- 04 Oct, 2010 13 commits
-
-
Eric Dumazet authored
While doing stress tests with IP route cache disabled, and multi queue devices, I noticed a very high contention on one rwlock used in neighbour code. When many cpus are trying to send frames (possibly using a high performance multiqueue device) to the same neighbour, they fight for the neigh->lock rwlock in order to call neigh_hh_init(), and fight on hh->hh_refcnt (a pair of atomic_inc/atomic_dec_and_test()) But we dont need to call neigh_hh_init() for dst that are used only once. It costs four atomic operations at least, on two contended cache lines, plus the high contention on neigh->lock rwlock. Introduce a new dst flag, DST_NOCACHE, that is set when dst was not inserted in route cache. With the stress test bench, sending 160000000 frames on one neighbour, results are : Before patch: real 2m28.406s user 0m11.781s sys 36m17.964s After patch: real 1m26.532s user 0m12.185s sys 20m3.903s Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
David S. Miller authored
Signed-off-by: -
Sathya Perla authored
This patch adds multiple RX queue support to be2net. There are upto 4 extra rx-queues per port into which TCP/UDP traffic can be hashed into. Some of the ethtool stats are now displayed on a per queue basis. Signed-off-by:
Sathya Perla <sathya.perla@emulex.com> Signed-off-by:
-
David S. Miller authored
Merge branch 'for-davem' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next-2.6
-
Ursula Braun authored
This patch adapts qeth to handle tagged frames with VLAN-ID 0 and with or without priority information in the tag. It enables qeth to receive priority-tagged frames on a base interface, for example from z/OS, without configuring an additional VLAN interface. Signed-off-by:
Ursula Braun <ursula.braun@de.ibm.com> Signed-off-by:
Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by:
-
Dimitris Michailidis authored
Remove a bogus PCI function number check from the driver's .remove method that causes pci_release_regions not to be called for function 0 if additional functions are attached and one of them is used as primary. Signed-off-by:
Dimitris Michailidis <dm@chelsio.com> Signed-off-by:
-
Julia Lawall authored
This code does not call deinit_card(card); in an error case, as done in other error-handling code in the same function. But actually, the called function init_sram can only return 0, so there is no need for the error check at all. init_sram is also given a void return type, and its single return statement at the end of the function is dropped. A simplified version of the sematic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @r exists@ @r@ statement S1,S2,S3; constant C1,C2,C3; @@ *if (...) {... S1 return -C1;} ... *if (...) {... when != S1 return -C2;} ... *if (...) {... S1 return -C3;} // </smpl> Signed-off-by:
Julia Lawall <julia@diku.dk> Signed-off-by:
-
Andrew Morton authored
ERROR: trailing statements should be on next line #23: FILE: drivers/net/tulip/de4x5.c:5477: + if (copy_to_user(ioc->data, tmp.lval, ioc->len)) return -EFAULT; total: 1 errors, 0 warnings, 8 lines checked ./patches/drivers-net-tulip-de4x5c-fix-copy-length-in-de4x5_ioctl.patch has style problems, please review. If any of these errors are false positives report them to the maintainer, see CHECKPATCH in MAINTAINERS. Please run checkpatch prior to sending patches Cc: Dan Rosenberg <dan.j.rosenberg@gmail.com> Cc: Grant Grundler <grundler@parisc-linux.org> Cc: Jeff Mahoney <jeffm@suse.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
-
Eric Dumazet authored
Various code style cleanups Signed-off-by:
-
Eric Dumazet authored
Use RCU & RTNL protection for mfc_cache_array[] ipmr_cache_find() is called under rcu_read_lock(); Signed-off-by:
-
Eric Dumazet authored
Use RCU and RTNL to protect (struct mr_table)->mroute_sk Readers use RCU, writers use RTNL. ip_ra_control() already use an RCU grace period before ip_ra_destroy_rcu(), so we dont need synchronize_rcu() in mrtsock_destruct() Signed-off-by:
-
Eric Dumazet authored
No need to get a reference on reg_dev and release it, we are in a rcu_read_lock() protected section. Signed-off-by:
-
stephen hemminger authored
This table is only used in gre.c Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Signed-off-by:
-
- 01 Oct, 2010 27 commits
-
-
John W. Linville authored
Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next-2.6 into for-davem
-
Vasanthy Kolluri authored
Update MAINTAINERS list Signed-off-by:
Vasanthy Kolluri <vkolluri@cisco.com> Signed-off-by:
Roopa Prabhu <roprabhu@cisco.com> Signed-off-by:
David Wang <dwang2@cisco.com> Signed-off-by:
-
Vasanthy Kolluri authored
Make functions used locally in a file as static Signed-off-by:
-
Vasanthy Kolluri authored
Removed code that is unused Signed-off-by:
-
Eric Dumazet authored
On 64bit arches, there are two 32bit holes that we can remove. sizeof(struct neighbour) shrinks from 0xf8 to 0xf0 bytes Signed-off-by:
-
Tilman Schmidt authored
Rephrase some USB error messages to make them clearer and more consistent. Downgrade some warning messages that may occur during normal operation to debug messages. Signed-off-by:
Tilman Schmidt <tilman@imap.cc> Signed-off-by:
-
Tilman Schmidt authored
Rework the handling of USB errors in interrupt input reads to clear halts correctly, delay URB resubmission after errors, limit retries, and improve error recovery. Signed-off-by:
-
Tilman Schmidt authored
Recover from a lost HD_RECEIVEATDATA_ACK message by sending a zero-length HD_READ_ATMESSAGE command when ev_layer sends "+++". Signed-off-by:
-
Tilman Schmidt authored
