_skb, 0x0, NUM_TX_DESC * sizeof(struct ring_info));
+
+ /* Paranoid debug helping stuff ony beyond this point. */
+
+#define RTL_DBG_SKB_LEN 80
+
+ skb = netdev_alloc_skb_ip_align(tp->dev, RTL_DBG_SKB_LEN);
+ if (!skb)
+ return -ENOMEM;
+
+ for (i = 0; i < RTL_DBG_SKB_LEN; i++)
+ skb->data[i] = 0xaa;
+
+ mapping = dma_map_single(d, skb->data, RTL_DBG_SKB_LEN, DMA_TO_DEVICE);
+ if (unlikely(dma_mapping_error(d, mapping))) {
+ dev_kfree_skb(skb);
+ return -EIO;
+ }
+
+ for (i = NUM_TX_DESC; i < (NUM_TX_DESC + NUM_TX_DESC_DBG); i++) {
+ struct TxDesc *txd = tp->TxDescArray + i;
+
+ txd->addr = cpu_to_le64(mapping);
+ txd->opts1 = cpu_to_le32(DescOwn | FirstFrag | LastFrag |
+ RTL_DBG_SKB_LEN);
+ }
+ tp->TxDescArray[i - 1].opts1 |= RingEnd;
+
+ tp->guard.skb = skb;
+ tp->guard.mapping = mapping;
+
+ return 0;
+}
+
+static void rtl_clean_ring_tx(struct rtl8169_private *tp)
+{
+ struct device *d = &tp->pci_dev->dev;
+
+ dev_kfree_skb(tp->guard.skb);
+ dma_unmap_single(d, tp->guard.mapping, RTL_DBG_SKB_LEN, DMA_TO_DEVICE);
+}
+
+static int rtl_init_ring_rx(struct rtl8169_private *tp)
+{
+ memset(tp->Rx_databuff, 0x0, NUM_RX_DESC * sizeof(void *));
+ return rtl8169_rx_fill(tp);
+}
+
+static int rtl8169_init_rings(struct net_device *dev)
{
struct rtl8169_private *tp = netdev_priv(dev);
+ int rc;
rtl8169_init_ring_indexes(tp);
- memset(tp->tx_skb, 0x0, NUM_TX_DESC * sizeof(struct ring_info));
- memset(tp->Rx_databuff, 0x0, NUM_RX_DESC * sizeof(void *));
+ rc = rtl_init_ring_tx(tp);
+ if (rc < 0)
+ return rc;
- return rtl8169_rx_fill(tp);
+ rc = rtl_init_ring_rx(tp);
+ if (rc < 0)
+ rtl_clean_ring_tx(tp);
+
+ return rc;
}
-static void rtl8169_unmap_tx_skb(struct device *d, struct ring_info *tx_skb,
- struct TxDesc *desc)
+static void rtl_unmap_tx_skb(struct rtl8169_private *tp,
+ struct ring_info *tx_skb, struct TxDesc *desc)
{
+ struct device *d = &tp->pci_dev->dev;
unsigned int len = tx_skb->len;
dma_unmap_single(d, le64_to_cpu(desc->addr), len, DMA_TO_DEVICE);
desc->opts1 &= cpu_to_le32(RingEnd);
desc->opts2 = 0x00;
- desc->addr = 0x00;
+ desc->addr = cpu_to_le64(tp->guard.mapping);
tx_skb->len = 0;
}
@@ -5364,8 +5433,7 @@ static void rtl8169_tx_clear_range(struct rtl8169_private *tp, u32 start,
if (len) {
struct sk_buff *skb = tx_skb->skb;
- rtl8169_unmap_tx_skb(&tp->pci_dev->dev, tx_skb,
- tp->TxDescArray + entry);
+ rtl_unmap_tx_skb(tp, tx_skb, tp->TxDescArray + entry);
if (skb) {
tp->dev->stats.tx_dropped++;
dev_kfree_skb(skb);
@@ -5398,7 +5466,7 @@ static void rtl_print_ring(struct rtl8169_private *tp)
netif_info(tp, drv, dev, "Tx ring: dirty=%08x current=%08x\n",
tp->dirty_tx, tp->cur_tx);
- for (i = 0; i < NUM_TX_DESC; i++) {
+ for (i = 0; i < (NUM_TX_DESC + NUM_TX_DESC_DBG); i++) {
struct TxDesc *desc = tp->TxDescArray + i;
netif_info(tp, drv, dev, "[%03d] %08x %08x %0llx\n", i,
@@ -5600,7 +5668,7 @@ static netdev_tx_t rtl8169_start_xmit(struct sk_buff *skb,
return NETDEV_TX_OK;
err_dma_1:
- rtl8169_unmap_tx_skb(d, tp->tx_skb + entry, txd);
+ rtl_unmap_tx_skb(tp, tp->tx_skb + entry, txd);
err_dma_0:
dev_kfree_skb(skb);
dev->stats.tx_dropped++;
@@ -5698,8 +5766,7 @@ xmit_race:
break;
}
- rtl8169_unmap_tx_skb(&tp->pci_dev->dev, tx_skb,
- tp->TxDescArray + entry);
+ rtl_unmap_tx_skb(tp, tx_skb, tp->TxDescArray + entry);
if (status & LastFrag) {
struct sk_buff *skb = tx_skb->skb;
@@ -6034,6 +6101,8 @@ static void rtl8169_down(struct net_device *dev)
rtl8169_rx_clear(tp);
+ rtl_clean_ring_tx(tp);
+
rtl_pll_power_down(tp);
}
@@ -6099,7 +6168,7 @@ static int rtl_open(struct net_device *dev)
if (!tp->RxDescArray)
goto err_free_tx_0;
- retval = rtl8169_init_ring(dev);
+ retval = rtl8169_init_rings(dev);
if (retval < 0)
goto err_free_rx_1;
@@ -6143,6 +6212,7 @@ out:
err_release_fw_2:
rtl_release_firmware(tp);
rtl8169_rx_clear(tp);
+ rtl_clean_ring_tx(tp);
err_free_rx_1:
dma_free_coherent(&pdev->dev, R8169_RX_RING_BYTES, tp->RxDescArray,
tp->RxPhyAddr);
--
1.7.7.6
[-- Attachment #6: 0005-r8169-debugfs-support.patch --]
[-- Type: text/plain, Size: 5535 bytes --]
>From bf93ca2cd16948911a5cb978e1fc46f468464156 Mon Sep 17 00:00:00 2001
Message-Id: <bf93ca2cd16948911a5cb978e1fc46f468464156.1338026442.git.romieu@fr.zoreil.com>
In-Reply-To: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
References: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Fri, 25 May 2012 11:18:03 +0200
Subject: [PATCH 5/6] r8169: debugfs support.
X-Organisation: Land of Sunshine Inc.
