From 3da0ae6298fa10d4f5e2855ae400b2470bc6693d Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 25 Mar 2010 20:32:39 -0700 Subject: isdn: Add netdev to lists in MAINTAINERS entry. Signed-off-by: David S. Miller diff --git a/MAINTAINERS b/MAINTAINERS index 5c99bd6..9ff6341 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3083,6 +3083,7 @@ F: include/scsi/*iscsi* ISDN SUBSYSTEM M: Karsten Keil L: isdn4linux@listserv.isdn4linux.de (subscribers-only) +L: netdev@vger.kernel.org W: http://www.isdn4linux.de T: git git://git.kernel.org/pub/scm/linux/kernel/git/kkeil/isdn-2.6.git S: Maintained -- cgit v0.10.2 From 4300e8c7f64d95a80ffa7d98d98738f41546bc30 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 26 Mar 2010 10:23:30 -0700 Subject: Revert "r8169: enable 64-bit DMA by default for PCI Express devices (v2)" This reverts commit 353176888386d9025062a12dcec08d49af10cf2c. People are reporting problems due to this change and there is no anticipation that the cause will be tracked down any time soon. We can try next time to selectively re-enable this based upon chip type, or have a black list of some sort. Signed-off-by: David S. Miller diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c index 9d3ebf3..b93fd23 100644 --- a/drivers/net/r8169.c +++ b/drivers/net/r8169.c @@ -187,7 +187,7 @@ static DEFINE_PCI_DEVICE_TABLE(rtl8169_pci_tbl) = { MODULE_DEVICE_TABLE(pci, rtl8169_pci_tbl); static int rx_copybreak = 200; -static int use_dac = -1; +static int use_dac; static struct { u32 msg_enable; } debug = { -1 }; @@ -511,8 +511,7 @@ MODULE_DESCRIPTION("RealTek RTL-8169 Gigabit Ethernet driver"); module_param(rx_copybreak, int, 0); MODULE_PARM_DESC(rx_copybreak, "Copy breakpoint for copy-only-tiny-frames"); module_param(use_dac, int, 0); -MODULE_PARM_DESC(use_dac, "Enable PCI DAC. -1 defaults on for PCI Express only." -" Unsafe on 32 bit PCI slot."); +MODULE_PARM_DESC(use_dac, "Enable PCI DAC. Unsafe on 32 bit PCI slot."); module_param_named(debug, debug.msg_enable, int, 0); MODULE_PARM_DESC(debug, "Debug verbosity level (0=none, ..., 16=all)"); MODULE_LICENSE("GPL"); @@ -2974,7 +2973,6 @@ rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) void __iomem *ioaddr; unsigned int i; int rc; - int this_use_dac = use_dac; if (netif_msg_drv(&debug)) { printk(KERN_INFO "%s Gigabit Ethernet driver %s loaded\n", @@ -3040,17 +3038,8 @@ rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) tp->cp_cmd = PCIMulRW | RxChkSum; - tp->pcie_cap = pci_find_capability(pdev, PCI_CAP_ID_EXP); - if (!tp->pcie_cap) - netif_info(tp, probe, dev, "no PCI Express capability\n"); - - if (this_use_dac < 0) - this_use_dac = tp->pcie_cap != 0; - if ((sizeof(dma_addr_t) > 4) && - this_use_dac && - !pci_set_dma_mask(pdev, DMA_BIT_MASK(64))) { - netif_info(tp, probe, dev, "using 64-bit DMA\n"); + !pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) && use_dac) { tp->cp_cmd |= PCIDAC; dev->features |= NETIF_F_HIGHDMA; } else { @@ -3069,6 +3058,10 @@ rtl8169_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) goto err_out_free_res_4; } + tp->pcie_cap = pci_find_capability(pdev, PCI_CAP_ID_EXP); + if (!tp->pcie_cap) + netif_info(tp, probe, dev, "no PCI Express capability\n"); + RTL_W16(IntrMask, 0x0000); /* Soft reset the chip. */ -- cgit v0.10.2 From bb2792e0383793d5135ba777e93f0a918371394b Mon Sep 17 00:00:00 2001 From: Amit Kumar Salecha Date: Fri, 26 Mar 2010 00:30:07 +0000 Subject: netxen: fix bios version calculation Bios sub version from unified fw image is calculated incorrect. Signed-off-by: Amit Kumar Salecha Signed-off-by: David S. Miller diff --git a/drivers/net/netxen/netxen_nic_init.c b/drivers/net/netxen/netxen_nic_init.c index 1c63610e..7eb925a 100644 --- a/drivers/net/netxen/netxen_nic_init.c +++ b/drivers/net/netxen/netxen_nic_init.c @@ -761,7 +761,7 @@ nx_get_bios_version(struct netxen_adapter *adapter) if (adapter->fw_type == NX_UNIFIED_ROMIMAGE) { bios_ver = cpu_to_le32(*((u32 *) (&fw->data[prd_off]) + NX_UNI_BIOS_VERSION_OFF)); - return (bios_ver << 24) + ((bios_ver >> 8) & 0xff00) + + return (bios_ver << 16) + ((bios_ver >> 8) & 0xff00) + (bios_ver >> 24); } else return cpu_to_le32(*(u32 *)&fw->data[NX_BIOS_VERSION_OFFSET]); -- cgit v0.10.2 From 77c553900c58c3e4f475e233ad4ff6aeb282deb4 Mon Sep 17 00:00:00 2001 From: Amit Kumar Salecha Date: Fri, 26 Mar 2010 00:30:08 +0000 Subject: netxen: fix warning in ioaddr for NX3031 chip Signed-off-by: Amit Kumar Salecha crb_intr_mask/crb_sts_consumer is predefined for NX2031 not for NX3031. For NX3031, these values get defined in rx context creation. Signed-off-by: David S. Miller diff --git a/drivers/net/netxen/netxen_nic_ctx.c b/drivers/net/netxen/netxen_nic_ctx.c index 2a8ef5f..f26e547 100644 --- a/drivers/net/netxen/netxen_nic_ctx.c +++ b/drivers/net/netxen/netxen_nic_ctx.c @@ -669,13 +669,15 @@ int netxen_alloc_hw_resources(struct netxen_adapter *adapter) } sds_ring->desc_head = (struct status_desc *)addr; - sds_ring->crb_sts_consumer = - netxen_get_ioaddr(adapter, - recv_crb_registers[port].crb_sts_consumer[ring]); + if (NX_IS_REVISION_P2(adapter->ahw.revision_id)) { + sds_ring->crb_sts_consumer = + netxen_get_ioaddr(adapter, + recv_crb_registers[port].crb_sts_consumer[ring]); - sds_ring->crb_intr_mask = - netxen_get_ioaddr(adapter, - recv_crb_registers[port].sw_int_mask[ring]); + sds_ring->crb_intr_mask = + netxen_get_ioaddr(adapter, + recv_crb_registers[port].sw_int_mask[ring]); + } } -- cgit v0.10.2 From afbe5cd6c40e0f20fa8832d17fa44ae605591ce1 Mon Sep 17 00:00:00 2001 From: Amit Kumar Salecha Date: Fri, 26 Mar 2010 00:30:09 +0000 Subject: netxen: added sanity check for pci map Signed-off-by: Amit Kumar Salecha Return value of ioremap is not checked, NULL check added. Signed-off-by: Amit Kumar Salecha Signed-off-by: David S. Miller diff --git a/drivers/net/netxen/netxen_nic_main.c b/drivers/net/netxen/netxen_nic_main.c index 9a7a0f3..01808b2 100644 --- a/drivers/net/netxen/netxen_nic_main.c +++ b/drivers/net/netxen/netxen_nic_main.c @@ -604,16 +604,14 @@ netxen_cleanup_pci_map(struct netxen_adapter *adapter) static int netxen_setup_pci_map(struct netxen_adapter *adapter) { - void __iomem *mem_ptr0 = NULL; - void __iomem *mem_ptr1 = NULL; - void __iomem *mem_ptr2 = NULL; void __iomem *db_ptr = NULL; resource_size_t mem_base, db_base; - unsigned long mem_len, db_len = 0, pci_len0 = 0; + unsigned long mem_len, db_len = 0; struct pci_dev *pdev = adapter->pdev; int pci_func = adapter->ahw.pci_func; + struct netxen_hardware_context *ahw = &adapter->ahw; int err = 0; @@ -630,24 +628,40 @@ netxen_setup_pci_map(struct netxen_adapter *adapter) /* 128 Meg of memory */ if (mem_len == NETXEN_PCI_128MB_SIZE) { - mem_ptr0 = ioremap(mem_base, FIRST_PAGE_GROUP_SIZE); - mem_ptr1 = ioremap(mem_base + SECOND_PAGE_GROUP_START, + + ahw->pci_base0 = ioremap(mem_base, FIRST_PAGE_GROUP_SIZE); + ahw->pci_base1 = ioremap(mem_base + SECOND_PAGE_GROUP_START, SECOND_PAGE_GROUP_SIZE); - mem_ptr2 = ioremap(mem_base + THIRD_PAGE_GROUP_START, + ahw->pci_base2 = ioremap(mem_base + THIRD_PAGE_GROUP_START, THIRD_PAGE_GROUP_SIZE); - pci_len0 = FIRST_PAGE_GROUP_SIZE; + if (ahw->pci_base0 == NULL || ahw->pci_base1 == NULL || + ahw->pci_base2 == NULL) { + dev_err(&pdev->dev, "failed to map PCI bar 0\n"); + err = -EIO; + goto err_out; + } + + ahw->pci_len0 = FIRST_PAGE_GROUP_SIZE; + } else if (mem_len == NETXEN_PCI_32MB_SIZE) { - mem_ptr1 = ioremap(mem_base, SECOND_PAGE_GROUP_SIZE); - mem_ptr2 = ioremap(mem_base + THIRD_PAGE_GROUP_START - + + ahw->pci_base1 = ioremap(mem_base, SECOND_PAGE_GROUP_SIZE); + ahw->pci_base2 = ioremap(mem_base + THIRD_PAGE_GROUP_START - SECOND_PAGE_GROUP_START, THIRD_PAGE_GROUP_SIZE); + if (ahw->pci_base1 == NULL || ahw->pci_base2 == NULL) { + dev_err(&pdev->dev, "failed to map PCI bar 0\n"); + err = -EIO; + goto err_out; + } + } else if (mem_len == NETXEN_PCI_2MB_SIZE) { - mem_ptr0 = pci_ioremap_bar(pdev, 0); - if (mem_ptr0 == NULL) { + ahw->pci_base0 = pci_ioremap_bar(pdev, 0); + if (ahw->pci_base0 == NULL) { dev_err(&pdev->dev, "failed to map PCI bar 0\n"); return -EIO; } - pci_len0 = mem_len; + ahw->pci_len0 = mem_len; } else { return -EIO; } @@ -656,11 +670,6 @@ netxen_setup_pci_map(struct netxen_adapter *adapter) dev_info(&pdev->dev, "%dMB memory map\n", (int)(mem_len>>20)); - adapter->ahw.pci_base0 = mem_ptr0; - adapter->ahw.pci_len0 = pci_len0; - adapter->ahw.pci_base1 = mem_ptr1; - adapter->ahw.pci_base2 = mem_ptr2; - if (NX_IS_REVISION_P3P(adapter->ahw.revision_id)) { adapter->ahw.ocm_win_crb = netxen_get_ioaddr(adapter, NETXEN_PCIX_PS_REG(PCIX_OCM_WINDOW_REG(pci_func))); -- cgit v0.10.2 From 48c11a59c4c1d9926be34920d45da037516eb7b8 Mon Sep 17 00:00:00 2001 From: Amit Kumar Salecha Date: Fri, 26 Mar 2010 00:30:10 +0000 Subject: netxen: update version to 4.0.73 Signed-off-by: Amit Kumar Salecha Signed-off-by: David S. Miller diff --git a/drivers/net/netxen/netxen_nic.h b/drivers/net/netxen/netxen_nic.h index 144d2e8..0f70383 100644 --- a/drivers/net/netxen/netxen_nic.h +++ b/drivers/net/netxen/netxen_nic.h @@ -53,8 +53,8 @@ #define _NETXEN_NIC_LINUX_MAJOR 4 #define _NETXEN_NIC_LINUX_MINOR 0 -#define _NETXEN_NIC_LINUX_SUBVERSION 72 -#define NETXEN_NIC_LINUX_VERSIONID "4.0.72" +#define _NETXEN_NIC_LINUX_SUBVERSION 73 +#define NETXEN_NIC_LINUX_VERSIONID "4.0.73" #define NETXEN_VERSION_CODE(a, b, c) (((a) << 24) + ((b) << 16) + (c)) #define _major(v) (((v) >> 24) & 0xff) -- cgit v0.10.2 From 65deeed7b34bc5b8d3cbff495e8fa2ae7b563480 Mon Sep 17 00:00:00 2001 From: Greg Rose Date: Wed, 24 Mar 2010 09:35:42 +0000 Subject: ixgbevf: Fix signed/unsigned int error In the Tx mapping function if a DMA error occurred then the unwind of previously mapped sections would improperly check an unsigned int if it was less than zero. Changed the index variable to signed to avoid the error. Signed-off-by: Greg Rose Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbevf/ixgbevf_main.c b/drivers/net/ixgbevf/ixgbevf_main.c index d6cbd94..1bbbef3 100644 --- a/drivers/net/ixgbevf/ixgbevf_main.c +++ b/drivers/net/ixgbevf/ixgbevf_main.c @@ -2943,9 +2943,10 @@ static int ixgbevf_tx_map(struct ixgbevf_adapter *adapter, struct ixgbevf_tx_buffer *tx_buffer_info; unsigned int len; unsigned int total = skb->len; - unsigned int offset = 0, size, count = 0, i; + unsigned int offset = 0, size, count = 0; unsigned int nr_frags = skb_shinfo(skb)->nr_frags; unsigned int f; + int i; i = tx_ring->next_to_use; -- cgit v0.10.2 From 5809a1ae77721931ca7bd7aeacb37fdabe6f07c0 Mon Sep 17 00:00:00 2001 From: Greg Rose Date: Wed, 24 Mar 2010 09:36:08 +0000 Subject: ixgbe: In SR-IOV mode insert delay before bring the adapter up VFs running in guest VMs do not respond in as timely a manner to PF indication it is going down as they do when running in the host domain. If the adapter is in SR-IOV mode insert a two second delay to guarantee that all VFs have had time to respond to the PF reset. In any case resetting the PF while VFs are active should be discouraged but if it must be done then there will be a two second delay to help synchronize resets among the PF and all the VFs. Signed-off-by: Greg Rose Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c index d75c46f..d2cda9e 100644 --- a/drivers/net/ixgbe/ixgbe_main.c +++ b/drivers/net/ixgbe/ixgbe_main.c @@ -3056,6 +3056,14 @@ void ixgbe_reinit_locked(struct ixgbe_adapter *adapter) while (test_and_set_bit(__IXGBE_RESETTING, &adapter->state)) msleep(1); ixgbe_down(adapter); + /* + * If SR-IOV enabled then wait a bit before bringing the adapter + * back up to give the VFs time to respond to the reset. The + * two second wait is based upon the watchdog timer cycle in + * the VF driver. + */ + if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) + msleep(2000); ixgbe_up(adapter); clear_bit(__IXGBE_RESETTING, &adapter->state); } -- cgit v0.10.2 From 581d1aa777580c1c22169538ffb46676b13c408e Mon Sep 17 00:00:00 2001 From: Greg Rose Date: Wed, 24 Mar 2010 09:36:27 +0000 Subject: ixgbe: Change where clear_to_send_flag is reset to zero. The clear_to_send flag is being cleared before the call to ping all the VFs. It should be called after pinging all the VFs. Signed-off-by: Greg Rose Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c index d2cda9e..1066d53 100644 --- a/drivers/net/ixgbe/ixgbe_main.c +++ b/drivers/net/ixgbe/ixgbe_main.c @@ -3244,13 +3244,15 @@ void ixgbe_down(struct ixgbe_adapter *adapter) /* disable receive for all VFs and wait one second */ if (adapter->num_vfs) { - for (i = 0 ; i < adapter->num_vfs; i++) - adapter->vfinfo[i].clear_to_send = 0; - /* ping all the active vfs to let them know we are going down */ ixgbe_ping_all_vfs(adapter); + /* Disable all VFTE/VFRE TX/RX */ ixgbe_disable_tx_rx(adapter); + + /* Mark all the VFs as inactive */ + for (i = 0 ; i < adapter->num_vfs; i++) + adapter->vfinfo[i].clear_to_send = 0; } /* disable receives */ -- cgit v0.10.2 From e0fce6950b822aba7840d82c2d2018f1e1b8276b Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Wed, 24 Mar 2010 10:01:45 +0000 Subject: ixgbe: cleanup maximum number of tx queues In the last patch I missed an unecessary min_t comparison. This patch removes it, the path allocates at most 72 tx queues for 82599 and 24 for 82598 there is no need for this check. Additionally this sets MAX_[TX|RX]_QUEUES to 72. Which is used as the size for the tx/rx_ring arrays. There is no reason to have more tx_rings/rx_rings then num_tx_queues. Signed-off-by: John Fastabend Signed-off-by: Jeff Kirsher Acked-by: Eric Dumazet Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe.h b/drivers/net/ixgbe/ixgbe.h index 19e94ee..79c35ae 100644 --- a/drivers/net/ixgbe/ixgbe.h +++ b/drivers/net/ixgbe/ixgbe.h @@ -204,14 +204,17 @@ enum ixgbe_ring_f_enum { #define IXGBE_MAX_FDIR_INDICES 64 #ifdef IXGBE_FCOE #define IXGBE_MAX_FCOE_INDICES 8 +#define MAX_RX_QUEUES (IXGBE_MAX_FDIR_INDICES + IXGBE_MAX_FCOE_INDICES) +#define MAX_TX_QUEUES (IXGBE_MAX_FDIR_INDICES + IXGBE_MAX_FCOE_INDICES) +#else +#define MAX_RX_QUEUES IXGBE_MAX_FDIR_INDICES +#define MAX_TX_QUEUES IXGBE_MAX_FDIR_INDICES #endif /* IXGBE_FCOE */ struct ixgbe_ring_feature { int indices; int mask; } ____cacheline_internodealigned_in_smp; -#define MAX_RX_QUEUES 128 -#define MAX_TX_QUEUES 128 #define MAX_RX_PACKET_BUFFERS ((adapter->flags & IXGBE_FLAG_DCB_ENABLED) \ ? 