Jon Mason [Fri, 10 Dec 2010 15:40:02 +0000 (15:40 +0000)]
s2io: make strings at tables const
Put immutable data in read/only section.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 15:40:01 +0000 (15:40 +0000)]
s2io: rx_ring_sz bounds checking
modparm rx_ring_sz can be set to be greater than the maximum allowable
number of blocks. This results in an array overrun when probing the
driver, and causes memory corruption.
Also, the MAX_RX_DESC_1 multiply the max number of rings by max number
of blocker per ring by 127, but the driver does the same calculation
with 127+1. This results in the possibility of the value being set
being larger than the maximum allowable value.
Finally, clean-up the s2io_ethtool_gringparam code to be more
intuitive.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Carolyn Wyborny [Mon, 22 Nov 2010 17:17:21 +0000 (17:17 +0000)]
igb: Add new function to read part number from EEPROM in string format
New adapters will have part numbers stored in string format rather than
simple hex format. This function will read part number formats in either
hex or string.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Wed, 1 Dec 2010 08:45:24 +0000 (08:45 +0000)]
e1000e: increment the driver version
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>
Bruce Allan [Thu, 9 Dec 2010 23:04:25 +0000 (23:04 +0000)]
e1000e: static analysis tools complain of a possible null ptr p dereference
Adding this default case resolves the issue.
v2- Removed "break" in default case based on feedback
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Wed, 24 Nov 2010 06:02:06 +0000 (06:02 +0000)]
e1000e: minor error message corrections
Correct error messages when setting up Rx resources and when checking
module parameters.
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>
Bruce Allan [Wed, 24 Nov 2010 07:42:43 +0000 (07:42 +0000)]
e1000e: prevent null ptr dereference in e1000_tx_queue()
tx_desc can be dereferenced as a null pointer when count is passed in
as 0.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Wed, 24 Nov 2010 06:01:51 +0000 (06:01 +0000)]
e1000e: support new PBA format from EEPROM
Provide support to e1000e for displaying the new format of the PBA found
in the EEPROM. The unique PBA identifier is no longer restricted to
hexadecimal numbers and must now be read and displayed as a string.
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>
Bruce Allan [Wed, 24 Nov 2010 06:01:46 +0000 (06:01 +0000)]
e1000e: 82579 PHY incorrectly identified during init
During init, reading the 2 PHY ID registers back-to-back in the default
fast mode could return invalid data (all F's) and in slow mode could
return data to the second read the data from the first read. To resolve
the issue in fast mode, set to slow mode before any PHY accesses; to
resolve the issue in slow mode, put in a delay for every 82579 PHY access.
Since this PHY is currently only paired with the pch2lan MAC and the PHY
type is not known before the first PHY access which can fail this way,
check for this based on MAC-type.
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>
Bruce Allan [Wed, 24 Nov 2010 06:01:41 +0000 (06:01 +0000)]
e1000e: 82577/8/9 mis-configured OEM bits during S0->Sx
The LPLU (Low Power Link Up) and Gigabit Disable bits (a.k.a. OEM bits)
were being configured incorrectly when device goes to D3 state.
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>
Bruce Allan [Wed, 24 Nov 2010 06:01:35 +0000 (06:01 +0000)]
e1000e: 82571 Serdes can fail to get link
When link partner is sending continuous Config symbols, the 82571 Serdes
FIFO can overflow resulting in Invalid bit getting set. To resolve this,
if Sync and Config bits are both 1 ignore the Invalid bit and restart
auto-negotiation.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Wed, 24 Nov 2010 06:01:30 +0000 (06:01 +0000)]
e1000e: 82577/8 must acquire h/w semaphore before workaround
The workaround function e1000_configure_k1_pchlan() assumes the h/w
semaphore is already acquired. This was originally missed when setting up
the part for the ethtool loopback test.