Almost completely copied from Stephen Hemminger's sky2.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/Kconfig | 9 +++
drivers/net/ethernet/realtek/r8169.c | 126 ++++++++++++++++++++++++++++++++++
2 files changed, 135 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/realtek/Kconfig b/drivers/net/ethernet/realtek/Kconfig
index 5821966..ae48c78 100644
--- a/drivers/net/ethernet/realtek/Kconfig
+++ b/drivers/net/ethernet/realtek/Kconfig
@@ -115,4 +115,13 @@ config R8169
To compile this driver as a module, choose M here: the module
will be called r8169. This is recommended.
+config R8169_DEBUGFS
+ bool "Realtek 8169 gigabit ethernet debugfs support"
+ depends on R8169
+ ---help---
+ Say Y here if you want to see the content of the descriptor rings
+ and misc undocumented debug information.
+
+ Debug data is available in the /sys/kernel/debug/r8169/ethX file.
+
endif # NET_VENDOR_REALTEK
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 4812515..834ff62 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -28,6 +28,7 @@
#include <linux/firmware.h>
#include <linux/pci-aspm.h>
#include <linux/prefetch.h>
+#include <linux/debugfs.h>
#include <asm/io.h>
#include <asm/irq.h>
@@ -60,6 +61,10 @@
#define dprintk(fmt, args...) do {} while (0)
#endif /* RTL8169_DEBUG */
+#ifdef CONFIG_R8169_DEBUGFS
+static struct dentry *rtl_debug;
+#endif
+
#define R8169_MSG_DEFAULT \
(NETIF_MSG_DRV | NETIF_MSG_PROBE | NETIF_MSG_IFUP | NETIF_MSG_IFDOWN)
@@ -781,6 +786,10 @@ struct rtl8169_private {
struct sk_buff *skb;
dma_addr_t mapping;
} guard;
+
+#ifdef CONFIG_R8169_DEBUGFS
+ struct dentry *debugfs;
+#endif
};
MODULE_AUTHOR("Realtek and the Linux r8169 crew <netdev@vger.kernel.org>");
@@ -6794,14 +6803,131 @@ static struct pci_driver rtl8169_pci_driver = {
.driver.pm = RTL8169_PM_OPS,
};
+#ifdef CONFIG_R8169_DEBUGFS
+
+static struct dentry *rtl_debug;
+
+static int rtl_debug_show(struct seq_file *seq, void *v)
+{
+ struct net_device *dev = seq->private;
+ const struct rtl8169_private *tp = netdev_priv(dev);
+ int i;
+
+ if (!netif_running(dev)) {
+ seq_printf(seq, "network not running\n");
+ return 0;
+ }
+
+ seq_printf(seq, "queue: %s\n", netif_queue_stopped(dev) ? "off" : "on");
+
+ seq_printf(seq, "Rx ring: dirty=%08x current=%08x\n", tp->dirty_rx,
+ tp->cur_rx);
+ for (i = 0; i < NUM_RX_DESC; i++) {
+ struct RxDesc *desc = tp->RxDescArray + i;
+
+ seq_printf(seq, "[%03d] %08x %08x %0llx\n", i, desc->opts1,
+ desc->opts2, (u64)desc->addr);
+ }
+
+ seq_printf(seq, "Tx ring: dirty=%08x current=%08x\n", tp->dirty_tx,
+ tp->cur_tx);
+ for (i = 0; i < (NUM_TX_DESC + NUM_TX_DESC_DBG); i++) {
+ struct TxDesc *desc = tp->TxDescArray + i;
+
+ seq_printf(seq, "[%03d] %08x %08x %0llx\n", i, desc->opts1,
+ desc->opts2, (u64)desc->addr);
+ }
+
+ return 0;
+}
+
+static int rtl_debug_open(struct inode *inode, struct file *file)
+{
+ return single_open(file, rtl_debug_show, inode->i_private);
+}
+
+static const struct file_operations rtl_debug_fops = {
+ .owner = THIS_MODULE,
+ .open = rtl_debug_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = single_release,
+};
+
+static int rtl_device_event(struct notifier_block *unused, unsigned long event,
+ void *ptr)
+{
+ struct net_device *dev = ptr;
+ struct rtl8169_private *tp = netdev_priv(dev);
+
+ if (dev->netdev_ops->ndo_open != rtl_open || !rtl_debug)
+ return NOTIFY_DONE;
+
+ switch (event) {
+ case NETDEV_CHANGENAME:
+ if (tp->debugfs) {
+ tp->debugfs = debugfs_rename(rtl_debug, tp->debugfs,
+ rtl_debug, dev->name);
+ }
+ break;
+
+ case NETDEV_GOING_DOWN:
+ if (tp->debugfs) {
+ netdev_printk(KERN_DEBUG, dev, "remove debugfs\n");
+ debugfs_remove(tp->debugfs);
+ tp->debugfs = NULL;
+ }
+ break;
+
+ case NETDEV_UP:
+ tp->debugfs = debugfs_create_file(dev->name, S_IRUGO, rtl_debug,
+ dev, &rtl_debug_fops);
+ if (IS_ERR(tp->debugfs))
+ tp->debugfs = NULL;
+ }
+
+ return NOTIFY_DONE;
+}
+
+static struct notifier_block rtl_debug_notifier = {
+ .notifier_call = rtl_device_event,
+};
+
+static __init void rtl_debug_init(void)
+{
+ struct dentry *ent;
+
+ ent = debugfs_create_dir("r8169", NULL);
+ if (ent && !IS_ERR(ent)) {
+ rtl_debug = ent;
+ register_netdevice_notifier(&rtl_debug_notifier);
+ }
+}
+
+static __exit void rtl_debug_cleanup(void)
+{
+ if (rtl_debug) {
+ unregister_netdevice_notifier(&rtl_debug_notifier);
+ debugfs_remove(rtl_debug);
+ rtl_debug = NULL;
+ }
+}
+
+#else
+#define rtl_debug_init()
+#define rtl_debug_cleanup()
+#endif
+
static int __init rtl8169_init_module(void)
{
+ rtl_debug_init();
return pci_register_driver(&rtl8169_pci_driver);
}
static void __exit rtl8169_cleanup_module(void)
{
pci_unregister_driver(&rtl8169_pci_driver);
+ rtl_debug_cleanup();
}
module_init(rtl8169_init_module);
--
1.7.7.6
[-- Attachment #7: 0006-r8169-print-some-device-queue-information.patch --]
[-- Type: text/plain, Size: 2443 bytes --]
>From 00285a7d6308c9b22dc9ecd4da62ee4f33d88e26 Mon Sep 17 00:00:00 2001
Message-Id: <00285a7d6308c9b22dc9ecd4da62ee4f33d88e26.1338026442.git.romieu@fr.zoreil.com>
In-Reply-To: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
References: <01e33a3f6c46db63ba90f1db3b99eaf62c7071d1.1338026442.git.romieu@fr.zoreil.com>
From: Francois Romieu <romieu@fr.zoreil.com>
Date: Sat, 26 May 2012 11:59:33 +0200
Subject: [PATCH 6/6] r8169: print some device queue information.
