* Re: [PATCH NEXT 3/3] netxen: add hardware LRO support
From: David Miller @ 2009-08-05 19:41 UTC (permalink / raw)
To: dhananjay; +Cc: davem, netdev, narender.kumar
In-Reply-To: <1249493649-26336-3-git-send-email-dhananjay@netxen.com>
From: Dhananjay Phadke <dhananjay@netxen.com>
Date: Wed, 5 Aug 2009 10:34:09 -0700
> Add support to handle aggregate packets from firmware.
> Local TCP flows are automatically identified by firmware
> based on the dest IP hash added by driver for local IP
> addresses.
>
> The packets are sent down on the jumbo rx ring.
>
> Signed-off-by: Narender Kumar <narender.kumar@qlogic.com>
> Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Applied.
^ permalink raw reply
* Re: [PATCH NEXT 2/3] netxen: remove unnecessary structures
From: David Miller @ 2009-08-05 19:41 UTC (permalink / raw)
To: dhananjay; +Cc: davem, netdev
In-Reply-To: <1249493649-26336-2-git-send-email-dhananjay@netxen.com>
From: Dhananjay Phadke <dhananjay@netxen.com>
Date: Wed, 5 Aug 2009 10:34:08 -0700
> Remove unnecessary offsetof calulations on these structures:
> netxen_board_info, netxen_user_old_info, netxen_new_user_info.
>
> The offsets into the flash are fixed, don't need to be calculated.
>
> Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Applied.
^ permalink raw reply
* Re: [PATCH NEXT 1/3] netxen: fix vlan tso case
From: David Miller @ 2009-08-05 19:41 UTC (permalink / raw)
To: dhananjay; +Cc: davem, netdev
In-Reply-To: <1249493649-26336-1-git-send-email-dhananjay@netxen.com>
From: Dhananjay Phadke <dhananjay@netxen.com>
Date: Wed, 5 Aug 2009 10:34:07 -0700
> Fix the calculation of remaining header length in TSO
> over vlan device case. This was inadvertently missed
> out in patch 028afe719855a157e32450c ("netxen: add vlan
> tx acceleration support").
>
> Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Applied.
^ permalink raw reply
* Re: [PATCH 6/5] cpmac: unmark as broken
From: David Miller @ 2009-08-05 19:24 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042317.49497.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 23:17:49 +0200
> Starting with version 0.5.1, cpmac is no longer broken.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 5/5] cpmac: bump version to 0.5.1
From: David Miller @ 2009-08-05 19:23 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042253.00563.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 22:53:00 +0200
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 4/5] cpmac: wait longer after MDIO reset
From: David Miller @ 2009-08-05 19:23 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042252.57848.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 22:52:57 +0200
> This patch slows down the MDIO_ALIVE busy waiting to let
> switches and PHY come up after reset. Previous loop was
> too quick for IC+175C and ADM6996C/L switches to come up.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 3/5] cpmac: add support for fixed PHY
From: David Miller @ 2009-08-05 19:23 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042252.53068.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 22:52:52 +0200
> This patch adds support for fixed PHY connected in MII mode
> to cpmac. We allow external and dumb_switch module parameters
> to override the PHY detection process since they are always connected
> with MDIO bus identifier 0. This lets fixed PHYs to be detected
> correctly and be connected to the their corresponding MDIO
> bus identifier.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 2/5] ar7: add fixed PHY support for the two on-board cpmac
From: David Miller @ 2009-08-05 19:23 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042252.47549.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 22:52:47 +0200
> This patch adds fixed PHY support for the two on-chip
> cpmac Ethernet adapters.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 1/5] cpmac: fix wrong MDIO bus identifier
From: David Miller @ 2009-08-05 19:23 UTC (permalink / raw)
To: florian; +Cc: ralf, linux-mips, netdev
In-Reply-To: <200908042252.42254.florian@openwrt.org>
From: Florian Fainelli <florian@openwrt.org>
Date: Tue, 4 Aug 2009 22:52:41 +0200
> This patch fixes the wrong MDIO bus identifier which was
> set to 0 unconditionaly, suitable for external switches while
> it is actually 1 for PHYs different than external switches
> which are autodetected.
>
> Signed-off-by: Florian Fainelli <florian@openwrt.org>
Applied to net-next-2.6
^ permalink raw reply
* Re: [PATCH 2/3] af_ieee802154: provide dummy get/setsockopt
From: David Miller @ 2009-08-05 19:17 UTC (permalink / raw)
To: dbaryshkov; +Cc: netdev, slapin, linux-zigbee-devel
In-Reply-To: <1249457297-8355-3-git-send-email-dbaryshkov@gmail.com>
From: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Date: Wed, 5 Aug 2009 11:28:16 +0400
> Provide dummt get/setsockopt implementations to stop these
> syscalls from oopsing on our sockets.
>
> Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
See "sock_no_getsockopt()" and "sock_no_setsockopt()" which are
provided specifically for this situation.
^ permalink raw reply
* Re: [PATCH] xfrm: xfrm hash to use Jenkins' hash
From: David Miller @ 2009-08-05 19:08 UTC (permalink / raw)
To: joamaki; +Cc: netdev
In-Reply-To: <ce10d01f0908050041p43c308fav72ac640f665e161f@mail.gmail.com>
From: Jussi Mäki <joamaki@gmail.com>
Date: Wed, 5 Aug 2009 10:41:42 +0300
> Hi,
>
> The current xfrm hash functions perform very poorly when a number of
> policies have the same
> last byte in source and destination addresses.
>
> For example with __xfrm_dst_hash, hmask of 0xfff:
>
> 192.168.0.1-172.16.0.1 hashes to 3258
> 192.168.0.2-172.16.0.2 hashes to 3258
> ... and so on.
>
> This patch addresses the issue by rewriting the xfrm
> hash functions to use the Jenkins' hash function.
>
> Signed-off-by: Jussi Maki <joamaki@gmail.com>
jhash expands to a lot of code, and given your description of the
problem, you could have fixed it by adding 2 instructions (see below)
instead of 20 or 30 (jhash instruction count) at every hash
calculation site.
Simply change every instance of:
(h >> 16)
with
((h >> 16) ^ (h >> 24))
As much as I love jhash, it's overkill for fixing this problem.
And if we do end up using jhash, it should get inlined into a
seperate non-inline function instead of expanding that monster
4 or 5 times throughout the XFRM code.
I'm not applying this, either make the simple one-liner fix I
suggested above work or move the jhash into a non-inline expansion.