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>
Bruce Allan [Wed, 24 Nov 2010 06:01:25 +0000 (06:01 +0000)]
e1000e: 82574/82583 performance improvement
Increasing the transmit fifo by 4K (by decreasing the receive fifo size
specified in .pba by the same amount) increases Tx performance.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Bruce Allan [Wed, 24 Nov 2010 06:01:20 +0000 (06:01 +0000)]
e1000e: 82571-based mezzanine card can fail ethtool link test
On certain 82571-based mezzanine NICs in some blade servers, the ethtool
link test can fail due to the serdes_has_link flag not set correctly.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The kernel goes BUG() at the time 'ethtool -p eth0 3' comes
back, which is due to adapter->led_blink_task initialized
several times. At the time it is still running this results
in a corrupted task_list of the associated workqueue.
The fix is to move the workqueue initialization to the
probe function instead.
Signed-off-by: Holger Eitzenberger <holger@eitzenberger.org> Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Fri, 3 Dec 2010 09:33:54 +0000 (09:33 +0000)]
ixgbe: cleanup string function calls to use bound checking versions.
Some minor cleanup to use string calls that use bound checks just to
be extra safe.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Stephen Ko <stephen.s.ko@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch puts it back on to allow for setting ntuple via ethtool.
CC: Jesse Gross <jesse@nicira.com> Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com> Tested-by: Stephen Ko <stephen.s.ko@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Thu, 9 Dec 2010 06:55:19 +0000 (06:55 +0000)]
ixgbe: fix X540 to use it's own info struct
This patch enables X540 hardware to use it's own set of support
functions. This is useful as it has no need of SFP+ support. A
couple minor bugs with the eeprom semaphore were also cleaned up.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Stephen Ko <stephen.s.ko@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Don Skidmore [Thu, 9 Dec 2010 06:55:14 +0000 (06:55 +0000)]
ixgbe: fix X540 phy id to correct value
The existing PHY ID for X540 was from early production hardware and
is no longer correct. This patch corrects that.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com> Tested-by: Stephen Ko <stephen.s.ko@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Dean Nelson [Thu, 11 Nov 2010 05:50:25 +0000 (05:50 +0000)]
e1000: fix return value not set on error
Dean noticed that 'err' wasn't being set when the "goto err_dma"
statement is executed in the following hunk from the commit. It's value
will be zero as a result of a successful call to e1000_init_hw_struct().
This patch changes the error condition to be correctly propagated.
CC: stable@kernel.org Signed-off-by: Dean Nelson <dnelson@redhat.com> Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Tested-by: Emil Tantilov <emil.s.tantilov@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Taku Izumi [Thu, 9 Dec 2010 15:17:13 +0000 (15:17 +0000)]
bonding: add the debugfs facility to the bonding driver
This patch provides the debugfs facility to the bonding driver.
The "bonding" directory is created in the debugfs root and directories of
each bonding interface (like bond0, bond1...) are created in that.
# mount -t debugfs none /sys/kernel/debug
# ls /sys/kernel/debug/bonding
bond0 bond1
Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com> Signed-off-by: Jay Vosburgh <fubar@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Roopa Prabhu [Fri, 10 Dec 2010 12:02:33 +0000 (12:02 +0000)]
enic: Move enic port profile handling code to a new 802.1Qbh provisioning info type
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com> Signed-off-by: David Wang <dwang2@cisco.com> Signed-off-by: Christian Benvenuti <benve@cisco.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 14:03:01 +0000 (14:03 +0000)]
vxge: independent interrupt moderation
Configure the workload clock register and TIM register for independent
interrupt moderation based on the individual vpath utilization instead
of common link utilization. This greatly improves latency.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: Ram Vepa <ram.vepa@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 14:03:00 +0000 (14:03 +0000)]
vxge: hotplug stall
When hot-unplugging a vxge adapter while running, the driver's remove
routine prints warning and then stalls the calling thread. This is due
to vxge_remove calling vxge_device_unregister to unregister the netdev
before calling flush_scheduled_work clear any pending work. Swapping
the order of these two functions resolves the issue.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: Ram Vepa <ram.vepa@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 14:02:59 +0000 (14:02 +0000)]
vxge: transmit timeout deadlock
Use a workqueue to handle the device reset during a transmit timeout, as
there can be a deadlock during bringup. Also, set the netif carrier off
before the watchdog reset is started to prevent the timeout from
reoccurring while still processing the first.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: Ram Vepa <ram.vepa@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 14:02:58 +0000 (14:02 +0000)]
vxge: use pci_request_region()
Only BAR0 is ever accessed, thus making the calls to pci_request_regions
overkill. Change calls of pci_request_regions to pci_request_region to
reduce the size of the mapped area.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: Ram Vepa <ram.vepa@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jon Mason [Fri, 10 Dec 2010 14:02:57 +0000 (14:02 +0000)]
vxge: fix crash of VF when unloading PF
Calling pci_disable_sriov when unloading a SR-IOV physical function
driver from a host when a guest is using a virtual function from that
device can cause a host crash or VM crash. The crash is caused by the
virtual config space no longer being present when PF is removed (due to
the pci_disable_sriov). This can be avoided by not calling
pci_disable_sriov to disable the PCI space when shutting down the PF.