8 : 1) diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c index 1066d53..208fb4a 100644 --- a/drivers/net/ixgbe/ixgbe_main.c +++ b/drivers/net/ixgbe/ixgbe_main.c @@ -6061,7 +6061,6 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev, indices += min_t(unsigned int, num_possible_cpus(), IXGBE_MAX_FCOE_INDICES); #endif - indices = min_t(unsigned int, indices, MAX_TX_QUEUES); netdev = alloc_etherdev_mq(sizeof(struct ixgbe_adapter), indices); if (!netdev) { err = -ENOMEM; -- cgit v0.10.2 From a7551b75fe47fb6fb70f679935845e741c5e0855 Mon Sep 17 00:00:00 2001 From: Robert Love Date: Wed, 24 Mar 2010 10:02:04 +0000 Subject: ixgbe: Don't allow user buffer count to exceed 256 If the user buffer count was 256 the shift would place a 1 in the offset region leading to errors. It also overwrites the uers buffer list. This patch makes sure that at most 256 user buffers are allowed for DDP and the buffer count is masked properly such that it doesn't overwrite the offset when shifting the bits. Signed-off-by: Robert Love Signed-off-by: Yi Zou Signed-off-by: Frank Zhang Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_fcoe.c b/drivers/net/ixgbe/ixgbe_fcoe.c index 700cfc0..e1978da 100644 --- a/drivers/net/ixgbe/ixgbe_fcoe.c +++ b/drivers/net/ixgbe/ixgbe_fcoe.c @@ -202,6 +202,15 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid, addr = sg_dma_address(sg); len = sg_dma_len(sg); while (len) { + /* max number of buffers allowed in one DDP context */ + if (j >= IXGBE_BUFFCNT_MAX) { + netif_err(adapter, drv, adapter->netdev, + "xid=%x:%d,%d,%d:addr=%llx " + "not enough descriptors\n", + xid, i, j, dmacount, (u64)addr); + goto out_noddp_free; + } + /* get the offset of length of current buffer */ thisoff = addr & ((dma_addr_t)bufflen - 1); thislen = min((bufflen - thisoff), len); @@ -227,20 +236,13 @@ int ixgbe_fcoe_ddp_get(struct net_device *netdev, u16 xid, len -= thislen; addr += thislen; j++; - /* max number of buffers allowed in one DDP context */ - if (j > IXGBE_BUFFCNT_MAX) { - DPRINTK(DRV, ERR, "xid=%x:%d,%d,%d:addr=%llx " - "not enough descriptors\n", - xid, i, j, dmacount, (u64)addr); - goto out_noddp_free; - } } } /* only the last buffer may have non-full bufflen */ lastsize = thisoff + thislen; fcbuff = (IXGBE_FCBUFF_4KB << IXGBE_FCBUFF_BUFFSIZE_SHIFT); - fcbuff |= (j << IXGBE_FCBUFF_BUFFCNT_SHIFT); + fcbuff |= ((j & 0xff) << IXGBE_FCBUFF_BUFFCNT_SHIFT); fcbuff |= (firstoff << IXGBE_FCBUFF_OFFSET_SHIFT); fcbuff |= (IXGBE_FCBUFF_VALID); -- cgit v0.10.2 From ca77cd59d28456b4061afa5254972ec47fa8baf5 Mon Sep 17 00:00:00 2001 From: Robert Love Date: Wed, 24 Mar 2010 12:45:00 +0000 Subject: ixgbe: Priority tag FIP frames Currently FIP (FCoE Initialization Protocol) frames are going untagged. This causes various problems with FCFs (switches) that have negotiated a priority over dcbx. This patch tags FIP frames with the same priority as the FCoE frames. Signed-off-by: Robert Love Signed-off-by: Chris Leech Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_main.c b/drivers/net/ixgbe/ixgbe_main.c index 208fb4a..0c553f6 100644 --- a/drivers/net/ixgbe/ixgbe_main.c +++ b/drivers/net/ixgbe/ixgbe_main.c @@ -5648,7 +5648,8 @@ static u16 ixgbe_select_queue(struct net_device *dev, struct sk_buff *skb) #ifdef IXGBE_FCOE if ((adapter->flags & IXGBE_FLAG_FCOE_ENABLED) && - (skb->protocol == htons(ETH_P_FCOE))) { + ((skb->protocol == htons(ETH_P_FCOE)) || + (skb->protocol == htons(ETH_P_FIP)))) { txq &= (adapter->ring_feature[RING_F_FCOE].indices - 1); txq += adapter->ring_feature[RING_F_FCOE].mask; return txq; @@ -5695,18 +5696,25 @@ static netdev_tx_t ixgbe_xmit_frame(struct sk_buff *skb, tx_ring = adapter->tx_ring[skb->queue_mapping]; - if ((adapter->flags & IXGBE_FLAG_FCOE_ENABLED) && - (skb->protocol == htons(ETH_P_FCOE))) { - tx_flags |= IXGBE_TX_FLAGS_FCOE; #ifdef IXGBE_FCOE + if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED) { #ifdef CONFIG_IXGBE_DCB - tx_flags &= ~(IXGBE_TX_FLAGS_VLAN_PRIO_MASK - << IXGBE_TX_FLAGS_VLAN_SHIFT); - tx_flags |= ((adapter->fcoe.up << 13) - << IXGBE_TX_FLAGS_VLAN_SHIFT); -#endif + /* for FCoE with DCB, we force the priority to what + * was specified by the switch */ + if ((skb->protocol == htons(ETH_P_FCOE)) || + (skb->protocol == htons(ETH_P_FIP))) { + tx_flags &= ~(IXGBE_TX_FLAGS_VLAN_PRIO_MASK + << IXGBE_TX_FLAGS_VLAN_SHIFT); + tx_flags |= ((adapter->fcoe.up << 13) + << IXGBE_TX_FLAGS_VLAN_SHIFT); + } #endif + /* flag for FCoE offloads */ + if (skb->protocol == htons(ETH_P_FCOE)) + tx_flags |= IXGBE_TX_FLAGS_FCOE; } +#endif + /* four things can cause us to need a context descriptor */ if (skb_is_gso(skb) || (skb->ip_summed == CHECKSUM_PARTIAL) || -- cgit v0.10.2 From af06393bbde6e8d474622a0517cffc662676e3fe Mon Sep 17 00:00:00 2001 From: Chris Leech Date: Wed, 24 Mar 2010 12:45:21 +0000 Subject: ixgbe: filter FIP frames into the FCoE offload queues During FCF solicitation, the switch is supposed to pad the solicited advertisement out to the endpoints specified maximum FCoE frame size. That means that we need to receive FIP frames that are larger than the standard MTU. To make sure the receive queue is configured correctly, we should be filtering FIP traffic into the FCoE queues. Signed-off-by: Chris Leech Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_fcoe.c b/drivers/net/ixgbe/ixgbe_fcoe.c index e1978da..9276d59 100644 --- a/drivers/net/ixgbe/ixgbe_fcoe.c +++ b/drivers/net/ixgbe/ixgbe_fcoe.c @@ -522,6 +522,9 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) /* Enable L2 eth type filter for FCoE */ IXGBE_WRITE_REG(hw, IXGBE_ETQF(IXGBE_ETQF_FILTER_FCOE), (ETH_P_FCOE | IXGBE_ETQF_FCOE | IXGBE_ETQF_FILTER_EN)); + /* Enable L2 eth type filter for FIP */ + IXGBE_WRITE_REG(hw, IXGBE_ETQF(IXGBE_ETQF_FILTER_FIP), + (ETH_P_FIP | IXGBE_ETQF_FILTER_EN)); if (adapter->ring_feature[RING_F_FCOE].indices) { /* Use multiple rx queues for FCoE by redirection table */ for (i = 0; i < IXGBE_FCRETA_SIZE; i++) { @@ -532,6 +535,12 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) } IXGBE_WRITE_REG(hw, IXGBE_FCRECTL, IXGBE_FCRECTL_ENA); IXGBE_WRITE_REG(hw, IXGBE_ETQS(IXGBE_ETQF_FILTER_FCOE), 0); + fcoe_i = f->mask; + fcoe_i &= IXGBE_FCRETA_ENTRY_MASK; + fcoe_q = adapter->rx_ring[fcoe_i]->reg_idx; + IXGBE_WRITE_REG(hw, IXGBE_ETQS(IXGBE_ETQF_FILTER_FIP), + IXGBE_ETQS_QUEUE_EN | + (fcoe_q << IXGBE_ETQS_RX_QUEUE_SHIFT)); } else { /* Use single rx queue for FCoE */ fcoe_i = f->mask; @@ -541,6 +550,12 @@ void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter) IXGBE_ETQS_QUEUE_EN | (fcoe_q << IXGBE_ETQS_RX_QUEUE_SHIFT)); } + /* send FIP frames to the first FCoE queue */ + fcoe_i = f->mask; + fcoe_q = adapter->rx_ring[fcoe_i]->reg_idx; + IXGBE_WRITE_REG(hw, IXGBE_ETQS(IXGBE_ETQF_FILTER_FIP), + IXGBE_ETQS_QUEUE_EN | + (fcoe_q << IXGBE_ETQS_RX_QUEUE_SHIFT)); IXGBE_WRITE_REG(hw, IXGBE_FCRXCTRL, IXGBE_FCRXCTRL_FCOELLI | diff --git a/drivers/net/ixgbe/ixgbe_type.h b/drivers/net/ixgbe/ixgbe_type.h index 0ed5ab3..4ec6dc1 100644 --- a/drivers/net/ixgbe/ixgbe_type.h +++ b/drivers/net/ixgbe/ixgbe_type.h @@ -1298,6 +1298,7 @@ #define IXGBE_ETQF_FILTER_BCN 1 #define IXGBE_ETQF_FILTER_FCOE 2 #define IXGBE_ETQF_FILTER_1588 3 +#define IXGBE_ETQF_FILTER_FIP 4 /* VLAN Control Bit Masks */ #define IXGBE_VLNCTRL_VET 0x0000FFFF /* bits 0-15 */ #define IXGBE_VLNCTRL_CFI 0x10000000 /* bit 28 */ -- cgit v0.10.2 From a6d36d5689b1806a3365c909192e9f03a43a632b Mon Sep 17 00:00:00 2001 From: Ben Menchaca Date: Wed, 24 Mar 2010 05:05:02 +0000 Subject: gianfar: fix undo of reserve() Fix undo of reserve() before RX recycle gfar_new_skb reserve()s space in the SKB to align it. If an error occurs, and the skb needs to be returned to the RX recycle queue, the current code