^ permalink raw reply
* Re: r8169 (+others ?) and note_interrupt performance hit on 2.6.30.x
From: Michal Soltys @ 2009-08-05 18:54 UTC (permalink / raw)
To: netdev; +Cc: Jarek Poplawski, Francois Romieu
In-Reply-To: <4A6E21B4.7090103@gmail.com>
Jarek Poplawski wrote:
> Michal Soltys wrote, On 07/26/2009 08:43 PM:
>
> ...
>> Anwyay, not sure if this is useful, so if there is anything better
>> I could do, just push me in the right direction.
>
> Short test of 2.6.29? (Unless you prefer a bisection...)
>
> Jarek P.
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
I did first bisection, and the whole cpu load increases seem a bit more
complicated than I originally thought.
I noticed at least 2 (possibly 3) moments where it increased (from
original full 1 gbit at 20% - 50% load of one core)
(1) jump from 20% - 50% to 80% - 90%
(2) commit f11a377b3f4e897d11f0e8d1fc688667e2f19708
r8169: avoid losing MSI interrupts
After the above commit, one core is not enough to handle 1 gbit speed
with 1k packets, it ends with with ~760mbit and one core drained at 100%
. The situation right before this commit is as in (1)
All tests were done, as mentioned earlier - with nc process and eth
interrupts pinned to the same core, and besides nc process - idle
machine. Tested with udp only, standard pfifo_fast.
The changes happened between 2.6.29 and 2.6.30-rc6 .
I'll try to pinpoint when (1) happens now. Lookint at mpstat output,
those may actually be two different commits.
^ permalink raw reply
* [PATCH 5/5] net: smsc911x: switch to new dev_pm_ops
From: Daniel Mack @ 2009-08-05 18:29 UTC (permalink / raw)
To: linux-kernel
Cc: linux-pm, Daniel Mack, Steve Glendinning, David S. Miller, netdev
In-Reply-To: <1249496971-9019-4-git-send-email-daniel@caiaq.de>
Hibernation is unsupported for now, which meets the actual
implementation in the driver. For free/thaw, the chip's D2 state should
be entered.
Signed-off-by: Daniel Mack <daniel@caiaq.de>
Cc: Steve Glendinning <steve.glendinning@smsc.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
drivers/net/smsc911x.c | 31 ++++++++++++++++++++-----------
1 files changed, 20 insertions(+), 11 deletions(-)
diff --git a/drivers/net/smsc911x.c b/drivers/net/smsc911x.c
index 94b6d26..c266785 100644
--- a/drivers/net/smsc911x.c
+++ b/drivers/net/smsc911x.c
@@ -50,6 +50,7 @@
#include <linux/swab.h>
#include <linux/phy.h>
#include <linux/smsc911x.h>
+#include <linux/device.h>
#include "smsc911x.h"
#define SMSC_CHIPNAME "smsc911x"
@@ -2114,10 +2115,12 @@ out_0:
/* This implementation assumes the devices remains powered on its VDDVARIO
* pins during suspend. */
-static int smsc911x_suspend(struct platform_device *pdev, pm_message_t state)
+/* TODO: implement freeze/thaw callbacks for hibernation.*/
+
+static int smsc911x_suspend(struct device *dev)
{
- struct net_device *dev = platform_get_drvdata(pdev);
- struct smsc911x_data *pdata = netdev_priv(dev);
+ struct net_device *ndev = dev_get_drvdata(dev);
+ struct smsc911x_data *pdata = netdev_priv(ndev);
/* enable wake on LAN, energy detection and the external PME
* signal. */
@@ -2128,10 +2131,10 @@ static int smsc911x_suspend(struct platform_device *pdev, pm_message_t state)
return 0;
}
-static int smsc911x_resume(struct platform_device *pdev)
+static int smsc911x_resume(struct device *dev)
{
- struct net_device *dev = platform_get_drvdata(pdev);
- struct smsc911x_data *pdata = netdev_priv(dev);
+ struct net_device *ndev = dev_get_drvdata(dev);
+ struct smsc911x_data *pdata = netdev_priv(ndev);
unsigned int to = 100;
/* Note 3.11 from the datasheet:
@@ -2149,19 +2152,25 @@ static int smsc911x_resume(struct platform_device *pdev)
return (to == 0) ? -EIO : 0;
}
+static struct dev_pm_ops smsc911x_pm_ops = {
+ .suspend = smsc911x_suspend,
+ .resume = smsc911x_resume,
+};
+
+#define SMSC911X_PM_OPS (&smsc911x_pm_ops)
+
#else
-#define smsc911x_suspend NULL
-#define smsc911x_resume NULL
+#define SMSC911X_PM_OPS NULL
#endif
static struct platform_driver smsc911x_driver = {
.probe = smsc911x_drv_probe,
.remove = __devexit_p(smsc911x_drv_remove),
.driver = {
- .name = SMSC_CHIPNAME,
+ .name = SMSC_CHIPNAME,
+ .owner = THIS_MODULE,
+ .pm = SMSC911X_PM_OPS,
},
- .suspend = smsc911x_suspend,
- .resume = smsc911x_resume,
};
/* Entry point for loading the module */
--
1.6.3.3
^ permalink raw reply related
* Re: net: mark read-only arrays as const
From: David Miller @ 2009-08-05 17:43 UTC (permalink / raw)
To: jengelh; +Cc: netdev
In-Reply-To: <alpine.LSU.2.00.0908051708340.11618@fbirervta.pbzchgretzou.qr>
From: Jan Engelhardt <jengelh@medozas.de>
Date: Wed, 5 Aug 2009 17:10:36 +0200 (CEST)
> net: mark read-only arrays as const
>
> String literals are constant, and usually, we can also tag the array
> of pointers const too, moving it to the .rodata section.
>
> Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Looks good, applied.
^ permalink raw reply
* [PATCH NEXT 3/3] netxen: add hardware LRO support
From: Dhananjay Phadke @ 2009-08-05 17:34 UTC (permalink / raw)
To: davem; +Cc: netdev, Narender Kumar
In-Reply-To: <1249493649-26336-1-git-send-email-dhananjay@netxen.com>
Add support to handle aggregate packets from firmware.
Local TCP flows are automatically identified by firmware
based on the dest IP hash added by driver for local IP
addresses.
The packets are sent down on the jumbo rx ring.