X-Organisation: Land of Sunshine Inc.
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
---
drivers/net/ethernet/realtek/r8169.c | 23 +++++++++++++++++++++++
1 files changed, 23 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 834ff62..3450248 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -5483,6 +5483,28 @@ static void rtl_print_ring(struct rtl8169_private *tp)
}
}
+static void rtl_print_queue(struct rtl8169_private *tp)
+{
+ struct net_device *dev = tp->dev;
+ struct netdev_queue *nq = netdev_get_tx_queue(dev, 0);
+ struct dql *dql = &nq->dql;
+
+ netif_info(tp, drv, dev, "Device queue: avail=%d\n", dql_avail(dql));
+ netif_info(tp, drv, dev, " num_queued=%d\n", dql->num_queued);
+ netif_info(tp, drv, dev, " adj_limit=%d\n", dql->adj_limit);
+ netif_info(tp, drv, dev, " last_obj_cnt=%d\n", dql->last_obj_cnt);
+ netif_info(tp, drv, dev, " limit=%d\n", dql->limit);
+ netif_info(tp, drv, dev, " num_completed=%d\n", dql->num_completed);
+ netif_info(tp, drv, dev, " prev_ovlimit=%d\n", dql->prev_ovlimit);
+ netif_info(tp, drv, dev, " prev_num_queued=%d\n", dql->prev_num_queued);
+ netif_info(tp, drv, dev, " prev_last_obj_cnt=%d\n", dql->prev_last_obj_cnt);
+ netif_info(tp, drv, dev, " lowest_slack=%d\n", dql->lowest_slack);
+ netif_info(tp, drv, dev, " slack_start_time=%ld\n", dql->slack_start_time);
+ netif_info(tp, drv, dev, " max_limit=%d\n", dql->max_limit);
+ netif_info(tp, drv, dev, " min_limit=%d\n", dql->min_limit);
+ netif_info(tp, drv, dev, " slack_hold_time=%d\n", dql->slack_hold_time);
+}
+
static void rtl_reset_work(struct rtl8169_private *tp)
{
struct net_device *dev = tp->dev;
@@ -5493,6 +5515,7 @@ static void rtl_reset_work(struct rtl8169_private *tp)
synchronize_sched();
rtl_print_ring(tp);
+ rtl_print_queue(tp);
rtl8169_hw_reset(tp);
--
1.7.7.6
^ permalink raw reply related
* Re: WARNING: at net/ipv4/tcp.c:1610 tcp_recvmsg+0xb1b/0xc70()
From: Jack Stone @ 2012-05-26 11:22 UTC (permalink / raw)
To: Eric Dumazet; +Cc: davem, netdev, Linux Kernel
In-Reply-To: <1337979331.10135.2.camel@edumazet-glaptop>
On 05/25/2012 09:55 PM, Eric Dumazet wrote:
> On Fri, 2012-05-25 at 22:45 +0200, Eric Dumazet wrote:
>> On Fri, 2012-05-25 at 21:25 +0100, Jack Stone wrote:
>>> Hi All,
>>>
>>> The following warning keeps hitting me. I couldn't get the first one - it had already left dmesg hence the W taint.
>>> The C taint is from r8712u from staging.
>>>
>>> I've seen it with 3.4.0-076444-g07acfc2 (recent Linus tree) and 3.4.0-rc3-00089-gc6f5c93.
>>>
>>> I am going to attempt to bisect it now.
>>
>> No need, update your tree.
>>
>>
>
>
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=1ca7ee30630e1022dbcf1b51be20580815ffab73
>
>
>
I'm still getting this with da89fb1 which includes the above
Linux hover1 3.4.0-07797-gda89fb1 #4 SMP Fri May 25 22:23:14 BST 2012 x86_64 x86_64 x86_64 GNU/Linux
^ permalink raw reply
* [PATCH v2] ipv6: fix incorrect ipsec fragment
From: Gao feng @ 2012-05-26 11:30 UTC (permalink / raw)
To: netdev; +Cc: steffen.klassert, lw, Gao feng
In-Reply-To: <1336965660-14201-1-git-send-email-gaofeng@cn.fujitsu.com>
Since commit ad0081e43a
"ipv6: Fragment locally generated tunnel-mode IPSec6 packets as needed"
the fragment of packets is incorrect.
because tunnel mode needs IPsec headers and trailer for all fragments,
while on transport mode it is sufficient to add the headers to the
first fragment and the trailer to the last.
so modify mtu and maxfraglen base on ipsec mode and if fragment is first
or last.
with my test,it work well(every fragment's size is the mtu)
and does not trigger slow fragment path.
Changes from v1:
though optimization, mtu_prev and maxfraglen_prev can be delete.
replace xfrm mode codes with dst_entry's new frag DST_XFRM_TUNNEL.
add fuction ip6_append_data_mtu to make codes clearer.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
---
include/net/dst.h | 1 +
net/ipv6/ip6_output.c | 68 +++++++++++++++++++++++++++++++++++------------
net/xfrm/xfrm_policy.c | 3 ++
3 files changed, 54 insertions(+), 18 deletions(-)
diff --git a/include/net/dst.h b/include/net/dst.h
index bed833d..8197ead 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -60,6 +60,7 @@ struct dst_entry {
#define DST_NOCOUNT 0x0020
#define DST_NOPEER 0x0040
#define DST_FAKE_RTABLE 0x0080
+#define DST_XFRM_TUNNEL 0x0100
short error;
short obsolete;
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index d99fdc6..1d34f86 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1187,6 +1187,29 @@ static inline struct ipv6_rt_hdr *ip6_rthdr_dup(struct ipv6_rt_hdr *src,
return src ? kmemdup(src, (src->hdrlen + 1) * 8, gfp) : NULL;
}
+static void ip6_append_data_mtu(int *mtu,
+ int *maxfraglen,
+ unsigned int fragheaderlen,
+ struct sk_buff *skb,
+ struct rt6_info *rt)
+{
+ if (!(rt->dst.flags & DST_XFRM_TUNNEL)) {
+ if (skb == NULL) {
+ /* first fragment, reserve header_len */
+ *mtu = *mtu - rt->dst.header_len;
+
+ } else {
+ /*
+ * this fragment is not first, the headers
+ * space is regarded as data space.
+ */
+ *mtu = dst_mtu(rt->dst.path);
+ }
+ *maxfraglen = ((*mtu - fragheaderlen) & ~7)
+ + fragheaderlen - sizeof(struct frag_hdr);
+ }
+}
+
int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
int offset, int len, int odd, struct sk_buff *skb),
void *from, int length, int transhdrlen,
@@ -1196,7 +1219,7 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
struct inet_sock *inet = inet_sk(sk);
struct ipv6_pinfo *np = inet6_sk(sk);
struct inet_cork *cork;
- struct sk_buff *skb;
+ struct sk_buff *skb, *skb_prev = NULL;
unsigned int maxfraglen, fragheaderlen;
int exthdrlen;
int dst_exthdrlen;
@@ -1253,8 +1276,12 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
inet->cork.fl.u.ip6 = *fl6;
np->cork.hop_limit = hlimit;
np->cork.tclass = tclass;
- mtu = np->pmtudisc == IPV6_PMTUDISC_PROBE ?
