- 15 Nov, 2010 6 commits
-
-
Changli Gao authored
Signed-off-by:
Changli Gao <xiaosuo@gmail.com> Signed-off-by:
Patrick McHardy <kaber@trash.net>
-
Changli Gao authored
Less IDs make nf_ct_ext smaller. Signed-off-by:
-
Changli Gao authored
As we own the conntrack and the others can't see it until we confirm it, we don't need to use atomic bit operation on ct->status. Signed-off-by:
-
Changli Gao authored
Signed-off-by:
-
Changli Gao authored
In function update_alloc_size(), sizeof(struct nf_ct_ext) is added twice wrongly. Signed-off-by:
-
Jan Engelhardt authored
I am observing consistent behavior even with bridges, so let's unlock this. xt_mac is already usable in FORWARD, too. Section 9 of http://ebtables.sourceforge.net/br_fw_ia/br_fw_ia.html#section9 says the MAC source address is changed, but my observation does not match that claim -- the MAC header is retained. Signed-off-by:
Jan Engelhardt <jengelh@medozas.de> [Patrick; code inspection seems to confirm this] Signed-off-by:
-
- 12 Nov, 2010 3 commits
-
-
Changli Gao authored
Signed-off-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
-
Changli Gao authored
ct->proto is big(60 bytes) due to structure ip_ct_tcp, and we don't need to initialize the whole for all the other protocols. This patch moves proto to the end of structure nf_conn, and pushes the initialization down to the individual protocols. Signed-off-by:
-
David S. Miller authored
When we test rt->fl.iif against zero, we're seeing if it's an output or an input route. Make that explicit with some helper functions. Signed-off-by:David S. Miller <davem@davemloft.net>
-
- 11 Nov, 2010 15 commits
-
-
-
Eric Dumazet authored
It seems idev field in struct rtable has no special purpose, but adding extra atomic ops. We hold refcounts on the device itself (using percpu data, so pretty cheap in current kernel). infiniband case is solved using dst.dev instead of idev->dev Removal of this field means routing without route cache is now using shared data, percpu data, and only potential contention is a pair of atomic ops on struct neighbour per forwarded packet. About 5% speedup on routing test. Signed-off-by:
-
Eric Dumazet authored
It is important to move nud_state outside of the often modified cache line (because of refcnt), to reduce false sharing in neigh_event_send() This is a followup of commit 0ed8ddf4 (neigh: Protect neigh->ha[] with a seqlock) This gives a 7% speedup on routing test with IP route cache disabled. Signed-off-by:
-
Jon Mason authored
Update vxge driver version Signed-off-by:
Jon Mason <jon.mason@exar.com> Signed-off-by:
Ram Vepa <ram.vepa@exar.com> Signed-off-by:
-
Jon Mason authored
Correct issues found by running sparse on the vxge driver, as well as other miscellaneous cleanups. Signed-off-by:
-
Jon Mason authored
Update Kconfig to reflect Exar's purchase of Neterion (formerly S2IO). Signed-off-by:
-
Jon Mason authored
The values used to determined if the adapter is running in single or multi-function mode were previously modified to the values necessary when making the VXGE_HW_FW_API_GET_FUNC_MODE firmware call. However, the firmware call was not modified. This had the driver printing out on probe that the adapter was in multi-function mode when in single function mode and vice versa. Signed-off-by:
-
Jon Mason authored
Detect if the adapter is Titan or Titan1A, and tune the driver for this hardware. Also, remove unnecessary function __vxge_hw_device_id_get. Signed-off-by:
-
Jon Mason authored
Propagate the return code of the call to vxge_hw_vpath_fw_api and __vxge_hw_vpath_pci_func_mode_get. This enables the proper handling of error conditions when querying the function mode of the device during probe. Signed-off-by:
-
Jon Mason authored
Add support for enable/disabling hardware timestamping on receive packets via ioctl call. When enabled, the hardware timestamp replaces the FCS in the payload. Signed-off-by:
-
Jon Mason authored
Add the ability in the vxge driver to flash firmware via ethtool. Updated to include comments from Ben Hutchings. Signed-off-by:
-
Jon Mason authored
It is possible for multiple callers to access the firmware interface for the same vpath simultaneously, resulting in uncertain output. Add locks to serialize access. Also, make functions only accessed locally static, thus requiring some movement of code blocks. Signed-off-by:
-
Jon Mason authored
Remove all of the unnecessary debug printk indirection and temporary variables for vxge_debug_ll and vxge_assert. Signed-off-by:
-
Jon Mason authored
Wait for the receive traffic to become idle before attempting to close or reset the adapter. To enable the processing of packets while Receive Idle, move the clearing of __VXGE_STATE_CARD_UP bit in vxge_close to after it. Also, modify the return value of the ISR when the adapter is down to IRQ_HANDLED. Otherwise there are unhandled interrupts for the device. Signed-off-by:
-
Jon Mason authored
Enable RSS hashing and add ability to pass up the adapter calculated rx hash up the network stack (if feature is available). Add the ability to enable/disable feature via ethtool, which requires that the adapter is not running at the time. Other miscellaneous cleanups and fixes required to get RSS working. Signed-off-by:
-
- 10 Nov, 2010 5 commits
-
-
Gerrit Renker authored