Signed-off-by: Narender Kumar <narender.kumar@qlogic.com>
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
---
drivers/net/netxen/netxen_nic.h | 21 ++++++
drivers/net/netxen/netxen_nic_ctx.c | 2 +
drivers/net/netxen/netxen_nic_init.c | 124 ++++++++++++++++++++++++++++------
3 files changed, 125 insertions(+), 22 deletions(-)
diff --git a/drivers/net/netxen/netxen_nic.h b/drivers/net/netxen/netxen_nic.h
index bb4aa4f..ae81f70 100644
--- a/drivers/net/netxen/netxen_nic.h
+++ b/drivers/net/netxen/netxen_nic.h
@@ -365,6 +365,7 @@ struct rcv_desc {
#define NETXEN_NIC_RXPKT_DESC 0x04
#define NETXEN_OLD_RXPKT_DESC 0x3f
#define NETXEN_NIC_RESPONSE_DESC 0x05
+#define NETXEN_NIC_LRO_DESC 0x12
/* for status field in status_desc */
#define STATUS_NEED_CKSUM (1)
@@ -398,6 +399,24 @@ struct rcv_desc {
#define netxen_get_sts_opcode(sts_data) \
(((sts_data) >> 58) & 0x03F)
+#define netxen_get_lro_sts_refhandle(sts_data) \
+ ((sts_data) & 0x0FFFF)
+#define netxen_get_lro_sts_length(sts_data) \
+ (((sts_data) >> 16) & 0x0FFFF)
+#define netxen_get_lro_sts_l2_hdr_offset(sts_data) \
+ (((sts_data) >> 32) & 0x0FF)
+#define netxen_get_lro_sts_l4_hdr_offset(sts_data) \
+ (((sts_data) >> 40) & 0x0FF)
+#define netxen_get_lro_sts_timestamp(sts_data) \
+ (((sts_data) >> 48) & 0x1)
+#define netxen_get_lro_sts_type(sts_data) \
+ (((sts_data) >> 49) & 0x7)
+#define netxen_get_lro_sts_push_flag(sts_data) \
+ (((sts_data) >> 52) & 0x1)
+#define netxen_get_lro_sts_seq_number(sts_data) \
+ ((sts_data) & 0x0FFFFFFFF)
+
+
struct status_desc {
__le64 status_desc_data[2];
} __attribute__ ((aligned(16)));
@@ -712,6 +731,7 @@ struct netxen_recv_context {
#define NX_CAP0_LSO NX_CAP_BIT(0, 6)
#define NX_CAP0_JUMBO_CONTIGUOUS NX_CAP_BIT(0, 7)
#define NX_CAP0_LRO_CONTIGUOUS NX_CAP_BIT(0, 8)
+#define NX_CAP0_HW_LRO NX_CAP_BIT(0, 10)
/*
* Context state
@@ -969,6 +989,7 @@ typedef struct {
#define NX_FW_CAPABILITY_PEXQ (1 << 7)
#define NX_FW_CAPABILITY_BDG (1 << 8)
#define NX_FW_CAPABILITY_FVLANTX (1 << 9)
+#define NX_FW_CAPABILITY_HW_LRO (1 << 10)
/* module types */
#define LINKEVENT_MODULE_NOT_PRESENT 1
diff --git a/drivers/net/netxen/netxen_nic_ctx.c b/drivers/net/netxen/netxen_nic_ctx.c
index 9e04696..412d658 100644
--- a/drivers/net/netxen/netxen_nic_ctx.c
+++ b/drivers/net/netxen/netxen_nic_ctx.c
@@ -203,6 +203,8 @@ nx_fw_cmd_create_rx_ctx(struct netxen_adapter *adapter)
cap = (NX_CAP0_LEGACY_CONTEXT | NX_CAP0_LEGACY_MN);
cap |= (NX_CAP0_JUMBO_CONTIGUOUS | NX_CAP0_LRO_CONTIGUOUS);
+ if (adapter->capabilities & NX_FW_CAPABILITY_HW_LRO)
+ cap |= NX_CAP0_HW_LRO;
prq->capabilities[0] = cpu_to_le32(cap);
prq->host_int_crb_mode =
diff --git a/drivers/net/netxen/netxen_nic_init.c b/drivers/net/netxen/netxen_nic_init.c
index 81253ab..5828287 100644
--- a/drivers/net/netxen/netxen_nic_init.c
+++ b/drivers/net/netxen/netxen_nic_init.c
@@ -1242,20 +1242,31 @@ no_skb:
static struct netxen_rx_buffer *
netxen_process_rcv(struct netxen_adapter *adapter,
- int ring, int index, int length, int cksum, int pkt_offset,
- struct nx_host_sds_ring *sds_ring)
+ struct nx_host_sds_ring *sds_ring,
+ int ring, u64 sts_data0)
{
struct net_device *netdev = adapter->netdev;
struct netxen_recv_context *recv_ctx = &adapter->recv_ctx;
struct netxen_rx_buffer *buffer;
struct sk_buff *skb;
- struct nx_host_rds_ring *rds_ring = &recv_ctx->rds_rings[ring];
+ struct nx_host_rds_ring *rds_ring;
+ int index, length, cksum, pkt_offset;
- if (unlikely(index > rds_ring->num_desc))
+ if (unlikely(ring >= adapter->max_rds_rings))
+ return NULL;
+
+ rds_ring = &recv_ctx->rds_rings[ring];
+
+ index = netxen_get_sts_refhandle(sts_data0);
+ if (unlikely(index >= rds_ring->num_desc))
return NULL;
buffer = &rds_ring->rx_buf_arr[index];
+ length = netxen_get_sts_totallength(sts_data0);
+ cksum = netxen_get_sts_status(sts_data0);
+ pkt_offset = netxen_get_sts_pkt_offset(sts_data0);
+
skb = netxen_process_rxbuf(adapter, rds_ring, index, cksum);
if (!skb)
return buffer;
@@ -1279,6 +1290,78 @@ netxen_process_rcv(struct netxen_adapter *adapter,
return buffer;
}
+#define TCP_HDR_SIZE 20
+#define TCP_TS_OPTION_SIZE 12
+#define TCP_TS_HDR_SIZE (TCP_HDR_SIZE + TCP_TS_OPTION_SIZE)
+
+static struct netxen_rx_buffer *
+netxen_process_lro(struct netxen_adapter *adapter,