Each function in the X3100 operates independently and in this case will
operate properly in the absence of the PF.
Also, added improved logic in the detection of SR-IOV initialization.
Signed-off-by: Jon Mason <jon.mason@exar.com> Signed-off-by: Ram Vepa <ram.vepa@exar.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu [Thu, 9 Dec 2010 17:38:11 +0000 (17:38 +0000)]
bridge: Use consistent NF_DROP returns in nf_pre_routing
The nf_pre_routing functions in bridging have collected two
distinct ways of returning NF_DROP over the years, inline and
via goto. There is no reason for preferring either one.
So this patch arbitrarily picks the inline variant and converts
the all the gotos.
Also removes a redundant comment.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings [Thu, 9 Dec 2010 12:10:25 +0000 (12:10 +0000)]
netdev: Use default implementation of ethtool_ops::get_link where possible
Various drivers are using implementations of ethtool_ops::get_link
that are equivalent to the default ethtool_op_get_link(). Change
them to use that instead.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ben Hutchings [Thu, 9 Dec 2010 12:08:35 +0000 (12:08 +0000)]
ethtool: Report link-down while interface is down
While an interface is down, many implementations of
ethtool_ops::get_link, including the default, ethtool_op_get_link(),
will report the last link state seen while the interface was up. In
general the current physical link state is not available if the
interface is down.
Define ETHTOOL_GLINK to reflect whether the interface *and* any
physical port have a working link, and consistently return 0 when the
interface is down.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Roopa Prabhu [Wed, 8 Dec 2010 13:54:03 +0000 (13:54 +0000)]
enic: Use VF mac set by IFLA_VF_MAC in port profile provisioning data
This patch adds support in enic 802.1Qbh port profile provisioning code
to use the mac address set by IFLA_VF_MAC. For now we handle this mac as a
special case for a VM mac address sent to us by libvirt. The VM mac address
is sent to the switch along with the rest of the port profile provisioning
data. This patch also adds calls to register and deregister the mac address
during port profile association/deassociation.
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com> Signed-off-by: David Wang <dwang2@cisco.com> Signed-off-by: Christian Benvenuti <benve@cisco.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Roopa Prabhu [Wed, 8 Dec 2010 13:53:58 +0000 (13:53 +0000)]
enic: Add ndo_set_vf_mac support for enic dynamic devices
This patch implements the ndo_set_vf_mac netdev operation for enic
dynamic devices. It treats the mac address set by IFLA_VF_MAC as a
special case to use it in the port profile provisioning data.
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com> Signed-off-by: David Wang <dwang2@cisco.com> Signed-off-by: Christian Benvenuti <benve@cisco.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Roopa Prabhu [Wed, 8 Dec 2010 13:19:58 +0000 (13:19 +0000)]
enic: Add ndo_set_rx_mode support for enic vnics
Add ndo_set_rx_mode support to register unicast and multicast
address filters for enic devices
Signed-off-by: Roopa Prabhu <roprabhu@cisco.com> Signed-off-by: Vasanthy Kolluri <vkolluri@cisco.com> Signed-off-by: David Wang <dwang2@cisco.com> Signed-off-by: Christian Benvenuti <benve@cisco.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Junchang Wang [Wed, 8 Dec 2010 16:55:16 +0000 (16:55 +0000)]
pktgen: adding prefetchw() call
We know for sure pktgen is going to write skb->data right after
*_alloc_skb, causing unnecessary cache misses.
