- 11 May, 2012 12 commits
-
-
Eric Dumazet authored
An implementation of CoDel AQM, from Kathleen Nichols and Van Jacobson. http://queue.acm.org/detail.cfm?id=2209336 This AQM main input is no longer queue size in bytes or packets, but the delay packets stay in (FIFO) queue. As we don't have infinite memory, we still can drop packets in enqueue() in case of massive load, but mean of CoDel is to drop packets in dequeue(), using a control law based on two simple parameters : target : target sojourn time (default 5ms) interval : width of moving time window (default 100ms) Based on initial work from Dave Taht. Refactored to help future codel inclusion as a plugin for other linux qdisc (FQ_CODEL, ...), like RED. include/net/codel.h contains codel algorithm as close as possible than Kathleen reference. net/sched/sch_codel.c contains the linux qdisc specific glue. Separate structures permit a memory efficient implementation of fq_codel (to be sent as a separate work) : Each flow has its own struct codel_vars. timestamps are taken at enqueue() time with 1024 ns precision, allowing a range of 2199 seconds in queue, and 100Gb links support. iproute2 uses usec as base unit. Selected packets are dropped, unless ECN is enabled and packets can get ECN mark instead. Tested from 2Mb to 10Gb speeds with no particular problems, on ixgbe and tg3 drivers (BQL enabled). Usage: tc qdisc ... codel [ limit PACKETS ] [ target TIME ] [ interval TIME ] [ ecn ] qdisc codel 10: parent 1:1 limit 2000p target 3.0ms interval 60.0ms ecn Sent 13347099587 bytes 8815805 pkt (dropped 0, overlimits 0 requeues 0) rate 202365Kbit 16708pps backlog 113550b 75p requeues 0 count 116 lastcount 98 ldelay 4.3ms dropping drop_next 816us maxpacket 1514 ecn_mark 84399 drop_overlimit 0 CoDel must be seen as a base module, and should be used keeping in mind there is still a FIFO queue. So a typical setup will probably need a hierarchy of several qdiscs and packet classifiers to be able to meet whatever constraints a user might have. One possible example would be to use fq_codel, which combines Fair Queueing and CoDel, in replacement of sfq / sfq_red. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
Dave Taht <dave.taht@bufferbloat.net> Cc: Kathleen Nichols <nichols@pollere.com> Cc: Van Jacobson <van@pollere.net> Cc: Tom Herbert <therbert@google.com> Cc: Matt Mathis <mattmathis@google.com> Cc: Yuchung Cheng <ycheng@google.com> Cc: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Eric Dumazet authored
Class bytes/packets stats can be misleading because they are updated in enqueue() while packet might be dropped later. We already fixed all qdiscs but sch_atm. This patch makes the final cleanup. class rate estimators can now match qdisc ones. Signed-off-by:
-
Joe Perches authored
Use the new bool function ether_addr_equal_64bits to add some clarity and reduce the likelihood for misuse of compare_ether_addr_64bits for sorting. Done via cocci script: $ cat compare_ether_addr_64bits.cocci @@ expression a,b; @@ - !compare_ether_addr_64bits(a, b) + ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - compare_ether_addr_64bits(a, b) + !ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - !ether_addr_equal_64bits(a, b) == 0 + ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - !ether_addr_equal_64bits(a, b) != 0 + !ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - ether_addr_equal_64bits(a, b) == 0 + !ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - ether_addr_equal_64bits(a, b) != 0 + ether_addr_equal_64bits(a, b) @@ expression a,b; @@ - !!ether_addr_equal_64bits(a, b) + ether_addr_equal_64bits(a, b) Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Joe Perches authored
Add an optimized boolean function to check if 2 ethernet addresses are the same. This is to avoid any confusion about compare_ether_addr_64bits returning an unsigned, and not being able to use the compare_ether_addr_64bits function for sorting ala memcmp. Signed-off-by:
-
Joe Perches authored