+ struct nx_host_sds_ring *sds_ring,
+ int ring, u64 sts_data0, u64 sts_data1)
+{
+ struct net_device *netdev = adapter->netdev;
+ struct netxen_recv_context *recv_ctx = &adapter->recv_ctx;
+ struct netxen_rx_buffer *buffer;
+ struct sk_buff *skb;
+ struct nx_host_rds_ring *rds_ring;
+ struct iphdr *iph;
+ struct tcphdr *th;
+ bool push, timestamp;
+ int l2_hdr_offset, l4_hdr_offset;
+ int index;
+ u16 lro_length, length, data_offset;
+ u32 seq_number;
+
+ if (unlikely(ring > adapter->max_rds_rings))
+ return NULL;
+
+ rds_ring = &recv_ctx->rds_rings[ring];
+
+ index = netxen_get_lro_sts_refhandle(sts_data0);
+ if (unlikely(index > rds_ring->num_desc))
+ return NULL;
+
+ buffer = &rds_ring->rx_buf_arr[index];
+
+ timestamp = netxen_get_lro_sts_timestamp(sts_data0);
+ lro_length = netxen_get_lro_sts_length(sts_data0);
+ l2_hdr_offset = netxen_get_lro_sts_l2_hdr_offset(sts_data0);
+ l4_hdr_offset = netxen_get_lro_sts_l4_hdr_offset(sts_data0);
+ push = netxen_get_lro_sts_push_flag(sts_data0);
+ seq_number = netxen_get_lro_sts_seq_number(sts_data1);
+
+ skb = netxen_process_rxbuf(adapter, rds_ring, index, STATUS_CKSUM_OK);
+ if (!skb)
+ return buffer;
+
+ if (timestamp)
+ data_offset = l4_hdr_offset + TCP_TS_HDR_SIZE;
+ else
+ data_offset = l4_hdr_offset + TCP_HDR_SIZE;
+
+ skb_put(skb, lro_length + data_offset);
+
+ skb->truesize = (skb->len + sizeof(struct sk_buff) +
+ ((unsigned long)skb->data - (unsigned long)skb->head));
+
+ skb_pull(skb, l2_hdr_offset);
+ skb->protocol = eth_type_trans(skb, netdev);
+
+ iph = (struct iphdr *)skb->data;
+ th = (struct tcphdr *)(skb->data + (iph->ihl << 2));
+
+ length = (iph->ihl << 2) + (th->doff << 2) + lro_length;
+ iph->tot_len = htons(length);
+ iph->check = 0;
+ iph->check = ip_fast_csum((unsigned char *)iph, iph->ihl);
+ th->psh = push;
+ th->seq = htonl(seq_number);
+
+ netif_receive_skb(skb);
+
+ return buffer;
+}
+
#define netxen_merge_rx_buffers(list, head) \
do { list_splice_tail_init(list, head); } while (0);
@@ -1295,28 +1378,33 @@ netxen_process_rcv_ring(struct nx_host_sds_ring *sds_ring, int max)
u32 consumer = sds_ring->consumer;
int count = 0;
- u64 sts_data;
- int opcode, ring, index, length, cksum, pkt_offset, desc_cnt;
+ u64 sts_data0, sts_data1;
+ int opcode, ring = 0, desc_cnt;
while (count < max) {
desc = &sds_ring->desc_head[consumer];
- sts_data = le64_to_cpu(desc->status_desc_data[0]);
+ sts_data0 = le64_to_cpu(desc->status_desc_data[0]);
- if (!(sts_data & STATUS_OWNER_HOST))
+ if (!(sts_data0 & STATUS_OWNER_HOST))
break;
- desc_cnt = netxen_get_sts_desc_cnt(sts_data);
- ring = netxen_get_sts_type(sts_data);
+ desc_cnt = netxen_get_sts_desc_cnt(sts_data0);
- if (ring > RCV_RING_JUMBO)
- goto skip;
-
- opcode = netxen_get_sts_opcode(sts_data);
+ opcode = netxen_get_sts_opcode(sts_data0);
switch (opcode) {
case NETXEN_NIC_RXPKT_DESC:
case NETXEN_OLD_RXPKT_DESC:
case NETXEN_NIC_SYN_OFFLOAD:
+ ring = netxen_get_sts_type(sts_data0);
+ rxbuf = netxen_process_rcv(adapter, sds_ring,
+ ring, sts_data0);
+ break;
+ case NETXEN_NIC_LRO_DESC:
+ ring = netxen_get_lro_sts_type(sts_data0);
+ sts_data1 = le64_to_cpu(desc->status_desc_data[1]);
+ rxbuf = netxen_process_lro(adapter, sds_ring,
+ ring, sts_data0, sts_data1);
break;
case NETXEN_NIC_RESPONSE_DESC:
netxen_handle_fw_message(desc_cnt, consumer, sds_ring);
@@ -1326,14 +1414,6 @@ netxen_process_rcv_ring(struct nx_host_sds_ring *sds_ring, int max)
WARN_ON(desc_cnt > 1);
- index = netxen_get_sts_refhandle(sts_data);
- length = netxen_get_sts_totallength(sts_data);
- cksum = netxen_get_sts_status(sts_data);
- pkt_offset = netxen_get_sts_pkt_offset(sts_data);
-
- rxbuf = netxen_process_rcv(adapter, ring, index,
- length, cksum, pkt_offset, sds_ring);
-
if (rxbuf)
list_add_tail(&rxbuf->list, &sds_ring->free_list[ring]);
--
1.6.0.2
^ permalink raw reply related
* [PATCH NEXT 2/3] netxen: remove unnecessary structures
From: Dhananjay Phadke @ 2009-08-05 17:34 UTC (permalink / raw)
To: davem; +Cc: netdev
In-Reply-To: <1249493649-26336-1-git-send-email-dhananjay@netxen.com>
Remove unnecessary offsetof calulations on these structures:
netxen_board_info, netxen_user_old_info, netxen_new_user_info.
The offsets into the flash are fixed, don't need to be calculated.