Idea is to add a prefetchw() call to prefetch the first cache line
indicated by skb->data. On systems with Adjacent Cache Line Prefetch,
it's probably two cache lines are prefetched.
With this patch, pktgen on Intel SR1625 server with two E5530
quad-core processors and a single ixgbe-based NIC went from 8.63Mpps
to 9.03Mpps, with 4.6% improvement.
Signed-off-by: Junchang Wang <junchangwang@gmail.com> Acked-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This driver now uses Generic Receive Offload, not the older LRO.
Change references to LRO in names and comments.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Acked-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Martin Willi [Wed, 8 Dec 2010 04:37:51 +0000 (04:37 +0000)]
xfrm: Traffic Flow Confidentiality for IPv6 ESP
Add TFC padding to all packets smaller than the boundary configured
on the xfrm state. If the boundary is larger than the PMTU, limit
padding to the PMTU.
Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Martin Willi [Wed, 8 Dec 2010 04:37:50 +0000 (04:37 +0000)]
xfrm: Traffic Flow Confidentiality for IPv4 ESP
Add TFC padding to all packets smaller than the boundary configured
on the xfrm state. If the boundary is larger than the PMTU, limit
padding to the PMTU.
Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
The XFRMA_TFCPAD attribute for XFRM state installation configures
Traffic Flow Confidentiality by padding ESP packets to a specified
length.
Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Jan Glauber [Wed, 8 Dec 2010 02:58:01 +0000 (02:58 +0000)]
qeth: buffer count imbalance
The used buffers counter is not incremented in case of an error so
the counter can become negative. Increment the used buffers counter
before checking for errors.
Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Einar Lueck [Wed, 8 Dec 2010 02:57:59 +0000 (02:57 +0000)]
qeth: support VIPA add/del in offline mode
Only work through the IP adddress to do list if the card is UP or
SOFTSETUP. Enables to configure VIPA add/del in offline mode.
Signed-off-by: Einar Lueck <elelueck@de.ibm.com> Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Einar Lueck [Wed, 8 Dec 2010 02:57:58 +0000 (02:57 +0000)]
qeth: support ipv6 query arp cache for HiperSockets
Function qeth_l3_arp_query now queries for IPv6 addresses, too, if
QETH_QARP_WITH_IPV6 is passed as parameter to the ioctl. HiperSockets
and GuestLAN in HiperSockets mode provide corresponding entries.
Signed-off-by: Einar Lueck <elelueck@de.ibm.com> Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Tobias Klauser [Fri, 10 Dec 2010 03:18:04 +0000 (03:18 +0000)]
bridge: Fix return values of br_multicast_add_group/br_multicast_new_group
If br_multicast_new_group returns NULL, we would return 0 (no error) to
the caller of br_multicast_add_group, which is not what we want. Instead
br_multicast_new_group should return ERR_PTR(-ENOMEM) in this case.
Also propagate the error number returned by br_mdb_rehash properly.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Shan Wei [Fri, 10 Dec 2010 11:49:23 +0000 (12:49 +0100)]
dccp: remove unused macros
Remove macros which have been unused since the initial implementation
(commit 7c657876b63cb1d8a2ec06f8fc6c37bb8412e66c, [DCCP]: Initial
implementation from Tue Aug 9 20:14:34 2005 -0700).
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>
bnx2x_src_init_t2() is used only when BCM_CNIC is defined.
So, to avoid a compilation warning, we won't define it unless
BCM_CNIC is defined.
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>
bnx2x: Use dma_alloc_coherent() semantics for ILT memory allocation
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>
Make the LSO code work on BE platforms: parsing_data field of
a parsing BD (PBD) for 57712 was improperly composed which made FW read wrong
values for TCP header's length and offset and, as a result, the corresponding
PCI device was performing bad DMA reads triggering EEH.
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>
The current jhash.h implements the lookup2() hash function by Bob Jenkins.