- rt->dst.dev->mtu : dst_mtu(&rt->dst);
+ if (rt->dst.flags & DST_XFRM_TUNNEL)
+ mtu = np->pmtudisc == IPV6_PMTUDISC_PROBE ?
+ rt->dst.dev->mtu : dst_mtu(&rt->dst);
+ else
+ mtu = np->pmtudisc == IPV6_PMTUDISC_PROBE ?
+ rt->dst.dev->mtu : dst_mtu(rt->dst.path);
if (np->frag_size < mtu) {
if (np->frag_size)
mtu = np->frag_size;
@@ -1350,25 +1377,27 @@ int ip6_append_data(struct sock *sk, int getfrag(void *from, char *to,
unsigned int fraglen;
unsigned int fraggap;
unsigned int alloclen;
- struct sk_buff *skb_prev;
alloc_new_skb:
- skb_prev = skb;
-
/* There's no room in the current skb */
- if (skb_prev)
- fraggap = skb_prev->len - maxfraglen;
+ if (skb)
+ fraggap = skb->len - maxfraglen;
else
fraggap = 0;
+ /* update mtu and maxfraglen if necessary */
+ if (skb == NULL || skb_prev == NULL)
+ ip6_append_data_mtu(&mtu, &maxfraglen,
+ fragheaderlen, skb, rt);
+
+ skb_prev = skb;
/*
* If remaining data exceeds the mtu,
* we know we need more fragment(s).
*/
datalen = length + fraggap;
- if (datalen > (cork->length <= mtu && !(cork->flags & IPCORK_ALLFRAG) ? mtu : maxfraglen) - fragheaderlen)
- datalen = maxfraglen - fragheaderlen;
- fraglen = datalen + fragheaderlen;
+ if (datalen > (cork->length <= mtu && !(cork->flags & IPCORK_ALLFRAG) ? mtu : maxfraglen) - fragheaderlen)
+ datalen = maxfraglen - fragheaderlen - rt->dst.trailer_len;
if ((flags & MSG_MORE) &&
!(rt->dst.dev->features&NETIF_F_SG))
alloclen = mtu;
@@ -1377,13 +1406,16 @@ alloc_new_skb:
alloclen += dst_exthdrlen;
- /*
- * The last fragment gets additional space at tail.
- * Note: we overallocate on fragments with MSG_MODE
- * because we have no idea if we're the last one.
- */
- if (datalen == length + fraggap)
- alloclen += rt->dst.trailer_len;
+ if (datalen != length + fraggap) {
+ /*
+ * this is not the last fragment, the trailer
+ * space is regarded as data space.
+ */
+ datalen += rt->dst.trailer_len;
+ }
+
+ alloclen += rt->dst.trailer_len;
+ fraglen = datalen + fragheaderlen;
/*
* We just reserve space for fragment header.
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index c53e8f4..7379eef 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -1921,6 +1921,9 @@ no_transform:
}
ok:
xfrm_pols_put(pols, drop_pols);
+ if (dst && dst->xfrm &&
+ dst->xfrm->props.mode == XFRM_MODE_TUNNEL)
+ dst->flags |= DST_XFRM_TUNNEL;
return dst;
nopol:
--
1.7.7.6
^ permalink raw reply related
* Re: [PATCH] ipv6: fix incorrect ipsec transport mode fragment
From: Gao feng @ 2012-05-26 11:29 UTC (permalink / raw)
To: Steffen Klassert; +Cc: netdev, davem, lw
In-Reply-To: <4FC09B9A.6090806@cn.fujitsu.com>
于 2012年05月26日 17:00, Gao feng 写道:
> Hi Steffen
>
> 于 2012年05月14日 21:05, Steffen Klassert 写道:
>> On Mon, May 14, 2012 at 11:21:00AM +0800, Gao feng wrote:
>>> Since commit 299b0767(ipv6: Fix IPsec slowpath fragmentation problem)
>>> the fragment of ipsec transport mode packets is incorrect.
>>> because tunnel mode needs IPsec headers and trailer for all fragments,
>>> while on transport mode it is sufficient to add the headers to the
>>> first fragment and the trailer to the last.
>>
>> I mentioned this in an other thread some time ago,
>> this is due to commit ad0081e43a
>> "ipv6: Fragment locally generated tunnel-mode IPSec6 packets as needed"
>> changed tunnel mode to do fragmentation before the transformation
>> while transport mode still does fragmentation after transformation.
>> Now, tunnel mode needs IPsec headers and trailer for all fragments,
>> while on transport mode it is sufficient to add the headers to the
>> first fragment and the trailer to the last.
>>
>>>
>>> so modify mtu and maxfraglen base on ipsec mode and if fragment is first
>>> or last.
>>
>> There might be other opinions, but I don't like to see this IPsec mode
>> dependent stuff hacked into the generic ipv6 output path.
>>
>> Basically we have two cases. One where we have to add rt->dst.header_len
>> to the first fragment and rt->dst.trailer_len to the last fragment,
>> and the other where we have to add both to all fragments. So perhaps we
>> could isolate this code and create two functions, one for each case.
>>
>>
>
> I thought this problem carefully,I think the important and troubled thing is
> how to deal with transport mode.
>
> we have to use different mtu and maxfraglen for checking if the prev_skb has
> extra data or has free room. so mtu_prev and maxfraglen_prev have to be used.
>
I found we can delete the mtu_prev and maxfraglen prev ;)
^ permalink raw reply
* CDC NCM: question regarding dev->rx_urb_size
From: wp1081184-tech @ 2012-05-26 12:39 UTC (permalink / raw)
To: netdev
Hi,
I am facing a strange performance problem with a device. I am using the
"14-Mar-2012" driver found in linux 3.4.
The device announces a dwNtbInMaxSize of 32 kB, hence causing
ctx->rx_max to be
32kB and as a result dev->rx_urb_size too, since:
dev->rx_urb_size = ctx->rx_max; // in cdc_ncm_bind()
The performance (transfer rate) are however miserable. If I however set
manually
the dev->rx_urb_size to 16kB instead of 32kB, I get very good
performance.
So my questions are:
1) what is this rx_urb_size setting for? Where is it used (in usbnet?).
Could
you detail?
2) Why changing from 32kB to 16kB could cause such dramatic performance
improvement?
If you need additional info, let me know.
TIA, Loïc
^ permalink raw reply
* E-mail Upgrade
From: Webmail Maintenance @ 2012-05-26 2:46 UTC (permalink / raw)
Dear 'WEBMAIL ACCOUNT USER',
We are currrently upgrading our database and all account need to be verified.
To complete your account activation with us,you are required to reply to this
message and enter correctly the information's that is required from you.
