- 21 May, 2009 14 commits
-
-
Neil Horman authored
Patch to add the ability to detect drops in hardware interfaces via dropwatch. Adds a tracepoint to net_rx_action to signal everytime a napi instance is polled. The dropmon code then periodically checks to see if the rx_frames counter has changed, and if so, adds a drop notification to the netlink protocol, using the reserved all-0's vector to indicate the drop location was in hardware, rather than somewhere in the code. Signed-off-by:
Neil Horman <nhorman@tuxdriver.com> include/linux/net_dropmon.h | 8 ++ include/trace/napi.h | 11 +++ net/core/dev.c | 5 + net/core/drop_monitor.c | 124 ++++++++++++++++++++++++++++++++++++++++++-- net/core/net-traces.c | 4 + net/core/netpoll.c | 2 6 files changed, 149 insertions(+), 5 deletions(-) Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Eric W. Biederman authored
The the PACKET_ADD_MEMBERSHIP and the PACKET_DROP_MEMBERSHIP setsockopt calls for af_packet already has all of the infrastructure needed to subscribe to multiple mac addresses. All that is missing is a flag to say that the address we want to listen on is a unicast address. So introduce PACKET_MR_UNICAST and wire it up to dev_unicast_add and dev_unicast_delete. Additionally I noticed that errors from dev_mc_add were not propagated from packet_dev_mc so fix that. Signed-off-by:
Eric W. Biederman <ebiederm@aristanetworks.com> Signed-off-by:
-
Stephen Hemminger authored
The net_ns_init code can be simplified. No need to save error code if it is only going to panic if it is set 4 lines later. Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Acked-by:
"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Stephen Hemminger authored
Signed-off-by:
-
Divy Le Ray authored
mdio's dev field needs to be set before mdio ops occur. Signed-off-by:
Divy Le Ray <divy@chelsio.com> Signed-off-by:
-
Divy Le Ray authored
mdio's dev field needs to be set before mdio ops occur. Signed-off-by:
-
Ben Hutchings authored
Signed-off-by:
Ben Hutchings <bhutchings@solarflare.com> Acked-by:
-
Ben Hutchings authored
Signed-off-by:
-
Klaus-Dieter Wacket authored
Function qeth_clear_output_buffer for HiperSockets may not clear all 16 SBALEs, but only the used ones. The error flag in SBALF15 has to be cleared in any case. Signed-off-by:
Klaus-Dieter Wacker <kdwacker@de.ibm.com> Signed-off-by:
Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by:
-
Ursula Braun authored
For HiperSocket devices receive-path checksumming is not required. Thus NO_CHECKSUMMING is used as default for HiperSocket interfaces. For layer3 devices configured with NO_CHECKSUMMING received skbs should have set their ip_summed field to CHECKSUM_UNNECESSARY. Signed-off-by:
Ursula Braun <ursula.braun@de.ibm.com> Signed-off-by:
-
Ursula Braun authored
z/VM Virtual Switch Port Isolation allows guests on a VLAN UNAWARE virtual switch to be isolated from other guests on the VSWITCH. (See z/VM Apars VM64281 and VM64463). The Linux qeth driver is affected, because it has to handle new error codes introduced with the z/VM VSWITCH Port Isolation support. Signed-off-by:
-
Ursula Braun authored
Channels are already removed when setting a ctcm-device offline. Thus ctcm_remove_device must not refer to channel information. Solution: delete channel information from the trace call in ctcm_remove_device. Signed-off-by:
-
Ursula Braun authored
If a qeth device is plugged off, setting the device online stops in state HARDSETUP and a failure is reported to the base cio-layer causing halt/clear to be invoked. Replugging the device again triggers a qeth recovery without notification of the cio-layer. If a device is ungrouped in this state, the qeth set_offline function is not invoked, because the corresponding ccwgroup device is not in state ONLINE. Then incoming traffic is still handled by the qdio layer resulting in a crash in qeth_l<x>_qdio_input_handler, because (part of) the qeth data structures for this device are already removed. Solution: After replugging the device qeth recovery should lead to a working net device. Thus a "LAN offline" result when setting a qeth device online must not report a failure to the base cio-layer. Signed-off-by:
-
Rami Rosen authored
The netlink message header (struct nlmsghdr) is an unused parameter in fill method of fib_rules_ops struct. This patch removes this parameter from this method and fixes the places where this method is called. (include/net/fib_rules.h) Signed-off-by:
Rami Rosen <ramirose@gmail.com> Signed-off-by:
-
- 20 May, 2009 1 commit
-
-
Eric W. Biederman authored
Stephen Rothwell <sfr@canb.auug.org.au> writes: > Today's linux-next build of at least some av32 and arm configs failed like this: > > arch/avr32/kernel/signal.c:216: error: conflicting types for 'restart_syscall' > include/linux/sched.h:2184: error: previous definition of 'restart_syscall' was here > > Caused by commit 690cc3ff ("syscall: > Implement a convinience function restart_syscall") from the net tree. Grrr. Some days it feels like all of the good names are already taken. Let's just rename the two static users in arm and avr32 to get this sorted out. Signed-off-by:
-
- 19 May, 2009 25 commits
-
-
Ben Hutchings authored
GRO/LRO can be controlled through ethtool so this is unnecessary. Signed-off-by:
-
Sascha Hlusiak authored
be sent periodically. The rs_delay can be speficied when adding the PRL entry and defaults to 15 minutes. The RS is sent from every link local adress that's assigned to the tunnel interface. It's directed to the (guessed) linklocal address of the router and is sent through the tunnel. Better: send to ff02::2 encapsuled in unicast directed to router-v4. Signed-off-by:
Sascha Hlusiak <contact@saschahlusiak.de> Signed-off-by:
-
Sascha Hlusiak authored
A tunnel with no local ipv4 endpoint would otherwise use the ISATAP linklocal address fe80::5efe:0:0, which is invalid. Rather not add a linklocal address at all. Signed-off-by:
-
Sascha Hlusiak authored
Typo. When deleting a PRL entry, return status to userspace instead of success. Signed-off-by:
-
Sascha Hlusiak authored
Check link device when looking up a tunnel. When a tunnel is linked to a interface, traffic from a different interface must not reach the tunnel. This also allows creating of multiple tunnels with the same endpoints, if the link device differs. Signed-off-by:
-
Sascha Hlusiak authored
When locating the tunnel, do not continue if it is found. Otherwise a different tunnel with similar configuration would be returned and parts could be overwritten. Signed-off-by:
-
Chris Friesen authored
The DHCP spec allows the server to specify the MTU. This can be useful for netbooting with UDP-based NFS-root on a network using jumbo frames. This patch allows the kernel IP autoconfiguration to handle this option correctly. It would be possible to use initramfs and add a script to set the MTU, but that seems like a complicated solution if no initramfs is otherwise necessary, and would bloat the kernel image more than this code would. This patch was originally submitted to LKML in 2003 by Hans-Peter Jansen. Signed-off-by:
Chris Friesen <cfriesen@nortel.com> Signed-off-by:
-
Pablo Neira Ayuso authored
nlmsg_new() adds the size of the netlink header to the value that has been passed as parameter. If NLMSG_GOODSIZE is selected, we request an allocation of one memory page plus the size of the header. Instead, NLMSG_DEFAULT_SIZE should be used since it already substracts the size of the Netlink header. I have the impression that the similar naming in both constant is error prone when using it with nlmsg_new(). This is already documented in include/net/netlink.h Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
-
Brice Goglin authored
Update myri10ge driver version to 1.5.0-1.415. Signed-off-by:
Brice Goglin <brice@myri.com> Signed-off-by:
-
Brice Goglin authored
Allow myri10ge LRO to be enabled/disabled via ethtool (and by the stack for packet forwarding). Signed-off-by:
-
Eric Dumazet authored
We can slightly reduce size of teqlN structure, not duplicating stats structure in teql_master but using stats field from net_device.stats for tx_errors and from netdev_queue for tx_bytes/tx_packets/tx_dropped values. Signed-off-by:
Eric Dumazet <dada1@cosmosbay.com> Signed-off-by:
-
Jesse Brandeburg authored
This is purely a cleanup patch. This collapses some of the code required when we configure our Tx and Rx feature sets, and makes the code more readable and maintainable. Signed-off-by:
Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Peter P Waskiewicz Jr authored
The SFF specification for Direct Attach cable detection has now been ratified. Previously, DA cable detect was looking at the Twinaxial bit in byte 9 of the SFP+ EEPROM. The spec now defines active and passive DA cables in byte 8 of the SFP+ EEPROM. This patch changes the cable detection for both 82598 and 82599 SFP+ adapters to conform to the new spec. Signed-off-by:
-
Peter P Waskiewicz Jr authored