However, lookup2() is outdated as Bob wrote a new hash function called
lookup3(). The patch replaces the lookup2() implementation of the 'jhash*'
functions with that of lookup3().
You can read a longer comparison of the two and other hash functions at
http://burtleburtle.net/bob/hash/doobs.html.
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu> Acked-by: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 30 Nov 2010 19:04:07 +0000 (19:04 +0000)]
net: optimize INET input path further
Followup of commit b178bb3dfc30 (net: reorder struct sock fields)
Optimize INET input path a bit further, by :
1) moving sk_refcnt close to sk_lock.
This reduces number of dirtied cache lines by one on 64bit arches (and
64 bytes cache line size).
2) moving inet_daddr & inet_rcv_saddr at the beginning of sk
(same cache line than hash / family / bound_dev_if / nulls_node)
This reduces number of accessed cache lines in lookups by one, and dont
increase size of inet and timewait socks.
inet and tw sockets now share same place-holder for these fields.
David S. Miller [Thu, 9 Dec 2010 05:16:57 +0000 (21:16 -0800)]
net: Abstract away all dst_entry metrics accesses.
Use helper functions to hide all direct accesses, especially writes,
to dst_entry metrics values.
This will allow us to:
1) More easily change how the metrics are stored.
2) Implement COW for metrics.
In particular this will help us put metrics into the inetpeer
cache if that is what we end up doing. We can make the _metrics
member a pointer instead of an array, initially have it point
at the read-only metrics in the FIB, and then on the first set
grab an inetpeer entry and point the _metrics member there.
Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Eric Dumazet [Tue, 7 Dec 2010 12:20:47 +0000 (12:20 +0000)]
tcp: protect sysctl_tcp_cookie_size reads
Make sure sysctl_tcp_cookie_size is read once in
tcp_cookie_size_check(), or we might return an illegal value to caller
if sysctl_tcp_cookie_size is changed by another cpu.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Ben Hutchings <bhutchings@solarflare.com> Cc: William Allen Simpson <william.allen.simpson@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 7 Dec 2010 12:03:55 +0000 (12:03 +0000)]
tcp: avoid a possible divide by zero
sysctl_tcp_tso_win_divisor might be set to zero while one cpu runs in
tcp_tso_should_defer(). Make sure we dont allow a divide by zero by
reading sysctl_tcp_tso_win_divisor exactly once.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Breno Leitao [Wed, 8 Dec 2010 20:19:14 +0000 (12:19 -0800)]
ehea: Fixing LRO configuration
In order to set LRO on ehea, the user must set a module parameter, which
is not the standard way to do so. This patch adds a way to set LRO using
the ethtool tool.
Signed-off-by: Breno Leitao <leitao@linux.vnet.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Tom Herbert [Wed, 8 Dec 2010 20:16:33 +0000 (12:16 -0800)]
tcp: Replace time wait bucket msg by counter
Rather than printing the message to the log, use a mib counter to keep
track of the count of occurences of time wait bucket overflow. Reduces
spam in logs.
Signed-off-by: Tom Herbert <therbert@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
x25 does not decrement the network device reference counts on module unload.
Thus unregistering any pre-existing interface after unloading the x25 module
hangs and results in
unregister_netdevice: waiting for tap0 to become free. Usage count = 1
This patch decrements the reference counts of all interfaces in x25_link_free,
the way it is already done in x25_link_device_down for NETDEV_DOWN events.
Signed-off-by: Apollon Oikonomopoulos <apollon@noc.grnet.gr> Signed-off-by: David S. Miller <davem@davemloft.net>
Regarding benet be_cmd_multicast_set() function, now using
netdev_for_each_mc_addr() helper for mac address copy, but
when copying to req->mac[] did not increase of the index.
Cc: Sathya Perla <sathyap@serverengines.com> Cc: Subbu Seetharaman <subbus@serverengines.com> Cc: Sarveshwar Bandi <sarveshwarb@serverengines.com> Cc: Ajit Khaparde <ajitk@serverengines.com> Signed-off-by: Joe Jin <joe.jin@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Roger Luethi [Mon, 6 Dec 2010 00:59:40 +0000 (00:59 +0000)]
via-rhine: hardware VLAN support
This patch adds VLAN hardware support for Rhine chips.
