- Sep 21, 2011
-
-
-
Jason Wang authored
Commit d1b08284 use new frag API but would leave f to be used uninitialized, this patch fix it. Signed-off-by:
Jason Wang <jasowang@redhat.com> Acked-by:
Michael S. Tsirkin <mst@redhat.com> Acked-by:
Ian Campbell <ian.campbell@citrix.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
-
- Sep 20, 2011
-
-
Alexander Duyck authored
Instead of using the multi_tx_table to map possible Tx queues to Tx rings we can just do simple subtraction for the unlikely event that the Tx queue provided exceeds the number of Tx rings. Signed-off-by:
Alexander Duyck <alexander.h.duyck@intel.com> Tested-by:
Aaron Brown <aaron.f.brown@intel.com> Signed-off-by:
Jeff Kirsher <jeffrey.t.kirsher@intel.com>
-
Alexander Duyck authored
Since igb only uses advanced descriptors we might as well just use an IGB specific define and drop the _ADV suffix for the descriptor declarations. In addition this can be further reduced by assuming that it will be working on pointers since that is normally how the Tx descriptors are handled. Signed-off-by:
-
Alexander Duyck authored
Many of the function names in the hot path are carrying an extra "_adv" suffix on the end of them to represent the fact that they are using advanced descriptors instead of legacy descriptors. However since all igb uses are advanced descriptors adding the extra suffix doesn't really add any additional data. Since this is the case it is easiest to just drop the suffix and save us from having to store the extra characters. Signed-off-by:
-
Alexander Duyck authored
This change is meant to be a general cleanup and performance improvement for clean_rx_irq. The previous patch should have updated the allocation so that the rings can be treated as read-only within the clean_rx_irq function. In addition I am re-ordering the operations such that several goals are accomplished including reducing the overhead for packet accounting, reducing the number of items on the stack, and improving overall performance. Signed-off-by:
-
Alexander Duyck authored
This change is meant to improve performance by splitting the Tx and Rx rings into 3 sections. The first is primarily a read only section containing basic things like the indexes, a pointer to the dev and netdev structures, and basic information. The second section contains the stats and next_to_use and next_to_clean values. The third section is primarily unused values that can just be placed at the end of the ring and are not used in the hot path. The adapter structure has several sections that are read in the hot path. In order to improve performance there I am combining the frequent read hot path items into a single cache line. Signed-off-by:
-
Alexander Duyck authored
This change is meant to streamline the Rx buffer allocation and cleanup. This is accomplished by reducing the number of writes by only having the Rx descriptor ring written by software during allocation, and it will only be read during cleanup. Signed-off-by:
-
Alexander Duyck authored
This change removes support for single buffer mode from igb and makes the driver function in packet split always. The advantage to doing this is that we can reduce total memory allocation overhead significantly as we will only need to allocate one 1K slab per packet and then make use of a reusable half page instead of allocating a 2K slab per packet. Signed-off-by:
-
Alexander Duyck authored
This patch modifies the max_frame_size in order account for an optional VLAN tag. In order to support this we must also increase the MAX_STD_JUMBO_FRAME_SIZE to account for the 4 extra bytes. Signed-off-by:
-
Alexander Duyck authored
This change cleans up the RXDCTL and TXDCTL configurations and optimizes RX performance by allowing back write-backs on all hardware other than 82576. Signed-off-by:
-
-
- Sep 18, 2011
-
-
Ying Xue authored
Elimintes prototype link event tracking functionality that has never been fleshed out and doesn't do anything useful at the current time. Signed-off-by:
Ying Xue <ying.xue@windriver.com> Signed-off-by:
Allan Stephens <allan.stephens@windriver.com> Signed-off-by:
Paul Gortmaker <paul.gortmaker@windriver.com>
-
Ying Xue authored
Eliminate the "event_cb" member from TIPC's "subscription" structure since the function pointer it holds always points to subscr_send_event(). Signed-off-by:
-
Ying Xue authored
Modifies the proto_ops structure used by TIPC DGRAM and RDM sockets so that calls to listen() and accept() are handled by existing kernel "unsupported operation" routines, and eliminates the related checks in the listen and accept routines used by SEQPACKET and STREAM sockets that are no longer needed. Signed-off-by:
-
Ying Xue authored
Adds support for the SO_SNDTIMEO socket option. (This complements the existing support for SO_RCVTIMEO that is already present.) Signed-off-by:
-
Allan Stephens authored
Modifies the initial transfer of name table entries to a new neighboring node so that the messages are enqueued as a unit, rather than individually. The revised algorithm now locates the link carrying the message only once, and eliminates unnecessary checks for link congestion, message fragmentation, and message bundling that are not required when sending these messages. Signed-off-by:
-
Paul Gortmaker authored
Functions like this are called using unsigned longs from function pointers. In this case, the function is passed in a node which is normally internally treated as a u32 by TIPC. Rather than add more casts into this function in the future for each added use of node within, move the cast to a single place on a local. Signed-off-by: -
Allan Stephens authored
Reduces the maximum size of messages sent during the initial exchange of name table information between two nodes to be no larger than the MTU of the first link established between the nodes. This ensures that messages will never need to be fragmented, which would add unnecessary overhead to the name table synchronization mechanism. Signed-off-by:
-
Allan Stephens authored
Reduces the number of bearers a node can support to 2, which can use identical or non-identical media. This change won't impact users, since they are currently limited to a maximum of 2 Ethernet bearers, and will save memory by eliminating a number of unused entries in TIPC's media and bearer arrays. Signed-off-by:
-
Allan Stephens authored
Removes obsolete code that searches for an Ethernet bearer structure entry to use for a newly enabled bearer, since this search is now performed at the start of the enabling algorithm. Signed-off-by:
-
Allan Stephens authored
Ensures that the device list lock is held while trying to locate the Ethernet device used by a newly enabled bearer, so that the addition or removal of a device does not cause problems. Signed-off-by:
-
Allan Stephens authored
Enhances TIPC to ensure that a node that loses contact with a neighboring node does not allow contact to be re-established until it sees that its peer has also recognized the loss of contact. Previously, nodes that were connected by two or more links could encounter a situation in which node A would lose contact with node B on all of its links, purge its name table of names published by B, and then fail to repopulate those names once contact with B was restored. This would happen because B was able to re-establish one or more links so quickly that it never reached a point where it had no links to A -- meaning that B never saw a loss of contact with A, and consequently didn't re-publish its names to A. This problem is now prevented by enhancing the cleanup done by TIPC following a loss of contact with a neighboring node to ensure that node A ignores all messages sent by B until it receives a LINK_PROTOCOL message that indicates B has lost contact with A, thereby preventing the (re)establishment of links between the nodes. The loss of contact is recognized when a RESET or ACTIVATE message is received that has a "redundant link exists" field of 0, indicating that B's sending link endpoint is in a reset state and that B has no other working links. Additionally, TIPC now suppresses the sending of (most) link protocol messages to a neighboring node while it is cleaning up after an earlier loss of contact with that node. This stops the peer node from prematurely activating its link endpoint, which would prevent TIPC from later activating its own end. TIPC still allows outgoing RESET messages to occur during cleanup, to avoid problems if its own node recognizes the loss of contact first and tries to notify the peer of the situation. Finally, TIPC now recognizes an impending loss of contact with a peer node as soon as it receives a RESET message on a working link that is the peer's only link to the node, and ensures that the link protocol suppression mentioned above goes into effect right away -- that is, even before its own link endpoints have failed. This is necessary to ensure correct operation when there are redundant links between the nodes, since otherwise TIPC would send an ACTIVATE message upon receiving a RESET on its first link and only begin suppressing when a RESET on its second link was received, instead of initiating suppression with the first RESET message as it needs to. Note: The reworked cleanup code also eliminates a check that prevented a link endpoint's discovery object from responding to incoming messages while stale name table entries are being purged. This check is now unnecessary and would have slowed down re-establishment of communication between the nodes in some situations. Signed-off-by:
-
- Sep 17, 2011
-
-
Eric Dumazet authored
tcp_md5sig_pool is currently an 'array' (a percpu object) of pointers to struct tcp_md5sig_pool. Only the pointers are NUMA aware, but objects themselves are all allocated on a single node. Remove this extra indirection to get proper percpu memory (NUMA aware) and make code simpler. Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
Rasesh Mody authored
Change details: - In a continuous sequence of ifconfig up/down operations, there is a small window of race between bnad_set_rx_mode() and bnad_cleanup_rx() while the former tries to access rx_info->rx & the latter sets it to NULL. This race could lead to bna_rx_mode_set() being called with a NULL (rx_info->rx) pointer and a crash. - Hold bnad->bna_lock while setting / unsetting rx_info->rx in bnad_setup_rx() & bnad_cleanup_rx(), thereby eliminating the race described above. Signed-off-by:
Gurunatha Karaje <gkaraje@brocade.com> Signed-off-by:
Rasesh Mody <rmody@brocade.com> Signed-off-by:
-
Rasesh Mody authored
When Rx queue size is changed, queues are torn down and setup with the new queue size. During this operation, clear promiscuous mode and restore the original VLAN filter. Signed-off-by:
-
Rasesh Mody authored
Remove a BUG_ON() as it is not required. Change the unconditional write to release a semaphore to read sem first and then write. This will eliminate the possibility of sem getting locked while trying to release it in case if previous sem_get operation failed. Signed-off-by:
-
Emil Tantilov authored
netif_tx_start_all_queues() is already called in ixgbe_up_complete, no need to do it twice. Signed-off-by:
Emil Tantilov <emil.s.tantilov@intel.com> Tested-by:
Phil Schmitt <phillip.j.schmitt@intel.com> Signed-off-by:
-
Emil Tantilov authored
Signed-off-by:
-
Emil Tantilov authored
Remove duplicate inc of hwstats->ruc Introduce separate loops for 8 and 16 register reads. Consolidate mac checks under one case. Make sure registers are cleared on read. Reported-by:
Jonathan Lynch <jonathan.lynch@thenowfactory.com> Signed-off-by:
-
Alexander Duyck authored
This patch improves the memory utilization with RSC when in one-buffer mode. This is accomplished by making the default buffer sizes match up with the standard memory allocation sizes minus 1K for shared info and padding overhead. By doing this CPU utilization when doing large receives can be reduced by as much as 8%. Signed-off-by:
-
Alexander Duyck authored
The adapter structure was removed from the call so it can be dropped from the ixgbe_fso documentation. Signed-off-by:
Ross Brattain <ross.b.brattain@intel.com> Signed-off-by:
-
Alexander Duyck authored
One of the 82598 phys was not being correctly identified as being SFP. This change corrects that. Signed-off-by:
-
Alexander Duyck authored
This change adds a small bit of missing code for enabling the overheat sensor Signed-off-by:
-
Alexander Duyck authored
ixgbe_up and ixgbe_up_complete will always return 0. Since this doesn't provide any useful information we might as well just make them both void and save ourselves from having to return an unused value. Signed-off-by:
-
Alexander Duyck authored
This change fixes an issue in which the incorrect amount of headroom was being reserved for flow director filters. Signed-off-by:
-
Alexander Duyck authored
This change fixes a minor redundancy in that tx_sample_rate was set twice. Signed-off-by:
-
Michał Mirosław authored
Private rx_csum flags are now duplicate of netdev->features & NETIF_F_RXCSUM. Removing this needs deeper surgery. Things noticed: - ixgb has RX csum disabled by default - HW VLAN acceleration probably can be toggled, but it's left as is - the resets on RX csum offload change can probably be avoided - there is A LOT of copy-and-pasted code here Signed-off-by:
Michał Mirosław <mirq-linux@rere.qmqm.pl> Signed-off-by:
-
David S. Miller authored
This reverts commit 0856a304 . As requested by Eric Dumazet, it has various ref-counting problems and has introduced regressions. Eric will add a more suitable version of this performance fix. Signed-off-by:
-