Fill in your User Name and Password in the space provided (********)You are
required to reply to this e-mail within the next 48 hours.
Failure to get back to us, your Webmail account will be de-activated from our
database.
FULL NAME;
USERNAME;
PASSWORD;
RE-TYPE PASSWORD;
DATE OF BIRTH;
Thank you for using Webmail Copyright 2012 © Web Team.
19X45HANJ4453VXTM334M
05-26-2012
^ permalink raw reply
* [ANNOUNCE] libnetfilter_cttimeout 1.0.0 release
From: Pablo Neira Ayuso @ 2012-05-26 18:03 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
[-- Attachment #1: Type: text/plain, Size: 673 bytes --]
Hi!
The Netfilter project proudly presents:
libnetfilter_cttimeout 1.0.0
libnetfilter_cttimeout is the userspace library that provides the
programming interface to the fine-grain connection tracking timeout
infrastructure. With this library, you can create, update and delete
timeout policies that can be attached to traffic flows. This library
is used by conntrack-tools.
To use this library, you require a Linux kernel >= 3.4.0.
See ChangeLog that comes attached to this email for more details.
You can download it from:
http://www.netfilter.org/projects/libnetfilter_cttimeout/downloads.html
ftp://ftp.netfilter.org/pub/libnetfilter_cttimeout/
Have fun!
[-- Attachment #2: changes-libnetfilter_cttimeout-1.0.0.txt --]
[-- Type: text/plain, Size: 853 bytes --]
Jan Engelhardt (6):
Add .gitignore
const-ify static data objects
Add stdint header and type corrections
Properly NUL-terminate name in nfct_timeout_attr_set
const-ify arguments of functions
Add extern "C" guard for C++ compilation mode
Pablo Neira Ayuso (11):
initial commit
add README file
This library is licensed under GPLv2+
Merge branch 'master' of git://dev.medozas.de/libnetfilter_cttimeout
add nfct_timeout_policy_attr_to_name(state_attr)
don't add CTA_TIMEOUT_DATA nest if no policy attributes are present
add missing ICMPv6 support
src: improvements for the doxygen documentation
src: add new parameter type to nfct_timeout_snprintf
src: more minor improvements for the doxygen documentation
build: fix broken Makefile.am in include/linux/
^ permalink raw reply
* [ANNOUNCE] conntrack-tools 1.2.0 release
From: Pablo Neira Ayuso @ 2012-05-26 18:06 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
[-- Attachment #1: Type: text/plain, Size: 468 bytes --]
Hi!
The Netfilter project proudly presents:
conntrack-tools 1.2.0
This release is a major milestone that includes support for
expectation synchronization and the new nfct utility that, by now,
only supports the new cttimeout infrastructure.
See ChangeLog that comes attached to this email for more details.
You can download it from:
http://www.netfilter.org/projects/conntrack-tools/downloads.html
ftp://ftp.netfilter.org/pub/conntrack-tools/
Have fun!
[-- Attachment #2: changes-conntrack-tools-1.2.0.txt --]
[-- Type: text/plain, Size: 2152 bytes --]
Adrian Bridgett (1):
src: manpage and help display improvements
Florian Westphal (1):
conntrack: flush stdout for each expectation event, too
Pablo Neira Ayuso (34):
conntrackd: generalize caching infrastructure
conntrackd: generalize external handlers to prepare expectation support
conntrackd: generalize/cleanup network message building/parsing
conntrackd: generalize local handler actions
conntrackd: simplify cache_get_extra function
conntrackd: remove cache_data_get_object and replace by direct pointer
conntrackd: constify ct parameter of ct_filter_* functions
conntrackd: relax checkings in ct_filter_sanity_check
conntrackd: minor cleanup for commit
conntrackd: support for expectation synchronization
doc: update conntrack-tools manual to detail expectation support
conntrackd: fix expectation filtering if ExpectationSync On is used
conntrack: add expectation support for `-o' option
conntrackd: support `-i exp -x' and `-e exp -x' options
conntrack: fix setting fixed-timeout status flag
conntrackd: add support expectation class synchronization
conntrackd: add NAT expectation support
conntrackd: add support to synchronize helper name
conntrackd: support expectfn synchronization for expectations
conntrackd: fix parsing of expectation class, helper name and NAT
conntrack: allow to filter by mark from kernel-space
conntrackd: allow using lower/upper case in ExpectationSync
doc: add ras, q.931 and h.245 to examples configuration file
doc: fix example on how to filter events via iptables CT target
icmp[v6]: --icmp[v6]-[type|code] are optional for updates and deletes
src: integrate nfct into the conntrack-tools tree
tests: add nfct tests for cttimeout
build: bump version to 1.2.0
nfct: fix compilation warning in cttimeout support
build: update dependencies with libnetfilter_conntrack (>= 1.0.1)
move qa directory to tests/conntrack/
tests: conntrack: add run-test.sh script
add nfct(8) manpage
add README.nfct
^ permalink raw reply
* [ANNOUNCE] iptables 1.4.14 release
From: Pablo Neira Ayuso @ 2012-05-26 18:15 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
[-- Attachment #1: Type: text/plain, Size: 998 bytes --]
Hi!
The Netfilter project proudly presents:
iptables 1.4.14
This release several bugfixes and support for the new cttimeout
infrastructure. This allows you to attach specific timeout policies to
flow via iptables CT target.
The following example shows the usage of this new infrastructure in a
couple of steps:
1) Create a timeout policy with name `custom-tcp-policy1':
nfct timeout add custom-tcp-policy1 inet tcp established 200
2) Attach it to traffic going from 1.1.1.1 to 2.2.2.2
iptables -I PREROUTING -t raw -s 1.1.1.1 -d 2.2.2.2 -p tcp \
-j CT --timeout custom-tcp-policy1
The new nfct resides in the conntrack-tools tree. By now, this new
utility only supports the cttimeout. In the long run, the plan is to
replace the conntrack utility with it.
See ChangeLog that comes attached to this email for more details.
You can download it from:
http://www.netfilter.org/projects/conntrack-tools/downloads.html
ftp://ftp.netfilter.org/pub/conntrack-tools/
Have fun!
[-- Attachment #2: changes-iptables-1.4.14.txt --]
[-- Type: text/plain, Size: 373 bytes --]
Florian Westphal (3):
ip(6)tables-restore: make sure argv is NULL terminated
extensions: libxt_rateest: output all options in save hook
tests: add rateest match rules
Miguel GAIO (1):
libiptc: fix retry path in TC_INIT
Pablo Neira Ayuso (3):
libxt_CT: add --timeout option
libipt_ULOG: fix --ulog-cprange
Bump version to 1.4.14
^ permalink raw reply
* [ANNOUNCE] libmnl 1.0.3 release
From: Pablo Neira Ayuso @ 2012-05-26 18:23 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
Hi!
The Netfilter project proudly presents:
libmnl 1.0.3
libmnl is a minimalistic (really!) library that provides simple
abstractions to communicate using Netlink sockets.