attempts to reset head, but did not reset tail. This patch remembers the alignment amount, and reverses the reserve() when needed. Signed-off-by: Ben Menchaca Signed-off-by: David S. Miller diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c index b671555..669de02 100644 --- a/drivers/net/gianfar.c +++ b/drivers/net/gianfar.c @@ -2393,6 +2393,7 @@ struct sk_buff * gfar_new_skb(struct net_device *dev) * as many bytes as needed to align the data properly */ skb_reserve(skb, alignamount); + GFAR_CB(skb)->alignamount = alignamount; return skb; } @@ -2533,13 +2534,13 @@ int gfar_clean_rx_ring(struct gfar_priv_rx_q *rx_queue, int rx_work_limit) newskb = skb; else if (skb) { /* - * We need to reset ->data to what it + * We need to un-reserve() the skb to what it * was before gfar_new_skb() re-aligned * it to an RXBUF_ALIGNMENT boundary * before we put the skb back on the * recycle list. */ - skb->data = skb->head + NET_SKB_PAD; + skb_reserve(skb, -GFAR_CB(skb)->alignamount); __skb_queue_head(&priv->rx_recycle, skb); } } else { diff --git a/drivers/net/gianfar.h b/drivers/net/gianfar.h index 3d72dc4..17d25e7 100644 --- a/drivers/net/gianfar.h +++ b/drivers/net/gianfar.h @@ -566,6 +566,12 @@ struct rxfcb { u16 vlctl; /* VLAN control word */ }; +struct gianfar_skb_cb { + int alignamount; +}; + +#define GFAR_CB(skb) ((struct gianfar_skb_cb *)((skb)->cb)) + struct rmon_mib { u32 tr64; /* 0x.680 - Transmit and Receive 64-byte Frame Counter */ -- cgit v0.10.2 From ac90a149361a331f697d5aa500bedcff22054669 Mon Sep 17 00:00:00 2001 From: Kyle McMartin Date: Fri, 27 Mar 2009 17:23:32 +0000 Subject: tulip: Fix null dereference in uli526x_rx_packet() Acked-by: Grant Grundler Signed-off-by: David S. Miller diff --git a/drivers/net/tulip/uli526x.c b/drivers/net/tulip/uli526x.c index 0ab05af..90be57b 100644 --- a/drivers/net/tulip/uli526x.c +++ b/drivers/net/tulip/uli526x.c @@ -851,13 +851,15 @@ static void uli526x_rx_packet(struct net_device *dev, struct uli526x_board_info if ( !(rdes0 & 0x8000) || ((db->cr6_data & CR6_PM) && (rxlen>6)) ) { + struct sk_buff *new_skb = NULL; + skb = rxptr->rx_skb_ptr; /* Good packet, send to upper layer */ /* Shorst packet used new SKB */ - if ( (rxlen < RX_COPY_SIZE) && - ( (skb = dev_alloc_skb(rxlen + 2) ) - != NULL) ) { + if ((rxlen < RX_COPY_SIZE) && + ((new_skb = dev_alloc_skb(rxlen + 2) != NULL))) { + skb = new_skb; /* size less than COPY_SIZE, allocate a rxlen SKB */ skb_reserve(skb, 2); /* 16byte align */ memcpy(skb_put(skb, rxlen), -- cgit v0.10.2 From 4b97efdf392563bf03b4917a0b5add2df65de39a Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Fri, 26 Mar 2010 20:27:49 -0700 Subject: net: fix netlink address dumping in IPv4/IPv6 When a dump is interrupted at the last device in a hash chain and then continued, "idx" won't get incremented past s_idx, so s_ip_idx is not reset when moving on to the next device. This means of all following devices only the last n - s_ip_idx addresses are dumped. Tested-by: Pawel Staszewski Signed-off-by: Patrick McHardy diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 51ca946..3feb2b3 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -1194,7 +1194,7 @@ static int inet_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb) hlist_for_each_entry_rcu(dev, node, head, index_hlist) { if (idx < s_idx) goto cont; - if (idx > s_idx) + if (h > s_h || idx > s_idx) s_ip_idx = 0; in_dev = __in_dev_get_rcu(dev); if (!in_dev) diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index 3381b43..7e567ae 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -3610,7 +3610,7 @@ static int inet6_dump_addr(struct sk_buff *skb, struct netlink_callback *cb, hlist_for_each_entry_rcu(dev, node, head, index_hlist) { if (idx < s_idx) goto cont; - if (idx > s_idx) + if (h > s_h || idx > s_idx) s_ip_idx = 0; ip_idx = 0; if ((idev = __in6_dev_get(dev)) == NULL) -- cgit v0.10.2 From b35ecb5d404c00f2420221ccbb1bbba1139353a4 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 24 Mar 2010 07:43:17 +0000 Subject: ipv4: Cleanup struct net dereference in rt_intern_hash There's no need in getting it 3 times and gcc isn't smart enough to understand this himself. This is just a cleanup before the fix (next patch). Signed-off-by: Pavel Emelyanov Signed-off-by: David S. Miller diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 54fd68c..124af16 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1212,11 +1212,11 @@ restart: slow_chain_length(rt_hash_table[hash].chain) > rt_chain_length_max) { struct net *net = dev_net(rt->u.dst.dev); int num = ++net->ipv4.current_rt_cache_rebuild_count; - if (!rt_caching(dev_net(rt->u.dst.dev))) { + if (!rt_caching(net)) { printk(KERN_WARNING "%s: %d rebuilds is over limit, route caching disabled\n", rt->u.dst.dev->name, num); } - rt_emergency_hash_rebuild(dev_net(rt->u.dst.dev)); + rt_emergency_hash_rebuild(net); } } -- cgit v0.10.2 From 6a2bad70d546cf30a46bc6d9ec0cb9a0891a38eb Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 24 Mar 2010 21:51:22 +0000 Subject: ipv4: Restart rt_intern_hash after emergency rebuild (v2) The the rebuild changes the genid which in turn is used at the hash calculation. Thus if we don't restart and go on with inserting the rt will happen in wrong chain. (Fixed Neil's comment about the index passed into the rt_intern_hash) Signed-off-by: Pavel Emelyanov Reviewed-by: Neil Horman Signed-off-by: David S. Miller diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 124af16..d413b57 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -1097,7 +1097,7 @@ static int slow_chain_length(const struct rtable *head) } static int rt_intern_hash(unsigned hash, struct rtable *rt, - struct rtable **rp, struct sk_buff *skb) + struct rtable **rp, struct sk_buff *skb, int ifindex) { struct rtable *rth, **rthp; unsigned long now; @@ -1217,6 +1217,11 @@ restart: rt->u.dst.dev->name, num); } rt_emergency_hash_rebuild(net); + spin_unlock_bh(rt_hash_lock_addr(hash)); + + hash = rt_hash(rt->fl.fl4_dst, rt->fl.fl4_src, + ifindex, rt_genid(net)); + goto restart; } } @@ -1477,7 +1482,7 @@ void ip_rt_redirect(__be32 old_gw, __be32 daddr, __be32 new_gw, &netevent); rt_del(hash, rth); - if (!rt_intern_hash(hash, rt, &rt, NULL)) + if (!rt_intern_hash(hash, rt, &rt, NULL, rt->fl.oif)) ip_rt_put(rt); goto do_next; } @@ -1931,7 +1936,7 @@ static int ip_route_input_mc(struct sk_buff *skb, __be32 daddr, __be32 saddr, in_dev_put(in_dev); hash = rt_hash(daddr, saddr, dev->ifindex, rt_genid(dev_net(dev))); - return rt_intern_hash(hash, rth, NULL, skb); + return rt_intern_hash(hash, rth, NULL, skb, dev->ifindex); e_nobufs: in_dev_put(in_dev); @@ -2098,7 +2103,7 @@ static int ip_mkroute_input(struct sk_buff *skb, /* put it into the cache */ hash = rt_hash(daddr, saddr, fl->iif, rt_genid(dev_net(rth->u.dst.dev))); - return rt_intern_hash(hash, rth, NULL, skb); + return rt_intern_hash(hash, rth, NULL, skb, fl->iif); } /* @@ -2255,7 +2260,7 @@ local_input: } rth->rt_type = res.type; hash = rt_hash(daddr, saddr, fl.iif, rt_genid(net)); - err = rt_intern_hash(hash, rth, NULL, skb); + err = rt_intern_hash(hash, rth, NULL, skb, fl.iif); goto done; no_route: @@ -2502,7 +2507,7 @@ static int ip_mkroute_output(struct rtable **rp, if (err == 0) { hash = rt_hash(oldflp->fl4_dst, oldflp->fl4_src, oldflp->oif, rt_genid(dev_net(dev_out))); - err = rt_intern_hash(hash, rth, rp, NULL); + err = rt_intern_hash(hash, rth, rp, NULL, oldflp->oif); } return err; -- cgit v0.10.2 From a08af745e4c711d22aeadc2adade36958fe03ce8 Mon Sep 17 00:00:00 2001 From: Emil Tantilov Date: Thu, 25 Mar 2010 12:11:48 +0000 Subject: igbvf: do not modify tx_queue_len on link speed change Previously the driver tweaked txqueuelen to avoid false Tx hang reports seen at half duplex. This had the effect of overriding user set values on link change/reset. Testing shows that adjusting only the timeout factor is sufficient to prevent Tx hang reports at half duplex. Based on e1000e patch by Franco Fichtner CC: Franco Fichtner Signed-off-by: Emil Tantilov Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/igbvf/igbvf.h b/drivers/net/igbvf/igbvf.h index a1774b2..debeee2 100644 --- a/drivers/net/igbvf/igbvf.h +++ b/drivers/net/igbvf/igbvf.h @@ -198,7 +198,6 @@ struct igbvf_adapter { struct igbvf_ring *tx_ring /* One per active queue */ ____cacheline_aligned_in_smp; - unsigned long tx_queue_len; unsigned int restart_queue; u32 txd_cmd; diff --git a/drivers/net/igbvf/netdev.c b/drivers/net/igbvf/netdev.c index a77afd8..b41037e 100644 --- a/drivers/net/igbvf/netdev.c +++ b/drivers/net/igbvf/netdev.c @@ -1304,8 +1304,6 @@ static void igbvf_configure_tx(struct igbvf_adapter *adapter) /* enable Report Status bit */ adapter->txd_cmd |= E1000_ADVTXD_DCMD_RS; - - adapter->tx_queue_len = adapter->netdev->tx_queue_len; } /** @@ -1524,7 +1522,6 @@ void igbvf_down(struct igbvf_adapter *adapter) del_timer_sync(&adapter->watchdog_timer); - netdev->tx_queue_len = adapter->tx_queue_len; netif_carrier_off(netdev); /* record the stats before reset*/ @@ -1857,21 +1854,15 @@ static void igbvf_watchdog_task(struct work_struct *work) &adapter->link_duplex); igbvf_print_link_info(adapter); - /* - * tweak tx_queue_len according to speed/duplex - * and adjust the timeout factor - */ - netdev->tx_queue_len = adapter->tx_queue_len; + /* adjust timeout factor according to speed/duplex */ adapter->tx_timeout_factor = 1; switch (adapter->link_speed) { case SPEED_10: txb2b = 0; - netdev->tx_queue_len = 10; adapter->tx_timeout_factor = 16; break; case SPEED_100: txb2b = 0; - netdev->tx_queue_len = 100; /* maybe add some timeout factor ? */ break; } -- cgit v0.10.2 From f49c57e141c7f53353e4265a31dc2324e6215037 Mon Sep 17 00:00:00 2001 From: Emil Tantilov Date: Wed, 24 Mar 2010 12:55:02 +0000 Subject: e1000e: do not modify tx_queue_len on link speed change Previously the driver tweaked txqueuelen to avoid false Tx hang reports seen at half duplex. This had the effect of overriding user set values on link change/reset. Testing shows that adjusting only the timeout factor is sufficient to prevent Tx hang reports at half duplex. This patch removes all instances of tx_queue_len in the driver. Originally reported and patched by Franco Fichtner CC: Franco Fichtner Signed-off-by: Emil Tantilov Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h index c2ec095..118bdf4 100644 --- a/drivers/net/e1000e/e1000.h +++ b/drivers/net/e1000e/e1000.h @@ -279,7 +279,6 @@ struct e1000_adapter { struct napi_struct napi; - unsigned long tx_queue_len; unsigned int restart_queue; u32 txd_cmd; diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c index 88d54d3..e1cceb6 100644 --- a/drivers/net/e1000e/netdev.c +++ b/drivers/net/e1000e/netdev.c @@ -2289,8 +2289,6 @@ static void e1000_configure_tx(struct e1000_adapter *adapter) ew32(TCTL, tctl); e1000e_config_collision_dist(hw); - - adapter->tx_queue_len = adapter->netdev->tx_queue_len; } /** @@ -2877,7 +2875,6 @@ void e1000e_down(struct e1000_adapter *adapter) del_timer_sync(&adapter->watchdog_timer); del_timer_sync(&adapter->phy_info_timer); - netdev->tx_queue_len = adapter->tx_queue_len; netif_carrier_off(netdev); adapter->link_speed = 0; adapter->link_duplex = 0; @@ -3588,21 +3585,15 @@ static void e1000_watchdog_task(struct work_struct *work) "link gets many collisions.\n"); } - /* - * tweak tx_queue_len according to speed/duplex - * and adjust the timeout factor - */ - netdev->tx_queue_len = adapter->tx_queue_len; + /* adjust timeout factor according to speed/duplex */ adapter->tx_timeout_factor = 1; switch (adapter->link_speed) { case SPEED_10: txb2b = 0; - netdev->tx_queue_len = 10; adapter->tx_timeout_factor = 16; break; case SPEED_100: txb2b = 0; - netdev->tx_queue_len = 100; adapter->tx_timeout_factor = 10; break; } -- cgit v0.10.2 From 71c5c1595c04852d6fbf3c4882b47b30b61a4d32 Mon Sep 17 00:00:00 2001 From: Brandon L Black Date: Fri, 26 Mar 2010 16:18:03 +0000 Subject: net: Add MSG_WAITFORONE flag to recvmmsg Add new flag MSG_WAITFORONE for the recvmmsg() syscall. When this flag is specified for a blocking socket, recvmmsg() will only block until at least 1 packet is available. The default behavior is to block until all vlen packets are available. This flag has no effect on non-blocking sockets or when used in combination with MSG_DONTWAIT. Signed-off-by: Brandon L Black Acked-by: Ulrich Drepper Acked-by: Eric Dumazet Acked-by: Arnaldo Carvalho de Melo Signed-off-by: David S. Miller diff --git a/include/linux/socket.h b/include/linux/socket.h index 7b3aae2..354cc56 100644 --- a/include/linux/socket.h +++ b/include/linux/socket.h @@ -255,6 +255,7 @@ struct ucred { #define MSG_ERRQUEUE 0x2000 /* Fetch message from error queue */ #define MSG_NOSIGNAL 0x4000 /* Do not generate SIGPIPE */ #define MSG_MORE 0x8000 /* Sender will send more */ +#define MSG_WAITFORONE 0x10000 /* recvmmsg(): block until 1+ packets avail */ #define MSG_EOF MSG_FIN diff --git a/net/socket.c b/net/socket.c index 769c386..f55ffe9 100644 --- a/net/socket.c +++ b/net/socket.c @@ -2135,6 +2135,10 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen, break; ++datagrams; + /* MSG_WAITFORONE turns on MSG_DONTWAIT after one packet */ + if (flags & MSG_WAITFORONE) + flags |= MSG_DONTWAIT; + if (timeout) { ktime_get_ts(timeout); *timeout = timespec_sub(end_time, *timeout); -- cgit v0.10.2 From c0e4d4bad4e8cf0aa787a3045392f949d76b5886 Mon Sep 17 00:00:00 2001 From: wzt wzt Date: Thu, 25 Mar 2010 20:12:59 +0000 Subject: benet: Fix compile warnnings in drivers/net/benet/be_ethtool.c Fix the following warnings: be_ethtool.c:493: warning: integer constant is too large for 'long' type be_ethtool.c:493: warning: integer constant is too large for 'long' type Signed-off-by: Zhitong Wang Acked-by: Ajit Khaparde Signed-off-by: David S. Miller diff --git a/drivers/net/benet/be_ethtool.c b/drivers/net/benet/be_ethtool.c index 9560d48..51e1065 100644 --- a/drivers/net/benet/be_ethtool.c +++ b/drivers/net/benet/be_ethtool.c @@ -490,7 +490,7 @@ be_test_ddr_dma(struct be_adapter *adapter) { int ret, i; struct be_dma_mem ddrdma_cmd; - u64 pattern[2] = {0x5a5a5a5a5a5a5a5a, 0xa5a5a5a5a5a5a5a5}; + u64 pattern[2] = {0x5a5a5a5a5a5a5a5aULL, 0xa5a5a5a5a5a5a5a5ULL}; ddrdma_cmd.size = sizeof(struct be_cmd_req_ddrdma_test); ddrdma_cmd.va = pci_alloc_consistent(adapter->pdev, ddrdma_cmd.size, -- cgit v0.10.2 From e017b60316468f21a63bdd4affefaf81a7f988fd Mon Sep 17 00:00:00 2001 From: Alexander Duyck Date: Thu, 25 Mar 2010 17:15:06 +0000 Subject: igb: use correct bits to identify if managability is enabled igb was previously checking the wrong bits in the MANC register to determine if managability was enabled. As a result it was incorrectly powering down and resetting the phy when it didn't need to. Signed-off-by: Alexander Duyck Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/igb/e1000_mac.c b/drivers/net/igb/e1000_mac.c index 2a8a886..be8d010 100644 --- a/drivers/net/igb/e1000_mac.c +++ b/drivers/net/igb/e1000_mac.c @@ -1367,7 +1367,8 @@ out: * igb_enable_mng_pass_thru - Enable processing of ARP's * @hw: pointer to the HW structure * - * Verifies the hardware needs to allow ARPs to be processed by the host. + * Verifies the hardware needs to leave interface enabled so that frames can + * be directed to and from the management interface. **/ bool igb_enable_mng_pass_thru(struct e1000_hw *hw) { @@ -1380,8 +1381,7 @@ bool igb_enable_mng_pass_thru(struct e1000_hw *hw) manc = rd32(E1000_MANC); - if (!