summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2010-10-18Merge branch 'can/mcp251x-for-net-next' of ↵David S. Miller
git://git.pengutronix.de/git/mkl/linux-2.6
2010-10-18bnx2x: update version to 1.60.00-2Dmitry Kravkov
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18bnx2x: remove unnecessary FUNC_FLG_RSS flag and relatedDmitry Kravkov
As suggested by: Joe Perches <joe@perches.com> Although RSS is meaningless when there is a single HW queue we still need it enabled in order to have HW Rx hash generated. Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18bnx2x: Use correct FW constant for header paddingDmitry Kravkov
the value of the constant is the same, but it's clearer to use original constant provided by HSI Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18bnx2x: do not deal with power if no capabilityDmitry Kravkov
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18bnx2x: remove redundant commands during error handlingDmitry Kravkov
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18bnx2x: Optimized the branching in the bnx2x_rx_int()Vladislav Zolotarov
Optimized the branching in the bnx2x_rx_int() based on the fact that FP CQE will always have at least one of START or STOP flags set, so if not both bits are set and START bit is not set, then it's a STOP bit that is set. Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com> Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com> Signed-off-by: Eilon Greenstein <eilong@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18can: mcp251x: optimize 2515, rx int gets cleared automaticallyMarc Kleine-Budde
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: define helper functions mcp251x_is_2510, mcp251x_is_2515Marc Kleine-Budde
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: Don't use pdata->model for chip selection anymoreMarc Kleine-Budde
Since commit e446630c960946b5c1762e4eadb618becef599e7, i.e. v2.6.35-rc1, the mcp251x chip model can be selected via the modalias member in the struct spi_board_info. The driver stores the actual model in the struct mcp251x_platform_data. From the driver point of view the platform_data should be read only. Since all in-tree users of the mcp251x have already been converted to the modalias method, this patch moves the "model" member from the struct mcp251x_platform_data to the driver's private data structure. Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Cc: Christian Pellegrin <chripell@fsfe.org> Cc: Marc Zyngier <maz@misterjones.org>
2010-10-18can: mcp251x: write intf only when neededMarc Kleine-Budde
This patch introduces a variable "clear_intf" that hold the bits that should be cleared. Only read-modify-write register if "clear_intf" is set. Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: read-modify-write eflag only when neededSascha Hauer
Use read-modify-write instead of a simple write to change the register contents, to close existing the race window between the original manual read and write. Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: allow to read two registers in one spi transferSascha Hauer
This patch bases on work done earlier by David Jander. Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Acked-by: David Jander <david@protonic.nl> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: increase rx_errors on overflow, not only rx_over_errorsSascha Hauer
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18can: mcp251x: fix NOHZ local_softirq_pending 08 warningMarc Kleine-Budde
This patch replaces netif_rx() with netif_rx_ni() which has to be used from the threaded interrupt i.e. process context context. Thanks to Christian Pellegrin for pointing at the right fix: 481a8199142c050b72bff8a1956a49fd0a75bbe0 by Oliver Hartkopp. Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Acked-by: Wolfgang Grandegger <wg@grandegger.com>
2010-10-18ns83820: spin_lock_irq() => spin_lock()Dan Carpenter
This is essentially cosmetic. At this point the IRQs are already disabled because we called spin_lock_irq(&dev->rx_info.lock). The real bug here was fixed back in 2006 in 3a10ccebe: "[PATCH] lock validator: fix ns83820.c irq-flags bug". Prior to that patch, it was a "spin_lock_irq is not nestable" type bug. The 2006 patch changes the unlock to not re-enable IRQs, which eliminates the potential deadlock. But this bit was missed. We should change the lock function as well so it balances nicely. Signed-off-by: Dan Carpenter <error27@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18tipc: Simplify bearer shutdown logicAllan Stephens
Optimize processing in TIPC's bearer shutdown code, including: 1. Remove an unnecessary check to see if TIPC bearer's can exist. 2. Don't release spinlocks before calling a media-specific disabling routine, since the routine can't sleep. 3. Make bearer_disable() operate directly on a struct bearer, instead of needlessly taking a name and then mapping that to the struct. Signed-off-by: Allan Stephens <allan.stephens@windriver.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Reviewed-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-18tipc: Kill tipc_get_mode() completely.David S. Miller