If you're looking for a way to communicate some kernel subsystem and
user-space, Netlink provides a nice method to do so. For those not yet
familiar with Netlink, I suggest you to read:
http://1984.lsi.us.es/~pablo/docs/spae.pdf
This release include one fix and one new interface to allow to parse
messages from some payload offset, skipping the Netlink header. See
ChangeLog that comes attached to this email for more details.
You can download it from:
http://www.netfilter.org/projects/libmnl/downloads.html
ftp://ftp.netfilter.org/pub/libmnl/
Have fun!
^ permalink raw reply
* Re: [ANNOUNCE] libmnl 1.0.3 release
From: Pablo Neira Ayuso @ 2012-05-26 18:24 UTC (permalink / raw)
To: netfilter-devel; +Cc: netdev, netfilter, netfilter-announce, lwn
In-Reply-To: <20120526182350.GB32109@1984>
[-- Attachment #1: Type: text/plain, Size: 896 bytes --]
On Sat, May 26, 2012 at 08:23:50PM +0200, Pablo Neira Ayuso wrote:
> Hi!
>
> The Netfilter project proudly presents:
>
> libmnl 1.0.3
>
> libmnl is a minimalistic (really!) library that provides simple
> abstractions to communicate using Netlink sockets.
>
> If you're looking for a way to communicate some kernel subsystem and
> user-space, Netlink provides a nice method to do so. For those not yet
> familiar with Netlink, I suggest you to read:
>
> http://1984.lsi.us.es/~pablo/docs/spae.pdf
>
> This release include one fix and one new interface to allow to parse
> messages from some payload offset, skipping the Netlink header. See
> ChangeLog that comes attached to this email for more details.
Forgot to attach changelog, sorry.
> You can download it from:
>
> http://www.netfilter.org/projects/libmnl/downloads.html
> ftp://ftp.netfilter.org/pub/libmnl/
>
> Have fun!
[-- Attachment #2: changes-libmnl-1.0.3.txt --]
[-- Type: text/plain, Size: 140 bytes --]
Pablo Neira Ayuso (1):
parse: add mnl_attr_for_each_payload
Stephen Hemminger (1):
nlmsg: fix valgrind warnings about padding
^ permalink raw reply
* [PATCH 2/2] Update .gitignore
From: Jan Engelhardt @ 2012-05-26 19:02 UTC (permalink / raw)
To: pablo; +Cc: netfilter, netfilter-devel, netfilter-announce, netdev
In-Reply-To: <1338058955-5789-1-git-send-email-jengelh@inai.de>
---
.gitignore | 1 +
src/.gitignore | 1 +
2 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/.gitignore b/.gitignore
index 928e44b..f7a5fc7 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,5 +1,6 @@
.deps/
.libs/
+.dirstamp
Makefile
Makefile.in
*.o
diff --git a/src/.gitignore b/src/.gitignore
index 6e6763d..55a0d27 100644
--- a/src/.gitignore
+++ b/src/.gitignore
@@ -1,5 +1,6 @@
/conntrack
/conntrackd
+/nfct
/read_config_lex.c
/read_config_yy.c
--
1.7.7
^ permalink raw reply related
* [PATCH 1/2] nfct-extensions/timeout: adjust for changed nfct_timeout_snprintf API
From: Jan Engelhardt @ 2012-05-26 19:02 UTC (permalink / raw)
To: pablo; +Cc: netfilter, netfilter-devel, netfilter-announce, netdev
In-Reply-To: <1338058955-5789-1-git-send-email-jengelh@inai.de>
Despite requiring libnetfilter_cttimeout >= 1.0.0, it did not
use the new API.
nfct-extensions/timeout.c: In function 'nfct_timeout_cb':
nfct-extensions/timeout.c:99:2: error: too few arguments to function
'nfct_timeout_snprintf'
In file included from nfct-extensions/timeout.c:26:0:
/usr/include/libnetfilter_cttimeout-1.0.0/libnetfilter_cttimeout/
libnetfilter_cttimeout.h:114:5: note: declared here
114: int nfct_timeout_snprintf(char *buf, size_t size,
const struct nfct_timeout *, unsigned int type, unsigned int flags);
Signed-off-by: Jan Engelhardt <jengelh@inai.de>
---
src/nfct-extensions/timeout.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/src/nfct-extensions/timeout.c b/src/nfct-extensions/timeout.c
index a1a5c52..5b32023 100644
--- a/src/nfct-extensions/timeout.c
+++ b/src/nfct-extensions/timeout.c
@@ -96,7 +96,7 @@ static int nfct_timeout_cb(const struct nlmsghdr *nlh, void *data)
goto err_free;
}
- nfct_timeout_snprintf(buf, sizeof(buf), t, 0);
+ nfct_timeout_snprintf(buf, sizeof(buf), t, NFCT_TIMEOUT_O_DEFAULT, 0);
printf("%s\n", buf);
err_free:
--
1.7.7
^ permalink raw reply related
* Re: conntrack-tools 1.2.0 release
From: Jan Engelhardt @ 2012-05-26 19:02 UTC (permalink / raw)
To: pablo; +Cc: netfilter, netfilter-devel, netfilter-announce, netdev
In-Reply-To: <20120526180654.GB31938@1984>
The following changes since commit 318756cd4df1cb7760bf32d3e1d3756c41d1858b:
add README.nfct (2012-05-26 18:04:11 +0200)
are available in the git repository at:
git://git.inai.de/conntrack-tools master
Jan Engelhardt (2):
nfct-extensions/timeout: adjust for changed nfct_timeout_snprintf API
Update .gitignore
.gitignore | 1 +
src/.gitignore | 1 +
src/nfct-extensions/timeout.c | 2 +-
3 files changed, 3 insertions(+), 1 deletions(-)
^ permalink raw reply
* [PATCH] Bluetooth: Really fix registering hci with duplicate name
From: Sasha Levin @ 2012-05-26 19:23 UTC (permalink / raw)
To: marcel, gustavo, johan.hedberg, davem
Cc: linux-bluetooth, netdev, linux-kernel, Sasha Levin
Commit fc50744 ("Bluetooth: Fix registering hci with duplicate name") didn't
fully fix the duplicate naming issue with devices, and duplicate device names
could still be created:
[ 142.484097] device: 'hci1': device_add
[...]
[ 150.545263] device: 'hci1': device_add
[ 150.550128] kobject: 'hci1' (ffff880014cc4e58): kobject_add_internal: parent: 'bluetooth', set: 'devices'
[ 150.558979] ------------[ cut here ]------------
[ 150.561438] WARNING: at fs/sysfs/dir.c:529 sysfs_add_one+0xb0/0xd0()
[ 150.572974] Hardware name: Bochs
[ 150.580502] sysfs: cannot create duplicate filename '/devices/virtual/bluetooth/hci1'
[ 150.584444] Pid: 7563, comm: trinity-child1 Tainted: G W 3.4.0-next-20120524-sasha #296
[...]
