- 30 Jun, 2010 8 commits
-
-
Flavio Leitner authored
When two systems using bonding devices in adaptive load balancing (ALB) communicates with each other, an endless ping-pong of ARP replies starts between these two systems. What happens? In the ALB mode, bonding driver keeps track of each client connected in a hash table, so it can do the receive load balancing (RLB). This hash table is updated when an ARP reply is received, then it scans for the client entry, updates its MAC address and flag it to be announced later. Therefore, two seconds later, the alb monitor runs and send for each updated client entry two ARP replies updating this specific client. The same process happens on the receiving system, causing the endless ping-pong of arp replies. See more information including the relevant functions below: System 1 System 2 bond0 bond0 ping <system2> ARP request ---------> <--------- ARP reply +->rlb_arp_recv <---------------------+ <--- loop begins | rlb_update_entry_from_arp | | client_info->ntt = 1; | | bond_info->rx_ntt = 1; | | | | <communication succeed> | | | | bond_alb_monitor | | rlb_update_rx_clients | | rlb_update_client | | arp_create(ARPOP_REPLY) | | send ARP reply --------------> V | send ARP reply --------------> | rlb_arp_recv | rlb_update_entry_from_arp | client_info->ntt = 1; | bond_info->rx_ntt = 1; | < snipped, same as in system 1> +------- <-------------- send ARP reply <-------------- send ARP reply Besides the unneeded networking traffic, this loop breaks a cluster because a backup system can't take over the IP address. There is always one system sending an ARP reply poisoning the network. This patch fixes the problem adding a check for the MAC address before updating it. Thus, if the MAC address didn't change, there is no need to update neither to announce it later. Signed-off-by:Flavio Leitner <fleitner@redhat.com> Signed-off-by:
Jay Vosburgh <fubar@us.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David Howells authored
Fix abuse of the preincrement operator as detected when building with gcc 4.6.0: CC [M] drivers/bluetooth/hci_bcsp.o drivers/bluetooth/hci_bcsp.c: In function 'bcsp_prepare_pkt': drivers/bluetooth/hci_bcsp.c:247:20: warning: operation on 'bcsp->msgq_txseq' may be undefined Reported-by:
Justin P. Mattock <justinmattock@gmail.com> Signed-off-by:
David Howells <dhowells@redhat.com> Acked-by:
Gustavo F. Padovan <padovan@profusion.mobi> Signed-off-by:
-
Saeed Bishara authored
Some controllers (KW, Dove) limits the TX IP/layer4 checksum offloading to a max size. Signed-off-by:
Saeed Bishara <saeed@marvell.com> Acked-by:
Lennert Buytenhek <buytenh@wantstofly.org> Signed-off-by:
-
-
Andreas Steffen authored
Determine the size of the xfrm_mark struct, not of its pointer. Signed-off-by:
Andreas Steffen <andreas.steffen@strongswan.org> Acked-by:
Jamal Hadi Salim <hadi@cyberus.ca> Signed-off-by:
-
Guillaume Gaudonville authored
In driver ixgbe, ixgbe_atr may cause crashes for non-ipv4 packets. Just add a test to check skb->protocol. It may crash on short packets due to ip_hdr() access. Signed-off-by:
Guillaume Gaudonville <guillaume.gaudonville@6wind.com> Acked-by:
Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by:
Don Skidmore <donald.c.skidmore@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
John Fastabend authored
Disabling the tx laser while receiving DMA requests can hang the device. After this occurs the device is in a bad state. The GPIO bit never clears when PCI master access is disabled and a reboot is required to get the device in a good state again. Signed-off-by:
John Fastabend <john.r.fastabend@intel.com> Acked-by:
-
Andy Gospodarek authored
This patch added to 2.6.34: commit 5f6c0181 Author: Jesse Brandeburg <jesse.brandeburg@intel.com> Date: Wed Apr 14 16:04:23 2010 -0700 ixgbe: fix bug with vlan strip in promsic mode among other things added a function called ixgbe_vlan_filter_enable. This new function wants to access and set some rx_ring parameters, but adapter->rx_ring has already been freed. This simply moves the free until after the access and makes __ixgbe_shutdown look more like ixgbe_remove. Signed-off-by:
Andy Gospodarek <andy@greyhouse.net> Acked-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by:
Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by:
-
- 29 Jun, 2010 3 commits
-
-
Ben Hutchings authored
struct ethtool_rxnfc was originally defined in 2.6.27 for the ETHTOOL_{G,S}RXFH command with only the cmd, flow_type and data fields. It was then extended in 2.6.30 to support various additional commands. These commands should have been defined to use a new structure, but it is too late to change that now. Since user-space may still be using the old structure definition for the ETHTOOL_{G,S}RXFH commands, and since they do not need the additional fields, only copy the originally defined fields to and from user-space. Signed-off-by:Ben Hutchings <bhutchings@solarflare.com> Cc: stable@kernel.org Signed-off-by:
-
Ben Hutchings authored
On a 32-bit machine, info.rule_cnt >= 0x40000000 leads to integer overflow and the buffer may be smaller than needed. Since ETHTOOL_GRXCLSRLALL is unprivileged, this can presumably be used for at least denial of service. Signed-off-by:
-
Andy Gospodarek authored
Support for netpoll over bonded interfaces was added here: commit f6dc31a8 Author: WANG Cong <amwang@redhat.com> Date: Thu May 6 00:48:51 2010 -0700 bonding: make bonding support netpoll but it is bad enough that we should probably just disable netpoll over bonding until some of the locking logic in the bonding driver is changed or converted completely to RCU. Simple actions like changing the active slave in active-backup mode will hang the box if a high enough printk debugging level is enabled. Keeping the old code around will be good for anyone that wants to work on it (and for after the RCU conversion), so I propose this small patch rather than ripping it all out. Signed-off-by:
-
- 27 Jun, 2010 3 commits
-
-
David Woodhouse authored
Commit e13647c1 (phylib: Add support for the LXT973 phy.) added a new ID but neglected to also add it to the MODULE_DEVICE_TABLE. Signed-off-by:
David Woodhouse <David.Woodhouse@intel.com> Signed-off-by:
-
Jiri Slaby authored
Stanse found that lp is dereferenced earlier than checked for being NULL in hysdn_rx_netpkt. Move the initialization below the test. Signed-off-by:
Jiri Slaby <jslaby@suse.cz> Cc: Karsten Keil <isdn@linux-pingi.de> Cc: "David S. Miller" <davem@davemloft.net> Cc: Stephen Hemminger <shemminger@vyatta.com> Cc: Patrick McHardy <kaber@trash.net> Cc: netdev@vger.kernel.org Signed-off-by:
-
Michal Schmidt authored
When pci_enable_msix() returned ret<0, entries and vxge_entries were leaked. While at it, use the centralized exit idiom in the function. Signed-off-by:
Michal Schmidt <mschmidt@redhat.com> Acked-by:
Ram Vepa <ram.vepa@exar.com> Signed-off-by:
-
- 26 Jun, 2010 8 commits
-
-
Tilman Schmidt authored
CAPI applications can handle several connections in parallel, so one connection state per application isn't sufficient. Store the connection state in the channel structure instead. Impact: bugfix Signed-off-by:
Tilman Schmidt <tilman@imap.cc> Signed-off-by:
-
Tilman Schmidt authored
Adapt to buggy device firmware which accepts setting HLC only in the same command line as BC, by encoding HLC and BC in a single command if both are specified, and rejecting HLC without BC. Impact: bugfix Signed-off-by:
-
Tilman Schmidt authored
The Gigaset CAPI driver handled all DATA_B3_REQ messages as if the Delivery Confirmation flag bit was set, delaying the emission of the DATA_B3_CONF reply until the data was actually transmitted. Some CAPI applications (notably Asterisk) aren't happy with that behaviour. Change it to actually evaluate the Delivery Confirmation flag as described the CAPI specification. Impact: bugfix Signed-off-by:
-
Tilman Schmidt authored
Make the Gigaset CAPI driver select L2_VOICE (AT^SBPR=2) as the layer 2 encoding for transparent connections, like the ISDN4Linux variant. L2_BITSYNC (AT^SBPR=0) mutes internal connections and distorts external ones. Impact: bugfix Signed-off-by:
-
Tilman Schmidt authored
Fix the Gigaset CAPI driver to limit the length of a connection's payload data receive buffers to the corresponding CAPI application's data buffer size, as some real-life CAPI applications tend to be rather unhappy if they receive bigger data blocks than requested. Impact: bugfix Signed-off-by:
-
Florian Fainelli authored
If the call to phy_connect fails, we will return directly instead of freeing the previously allocated struct net_device. Signed-off-by:
Florian Fainelli <florian@openwrt.org> CC: stable@kernel.org Signed-off-by:
-
Ken Kawasaki authored
smc91c92_cs: Fix the problem that lan & modem does not work simultaneously in the Megahertz multi-function card. We need to write MEGAHERTZ_ISR to retrigger interrupt. Signed-off-by:
Ken Kawasaki <ken_kawasaki@spring.nifty.jp> Signed-off-by:
-
stephen hemminger authored
The addition of TLLAO option created a kernel OOPS regression for the case where neighbor advertisement is being sent via proxy path. When using proxy, ipv6_get_ifaddr() returns NULL causing the NULL dereference. Change causing the bug was: commit f7734fdf Author: Octavian Purdila <opurdila@ixiacom.com> Date: Fri Oct 2 11:39:15 2009 +0000 make TLLAO option for NA packets configurable Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Acked-by:
YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> Signed-off-by:
-
- 25 Jun, 2010 3 commits
-
-
Wey-Yi Guy authored
When building tx command, always set TX_CMD_FLAG_PROT_REQUIRE_MSK for 5000 series and up. Without setting this bit the firmware will not examine the RTS/CTS setting and thus not send traffic with the appropriate protection. RTS/CTS is is required for HT traffic in a noisy environment where, without this setting, connections will stall on some hardware as documented in the patch that initially attempted to address this: commit 1152dcc2 Author: Wey-Yi Guy <wey-yi.w.guy@intel.com> Date: Fri Jan 15 13:42:58 2010 -0800 iwlwifi: Fix throughput stall issue in HT mode for 5000 Similar to 6000 and 1000 series, RTS/CTS is the recommended protection mechanism for 5000 series in HT mode based on the HW design. Using RTS/CTS will better protect the inner exchange from interference, especially in highly-congested environment, it also prevent uCode encounter TX FIFO underrun and other HT mode related performance issues. For 3945 and 4965, different flags are used for RTS/CTS or CTS-to-Self protection. Signed-off-by:Wey-Yi Guy <wey-yi.w.guy@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com>
-
