* [net-next-2.6 06/14] ixgbe: add support to FCoE DDP in target mode
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Yi Zou, netdev, gospo, bphilips, Kiran Patil, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Yi Zou <yi.zou@intel.com>
Add support to the ndo_fcoe_ddp_target() to allow the Intel 82599 device to
also provide DDP offload capability when the upper FCoE protocol stack is
operating as a target.
Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe.h | 2 +
drivers/net/ixgbe/ixgbe_fcoe.c | 86 ++++++++++++++++++++++++++++++++++------
drivers/net/ixgbe/ixgbe_fcoe.h | 4 ++
drivers/net/ixgbe/ixgbe_main.c | 1 +
4 files changed, 81 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe.h b/drivers/net/ixgbe/ixgbe.h
index b60b81b..1e546fc 100644
--- a/drivers/net/ixgbe/ixgbe.h
+++ b/drivers/net/ixgbe/ixgbe.h
@@ -552,6 +552,8 @@ extern int ixgbe_fcoe_ddp(struct ixgbe_adapter *adapter,
struct sk_buff *skb);
extern int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
struct scatterlist *sgl, unsigned int sgc);
+extern int ixgbe_fcoe_ddp_target(struct net_device *netdev, u16 xid,
+ struct scatterlist *sgl, unsigned int sgc);
extern int ixgbe_fcoe_ddp_put(struct net_device *netdev, u16 xid);
extern int ixgbe_fcoe_enable(struct net_device *netdev);
extern int ixgbe_fcoe_disable(struct net_device *netdev);
diff --git a/drivers/net/ixgbe/ixgbe_fcoe.c b/drivers/net/ixgbe/ixgbe_fcoe.c
index 27203c8..00af15a 100644
--- a/drivers/net/ixgbe/ixgbe_fcoe.c
+++ b/drivers/net/ixgbe/ixgbe_fcoe.c
@@ -135,22 +135,19 @@ out_ddp_put:
return len;
}
+
/**
- * ixgbe_fcoe_ddp_get - called to set up ddp context
+ * ixgbe_fcoe_ddp_setup - called to set up ddp context
* @netdev: the corresponding net_device
* @xid: the exchange id requesting ddp
* @sgl: the scatter-gather list for this request
* @sgc: the number of scatter-gather items
*
- * This is the implementation of net_device_ops.ndo_fcoe_ddp_setup
- * and is expected to be called from ULD, e.g., FCP layer of libfc
- * to set up ddp for the corresponding xid of the given sglist for
- * the corresponding I/O.
- *
* Returns : 1 for success and 0 for no ddp
*/
-int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
- struct scatterlist *sgl, unsigned int sgc)
+static int ixgbe_fcoe_ddp_setup(struct net_device *netdev, u16 xid,
+ struct scatterlist *sgl, unsigned int sgc,
+ int target_mode)
{
struct ixgbe_adapter *adapter;
struct ixgbe_hw *hw;
@@ -164,7 +161,7 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
unsigned int lastsize;
unsigned int thisoff = 0;
unsigned int thislen = 0;
- u32 fcbuff, fcdmarw, fcfltrw;
+ u32 fcbuff, fcdmarw, fcfltrw, fcrxctl;
dma_addr_t addr = 0;
if (!netdev || !sgl)
@@ -275,6 +272,9 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
fcbuff = (IXGBE_FCBUFF_4KB << IXGBE_FCBUFF_BUFFSIZE_SHIFT);
fcbuff |= ((j & 0xff) << IXGBE_FCBUFF_BUFFCNT_SHIFT);
fcbuff |= (firstoff << IXGBE_FCBUFF_OFFSET_SHIFT);
+ /* Set WRCONTX bit to allow DDP for target */
+ if (target_mode)
+ fcbuff |= (IXGBE_FCBUFF_WRCONTX);
fcbuff |= (IXGBE_FCBUFF_VALID);
fcdmarw = xid;
@@ -287,6 +287,16 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
/* program DMA context */
hw = &adapter->hw;
spin_lock_bh(&fcoe->lock);
+
+ /* turn on last frame indication for target mode as FCP_RSPtarget is
+ * supposed to send FCP_RSP when it is done. */
+ if (target_mode && !test_bit(__IXGBE_FCOE_TARGET, &fcoe->mode)) {
+ set_bit(__IXGBE_FCOE_TARGET, &fcoe->mode);
+ fcrxctl = IXGBE_READ_REG(hw, IXGBE_FCRXCTRL);
+ fcrxctl |= IXGBE_FCRXCTRL_LASTSEQH;
+ IXGBE_WRITE_REG(hw, IXGBE_FCRXCTRL, fcrxctl);
+ }
+
IXGBE_WRITE_REG(hw, IXGBE_FCPTRL, ddp->udp & DMA_BIT_MASK(32));
IXGBE_WRITE_REG(hw, IXGBE_FCPTRH, (u64)ddp->udp >> 32);
IXGBE_WRITE_REG(hw, IXGBE_FCBUFF, fcbuff);
@@ -295,6 +305,7 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
IXGBE_WRITE_REG(hw, IXGBE_FCPARAM, 0);
IXGBE_WRITE_REG(hw, IXGBE_FCFLT, IXGBE_FCFLT_VALID);
IXGBE_WRITE_REG(hw, IXGBE_FCFLTRW, fcfltrw);
+
spin_unlock_bh(&fcoe->lock);
return 1;
@@ -309,6 +320,47 @@ out_noddp_unmap:
}
/**
+ * ixgbe_fcoe_ddp_get - called to set up ddp context in initiator mode
+ * @netdev: the corresponding net_device
+ * @xid: the exchange id requesting ddp
+ * @sgl: the scatter-gather list for this request
+ * @sgc: the number of scatter-gather items
+ *
+ * This is the implementation of net_device_ops.ndo_fcoe_ddp_setup
+ * and is expected to be called from ULD, e.g., FCP layer of libfc
+ * to set up ddp for the corresponding xid of the given sglist for
+ * the corresponding I/O.
+ *
+ * Returns : 1 for success and 0 for no ddp
+ */
+int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid,
+ struct scatterlist *sgl, unsigned int sgc)
+{
+ return ixgbe_fcoe_ddp_setup(netdev, xid, sgl, sgc, 0);
+}
+
+/**
+ * ixgbe_fcoe_ddp_target - called to set up ddp context in target mode
+ * @netdev: the corresponding net_device
+ * @xid: the exchange id requesting ddp
+ * @sgl: the scatter-gather list for this request
+ * @sgc: the number of scatter-gather items
+ *
+ * This is the implementation of net_device_ops.ndo_fcoe_ddp_target
+ * and is expected to be called from ULD, e.g., FCP layer of libfc
+ * to set up ddp for the corresponding xid of the given sglist for
+ * the corresponding I/O. The DDP in target mode is a write I/O request
+ * from the initiator.
+ *
+ * Returns : 1 for success and 0 for no ddp
+ */
+int ixgbe_fcoe_ddp_target(struct net_device *netdev, u16 xid,
+ struct scatterlist *sgl, unsigned int sgc)
+{
+ return ixgbe_fcoe_ddp_setup(netdev, xid, sgl, sgc, 1);
+}
+
+/**
* ixgbe_fcoe_ddp - check ddp status and mark it done
* @adapter: ixgbe adapter
* @rx_desc: advanced rx descriptor
@@ -331,6 +383,7 @@ int ixgbe_fcoe_ddp(struct ixgbe_adapter *adapter,
struct ixgbe_fcoe *fcoe;
struct ixgbe_fcoe_ddp *ddp;
struct fc_frame_header *fh;
+ struct fcoe_crc_eof *crc;
if (!ixgbe_rx_is_fcoe(rx_desc))
goto ddp_out;
@@ -384,7 +437,18 @@ int ixgbe_fcoe_ddp(struct ixgbe_adapter *adapter,
else if (ddp->len)
rc = ddp->len;
}
-
+ /* In target mode, check the last data frame of the sequence.
+ * For DDP in target mode, data is already DDPed but the header
+ * indication of the last data frame ould allow is to tell if we
+ * got all the data and the ULP can send FCP_RSP back, as this is
+ * not a full fcoe frame, we fill the trailer here so it won't be
+ * dropped by the ULP stack.
+ */
+ if ((fh->fh_r_ctl == FC_RCTL_DD_SOL_DATA) &&
+ (fctl & FC_FC_END_SEQ)) {
+ crc = (struct fcoe_crc_eof *)skb_put(skb, sizeof(*crc));
+ crc->fcoe_eof = FC_EOF_T;
+ }
ddp_out:
return rc;
}
@@ -840,5 +904,3 @@ int ixgbe_fcoe_get_wwn(struct net_device *netdev, u64 *wwn, int type)
}
return rc;
}
-
-
diff --git a/drivers/net/ixgbe/ixgbe_fcoe.h b/drivers/net/ixgbe/ixgbe_fcoe.h
index 02a00d2..5a650a4 100644
--- a/drivers/net/ixgbe/ixgbe_fcoe.h
+++ b/drivers/net/ixgbe/ixgbe_fcoe.h
@@ -52,6 +52,9 @@
/* fcerr */
#define IXGBE_FCERR_BADCRC 0x00100000
+/* FCoE DDP for target mode */
+#define __IXGBE_FCOE_TARGET 1
+
struct ixgbe_fcoe_ddp {
int len;
u32 err;
@@ -66,6 +69,7 @@ struct ixgbe_fcoe {
u8 tc;
u8 up;
#endif
+ unsigned long mode;
atomic_t refcnt;
spinlock_t lock;
struct pci_pool *pool;
diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c
index 32231ff..5e8c39d 100644
--- a/drivers/net/ixgbe/ixgbe_main.c
+++ b/drivers/net/ixgbe/ixgbe_main.c
@@ -7019,6 +7019,7 @@ static const struct net_device_ops ixgbe_netdev_ops = {
#endif
#ifdef IXGBE_FCOE
.ndo_fcoe_ddp_setup = ixgbe_fcoe_ddp_get,
+ .ndo_fcoe_ddp_target = ixgbe_fcoe_ddp_target,
.ndo_fcoe_ddp_done = ixgbe_fcoe_ddp_put,
.ndo_fcoe_enable = ixgbe_fcoe_enable,
.ndo_fcoe_disable = ixgbe_fcoe_disable,
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 08/14] ixgbe: clear correct counters for flow control on 82599
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
The 82599 was not correctly having some of it's counters cleared for flow
control. This change corrects that.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_common.c | 34 +++++++++++++++++++++++++++-------
drivers/net/ixgbe/ixgbe_type.h | 2 ++
2 files changed, 29 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index 94a5621..85cc301 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -136,17 +136,29 @@ s32 ixgbe_clear_hw_cntrs_generic(struct ixgbe_hw *hw)
IXGBE_READ_REG(hw, IXGBE_MRFC);
IXGBE_READ_REG(hw, IXGBE_RLEC);
IXGBE_READ_REG(hw, IXGBE_LXONTXC);
- IXGBE_READ_REG(hw, IXGBE_LXONRXC);
IXGBE_READ_REG(hw, IXGBE_LXOFFTXC);
- IXGBE_READ_REG(hw, IXGBE_LXOFFRXC);
+ if (hw->mac.type >= ixgbe_mac_82599EB) {
+ IXGBE_READ_REG(hw, IXGBE_LXONRXCNT);
+ IXGBE_READ_REG(hw, IXGBE_LXOFFRXCNT);
+ } else {
+ IXGBE_READ_REG(hw, IXGBE_LXONRXC);
+ IXGBE_READ_REG(hw, IXGBE_LXOFFRXC);
+ }
for (i = 0; i < 8; i++) {
IXGBE_READ_REG(hw, IXGBE_PXONTXC(i));
- IXGBE_READ_REG(hw, IXGBE_PXONRXC(i));
IXGBE_READ_REG(hw, IXGBE_PXOFFTXC(i));
- IXGBE_READ_REG(hw, IXGBE_PXOFFRXC(i));
+ if (hw->mac.type >= ixgbe_mac_82599EB) {
+ IXGBE_READ_REG(hw, IXGBE_PXONRXCNT(i));
+ IXGBE_READ_REG(hw, IXGBE_PXOFFRXCNT(i));
+ } else {
+ IXGBE_READ_REG(hw, IXGBE_PXONRXC(i));
+ IXGBE_READ_REG(hw, IXGBE_PXOFFRXC(i));
+ }
}
-
+ if (hw->mac.type >= ixgbe_mac_82599EB)
+ for (i = 0; i < 8; i++)
+ IXGBE_READ_REG(hw, IXGBE_PXON2OFFCNT(i));
IXGBE_READ_REG(hw, IXGBE_PRC64);
IXGBE_READ_REG(hw, IXGBE_PRC127);
IXGBE_READ_REG(hw, IXGBE_PRC255);
@@ -184,9 +196,17 @@ s32 ixgbe_clear_hw_cntrs_generic(struct ixgbe_hw *hw)
IXGBE_READ_REG(hw, IXGBE_BPTC);
for (i = 0; i < 16; i++) {
IXGBE_READ_REG(hw, IXGBE_QPRC(i));
- IXGBE_READ_REG(hw, IXGBE_QBRC(i));
IXGBE_READ_REG(hw, IXGBE_QPTC(i));
- IXGBE_READ_REG(hw, IXGBE_QBTC(i));
+ if (hw->mac.type >= ixgbe_mac_82599EB) {
+ IXGBE_READ_REG(hw, IXGBE_QBRC_L(i));
+ IXGBE_READ_REG(hw, IXGBE_QBRC_H(i));
+ IXGBE_READ_REG(hw, IXGBE_QBTC_L(i));
+ IXGBE_READ_REG(hw, IXGBE_QBTC_H(i));
+ IXGBE_READ_REG(hw, IXGBE_QPRDC(i));
+ } else {
+ IXGBE_READ_REG(hw, IXGBE_QBRC(i));
+ IXGBE_READ_REG(hw, IXGBE_QBTC(i));
+ }
}
return 0;
diff --git a/drivers/net/ixgbe/ixgbe_type.h b/drivers/net/ixgbe/ixgbe_type.h
index 013751d..13a45c7 100644
--- a/drivers/net/ixgbe/ixgbe_type.h
+++ b/drivers/net/ixgbe/ixgbe_type.h
@@ -659,6 +659,8 @@
#define IXGBE_QPTC(_i) (0x06030 + ((_i) * 0x40)) /* 16 of these */
#define IXGBE_QBRC(_i) (0x01034 + ((_i) * 0x40)) /* 16 of these */
#define IXGBE_QBTC(_i) (0x06034 + ((_i) * 0x40)) /* 16 of these */
+#define IXGBE_QBRC_L(_i) (0x01034 + ((_i) * 0x40)) /* 16 of these */
+#define IXGBE_QBRC_H(_i) (0x01038 + ((_i) * 0x40)) /* 16 of these */
#define IXGBE_QPRDC(_i) (0x01430 + ((_i) * 0x40)) /* 16 of these */
#define IXGBE_QBTC_L(_i) (0x08700 + ((_i) * 0x8)) /* 16 of these */
#define IXGBE_QBTC_H(_i) (0x08704 + ((_i) * 0x8)) /* 16 of these */
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 01/14] e1000e: fix build issue due to undefined reference to crc32_le
From: Jeff Kirsher @ 2011-03-08 2:24 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
kernel build fails with:
drivers/built-in.o: In function `e1000_lv_jumbo_workaround_ich8lan':
(.text+0x3e7a8): undefined reference to `crc32_le'
when CONFIG_CRC32 is not set or does not match the CONFIG_E1000E
selection.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Reviewed-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/Kconfig | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 7ce3dd0..925c25c 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -2109,6 +2109,7 @@ config E1000
config E1000E
tristate "Intel(R) PRO/1000 PCI-Express Gigabit Ethernet support"
depends on PCI && (!SPARC32 || BROKEN)
+ select CRC32
---help---
This driver supports the PCI-Express Intel(R) PRO/1000 gigabit
ethernet family of adapters. For PCI or PCI-X e1000 adapters,
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 03/14] ixgbe: fix missing function pointer conversion
From: Jeff Kirsher @ 2011-03-08 2:24 UTC (permalink / raw)
To: davem; +Cc: Don Skidmore, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Don Skidmore <donald.c.skidmore@intel.com>
In the previous commit:
commit 5e655105e3e19d746f9e95c514b014c11c3d1b6a
Author: Don Skidmore <donald.c.skidmore@intel.com>
Date: Fri Feb 25 01:58:04 2011 +0000
ixgbe: add function pointer for semaphore function
there was one release of the semaphore function call which
did not get converted to a function pointer.