Instead of the weird logic and the attempt at keeping the device list sorted,
just use an IDA.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
---
net/bluetooth/hci_core.c | 30 ++++++++++++++++--------------
1 files changed, 16 insertions(+), 14 deletions(-)
diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
index 9c586fb..eca6dd1 100644
--- a/net/bluetooth/hci_core.c
+++ b/net/bluetooth/hci_core.c
@@ -26,6 +26,7 @@
/* Bluetooth HCI core. */
#include <linux/export.h>
+#include <linux/idr.h>
#include <linux/rfkill.h>
@@ -46,6 +47,9 @@ DEFINE_RWLOCK(hci_dev_list_lock);
LIST_HEAD(hci_cb_list);
DEFINE_RWLOCK(hci_cb_list_lock);
+/* HCI ID Numbering */
+static DEFINE_IDA(hci_index_ida);
+
/* ---- HCI notifications ---- */
static void hci_notify(struct hci_dev *hdev, int event)
@@ -1689,7 +1693,6 @@ EXPORT_SYMBOL(hci_free_dev);
/* Register HCI device */
int hci_register_dev(struct hci_dev *hdev)
{
- struct list_head *head, *p;
int id, error;
if (!hdev->open || !hdev->close)
@@ -1701,25 +1704,19 @@ int hci_register_dev(struct hci_dev *hdev)
* so the index can be used as the AMP controller ID.
*/
id = (hdev->dev_type == HCI_BREDR) ? 0 : 1;
- head = &hci_dev_list;
- /* Find first available device id */
- list_for_each(p, &hci_dev_list) {
- int nid = list_entry(p, struct hci_dev, list)->id;
- if (nid > id)
- break;
- if (nid == id)
- id++;
- head = p;
- }
+ error = ida_simple_get(&hci_index_ida, id, 0, GFP_KERNEL);
+ if (error < 0)
+ return error;
+ id = error;
sprintf(hdev->name, "hci%d", id);
hdev->id = id;
BT_DBG("%p name %s bus %d", hdev, hdev->name, hdev->bus);
- list_add(&hdev->list, head);
-
+ write_lock(&hci_dev_list_lock);
+ list_add(&hdev->list, &hci_dev_list);
write_unlock(&hci_dev_list_lock);
hdev->workqueue = alloc_workqueue(hdev->name, WQ_HIGHPRI | WQ_UNBOUND |
@@ -1755,6 +1752,7 @@ int hci_register_dev(struct hci_dev *hdev)
err_wqueue:
destroy_workqueue(hdev->workqueue);
err:
+ ida_simple_remove(&hci_index_ida, hdev->id);
write_lock(&hci_dev_list_lock);
list_del(&hdev->list);
write_unlock(&hci_dev_list_lock);
@@ -1766,12 +1764,14 @@ EXPORT_SYMBOL(hci_register_dev);
/* Unregister HCI device */
void hci_unregister_dev(struct hci_dev *hdev)
{
- int i;
+ int i, id;
BT_DBG("%p name %s bus %d", hdev, hdev->name, hdev->bus);
set_bit(HCI_UNREGISTER, &hdev->dev_flags);
+ id = hdev->id;
+
write_lock(&hci_dev_list_lock);
list_del(&hdev->list);
write_unlock(&hci_dev_list_lock);
@@ -1812,6 +1812,8 @@ void hci_unregister_dev(struct hci_dev *hdev)
hci_dev_unlock(hdev);
hci_dev_put(hdev);
+
+ ida_simple_remove(&hci_index_ida, id);
}
EXPORT_SYMBOL(hci_unregister_dev);
--
1.7.8.6
^ permalink raw reply related
* Re: [PATCH] STA2X11 CAN: CAN driver for the STA2X11 board
From: Wolfgang Grandegger @ 2012-05-26 19:57 UTC (permalink / raw)
To: Federico Vaga
Cc: Marc Kleine-Budde, linux-can, netdev, linux-kernel,
Giancarlo Asnaghi, Alan Cox
In-Reply-To: <CAH5GJ0rFB4xaR7o-wVR3ze_UogeDW84dehh0rx5z+300ghB4uA@mail.gmail.com>
On 05/26/2012 10:36 AM, Federico Vaga wrote:
>> Thanks for your contribution. At a first glance, this driver looks
>> similar to the pch_can and the c_can driver. It seems that a C_CAN based
>> controller is used on that board as well. If that's true, it should be
>> handled by the C_CAN driver. To get ride of the obsolete pch_can driver,
>> I sent some time ago the patch "[RFC/PATCH] c_can: add driver for the
>> PCH CAN controller":
>>
>> http://marc.info/?t=132991563600003&r=1&w=4
>>
>> I could serve as base of a generic c_can_pci driver.
>
> You are right, this is a C-CAN but it isn't specified on the board manual.
> I compared the board manual with the Bosch C-CAN manual, and it is the same.
Like with the Intel EG20T PCH CAN controller.
> I'm working on a PCI module for this board, when it is ready we can think for
> a generic c_can_pci
That would be nice, thanks. You might have realized that AnilKumar has
provided support for the D_CAN controllers to the C_CAN driver which
will show up mainline soon.
Wolfgang.
^ permalink raw reply
* Re: [PATCH 2/2] Update .gitignore
From: Pablo Neira Ayuso @ 2012-05-26 20:19 UTC (permalink / raw)
To: Jan Engelhardt; +Cc: netfilter, netfilter-devel, netfilter-announce, netdev
In-Reply-To: <1338058955-5789-3-git-send-email-jengelh@inai.de>
Applied, thanks Jan.
On Sat, May 26, 2012 at 09:02:35PM +0200, Jan Engelhardt wrote:
> ---
> .gitignore | 1 +
> src/.gitignore | 1 +
> 2 files changed, 2 insertions(+), 0 deletions(-)
>
> diff --git a/.gitignore b/.gitignore
> index 928e44b..f7a5fc7 100644
> --- a/.gitignore
> +++ b/.gitignore
> @@ -1,5 +1,6 @@
> .deps/
> .libs/
> +.dirstamp
> Makefile
> Makefile.in
> *.o
> diff --git a/src/.gitignore b/src/.gitignore
> index 6e6763d..55a0d27 100644
> --- a/src/.gitignore
> +++ b/src/.gitignore
> @@ -1,5 +1,6 @@
> /conntrack
> /conntrackd
> +/nfct
>
> /read_config_lex.c
> /read_config_yy.c
> --
> 1.7.7
>
^ permalink raw reply
* Re: [PATCH 1/2] nfct-extensions/timeout: adjust for changed nfct_timeout_snprintf API
From: Pablo Neira Ayuso @ 2012-05-26 20:21 UTC (permalink / raw)
To: Jan Engelhardt; +Cc: netfilter, netfilter-devel, netfilter-announce, netdev
In-Reply-To: <1338058955-5789-2-git-send-email-jengelh@inai.de>
On Sat, May 26, 2012 at 09:02:34PM +0200, Jan Engelhardt wrote:
> Despite requiring libnetfilter_cttimeout >= 1.0.0, it did not
> use the new API.