Use the new bool function ether_addr_equal to add some clarity and reduce the likelihood for misuse of compare_ether_addr for sorting. Done via cocci script: $ cat compare_ether_addr.cocci @@ expression a,b; @@ - !compare_ether_addr(a, b) + ether_addr_equal(a, b) @@ expression a,b; @@ - compare_ether_addr(a, b) + !ether_addr_equal(a, b) @@ expression a,b; @@ - !ether_addr_equal(a, b) == 0 + ether_addr_equal(a, b) @@ expression a,b; @@ - !ether_addr_equal(a, b) != 0 + !ether_addr_equal(a, b) @@ expression a,b; @@ - ether_addr_equal(a, b) == 0 + !ether_addr_equal(a, b) @@ expression a,b; @@ - ether_addr_equal(a, b) != 0 + ether_addr_equal(a, b) @@ expression a,b; @@ - !!ether_addr_equal(a, b) + ether_addr_equal(a, b) Signed-off-by:
-
Sathya Perla authored
Calling pci_disable_sriov() while VFs are assigned to VMs causes kernel panic. This patch uses PCI_DEV_FLAGS_ASSIGNED bit state of the VF's pci_dev to avoid this. Also, the unconditional function reset cmd issued on a PF probe can delete the VF configuration for the previously enabled VFs. A scratchpad register is now used to issue a function reset only when needed (i.e., in a crash dump scenario.) Signed-off-by:
Sathya Perla <sathya.perla@emulex.com> Signed-off-by:
-
James Chapman authored
If enabled, L2TP data packets have sequence numbers which a receiver can use to drop out of sequence frames or try to reorder them. The first frame has sequence number 0, but the L2TP code currently expects it to be 1. This results in the first data frame being handled as out of sequence. This one-line patch fixes the problem. Signed-off-by:
James Chapman <jchapman@katalix.com> Signed-off-by:
-
James Chapman authored
When L2TP data packet reordering is enabled, packets are held in a queue while waiting for out-of-sequence packets. If a packet gets lost, packets will be held until the reorder timeout expires, when we are supposed to then advance to the sequence number of the next packet but we don't currently do so. As a result, the data channel is stuck because we are waiting for a packet that will never arrive - all packets age out and none are passed. The fix is to add a flag to the session context, which is set when the reorder timeout expires and tells the receive code to reset the next expected sequence number to that of the next packet in the queue. Tested in a production L2TP network with Starent and Nortel L2TP gear. Signed-off-by:
-
Pavel Emelyanov authored
As proposed by Eric, make the tcp_input.o thinner. add/remove: 1/1 grow/shrink: 1/4 up/down: 868/-1329 (-461) function old new delta tcp_try_rmem_schedule - 864 +864 tcp_ack 4811 4815 +4 tcp_validate_incoming 817 815 -2 tcp_collapse 860 858 -2 tcp_send_rcvq 555 353 -202 tcp_data_queue 3435 3033 -402 tcp_prune_queue 721 - -721 Signed-off-by:
Pavel Emelyanov <xemul@parallels.com> Acked-by:
-
Pavel Emelyanov authored
As noted by Eric, no checks are performed on the data size we're putting in the read queue during repair. Thus, validate the given data size with the common rmem management routine. Signed-off-by:
-
Pavel Emelyanov authored
It actually works on the input queue and will use its read mem routines, thus it's better to have in in the tcp_input.c file. Signed-off-by:
-
-
- 10 May, 2012 28 commits
-
-
Don Skidmore authored
Update version number to better match the version of the out of tree driver with similar functionality. Signed-off-by:
Don Skidmore <donald.c.skidmore@intel.com> Tested-by:
Phil Schmitt <phillip.j.schmitt@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Don Skidmore authored
When the hwmon code was initially added it was with the assumption that a sysfs patch would be also coming soon. Since that isn't the case some clean up needs to be done. This patch does that. Signed-off-by:
Stephen Ko <stephen.s.ko@intel.com> Signed-off-by:
-
Jacob Keller authored
Kernel software timestamping requires that the driver calls skb_tx_timestamp just before passing the skb to the MAC, in order to provide the best software timestamps. This patch adds this call for that support. Signed-off-by:
Jacob Keller <jacob.e.keller@intel.com> Tested-by:
-
Jacob Keller authored
This patch adds support for the ethtool get_ts_info operation, which enables access of available timestamp/timesync support for that device. It can query which ptp clock device is associated with the particular port. Signed-off-by:
-
Jacob Keller authored
The current value of the udelay timeout for ixgbe_disable_rx_buff is too short. This causes the security path to not not be properly disabled during the section that is meant to have it turned off. The end result causes a race condition that results in RX issues. Signed-off-by:
-
Jacob E Keller authored