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
---
drivers/net/netxen/netxen_nic.h | 167 ++----------------------------------
drivers/net/netxen/netxen_nic_hw.c | 27 ++----
2 files changed, 18 insertions(+), 176 deletions(-)
diff --git a/drivers/net/netxen/netxen_nic.h b/drivers/net/netxen/netxen_nic.h
index e22d086..bb4aa4f 100644
--- a/drivers/net/netxen/netxen_nic.h
+++ b/drivers/net/netxen/netxen_nic.h
@@ -441,154 +441,6 @@ struct status_desc {
#define NETXEN_BRDTYPE_P3_10G_XFP 0x0032
#define NETXEN_BRDTYPE_P3_10G_TP 0x0080
-struct netxen_board_info {
- u32 header_version;
-
- u32 board_mfg;
- u32 board_type;
- u32 board_num;
- u32 chip_id;
- u32 chip_minor;
- u32 chip_major;
- u32 chip_pkg;
- u32 chip_lot;
-
- u32 port_mask; /* available niu ports */
- u32 peg_mask; /* available pegs */
- u32 icache_ok; /* can we run with icache? */
- u32 dcache_ok; /* can we run with dcache? */
- u32 casper_ok;
-
- u32 mac_addr_lo_0;
- u32 mac_addr_lo_1;
- u32 mac_addr_lo_2;
- u32 mac_addr_lo_3;
-
- /* MN-related config */
- u32 mn_sync_mode; /* enable/ sync shift cclk/ sync shift mclk */
- u32 mn_sync_shift_cclk;
- u32 mn_sync_shift_mclk;
- u32 mn_wb_en;
- u32 mn_crystal_freq; /* in MHz */
- u32 mn_speed; /* in MHz */
- u32 mn_org;
- u32 mn_depth;
- u32 mn_ranks_0; /* ranks per slot */
- u32 mn_ranks_1; /* ranks per slot */
- u32 mn_rd_latency_0;
- u32 mn_rd_latency_1;
- u32 mn_rd_latency_2;
- u32 mn_rd_latency_3;
- u32 mn_rd_latency_4;
- u32 mn_rd_latency_5;
- u32 mn_rd_latency_6;
- u32 mn_rd_latency_7;
- u32 mn_rd_latency_8;
- u32 mn_dll_val[18];
- u32 mn_mode_reg; /* MIU DDR Mode Register */
- u32 mn_ext_mode_reg; /* MIU DDR Extended Mode Register */
- u32 mn_timing_0; /* MIU Memory Control Timing Rgister */
- u32 mn_timing_1; /* MIU Extended Memory Ctrl Timing Register */
- u32 mn_timing_2; /* MIU Extended Memory Ctrl Timing2 Register */
-
- /* SN-related config */
- u32 sn_sync_mode; /* enable/ sync shift cclk / sync shift mclk */
- u32 sn_pt_mode; /* pass through mode */
- u32 sn_ecc_en;
- u32 sn_wb_en;
- u32 sn_crystal_freq;
- u32 sn_speed;
- u32 sn_org;
- u32 sn_depth;
- u32 sn_dll_tap;
- u32 sn_rd_latency;
-
- u32 mac_addr_hi_0;
- u32 mac_addr_hi_1;
- u32 mac_addr_hi_2;
- u32 mac_addr_hi_3;
-
- u32 magic; /* indicates flash has been initialized */
-
- u32 mn_rdimm;
- u32 mn_dll_override;
-
-};
-
-#define FLASH_NUM_PORTS (4)
-
-struct netxen_flash_mac_addr {
- u32 flash_addr[32];
-};
-
-struct netxen_user_old_info {
- u8 flash_md5[16];
- u8 crbinit_md5[16];
- u8 brdcfg_md5[16];
- /* bootloader */
- u32 bootld_version;
- u32 bootld_size;
- u8 bootld_md5[16];
- /* image */
- u32 image_version;
- u32 image_size;
- u8 image_md5[16];
- /* primary image status */
- u32 primary_status;
- u32 secondary_present;
-
- /* MAC address , 4 ports */
- struct netxen_flash_mac_addr mac_addr[FLASH_NUM_PORTS];
-};
-#define FLASH_NUM_MAC_PER_PORT 32
-struct netxen_user_info {
- u8 flash_md5[16 * 64];
- /* bootloader */
- u32 bootld_version;
- u32 bootld_size;
- /* image */
- u32 image_version;
- u32 image_size;
- /* primary image status */
- u32 primary_status;
- u32 secondary_present;
-
- /* MAC address , 4 ports, 32 address per port */
- u64 mac_addr[FLASH_NUM_PORTS * FLASH_NUM_MAC_PER_PORT];
- u32 sub_sys_id;
- u8 serial_num[32];
-
- /* Any user defined data */
-};
-
-/*
- * Flash Layout - new format.
- */
-struct netxen_new_user_info {
- u8 flash_md5[16 * 64];
- /* bootloader */
- u32 bootld_version;
- u32 bootld_size;
- /* image */
- u32 image_version;
- u32 image_size;
- /* primary image status */
- u32 primary_status;
- u32 secondary_present;
-
- /* MAC address , 4 ports, 32 address per port */
- u64 mac_addr[FLASH_NUM_PORTS * FLASH_NUM_MAC_PER_PORT];
- u32 sub_sys_id;
- u8 serial_num[32];
-
- /* Any user defined data */
-};
-
-#define SECONDARY_IMAGE_PRESENT 0xb3b4b5b6
-#define SECONDARY_IMAGE_ABSENT 0xffffffff
-#define PRIMARY_IMAGE_GOOD 0x5a5a5a5a
-#define PRIMARY_IMAGE_BAD 0xffffffff
-
/* Flash memory map */
#define NETXEN_CRBINIT_START 0 /* crbinit section */
#define NETXEN_BRDCFG_START 0x4000 /* board config */
@@ -599,28 +451,25 @@ struct netxen_new_user_info {
#define NETXEN_PXE_START 0x3E0000 /* PXE boot rom */
#define NETXEN_USER_START 0x3E8000 /* Firmare info */
#define NETXEN_FIXED_START 0x3F0000 /* backup of crbinit */
+#define NETXEN_USER_START_OLD NETXEN_PXE_START /* very old flash */
+#define NX_OLD_MAC_ADDR_OFFSET (NETXEN_USER_START)
#define NX_FW_VERSION_OFFSET (NETXEN_USER_START+0x408)
#define NX_FW_SIZE_OFFSET (NETXEN_USER_START+0x40c)
+#define NX_FW_MAC_ADDR_OFFSET (NETXEN_USER_START+0x418)
+#define NX_FW_SERIAL_NUM_OFFSET (NETXEN_USER_START+0x81c)
#define NX_BIOS_VERSION_OFFSET (NETXEN_USER_START+0x83c)
+
+#define NX_HDR_VERSION_OFFSET (NETXEN_BRDCFG_START)
+#define NX_BRDTYPE_OFFSET (NETXEN_BRDCFG_START+0x8)
#define NX_FW_MAGIC_OFFSET (NETXEN_BRDCFG_START+0x128)
+
#define NX_FW_MIN_SIZE (0x3fffff)
#define NX_P2_MN_ROMIMAGE 0
#define NX_P3_CT_ROMIMAGE 1
#define NX_P3_MN_ROMIMAGE 2
#define NX_FLASH_ROMIMAGE 3
-#define NETXEN_USER_START_OLD NETXEN_PXE_START /* for backward compatibility */
-
-#define NETXEN_FLASH_START (NETXEN_CRBINIT_START)
-#define NETXEN_INIT_SECTOR (0)