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_common.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index a7fb2e0..94a5621 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -1075,7 +1075,7 @@ static void ixgbe_release_eeprom(struct ixgbe_hw *hw)
eec &= ~IXGBE_EEC_REQ;
IXGBE_WRITE_REG(hw, IXGBE_EEC, eec);
- ixgbe_release_swfw_sync(hw, IXGBE_GSSR_EEP_SM);
+ hw->mac.ops.release_swfw_sync(hw, IXGBE_GSSR_EEP_SM);
/* Delay before attempt to obtain semaphore again to allow FW access */
msleep(hw->eeprom.semaphore_delay);
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 04/14] net: add ndo_fcoe_ddp_target() to support FCoE DDP in target mode
From: Jeff Kirsher @ 2011-03-08 2:24 UTC (permalink / raw)
To: davem; +Cc: Yi Zou, netdev, gospo, bphilips, Kiran Patil, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Yi Zou <yi.zou@intel.com>
The Fiber Channel over Ethernet (FCoE) Direct Data Placement (DDP) can also be
used for FCoE target, where the DDP used for read I/O on an initiator can be
used on an FCoE target to speed up the write I/O to the target from the initiator.
The added ndo_fcoe_ddp_target() works in the similar way as the existing
ndo_fcoe_ddp_setup() to allow the underlying hardware set up the DDP context
accordingly when it gets called from the FCoE target implementation on top
the existing Open-FCoE fcoe/libfc protocol stack so without losing the ability
to provide DDP for read I/O as an initiator, it can also provide DDP offload
to the write I/O coming from the initiator as a target.
Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Kiran Patil <kiran.patil@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
include/linux/netdevice.h | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 71563e7..6bd5d46 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -871,6 +871,10 @@ struct net_device_ops {
unsigned int sgc);
int (*ndo_fcoe_ddp_done)(struct net_device *dev,
u16 xid);
+ int (*ndo_fcoe_ddp_target)(struct net_device *dev,
+ u16 xid,
+ struct scatterlist *sgl,
+ unsigned int sgc);
#define NETDEV_FCOE_WWNN 0
#define NETDEV_FCOE_WWPN 1
int (*ndo_fcoe_get_wwn)(struct net_device *dev,
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 10/14] ixgbe: Enable flow control pause parameter auto-negotiation support
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
This patch enables flow control pause parameters auto-negotiation support
to 82599 based 10G Base-T, backplane devices and multi-speed fiber optics
modules at 1G speed
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_82598.c | 7 +-
drivers/net/ixgbe/ixgbe_82599.c | 3 +
drivers/net/ixgbe/ixgbe_common.c | 460 ++++++++++++++++++++++----------------
drivers/net/ixgbe/ixgbe_main.c | 3 +-
drivers/net/ixgbe/ixgbe_phy.h | 4 +
drivers/net/ixgbe/ixgbe_type.h | 3 +
6 files changed, 281 insertions(+), 199 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_82598.c b/drivers/net/ixgbe/ixgbe_82598.c
index dc977b1..ff23907 100644
--- a/drivers/net/ixgbe/ixgbe_82598.c
+++ b/drivers/net/ixgbe/ixgbe_82598.c
@@ -366,7 +366,7 @@ static s32 ixgbe_fc_enable_82598(struct ixgbe_hw *hw, s32 packetbuf_num)
/* Negotiate the fc mode to use */
ret_val = ixgbe_fc_autoneg(hw);
- if (ret_val)
+ if (ret_val == IXGBE_ERR_FLOW_CONTROL)
goto out;
/* Disable any previous flow control settings */
@@ -384,10 +384,10 @@ static s32 ixgbe_fc_enable_82598(struct ixgbe_hw *hw, s32 packetbuf_num)
* 2: Tx flow control is enabled (we can send pause frames but
* we do not support receiving pause frames).
* 3: Both Rx and Tx flow control (symmetric) are enabled.
- * other: Invalid.
#ifdef CONFIG_DCB
* 4: Priority Flow Control is enabled.
#endif
+ * other: Invalid.
*/
switch (hw->fc.current_mode) {
case ixgbe_fc_none:
@@ -444,9 +444,10 @@ static s32 ixgbe_fc_enable_82598(struct ixgbe_hw *hw, s32 packetbuf_num)
reg = (rx_pba_size - hw->fc.low_water) << 6;
if (hw->fc.send_xon)
reg |= IXGBE_FCRTL_XONE;
+
IXGBE_WRITE_REG(hw, IXGBE_FCRTL(packetbuf_num), reg);
- reg = (rx_pba_size - hw->fc.high_water) << 10;
+ reg = (rx_pba_size - hw->fc.high_water) << 6;
reg |= IXGBE_FCRTH_FCEN;
IXGBE_WRITE_REG(hw, IXGBE_FCRTH(packetbuf_num), reg);
diff --git a/drivers/net/ixgbe/ixgbe_82599.c b/drivers/net/ixgbe/ixgbe_82599.c
index 2f8b9f4..00aeba3 100644
--- a/drivers/net/ixgbe/ixgbe_82599.c
+++ b/drivers/net/ixgbe/ixgbe_82599.c
@@ -772,6 +772,9 @@ static s32 ixgbe_setup_mac_link_82599(struct ixgbe_hw *hw,
/* Check to see if speed passed in is supported. */
hw->mac.ops.get_link_capabilities(hw, &link_capabilities, &autoneg);
+ if (status != 0)
+ goto out;
+
speed &= link_capabilities;
if (speed == IXGBE_LINK_SPEED_UNKNOWN) {
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index 561f666..7a6a3fb 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -47,6 +47,12 @@ static void ixgbe_lower_eeprom_clk(struct ixgbe_hw *hw, u32 *eec);
static void ixgbe_release_eeprom(struct ixgbe_hw *hw);
static s32 ixgbe_mta_vector(struct ixgbe_hw *hw, u8 *mc_addr);
+static s32 ixgbe_fc_autoneg_fiber(struct ixgbe_hw *hw);
+static s32 ixgbe_fc_autoneg_backplane(struct ixgbe_hw *hw);
+static s32 ixgbe_fc_autoneg_copper(struct ixgbe_hw *hw);
+static s32 ixgbe_device_supports_autoneg_fc(struct ixgbe_hw *hw);
+static s32 ixgbe_negotiate_fc(struct ixgbe_hw *hw, u32 adv_reg, u32 lp_reg,
+ u32 adv_sym, u32 adv_asm, u32 lp_sym, u32 lp_asm);
static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num);
/**
@@ -1566,7 +1572,7 @@ s32 ixgbe_fc_enable_generic(struct ixgbe_hw *hw, s32 packetbuf_num)
#endif /* CONFIG_DCB */
/* Negotiate the fc mode to use */
ret_val = ixgbe_fc_autoneg(hw);
- if (ret_val)
+ if (ret_val == IXGBE_ERR_FLOW_CONTROL)
goto out;
/* Disable any previous flow control settings */
@@ -1674,12 +1680,13 @@ out:
**/
s32 ixgbe_fc_autoneg(struct ixgbe_hw *hw)
{
- s32 ret_val = 0;
+ s32 ret_val = IXGBE_ERR_FC_NOT_NEGOTIATED;
ixgbe_link_speed speed;
- u32 pcs_anadv_reg, pcs_lpab_reg, linkstat;
- u32 links2, anlp1_reg, autoc_reg, links;
bool link_up;
+ if (hw->fc.disable_fc_autoneg)
+ goto out;
+
/*
* AN should have completed when the cable was plugged in.
* Look for reasons to bail out. Bail out if:
@@ -1690,153 +1697,199 @@ s32 ixgbe_fc_autoneg(struct ixgbe_hw *hw)
* So use link_up_wait_to_complete=false.
*/
hw->mac.ops.check_link(hw, &speed, &link_up, false);
-
- if (hw->fc.disable_fc_autoneg || (!link_up)) {
- hw->fc.fc_was_autonegged = false;
- hw->fc.current_mode = hw->fc.requested_mode;
+ if (!link_up) {
+ ret_val = IXGBE_ERR_FLOW_CONTROL;
goto out;
}
- /*
- * On backplane, bail out if
- * - backplane autoneg was not completed, or if
- * - we are 82599 and link partner is not AN enabled
- */
- if (hw->phy.media_type == ixgbe_media_type_backplane) {
- links = IXGBE_READ_REG(hw, IXGBE_LINKS);
- if ((links & IXGBE_LINKS_KX_AN_COMP) == 0) {
- hw->fc.fc_was_autonegged = false;
- hw->fc.current_mode = hw->fc.requested_mode;
- goto out;
- }
+ switch (hw->phy.media_type) {
+ /* Autoneg flow control on fiber adapters */
+ case ixgbe_media_type_fiber:
+ if (speed == IXGBE_LINK_SPEED_1GB_FULL)
+ ret_val = ixgbe_fc_autoneg_fiber(hw);
+ break;
- if (hw->mac.type == ixgbe_mac_82599EB) {
- links2 = IXGBE_READ_REG(hw, IXGBE_LINKS2);
- if ((links2 & IXGBE_LINKS2_AN_SUPPORTED) == 0) {
- hw->fc.fc_was_autonegged = false;
- hw->fc.current_mode = hw->fc.requested_mode;
- goto out;
- }
- }
+ /* Autoneg flow control on backplane adapters */
+ case ixgbe_media_type_backplane:
+ ret_val = ixgbe_fc_autoneg_backplane(hw);
+ break;
+
+ /* Autoneg flow control on copper adapters */
+ case ixgbe_media_type_copper:
+ if (ixgbe_device_supports_autoneg_fc(hw) == 0)
+ ret_val = ixgbe_fc_autoneg_copper(hw);
+ break;
+
+ default:
+ break;
+ }
+
+out:
+ if (ret_val == 0) {
+ hw->fc.fc_was_autonegged = true;
+ } else {
+ hw->fc.fc_was_autonegged = false;
+ hw->fc.current_mode = hw->fc.requested_mode;
}
+ return ret_val;
+}
+
+/**
+ * ixgbe_fc_autoneg_fiber - Enable flow control on 1 gig fiber
+ * @hw: pointer to hardware structure
+ *
+ * Enable flow control according on 1 gig fiber.
+ **/
+static s32 ixgbe_fc_autoneg_fiber(struct ixgbe_hw *hw)
+{
+ u32 pcs_anadv_reg, pcs_lpab_reg, linkstat;
+ s32 ret_val;
/*
* On multispeed fiber at 1g, bail out if
* - link is up but AN did not complete, or if
* - link is up and AN completed but timed out
*/
- if (hw->phy.multispeed_fiber && (speed == IXGBE_LINK_SPEED_1GB_FULL)) {
- linkstat = IXGBE_READ_REG(hw, IXGBE_PCS1GLSTA);
- if (((linkstat & IXGBE_PCS1GLSTA_AN_COMPLETE) == 0) ||
- ((linkstat & IXGBE_PCS1GLSTA_AN_TIMED_OUT) == 1)) {
- hw->fc.fc_was_autonegged = false;
- hw->fc.current_mode = hw->fc.requested_mode;
- goto out;
- }
+
+ linkstat = IXGBE_READ_REG(hw, IXGBE_PCS1GLSTA);
+ if (((linkstat & IXGBE_PCS1GLSTA_AN_COMPLETE) == 0) ||
+ ((linkstat & IXGBE_PCS1GLSTA_AN_TIMED_OUT) == 1)) {
+ ret_val = IXGBE_ERR_FC_NOT_NEGOTIATED;
+ goto out;
}
+ pcs_anadv_reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANA);
+ pcs_lpab_reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANLP);
+
+ ret_val = ixgbe_negotiate_fc(hw, pcs_anadv_reg,
+ pcs_lpab_reg, IXGBE_PCS1GANA_SYM_PAUSE,
+ IXGBE_PCS1GANA_ASM_PAUSE,
+ IXGBE_PCS1GANA_SYM_PAUSE,
+ IXGBE_PCS1GANA_ASM_PAUSE);
+
+out:
+ return ret_val;
+}
+
+/**
+ * ixgbe_fc_autoneg_backplane - Enable flow control IEEE clause 37
+ * @hw: pointer to hardware structure
+ *
+ * Enable flow control according to IEEE clause 37.
+ **/
+static s32 ixgbe_fc_autoneg_backplane(struct ixgbe_hw *hw)
+{
+ u32 links2, anlp1_reg, autoc_reg, links;
+ s32 ret_val;
+
/*
- * Bail out on
- * - copper or CX4 adapters
- * - fiber adapters running at 10gig
+ * On backplane, bail out if
+ * - backplane autoneg was not completed, or if
+ * - we are 82599 and link partner is not AN enabled
*/
- if ((hw->phy.media_type == ixgbe_media_type_copper) ||
- (hw->phy.media_type == ixgbe_media_type_cx4) ||
- ((hw->phy.media_type == ixgbe_media_type_fiber) &&
- (speed == IXGBE_LINK_SPEED_10GB_FULL))) {
+ links = IXGBE_READ_REG(hw, IXGBE_LINKS);
+ if ((links & IXGBE_LINKS_KX_AN_COMP) == 0) {
hw->fc.fc_was_autonegged = false;
hw->fc.current_mode = hw->fc.requested_mode;
+ ret_val = IXGBE_ERR_FC_NOT_NEGOTIATED;
goto out;
}
+ if (hw->mac.type == ixgbe_mac_82599EB) {
+ links2 = IXGBE_READ_REG(hw, IXGBE_LINKS2);
+ if ((links2 & IXGBE_LINKS2_AN_SUPPORTED) == 0) {
+ hw->fc.fc_was_autonegged = false;
+ hw->fc.current_mode = hw->fc.requested_mode;
+ ret_val = IXGBE_ERR_FC_NOT_NEGOTIATED;
+ goto out;
+ }
+ }
/*
- * Read the AN advertisement and LP ability registers and resolve
+ * Read the 10g AN autoc and LP ability registers and resolve
* local flow control settings accordingly
*/
- if ((speed == IXGBE_LINK_SPEED_1GB_FULL) &&
- (hw->phy.media_type != ixgbe_media_type_backplane)) {
- pcs_anadv_reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANA);
- pcs_lpab_reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANLP);
- if ((pcs_anadv_reg & IXGBE_PCS1GANA_SYM_PAUSE) &&
- (pcs_lpab_reg & IXGBE_PCS1GANA_SYM_PAUSE)) {
- /*
- * Now we need to check if the user selected Rx ONLY
- * of pause frames. In this case, we had to advertise
- * FULL flow control because we could not advertise RX
- * ONLY. Hence, we must now check to see if we need to
- * turn OFF the TRANSMISSION of PAUSE frames.
- */
- if (hw->fc.requested_mode == ixgbe_fc_full) {
- hw->fc.current_mode = ixgbe_fc_full;
- hw_dbg(hw, "Flow Control = FULL.\n");
- } else {
- hw->fc.current_mode = ixgbe_fc_rx_pause;
- hw_dbg(hw, "Flow Control=RX PAUSE only\n");
- }
- } else if (!(pcs_anadv_reg & IXGBE_PCS1GANA_SYM_PAUSE) &&
- (pcs_anadv_reg & IXGBE_PCS1GANA_ASM_PAUSE) &&
- (pcs_lpab_reg & IXGBE_PCS1GANA_SYM_PAUSE) &&
- (pcs_lpab_reg & IXGBE_PCS1GANA_ASM_PAUSE)) {
- hw->fc.current_mode = ixgbe_fc_tx_pause;
- hw_dbg(hw, "Flow Control = TX PAUSE frames only.\n");
- } else if ((pcs_anadv_reg & IXGBE_PCS1GANA_SYM_PAUSE) &&
- (pcs_anadv_reg & IXGBE_PCS1GANA_ASM_PAUSE) &&
- !(pcs_lpab_reg & IXGBE_PCS1GANA_SYM_PAUSE) &&
- (pcs_lpab_reg & IXGBE_PCS1GANA_ASM_PAUSE)) {
- hw->fc.current_mode = ixgbe_fc_rx_pause;
- hw_dbg(hw, "Flow Control = RX PAUSE frames only.\n");
- } else {
- hw->fc.current_mode = ixgbe_fc_none;
- hw_dbg(hw, "Flow Control = NONE.\n");
- }
- }
+ autoc_reg = IXGBE_READ_REG(hw, IXGBE_AUTOC);
+ anlp1_reg = IXGBE_READ_REG(hw, IXGBE_ANLP1);
- if (hw->phy.media_type == ixgbe_media_type_backplane) {
+ ret_val = ixgbe_negotiate_fc(hw, autoc_reg,
+ anlp1_reg, IXGBE_AUTOC_SYM_PAUSE, IXGBE_AUTOC_ASM_PAUSE,
+ IXGBE_ANLP1_SYM_PAUSE, IXGBE_ANLP1_ASM_PAUSE);
+
+out:
+ return ret_val;
+}
+
+/**
+ * ixgbe_fc_autoneg_copper - Enable flow control IEEE clause 37
+ * @hw: pointer to hardware structure
+ *
+ * Enable flow control according to IEEE clause 37.