It's completely unused and exporting a static symbol makes no sense and breaks the build. Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Update version to 3.115Matt Carlson
This patch updates the tg3 version to 3.115. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Report invalid link from tg3_get_settings()Matt Carlson
Currently the tg3 driver leaves the speed and duplex fields uninitialized in tg3_get_settings() if the device is not up. This can lead to some strange deductions in certain versions of ethtool. When the device is up and the link is down, the driver reports SPEED_INVALID and DUPLEX_INVALID for these fields. This patch makes the presentation consistent by returning SPEED_INVALID and DUPLEX_INVALID when the device has not been brought up as well. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Don't allocate jumbo ring for 5780 class devsMatt Carlson
The 5714, 5715, and 5780 devices do not have a separate rx jumbo producer ring. This patch changes the code so that resources are not allocated for it. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Cleanup tg3_alloc_rx_skb()Matt Carlson
src_map is no longer used in tg3_alloc_rx_skb(). Remove it. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Add EEE supportMatt Carlson
This patch adds Energy Efficient Ethernet (EEE) support for the 5718 device ID and the 57765 B0 asic revision. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Add clause 45 register accessor methodsMatt Carlson
This patch adds clause 45 register access methods. They will be used in the following patch. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Disable unused transmit ringsMatt Carlson
This patch allows the driver to disable the additional transmit rings available on the 5717 and 5719 devices. This is not strictly necessary, but is done anyways for correctness. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tg3: Add support for selfboot format 1 v6Matt Carlson
5718 B0 and 5719 devices will use a new selfboot firmware format. This patch adds code to detect the new format so that bootcode versions get reported correctly. Signed-off-by: Matt Carlson <mcarlson@broadcom.com> Reviewed-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17fib_hash: RCU conversion phase 2Eric Dumazet
Get rid of fib_hash_lock rwlock. The fn_zone hash table resize is the noticeable part of this patch. I added a seqlock per fn_zone, so that readers can restart their lookup in the (very rare) case a writer expanded the hash table. Add rcu heads in fib_alias and fib_node, use call_rcu() to defer their freeing, and use appropriate _rcu list manipulations. Stress test (160.000.000 udp frames sent, IP route cache disabled to mimic DDOS attack, FIB_HASH) Before: real 0m41.191s user 0m13.137s sys 8m55.241s After: real 0m38.091s user 0m13.189s sys 7m53.018s Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17fib_hash: RCU conversion phase 1Eric Dumazet
First step for RCU conversion of fib_hash : struct fn_zone are created and never deleted. Very classic conversion, using rcu_assign_pointer(), rcu_dereference() and rtnl_dereference() verbs. __rcu markers on fz_next and fn_zone_list They are created under RTNL, we dont need fib_hash_lock anymore in fn_new_zone(). Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17fib_hash: embed initial hash table in fn_zoneEric Dumazet
While looking for false sharing problems, I noticed sizeof(struct fn_zone) was small (28 bytes) and possibly sharing a cache line with an often written kernel structure. Most of the time, fn_zone uses its initial hash table of 16 slots. We can avoid the false sharing problem by embedding this initial hash table in fn_zone itself, so that sizeof(fn_zone) > L1_CACHE_BYTES We did a similar optimization in commit a6501e080c (Reduce memory needs and speedup lookups) Add a fz_revorder field to speedup fn_hash() a bit. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17drivers/net/pch_gbe: Use DEFINE_PCI_DEVICE_TABLEJoe Perches
Use the standard macro to put this table in __devinitconst. Compiled, untested. Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17netns: reorder fields in struct netEric Dumazet
In a network bench, I noticed an unfortunate false sharing between 'loopback_dev' and 'count' fields in "struct net". 'count' is written each time a socket is created or destroyed, while loopback_dev might be often read in routing code. Move loopback_dev in a read mostly section of "struct net" Note: struct netns_xfrm is cache line aligned on SMP. (It contains a "struct dst_ops") Move it at the end to avoid holes, and reduce sizeof(struct net) by 128 bytes on ia32. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tcp: use correct counters in CA_CWR state tooIlpo Järvinen