-#define NETXEN_PRIMARY_START (NETXEN_BOOTLD_START)
-#define NETXEN_FLASH_CRBINIT_SIZE (0x4000)
-#define NETXEN_FLASH_BRDCFG_SIZE (sizeof(struct netxen_board_info))
-#define NETXEN_FLASH_USER_SIZE (sizeof(struct netxen_user_info)/sizeof(u32))
-#define NETXEN_FLASH_SECONDARY_SIZE (NETXEN_USER_START-NETXEN_SECONDARY_START)
-#define NETXEN_NUM_PRIMARY_SECTORS (0x20)
-#define NETXEN_NUM_CONFIG_SECTORS (1)
extern char netxen_nic_driver_name[];
/* Number of status descriptors to handle per interrupt */
diff --git a/drivers/net/netxen/netxen_nic_hw.c b/drivers/net/netxen/netxen_nic_hw.c
index ddb9deb..673dcf5 100644
--- a/drivers/net/netxen/netxen_nic_hw.c
+++ b/drivers/net/netxen/netxen_nic_hw.c
@@ -816,18 +816,15 @@ int netxen_get_flash_mac_addr(struct netxen_adapter *adapter, __le64 *mac)
__le32 *pmac = (__le32 *) mac;
u32 offset;
- offset = NETXEN_USER_START +
- offsetof(struct netxen_new_user_info, mac_addr) +
- adapter->portnum * sizeof(u64);
+ offset = NX_FW_MAC_ADDR_OFFSET + (adapter->portnum * sizeof(u64));
if (netxen_get_flash_block(adapter, offset, sizeof(u64), pmac) == -1)
return -1;
if (*mac == cpu_to_le64(~0ULL)) {
- offset = NETXEN_USER_START_OLD +
- offsetof(struct netxen_user_old_info, mac_addr) +
- adapter->portnum * sizeof(u64);
+ offset = NX_OLD_MAC_ADDR_OFFSET +
+ (adapter->portnum * sizeof(u64));
if (netxen_get_flash_block(adapter,
offset, sizeof(u64), pmac) == -1)
@@ -1857,13 +1854,11 @@ int netxen_nic_get_board_info(struct netxen_adapter *adapter)
int offset, board_type, magic, header_version;
struct pci_dev *pdev = adapter->pdev;
- offset = NETXEN_BRDCFG_START +
- offsetof(struct netxen_board_info, magic);
+ offset = NX_FW_MAGIC_OFFSET;
if (netxen_rom_fast_read(adapter, offset, &magic))
return -EIO;
- offset = NETXEN_BRDCFG_START +
- offsetof(struct netxen_board_info, header_version);
+ offset = NX_HDR_VERSION_OFFSET;
if (netxen_rom_fast_read(adapter, offset, &header_version))
return -EIO;
@@ -1875,8 +1870,7 @@ int netxen_nic_get_board_info(struct netxen_adapter *adapter)
return -EIO;
}
- offset = NETXEN_BRDCFG_START +
- offsetof(struct netxen_board_info, board_type);
+ offset = NX_BRDTYPE_OFFSET;
if (netxen_rom_fast_read(adapter, offset, &board_type))
return -EIO;
@@ -2022,23 +2016,22 @@ void netxen_nic_get_firmware_info(struct netxen_adapter *adapter)
u32 fw_major, fw_minor, fw_build;
char brd_name[NETXEN_MAX_SHORT_NAME];
char serial_num[32];
- int i, addr, val;
+ int i, offset, val;
int *ptr32;
struct pci_dev *pdev = adapter->pdev;
adapter->driver_mismatch = 0;
ptr32 = (int *)&serial_num;
- addr = NETXEN_USER_START +
- offsetof(struct netxen_new_user_info, serial_num);
+ offset = NX_FW_SERIAL_NUM_OFFSET;
for (i = 0; i < 8; i++) {
- if (netxen_rom_fast_read(adapter, addr, &val) == -1) {
+ if (netxen_rom_fast_read(adapter, offset, &val) == -1) {
dev_err(&pdev->dev, "error reading board info\n");
adapter->driver_mismatch = 1;
return;
}
ptr32[i] = cpu_to_le32(val);
- addr += sizeof(u32);
+ offset += sizeof(u32);
}
fw_major = NXRD32(adapter, NETXEN_FW_VERSION_MAJOR);
--
1.6.0.2
^ permalink raw reply related
* [PATCH NEXT 1/3] netxen: fix vlan tso case
From: Dhananjay Phadke @ 2009-08-05 17:34 UTC (permalink / raw)
To: davem; +Cc: netdev
Fix the calculation of remaining header length in TSO
over vlan device case. This was inadvertently missed
out in patch 028afe719855a157e32450c ("netxen: add vlan
tx acceleration support").
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
---
drivers/net/netxen/netxen_nic_main.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/netxen/netxen_nic_main.c b/drivers/net/netxen/netxen_nic_main.c
index 40549a0..c91e022 100644
--- a/drivers/net/netxen/netxen_nic_main.c
+++ b/drivers/net/netxen/netxen_nic_main.c
@@ -1450,7 +1450,7 @@ netxen_tso_check(struct net_device *netdev,
skb_copy_from_linear_data_offset(skb, 12,
(char *)vh + 16, copy_len - 16);
- copied = copy_len;
+ copied = copy_len - VLAN_HLEN;
offset = 0;
producer = get_next_index(producer, tx_ring->num_desc);
--
1.6.0.2
^ permalink raw reply related
* Re: [PATCH] dccp: missing destroy of percpu counter variable while unload module
From: David Miller @ 2009-08-05 17:22 UTC (permalink / raw)
To: eric.dumazet; +Cc: yjwei, gerrit, dccp, netdev
In-Reply-To: <4A793BE8.1060101@gmail.com>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Wed, 05 Aug 2009 09:59:36 +0200
> Wei Yongjun a écrit :
>> Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
...
> Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
>
> This was added by commit dd24c001 (net: Use a percpu_counter for orphan_count)
> (Tue Nov 25 21:17:14 2008), so this could be a stable candidate (2.6.29 & 2.6.30)
Applied and queued up for -stable, thanks.
^ permalink raw reply
* Re: [PATCH] netxen: fix dma mask update calculation
From: David Miller @ 2009-08-05 17:22 UTC (permalink / raw)
To: dhananjay; +Cc: davem, netdev
In-Reply-To: <1249418343-15741-1-git-send-email-dhananjay@netxen.com>
From: Dhananjay Phadke <dhananjay@netxen.com>
Date: Tue, 4 Aug 2009 13:39:03 -0700
> Fix dma mask calculation that caps at 63-bit addressing even
> when firmware advertises full 64-bit support.
>
> Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Applied.
