- 19 Dec, 2009 8 commits
-
-
Breno Leitao authored
If bnx2 schedules a reset via the reset_task, e.g., due to a TX timeout, it's possible for the NIC to be disabled with packets pending for transmit. In this case, napi_disable will loop forever, eventually crashing the kernel. This patch moves the disable of the device to after the napi_disable call. Signed-off-by:
Breno Leitao <leitao@linux.vnet.ibm.com> Acked-by:
Michael Chan <mchan@broadcom.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Yang Hongyang authored
When I do an ipv6 module force unload,I got the following oops: #rmmod -f ipv6 ------------[ cut here ]------------ kernel BUG at mm/slub.c:2969! invalid opcode: 0000 [#1] SMP last sysfs file: /sys/devices/pci0000:00/0000:00:11.0/0000:02:03.0/net/eth2/ifindex Modules linked in: ipv6(-) dm_multipath uinput ppdev tpm_tis tpm tpm_bios pcspkr pcnet32 mii parport_pc i2c_piix4 parport i2c_core floppy mptspi mptscsih mptbase scsi_transport_spi Pid: 2530, comm: rmmod Tainted: G R 2.6.32 #2 440BX Desktop Reference Platform/VMware Virtual Platform EIP: 0060:[<c04b73f2>] EFLAGS: 00010246 CPU: 0 EIP is at kfree+0x6a/0xdd EAX: 00000000 EBX: c09e86bc ECX: c043e4dd EDX: c14293e0 ESI: e141f1d8 EDI: e140fc31 EBP: dec58ef0 ESP: dec58ed0 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 Process rmmod (pid: 2530, ti=dec58000 task=decb1940 task.ti=dec58000) Stack: c14293e0 00000282 df624240 c0897d08 c09e86bc c09e86bc e141f1d8 dec58f1c <0> dec58f00 e140fc31 c09e84c4 e141f1bc dec58f14 c0689d21 dec58f1c e141f1bc <0> 00000000 dec58f2c c0689eff c09e84d8 c09e84d8 e141f1bc bff33a90 dec58f38 Call Trace: [<e140fc31>] ? ipv6_frags_exit_net+0x22/0x32 [ipv6] [<c0689d21>] ? ops_exit_list+0x19/0x3d [<c0689eff>] ? unregister_pernet_operations+0x2a/0x51 [<c0689f70>] ? unregister_pernet_subsys+0x17/0x24 [<e140fbfe>] ? ipv6_frag_exit+0x21/0x32 [ipv6] [<e141a361>] ? inet6_exit+0x47/0x122 [ipv6] [<c045f5de>] ? sys_delete_module+0x198/0x1f6 [<c04a8acf>] ? remove_vma+0x57/0x5d [<c070f63f>] ? do_page_fault+0x2e7/0x315 [<c0403218>] ? sysenter_do_call+0x12/0x28 Code: 86 00 00 00 40 c1 e8 0c c1 e0 05 01 d0 89 45 e0 66 83 38 00 79 06 8b 40 0c 89 45 e0 8b 55 e0 8b 02 84 c0 78 14 66 a9 00 c0 75 04 <0f> 0b eb fe 8b 45 e0 e8 35 15 fe ff eb 5d 8b 45 04 8b 55 e0 89 EIP: [<c04b73f2>] kfree+0x6a/0xdd SS:ESP 0068:dec58ed0 ---[ end trace 4475d1a5b0afa7e5 ]--- It's because in ip6_frags_ns_sysctl_register, "table" only alloced when "net" is not equals to "init_net".So when we free "table" in ip6_frags_ns_sysctl_unregister,we should check this first. This patch fix the problem. Signed-off-by:
Yang Hongyang <yanghy@cn.fujitsu.com> Signed-off-by:
-
Nageswari Srinivasan authored
There was a typo in "if condition" checking for validity of MDIO bus frequency passed as part of platform data. Bitwise AND was being used instead of a Logical AND. Tested on: DM6467 EVM Signed-off-by:
Nageswari Srinivasan <nageswari@ti.com> Acked-by:
Anant Gole <anantgole@ti.com> Signed-off-by:
-
Roger Oksanen authored
Alan Stern noticed that e100 caused slab corruption. commit 98468efd changed the allocation of cbs to use dma pools that don't return zeroed memory, especially the cb->status field used to track which cb to clean, causing (the visible) double freeing of skbs and a wrong free cbs count. Now the cbs are explicitly zeroed at allocation time. Reported-by:
Alan Stern <stern@rowland.harvard.edu> Tested-by:
Roger Oksanen <roger.oksanen@cs.helsinki.fi> Acked-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by:
-
Roel Kluin authored
If not signed read errors are ignored. Signed-off-by:
Roel Kluin <roel.kluin@gmail.com> Acked-by:
Matt Carlson <mcarlson@broadcom.com> Signed-off-by:
-
Bruce Allan authored
Do not override the customizable LED configuration set in the EEPROM. Signed-off-by:
Bruce Allan <bruce.w.allan@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Stephen Hemminger authored
Netxen driver is doing this bogus thing to create a control file. This fails if device doesn't exist, and overall is a bad way to do the module parameter. Rather than fix borked code, just rewrite. Just using a writeable module parameter of 0/1 is the correct way Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Signed-off-by:
-
Alexey Dobriyan authored