+ **/
+static s32 ixgbe_fc_autoneg_copper(struct ixgbe_hw *hw)
+{
+ u16 technology_ability_reg = 0;
+ u16 lp_technology_ability_reg = 0;
+
+ hw->phy.ops.read_reg(hw, MDIO_AN_ADVERTISE,
+ MDIO_MMD_AN,
+ &technology_ability_reg);
+ hw->phy.ops.read_reg(hw, MDIO_AN_LPA,
+ MDIO_MMD_AN,
+ &lp_technology_ability_reg);
+
+ return ixgbe_negotiate_fc(hw, (u32)technology_ability_reg,
+ (u32)lp_technology_ability_reg,
+ IXGBE_TAF_SYM_PAUSE, IXGBE_TAF_ASM_PAUSE,
+ IXGBE_TAF_SYM_PAUSE, IXGBE_TAF_ASM_PAUSE);
+}
+
+/**
+ * ixgbe_negotiate_fc - Negotiate flow control
+ * @hw: pointer to hardware structure
+ * @adv_reg: flow control advertised settings
+ * @lp_reg: link partner's flow control settings
+ * @adv_sym: symmetric pause bit in advertisement
+ * @adv_asm: asymmetric pause bit in advertisement
+ * @lp_sym: symmetric pause bit in link partner advertisement
+ * @lp_asm: asymmetric pause bit in link partner advertisement
+ *
+ * Find the intersection between advertised settings and link partner's
+ * advertised settings
+ **/
+static s32 ixgbe_negotiate_fc(struct ixgbe_hw *hw, u32 adv_reg, u32 lp_reg,
+ u32 adv_sym, u32 adv_asm, u32 lp_sym, u32 lp_asm)
+{
+ if ((!(adv_reg)) || (!(lp_reg)))
+ return IXGBE_ERR_FC_NOT_NEGOTIATED;
+
+ if ((adv_reg & adv_sym) && (lp_reg & lp_sym)) {
/*
- * Read the 10g AN autoc and LP ability registers and resolve
- * local flow control settings accordingly
+ * Now we need to check if the user selected Rx ONLY
+ * of pause frames. In this case, we had to advertise
+ * FULL flow control because we could not advertise RX
+ * ONLY. Hence, we must now check to see if we need to
+ * turn OFF the TRANSMISSION of PAUSE frames.
*/
- autoc_reg = IXGBE_READ_REG(hw, IXGBE_AUTOC);
- anlp1_reg = IXGBE_READ_REG(hw, IXGBE_ANLP1);
-
- if ((autoc_reg & IXGBE_AUTOC_SYM_PAUSE) &&
- (anlp1_reg & IXGBE_ANLP1_SYM_PAUSE)) {
- /*
- * Now we need to check if the user selected Rx ONLY
- * of pause frames. In this case, we had to advertise
- * FULL flow control because we could not advertise RX
- * ONLY. Hence, we must now check to see if we need to
- * turn OFF the TRANSMISSION of PAUSE frames.
- */
- if (hw->fc.requested_mode == ixgbe_fc_full) {
- hw->fc.current_mode = ixgbe_fc_full;
- hw_dbg(hw, "Flow Control = FULL.\n");
- } else {
- hw->fc.current_mode = ixgbe_fc_rx_pause;
- hw_dbg(hw, "Flow Control=RX PAUSE only\n");
- }
- } else if (!(autoc_reg & IXGBE_AUTOC_SYM_PAUSE) &&
- (autoc_reg & IXGBE_AUTOC_ASM_PAUSE) &&
- (anlp1_reg & IXGBE_ANLP1_SYM_PAUSE) &&
- (anlp1_reg & IXGBE_ANLP1_ASM_PAUSE)) {
- hw->fc.current_mode = ixgbe_fc_tx_pause;
- hw_dbg(hw, "Flow Control = TX PAUSE frames only.\n");
- } else if ((autoc_reg & IXGBE_AUTOC_SYM_PAUSE) &&
- (autoc_reg & IXGBE_AUTOC_ASM_PAUSE) &&
- !(anlp1_reg & IXGBE_ANLP1_SYM_PAUSE) &&
- (anlp1_reg & IXGBE_ANLP1_ASM_PAUSE)) {
- hw->fc.current_mode = ixgbe_fc_rx_pause;
- hw_dbg(hw, "Flow Control = RX PAUSE frames only.\n");
+ if (hw->fc.requested_mode == ixgbe_fc_full) {
+ hw->fc.current_mode = ixgbe_fc_full;
+ hw_dbg(hw, "Flow Control = FULL.\n");
} else {
- hw->fc.current_mode = ixgbe_fc_none;
- hw_dbg(hw, "Flow Control = NONE.\n");
+ hw->fc.current_mode = ixgbe_fc_rx_pause;
+ hw_dbg(hw, "Flow Control=RX PAUSE frames only\n");
}
+ } else if (!(adv_reg & adv_sym) && (adv_reg & adv_asm) &&
+ (lp_reg & lp_sym) && (lp_reg & lp_asm)) {
+ hw->fc.current_mode = ixgbe_fc_tx_pause;
+ hw_dbg(hw, "Flow Control = TX PAUSE frames only.\n");
+ } else if ((adv_reg & adv_sym) && (adv_reg & adv_asm) &&
+ !(lp_reg & lp_sym) && (lp_reg & lp_asm)) {
+ hw->fc.current_mode = ixgbe_fc_rx_pause;
+ hw_dbg(hw, "Flow Control = RX PAUSE frames only.\n");
+ } else {
+ hw->fc.current_mode = ixgbe_fc_none;
+ hw_dbg(hw, "Flow Control = NONE.\n");
}
- /* Record that current_mode is the result of a successful autoneg */
- hw->fc.fc_was_autonegged = true;
-
-out:
- return ret_val;
+ return 0;
}
/**
@@ -1848,7 +1901,8 @@ out:
static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
{
s32 ret_val = 0;
- u32 reg;
+ u32 reg = 0, reg_bp = 0;
+ u16 reg_cu = 0;
#ifdef CONFIG_DCB
if (hw->fc.requested_mode == ixgbe_fc_pfc) {
@@ -1856,7 +1910,7 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
goto out;
}
-#endif
+#endif /* CONFIG_DCB */
/* Validate the packetbuf configuration */
if (packetbuf_num < 0 || packetbuf_num > 7) {
hw_dbg(hw, "Invalid packet buffer number [%d], expected range "
@@ -1894,11 +1948,26 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
hw->fc.requested_mode = ixgbe_fc_full;
/*
- * Set up the 1G flow control advertisement registers so the HW will be
- * able to do fc autoneg once the cable is plugged in. If we end up
- * using 10g instead, this is harmless.
+ * Set up the 1G and 10G flow control advertisement registers so the
+ * HW will be able to do fc autoneg once the cable is plugged in. If
+ * we link at 10G, the 1G advertisement is harmless and vice versa.
*/
- reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANA);
+
+ switch (hw->phy.media_type) {
+ case ixgbe_media_type_fiber:
+ case ixgbe_media_type_backplane:
+ reg = IXGBE_READ_REG(hw, IXGBE_PCS1GANA);
+ reg_bp = IXGBE_READ_REG(hw, IXGBE_AUTOC);
+ break;
+
+ case ixgbe_media_type_copper:
+ hw->phy.ops.read_reg(hw, MDIO_AN_ADVERTISE,
+ MDIO_MMD_AN, ®_cu);
+ break;
+
+ default:
+ ;
+ }
/*
* The possible values of fc.requested_mode are:
@@ -1917,6 +1986,11 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
case ixgbe_fc_none:
/* Flow control completely disabled by software override. */
reg &= ~(IXGBE_PCS1GANA_SYM_PAUSE | IXGBE_PCS1GANA_ASM_PAUSE);
+ if (hw->phy.media_type == ixgbe_media_type_backplane)
+ reg_bp &= ~(IXGBE_AUTOC_SYM_PAUSE |
+ IXGBE_AUTOC_ASM_PAUSE);
+ else if (hw->phy.media_type == ixgbe_media_type_copper)
+ reg_cu &= ~(IXGBE_TAF_SYM_PAUSE | IXGBE_TAF_ASM_PAUSE);
break;
case ixgbe_fc_rx_pause:
/*
@@ -1928,6 +2002,11 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
* disable the adapter's ability to send PAUSE frames.
*/
reg |= (IXGBE_PCS1GANA_SYM_PAUSE | IXGBE_PCS1GANA_ASM_PAUSE);
+ if (hw->phy.media_type == ixgbe_media_type_backplane)
+ reg_bp |= (IXGBE_AUTOC_SYM_PAUSE |
+ IXGBE_AUTOC_ASM_PAUSE);
+ else if (hw->phy.media_type == ixgbe_media_type_copper)
+ reg_cu |= (IXGBE_TAF_SYM_PAUSE | IXGBE_TAF_ASM_PAUSE);
break;
case ixgbe_fc_tx_pause:
/*
@@ -1936,10 +2015,22 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
*/
reg |= (IXGBE_PCS1GANA_ASM_PAUSE);
reg &= ~(IXGBE_PCS1GANA_SYM_PAUSE);
+ if (hw->phy.media_type == ixgbe_media_type_backplane) {
+ reg_bp |= (IXGBE_AUTOC_ASM_PAUSE);
+ reg_bp &= ~(IXGBE_AUTOC_SYM_PAUSE);
+ } else if (hw->phy.media_type == ixgbe_media_type_copper) {
+ reg_cu |= (IXGBE_TAF_ASM_PAUSE);
+ reg_cu &= ~(IXGBE_TAF_SYM_PAUSE);
+ }
break;
case ixgbe_fc_full:
/* Flow control (both Rx and Tx) is enabled by SW override. */
reg |= (IXGBE_PCS1GANA_SYM_PAUSE | IXGBE_PCS1GANA_ASM_PAUSE);
+ if (hw->phy.media_type == ixgbe_media_type_backplane)
+ reg_bp |= (IXGBE_AUTOC_SYM_PAUSE |
+ IXGBE_AUTOC_ASM_PAUSE);
+ else if (hw->phy.media_type == ixgbe_media_type_copper)
+ reg_cu |= (IXGBE_TAF_SYM_PAUSE | IXGBE_TAF_ASM_PAUSE);
break;
#ifdef CONFIG_DCB
case ixgbe_fc_pfc:
@@ -1953,80 +2044,37 @@ static s32 ixgbe_setup_fc(struct ixgbe_hw *hw, s32 packetbuf_num)
break;
}
- IXGBE_WRITE_REG(hw, IXGBE_PCS1GANA, reg);
- reg = IXGBE_READ_REG(hw, IXGBE_PCS1GLCTL);
-
- /* Disable AN timeout */
- if (hw->fc.strict_ieee)
- reg &= ~IXGBE_PCS1GLCTL_AN_1G_TIMEOUT_EN;
+ if (hw->mac.type != ixgbe_mac_X540) {
+ /*
+ * Enable auto-negotiation between the MAC & PHY;
+ * the MAC will advertise clause 37 flow control.
+ */
+ IXGBE_WRITE_REG(hw, IXGBE_PCS1GANA, reg);
+ reg = IXGBE_READ_REG(hw, IXGBE_PCS1GLCTL);
- IXGBE_WRITE_REG(hw, IXGBE_PCS1GLCTL, reg);
- hw_dbg(hw, "Set up FC; PCS1GLCTL = 0x%08X\n", reg);
+ /* Disable AN timeout */
+ if (hw->fc.strict_ieee)
+ reg &= ~IXGBE_PCS1GLCTL_AN_1G_TIMEOUT_EN;
- /*
- * Set up the 10G flow control advertisement registers so the HW
- * can do fc autoneg once the cable is plugged in. If we end up
- * using 1g instead, this is harmless.
- */
- reg = IXGBE_READ_REG(hw, IXGBE_AUTOC);
+ IXGBE_WRITE_REG(hw, IXGBE_PCS1GLCTL, reg);
+ hw_dbg(hw, "Set up FC; PCS1GLCTL = 0x%08X\n", reg);
+ }
/*
- * The possible values of fc.requested_mode are:
- * 0: Flow control is completely disabled
- * 1: Rx flow control is enabled (we can receive pause frames,
- * but not send pause frames).
- * 2: Tx flow control is enabled (we can send pause frames but
- * we do not support receiving pause frames).
- * 3: Both Rx and Tx flow control (symmetric) are enabled.
- * other: Invalid.
+ * AUTOC restart handles negotiation of 1G and 10G on backplane
+ * and copper. There is no need to set the PCS1GCTL register.
+ *
*/
- switch (hw->fc.requested_mode) {
- case ixgbe_fc_none:
- /* Flow control completely disabled by software override. */
- reg &= ~(IXGBE_AUTOC_SYM_PAUSE | IXGBE_AUTOC_ASM_PAUSE);
- break;
- case ixgbe_fc_rx_pause:
- /*
- * Rx Flow control is enabled and Tx Flow control is
- * disabled by software override. Since there really
- * isn't a way to advertise that we are capable of RX
- * Pause ONLY, we will advertise that we support both
- * symmetric and asymmetric Rx PAUSE. Later, we will
- * disable the adapter's ability to send PAUSE frames.
- */
- reg |= (IXGBE_AUTOC_SYM_PAUSE | IXGBE_AUTOC_ASM_PAUSE);
- break;
- case ixgbe_fc_tx_pause:
- /*
- * Tx Flow control is enabled, and Rx Flow control is
- * disabled by software override.
- */
- reg |= (IXGBE_AUTOC_ASM_PAUSE);
- reg &= ~(IXGBE_AUTOC_SYM_PAUSE);
- break;
- case ixgbe_fc_full:
- /* Flow control (both Rx and Tx) is enabled by SW override. */
- reg |= (IXGBE_AUTOC_SYM_PAUSE | IXGBE_AUTOC_ASM_PAUSE);
- break;
-#ifdef CONFIG_DCB
- case ixgbe_fc_pfc:
- goto out;
- break;
-#endif /* CONFIG_DCB */
- default:
- hw_dbg(hw, "Flow control param set incorrectly\n");
- ret_val = IXGBE_ERR_CONFIG;
- goto out;
- break;
+ if (hw->phy.media_type == ixgbe_media_type_backplane) {
+ reg_bp |= IXGBE_AUTOC_AN_RESTART;
+ IXGBE_WRITE_REG(hw, IXGBE_AUTOC, reg_bp);
+ } else if ((hw->phy.media_type == ixgbe_media_type_copper) &&
+ (ixgbe_device_supports_autoneg_fc(hw) == 0)) {
+ hw->phy.ops.write_reg(hw, MDIO_AN_ADVERTISE,
+ MDIO_MMD_AN, reg_cu);
}
- /*
- * AUTOC restart handles negotiation of 1G and 10G. There is
- * no need to set the PCS1GCTL register.
- */
- reg |= IXGBE_AUTOC_AN_RESTART;
- IXGBE_WRITE_REG(hw, IXGBE_AUTOC, reg);
- hw_dbg(hw, "Set up FC; IXGBE_AUTOC = 0x%08X\n", reg);
+ hw_dbg(hw, "Set up FC; IXGBE_AUTOC = 0x%08X\n", reg);
out:
return ret_val;
}
@@ -2751,6 +2799,28 @@ wwn_prefix_out:
}
/**
+ * ixgbe_device_supports_autoneg_fc - Check if phy supports autoneg flow
+ * control
+ * @hw: pointer to hardware structure
+ *
+ * There are several phys that do not support autoneg flow control. This
+ * function check the device id to see if the associated phy supports
+ * autoneg flow control.