This patch enables the PPS system in the PHC framework, by enabling the clock-out feature on the X540 device. Causes the SDP0 to be set as a 1Hz clock. Also configures the timesync interrupt cause in order to report each pulse to the PPS via the PHC framework, which can be used for general system clock synchronization. (This allows a stable method for tuning the general system time via the on-board SYSTIM register based clock.) Signed-off-by:
-
Jacob Keller authored
This patch enables hardware timestamping for use with PTP software by extracting a ns counter from an arbitrary fixed point cycles counter. The hardware generates SYSTIME registers using the DMA tick which changes based on the current link speed. These SYSTIME registers are converted to ns using the cyclecounter and timecounter structures provided by the kernel. Using the SO_TIMESTAMPING api, software can enable and access timestamps for PTP packets. The SO_TIMESTAMPING API has space for 3 different kinds of timestamps, SYS, RAW, and SOF. SYS hardware timestamps are hardware ns values that are then scaled to the software clock. RAW hardware timestamps are the direct raw value of the ns counter. SOF software timestamps are the software timestamp calculated as close as possible to the software transmit, but are not offloaded to the hardware. This patch only supports the RAW hardware timestamps due to inefficiency of the SYS design. This patch also enables the PHC subsystem features for atomically adjusting the cycle register, and adjusting the clock frequency in parts per billion. This frequency adjustment works by slightly adjusting the value added to the cycle registers each DMA tick. This causes the hardware registers to overflow rapidly (approximately once every 34 seconds, when at 10gig link). To solve this, the timecounter structure is used, along with a timer set for every 25 seconds. This allows for detecting register overflow and converting the cycle counter registers into ns values needed for providing useful timestamps to the network stack. Only the basic required clock functions are supported at this time, although the hardware supports some ancillary features and these could easily be enabled in the future. Note that use of this hardware timestamping requires modifying daemon software to use the SO_TIMESTAMPING API for timestamps, and the ptp_clock PHC framework for accessing the clock. The timestamps have no relation to the system time at all, so software must use the posix clock generated by the PHC framework instead. Signed-off-by:
-
Greg Rose authored
If the VF sends a MACVLAN request with index of zero then it is not actually trying to add a filter. Check the index value and only indicate that operation is not allowed when the VF is actually trying to add a filter. Signed-off-by:
Greg Rose <gregory.v.rose@intel.com> Tested-by:
Sibai Li <sibai.li@intel.com> Signed-off-by:
-
Alexander Duyck authored
The drop enable bit can be used to improve the performance of the adapter in the case of multiple queues being present. This performance gain is due to the fact that some slower CPUs can cause the FIFO to backfill preventing faster CPUs from receiving additional work. By setting the drop enable bit we prevent this and instead just drop the packets that would have been bound for the slower CPU. Signed-off-by:
Alexander Duyck <alexander.h.duyck@intel.com> Tested-by:
Ross Brattain <ross.b.brattain@intel.com> Signed-off-by:
-
Alexander Duyck authored
This change cleans up the logic in the priority based flow control configuration routines. Both the 82599 and 82598 based routines perform similar functions however they are both arranged completely differently. This patch goes over both of them to clean up the code. In addition I am dropping the ixgbe_fc_pfc flow control mode and instead just replacing it with checks for if priority flow control is enabled. This allows us to maintain some of the link flow control information which allows for an easier transition between link and priority flow control. Signed-off-by:
-
Alexander Duyck authored
Previously we would get a mailbox error and still process the message. Instead we should exit on error. In addition we should also be flushing the ACK of the message so that we can guarantee that the other end is aware we have received the message while we are processing it. Signed-off-by:
-
Koki Sanagi authored
Current igb outputs registers related to TX/RX queues(ex. RDT, RDH, TDT, TDH). But it thinks the number of RX/TX queues is 4. But 82576 has 16 RX/TX queues. This patch modifies igb to output the rest of the registers if the device is 82576. Signed-off-by:
Koki Sanagi <sanagi.koki@jp.fujitsu.com> Acked-by:
Carolyn Wyborny <carolyn.wyborny@intel.com> Tested-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
-
-
Rajesh Borundia authored
o Linux stack estimates MSS from skb->len or skb_shinfo(skb)->gso_size. In case of LRO skb->len is aggregate of len of number of packets hence MSS obtained using skb->len would be incorrect. Incorrect estimation of recv MSS would lead to delayed acks in some traffic patterns (which sends two or three packets and wait for ack and only then send remaining packets). This leads to drop in performance. Hence we need to set gso_size to MSS obtained from firmware. o This is fixed recently in firmware hence the MSS is obtained based on capability. If fw is capable of sending the MSS then only driver sets the gso_size. Signed-off-by:
Rajesh Borundia <rajesh.borundia@qlogic.com> Signed-off-by:
-
Sucheta Chakraborty authored
Driver queries DIMM information from firmware and accordingly sets "presence" field of the structure. "presence" field when set to 0xff denotes invalid flag. And when set to 0x0 denotes DIMM memory is not present. Signed-off-by:
Sucheta Chakraborty <sucheta.chakraborty@qlogic.com> Signed-off-by:
-
Manish chopra authored
o 0x3, 0x7, 0xF, 0x1F, 0x3F, 0x7F and 0xFF are the allowed capture masks. Signed-off-by:
Manish chopra <manish.chopra@qlogic.com> Signed-off-by:
-
Sritej Velaga authored
disable fw dump by default at start up. Signed-off-by:
Sritej Velaga <sritej.velaga@qlogic.com> Signed-off-by:
-
-
Ben Hutchings authored
Signed-off-by:Ben Hutchings <bhutchings@solarflare.com>
-
Stuart Hodgson authored
Currently allows for SFP+ eeprom to be returned using the ethtool API. This can be extended in future to handle different eeprom formats and sizes Signed-off-by:
Stuart Hodgson <smhodgson@solarflare.com> [bwh: Drop redundant validation, comment, whitespace] Signed-off-by:
-
Stuart Hodgson authored
ETHTOOL_GMODULEINFO returns a new struct ethtool_modinfo that will return the type and size of plug-in module eeprom (such as SFP+) for parsing by userland program. ETHTOOL_GMODULEEEPROM returns the raw eeprom information using the existing ethtool_eeprom structture to return the data Signed-off-by:
-
Ben Hutchings authored
We want to support reading module (SFP+, XFP, ...) EEPROMs as well as NIC EEPROMs. They will need a different command number and driver operation, but the structure and arguments will be the same and so we can share most of the code here. Signed-off-by: -
David Riddoch authored
Previously we refilled with much larger batches, which caused large latency spikes. We now have many more much much smaller spikes! Signed-off-by: -
David Riddoch authored
There was no runtime control of the fast_fill_limit in any case, so purged that field. Signed-off-by: -
Ben Hutchings authored
We need to clear the private data pointer in the PCI device. Also reorder cleanup in efx_pci_remove() for symmetry. Signed-off-by: -
Stuart Hodgson authored
efx_nic_fatal_interrupt() disables DMA before scheduling a reset. After this, we need not and *cannot* flush queues. Signed-off-by: -
Joe Perches authored
Use the new bool function ether_addr_equal to add some clarity and reduce the likelihood for misuse of compare_ether_addr for sorting. Done via cocci script: $ cat compare_ether_addr.cocci @@ expression a,b; @@ - !compare_ether_addr(a, b) + ether_addr_equal(a, b) @@ expression a,b; @@ - compare_ether_addr(a, b) + !ether_addr_equal(a, b) @@ expression a,b; @@ - !ether_addr_equal(a, b) == 0 + ether_addr_equal(a, b) @@ expression a,b; @@ - !ether_addr_equal(a, b) != 0 + !ether_addr_equal(a, b) @@ expression a,b; @@ - ether_addr_equal(a, b) == 0 + !ether_addr_equal(a, b) @@ expression a,b; @@ - ether_addr_equal(a, b) != 0 + ether_addr_equal(a, b) @@ expression a,b; @@ - !!ether_addr_equal(a, b) + ether_addr_equal(a, b) Signed-off-by:
-
Joe Perches authored
spatch/coccinelle isn't perfect. It doesn't understand __aligned(x) and doesn't convert functions it can't parse. Convert the remaining compare_ether_addr uses. Signed-off-by:
-