(manc & E1000_MANC_RCV_TCO_EN) || - !(manc & E1000_MANC_EN_MAC_ADDR_FILTER)) + if (!(manc & E1000_MANC_RCV_TCO_EN)) goto out; if (hw->mac.arc_subsystem_valid) { -- cgit v0.10.2 From e7d481a6f3c13041446b7bb8f98ab861460076a3 Mon Sep 17 00:00:00 2001 From: Greg Rose Date: Thu, 25 Mar 2010 17:06:48 +0000 Subject: ixgbe: Do not run all Diagnostic offline tests when VFs are active When running the offline diagnostic tests check to see if any VFs are online. If so then only run the link test. This is necessary because the VFs running in guest VMs aren't aware of when the PF is taken offline for a diagnostic test. Also put a message to the system log telling the system administrator to take the VFs offline manually if (s)he wants to run a full diagnostic. Return 1 on each of the tests not run to alert the user of the condition. Signed-off-by: Greg Rose Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/ixgbe/ixgbe_ethtool.c b/drivers/net/ixgbe/ixgbe_ethtool.c index 7949a44..1959ef7 100644 --- a/drivers/net/ixgbe/ixgbe_ethtool.c +++ b/drivers/net/ixgbe/ixgbe_ethtool.c @@ -1853,6 +1853,26 @@ static void ixgbe_diag_test(struct net_device *netdev, if (ixgbe_link_test(adapter, &data[4])) eth_test->flags |= ETH_TEST_FL_FAILED; + if (adapter->flags & IXGBE_FLAG_SRIOV_ENABLED) { + int i; + for (i = 0; i < adapter->num_vfs; i++) { + if (adapter->vfinfo[i].clear_to_send) { + netdev_warn(netdev, "%s", + "offline diagnostic is not " + "supported when VFs are " + "present\n"); + data[0] = 1; + data[1] = 1; + data[2] = 1; + data[3] = 1; + eth_test->flags |= ETH_TEST_FL_FAILED; + clear_bit(__IXGBE_TESTING, + &adapter->state); + goto skip_ol_tests; + } + } + } + if (if_running) /* indicate we're in test mode */ dev_close(netdev); @@ -1908,6 +1928,7 @@ skip_loopback: clear_bit(__IXGBE_TESTING, &adapter->state); } +skip_ol_tests: msleep_interruptible(4 * 1000); } -- cgit v0.10.2 From 7438189baa0a2fe30084bdc97e3d540ebc5444f3 Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Thu, 25 Mar 2010 23:45:35 +0000 Subject: net: ipmr/ip6mr: prevent out-of-bounds vif_table access When cache is unresolved, c->mf[6]c_parent is set to 65535 and minvif, maxvif are not initialized, hence we must avoid to parse IIF and OIF. A second problem can happen when the user dumps a cache entry where a VIF, that was referenced at creation time, has been removed. Signed-off-by: Nicolas Dichtel Signed-off-by: David S. Miller diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 0b9d03c..d0a6092 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c @@ -1616,17 +1616,20 @@ ipmr_fill_mroute(struct sk_buff *skb, struct mfc_cache *c, struct rtmsg *rtm) int ct; struct rtnexthop *nhp; struct net *net = mfc_net(c); - struct net_device *dev = net->ipv4.vif_table[c->mfc_parent].dev; u8 *b = skb_tail_pointer(skb); struct rtattr *mp_head; - if (dev) - RTA_PUT(skb, RTA_IIF, 4, &dev->ifindex); + /* If cache is unresolved, don't try to parse IIF and OIF */ + if (c->mfc_parent > MAXVIFS) + return -ENOENT; + + if (VIF_EXISTS(net, c->mfc_parent)) + RTA_PUT(skb, RTA_IIF, 4, &net->ipv4.vif_table[c->mfc_parent].dev->ifindex); mp_head = (struct rtattr *)skb_put(skb, RTA_LENGTH(0)); for (ct = c->mfc_un.res.minvif; ct < c->mfc_un.res.maxvif; ct++) { - if (c->mfc_un.res.ttls[ct] < 255) { + if (VIF_EXISTS(net, ct) && c->mfc_un.res.ttls[ct] < 255) { if (skb_tailroom(skb) < RTA_ALIGN(RTA_ALIGN(sizeof(*nhp)) + 4)) goto rtattr_failure; nhp = (struct rtnexthop *)skb_put(skb, RTA_ALIGN(sizeof(*nhp))); diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 23e4ac0..27acfb5 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -1695,17 +1695,20 @@ ip6mr_fill_mroute(struct sk_buff *skb, struct mfc6_cache *c, struct rtmsg *rtm) int ct; struct rtnexthop *nhp; struct net *net = mfc6_net(c); - struct net_device *dev = net->ipv6.vif6_table[c->mf6c_parent].dev; u8 *b = skb_tail_pointer(skb); struct rtattr *mp_head; - if (dev) - RTA_PUT(skb, RTA_IIF, 4, &dev->ifindex); + /* If cache is unresolved, don't try to parse IIF and OIF */ + if (c->mf6c_parent > MAXMIFS) + return -ENOENT; + + if (MIF_EXISTS(net, c->mf6c_parent)) + RTA_PUT(skb, RTA_IIF, 4, &net->ipv6.vif6_table[c->mf6c_parent].dev->ifindex); mp_head = (struct rtattr *)skb_put(skb, RTA_LENGTH(0)); for (ct = c->mfc_un.res.minvif; ct < c->mfc_un.res.maxvif; ct++) { - if (c->mfc_un.res.ttls[ct] < 255) { + if (MIF_EXISTS(net, ct) && c->mfc_un.res.ttls[ct] < 255) { if (skb_tailroom(skb) < RTA_ALIGN(RTA_ALIGN(sizeof(*nhp)) + 4)) goto rtattr_failure; nhp = (struct rtnexthop *)skb_put(skb, RTA_ALIGN(sizeof(*nhp))); -- cgit v0.10.2 From 39ca5f033bb2ea18877632809185268eebbb37a9 Mon Sep 17 00:00:00 2001 From: Emil Tantilov Date: Fri, 26 Mar 2010 11:25:58 +0000 Subject: e1000: do not modify tx_queue_len on link speed change Previously the driver tweaked txqueuelen to avoid false Tx hang reports seen at half duplex. This had the effect of overriding user set values on link change/reset. Testing shows that adjusting only the timeout factor is sufficient to prevent Tx hang reports at half duplex. This patch removes all instances of tx_queue_len in the driver. Based on e1000e patch by Franco Fichtner CC: Franco Fichtner Signed-off-by: Emil Tantilov Acked-by: Jesse Brandeburg Signed-off-by: Jeff Kirsher Signed-off-by: David S. Miller diff --git a/drivers/net/e1000/e1000.h b/drivers/net/e1000/e1000.h index 9902b33..2f29c21 100644 --- a/drivers/net/e1000/e1000.h +++ b/drivers/net/e1000/e1000.h @@ -261,7 +261,6 @@ struct e1000_adapter { /* TX */ struct e1000_tx_ring *tx_ring; /* One per active queue */ unsigned int restart_queue; - unsigned long tx_queue_len; u32 txd_cmd; u32 tx_int_delay; u32 tx_abs_int_delay; diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c index 8be6fae..b15ece2 100644 --- a/drivers/net/e1000/e1000_main.c +++ b/drivers/net/e1000/e1000_main.c @@ -383,8 +383,6 @@ static void e1000_configure(struct e1000_adapter *adapter) adapter->alloc_rx_buf(adapter, ring, E1000_DESC_UNUSED(ring)); } - - adapter->tx_queue_len = netdev->tx_queue_len; } int e1000_up(struct e1000_adapter *adapter) @@ -503,7 +501,6 @@ void e1000_down(struct e1000_adapter *adapter) del_timer_sync(&adapter->watchdog_timer); del_timer_sync(&adapter->phy_info_timer); - netdev->tx_queue_len = adapter->tx_queue_len; adapter->link_speed = 0; adapter->link_duplex = 0; netif_carrier_off(netdev); @@ -2316,19 +2313,15 @@ static void e1000_watchdog(unsigned long data) E1000_CTRL_RFCE) ? "RX" : ((ctrl & E1000_CTRL_TFCE) ? "TX" : "None" ))); - /* tweak tx_queue_len according to speed/duplex - * and adjust the