+ **/
+static s32 ixgbe_device_supports_autoneg_fc(struct ixgbe_hw *hw)
+{
+
+ switch (hw->device_id) {
+ case IXGBE_DEV_ID_X540T:
+ return 0;
+ case IXGBE_DEV_ID_82599_T3_LOM:
+ return 0;
+ default:
+ return IXGBE_ERR_FC_NOT_SUPPORTED;
+ }
+}
+
+/**
* ixgbe_set_mac_anti_spoofing - Enable/Disable MAC anti-spoofing
* @hw: pointer to hardware structure
* @enable: enable or disable switch for anti-spoofing
diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c
index 5e8c39d..5998dc9 100644
--- a/drivers/net/ixgbe/ixgbe_main.c
+++ b/drivers/net/ixgbe/ixgbe_main.c
@@ -3775,7 +3775,8 @@ static int ixgbe_non_sfp_link_config(struct ixgbe_hw *hw)
if (ret)
goto link_cfg_out;
- if (hw->mac.ops.get_link_capabilities)
+ autoneg = hw->phy.autoneg_advertised;
+ if ((!autoneg) && (hw->mac.ops.get_link_capabilities))
ret = hw->mac.ops.get_link_capabilities(hw, &autoneg,
&negotiation);
if (ret)
diff --git a/drivers/net/ixgbe/ixgbe_phy.h b/drivers/net/ixgbe/ixgbe_phy.h
index 2327baf..9bf2783 100644
--- a/drivers/net/ixgbe/ixgbe_phy.h
+++ b/drivers/net/ixgbe/ixgbe_phy.h
@@ -58,6 +58,10 @@
#define IXGBE_I2C_EEPROM_STATUS_FAIL 0x2
#define IXGBE_I2C_EEPROM_STATUS_IN_PROGRESS 0x3
+/* Flow control defines */
+#define IXGBE_TAF_SYM_PAUSE 0x400
+#define IXGBE_TAF_ASM_PAUSE 0x800
+
/* Bit-shift macros */
#define IXGBE_SFF_VENDOR_OUI_BYTE0_SHIFT 24
#define IXGBE_SFF_VENDOR_OUI_BYTE1_SHIFT 16
diff --git a/drivers/net/ixgbe/ixgbe_type.h b/drivers/net/ixgbe/ixgbe_type.h
index 76bf21b..f190a4a 100644
--- a/drivers/net/ixgbe/ixgbe_type.h
+++ b/drivers/net/ixgbe/ixgbe_type.h
@@ -2698,6 +2698,9 @@ struct ixgbe_info {
#define IXGBE_ERR_EEPROM_VERSION -24
#define IXGBE_ERR_NO_SPACE -25
#define IXGBE_ERR_OVERTEMP -26
+#define IXGBE_ERR_FC_NOT_NEGOTIATED -27
+#define IXGBE_ERR_FC_NOT_SUPPORTED -28
+#define IXGBE_ERR_FLOW_CONTROL -29
#define IXGBE_ERR_SFP_SETUP_NOT_COMPLETE -30
#define IXGBE_ERR_PBA_SECTION -31
#define IXGBE_ERR_INVALID_ARGUMENT -32
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 12/14] ixgbe: improve logic in ixgbe_init_mbx_params_pf
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
Use if/then instead of an all-inclusive case statement.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_mbx.c | 27 ++++++++++++---------------
1 files changed, 12 insertions(+), 15 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_mbx.c b/drivers/net/ixgbe/ixgbe_mbx.c
index 2acacfa..3cf8aec 100644
--- a/drivers/net/ixgbe/ixgbe_mbx.c
+++ b/drivers/net/ixgbe/ixgbe_mbx.c
@@ -448,23 +448,20 @@ void ixgbe_init_mbx_params_pf(struct ixgbe_hw *hw)
{
struct ixgbe_mbx_info *mbx = &hw->mbx;
- switch (hw->mac.type) {
- case ixgbe_mac_82599EB:
- case ixgbe_mac_X540:
- mbx->timeout = 0;
- mbx->usec_delay = 0;
+ if (hw->mac.type != ixgbe_mac_82599EB &&
+ hw->mac.type != ixgbe_mac_X540)
+ return;
- mbx->size = IXGBE_VFMAILBOX_SIZE;
+ mbx->timeout = 0;
+ mbx->udelay = 0;
- mbx->stats.msgs_tx = 0;
- mbx->stats.msgs_rx = 0;
- mbx->stats.reqs = 0;
- mbx->stats.acks = 0;
- mbx->stats.rsts = 0;
- break;
- default:
- break;
- }
+ mbx->stats.msgs_tx = 0;
+ mbx->stats.msgs_rx = 0;
+ mbx->stats.reqs = 0;
+ mbx->stats.acks = 0;
+ mbx->stats.rsts = 0;
+
+ mbx->size = IXGBE_VFMAILBOX_SIZE;
}
#endif /* CONFIG_PCI_IOV */
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 09/14] ixgbe: Add x540 statistic counter definitions
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
Add defines to accumulate and display x540 PHY statistic counters on
transmit/receive.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_common.c | 9 +++++++++
drivers/net/ixgbe/ixgbe_type.h | 5 +++++
2 files changed, 14 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index 85cc301..561f666 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -209,6 +209,15 @@ s32 ixgbe_clear_hw_cntrs_generic(struct ixgbe_hw *hw)
}
}
+ if (hw->mac.type == ixgbe_mac_X540) {
+ if (hw->phy.id == 0)
+ hw->phy.ops.identify(hw);
+ hw->phy.ops.read_reg(hw, 0x3, IXGBE_PCRC8ECL, &i);
+ hw->phy.ops.read_reg(hw, 0x3, IXGBE_PCRC8ECH, &i);
+ hw->phy.ops.read_reg(hw, 0x3, IXGBE_LDPCECL, &i);
+ hw->phy.ops.read_reg(hw, 0x3, IXGBE_LDPCECH, &i);
+ }
+
return 0;
}
diff --git a/drivers/net/ixgbe/ixgbe_type.h b/drivers/net/ixgbe/ixgbe_type.h
index 13a45c7..76bf21b 100644
--- a/drivers/net/ixgbe/ixgbe_type.h
+++ b/drivers/net/ixgbe/ixgbe_type.h
@@ -671,6 +671,11 @@
#define IXGBE_FCOEDWRC 0x0242C /* Number of FCoE DWords Received */
#define IXGBE_FCOEPTC 0x08784 /* Number of FCoE Packets Transmitted */
#define IXGBE_FCOEDWTC 0x08788 /* Number of FCoE DWords Transmitted */
+#define IXGBE_PCRC8ECL 0x0E810
+#define IXGBE_PCRC8ECH 0x0E811
+#define IXGBE_PCRC8ECH_MASK 0x1F
+#define IXGBE_LDPCECL 0x0E820
+#define IXGBE_LDPCECH 0x0E821
/* Management */
#define IXGBE_MAVTV(_i) (0x05010 + ((_i) * 4)) /* 8 of these (0-7) */
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 13/14] ixgbe: fix spelling errors
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_common.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_common.c b/drivers/net/ixgbe/ixgbe_common.c
index 7a6a3fb..bcd9529 100644
--- a/drivers/net/ixgbe/ixgbe_common.c
+++ b/drivers/net/ixgbe/ixgbe_common.c
@@ -2162,7 +2162,7 @@ out:
* @hw: pointer to hardware structure
* @mask: Mask to specify which semaphore to acquire
*
- * Acquires the SWFW semaphore thought the GSSR register for the specified
+ * Acquires the SWFW semaphore through the GSSR register for the specified
* function (CSR, PHY0, PHY1, EEPROM, Flash)
**/
s32 ixgbe_acquire_swfw_sync(struct ixgbe_hw *hw, u16 mask)
@@ -2210,7 +2210,7 @@ s32 ixgbe_acquire_swfw_sync(struct ixgbe_hw *hw, u16 mask)
* @hw: pointer to hardware structure
* @mask: Mask to specify which semaphore to release
*
- * Releases the SWFW semaphore thought the GSSR register for the specified
+ * Releases the SWFW semaphore through the GSSR register for the specified
* function (CSR, PHY0, PHY1, EEPROM, Flash)
**/
void ixgbe_release_swfw_sync(struct ixgbe_hw *hw, u16 mask)
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 11/14] ixgbe: add function description
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
Add description for ixgbe_init_eeprom_params_X540 and whitespace fix.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_x540.c | 9 ++++++---
1 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_x540.c b/drivers/net/ixgbe/ixgbe_x540.c
index 2e3a2b4..f47e93f 100644
--- a/drivers/net/ixgbe/ixgbe_x540.c
+++ b/drivers/net/ixgbe/ixgbe_x540.c
@@ -251,8 +251,11 @@ static u32 ixgbe_get_supported_physical_layer_X540(struct ixgbe_hw *hw)
}
/**
- * ixgbe_init_eeprom_params_X540 - Initialize EEPROM params
- * @hw: pointer to hardware structure
+ * ixgbe_init_eeprom_params_X540 - Initialize EEPROM params
+ * @hw: pointer to hardware structure
+ *
+ * Initializes the EEPROM parameters ixgbe_eeprom_info within the
+ * ixgbe_hw struct in order to set up EEPROM access.
**/
static s32 ixgbe_init_eeprom_params_X540(struct ixgbe_hw *hw)
{
@@ -271,7 +274,7 @@ static s32 ixgbe_init_eeprom_params_X540(struct ixgbe_hw *hw)
IXGBE_EEPROM_WORD_SIZE_SHIFT);
hw_dbg(hw, "Eeprom params: type = %d, size = %d\n",
- eeprom->type, eeprom->word_size);
+ eeprom->type, eeprom->word_size);
}
return 0;
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 14/14] ixgbe: fix setting and reporting of advertised speeds
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
Add the ability to set 100/F on x540.
Fix reporting of advertised modes by adding check for phy.autoneg_advertised
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_ethtool.c | 41 ++++++++++++++++++-------------------
1 files changed, 20 insertions(+), 21 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_ethtool.c b/drivers/net/ixgbe/ixgbe_ethtool.c
index 83511c0..76380a2 100644
--- a/drivers/net/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ixgbe/ixgbe_ethtool.c
@@ -161,29 +161,25 @@ static int ixgbe_get_settings(struct net_device *netdev,
}
ecmd->advertising = ADVERTISED_Autoneg;
- if (hw->phy.autoneg_advertised & IXGBE_LINK_SPEED_100_FULL)
- ecmd->advertising |= ADVERTISED_100baseT_Full;
- if (hw->phy.autoneg_advertised & IXGBE_LINK_SPEED_10GB_FULL)
- ecmd->advertising |= ADVERTISED_10000baseT_Full;
- if (hw->phy.autoneg_advertised & IXGBE_LINK_SPEED_1GB_FULL)
- ecmd->advertising |= ADVERTISED_1000baseT_Full;
- /*
- * It's possible that phy.autoneg_advertised may not be
- * set yet. If so display what the default would be -
- * both 1G and 10G supported.
- */
- if (!(ecmd->advertising & (ADVERTISED_1000baseT_Full |
- ADVERTISED_10000baseT_Full)))
+ if (hw->phy.autoneg_advertised) {
+ if (hw->phy.autoneg_advertised &
+ IXGBE_LINK_SPEED_100_FULL)
+ ecmd->advertising |= ADVERTISED_100baseT_Full;
+ if (hw->phy.autoneg_advertised &
+ IXGBE_LINK_SPEED_10GB_FULL)
+ ecmd->advertising |= ADVERTISED_10000baseT_Full;
+ if (hw->phy.autoneg_advertised &
+ IXGBE_LINK_SPEED_1GB_FULL)
+ ecmd->advertising |= ADVERTISED_1000baseT_Full;
+ } else {
+ /*
+ * Default advertised modes in case
+ * phy.autoneg_advertised isn't set.
+ */
ecmd->advertising |= (ADVERTISED_10000baseT_Full |
ADVERTISED_1000baseT_Full);
-
- switch (hw->mac.type) {
- case ixgbe_mac_X540:
- if (!(ecmd->advertising & ADVERTISED_100baseT_Full))
- ecmd->advertising |= (ADVERTISED_100baseT_Full);
- break;
- default:
- break;
+ if (hw->mac.type == ixgbe_mac_X540)
+ ecmd->advertising |= ADVERTISED_100baseT_Full;
}
if (hw->phy.media_type == ixgbe_media_type_copper) {
@@ -336,6 +332,9 @@ static int ixgbe_set_settings(struct net_device *netdev,
if (ecmd->advertising & ADVERTISED_1000baseT_Full)
advertised |= IXGBE_LINK_SPEED_1GB_FULL;
+ if (ecmd->advertising & ADVERTISED_100baseT_Full)
+ advertised |= IXGBE_LINK_SPEED_100_FULL;
+
if (old == advertised)
return err;
/* this sets the link speed and restarts auto-neg */
--
1.7.4
^ permalink raw reply related
* [net-next-2.6 07/14] ixgbe: cleanup PHY init
From: Jeff Kirsher @ 2011-03-08 2:25 UTC (permalink / raw)
To: davem; +Cc: Emil Tantilov, netdev, gospo, bphilips, Jeff Kirsher
In-Reply-To: <1299551108-23663-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Emil Tantilov <emil.s.tantilov@intel.com>
This change cleans up several situations in which we were either stepping
over possible errors, or calling initialization routines multiple times.
Also includes whitespace fixes where applicable.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ixgbe/ixgbe_82598.c | 31 ++++++++--
drivers/net/ixgbe/ixgbe_82599.c | 121 +++++++++++++++++++++++----------------
drivers/net/ixgbe/ixgbe_phy.c | 21 ++++++-
3 files changed, 114 insertions(+), 59 deletions(-)
diff --git a/drivers/net/ixgbe/ixgbe_82598.c b/drivers/net/ixgbe/ixgbe_82598.c
index fc41329..dc977b1 100644
--- a/drivers/net/ixgbe/ixgbe_82598.c
+++ b/drivers/net/ixgbe/ixgbe_82598.c
@@ -280,10 +280,22 @@ static enum ixgbe_media_type ixgbe_get_media_type_82598(struct ixgbe_hw *hw)
{
enum ixgbe_media_type media_type;
+ /* Detect if there is a copper PHY attached. */
+ switch (hw->phy.type) {
+ case ixgbe_phy_cu_unknown:
+ case ixgbe_phy_tn:
+ case ixgbe_phy_aq:
+ media_type = ixgbe_media_type_copper;
+ goto out;
+ default:
+ break;
+ }
+
/* Media type for I82598 is based on device ID */
switch (hw->device_id) {
case IXGBE_DEV_ID_82598:
case IXGBE_DEV_ID_82598_BX:
+ /* Default device ID is mezzanine card KX/KX4 */
media_type = ixgbe_media_type_backplane;
break;
case IXGBE_DEV_ID_82598AF_DUAL_PORT:
@@ -306,7 +318,7 @@ static enum ixgbe_media_type ixgbe_get_media_type_82598(struct ixgbe_hw *hw)
media_type = ixgbe_media_type_unknown;
break;
}
-
+out:
return media_type;
}
@@ -632,7 +644,7 @@ out:
* @hw: pointer to hardware structure
* @speed: new link speed
* @autoneg: true if auto-negotiation enabled
- * @autoneg_wait_to_complete: true if waiting is needed to complete
+ * @autoneg_wait_to_complete: true when waiting for completion is needed
*
* Set the link speed in the AUTOC register and restarts link.