In error_reset(), if sending HD_RESET_INTERRUPT_PIPE to the device fails, try performing an USB reset. Also correct an error in the leading comment. Signed-off-by:
-
Tilman Schmidt authored
Use setup_timer() and mod_timer() instead of direct assignment to timer structure members, simplify the argument of one timer routine, and make extra sure all timers are stopped during suspend. Signed-off-by:
-
Tilman Schmidt authored
Remove the debug flag DEBUG_DRIVER and associated code. It doesn't serve any useful purpose anymore. Signed-off-by:
-
Tilman Schmidt authored
In transparent data reception, avoid a NULL pointer dereference in case an skbuff cannot be allocated, remove an inappropriate call to the HDLC flush routine, and correct the accounting of received bytes for continued buffers. Signed-off-by:
-
Tilman Schmidt authored
Rework the handling of USB errors in AT response reads to fix a possible infinite retry loop and a memory leak, and silence a few overly verbose kernel messages. Signed-off-by:
-
Tilman Schmidt authored
Unlock cs->lock before calling error_hangup() which is marked "cs->lock must not be held". Signed-off-by:
-
Matt Carlson authored
This patch updates the tg3 version to 3.114. Signed-off-by:
Matt Carlson <mcarlson@broadcom.com> Reviewed-by:
Michael Chan <mchan@broadcom.com> Signed-off-by:
-
Matt Carlson authored
This patch increases the rx ring sizes for those asic revs that support them. Signed-off-by:
-
Matt Carlson authored
This patch adds two new variables to track the size of the standard and jumbo rx producer ring sizes. The code is then pivoted to these variables from preprocessor constants. Signed-off-by:
-
Matt Carlson authored
There are other multiqueue modes 5717 and 5719 devices can assume. This patch makes sure that the loopback test is safe, should those other modes be enabled in the future. Signed-off-by:
-
Matt Carlson authored
This patch cleans up the default VPD partno section. New entries for 5717 asic rev devices were also added. Signed-off-by:
-
Matt Carlson authored
This product was never released to the public. Signed-off-by:
-
Matt Carlson authored
This patch enables a bit that prevents read DMA overflows and adjusts the txmbuf margin from the hardware default. The combination of these modifications prevents a tx data corruption issue we were seeing on the 5719. Signed-off-by:
-
Matt Carlson authored
Up until now the tg3 driver would call netif_napi_add() for the maximum number of NAPI instances the driver could use. The problem is that netpoll could call tg3_poll() on instances that are not active. The net effect is that the driver will crash attempting to dereference uninitialized pointers. The fix is to only allocate as many NAPI instances as the driver would use in tg3_open() and deleted them in tg3_close(). Signed-off-by:
-
Eric Dumazet authored
ip_route_output_slow() is enclosed in an rcu_read_lock() protected section, so that no references are taken/released on device, thanks to __ip_dev_find() & dev_get_by_index_rcu() Tested with ip route cache disabled, and a stress test : Before patch: elapsed time : real 1m38.347s user 0m11.909s sys 23m51.501s Profile: 13788.00 22.7% ip_route_output_slow [kernel] 7875.00 13.0% dst_destroy [kernel] 3925.00 6.5% fib_semantic_match [kernel] 3144.00 5.2% fib_rules_lookup [kernel] 3061.00 5.0% dst_alloc [kernel] 2276.00 3.7% rt_set_nexthop [kernel] 1762.00 2.9% fib_table_lookup [kernel] 1538.00 2.5% _raw_read_lock [kernel] 1358.00 2.2% ip_output [kernel] After patch: real 1m28.808s user 0m13.245s sys 20m37.293s 10950.00 17.2% ip_route_output_slow [kernel] 10726.00 16.9% dst_destroy [kernel] 5170.00 8.1% fib_semantic_match [kernel] 3937.00 6.2% dst_alloc [kernel] 3635.00 5.7% rt_set_nexthop [kernel] 2900.00 4.6% fib_rules_lookup [kernel] 2240.00 3.5% fib_table_lookup [kernel] 1427.00 2.2% _raw_read_lock [kernel] 1157.00 1.8% kmem_cache_alloc [kernel] Signed-off-by:
-
Eric Dumazet authored
ip_dev_find(net, addr) finds a device given an IPv4 source address and takes a reference on it. Introduce __ip_dev_find(), taking a third argument, to optionally take the device reference. Callers not asking the reference to be taken should be in an rcu_read_lock() protected section. Signed-off-by:
-
Bruce Allan authored
The initial support for 82579 was tuned poorly for performance. Adjust the packet buffer allocation appropriately for both standard and jumbo frames; and for jumbo frames increase the receive descriptor pre-fetch, disable adaptive interrupt moderation and set the DMA latency tolerance. Signed-off-by:
Bruce Allan <bruce.w.allan@intel.com> Tested-by:
Jeff Pieper <jeffrey.e.pieper@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by:
-
Jesse Brandeburg authored
Most e1000e parts support batching writebacks. The problem with this is that when some of the TADV or TIDV timers are not set, Tx can sit forever. This is solved in this patch with write flushes using the Flush Partial Descriptors (FPD) bit in TIDV and RDTR. This improves bus utilization and removes partial writes on e1000e, particularly from 82571 parts in S5500 chipset based machines. Only ES2LAN and 82571/2 parts are included in this optimization, to reduce testing load. Signed-off-by:
Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by:
Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by:
-
Emil Tantilov authored
Fix possible panic/hang with shared Legacy interrupts by not enabling interrupts when interface is down. Also fixes an intermittent link by enabling LSC upon exit from ixgbe_intr() This patch adds flags to ixgbe_irq_enable() to allow for some flexibility when enabling interrupts. Signed-off-by:
Stephen Ko <stephen.s.ko@intel.com> Signed-off-by:
-