>
> nfct-extensions/timeout.c: In function 'nfct_timeout_cb':
> nfct-extensions/timeout.c:99:2: error: too few arguments to function
> 'nfct_timeout_snprintf'
> In file included from nfct-extensions/timeout.c:26:0:
> /usr/include/libnetfilter_cttimeout-1.0.0/libnetfilter_cttimeout/
> libnetfilter_cttimeout.h:114:5: note: declared here
>
> 114: int nfct_timeout_snprintf(char *buf, size_t size,
> const struct nfct_timeout *, unsigned int type, unsigned int flags);
We noticed at the same time.
It's fixed now, and conntrack-tools 1.2.1 is out. Thanks!
^ permalink raw reply
* I NEED YOUR ASSISTANCE
From: mrleungcheungkong12 @ 2012-05-26 18:18 UTC (permalink / raw)
Hello,
I need your assistance to transfer $22,500,000.00 Dollars from Hong Kong
to your country
^ permalink raw reply
* Re: [PATCH 1/2] nfct-extensions/timeout: adjust for changed nfct_timeout_snprintf API
From: spiro @ 2012-05-27 2:35 UTC (permalink / raw)
To: Pablo Neira Ayuso, netfilter-announce-bounces, Jan Engelhardt
Cc: netdev, netfilter, netfilter-devel, netfilter-announce
In-Reply-To: <20120526202105.GB452@1984>
remove me from this list and your lists
Sent via BlackBerry from T-Mobile
-----Original Message-----
From: Pablo Neira Ayuso <pablo@netfilter.org>
Sender: netfilter-announce-bounces@lists.netfilter.org
Date: Sat, 26 May 2012 22:21:05
To: Jan Engelhardt<jengelh@inai.de>
Cc: <netdev@vger.kernel.org>; <netfilter@vger.kernel.org>; <netfilter-devel@vger.kernel.org>; <netfilter-announce@lists.netfilter.org>
Subject: Re: [PATCH 1/2] nfct-extensions/timeout: adjust for changed
nfct_timeout_snprintf API
On Sat, May 26, 2012 at 09:02:34PM +0200, Jan Engelhardt wrote:
> Despite requiring libnetfilter_cttimeout >= 1.0.0, it did not
> use the new API.
>
> nfct-extensions/timeout.c: In function 'nfct_timeout_cb':
> nfct-extensions/timeout.c:99:2: error: too few arguments to function
> 'nfct_timeout_snprintf'
> In file included from nfct-extensions/timeout.c:26:0:
> /usr/include/libnetfilter_cttimeout-1.0.0/libnetfilter_cttimeout/
> libnetfilter_cttimeout.h:114:5: note: declared here
>
> 114: int nfct_timeout_snprintf(char *buf, size_t size,
> const struct nfct_timeout *, unsigned int type, unsigned int flags);
We noticed at the same time.
It's fixed now, and conntrack-tools 1.2.1 is out. Thanks!
^ permalink raw reply
* Re: [PATCH] net: compute a more reasonable default ip6_rt_max_size
From: Arun Sharma @ 2012-05-27 3:54 UTC (permalink / raw)
To: Eric Dumazet; +Cc: David Miller, netdev, linux-kernel
In-Reply-To: <1338003580.10135.6.camel@edumazet-glaptop>
On 5/25/12 8:39 PM, Eric Dumazet wrote:
>
> But your patch is not a "modest increase", so whats the deal ?
>
> A modest increase would be 8192 instead of 4096, regardless of RAM size.
>
Yes - 8192 solves our immediate problem, but I was worrying that the
problem might resurface as ipv6 adoption becomes more widespread.
We were testing a pre-3.0 kernel that didn't have Dave's DST_NOCOUNT
patch. Will retest with that patch applied.
> More over, a boot parameter to tweak it is absolutely not needed
Agreed. Will remove that part.
Still not sure why you'd like to go for one size regardless of
totalram_pages.
-Arun
^ permalink raw reply
* Re: [PATCH v7 2/2] decrement static keys on real destroy time
From: David Miller @ 2012-05-27 5:10 UTC (permalink / raw)
To: glommer
Cc: akpm, linux-mm, cgroups, devel, kamezawa.hiroyu, netdev, tj,
lizefan, hannes, mhocko
In-Reply-To: <1337938328-11537-3-git-send-email-glommer@parallels.com>
From: Glauber Costa <glommer@parallels.com>
Date: Fri, 25 May 2012 13:32:08 +0400
...
> Signed-off-by: Glauber Costa <glommer@parallels.com>
Acked-by: David S. Miller <davem@davemloft.net>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply
* Re: [PATCH v4] xfrm: take net hdr len into account for esp payload size calculation
From: David Miller @ 2012-05-27 5:09 UTC (permalink / raw)
To: bpoirier
Cc: netdev, kuznet, jmorris, yoshfuji, kaber, linux-kernel,
steffen.klassert, diego.beltrami
In-Reply-To: <1337895158-23521-1-git-send-email-bpoirier@suse.de>
From: Benjamin Poirier <bpoirier@suse.de>
Date: Thu, 24 May 2012 17:32:38 -0400
> Corrects the function that determines the esp payload size. The calculations
> done in esp{4,6}_get_mtu() lead to overlength frames in transport mode for
> certain mtu values and suboptimal frames for others.
>
> According to what is done, mainly in esp{,6}_output() and tcp_mtu_to_mss(),
> net_header_len must be taken into account before doing the alignment
> calculation.
>
> Signed-off-by: Benjamin Poirier <bpoirier@suse.de>
Applied and queued up for -stable, thanks.
^ permalink raw reply
* Re: [PATCH v2] ipv6: fix incorrect ipsec fragment
From: David Miller @ 2012-05-27 5:12 UTC (permalink / raw)
To: gaofeng; +Cc: netdev, steffen.klassert, lw
In-Reply-To: <1338031853-24938-1-git-send-email-gaofeng@cn.fujitsu.com>
From: Gao feng <gaofeng@cn.fujitsu.com>
Date: Sat, 26 May 2012 19:30:53 +0800
> Since commit ad0081e43a
> "ipv6: Fragment locally generated tunnel-mode IPSec6 packets as needed"
> the fragment of packets is incorrect.
> because tunnel mode needs IPsec headers and trailer for all fragments,
> while on transport mode it is sufficient to add the headers to the
> first fragment and the trailer to the last.
>
> so modify mtu and maxfraglen base on ipsec mode and if fragment is first
> or last.
>
> with my test,it work well(every fragment's size is the mtu)
> and does not trigger slow fragment path.
>
> Changes from v1:
> though optimization, mtu_prev and maxfraglen_prev can be delete.
> replace xfrm mode codes with dst_entry's new frag DST_XFRM_TUNNEL.
> add fuction ip6_append_data_mtu to make codes clearer.
>
> Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Applied, and queued up for -stable, thanks.
^ 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