^ permalink raw reply
* Re: [PATCH] net: Fix spinlock use in alloc_netdev_mq()
From: Andrew Morton @ 2009-08-05 17:13 UTC (permalink / raw)
To: Jiri Pirko
Cc: Ingo Molnar, David Miller, Peter Zijlstra, torvalds, netdev,
linux-kernel
In-Reply-To: <20090805084746.GA3897@psychotron.englab.brq.redhat.com>
On Wed, 5 Aug 2009 10:47:47 +0200 Jiri Pirko <jpirko@redhat.com> wrote:
> >it's using an zero-initialized spinlock. This is a side-effect of:
> >
> > dev_unicast_init(dev);
> >
> >in alloc_netdev_mq() making use of dev->addr_list_lock.
> >
> >The device has just been allocated freshly, it's not accessible
> >anywhere yet so no locking is needed at all - in fact it's wrong
> >to lock it here (the lock isnt initialized yet).
>
> Yes this looks like the right approach. Sorry for this bug :(
Really?
> >--- a/net/core/dev.c
> >+++ b/net/core/dev.c
> >@@ -4007,9 +4007,7 @@ static void dev_unicast_flush(struct net_device *dev)
> >
> > static void dev_unicast_init(struct net_device *dev)
> > {
> >- netif_addr_lock_bh(dev);
> > __hw_addr_init(&dev->uc);
> >- netif_addr_unlock_bh(dev);
> > }
This means that the net_device is still floating around for quite a
long time with an uninitialised spinlock, so it will be quite easy for
the same problem to reoccur as the code evolves.
It would be more robust were we to initialise that lock close to the
netdev's allocation site.
^ permalink raw reply
* Re: [PATCH 1/1] veth: don't free priv->status until dev->destructor (v2)
From: Stephen Hemminger @ 2009-08-05 17:10 UTC (permalink / raw)
To: Eric W. Biederman
Cc: Serge E. Hallyn, Linux Containers, Sachin Sant, netdev,
David Miller, matthltc, lkml
In-Reply-To: <m1zlaeu5jk.fsf@fess.ebiederm.org>
On Tue, 04 Aug 2009 23:40:47 -0700
ebiederm@xmission.com (Eric W. Biederman) wrote:
> Stephen Hemminger <shemminger@vyatta.com> writes:
>
> > On Fri, 26 Jun 2009 11:24:18 -0500
> > "Serge E. Hallyn" <serue@us.ibm.com> wrote:
> >
> >> Based on the commit msg on ae0e8e82205c903978a79ebf5e31c670b61fa5b4, it looks
> >
> >> ether_setup(dev);
> >> @@ -306,7 +320,7 @@ static void veth_setup(struct net_device *dev)
> >> dev->netdev_ops = &veth_netdev_ops;
> >> dev->ethtool_ops = &veth_ethtool_ops;
> >> dev->features |= NETIF_F_LLTX;
> >> - dev->destructor = free_netdev;
> >> + dev->destructor = veth_dev_free;
> >>
> >
> > This is still going to oops if sysfs statistics referenced
> > after module unload because module is unloaded (code is gone)
> > and veth_dev_free no longer exists.
>
> Has anyone actually seen that cause an oops?
>
> The reason I am asking is that as I read the code we cannot have
> this problem. At worst the destructor callback is delayed until:
>
> veth_exit
> rtnl_link_unregister
> rtnl_unlock
> netdev_run_todo
> dev->destructor
>
>
> Similarly even if the sysfs filehandle is open we have called:
>
> netdev_unregister_kobject
> ...
> sysfs_addrm_finish
> sysfs_deactivate
>
> Which guarantees that sysfs_get_active_two will fail and all
> subsequent actions on that file will fail.
>
> Eric
Sysfs must be safer than it used to be.
^ permalink raw reply
* Re: [PATCH]bluetooth: rfcomm_init bug fix
From: Marcel Holtmann @ 2009-08-05 17:03 UTC (permalink / raw)
To: David Miller
Cc: hidave.darkstar-Re5JQEeQqe8AvxtiuMwx3w,
oliver-fJ+pQTUTwRTk1uMJSBkQmQ,
alan-qBU/x9rampVanCEyBjwyrvXRex20P6io,
netdev-u79uwXL29TY76Z2rM5mHXA,
linux-bluetooth-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20090805.093124.170466311.davem-fT/PcQaiUtIeIZ0/mPfg9Q@public.gmane.org>
Hi Dave,
> > Lets me make this perfectly clear here. Bluetooth patches are sent to
> > the linux-bluetooth mailing list. There is no need to send them to
> > netdev and/or LKML at the same time. Stop this cross-posting. They don't
> > get more review there anyway. I take care of sending them for net-2.6
> > inclusion if I feel that they are appropriate.
>
> Some of that isn't true, I was reading over his patches as he
> posted them. So cross-posting to netdev is absolutely not a
> waste of time IMHO.
this is great and netdev might make sense from time to time, but for
sure posting them to LKML is just noise. To be honest, if you wanna read
Bluetooth patches before they are acked by me, you could subscribe to
linux-bluetooth.
Regards
Marcel
^ permalink raw reply
* Re: [PATCH, RFC] bonding: prevent outgoing packets on inactive slaves
From: Jay Vosburgh @ 2009-08-05 17:01 UTC (permalink / raw)
To: Jiri Bohac; +Cc: davem, netdev
In-Reply-To: <20090805162429.GD16093@midget.suse.cz>
Jiri Bohac <jbohac@suse.cz> wrote:
>Applications exist that broadcast/multicast packets on all
>devices in the system (e.g. avahi-mdns).
>
>When a device is an inactive slave of a bond in active-backup
>mode, its MAC address may be set identical to other divecies in
>the bond. The broadcast/multicast packets may then confuse
>switches to direct packets to the inactive slave, rather than the
>active one.
>
>This patch makes sure the TX queues on inactive slaves are
>deactivated.
>
>Signed-off-by: Jiri Bohac <jbohac@suse.cz>
I'd love to include this patch; many times I've tracked down
"bonding" problems to some errant dingus confusing the switch, but I
think this patch will break some things, and therefore has to be a NAK.
Specifically, I suspect this will break users of some protocols
that intentionally (and legitimately) bind directly to the slave
underneath bonding, LLDP for one. I'm fairly sure there are such users,
because the inactive slave rx path was changed last year to permit
explicit binds to the inactive slaves to receive packets that normally
would be dropped:
commit 0d7a3681232f545c6a59f77e60f7667673ef0e93
Author: Joe Eykholt <jre@nuovasystems.com>
Date: Wed Jul 2 18:22:01 2008 -0700
net/core: Allow certain receives on inactive slave.
Allow a packet_type that specifies the exact device to receive
even on an inactive bonding slave devices. This is important for some
L2 protocols such as LLDP and FCoE. This can eventually be used
for the bonding special cases as well.