**/
@@ -671,7 +683,8 @@ static s32 ixgbe_setup_mac_link_82598(struct ixgbe_hw *hw,
* ixgbe_hw This will write the AUTOC register based on the new
* stored values
*/
- status = ixgbe_start_mac_link_82598(hw, autoneg_wait_to_complete);
+ status = ixgbe_start_mac_link_82598(hw,
+ autoneg_wait_to_complete);
}
return status;
@@ -1090,10 +1103,12 @@ static u32 ixgbe_get_supported_physical_layer_82598(struct ixgbe_hw *hw)
/* Copper PHY must be checked before AUTOC LMS to determine correct
* physical layer because 10GBase-T PHYs use LMS = KX4/KX */
- if (hw->phy.type == ixgbe_phy_tn ||
- hw->phy.type == ixgbe_phy_cu_unknown) {
- hw->phy.ops.read_reg(hw, MDIO_PMA_EXTABLE, MDIO_MMD_PMAPMD,
- &ext_ability);
+ switch (hw->phy.type) {
+ case ixgbe_phy_tn:
+ case ixgbe_phy_aq:
+ case ixgbe_phy_cu_unknown:
+ hw->phy.ops.read_reg(hw, MDIO_PMA_EXTABLE,
+ MDIO_MMD_PMAPMD, &ext_ability);
if (ext_ability & MDIO_PMA_EXTABLE_10GBT)
physical_layer |= IXGBE_PHYSICAL_LAYER_10GBASE_T;
if (ext_ability & MDIO_PMA_EXTABLE_1000BT)
@@ -1101,6 +1116,8 @@ static u32 ixgbe_get_supported_physical_layer_82598(struct ixgbe_hw *hw)
if (ext_ability & MDIO_PMA_EXTABLE_100BTX)
physical_layer |= IXGBE_PHYSICAL_LAYER_100BASE_TX;
goto out;
+ default:
+ break;
}
switch (autoc & IXGBE_AUTOC_LMS_MASK) {
diff --git a/drivers/net/ixgbe/ixgbe_82599.c b/drivers/net/ixgbe/ixgbe_82599.c
index 5ef968a..2f8b9f4 100644
--- a/drivers/net/ixgbe/ixgbe_82599.c
+++ b/drivers/net/ixgbe/ixgbe_82599.c
@@ -470,8 +470,6 @@ static void ixgbe_enable_tx_laser_multispeed_fiber(struct ixgbe_hw *hw)
**/
static void ixgbe_flap_tx_laser_multispeed_fiber(struct ixgbe_hw *hw)
{
- hw_dbg(hw, "ixgbe_flap_tx_laser_multispeed_fiber\n");
-
if (hw->mac.autotry_restart) {
ixgbe_disable_tx_laser_multispeed_fiber(hw);
ixgbe_enable_tx_laser_multispeed_fiber(hw);
@@ -494,17 +492,21 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
bool autoneg_wait_to_complete)
{
s32 status = 0;
- ixgbe_link_speed phy_link_speed;
+ ixgbe_link_speed link_speed = IXGBE_LINK_SPEED_UNKNOWN;
ixgbe_link_speed highest_link_speed = IXGBE_LINK_SPEED_UNKNOWN;
u32 speedcnt = 0;
u32 esdp_reg = IXGBE_READ_REG(hw, IXGBE_ESDP);
+ u32 i = 0;
bool link_up = false;
bool negotiation;
- int i;
/* Mask off requested but non-supported speeds */
- hw->mac.ops.get_link_capabilities(hw, &phy_link_speed, &negotiation);
- speed &= phy_link_speed;
+ status = hw->mac.ops.get_link_capabilities(hw, &link_speed,
+ &negotiation);
+ if (status != 0)
+ return status;
+
+ speed &= link_speed;
/*
* Try each speed one by one, highest priority first. We do this in
@@ -515,9 +517,12 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
highest_link_speed = IXGBE_LINK_SPEED_10GB_FULL;
/* If we already have link at this speed, just jump out */
- hw->mac.ops.check_link(hw, &phy_link_speed, &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed, &link_up,
+ false);
+ if (status != 0)
+ return status;
- if ((phy_link_speed == IXGBE_LINK_SPEED_10GB_FULL) && link_up)
+ if ((link_speed == IXGBE_LINK_SPEED_10GB_FULL) && link_up)
goto out;
/* Set the module link speed */
@@ -529,9 +534,9 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
msleep(40);
status = ixgbe_setup_mac_link_82599(hw,
- IXGBE_LINK_SPEED_10GB_FULL,
- autoneg,
- autoneg_wait_to_complete);
+ IXGBE_LINK_SPEED_10GB_FULL,
+ autoneg,
+ autoneg_wait_to_complete);
if (status != 0)
return status;
@@ -548,8 +553,11 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
msleep(100);
/* If we have link, just jump out */
- hw->mac.ops.check_link(hw, &phy_link_speed,
- &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed,
+ &link_up, false);
+ if (status != 0)
+ return status;
+
if (link_up)
goto out;
}
@@ -561,9 +569,12 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
highest_link_speed = IXGBE_LINK_SPEED_1GB_FULL;
/* If we already have link at this speed, just jump out */
- hw->mac.ops.check_link(hw, &phy_link_speed, &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed, &link_up,
+ false);
+ if (status != 0)
+ return status;
- if ((phy_link_speed == IXGBE_LINK_SPEED_1GB_FULL) && link_up)
+ if ((link_speed == IXGBE_LINK_SPEED_1GB_FULL) && link_up)
goto out;
/* Set the module link speed */
@@ -576,9 +587,9 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
msleep(40);
status = ixgbe_setup_mac_link_82599(hw,
- IXGBE_LINK_SPEED_1GB_FULL,
- autoneg,
- autoneg_wait_to_complete);
+ IXGBE_LINK_SPEED_1GB_FULL,
+ autoneg,
+ autoneg_wait_to_complete);
if (status != 0)
return status;
@@ -589,7 +600,11 @@ s32 ixgbe_setup_mac_link_multispeed_fiber(struct ixgbe_hw *hw,
msleep(100);
/* If we have link, just jump out */
- hw->mac.ops.check_link(hw, &phy_link_speed, &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed, &link_up,
+ false);
+ if (status != 0)
+ return status;
+
if (link_up)
goto out;
}
@@ -632,13 +647,10 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
bool autoneg_wait_to_complete)
{
s32 status = 0;
- ixgbe_link_speed link_speed;
+ ixgbe_link_speed link_speed = IXGBE_LINK_SPEED_UNKNOWN;
s32 i, j;
bool link_up = false;
u32 autoc_reg = IXGBE_READ_REG(hw, IXGBE_AUTOC);
- struct ixgbe_adapter *adapter = hw->back;
-
- hw_dbg(hw, "ixgbe_setup_mac_link_smartspeed.\n");
/* Set autoneg_advertised value based on input link speed */
hw->phy.autoneg_advertised = 0;
@@ -664,7 +676,7 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
for (j = 0; j < IXGBE_SMARTSPEED_MAX_RETRIES; j++) {
status = ixgbe_setup_mac_link_82599(hw, speed, autoneg,
autoneg_wait_to_complete);
- if (status)
+ if (status != 0)
goto out;
/*
@@ -677,8 +689,11 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
mdelay(100);
/* If we have link, just jump out */
- hw->mac.ops.check_link(hw, &link_speed,
- &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed,
+ &link_up, false);
+ if (status != 0)
+ goto out;
+
if (link_up)
goto out;
}
@@ -696,7 +711,7 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
hw->phy.smart_speed_active = true;
status = ixgbe_setup_mac_link_82599(hw, speed, autoneg,
autoneg_wait_to_complete);
- if (status)
+ if (status != 0)
goto out;
/*
@@ -709,8 +724,11 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
mdelay(100);
/* If we have link, just jump out */
- hw->mac.ops.check_link(hw, &link_speed,
- &link_up, false);
+ status = hw->mac.ops.check_link(hw, &link_speed,
+ &link_up, false);
+ if (status != 0)
+ goto out;
+
if (link_up)
goto out;
}
@@ -722,7 +740,7 @@ static s32 ixgbe_setup_mac_link_smartspeed(struct ixgbe_hw *hw,
out:
if (link_up && (link_speed == IXGBE_LINK_SPEED_1GB_FULL))
- e_info(hw, "Smartspeed has downgraded the link speed from "
+ hw_dbg(hw, "Smartspeed has downgraded the link speed from "
"the maximum advertised\n");
return status;
}
@@ -883,7 +901,7 @@ static s32 ixgbe_reset_hw_82599(struct ixgbe_hw *hw)
/* PHY ops must be identified and initialized prior to reset */
- /* Init PHY and function pointers, perform SFP setup */
+ /* Identify PHY and related function pointers */
status = hw->phy.ops.init(hw);
if (status == IXGBE_ERR_SFP_NOT_SUPPORTED)
@@ -895,6 +913,9 @@ static s32 ixgbe_reset_hw_82599(struct ixgbe_hw *hw)
hw->phy.sfp_setup_needed = false;
}
+ if (status == IXGBE_ERR_SFP_NOT_SUPPORTED)
+ goto reset_hw_out;
+
/* Reset PHY */
if (hw->phy.reset_disable == false && hw->phy.ops.reset != NULL)
hw->phy.ops.reset(hw);
@@ -2051,28 +2072,28 @@ static struct ixgbe_mac_operations mac_ops_82599 = {
};
static struct ixgbe_eeprom_operations eeprom_ops_82599 = {
- .init_params = &ixgbe_init_eeprom_params_generic,
- .read = &ixgbe_read_eerd_generic,
- .write = &ixgbe_write_eeprom_generic,
- .calc_checksum = &ixgbe_calc_eeprom_checksum_generic,
- .validate_checksum = &ixgbe_validate_eeprom_checksum_generic,
- .update_checksum = &ixgbe_update_eeprom_checksum_generic,
+ .init_params = &ixgbe_init_eeprom_params_generic,
+ .read = &ixgbe_read_eerd_generic,
+ .write = &ixgbe_write_eeprom_generic,
+ .calc_checksum = &ixgbe_calc_eeprom_checksum_generic,
+ .validate_checksum = &ixgbe_validate_eeprom_checksum_generic,
+ .update_checksum = &ixgbe_update_eeprom_checksum_generic,
};
static struct ixgbe_phy_operations phy_ops_82599 = {
- .identify = &ixgbe_identify_phy_82599,
- .identify_sfp = &ixgbe_identify_sfp_module_generic,
- .init = &ixgbe_init_phy_ops_82599,
- .reset = &ixgbe_reset_phy_generic,
- .read_reg = &ixgbe_read_phy_reg_generic,
- .write_reg = &ixgbe_write_phy_reg_generic,
- .setup_link = &ixgbe_setup_phy_link_generic,
- .setup_link_speed = &ixgbe_setup_phy_link_speed_generic,
- .read_i2c_byte = &ixgbe_read_i2c_byte_generic,
- .write_i2c_byte = &ixgbe_write_i2c_byte_generic,
- .read_i2c_eeprom = &ixgbe_read_i2c_eeprom_generic,
- .write_i2c_eeprom = &ixgbe_write_i2c_eeprom_generic,
- .check_overtemp = &ixgbe_tn_check_overtemp,
+ .identify = &ixgbe_identify_phy_82599,
+ .identify_sfp = &ixgbe_identify_sfp_module_generic,
+ .init = &ixgbe_init_phy_ops_82599,
+ .reset = &ixgbe_reset_phy_generic,
+ .read_reg = &ixgbe_read_phy_reg_generic,
+ .write_reg = &ixgbe_write_phy_reg_generic,
+ .setup_link = &ixgbe_setup_phy_link_generic,
+ .setup_link_speed = &ixgbe_setup_phy_link_speed_generic,
+ .read_i2c_byte = &ixgbe_read_i2c_byte_generic,
+ .write_i2c_byte = &ixgbe_write_i2c_byte_generic,
+ .read_i2c_eeprom = &ixgbe_read_i2c_eeprom_generic,
+ .write_i2c_eeprom = &ixgbe_write_i2c_eeprom_generic,
+ .check_overtemp = &ixgbe_tn_check_overtemp,
};
struct ixgbe_info ixgbe_82599_info = {
diff --git a/drivers/net/ixgbe/ixgbe_phy.c b/drivers/net/ixgbe/ixgbe_phy.c
index 197230b..9190a8f 100644
--- a/drivers/net/ixgbe/ixgbe_phy.c
+++ b/drivers/net/ixgbe/ixgbe_phy.c
@@ -57,6 +57,7 @@ s32 ixgbe_identify_phy_generic(struct ixgbe_hw *hw)
{
s32 status = IXGBE_ERR_PHY_ADDR_INVALID;
u32 phy_addr;
+ u16 ext_ability = 0;
if (hw->phy.type == ixgbe_phy_unknown) {
for (phy_addr = 0; phy_addr < IXGBE_MAX_PHY_ADDR; phy_addr++) {
@@ -65,12 +66,29 @@ s32 ixgbe_identify_phy_generic(struct ixgbe_hw *hw)
ixgbe_get_phy_id(hw);
hw->phy.type =
ixgbe_get_phy_type_from_id(hw->phy.id);
+
+ if (hw->phy.type == ixgbe_phy_unknown) {
+ hw->phy.ops.read_reg(hw,
+ MDIO_PMA_EXTABLE,
+ MDIO_MMD_PMAPMD,
+ &ext_ability);
+ if (ext_ability &
+ (MDIO_PMA_EXTABLE_10GBT |
+ MDIO_PMA_EXTABLE_1000BT))
+ hw->phy.type =
+ ixgbe_phy_cu_unknown;
+ else
+ hw->phy.type =
+ ixgbe_phy_generic;
+ }
+
status = 0;
break;
}
}
/* clear value if nothing found */
- hw->phy.mdio.prtad = 0;
+ if (status != 0)
+ hw->phy.mdio.prtad = 0;
} else {
status = 0;
}
@@ -823,7 +841,6 @@ s32 ixgbe_identify_sfp_module_generic(struct ixgbe_hw *hw)
goto out;
}
- /* This is guaranteed to be 82599, no need to check for NULL */
hw->mac.ops.get_device_caps(hw, &enforce_sfp);
if (!(enforce_sfp & IXGBE_DEVICE_CAPS_ALLOW_ANY_SFP) &&
!((hw->phy.sfp_type == ixgbe_sfp_type_1g_cu_core0) ||
--
1.7.4
^ permalink raw reply related
* Re: [net-next 08/12] ixgb: convert to new VLAN model
From: Jeff Kirsher @ 2011-03-08 2:30 UTC (permalink / raw)
To: Jesse Gross
Cc: Tantilov, Emil S, davem@davemloft.net, netdev@vger.kernel.org,
bphilips@novell.com, Pieper, Jeffrey E, Ben Hutchings
In-Reply-To: <AANLkTikkMTv4uCbpOFfdgHzt2jMa5u5ipCOQ3OOh3KLs@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 3131 bytes --]
On Mon, 2011-03-07 at 17:31 -0800, Jesse Gross wrote:
> On Tue, Jan 25, 2011 at 10:20 AM, Tantilov, Emil S
> <emil.s.tantilov@intel.com> wrote:
> >>-----Original Message-----
> >>From: Jesse Gross [mailto:jesse@nicira.com]
> >>Sent: Tuesday, January 25, 2011 9:23 AM
> >>To: Tantilov, Emil S
> >>Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> >>bphilips@novell.com; Pieper, Jeffrey E
> >>Subject: Re: [net-next 08/12] ixgb: convert to new VLAN model
> >>
> >>On Sun, Jan 23, 2011 at 4:25 PM, Tantilov, Emil S
> >><emil.s.tantilov@intel.com> wrote:
> >>> Jesse Gross wrote:
> >>>> On Thu, Jan 6, 2011 at 7:29 PM, <jeffrey.t.kirsher@intel.com> wrote:
> >>>>> +static int ixgb_set_flags(struct net_device *netdev, u32 data) +{
> >>>>> + struct ixgb_adapter *adapter = netdev_priv(netdev); +
> >>>>> bool need_reset; + int rc;
> >>>>> +
> >>>>> + /*
> >>>>> + * TX vlan insertion does not work per HW design when Rx
> >>>>> stripping is + * disabled. Disable txvlan when rxvlan is
> >>>>> off. + */ + if ((data & ETH_FLAG_RXVLAN) !=
> >>>>> (netdev->features & NETIF_F_HW_VLAN_RX)) + data ^=
> >>>>> ETH_FLAG_TXVLAN;
> >>>>
> >>>> Does this really do the right thing? If the RX vlan setting is
> >>>> changed, it will do the opposite of what the user requested for TX
> >>>> vlan?
> >>>>
> >>>> So if I start with both on (the default) and turn them both off in one
> >>>> command (a valid setting), I will get RX off and TX on (an invalid
> >>>> setting).
> >>>>
> >>>> Why not:
> >>>>
> >>>> if (!(data & ETH_FLAG_RXVLAN))
> >>>> data &= ~ETH_FLAG_TXVLAN;
> >>>
> >>> Yeah that works for disabling rxvlan, but what if rxvlan is disabled, and
> >>the user attempts to enable txvlan? At least our validation argued that we
> >>should make it work both ways. Perhaps something like the following?