sysctl table was copied, all right, but ->data for net.ipv6.route.gc_min_interval_ms was not reinitialized for "!= &init_net" case. In init_net everthing works by accident due to correct ->data initialization in source table. Signed-off-by:
Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by:
-
- 17 Dec, 2009 5 commits
-
-
Oliver Neukum authored
The device must be marked busy as it receives data. Signed-off-by:
Oliver Neukum <oliver@neukum.org> Tested-by:
Matthew Garrett <mjg@redhat.com> Signed-off-by:
Marcel Holtmann <marcel@holtmann.org>
-
Andrei Emeltchenko authored
When locking was introduced the error path branch was not taken into account. Error was found in sparse code checking. Kudos to Jani Nikula. Signed-off-by:
Andrei Emeltchenko <andrei.emeltchenko@nokia.com> Acked-by:
Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by:
-
Gustavo F. Padovan authored
Moving the Ack to before l2cap_retransmit_frame() we can avoid the case where txWindow is full and the packet can't be retransmited. Signed-off-by:
-
Gustavo F. Padovan authored
RemoteBusy flag need to be unset before l2cap_ertm_send(), otherwise l2cap_ertm_send() will return without sending packets because it checks that flag before start sending. Signed-off-by:
-
Roel Kluin authored
Return the PTR_ERR of the correct pointer. Signed-off-by:
-
- 16 Dec, 2009 8 commits
-
-
Magnus Damm authored
Fix sh_eth for sh7724 by adding NET_IP_ALIGN support V2. Without this patch the receive data is misaligned. Signed-off-by:
Magnus Damm <damm@opensource.se> Signed-off-by:
-
Gurucharan Shetty authored
When the 82598 is fed 802.1q packets, it chokes with an error of the form: ixgbe: eth0: ixgbe_tx_csum: partial checksum but proto=81! As the logic there was not smart enough to look into the vlan header to pick out the encapsulated protocol. There are times when we'd like to send these packets out without having to configure a vlan on the interface. Here we check for the vlan tag and allow the packet to go out with the correct hardware checksum. This patch is a clone of a previously submitted patch by Arthur Jones <ajones@riverbed.com> for igb (Commit - fa4a7ef3). Signed-off-by:
Gurucharan Shetty <gshetty@riverbed.com> Signed-off-by:
Arthur Jones <ajones@riverbed.com> Acked-by:
Mallikarjuna R Chilakala <mallikarjuna.chilakala@intel.com> Signed-off-by:
-
Mallikarjuna R Chilakala authored
Modified patch with Dave's comments to replace mdelay with proper msleep. Fix 82598 copper link issue, where the phy prematurely indicates link before it is ready to process packets. The new function looks for phy link and indicates that, when it is available. If phy is not ready within few seconds of MAC indicating link, the function will return failure which translates to link down indication. Signed-off-by:
Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by:
-
Mallikarjuna R Chilakala authored
Fix the restart_queue and non_eop_desc counters from being double-counted. They are cumulative in each ring, so we don't want to add them to the cumulative result in the adapter's master counter. Otherwise, the stats will be inaccurate Signed-off-by:
-
Florian Fainelli authored
This patch converts bcm63xx_enet to uset get_sset_count like the other drivers do. Signed-off-by:
Florian Fainelli <ffainelli@freebox.fr> Signed-off-by:
-
Eric Dumazet authored
commit 654d1f8a (packet: less dev_put() calls) introduced a problem, calling potentially sleeping functions from a rcu_read_lock() protected section. Fix this by releasing lock before the sock_wmalloc()/memcpy_fromiovec() calls. After skb allocation and copy from user space, we redo device lookup and appropriate tests. Reported-and-tested-by:
Frederic Weisbecker <fweisbec@gmail.com> Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
-
David S. Miller authored
It creates a regression, triggering badness for SYN_RECV sockets, for example: [19148.022102] Badness at net/ipv4/inet_connection_sock.c:293 [19148.022570] NIP: c02a0914 LR: c02a0904 CTR: 00000000 [19148.023035] REGS: eeecbd30 TRAP: 0700 Not tainted (2.6.32) [19148.023496] MSR: 00029032 <EE,ME,CE,IR,DR> CR: 24002442 XER: 00000000 [19148.024012] TASK = eee9a820[1756] 'privoxy' THREAD: eeeca000 This is likely caused by the change in the 'estab' parameter passed to tcp_parse_options() when invoked by the functions in net/ipv4/tcp_minisocks.c But even if that is fixed, the ->conn_request() changes made in this patch series is fundamentally wrong. They try to use the listening socket's 'dst' to probe the route settings. The listening socket doesn't even have a route, and you can't get the right route (the child request one) until much later after we setup all of the state, and it must be done by hand. This stuff really isn't ready, so the best thing to do is a full revert. This reverts the following commits: f55017a9 022c3f7d 1aba721e cda42ebd 345cda2f dc343475 05eaade2 6a2a2d6bSigned-off-by:
-
- 15 Dec, 2009 6 commits
-
-
Simon Horman authored
Make sure that any otherwise uninitialised fields of usvc are zero. This has been obvserved to cause a problem whereby the port of fwmark services may end up as a non-zero value which causes scheduling of a destination server to fail for persisitent services. As observed by Deon van der Merwe <dvdm@truteq.co.za>. This fix suggested by Julian Anastasov <ja@ssi.bg>. For good measure also zero udest. Cc: Deon van der Merwe <dvdm@truteq.co.za> Acked-by:
Julian Anastasov <ja@ssi.bg> Signed-off-by:
Simon Horman <horms@verge.net.au> Cc: stable@kernel.org Signed-off-by:
Patrick McHardy <kaber@trash.net>
-
Patrick McHardy authored
When fragments from bridge netfilter are passed to IPv4 or IPv6 conntrack and a reassembly queue with the same fragment key already exists from reassembling a similar packet received on a different device (f.i. with multicasted fragments), the reassembled packet might continue on a different codepath than where the head fragment originated. This can cause crashes in bridge netfilter when a fragment received on a non-bridge device (and thus with skb->nf_bridge == NULL) continues through the bridge netfilter code. Add a new reassembly identifier for packets originating from bridge netfilter and use it to put those packets in insolated queues. Fixes http://bugzilla.kernel.org/show_bug.cgi?id=14805Reported-and-Tested-by:
Chong Qiao <qiaochong@loongson.cn> Signed-off-by:
-
Patrick McHardy authored
Currently the same reassembly queue might be used for packets reassembled by conntrack in different positions in the stack (PREROUTING/LOCAL_OUT), as well as local delivery. This can cause "packet jumps" when the fragment completing a reassembled packet is queued from a different position in the stack than the previous ones. Add a "user" identifier to the reassembly queue key to seperate the queues of each caller, similar to what we do for IPv4. Signed-off-by: -
stephen hemminger authored
Since power management is done by PCI subsystem as well as driver, don't toggle the bit that disables PCI register writes. Signed-off-by:
Rafael J. Wysocki <rjw@sisk.pl> Signed-off-by:
-
stephen hemminger authored
Off by one in name lookup makes Optima display as (chip 0xbc) Signed-off-by:
-
Arnd Bergmann authored
On Monday 14 December 2009, andrew hendry wrote: > Thanks, I didn't know X.25 was actively maintained. I get bounces. > Is the the maintainers out of date? From looking at the posts on the x.25 mailing list and the changes that went into the kernel during the last three years in that area, I think it is safe to say that you are now the maintainer ;-). The last mail on this topic from Henner Eisen was around 2001. > AX.25 NETWORK LAYER > M: Ralf Baechle <ralf@linux-mips.org> > > X.25 NETWORK LAYER > M: Henner Eisen <eis@baty.hanse.de> How about this change? Signed-off-by:
Andrew Hendry <andrew.hendry@gmail.com> Signed-off-by:
-
- 14 Dec, 2009 13 commits
-
-
Xiaotian Feng authored
commit 9d3a0de7 makes slaves expire as they would do on the master with much shorter timeouts. But it introduces another problem: When we close a connection, on master server the connection became CLOSE_WAIT/TIME_WAIT, it was synced to slaves, but if master is finished within it's timeouts (CLOSE), it will not be synced to slaves. Then slaves will be kept on CLOSE_WAIT/TIME_WAIT until timeout reaches. Thus we should also sync with CLOSE. Cc: Wensong Zhang <wensong@linux-vs.org> Cc: Simon Horman <horms@verge.net.au> Cc: Julian Anastasov <ja@ssi.bg> Cc: David S. Miller <davem@davemloft.net> Signed-off-by:
Xiaotian Feng <dfeng@redhat.com> Acked-by:
-
Jan Engelhardt authored
For both .33 and .32-stable. Signed-off-by:
Jan Engelhardt <jengelh@medozas.de> Cc: stable@kernel.org Signed-off-by:
-
Joe Perches authored