Signed-off-by: Joe Eykholt <jre@nuovasystems.com>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
The fact that they're receiving on the inactive slave suggests
that there may be transmits on the same slave, and a quick read of the
LLDP spec seems to agree. I'm also unsure of exactly how FCoE operates
in this regard (whether it does anything that will break due to this
patch).
Anybody have better information about LLDP or FCoE?
-J
---
-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
^ permalink raw reply
* Re: [PATCH, RFC] bonding: prevent outgoing packets on inactive slaves
From: Eric Dumazet @ 2009-08-05 16:57 UTC (permalink / raw)
To: Jiri Bohac; +Cc: fubar, davem, netdev
In-Reply-To: <20090805162429.GD16093@midget.suse.cz>
Jiri Bohac a écrit :
> Applications exist that broadcast/multicast packets on all
> devices in the system (e.g. avahi-mdns).
>
> When a device is an inactive slave of a bond in active-backup
> mode, its MAC address may be set identical to other divecies in
> the bond. The broadcast/multicast packets may then confuse
> switches to direct packets to the inactive slave, rather than the
> active one.
>
> This patch makes sure the TX queues on inactive slaves are
> deactivated.
>
Wont this break ARP link monitoring ?
> Signed-off-by: Jiri Bohac <jbohac@suse.cz>
>
> --- a/drivers/net/bonding/bonding.h
> +++ b/drivers/net/bonding/bonding.h
> @@ -20,6 +20,7 @@
> #include <linux/if_bonding.h>
> #include <linux/kobject.h>
> #include <linux/in6.h>
> +#include <net/sch_generic.h>
> #include "bond_3ad.h"
> #include "bond_alb.h"
>
> @@ -291,12 +292,21 @@ static inline void bond_set_slave_inactive_flags(struct slave *slave)
> slave->dev->priv_flags |= IFF_SLAVE_INACTIVE;
> if (slave_do_arp_validate(bond, slave))
> slave->dev->priv_flags |= IFF_SLAVE_NEEDARP;
> +
> + /* prevent outgoing frames on inactive slaves from confusing switches */
> + if (bond->params.mode == BOND_MODE_ACTIVEBACKUP)
> + dev_deactivate_tx(slave->dev);
> }
>
> static inline void bond_set_slave_active_flags(struct slave *slave)
> {
> + struct bonding *bond = netdev_priv(slave->dev->master);
> +
> slave->state = BOND_STATE_ACTIVE;
> slave->dev->priv_flags &= ~(IFF_SLAVE_INACTIVE | IFF_SLAVE_NEEDARP);
> +
> + if (bond->params.mode == BOND_MODE_ACTIVEBACKUP)
> + dev_activate_tx(slave->dev);
> }
>
> static inline void bond_set_master_3ad_flags(struct bonding *bond)
> diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
> index 964ffa0..eee006a 100644
> --- a/include/net/sch_generic.h
> +++ b/include/net/sch_generic.h
> @@ -296,6 +296,8 @@ extern void dev_init_scheduler(struct net_device *dev);
> extern void dev_shutdown(struct net_device *dev);
> extern void dev_activate(struct net_device *dev);
> extern void dev_deactivate(struct net_device *dev);
> +extern void dev_activate_tx(struct net_device *dev);
> +extern void dev_deactivate_tx(struct net_device *dev);
> extern void qdisc_reset(struct Qdisc *qdisc);
> extern void qdisc_destroy(struct Qdisc *qdisc);
> extern void qdisc_tree_decrease_qlen(struct Qdisc *qdisc, unsigned int n);
> diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
> index 27d0381..514c73d 100644
> --- a/net/sched/sch_generic.c
> +++ b/net/sched/sch_generic.c
> @@ -628,7 +628,7 @@ static void transition_one_qdisc(struct net_device *dev,
> }
> }
>
> -void dev_activate(struct net_device *dev)
> +void dev_activate_tx(struct net_device *dev)
> {
> int need_watchdog;
>
> @@ -647,13 +647,19 @@ void dev_activate(struct net_device *dev)
>
> need_watchdog = 0;
> netdev_for_each_tx_queue(dev, transition_one_qdisc, &need_watchdog);
> - transition_one_qdisc(dev, &dev->rx_queue, NULL);
>
> if (need_watchdog) {
> dev->trans_start = jiffies;
> dev_watchdog_up(dev);
> }
> }
> +EXPORT_SYMBOL(dev_activate_tx);
> +
> +void dev_activate(struct net_device *dev)
> +{
> + dev_activate_tx(dev);
> + transition_one_qdisc(dev, &dev->rx_queue, NULL);
> +}
>
> static void dev_deactivate_queue(struct net_device *dev,
> struct netdev_queue *dev_queue,
> @@ -703,12 +709,18 @@ static bool some_qdisc_is_busy(struct net_device *dev)
> return false;
> }
>
> -void dev_deactivate(struct net_device *dev)
> +void dev_deactivate_tx(struct net_device *dev)
> {
> netdev_for_each_tx_queue(dev, dev_deactivate_queue, &noop_qdisc);
> - dev_deactivate_queue(dev, &dev->rx_queue, &noop_qdisc);
>
> dev_watchdog_down(dev);
> +}
> +EXPORT_SYMBOL(dev_deactivate_tx);
> +
> +void dev_deactivate(struct net_device *dev)
> +{
> + dev_deactivate_tx(dev);
> + dev_deactivate_queue(dev, &dev->rx_queue, &noop_qdisc);
>
> /* Wait for outstanding qdisc-less dev_queue_xmit calls. */
> synchronize_rcu();
>
>
>
^ permalink raw reply
* Re: [PATCH]bluetooth: rfcomm_init bug fix
From: David Miller @ 2009-08-05 16:31 UTC (permalink / raw)
To: marcel; +Cc: hidave.darkstar, oliver, alan, netdev, linux-bluetooth,
linux-kernel
In-Reply-To: <1249486555.3094.38.camel@localhost.localdomain>
From: Marcel Holtmann <marcel@holtmann.org>
Date: Wed, 05 Aug 2009 08:35:55 -0700
> Lets me make this perfectly clear here. Bluetooth patches are sent to
> the linux-bluetooth mailing list. There is no need to send them to
> netdev and/or LKML at the same time. Stop this cross-posting. They don't
> get more review there anyway. I take care of sending them for net-2.6
> inclusion if I feel that they are appropriate.
Some of that isn't true, I was reading over his patches as he
posted them. So cross-posting to netdev is absolutely not a
waste of time IMHO.
^ 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