> >>>
> >>> if (!(data & ETH_FLAG_RXVLAN) &&
> >>> (netdev->features & NETIF_F_HW_VLAN_TX))
> >>> data &= ~ETH_FLAG_TXVLAN;
> >>> else if (data & ETH_FLAG_TXVLAN)
> >>> data |= ETH_FLAG_RXVLAN;
> >>
> >>I think the logic above does what you describe and will always result
> >>in a consistent state. Turning dependent features on when needed is a
> >>little bit inconsistent with the rest of Ethtool (for example, turning
> >>on TSO when checksum offloading is off will not enable checksum
> >>offloading, it will produce an error). However, I know that drivers
> >
> > That is the reason I asked, as I don't want to keep bouncing the patch back and forth. Personally I like the idea of helping the user and adjusting the flags to something that works rather than a generic error message.
>
> Were you able to take a look at this? I think that we have something
> that is pretty close, so it would be nice to wrap it up.
I have a patch which has passed testing from Emil, I just need to check
with Emil that the current patch I have ready to push is the "latest"
version and does not need any more tweaks.
Cheers,
Jeff
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply
* Re: [net-next-2.6 04/14] net: add ndo_fcoe_ddp_target() to support FCoE DDP in target mode
From: Ben Hutchings @ 2011-03-08 2:43 UTC (permalink / raw)
To: Jeff Kirsher; +Cc: davem, Yi Zou, netdev, gospo, bphilips, Kiran Patil
In-Reply-To: <1299551108-23663-5-git-send-email-jeffrey.t.kirsher@intel.com>
On Mon, 2011-03-07 at 18:24 -0800, Jeff Kirsher wrote:
> From: Yi Zou <yi.zou@intel.com>
>
> The Fiber Channel over Ethernet (FCoE) Direct Data Placement (DDP) can also be
> used for FCoE target, where the DDP used for read I/O on an initiator can be
> used on an FCoE target to speed up the write I/O to the target from the initiator.
> The added ndo_fcoe_ddp_target() works in the similar way as the existing
> ndo_fcoe_ddp_setup() to allow the underlying hardware set up the DDP context
> accordingly when it gets called from the FCoE target implementation on top
> the existing Open-FCoE fcoe/libfc protocol stack so without losing the ability
> to provide DDP for read I/O as an initiator, it can also provide DDP offload
> to the write I/O coming from the initiator as a target.
>
> Signed-off-by: Yi Zou <yi.zou@intel.com>
> Signed-off-by: Kiran Patil <kiran.patil@intel.com>
> Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> ---
> include/linux/netdevice.h | 4 ++++
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> index 71563e7..6bd5d46 100644
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -871,6 +871,10 @@ struct net_device_ops {
> unsigned int sgc);
> int (*ndo_fcoe_ddp_done)(struct net_device *dev,
> u16 xid);
> + int (*ndo_fcoe_ddp_target)(struct net_device *dev,
> + u16 xid,
> + struct scatterlist *sgl,
> + unsigned int sgc);
This should be described in the comment heading net_device_ops.
(Really, all the existing FCoE operations ought to have been described
as well.)
Ben.
> #define NETDEV_FCOE_WWNN 0
> #define NETDEV_FCOE_WWPN 1
> int (*ndo_fcoe_get_wwn)(struct net_device *dev,
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* Re: [patch net-next-2.6] net: reinject arps into bonding slave instead of master
From: Andy Gospodarek @ 2011-03-08 2:43 UTC (permalink / raw)
To: Nicolas de Pesloüan
Cc: Andy Gospodarek, Jiri Pirko, netdev, davem, shemminger, kaber,
fubar, eric.dumazet
In-Reply-To: <4D7565BC.4050201@gmail.com>
On Tue, Mar 08, 2011 at 12:09:48AM +0100, Nicolas de Pesloüan wrote:
>
> After Jiri's last patch series, bond_arp_rcv() is not registered anymore
> as a protocol handler on bond0, but directly called from inside
> bond_handle_frame(), through bond->recv_probe.
>
> Because bond_handler_frame() is a rx_handler for the slave interfaces,
> bond_arp_rcv() is now called at the slave level and not a the master
> level anymore.
>
> Hence this patch and the reason I thought it should work.
>
> Did you tested this patch with Jiri's previous patches applied before?
>
No, I have not tested them yet. I looked at this problem first as
Jiri's patch did not fix a regression in code that is actually in
net-next right now.
I don't have a problem with the work you and Jiri are doing on this
latest patch series, but the frequency of patch revisions and now
regressions concerns me.
^ permalink raw reply
* Re: [net-next-2.6 05/14] vlan: add support to ndo_fcoe_ddp_target()
From: Ben Hutchings @ 2011-03-08 2:44 UTC (permalink / raw)
To: Jeff Kirsher; +Cc: davem, Yi Zou, netdev, gospo, bphilips, Kiran Patil
In-Reply-To: <1299551108-23663-6-git-send-email-jeffrey.t.kirsher@intel.com>
On Mon, 2011-03-07 at 18:24 -0800, Jeff Kirsher wrote:
> From: Yi Zou <yi.zou@intel.com>
>
> Add the new target ddp offload support ndo_fcoe_ddp_target().
>
> Signed-off-by: Yi Zou <yi.zou@intel.com>
> Signed-off-by: Kiran Patil <kiran.patil@intel.com>
> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> ---
> net/8021q/vlan_dev.c | 14 ++++++++++++++
> 1 files changed, 14 insertions(+), 0 deletions(-)
>
> diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
> index be73753..ae610f0 100644
> --- a/net/8021q/vlan_dev.c
> +++ b/net/8021q/vlan_dev.c
> @@ -625,6 +625,19 @@ static int vlan_dev_fcoe_get_wwn(struct net_device *dev, u64 *wwn, int type)
> rc = ops->ndo_fcoe_get_wwn(real_dev, wwn, type);
> return rc;
> }
> +
> +static int vlan_dev_fcoe_ddp_target(struct net_device *dev, u16 xid,
> + struct scatterlist *sgl, unsigned int sgc)
> +{
> + struct net_device *real_dev = vlan_dev_info(dev)->real_dev;
> + const struct net_device_ops *ops = real_dev->netdev_ops;
> + int rc = 0;
> +
> + if (ops->ndo_fcoe_ddp_target)
> + rc = ops->ndo_fcoe_ddp_target(real_dev, xid, sgl, sgc);
> +
> + return rc;
> +}
I can't see a caller or any documentation for this operation, so I don't
know what the semantics are supposed to be. But shouldn't the return
value be -EOPNOTSUPP if real_dev does not implement the operation?
Ben.
> #endif
>
> static void vlan_dev_change_rx_flags(struct net_device *dev, int change)
> @@ -858,6 +871,7 @@ static const struct net_device_ops vlan_netdev_ops = {
> .ndo_fcoe_enable = vlan_dev_fcoe_enable,
> .ndo_fcoe_disable = vlan_dev_fcoe_disable,
> .ndo_fcoe_get_wwn = vlan_dev_fcoe_get_wwn,
> + .ndo_fcoe_ddp_target = vlan_dev_fcoe_ddp_target,
> #endif
> };
>
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* RE: [net-next-2.6 04/14] net: add ndo_fcoe_ddp_target() to support FCoE DDP in target mode
From: Zou, Yi @ 2011-03-08 2:54 UTC (permalink / raw)
To: Ben Hutchings, Kirsher, Jeffrey T
Cc: davem@davemloft.net, netdev@vger.kernel.org, gospo@redhat.com,
bphilips@novell.com, Patil, Kiran
In-Reply-To: <1299552202.3967.2.camel@localhost>
> On Mon, 2011-03-07 at 18:24 -0800, Jeff Kirsher wrote:
> > From: Yi Zou <yi.zou@intel.com>
> >
> > The Fiber Channel over Ethernet (FCoE) Direct Data Placement (DDP) can
> also be
> > used for FCoE target, where the DDP used for read I/O on an initiator
> can be
> > used on an FCoE target to speed up the write I/O to the target from the
> initiator.
> > The added ndo_fcoe_ddp_target() works in the similar way as the
> existing
> > ndo_fcoe_ddp_setup() to allow the underlying hardware set up the DDP
> context
> > accordingly when it gets called from the FCoE target implementation on
> top
> > the existing Open-FCoE fcoe/libfc protocol stack so without losing the
> ability
> > to provide DDP for read I/O as an initiator, it can also provide DDP
> offload
> > to the write I/O coming from the initiator as a target.
> >
> > Signed-off-by: Yi Zou <yi.zou@intel.com>
> > Signed-off-by: Kiran Patil <kiran.patil@intel.com>
> > Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
> > Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> > ---
> > include/linux/netdevice.h | 4 ++++
> > 1 files changed, 4 insertions(+), 0 deletions(-)
> >
> > diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
> > index 71563e7..6bd5d46 100644
> > --- a/include/linux/netdevice.h
> > +++ b/include/linux/netdevice.h
> > @@ -871,6 +871,10 @@ struct net_device_ops {
> > unsigned int sgc);
> > int (*ndo_fcoe_ddp_done)(struct net_device *dev,
> > u16 xid);
> > + int (*ndo_fcoe_ddp_target)(struct net_device *dev,
> > + u16 xid,
> > + struct scatterlist *sgl,
> > + unsigned int sgc);
>
> This should be described in the comment heading net_device_ops.
> (Really, all the existing FCoE operations ought to have been described
> as well.)
>
> Ben.
Yes, you are right, it should be. Will it be ok if I send a separate
patch as follow-up to properly add all comments?
yi
>
> > #define NETDEV_FCOE_WWNN 0
> > #define NETDEV_FCOE_WWPN 1
> > int (*ndo_fcoe_get_wwn)(struct net_device *dev,
>
> --
> Ben Hutchings, Senior Software Engineer, Solarflare Communications
> Not speaking for my employer; that's the marketing department's job.
> They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* RE: [net-next-2.6 05/14] vlan: add support to ndo_fcoe_ddp_target()
From: Zou, Yi @ 2011-03-08 2:51 UTC (permalink / raw)
To: Ben Hutchings, Kirsher, Jeffrey T
Cc: davem@davemloft.net, netdev@vger.kernel.org, gospo@redhat.com,
bphilips@novell.com, Patil, Kiran
In-Reply-To: <1299552267.3967.3.camel@localhost>
> On Mon, 2011-03-07 at 18:24 -0800, Jeff Kirsher wrote:
> > From: Yi Zou <yi.zou@intel.com>
> >
> > Add the new target ddp offload support ndo_fcoe_ddp_target().
> >
> > Signed-off-by: Yi Zou <yi.zou@intel.com>
> > Signed-off-by: Kiran Patil <kiran.patil@intel.com>
> > Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> > ---
> > net/8021q/vlan_dev.c | 14 ++++++++++++++
> > 1 files changed, 14 insertions(+), 0 deletions(-)
> >
> > diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
> > index be73753..ae610f0 100644
> > --- a/net/8021q/vlan_dev.c
> > +++ b/net/8021q/vlan_dev.c
> > @@ -625,6 +625,19 @@ static int vlan_dev_fcoe_get_wwn(struct net_device
> *dev, u64 *wwn, int type)
> > rc = ops->ndo_fcoe_get_wwn(real_dev, wwn, type);
> > return rc;
> > }
> > +
> > +static int vlan_dev_fcoe_ddp_target(struct net_device *dev, u16 xid,
> > + struct scatterlist *sgl, unsigned int sgc)
> > +{
> > + struct net_device *real_dev = vlan_dev_info(dev)->real_dev;
> > + const struct net_device_ops *ops = real_dev->netdev_ops;
> > + int rc = 0;
> > +
> > + if (ops->ndo_fcoe_ddp_target)
> > + rc = ops->ndo_fcoe_ddp_target(real_dev, xid, sgl, sgc);
> > +
> > + return rc;
> > +}
>
> I can't see a caller or any documentation for this operation, so I don't
> know what the semantics are supposed to be. But shouldn't the return
> value be -EOPNOTSUPP if real_dev does not implement the operation?
>
> Ben.
This is similar to existing ndo_fcoe_ddp_setup(), but is for target mode,
the two are the same on semantics, but the LLD implementation will/may
be different depending on the driver as e.g. ,target works on RX_ID where
initiator works on OX_ID. The returning of rc being 0 indicates no DDP
was setup for this fc exchange, not necessarily EOPNOTSUPP.
Thanks,
yi
>
> > #endif
> >
> > static void vlan_dev_change_rx_flags(struct net_device *dev, int
> change)
> > @@ -858,6 +871,7 @@ static const struct net_device_ops vlan_netdev_ops
> = {
> > .ndo_fcoe_enable = vlan_dev_fcoe_enable,
> > .ndo_fcoe_disable = vlan_dev_fcoe_disable,
> > .ndo_fcoe_get_wwn = vlan_dev_fcoe_get_wwn,
> > + .ndo_fcoe_ddp_target = vlan_dev_fcoe_ddp_target,
> > #endif
> > };
> >
>
> --
> Ben Hutchings, Senior Software Engineer, Solarflare Communications
> Not speaking for my employer; that's the marketing department's job.
> They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* linux-next: manual merge of the net tree with the net-current tree
From: Stephen Rothwell @ 2011-03-08 3:09 UTC (permalink / raw)
To: David Miller, netdev
Cc: linux-next, linux-kernel, Dmitry Kravkov, Eilon Greenstein,
Vladislav Zolotarov
Hi all,
Today's linux-next merge of the net tree got a conflict in
drivers/net/bnx2x/bnx2x_cmn.c between commit
e3835b99333eb3ac7222f6fc0af5cae46074ac49 ("bnx2x: (NPAR) prevent HW
access in D3 state") from the net-current tree and commit
6e30dd4e3935ddb4e7dd27d5be7a6e5504e64a27 (bnx2x: Proper
netdev->ndo_set_rx_mode() implementation"") from the net tree.
Just overlapping additions (I think). I have fixed it up (see below) and
can carry the fix as necessary.
--
Cheers,
Stephen Rothwell sfr@canb.auug.org.au
diff --cc drivers/net/bnx2x/bnx2x_cmn.c
index a71b329,b01b622..0000000
--- a/drivers/net/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/bnx2x/bnx2x_cmn.c
@@@ -1481,11 -1489,15 +1506,20 @@@ int bnx2x_nic_load(struct bnx2x *bp, in
bnx2x_set_eth_mac(bp, 1);
+ if (bp->pending_max) {
+ bnx2x_update_max_mf_config(bp, bp->pending_max);
+ bp->pending_max = 0;
+ }
+
+ /* Clear MC configuration */
+ if (CHIP_IS_E1(bp))
+ bnx2x_invalidate_e1_mc_list(bp);
+ else
+ bnx2x_invalidate_e1h_mc_list(bp);
+
+ /* Clear UC lists configuration */
+ bnx2x_invalidate_uc_list(bp);
+
if (bp->port.pmf)
bnx2x_initial_phy_init(bp, load_mode);
^ permalink raw reply
* Re: [Patch] bonding: fix netpoll in active-backup mode
From: Cong Wang @ 2011-03-08 4:15 UTC (permalink / raw)
To: Neil Horman
Cc: linux-kernel, Jay Vosburgh, David S. Miller, Herbert Xu,
Paul E. McKenney, John W. Linville, Eric Dumazet, netdev
In-Reply-To: <20110307185038.GA31788@hmsreliant.think-freely.org>
于 2011年03月08日 02:50, Neil Horman 写道:
> On Mon, Mar 07, 2011 at 10:11:50PM +0800, Amerigo Wang wrote:
>> netconsole doesn't work in active-backup mode, because we don't do anything
>> for nic failover in active-backup mode. This patch fixes the problem by:
>>
>> 1) make slave_enable_netpoll() and slave_disable_netpoll() callable in softirq
>> context, that is, moving code after synchronize_rcu_bh() into call_rcu_bh()
>> callback function, teaching kzalloc() to use GFP_ATOMIC.
>>
>> 2) disable netpoll on old slave and enable netpoll on the new slave.
>>
>> Tested by ifdown the current active slave and ifup it again for several times,
>> netconsole works well.
>>
>> Signed-off-by: WANG Cong<amwang@redhat.com>
>>
> I may be missing soething but this seems way over-complicated to me. I presume
> the problem is that in active backup mode a failover results in the new active
> slave not having netpoll setup on it? If thats the case, why not just setup
> netpoll on all slaves when ndo_netpoll_setup is called on the bonding interface?