timeout factor */ - netdev->tx_queue_len = adapter->tx_queue_len; + /* adjust timeout factor according to speed/duplex */ adapter->tx_timeout_factor = 1; switch (adapter->link_speed) { case SPEED_10: txb2b = false; - netdev->tx_queue_len = 10; adapter->tx_timeout_factor = 16; break; case SPEED_100: txb2b = false; - netdev->tx_queue_len = 100; /* maybe add some timeout factor ? */ break; } -- cgit v0.10.2 From 44ebb95290afcc687511ad3f7fd6434e867c270a Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Fri, 26 Mar 2010 16:27:55 +0000 Subject: drivers/net: Fix continuation lines Signed-off-by: Joe Perches Signed-off-by: David S. Miller diff --git a/drivers/net/atlx/atl1.c b/drivers/net/atlx/atl1.c index 9ba5470..0ebd820 100644 --- a/drivers/net/atlx/atl1.c +++ b/drivers/net/atlx/atl1.c @@ -84,7 +84,7 @@ #define ATLX_DRIVER_VERSION "2.1.3" MODULE_AUTHOR("Xiong Huang , \ - Chris Snook , Jay Cliburn "); +Chris Snook , Jay Cliburn "); MODULE_LICENSE("GPL"); MODULE_VERSION(ATLX_DRIVER_VERSION); -- cgit v0.10.2 From a2fd940f4cff74b932728bd6ca12848da21a0234 Mon Sep 17 00:00:00 2001 From: Andy Gospodarek Date: Thu, 25 Mar 2010 14:49:05 +0000 Subject: bonding: fix broken multicast with round-robin mode Round-robin (mode 0) does nothing to ensure that any multicast traffic originally destined for the host will continue to arrive at the host when the link that sent the IGMP join or membership report goes down. One of the benefits of absolute round-robin transmit. Keeping track of subscribed multicast groups for each slave did not seem like a good use of resources, so I decided to simply send on the curr_active slave of the bond (typically the first enslaved device that is up). This makes failover management simple as IGMP membership reports only need to be sent when the curr_active_slave changes. I tested this patch and it appears to work as expected. Originally reported by Lon Hohberger . Signed-off-by: Andy Gospodarek CC: Lon Hohberger CC: Jay Vosburgh Signed-off-by: Jay Vosburgh Signed-off-by: David S. Miller diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 430c022..5b92fbf 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1235,6 +1235,11 @@ void bond_change_active_slave(struct bonding *bond, struct slave *new_active) write_lock_bh(&bond->curr_slave_lock); } } + + /* resend IGMP joins since all were sent on curr_active_slave */ + if (bond->params.mode == BOND_MODE_ROUNDROBIN) { + bond_resend_igmp_join_requests(bond); + } } /** @@ -4138,22 +4143,41 @@ static int bond_xmit_roundrobin(struct sk_buff *skb, struct net_device *bond_dev struct bonding *bond = netdev_priv(bond_dev); struct slave *slave, *start_at; int i, slave_no, res = 1; + struct iphdr *iph = ip_hdr(skb); read_lock(&bond->lock); if (!BOND_IS_OK(bond)) goto out; - /* - * Concurrent TX may collide on rr_tx_counter; we accept that - * as being rare enough not to justify using an atomic op here + * Start with the curr_active_slave that joined the bond as the + * default for sending IGMP traffic. For failover purposes one + * needs to maintain some consistency for the interface that will + * send the join/membership reports. The curr_active_slave found + * will send all of this type of traffic. */ - slave_no = bond->rr_tx_counter++ % bond->slave_cnt; + if ((iph->protocol == htons(IPPROTO_IGMP)) && + (skb->protocol == htons(ETH_P_IP))) { - bond_for_each_slave(bond, slave, i) { - slave_no--; - if (slave_no < 0) - break; + read_lock(&bond->curr_slave_lock); + slave = bond->curr_active_slave; + read_unlock(&bond->curr_slave_lock); + + if (!slave) + goto out; + } else { + /* + * Concurrent TX may collide on rr_tx_counter; we accept + * that as being rare enough not to justify using an + * atomic op here. + */ + slave_no = bond->rr_tx_counter++ % bond->slave_cnt; + + bond_for_each_slave(bond, slave, i) { + slave_no--; + if (slave_no < 0) + break; + } } start_at = slave; -- cgit v0.10.2 From 1546a713ae1f066f83469cdd99ebdf500d6a65e4 Mon Sep 17 00:00:00 2001 From: Ken Kawasaki Date: Sat, 27 Mar 2010 10:55:37 +0000 Subject: pcnet_cs: add new id pcnet_cs: *add new id (Allied Telesis LM33-PCM-T Lan&Modem multifunction card) *use PROD_ID for LA-PCM.(because LA-PCM and LM33-PCM-T use the same MANF_ID). Signed-off-by: Ken Kawasaki Signed-off-by: David S. Miller diff --git a/drivers/net/pcmcia/pcnet_cs.c b/drivers/net/pcmcia/pcnet_cs.c index 776cad2..1028fcb 100644 --- a/drivers/net/pcmcia/pcnet_cs.c +++ b/drivers/net/pcmcia/pcnet_cs.c @@ -1549,6 +1549,7 @@ static struct pcmcia_device_id pcnet_ids[] = { PCMCIA_PFC_DEVICE_MANF_CARD(0, 0x021b, 0x0101), PCMCIA_PFC_DEVICE_MANF_CARD(0, 0x08a1, 0xc0ab), PCMCIA_PFC_DEVICE_PROD_ID12(0, "AnyCom", "Fast Ethernet + 56K COMBO", 0x578ba6e7, 0xb0ac62c4), + PCMCIA_PFC_DEVICE_PROD_ID12(0, "ATKK", "LM33-PCM-T", 0xba9eb7e2, 0x077c174e), PCMCIA_PFC_DEVICE_PROD_ID12(0, "D-Link", "DME336T", 0x1a424a1c, 0xb23897ff), PCMCIA_PFC_DEVICE_PROD_ID12(0, "Grey Cell", "GCS3000", 0x2a151fac, 0x48b932ae), PCMCIA_PFC_DEVICE_PROD_ID12(0, "Linksys", "EtherFast 10&100 + 56K PC Card (PCMLM56)", 0x0733cc81, 0xb3765033), @@ -1740,7 +1741,7 @@ static struct pcmcia_device_id pcnet_ids[] = { PCMCIA_MFC_DEVICE_CIS_PROD_ID12(0, "DAYNA COMMUNICATIONS", "LAN AND MODEM MULTIFUNCTION", 0x8fdf8f89, 0xdd5ed9e8, "cis/DP83903.cis"), PCMCIA_MFC_DEVICE_CIS_PROD_ID4(0, "NSC MF LAN/Modem", 0x58fc6056, "cis/DP83903.cis"), PCMCIA_MFC_DEVICE_CIS_MANF_CARD(0, 0x0175, 0x0000, "cis/DP83903.cis"), - PCMCIA_DEVICE_CIS_MANF_CARD(0xc00f, 0x0002, "cis/LA-PCM.cis"), + PCMCIA_DEVICE_CIS_PROD_ID12("Allied Telesis,K.K", "Ethernet LAN Card", 0x2ad62f3c, 0x9fd2f0a2, "cis/LA-PCM.cis"), PCMCIA_DEVICE_CIS_PROD_ID12("KTI", "PE520 PLUS", 0xad180345, 0x9d58d392, "cis/PE520.cis"), PCMCIA_DEVICE_CIS_PROD_ID12("NDC", "Ethernet", 0x01c43ae1, 0x00b2e941, "cis/NE2K.cis"), PCMCIA_DEVICE_CIS_PROD_ID12("PMX ", "PE-200", 0x34f3f1c8, 0x10b59f8c, "cis/PE-200.cis"), diff --git a/drivers/serial/serial_cs.c b/drivers/serial/serial_cs.c index e91db4b..175d202 100644 --- a/drivers/serial/serial_cs.c +++ b/drivers/serial/serial_cs.c @@ -745,6 +745,7 @@ static struct pcmcia_device_id serial_ids[] = { PCMCIA_PFC_DEVICE_PROD_ID13(1, "Xircom", "REM10", 0x2e3ee845, 0x76df1d29), PCMCIA_PFC_DEVICE_PROD_ID13(1, "Xircom", "XEM5600", 0x2e3ee845, 0xf1403719), PCMCIA_PFC_DEVICE_PROD_ID12(1, "AnyCom", "Fast Ethernet + 56K COMBO", 0x578ba6e7, 0xb0ac62c4), + PCMCIA_PFC_DEVICE_PROD_ID12(1, "ATKK", "LM33-PCM-T", 0xba9eb7e2, 0x077c174e), PCMCIA_PFC_DEVICE_PROD_ID12(1, "D-Link", "DME336T", 0x1a424a1c, 0xb23897ff), PCMCIA_PFC_DEVICE_PROD_ID12(1, "Gateway 2000", "XJEM3336", 0xdd9989be, 0x662c394c), PCMCIA_PFC_DEVICE_PROD_ID12(1, "Grey Cell", "GCS3000", 0x2a151fac, 0x48b932ae), -- cgit v0.10.2 From 78f1cd02457252e1ffbc6caa44a17424a45286b8 Mon Sep 17 00:00:00 2001 From: Francois Romieu Date: Sat, 27 Mar 2010 19:35:46 -0700 Subject: r8169: fix broken register writes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is quite similar to b39fe41f481d20c201012e4483e76c203802dda7 though said registers are not even documented as 64-bit registers - as opposed to the initial TxDescStartAddress ones - but as single bytes which must be combined into 32 bits at the MMIO read/write level before being merged into a 64 bit logical entity. Credits go to Ben Hutchings for the MAR registers (aka "multicast is broken for ages on ARM) and