The driver uses up to 3 additional bytes of buffer space when extracting
802.1Q headers; PKT_BUF_SZ should still be sufficient.
The initial code was provided by David Lv. I reworked it to use standard
kernel facilities. Coding style clean up mostly follows via-velocity.
Adapted to new interface for VLAN acceleration (per request of Jesse Gross).
Signed-off-by: David Lv <DavidLv@viatech.com.cn> Signed-off-by: Roger Luethi <rl@hellgate.ch>
drivers/net/via-rhine.c | 326 +++++++++++++++++++++++++++++++++++++++++++++--
1 files changed, 312 insertions(+), 14 deletions(-) Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Sun, 5 Dec 2010 01:23:53 +0000 (01:23 +0000)]
net: RCU conversion of dev_getbyhwaddr() and arp_ioctl()
Le dimanche 05 décembre 2010 à 09:19 +0100, Eric Dumazet a écrit :
> Hmm..
>
> If somebody can explain why RTNL is held in arp_ioctl() (and therefore
> in arp_req_delete()), we might first remove RTNL use in arp_ioctl() so
> that your patch can be applied.
>
> Right now it is not good, because RTNL wont be necessarly held when you
> are going to call arp_invalidate() ?
While doing this analysis, I found a refcount bug in llc, I'll send a
patch for net-2.6
Meanwhile, here is the patch for net-next-2.6
Your patch then can be applied after mine.
Thanks
[PATCH] net: RCU conversion of dev_getbyhwaddr() and arp_ioctl()
dev_getbyhwaddr() was called under RTNL.
Rename it to dev_getbyhwaddr_rcu() and change all its caller to now use
RCU locking instead of RTNL.
Change arp_ioctl() to use RCU instead of RTNL locking.
Note: this fix a dev refcount bug in llc
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Sun, 5 Dec 2010 02:03:26 +0000 (02:03 +0000)]
llc: fix a device refcount imbalance
Le dimanche 05 décembre 2010 à 12:23 +0100, Eric Dumazet a écrit :
> Le dimanche 05 décembre 2010 à 09:19 +0100, Eric Dumazet a écrit :
>
> > Hmm..
> >
> > If somebody can explain why RTNL is held in arp_ioctl() (and therefore
> > in arp_req_delete()), we might first remove RTNL use in arp_ioctl() so
> > that your patch can be applied.
> >
> > Right now it is not good, because RTNL wont be necessarly held when you
> > are going to call arp_invalidate() ?
>
> While doing this analysis, I found a refcount bug in llc, I'll send a
> patch for net-2.6
Oh well, of course I must first fix the bug in net-2.6, and wait David
pull the fix in net-next-2.6 before sending this rcu conversion.
Note: this patch should be sent to stable teams (2.6.34 and up)
[PATCH net-2.6] llc: fix a device refcount imbalance
commit abf9d537fea225 (llc: add support for SO_BINDTODEVICE) added one
refcount imbalance in llc_ui_bind(), because dev_getbyhwaddr() doesnt
take a reference on device, while dev_get_by_index() does.
Fix this using RCU locking. And since an RCU conversion will be done for
2.6.38 for dev_getbyhwaddr(), put the rcu_read_lock/unlock exactly at
their final place.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: stable@kernel.org Cc: Octavian Purdila <opurdila@ixiacom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Changli Gao [Sat, 4 Dec 2010 14:09:08 +0000 (14:09 +0000)]
ifb: goto resched directly if error happens and dp->tq isn't empty
If we break the loop when there are still skbs in tq and no skb in
rq, the skbs will be left in txq until new skbs are enqueued into rq.
In rare cases, no new skb is queued, then these skbs will stay in rq
forever.
After this patch, if tq isn't empty when we break the loop, we goto
resched directly.
Signed-off-by: Changli Gao <xiaosuo@gmail.com> Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca> Signed-off-by: David S. Miller <davem@davemloft.net>
projects / linux-beck.git / log