As CWR is stronger than CA_Disorder state, we can miscount SACK/Reno failure into other timeouts. Not a bad problem as it can happen only due to ECN, FRTO detecting spurious RTO or xmit error which are the only callers of tcp_enter_cwr. And even then losses and RTO must still follow thereafter to actually end up into the relevant code paths. Compile tested. Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17tcp: sack lost marking fixesIlpo Järvinen
When only fast rexmit should be done, tcp_mark_head_lost marks L too far. Also, sacked_upto below 1 is perfectly valid number, the packets == 0 then needs to be trapped elsewhere. Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17stmmac: remove ifdef NETIF_F_TSO from stmmac_ethtool.cGiuseppe Cavallaro
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com> Reported-by: Armando Visconti <armando.visconti@st.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-17niu: introduce temp variables to avoid sparse warnings when swapping in-situHarvey Harrison
Suppress a large block of warnings like: drivers/net/niu.c:7094:38: warning: incorrect type in assignment (different base types) drivers/net/niu.c:7094:38: expected restricted __be32 [usertype] ip4src drivers/net/niu.c:7094:38: got unsigned long long drivers/net/niu.c:7104:17: warning: cast from restricted __be32 ... Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16net: move MII outside of NET_ETHERNET, fix kconfig warningRandy Dunlap
We have USB, PCMCIA, and gigabit ethernet drivers that select MII even though NET_ETHERNET is not enabled, so make MII not be dependent on NET_ETHERNET. It is still dependent on NET and NETDEVICES. Fixes kconfig unmet dependency warning (shortened, was very long string): warning: (ARM_AT91_ETHER && NETDEVICES && NET_ETHERNET && ARM && ARCH_AT91RM9200 || ARM_KS8695_ETHER && NETDEVICES && NET_ETHERNET && ARM && ARCH_KS8695 || ... || IP1000 && NETDEVICES && NETDEV_1000 && PCI && EXPERIMENTAL || HAMACHI && NETDEVICES && NETDEV_1000 && PCI || R8169 && NETDEVICES && NETDEV_1000 && PCI || SIS190 && NETDEVICES && NETDEV_1000 && PCI || VIA_VELOCITY && NETDEVICES && NETDEV_1000 && PCI || ATL1 && NETDEVICES && NETDEV_1000 && PCI || ATL1E && NETDEVICES && NETDEV_1000 && PCI && EXPERIMENTAL || ATL1C && NETDEVICES && NETDEV_1000 && PCI && EXPERIMENTAL || JME && NETDEVICES && NETDEV_1000 && PCI || STMMAC_ETH && NETDEV_1000 && NETDEVICES && HAS_IOMEM || USB_PEGASUS && NETDEVICES && USB && NET || USB_RTL8150 && NETDEVICES && USB && NET && EXPERIMENTAL || USB_USBNET && NETDEVICES && USB && NET || PCMCIA_SMC91C92 && NETDEVICES && NET_PCMCIA && PCMCIA) selects MII which has unmet direct dependencies (NETDEVICES && NET_ETHERNET) Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Acked-by: Jeff Garzik <jgarzik@pobox.com> [2006-NOV-30] Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16infiniband: fix mlx4 kconfig dependency warningRandy Dunlap
Fix kconfig dependency warning to satisfy dependencies: warning: (MLX4_EN && NETDEVICES && NETDEV_10000 && PCI && INET || MLX4_INFINIBAND && INFINIBAND) selects MLX4_CORE which has unmet direct dependencies (NETDEVICES && NETDEV_10000 && PCI) Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16stmmac: make function tables conststephen hemminger
These tables only contain function pointers. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Gustavo F. Padovan <padovan@profusion.mobi> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16stmmac: make ethtool functions localstephen hemminger
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Gustavo F. Padovan <padovan@profusion.mobi> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16tipc: cleanup function namespacestephen hemminger
Do some cleanups of TIPC based on make namespacecheck 1. Don't export unused symbols 2. Eliminate dead code 3. Make functions and variables local 4. Rename buf_acquire to tipc_buf_acquire since it is used in several files Compile tested only. This make break out of tree kernel modules that depend on TIPC routines. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16via-velocity: forced 1000 Mbps mode support.françois romieu
Full duplex only. Half duplex 1000 Mbps is not supported. Signed-off-by: David Lv <DavidLv@viatech.com.cn> Acked-by: Francois Romieu <romieu@fr.zoreil.com> Tested-by: Seguier Regis <rseguier@e-teleport.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16fib: avoid false sharing on fib_table_hashEric Dumazet
While doing profile analysis, I found fib_hash_table was sometime in a cache line shared by a possibly often written kernel structure. (CONFIG_IP_ROUTE_MULTIPATH || !CONFIG_IPV6_MULTIPLE_TABLES) It's hard to detect because not easily reproductible. Make sure we allocate a full cache line to keep this shared in all cpus caches. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16fib_trie: use fls() instead of open coded loopEric Dumazet