Johannes Berg authored
commit 3474ad63 Author: Johannes Berg <johannes.berg@intel.com> Date: Thu Apr 29 04:43:05 2010 -0700 iwlwifi: apply filter flags directly broke multicast. The reason, it turns out, is that the code previously checked if ALLMULTI _changed_, which the new code no longer did, and normally it _never_ changes. Had somebody changed it manually, the code prior to my patch there would have been broken already. The reason is that we always, unconditionally, ask the device to pass up all multicast frames, but the new code made it depend on ALLMULTI which broke it since now we'd pass up multicast frames depending on the default filter in the device, which isn't necessarily what we want (since we don't program it right now). Fix this by simply not checking allmulti as we have allmulti behaviour enabled already anyway. Reported-by:
Maxim Levitsky <maximlevitsky@gmail.com> Tested-by:
Johannes Berg <johannes.berg@intel.com>
-
Gustavo F. Padovan authored
commit ff6e2163 accidentally added a regression on the bnep code. Fixing it. Signed-off-by:
-
- 24 Jun, 2010 1 commit
-
-
Timo Teräs authored
Fix the bundle validation code to not assume having a valid policy. When we have multiple transformations for a xfrm policy, the bundle instance will be a chain of bundles with only the first one having the policy reference. When policy_genid is bumped it will expire the first bundle in the chain which is equivalent of expiring the whole chain. Reported-bisected-and-tested-by:
Timo Teräs <timo.teras@iki.fi> Signed-off-by:
-
- 23 Jun, 2010 5 commits
-
-
Brandon Philips authored
sky2_phy_reinit is called by the ethtool helpers sky2_set_settings, sky2_nway_reset and sky2_set_pauseparam when netif_running. However, at the end of sky2_phy_init GM_GP_CTRL has GM_GPCR_RX_ENA and GM_GPCR_TX_ENA cleared. So, doing these commands causes the device to stop working: $ ethtool -r eth0 $ ethtool -A eth0 autoneg off Fix this issue by enabling Rx/Tx after running sky2_phy_init in sky2_phy_reinit. Signed-off-by:
Brandon Philips <bphilips@suse.de> Tested-by:
Mike McCormack <mikem@ring3k.org> Signed-off-by:
-
Vasanthakumar Thiagarajan authored
There are few places where ANI is started without checking if it is right to start. This might lead to a case where ani timer would be left undeleted and cause improper memory acccess during module unload. This bug is clearly exposed with paprd support where the driver detects tx hang and does a chip reset. During this reset ani is (re)started without checking if it needs to be started. This would leave a timer scheduled even after all the resources are freed and cause a panic. This patch introduces a bit in sc_flags to indicate if ani needs to be started in sw_scan_start() and ath_reset(). This would fix the following panic. This issue is easily seen with ar9003 + paprd. BUG: unable to handle kernel paging request at 0000000000003f38 [<ffffffff81075391>] ? __queue_work+0x41/0x50 [<ffffffff8106afaa>] run_timer_softirq+0x17a/0x370 [<ffffffff81088be8>] ? tick_dev_program_event+0x48/0x110 [<ffffffff81061f69>] __do_softirq+0xb9/0x1f0 [<ffffffff810ba060>] ? handle_IRQ_event+0x50/0x160 [<ffffffff8100af5c>] call_softirq+0x1c/0x30 [<ffffffff8100c9f5>] do_softirq+0x65/0xa0 [<ffffffff81061e25>] irq_exit+0x85/0x90 [<ffffffff8155e095>] do_IRQ+0x75/0xf0 [<ffffffff815570d3>] ret_from_intr+0x0/0x11 <EOI> [<ffffffff812fd67b>] ? acpi_idle_enter_simple+0xe4/0x119 [<ffffffff812fd674>] ? acpi_idle_enter_simple+0xdd/0x119 [<ffffffff81441c87>] cpuidle_idle_call+0xa7/0x140 [<ffffffff81008da3>] cpu_idle+0xb3/0x110 [<ffffffff81550722>] start_secondary+0x1ee/0x1f5 Signed-off-by:
Vasanthakumar Thiagarajan <vasanth@atheros.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Dmitry Kravkov authored
Disable statistics initialization for eth clients that do not support statistics. This prevents memory corruption on bnx2x hw. Signed-off-by:
Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by:
Michael Chan <mchan@broadcom.com> Signed-off-by:
Eilon Greenstein <eilong@broadcom.com>
-
Anirban Chakraborty authored
netxen_nic and qlcnic driver depends on firmware_class module. Signed-off-by:
Anirban Chakraborty <anirban.chakraborty@qlogic.com> Signed-off-by:
Amit Kumar Salecha <amit.salecha@qlogic.com> Signed-off-by:
-
Eric Dumazet authored
commit aa2ea058 (tcp: fix outsegs stat for TSO segments) incorrectly assumed SNMP_ADD_STATS() was used from BH context. Fix this using mib[!in_softirq()] instead of mib[0] Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> CC: Tom Herbert <therbert@google.com> Signed-off-by:
-
- 22 Jun, 2010 2 commits
-
-
-
Filip Aben authored
This patch removes the setting of the low_latency flag. tty_flip_buffer_push() is occasionally being called in irq context, which causes a hang if the low_latency flag is set. Removing the low_latency flag only seems to impact the flush to ldisc, which will now be put on a workqueue. Signed-off-by:
Filip Aben <f.aben@option.com> Signed-off-by:
-
- 21 Jun, 2010 3 commits
-
-
Herbert Xu authored
It has been reported that the new UFO software fallback path fails under certain conditions with NFS. I tracked the problem down to the generation of UFO packets that are smaller than the MTU. The software fallback path simply discards these packets. This patch fixes the problem by not generating such packets on the UFO path. Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Reviewed-by:
Michael S. Tsirkin <mst@redhat.com> Signed-off-by:
-
Helge Deller authored
Fix commit 4cd24eaf (net: use netdev_mc_count and netdev_mc_empty when appropriate) Signed-off-by:
Helge Deller <deller@gmx.de> Signed-off-by:
-
Ralf Baechle authored
$ make CONFIG_DEBUG_SECTION_MISMATCH=y [...] WARNING: drivers/net/built-in.o(.data+0x0): Section mismatch in reference from the variable mipsnet_driver to the function .init.text:mipsnet_probe() The variable mipsnet_driver references the function __init mipsnet_probe() If the reference is valid then annotate the variable with __init* or __refdata (see linux/init.h) or name the variable: *_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console, [...] Fixed by making mipsnet_probe __devinit. Signed-off-by:
Ralf Baechle <ralf@linux-mips.org> drivers/net/mipsnet.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) Signed-off-by:
-
- 18 Jun, 2010 1 commit
-
-
Bob Copeland authored
ath5k assumes ah_current_channel is always a valid pointer in several places, but a newly created interface may not have a channel. To avoid null pointer dereferences, set it up to point to the first available channel until later reconfigured. This fixes the following oops: $ rmmod ath5k $ insmod ath5k $ iw phy0 set distance 11000 BUG: unable to handle kernel NULL pointer dereference at 00000006 IP: [<d0a1ff24>] ath5k_hw_set_coverage_class+0x74/0x1b0 [ath5k] *pde = 00000000 Oops: 0000 [#1] last sysfs file: /sys/devices/pci0000:00/0000:00:0e.0/ieee80211/phy0/index Modules linked in: usbhid option usb_storage usbserial usblp evdev lm90 scx200_acb i2c_algo_bit i2c_dev i2c_core via_rhine ohci_hcd ne2k_pci 8390 leds_alix2 xt_IMQ imq nf_nat_tftp nf_conntrack_tftp