The SFP+ NIC (device id 0x10fb) needs a semaphore to serialize PHY access, so our PHY init code must honor that same semaphore. Signed-off-by:
-
françois romieu authored
Due to mostly historic reasons, including a lack of reliability of the link handling (especially with the older 8169), the current r8169 driver emulates forced mode setting by limiting the advertised modes. With this change the driver allows real 10/100 forced mode settings on the 8169 and 8101/8102. Original idea by Vincent Steenhoute. The RTL_GIGA_MAC_VER_03 tweak was extracted from Realtek's r8169 v6.010.00 driver. Signed-off-by:
Francois Romieu <romieu@fr.zoreil.com> Tested-by:
Jean Delvare <jdelvare@suse.de> Cc: Edward Hsu <edward_hsu@realtek.com.tw> Signed-off-by:
-
françois romieu authored
Signed-off-by:
-
Eric Dumazet authored
Jarek pointed pppoe can call back dev_queue_xmit(), and might need skb->dst, so its safer to unset IFF_XMIT_DST_RELEASE on ppp devices. Signed-off-by:
-
Eric Dumazet authored
One point of contention in high network loads is the dst_release() performed when a transmited skb is freed. This is because NIC tx completion calls dev_kree_skb() long after original call to dev_queue_xmit(skb). CPU cache is cold and the atomic op in dst_release() stalls. On SMP, this is quite visible if one CPU is 100% handling softirqs for a network device, since dst_clone() is done by other cpus, involving cache line ping pongs. It seems right place to release dst is in dev_hard_start_xmit(), for most devices but ones that are virtual, and some exceptions. David Miller suggested to define a new device flag, set in alloc_netdev_mq() (so that most devices set it at init time), and carefuly unset in devices which dont want a NULL skb->dst in their ndo_start_xmit(). List of devices that must clear this flag is : - loopback device, because it calls netif_rx() and quoting Patrick : "ip_route_input() doesn't accept loopback addresses, so loopback packets already need to have a dst_entry attached." - appletalk/ipddp.c : needs skb->dst in its xmit function - And all devices that call again dev_queue_xmit() from their xmit function (as some classifiers need skb->dst) : bonding, vlan, macvlan, eql, ifb, hdlc_fr Signed-off-by: -
Eric W. Biederman authored
Sysfs files for a network device can not unconditionally take the rtnl_lock as the bonding sysfs files do. If someone accesses those sysfs files while the network device is being unregistered with the rtnl_lock held we will deadlock. So use trylock and restart_syscall to avoid this problem. Signed-off-by:
-
Eric W. Biederman authored
Network device sysfs files that grab the rtnl_lock unconditionally will deadlock if accessed when the network device is being unregistered. So use trylock and syscall_restart to avoid this deadlock. Signed-off-by:
-
Eric W. Biederman authored
Holding rtnl_lock when we are unregistering the sysfs files can deadlock if we unconditionally take rtnl_lock in a sysfs file. So fix it with the now familiar patter of: rtnl_trylock and syscall_restart() Signed-off-by:
-
Eric W. Biederman authored
sysctls are unregistered with the rntl_lock held making it unsafe to unconditionally grab the the rtnl_lock. Instead we need to call rtnl_trylock and restart the system call if we can not grab it. Otherwise we could deadlock at unregistration time. Signed-off-by:
-
Eric W. Biederman authored
Just returning -ERESTARTSYS without a signal pending is not good that will just leak it to userspace. We need return -ERESTARTNOINTR so we always restart and set signal pending so that we fall of the fast path of syscall return and setup the system call restart. So use restart_syscall() which does all of this for us. Signed-off-by:
-
Eric W. Biederman authored
The earlier patch to fix the deadlock between a network device going away and writing to sysfs attributes was incomplete. - It did not set signal_pending so we would leak ERSTARTSYS to user space. - It used ERESTARTSYS which only restarts if sigaction configures it to. - It did not cover store and show for ifalias. So fix all of these up and use the new helper restart_syscall so we get the details correct on what it takes. Signed-off-by:
-
Eric W. Biederman authored
Currently when we have a signal pending we have the functionality to restart that the current system call. There are other cases such as nasty lock ordering issues where it makes sense to have a simple fix that uses try lock and restarts the system call. Buying time to figure out how to rework the locking strategy. Signed-off-by:
-