fib_table_lookup() might use fls() to speedup an open coded loop. Noticed while doing a profile analysis. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16fib: remove a useless synchronize_rcu() callEric Dumazet
fib_nl_delrule() calls synchronize_rcu() for no apparent reason, while rtnl is held. I suspect it was done to avoid an atomic_inc_not_zero() in fib_rules_lookup(), which commit 7fa7cb7109d07 added anyway. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16fib6: use FIB_LOOKUP_NOREF in fib6_rule_lookup()Eric Dumazet
Avoid two atomic ops on found rule in fib6_rule_lookup() Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16sundance: Add initial ethtool stats supportDenis Kirjanov
Add ethtool stats support. Signed-off-by: Denis Kirjanov <dkirjanov@kernel.org> Acked-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16pch_gbe: fix if condition in set_settings()Dan Carpenter
There were no curly braces in this if condition so it always enabled full duplex. And ecmd->speed is an unsigned short so it is never equal to -1. The effect is that mii_ethtool_sset() fails with -EINVAL and an error is printed to dmesg. Signed-off-by: Dan Carpenter <error27@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16dnet: mark methods static and annotate for correct endiannessHarvey Harrison
Their doesn't appear to be bugs with the endianness handling here, just get the annotations right to keep sparse happy. Suppresses the following sparse warnings: drivers/net/dnet.c:30:5: warning: symbol 'dnet_readw_mac' was not declared. Should it be static? drivers/net/dnet.c:49:6: warning: symbol 'dnet_writew_mac' was not declared. Should it be static? drivers/net/dnet.c:364:5: warning: symbol 'dnet_phy_marvell_fixup' was not declared. Should it be static? drivers/net/dnet.c:66:13: warning: incorrect type in assignment (different base types) drivers/net/dnet.c:66:13: expected unsigned short [unsigned] [usertype] tmp drivers/net/dnet.c:66:13: got restricted __be16 [usertype] <noident> drivers/net/dnet.c:68:13: warning: incorrect type in assignment (different base types) drivers/net/dnet.c:68:13: expected unsigned short [unsigned] [usertype] tmp drivers/net/dnet.c:68:13: got restricted __be16 [usertype] <noident> drivers/net/dnet.c:70:13: warning: incorrect type in assignment (different base types) drivers/net/dnet.c:70:13: expected unsigned short [unsigned] [usertype] tmp drivers/net/dnet.c:70:13: got restricted __be16 [usertype] <noident> drivers/net/dnet.c:92:27: warning: cast to restricted __be16 drivers/net/dnet.c:94:33: warning: cast to restricted __be16 drivers/net/dnet.c:96:33: warning: cast to restricted __be16 Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16cxgb4vf: make single bit signed bitfields unsignedHarvey Harrison
Single bit signed bitfields don't make a lot of sense, noticed by sparse: drivers/net/cxgb4vf/t4vf_common.h:135:31: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:136:36: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:137:36: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:138:36: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:139:36: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:140:31: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:141:31: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:142:35: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:143:35: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:154:27: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:155:26: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:156:27: error: dubious one-bit signed bitfield drivers/net/cxgb4vf/t4vf_common.h:157:26: error: dubious one-bit signed bitfield Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-10-16net: allocate skbs on local nodeEric Dumazet
commit b30973f877 (node-aware skb allocation) spread a wrong habit of allocating net drivers skbs on a given memory node : The one closest to the NIC hardware. This is wrong because as soon as we try to scale network stack, we need to use many cpus to handle traffic and hit slub/slab management on cross-node allocations/frees when these cpus have to alloc/free skbs bound to a central node. skb allocated in RX path are ephemeral, they have a very short lifetime : Extra cost to maintain NUMA affinity is too expensive. What appeared as a nice idea four years ago is in fact a bad one. In 2010, NIC hardwares are multiqueue, or we use RPS to spread the load, and two 10Gb NIC might deliver more than 28 million packets per second, needing all the available cpus. Cost of cross-node handling in network and vm stacks outperforms the small benefit hardware had when doing its DMA transfert in its 'local' memory node at RX time. Even trying to differentiate the two allocations done for one skb (the sk_buff on local node, the data part on NIC hardware node) is not enough to bring good performance. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Acked-by: Tom Herbert <therbert@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
generated by cgit v0.10.2 at 2024-11-15 19:35:17 (GMT)