to Timo Teräs for the MAC registers. Signed-off-by: Francois Romieu Signed-off-by: David S. Miller diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c index b93fd23..7193afc 100644 --- a/drivers/net/r8169.c +++ b/drivers/net/r8169.c @@ -2820,8 +2820,8 @@ static void rtl_rar_set(struct rtl8169_private *tp, u8 *addr) spin_lock_irq(&tp->lock); RTL_W8(Cfg9346, Cfg9346_Unlock); - RTL_W32(MAC0, low); RTL_W32(MAC4, high); + RTL_W32(MAC0, low); RTL_W8(Cfg9346, Cfg9346_Lock); spin_unlock_irq(&tp->lock); @@ -4747,8 +4747,8 @@ static void rtl_set_rx_mode(struct net_device *dev) mc_filter[1] = swab32(data); } - RTL_W32(MAR0 + 0, mc_filter[0]); RTL_W32(MAR0 + 4, mc_filter[1]); + RTL_W32(MAR0 + 0, mc_filter[0]); RTL_W32(RxConfig, tmp); -- cgit v0.10.2 From 7855f761998893bb6bf861d55df95036fc9e36ab Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sun, 28 Mar 2010 18:56:34 -0700 Subject: tulip: Add missing parens. As reported by Stephen Rothwell. drivers/net/tulip/uli526x.c: In function 'uli526x_rx_packet': drivers/net/tulip/uli526x.c:861: warning: assignment makes pointer from integer without a cast Signed-off-by: David S. Miller diff --git a/drivers/net/tulip/uli526x.c b/drivers/net/tulip/uli526x.c index 90be57b..a4f09d4 100644 --- a/drivers/net/tulip/uli526x.c +++ b/drivers/net/tulip/uli526x.c @@ -858,7 +858,7 @@ static void uli526x_rx_packet(struct net_device *dev, struct uli526x_board_info /* Good packet, send to upper layer */ /* Shorst packet used new SKB */ if ((rxlen < RX_COPY_SIZE) && - ((new_skb = dev_alloc_skb(rxlen + 2) != NULL))) { + (((new_skb = dev_alloc_skb(rxlen + 2)) != NULL))) { skb = new_skb; /* size less than COPY_SIZE, allocate a rxlen SKB */ skb_reserve(skb, 2); /* 16byte align */ -- cgit v0.10.2 From 54c1a859efd9fd6cda05bc700315ba2519c14eba Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?YOSHIFUJI=20Hideaki=20/=20=E5=90=89=E8=97=A4=E8=8B=B1?= =?UTF-8?q?=E6=98=8E?= Date: Sun, 28 Mar 2010 07:15:45 +0000 Subject: ipv6: Don't drop cache route entry unless timer actually expired. This is ipv6 variant of the commit 5e016cbf6.. ("ipv4: Don't drop redirected route cache entry unless PTMU actually expired") by Guenter Roeck . Remove cache route entry in ipv6_negative_advice() only if the timer is expired. Signed-off-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 7fcb0e5..0d7713c 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -890,12 +890,17 @@ static struct dst_entry *ip6_negative_advice(struct dst_entry *dst) struct rt6_info *rt = (struct rt6_info *) dst; if (rt) { - if (rt->rt6i_flags & RTF_CACHE) - ip6_del_rt(rt); - else + if (rt->rt6i_flags & RTF_CACHE) { + if (rt6_check_expired(rt)) { + ip6_del_rt(rt); + dst = NULL; + } + } else { dst_release(dst); + dst = NULL; + } } - return NULL; + return dst; } static void ip6_link_failure(struct sk_buff *skb) -- cgit v0.10.2 From c0cd884af045338476b8e69a61fceb3f34ff22f1 Mon Sep 17 00:00:00 2001 From: Neil Horman Date: Mon, 29 Mar 2010 13:16:02 -0700 Subject: r8169: offical fix for CVE-2009-4537 (overlength frame DMAs) Official patch to fix the r8169 frame length check error. Based on this initial thread: http://marc.info/?l=linux-netdev&m=126202972828626&w=1 This is the official patch to fix the frame length problems in the r8169 driver. As noted in the previous thread, while this patch incurs a performance hit on the driver, its possible to improve performance dynamically by updating the mtu and rx_copybreak values at runtime to return performance to what it was for those NICS which are unaffected by the ideosyncracy (if there are any). Summary: A while back Eric submitted a patch for r8169 in which the proper allocated frame size was written to RXMaxSize to prevent the NIC from dmaing too much data. This was done in commit fdd7b4c3302c93f6833e338903ea77245eb510b4. A long time prior to that however, Francois posted 126fa4b9ca5d9d7cb7d46f779ad3bd3631ca387c, which expiclitly disabled the MaxSize setting due to the fact that the hardware behaved in odd ways when overlong frames were received on NIC's supported by this driver. This was mentioned in a security conference recently: http://events.ccc.de/congress/2009/Fahrplan//events/3596.en.html It seems that if we can't enable frame size filtering, then, as Eric correctly noticed, we can find ourselves DMA-ing too much data to a buffer, causing corruption. As a result is seems that we are forced to allocate a frame which is ready to handle a maximally sized receive. This obviously has performance issues with it, so to mitigate that issue, this patch does two things: 1) Raises the copybreak value to the frame allocation size, which should force appropriately sized packets to get allocated on rx, rather than a full new 16k buffer. 2) This patch only disables frame filtering initially (i.e., during the NIC open), changing the MTU results in ring buffer allocation of a size in relation to the new mtu (along with a warning indicating that this is dangerous). Because of item (2), individuals who can't cope with the performance hit (or can otherwise filter frames to prevent the bug), or who have hardware they are sure is unaffected by this issue, can manually lower the copybreak and reset the mtu such that performance is restored easily. Signed-off-by: Neil Horman Signed-off-by: David S. Miller diff --git a/drivers/net/r8169.c b/drivers/net/r8169.c index 7193afc..9674005 100644 --- a/drivers/net/r8169.c +++ b/drivers/net/r8169.c @@ -186,7 +186,12 @@ static DEFINE_PCI_DEVICE_TABLE(rtl8169_pci_tbl) = { MODULE_DEVICE_TABLE(pci, rtl8169_pci_tbl); -static int rx_copybreak = 200; +/* + * we set our copybreak very high so that we don't have + * to allocate 16k frames all the time (see note in + * rtl8169_open() + */ +static int rx_copybreak = 16383; static int use_dac; static struct { u32 msg_enable; @@ -3217,9 +3222,13 @@ static void __devexit rtl8169_remove_one(struct pci_dev *pdev) } static void rtl8169_set_rxbufsize(struct rtl8169_private *tp, - struct net_device *dev) + unsigned int mtu) { - unsigned int max_frame = dev->mtu + VLAN_ETH_HLEN + ETH_FCS_LEN; + unsigned int max_frame = mtu + VLAN_ETH_HLEN + ETH_FCS_LEN; + + if (max_frame != 16383) + printk(KERN_WARNING "WARNING! Changing of MTU on this NIC" + "May lead to frame reception errors!\n"); tp->rx_buf_sz = (max_frame > RX_BUF_SIZE) ? max_frame : RX_BUF_SIZE; } @@ -3231,7 +3240,17 @@ static int rtl8169_open(struct net_device *dev) int retval = -ENOMEM; - rtl8169_set_rxbufsize(tp, dev); + /* + * Note that we use a magic value here, its wierd I know + * its done because, some subset of rtl8169 hardware suffers from + * a problem in which frames received that are longer than + * the size set in RxMaxSize register return garbage sizes + * when received. To avoid this we need to turn off filtering, + * which is done by setting a value of 16383 in the RxMaxSize register + * and allocating 16k frames to handle the largest possible rx value + * thats what the magic math below does. + */ + rtl8169_set_rxbufsize(tp, 16383 - VLAN_ETH_HLEN - ETH_FCS_LEN); /* * Rx and Tx desscriptors needs 256 bytes alignment. @@ -3884,7 +3903,7 @@ static int rtl8169_change_mtu(struct net_device *dev, int new_mtu) rtl8169_down(dev); - rtl8169_set_rxbufsize(tp, dev); + rtl8169_set_rxbufsize(tp, dev->mtu); ret = rtl8169_init_ring(dev); if (ret < 0) -- cgit v0.10.2