> I don't see anything immeidately catastrophic that would happen as a result.
But we still need to clean up the netpoll on the failing slave, which still
needs to call slave_disable_netpoll() in monitor code, I see no big differences
with the solution I take.
> And then you wouldn't have to worry about disabling/enabling anything on a
> failover (or during a panic for that matter). As for the rcu bits? Why are
> they needed? One would presume that wouldn't (or at least shouldn't) be able to
> teardown our netpoll setup until such time as all the pending frames for that
> netpoll client have been transmitted. If we're not blocknig on that RCU isn't
> really going to help. Seems like the proper fix is take a reference to the
> appropriate npinfo struct in netpoll_send_skb, and drop it from the skbs
> destructor or some such.
I saw a "scheduling while in atomic" warning without touching the rcu bits.
Thanks!
^ permalink raw reply
* Re: dccp test-tree [RFC] [Patch 1/1] dccp: Only activate NN values after receiving the Confirm option
From: Samuel Jero @ 2011-03-08 4:50 UTC (permalink / raw)
To: Gerrit Renker; +Cc: dccp, netdev
In-Reply-To: <20110228112511.GE3620@gerrit.erg.abdn.ac.uk>
[-- Attachment #1: Type: text/plain, Size: 8053 bytes --]
> I am sending this as RFC since I have not yet deeply tested this. It makes
> the exchange of NN options in established state conform to RFC 4340, 6.6.1
> and thus actually is a bug fix.
I now had a chance to put this patch through extensive tests. I had to
modify two things about the patch to get good performance:
1)CCID2 now checks the congestion window against the currently
negotiating ack ratio value to determine whether to change the ack ratio
after an RTO, an idle period, or a loss. This prevents a problem where
an RTO or loss or idle period occurs and the ack ratio is less than the
congestion window but is in the process of being negotiated larger. That
negotiation will reach the sender on the first packet sent and RTOs will
result.
The primary change is adding a function called
dccp_feat_get_nn_next_val() to feat.c. This function returns the value
of the NN feature indicated that is is the process of being negotiated.
If the feature is not being negotiated, then the current value is
returned.
2)In a situation where the ack ratio has to be reduced because of an
RTO, idle period, or loss, CCID2 now sets the ack ratio to half of the
congestion window (or 1 if that's zero) instead of to the congestion
window. This should reduce the problems if one ack is lost (we have to
lose two acks to not acknowledge an entire congestion window and trigger
RTO)
Below is an improved patch. What are your thoughts on these changes?
Samuel Jero
Internetworking Research Group
Ohio University
>>>>>>>>>>>>>>>>>>>>>>>>>Patch<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
dccp: Only activate NN values after receiving the Confirm option
This defers changing local values using exchange of NN options in
established connection state by only activating the value after
receiving the Confirm option, as mandated by RFC 4340, 6.6.1.
Signed-off-by: Samuel Jero
diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c
--- a/net/dccp/ccids/ccid2.c
+++ b/net/dccp/ccids/ccid2.c
@@ -83,9 +82,13 @@ static int ccid2_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
return CCID_PACKET_SEND_AT_ONCE;
}
+static __u64 ccid2_ack_ratio_next(struct sock *sk)
+{
+ return dccp_feat_get_nn_next_val(sk, DCCPF_ACK_RATIO);
+}
+
static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val)
{
- struct dccp_sock *dp = dccp_sk(sk);
u32 max_ratio = DIV_ROUND_UP(ccid2_hc_tx_sk(sk)->tx_cwnd, 2);
/*
@@ -101,11 +104,10 @@ static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val)
if (val > DCCPF_ACK_RATIO_MAX)
val = DCCPF_ACK_RATIO_MAX;
- if (val == dp->dccps_l_ack_ratio)
+ if (val == ccid2_ack_ratio_next(sk))
return;
ccid2_pr_debug("changing local ack ratio to %u\n", val);
- dp->dccps_l_ack_ratio = val;
dccp_feat_signal_nn_change(sk, DCCPF_ACK_RATIO, val);
}
@@ -117,11 +119,9 @@ static void ccid2_change_l_seq_window(struct sock *sk, u64 val)
val = DCCPF_SEQ_WMIN;
if (val > DCCPF_SEQ_WMAX)
val = DCCPF_SEQ_WMAX;
- if (val == dp->dccps_l_seq_win)
- return;
- dp->dccps_l_seq_win = val;
- dccp_feat_signal_nn_change(sk, DCCPF_SEQUENCE_WINDOW, val);
+ if (val != dp->dccps_l_seq_win)
+ dccp_feat_signal_nn_change(sk, DCCPF_SEQUENCE_WINDOW, val);
}
static void ccid2_hc_tx_rto_expire(unsigned long data)
@@ -203,6 +203,10 @@ static void ccid2_cwnd_application_limited(struct sock *sk, const u32 now)
}
hc->tx_cwnd_used = 0;
hc->tx_cwnd_stamp = now;
+
+ /* Avoid spurious timeouts resulting from Ack Ratio > cwnd */
+ if (ccid2_ack_ratio_next(sk) > hc->tx_cwnd)
+ ccid2_change_l_ack_ratio(sk, hc->tx_cwnd/2 ? : 1U);
}
/* This borrows the code of tcp_cwnd_restart() */
@@ -221,6 +225,10 @@ static void ccid2_cwnd_restart(struct sock *sk, const u32 now)
hc->tx_cwnd_stamp = now;
hc->tx_cwnd_used = 0;
+
+ /* Avoid spurious timeouts resulting from Ack Ratio > cwnd */
+ if (ccid2_ack_ratio_next(sk) > hc->tx_cwnd)
+ ccid2_change_l_ack_ratio(sk, hc->tx_cwnd/2 ? : 1U);
}
static void ccid2_hc_tx_packet_sent(struct sock *sk, unsigned int len)
@@ -491,8 +531,8 @@ static void ccid2_congestion_event(struct sock *sk, struct ccid2_seq *seqp)
hc->tx_ssthresh = max(hc->tx_cwnd, 2U);
/* Avoid spurious timeouts resulting from Ack Ratio > cwnd */
- if (dccp_sk(sk)->dccps_l_ack_ratio > hc->tx_cwnd)
- ccid2_change_l_ack_ratio(sk, hc->tx_cwnd);
+ if (ccid2_ack_ratio_next(sk) > hc->tx_cwnd)
+ ccid2_change_l_ack_ratio(sk, hc->tx_cwnd/2 ? : 1U);
}
static int ccid2_hc_tx_parse_options(struct sock *sk, u8 packet_type,
diff --git a/net/dccp/dccp.h b/net/dccp/dccp.h
index 26cba68..e8cf5a7 100644
--- a/net/dccp/dccp.h
+++ b/net/dccp/dccp.h
@@ -491,6 +491,7 @@ static inline int dccp_ack_pending(const struct sock *sk)
}
extern int dccp_feat_signal_nn_change(struct sock *sk, u8 feat, u64 nn_val);
+extern __u64 dccp_feat_get_nn_next_val(struct sock *sk, u8 feat);
extern int dccp_feat_finalise_settings(struct dccp_sock *dp);
extern int dccp_feat_server_ccid_dependencies(struct dccp_request_sock *dreq);
extern int dccp_feat_insert_opts(struct dccp_sock*, struct dccp_request_sock*,
diff --git a/net/dccp/feat.c b/net/dccp/feat.c
index 9a8c2ba..bc26f33 100644
--- a/net/dccp/feat.c
+++ b/net/dccp/feat.c
@@ -775,12 +775,7 @@ int dccp_feat_register_sp(struct sock *sk, u8 feat, u8 is_local,
* @sk: DCCP socket of an established connection
* @feat: NN feature number from %dccp_feature_numbers
* @nn_val: the new value to use
- * This function is used to communicate NN updates out-of-band. The difference
- * to feature negotiation during connection setup is that values are activated
- * immediately after validation, i.e. we don't wait for the Confirm: either the
- * value is accepted by the peer (and then the waiting is futile), or it is not
- * (Reset or empty Confirm). We don't accept empty Confirms - transmitted values
- * are validated, and the peer "MUST accept any valid value" (RFC 4340, 6.3.2).
+ * This function is used to communicate NN updates out-of-band.
*/
int dccp_feat_signal_nn_change(struct sock *sk, u8 feat, u64 nn_val)
{
@@ -805,9 +800,6 @@ int dccp_feat_signal_nn_change(struct sock *sk, u8 feat, u64 nn_val)
dccp_feat_list_pop(entry);
}
- if (dccp_feat_activate(sk, feat, 1, &fval))
- return -EADV;
-
inet_csk_schedule_ack(sk);
return dccp_feat_push_change(fn, feat, 1, 0, &fval);
}
@@ -1356,6 +1348,9 @@ static u8 dccp_feat_handle_nn_established(struct sock *sk, u8 mandatory, u8 opt,
if (fval.nn != entry->val.nn)
return 0;
+ /* Only activate after receiving the Confirm option (6.6.1). */
+ dccp_feat_activate(sk, feat, local, &fval);
+
/* It has been confirmed - so remove the entry */
dccp_feat_list_pop(entry);
@@ -1374,6 +1369,39 @@ fast_path_failed:
: DCCP_RESET_CODE_OPTION_ERROR;
}
+/*
+* dccp_feat_get_nn_next_val - Get the currently negotiating value of an NN
+* feature. If the feature isn't being currently
+* negotiated, return it's current value.
+* @sk: DCCP socket of an established connection
+* @feat: NN feature number from %dccp_feature_numbers
+*/
+__u64 dccp_feat_get_nn_next_val(struct sock *sk, u8 feat)
+{
+ struct list_head *fn = &dccp_sk(sk)->dccps_featneg;
+ struct dccp_feat_entry *entry;
+ u8 type = dccp_feat_type(feat);
+
+ if (type == FEAT_UNKNOWN || type != FEAT_NN)
+ return 0;
+
+ entry = dccp_feat_list_lookup(fn, feat, 1);
+ if (entry != NULL)
+ return entry->val.nn;
+
+ switch (feat) {
+ case DCCPF_ACK_RATIO:
+ return dccp_sk(sk)->dccps_l_ack_ratio;
+ case DCCPF_SEQUENCE_WINDOW:
+ return dccp_sk(sk)->dccps_l_seq_win;
+ default:
+ dccp_pr_debug("Attempting to get value \
+ for unknown NN feature");
+ }
+return 0;
+}
+EXPORT_SYMBOL_GPL(dccp_feat_get_nn_next_val);
+
/**
* dccp_feat_parse_options - Process Feature-Negotiation Options
* @sk: for general use and used by the client during connection setup
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]
^ permalink raw reply related
* [PATCH] ipv4: Cache source address in nexthop entries.
From: David Miller @ 2011-03-08 4:58 UTC (permalink / raw)
To: netdev
When doing output route lookups, we have to select the source address
if the user has not specified an explicit one.
First, if the route has an explicit preferred source address
specified, then we use that.
Otherwise we search the route's outgoing interface for a suitable
address.
This search can be precomputed and cached at route insertion time.
The only missing part is that we have to refresh this precomputed
value any time addresses are added or removed from the interface, and
this is accomplished by fib_update_nh_saddrs().
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/net/ip_fib.h | 7 +++++--
net/ipv4/fib_frontend.c | 2 ++
net/ipv4/fib_semantics.c | 31 ++++++++++++++++++++++++-------
3 files changed, 31 insertions(+), 9 deletions(-)
diff --git a/include/net/ip_fib.h b/include/net/ip_fib.h
index 523a170..0e14083 100644
--- a/include/net/ip_fib.h
+++ b/include/net/ip_fib.h
@@ -60,6 +60,7 @@ struct fib_nh {
#endif
int nh_oif;
__be32 nh_gw;
+ __be32 nh_saddr;
};
/*
@@ -139,11 +140,13 @@ struct fib_result_nl {
#endif /* CONFIG_IP_ROUTE_MULTIPATH */
-#define FIB_RES_PREFSRC(res) ((res).fi->fib_prefsrc ? : __fib_res_prefsrc(&res))
+#define FIB_RES_SADDR(res) (FIB_RES_NH(res).nh_saddr)
#define FIB_RES_GW(res) (FIB_RES_NH(res).nh_gw)
#define FIB_RES_DEV(res) (FIB_RES_NH(res).nh_dev)
#define FIB_RES_OIF(res) (FIB_RES_NH(res).nh_oif)
+#define FIB_RES_PREFSRC(res) ((res).fi->fib_prefsrc ? : FIB_RES_SADDR(res))
+
struct fib_table {
struct hlist_node tb_hlist;
u32 tb_id;
@@ -224,8 +227,8 @@ extern void fib_select_default(struct fib_result *res);
extern int ip_fib_check_default(__be32 gw, struct net_device *dev);
extern int fib_sync_down_dev(struct net_device *dev, int force);
extern int fib_sync_down_addr(struct net *net, __be32 local);
+extern void fib_update_nh_saddrs(struct net_device *dev);
extern int fib_sync_up(struct net_device *dev);
-extern __be32 __fib_res_prefsrc(struct fib_result *res);
extern void fib_select_multipath(const struct flowi *flp, struct fib_result *res);
/* Exported by fib_trie.c */
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index ad0778a..1d2233c 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -890,10 +890,12 @@ static int fib_inetaddr_event(struct notifier_block *this, unsigned long event,
#ifdef CONFIG_IP_ROUTE_MULTIPATH
fib_sync_up(dev);
#endif
+ fib_update_nh_saddrs(dev);
rt_cache_flush(dev_net(dev), -1);
break;
case NETDEV_DOWN:
fib_del_ifaddr(ifa);
+ fib_update_nh_saddrs(dev);
if (ifa->ifa_dev->ifa_list == NULL) {
/* Last address was deleted from this interface.
* Disable IP.
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index 6349a21..952c737 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -853,6 +853,12 @@ struct fib_info *fib_create_info(struct fib_config *cfg)
goto err_inval;
}
+ change_nexthops(fi) {
+ nexthop_nh->nh_saddr = inet_select_addr(nexthop_nh->nh_dev,
+ nexthop_nh->nh_gw,
+ nexthop_nh->nh_scope);
+ } endfor_nexthops(fi)
+
link_it:
ofi = fib_find_info(fi);
if (ofi) {
@@ -898,13 +904,6 @@ failure:
return ERR_PTR(err);
}
-/* Find appropriate source address to this destination */
-
-__be32 __fib_res_prefsrc(struct fib_result *res)
-{
- return inet_select_addr(FIB_RES_DEV(*res), FIB_RES_GW(*res), res->scope);
-}
-
int fib_dump_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
u32 tb_id, u8 type, u8 scope, __be32 dst, int dst_len, u8 tos,
struct fib_info *fi, unsigned int flags)
@@ -1128,6 +1127,24 @@ out:
return;
}
+void fib_update_nh_saddrs(struct net_device *dev)
+{
+ struct hlist_head *head;
+ struct hlist_node *node;
+ struct fib_nh *nh;
+ unsigned int hash;
+
+ hash = fib_devindex_hashfn(dev->ifindex);
+ head = &fib_info_devhash[hash];
+ hlist_for_each_entry(nh, node, head, nh_hash) {
+ if (nh->nh_dev != dev)
+ continue;
+ nh->nh_saddr = inet_select_addr(nh->nh_dev,
+ nh->nh_gw,
+ nh->nh_scope);
+ }
+}
+
#ifdef CONFIG_IP_ROUTE_MULTIPATH
/*
--
1.7.4.1
^ permalink raw reply related
* Re: bonding can't change to another slave if you ifdown the active slave
From: Weiping Pan @ 2011-03-08 6:52 UTC (permalink / raw)
To: netdev
On 03/08/2011 05:15 AM, Nicolas de Pesloüan wrote:
> Le 07/03/2011 04:13, Weiping Pan a écrit :
>> On 03/05/2011 09:49 PM, Nicolas de Pesloüan wrote:
>>> Le 05/03/2011 03:53, Andy Gospodarek a écrit :
>>>> On Fri, Mar 04, 2011 at 10:15:17AM +0800, Weiping Pan wrote:
>>>>> Hi,
>>>>>
>>>>> I'm doing some Linux bonding driver test, and I find a problem in
>>>>> balance-rr mode.
>>>>> That's it can't change to another slave if you ifdown the active
>>>>> slave.
>>>>> Any comments are warmly welcomed!