nf_nat_irc nf_cc Pid: 1597, comm: iw Not tainted (2.6.32.14 #8) EIP: 0060:[<d0a1ff24>] EFLAGS: 00010296 CPU: 0 EIP is at ath5k_hw_set_coverage_class+0x74/0x1b0 [ath5k] EAX: 000000c2 EBX: 00000000 ECX: ffffffff EDX: c12d2080 ESI: 00000019 EDI: cf8c0000 EBP: d0a30edc ESP: cfa09bf4 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 Process iw (pid: 1597, ti=cfa09000 task=cf88a000 task.ti=cfa09000) Stack: d0a34f35 d0a353f8 d0a30edc 000000fe cf8c0000 00000000 1900063d cfa8c9e0 <0> cfa8c9e8 cfa8c0c0 cfa8c000 d0a27f0c 199d84b4 cfa8c200 00000010 d09bfdc7 <0> 00000000 00000000 ffffffff d08e0d28 cf9263c0 00000001 cfa09cc4 00000000 Call Trace: [<d0a27f0c>] ? ath5k_hw_attach+0xc8c/0x3c10 [ath5k] [<d09bfdc7>] ? __ieee80211_request_smps+0x1347/0x1580 [mac80211] [<d08e0d28>] ? nl80211_send_scan_start+0x7b8/0x4520 [cfg80211] [<c10f5db9>] ? nla_parse+0x59/0xc0 [<c11ca8d9>] ? genl_rcv_msg+0x169/0x1a0 [<c11ca770>] ? genl_rcv_msg+0x0/0x1a0 [<c11c7e68>] ? netlink_rcv_skb+0x38/0x90 [<c11c9649>] ? genl_rcv+0x19/0x30 [<c11c7c03>] ? netlink_unicast+0x1b3/0x220 [<c11c893e>] ? netlink_sendmsg+0x26e/0x290 [<c11a409e>] ? sock_sendmsg+0xbe/0xf0 [<c1032780>] ? autoremove_wake_function+0x0/0x50 [<c104d846>] ? __alloc_pages_nodemask+0x106/0x530 [<c1074933>] ? do_lookup+0x53/0x1b0 [<c10766f9>] ? __link_path_walk+0x9b9/0x9e0 [<c11acab0>] ? verify_iovec+0x50/0x90 [<c11a42b1>] ? sys_sendmsg+0x1e1/0x270 [<c1048e50>] ? find_get_page+0x10/0x50 [<c104a96f>] ? filemap_fault+0x5f/0x370 [<c1059159>] ? __do_fault+0x319/0x370 [<c11a55b4>] ? sys_socketcall+0x244/0x290 [<c101962c>] ? do_page_fault+0x1ec/0x270 [<c1019440>] ? do_page_fault+0x0/0x270 [<c1002ae5>] ? syscall_call+0x7/0xb Code: 00 b8 fe 00 00 00 b9 f8 53 a3 d0 89 5c 24 14 89 7c 24 10 89 44 24 0c 89 6c 24 08 89 4c 24 04 c7 04 24 35 4f a3 d0 e8 7c 30 60 f0 <0f> b7 43 06 ba 06 00 00 00 a8 10 75 0e 83 e0 20 83 f8 01 19 d2 EIP: [<d0a1ff24>] ath5k_hw_set_coverage_class+0x74/0x1b0 [ath5k] SS:ESP 0068:cfa09bf4 CR2: 0000000000000006 ---[ end trace 54f73d6b10ceb87b ]--- Cc: stable@kernel.org Reported-by:
Steve Brown <sbrown@cortland.com> Signed-off-by:
Bob Copeland <me@bobcopeland.com> Signed-off-by:
-
- 17 Jun, 2010 3 commits
-
-
stephen hemminger authored
It is common in end-node, non STP bridges to set forwarding delay to zero; which causes the forwarding database cleanup to run every clock tick. Change to run only as soon as needed or at next ageing timer interval which ever is sooner. Use round_jiffies_up macro rather than attempting round up by changing value. Signed-off-by:
-
FUJITA Tomonori authored
This removes dma_get_ops() prefetch optimization in bnx2. bnx2 uses dma_get_ops() to see if dma_sync_single_for_cpu() is noop. bnx2 does prefetch if it's noop. But dma_get_ops() isn't available on all the architectures (only the architectures that uses dma_map_ops struct have it). Using dma_get_ops() in drivers leads to compilation breakage on many architectures. This patch removes dma_get_ops() and changes bnx2 to do prefetch on all the architectures. This adds useless prefetch on non-coherent architectures but this is harmless. It is also unlikely to cause the performance drop. [ Remove now unused local variable 'pdev' -DaveM ] Signed-off-by:
FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> Acked-by:
-
Sergey Matyukevich authored
This patch implements a proper modification of RX skb buffers before recycling. Adjusting only skb->data is not enough because after that skb->tail and skb->len become incorrect. Signed-off-by:
Sergey Matyukevich <geomatsi@gmail.com> Signed-off-by:
-