Add #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt Remove DRV_NAME from pr_<level>s Consolidate long format strings Remove some extra tab indents Remove some unnecessary ()s from pr_<level>s arguments Align pr_<level> arguments Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Geert Uytterhoeven authored
When building for Sun 3: drivers/net/can/mcp251x.c:1074: undefined reference to `dma_free_coherent' drivers/net/can/mcp251x.c:976: undefined reference to `dma_alloc_coherent' drivers/net/can/mcp251x.c:1050: undefined reference to `dma_free_coherent' Signed-off-by:
Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by:
-
Julia Lawall authored
sizeof(dev->dev_addr) is the size of a pointer. A few lines above, the size of this field is obtained using netdev->addr_len for a call to memcpy, so do the same here. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression *x; expression f; type T; @@ *f(...,(T)x,...) // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Signed-off-by:
-
Julia Lawall authored
sizeof(dev->dev_addr) is the size of a pointer. On the other hand, sizeof(pdata->dev_addr) is the size of an array, so use that instead. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression *x; expression f; type T; @@ *f(...,(T)x,...) // </smpl> Signed-off-by:
-
Julia Lawall authored
The function efx_iterate_state contains the code memcpy(&payload->msg, payload_msg, sizeof(payload_msg)); This is the only use of payload_msg. The type of payload_msg is changed from a pointer to an array, so that the result of sizeof really is the length of the string. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression *x; expression f; type T; @@ *f(...,(T)x,...) // </smpl> Signed-off-by:
-
Julia Lawall authored
sizeof(TstSchedTbl) is just the size of the pointer. Change it to the size of the referenced data. A simplified version of the semantic patch that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @@ expression *x; expression f; type T; @@ *f(...,(T)x,...) // </smpl> Signed-off-by:
-
Ken Kawasaki authored
3c574_cs, 3c589_cs: disable irq before calling el3_interrupt in the media_check function. Signed-off-by:
Ken Kawasaki <ken_kawasaki@spring.nifty.jp> Signed-off-by:
-
roel kluin authored
The return value should be negative. Signed-off-by:
-
Oliver Hartkopp authored
A valid CAN dataframe can have a data length code (DLC) of 0 .. 8 data bytes. When reading the CAN controllers register the 4-bit value may contain values from 0 .. 15 which may exceed the reserved space in the socket buffer! The ISO 11898-1 Chapter 8.4.2.3 (DLC field) says that register values > 8 should be reduced to 8 without any error reporting or frame drop. This patch introduces a new helper macro to cast a given 4-bit data length code (dlc) to __u8 and ensure the DLC value to be max. 8 bytes. The different handlings in the rx path of the CAN netdevice drivers are fixed. Signed-off-by:
Oliver Hartkopp <oliver@hartkopp.net> Signed-off-by:
Wolfgang Grandegger <wg@grandegger.com> Signed-off-by:
-
Eric W. Biederman authored
I received some bug reports about userspace programs having problems because after RTM_NEWLINK was received they could not immediate access files under /proc/sys/net/ because they had not been registered yet. The original problem was trivially fixed by moving the userspace notification from rtnetlink_event() to the end of register_netdevice(). When testing that change I discovered I was still getting RTM_NEWLINK events before I could access proc and I was also getting RTM_NEWLINK events after I was seeing RTM_DELLINK. Things practically guaranteed to confuse userspace. After a little more investigation these extra notifications proved to be from the new notifiers NETDEV_POST_INIT and NETDEV_UNREGISTER_BATCH hitting the default case in rtnetlink_event, and triggering unnecessary RTM_NEWLINK messages. rtnetlink_event now explicitly handles NETDEV_UNREGISTER_BATCH and NETDEV_POST_INIT to avoid sending the incorrect userspace notifications. Signed-off-by:
Eric W. Biederman <ebiederm@aristanetworks.com> Signed-off-by:
-
Eric Dumazet authored
Now we can have a large udp hash table, udp_lib_get_port() loop should be converted to a do {} while (cond) form, or we dont enter it at all if hash table size is exactly 65536. Reported-by:Yinghai Lu <yinghai@kernel.org> Signed-off-by:
-