>>>>>
>>>>> regards
>>>>> Weiping Pan
>>>>>
>>>>> My host is Fedora 14, and I install VirtualBox (4.0.2), and enable 4
>>>>> nics for the guest system.
>>>>
>>>> Does this mean you are passing 4 NICs from your host to your guest
>>>> (maybe via direct pci-device assignment to the guest) or are you
>>>> creating 4 virtual devices on the host that are in a bridge group
>>>> on the
>>>> host?
>>>
>>> VirtualBox does not allow assignment of pci-device to the guest. The
>>> network interfaces on the guest are pure virtual one, with several
>>> modes available. In order to help you trouble shooting this problem,
>>> we need to know the mode form each of the virtual interfaces. Possible
>>> modes are NAT, bridged, internal-network, and host-only-network.
>>>
>>> Please provide the output of the following command:
>>>
>>> VBoxManage showvminfo <your-vm-uuid> | grep ^NIC
>>>
>>> To display your vm uuid, use the following command:
>>>
>>> VBoxManage list vms
>> [root@localhost ~]# VBoxManage showvminfo
>> 67b83c47-0ee2-46bc-b0ff-e0eb43edc1c2 |grep ^NIC
>> NIC 1: MAC: 0800270481A8, Attachment: Bridged Interface 'eth0', Cable
>> connected: on, Trace: off (file: none), Type: 82540EM, Reported speed: 0
>> Mbps, Boot priority: 0
>> NIC 2: MAC: 08002778F641, Attachment: Bridged Interface 'eth0', Cable
>> connected: on, Trace: off (file: none), Type: 82540EM, Reported speed: 0
>> Mbps, Boot priority: 0
>> NIC 3: MAC: 080027C408BA, Attachment: Bridged Interface 'eth0', Cable
>> connected: on, Trace: off (file: none), Type: 82540EM, Reported speed: 0
>> Mbps, Boot priority: 0
>> NIC 4: MAC: 080027DB339A, Attachment: Bridged Interface 'eth0', Cable
>> connected: on, Trace: off (file: none), Type: 82540EM, Reported speed: 0
>> Mbps, Boot priority: 0
>> NIC 5: disabled
>> NIC 6: disabled
>> NIC 7: disabled
>> NIC 8: disabled
>>
>> And when guest starts, i find that:
>> NIC 1: eth7
>> NIC 2: eth6
>> NIC 3: eth9
>> NIC 4: eth8
>
> Would you mind testing with "Host-only Interface 'vboxnet0'", instead
> of "Bridged Interface 'eth0'"?
>
> All the bonding tests I do use this setup and the link failure
> detection work well.
>
> Nicolas.
ok, I use "Host-only mode" and get
an conclusion, that if the first enslaved nic is pulled out, bonding
can't handle well.
First test.
I first enslave eth6, then pull it out, bonding doesn't work.
on host,
[root@localhost ~]# VBoxManage -v
4.0.2r69518
[root@localhost ~]# VBoxManage showvminfo
67b83c47-0ee2-46bc-b0ff-e0eb43edc1c2|grep ^NIC
NIC 1: MAC: 0800270481A8, Attachment: Host-only Interface
'vboxnet0', Cable connected: on, Trace: off (file: none), Type: 82540EM,
Reported speed: 0 Mbps, Boot priority: 0
NIC 2: MAC: 08002778F641, Attachment: Host-only Interface
'vboxnet0', Cable connected: on, Trace: off (file: none), Type: 82540EM,
Reported speed: 0 Mbps, Boot priority: 0
NIC 3: MAC: 080027C408BA, Attachment: Host-only Interface
'vboxnet0', Cable connected: on, Trace: off (file: none), Type: 82540EM,
Reported speed: 0 Mbps, Boot priority: 0
NIC 4: MAC: 080027DB339A, Attachment: Host-only Interface
'vboxnet0', Cable connected: on, Trace: off (file: none), Type: 82540EM,
Reported speed: 0 Mbps, Boot priority: 0
NIC 5: disabled
NIC 6: disabled
NIC 7: disabled
NIC 8: disabled
[root@localhost ~]# ifconfig vboxnet0
vboxnet0 Link encap:Ethernet HWaddr 0A:00:27:00:00:00
inet addr:192.168.56.1 Bcast:192.168.56.255 Mask:255.255.255.0
inet6 addr: fe80::800:27ff:fe00:0/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:268 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 b) TX bytes:52456 (51.2 KiB)
restart guest,
[root@localhost ~]# uname -a
Linux localhost.localdomain 2.6.35.11-83.fc14.i686 #1 SMP Mon Feb 7
07:04:18 UTC 2011 i686 i686 i386 GNU/Linux
[root@localhost ~]# ifconfig
eth6 Link encap:Ethernet HWaddr 08:00:27:78:F6:41
inet addr:192.168.56.101 Bcast:192.168.56.255
Mask:255.255.255.0
inet6 addr: fe80::a00:27ff:fe78:f641/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:14 errors:0 dropped:0 overruns:0 frame:0
TX packets:8 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:6772 (6.6 KiB) TX bytes:1152 (1.1 KiB)
eth7 Link encap:Ethernet HWaddr 08:00:27:04:81:A8
inet addr:192.168.56.102 Bcast:192.168.56.255
Mask:255.255.255.0
inet6 addr: fe80::a00:27ff:fe04:81a8/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:14 errors:0 dropped:0 overruns:0 frame:0
TX packets:8 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:6772 (6.6 KiB) TX bytes:1152 (1.1 KiB)
eth8 Link encap:Ethernet HWaddr 08:00:27:DB:33:9A
inet addr:192.168.56.104 Bcast:192.168.56.255
Mask:255.255.255.0
inet6 addr: fe80::a00:27ff:fedb:339a/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:63 errors:0 dropped:0 overruns:0 frame:0
TX packets:8 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:19456 (19.0 KiB) TX bytes:1152 (1.1 KiB)
eth9 Link encap:Ethernet HWaddr 08:00:27:C4:08:BA
inet addr:192.168.56.103 Bcast:192.168.56.255
Mask:255.255.255.0
inet6 addr: fe80::a00:27ff:fec4:8ba/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:63 errors:0 dropped:0 overruns:0 frame:0
TX packets:8 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:19456 (19.0 KiB) TX bytes:1152 (1.1 KiB)
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:16 errors:0 dropped:0 overruns:0 frame:0
TX packets:16 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:880 (880.0 b) TX bytes:880 (880.0 b)
so
NIC 1: eth7
NIC 2: eth6
NIC 3: eth9
NIC 4: eth8
on guest,
[root@localhost ~]# for i in eth{6..9}; do ifconfig $i down; done
[root@localhost ~]# modprobe bonding mode=0 miimon=100
[root@localhost ~]# ifconfig bond0 192.168.56.2 netmask 255.255.255.0 up
[root@localhost ~]# ifenslave bond0 eth6
[root@localhost ~]# ifenslave bond0 eth7
[root@localhost ~]# ping 192.168.56.1 -c 5
PING 192.168.56.1 (192.168.56.1) 56(84) bytes of data.
64 bytes from 192.168.56.1: icmp_req=1 ttl=64 time=1.83 ms
64 bytes from 192.168.56.1: icmp_req=2 ttl=64 time=0.149 ms
64 bytes from 192.168.56.1: icmp_req=3 ttl=64 time=0.204 ms
64 bytes from 192.168.56.1: icmp_req=4 ttl=64 time=0.294 ms
64 bytes from 192.168.56.1: icmp_req=5 ttl=64 time=0.412 ms
--- 192.168.56.1 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4003ms
rtt min/avg/max/mdev = 0.149/0.578/1.832/0.633 ms
[root@localhost ~]# arp
Address HWtype HWaddress Flags
Mask Iface
192.168.56.1 ether 0a:00:27:00:00:00
C bond0
on host,
[root@localhost ~]# arp
Address HWtype HWaddress Flags
Mask Iface
192.168.56.2 ether 08:00:27:78:f6:41
C vboxnet0
corerouter.nay.redhat.c ether 00:1d:45:20:d5:ff
C eth0
[root@localhost ~]# VBoxManage controlvm
67b83c47-0ee2-46bc-b0ff-e0eb43edc1c2 setlinkstate2 off
[root@localhost ~]# ping 192.168.56.1 -c 5
PING 192.168.56.1 (192.168.56.1) 56(84) bytes of data.
From 192.168.56.2 icmp_seq=2 Destination Host Unreachable
From 192.168.56.2 icmp_seq=3 Destination Host Unreachable
From 192.168.56.2 icmp_seq=4 Destination Host Unreachable
From 192.168.56.2 icmp_seq=5 Destination Host Unreachable
--- 192.168.56.1 ping statistics ---
5 packets transmitted, 0 received, +4 errors, 100% packet loss, time 4001ms
pipe 3
[root@localhost ~]# dmesg
[ 513.093249] e1000: eth6 NIC Link is Down
[ 513.129123] bonding: bond0: link status definitely down for interface
eth6, disabling it
[root@localhost ~]# ip route show
192.168.56.0/24 dev bond0 proto kernel scope link src 192.168.56.2
192.168.56.0/24 dev eth7 proto kernel scope link src 192.168.56.101
metric 1
[root@localhost ~]# ip neigh show
192.168.56.1 dev bond0 FAILED
on host,
[root@localhost ~]# ip route show
192.168.1.0/24 dev eth0 proto kernel scope link src 192.168.1.100
192.168.56.0/24 dev vboxnet0 proto kernel scope link src 192.168.56.1
10.66.64.0/23 dev eth0 proto kernel scope link src 10.66.65.228
metric 1
default via 10.66.65.254 dev eth0 proto static
[root@localhost ~]# ip neigh show
192.168.56.2 dev vboxnet0 lladdr 08:00:27:78:f6:41 STALE
10.66.65.254 dev eth0 lladdr 00:1d:45:20:d5:ff REACHABLE
Second test
I first enslave eth6, then pull eth7 out, bonding works well.
on guest,
[root@localhost ~]# modprobe bonding mode=0 miimon=100
[root@localhost ~]# ifconfig bond0 192.168.56.2 netmask 255.255.255.0 up
[root@localhost ~]# ifenslave bond0 eth6
[root@localhost ~]# ifenslave bond0 eth7
[root@localhost ~]# ping 192.168.56.1 -c 5
PING 192.168.56.1 (192.168.56.1) 56(84) bytes of data.
64 bytes from 192.168.56.1: icmp_req=1 ttl=64 time=0.902 ms
64 bytes from 192.168.56.1: icmp_req=2 ttl=64 time=0.260 ms
64 bytes from 192.168.56.1: icmp_req=3 ttl=64 time=0.237 ms
64 bytes from 192.168.56.1: icmp_req=4 ttl=64 time=0.335 ms
64 bytes from 192.168.56.1: icmp_req=5 ttl=64 time=0.170 ms
--- 192.168.56.1 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4001ms
rtt min/avg/max/mdev = 0.170/0.380/0.902/0.267 ms
[root@localhost ~]# dmesg
[ 1162.524825] bonding: MII link monitoring set to 100 ms
[ 1165.845586] ADDRCONF(NETDEV_UP): bond0: link is not ready
[ 1174.505912] e1000: eth6 NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: RX
[ 1174.512372] bonding: bond0: enslaving eth6 as an active interface
with an up link.
[ 1174.512897] ADDRCONF(NETDEV_CHANGE): bond0: link becomes ready
[ 1178.857322] ICMPv6 NA: someone advertises our address
fe80:0000:0000:0000:0a00:27ff:fe78:f641 on bond0!
[ 1178.858649] e1000: eth7 NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: RX
[ 1178.861391] bonding: bond0: enslaving eth7 as an active interface
with an up link.
[ 1184.682110] bond0: no IPv6 routers present
on host,
[root@localhost ~]# VBoxManage controlvm
67b83c47-0ee2-46bc-b0ff-e0eb43edc1c2 setlinkstate1 off
on guest,
[root@localhost ~]# ping 192.168.56.1 -c 5
PING 192.168.56.1 (192.168.56.1) 56(84) bytes of data.
64 bytes from 192.168.56.1: icmp_req=1 ttl=64 time=0.599 ms
64 bytes from 192.168.56.1: icmp_req=2 ttl=64 time=0.150 ms
64 bytes from 192.168.56.1: icmp_req=3 ttl=64 time=0.224 ms
64 bytes from 192.168.56.1: icmp_req=4 ttl=64 time=0.154 ms
64 bytes from 192.168.56.1: icmp_req=5 ttl=64 time=0.189 ms
--- 192.168.56.1 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4000ms
rtt min/avg/max/mdev = 0.150/0.263/0.599/0.170 ms
[root@localhost ~]# dmesg
[ 1281.421231] e1000: eth7 NIC Link is Down
[ 1281.492178] bonding: bond0: link status definitely down for interface
eth7, disabling it
many thanks
Weiping Pan
^ permalink raw reply
* Re: ANNOUNCE: debloat-testing kernel git tree
From: Pavel Machek @ 2011-03-08 6:58 UTC (permalink / raw)
To: Tianji Li
Cc: rick.jones2, Dave T?ht, sedat.dilek, netdev, linux-wireless,
bloat-devel, linux-kernel
In-Reply-To: <4D6FEFD4.9010608@nuim.ie>
On Thu 2011-03-03 13:45:24, Tianji Li wrote:
>
>
> On 03/03/2011 12:16 PM, Rick Jones wrote:
> >>For wireless routers and cable home gateways especially, this research
> >>shows that the total un-managed buffers in your system should be less
> >>than 32.
> >
> >Would it be a good thing to start describing these queues not so much in
> >terms of packets but in terms of delay (or bandwidth X delay)?
> >
>
> The unit of bandwidth is something like Mbps, that of delay can be
> second, so bandwidth X delay --> Mb, which is the unit of packet
> size. So both packets and delay should have the same effect for
> sizing buffers.
Bandwidth varies greatly with time. On cellphone, you can go from
5KB/sec GPRS to 300KB/sec HSDPA and back...
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
^ permalink raw reply
* Re: [RFC] [PATCH 06/11] esp4: Add support for IPsec extended sequence numbers
From: Steffen Klassert @ 2011-03-08 7:04 UTC (permalink / raw)
To: Herbert Xu
Cc: David Miller, Andreas Gruenbacher, Alex Badea, netdev,
linux-crypto
In-Reply-To: <20101202072947.GA22998@gondor.apana.org.au>
Sorry for the huge delay...
On Thu, Dec 02, 2010 at 03:29:47PM +0800, Herbert Xu wrote:
> On Mon, Nov 22, 2010 at 11:30:14AM +0100, Steffen Klassert wrote:
> >
> > @@ -205,11 +228,18 @@ static int esp_output(struct xfrm_state *x, struct sk_buff *skb)
> > skb_to_sgvec(skb, sg,
> > esph->enc_data + crypto_aead_ivsize(aead) - skb->data,
> > clen + alen);
> > - sg_init_one(asg, esph, sizeof(*esph));
> > +
> > + if ((x->props.flags & XFRM_STATE_ESN)) {
> > + sg_init_table(asg, 2);
> > + sg_set_buf(asg, esph, sizeof(*esph));
> > + *seqhi = htonl(XFRM_SKB_CB(skb)->seq.output.hi);
> > + sg_set_buf(asg + 1, seqhi, seqhilen);
> > + } else
> > + sg_init_one(asg, esph, sizeof(*esph));
>
> I think this is wrong for AEAD algorithms. You want the sequence
> number in network byte order for them so the high bits need to be
> inserted into the middle of the ESP header.
>
Yes, indeed.
> The other problem is that you're currently requiring the authencesn
> user to provide two SG entries which is fine for now. However,
> since this might be exported to user-space in future, authenecesn
> shouldn't really rely on that, or at least it shouldn't BUG.
>
> So one solution is to do it based on bytes in authencesn. That is,
> your associated input should always be 12 bytes long, and then you
> simply construct a new SG list for your actual processing with the
> middle 4 bytes taken out.
>
> For IPsec it could just provide an SG list with three entries,
> of 4 bytes each.
Ok, I've updated the patchset in this regard.
>
> Of course for simplicity, you could require this to be the case in
> authencesn and return -EINVAL (not BUG :) if it's not the case.
>
Doing BUG was a leftover from debugging the thing. On debugging it is
sometimes better to pull the emergency brake as soon as something
unexpected happens. I replaced BUG with return -EINVAL now.
I'll resend the patchset for a second round of review.
^ 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