This completes the implementation of a circular buffer for Ack Vectors, by extending the current (linear array-based) implementation. The changes are: (a) An `overflow' flag to deal with the case of overflow. As before, dynamic growth of the buffer will not be supported; but code will be added to deal robustly with overflowing Ack Vector buffers. (b) A `tail_seqno' field. When naively implementing the algorithm of Appendix A in RFC 4340, problems arise whenever subsequent Ack Vector records overlap, which can bring the entire run length calculation completely out of synch. (This is documented on http://www.erg.abdn.ac.uk/users/gerrit/dccp/notes/\ ack_vectors/tracking_tail_ackno/ .) (c) The buffer length is now computed dynamically (i.e. current fill level), as the span between head to tail. As a result, dccp_ackvec_pending() is now simpler - the #ifdef is no longer necessary since buf_empty is always true when IP_DCCP_ACKVEC is not configured. Signed-off-by:Gerrit Renker <gerrit@erg.abdn.ac.uk>
-
Gerrit Renker authored
This patch * separates Ack Vector housekeeping code from option-insertion code; * shifts option-specific code from ackvec.c into options.c; * introduces a dedicated routine to take care of the Ack Vector records; * simplifies the dccp_ackvec_insert_avr() routine: the BUG_ON was redundant, since the list is automatically arranged in descending order of ack_seqno. Signed-off-by: -
Gerrit Renker authored
This patch brings the Ack Vector interface up to date. Its main purpose is to lay the basis for the subsequent patches of this set, which will use the new data structure fields and routines. There are no real algorithmic changes, rather an adaptation: (1) Replaced the static Ack Vector size (2) with a #define so that it can be adapted (with low loss / Ack Ratio, a value of 1 works, so 2 seems to be sufficient for the moment) and added a solution so that computing the ECN nonce will continue to work - even with larger Ack Vectors. (2) Replaced the #defines for Ack Vector states with a complete enum. (3) Replaced #defines to compute Ack Vector length and state with general purpose routines (inlines), and updated code to use these. (4) Added a `tail' field (conversion to circular buffer in subsequent patch). (5) Updated the (outdated) documentation for Ack Vector struct. (6) All sequence number containers now trimmed to 48 bits. (7) Removal of unused bits: * removed dccpav_ack_nonce from struct dccp_ackvec, since this is already redundantly stored in the `dccpavr_ack_nonce' (of Ack Vector record); * removed Elapsed Time for Ack Vectors (it was nowhere used); * replaced semantics of dccpavr_sent_len with dccpavr_ack_runlen, since the code needs to be able to remember the old run length; * reduced the de-/allocation routines (redundant / duplicate tests). Signed-off-by: -
Ron Mercer authored
Signed-off-by:
Jitendra Kalsaria <jitendra.kalsaria@qlogic.com> Signed-off-by:
Ron Mercer <ron.mercer@qlogic.com> Signed-off-by:
-
Ron Mercer authored
By default we add firmware information to ethtool get regs. Optionally firmware info can instead be sent to log. Signed-off-by:
-
- 09 Nov, 2010 5 commits
-
-
Joe Perches authored
Coalesce long formats. Align arguments. Remove KERN_<level>. Signed-off-by:
Joe Perches <joe@perches.com> Signed-off-by:
-
Joe Perches authored
Coalesce long formats. Add missing newlines. Signed-off-by:
-
Joe Perches authored
Add missing newlines. Signed-off-by:
-
Joe Perches authored
Add missing newlines. Signed-off-by:
-
Eric Dumazet authored
Some network drivers use old TX_TIMEOUT definitions, assuming HZ=100 of old kernels. Convert these definitions to include HZ, since HZ can be 1000 these days. Signed-off-by:
-
- 08 Nov, 2010 6 commits
-
-
Eric Dumazet authored
unix_dgram_poll() is pretty expensive to check POLLOUT status, because it has to lock the socket to get its peer, take a reference on the peer to check its receive queue status, and queue another poll_wait on peer_wait. This all can be avoided if the process calling unix_dgram_poll() is not interested in POLLOUT status. It makes unix_dgram_recvmsg() faster by not queueing irrelevant pollers in peer_wait. On a test program provided by Alan Crequy : Before: real 0m0.211s user 0m0.000s sys 0m0.208s After: real 0m0.044s user 0m0.000s sys 0m0.040s Suggested-by:
Davide Libenzi <davidel@xmailserver.org> Reported-by:
Alban Crequy <alban.crequy@collabora.co.uk> Acked-by:
-
Eric Dumazet authored
Alban Crequy reported a problem with connected dgram af_unix sockets and provided a test program. epoll() would miss to send an EPOLLOUT event when a thread unqueues a packet from the other peer, making its receive queue not full. This is because unix_dgram_poll() fails to call sock_poll_wait(file, &unix_sk(other)->peer_wait, wait); if the socket is not writeable at the time epoll_ctl(ADD) is called. We must call sock_poll_wait(), regardless of 'writable' status, so that epoll can be notified later of states changes. Misc: avoids testing twice (sk->sk_shutdown & RCV_SHUTDOWN) Reported-by:
-
Eric Dumazet authored
Instead of wakeup all sleepers, use wake_up_interruptible_sync_poll() to wakeup only ones interested into writing the socket. This patch is a specialization of commit 37e5540b (epoll keyed wakeups: make sockets use keyed wakeups). On a test program provided by Alan Crequy : Before: real 0m3.101s user 0m0.000s sys 0m6.104s After: real 0m0.211s user 0m0.000s sys 0m0.208s Reported-by:
-
Eric Dumazet authored
While tracking dev_base_lock users, I found decnet used it in dnet_select_source(), but for a wrong purpose: Writers only hold RTNL, not dev_base_lock, so readers must use RCU if they cannot use RTNL. Adds an rcu_head in struct dn_ifaddr and handle proper RCU management. Adds __rcu annotation in dn_route as well. Signed-off-by:
Steven Whitehouse <swhiteho@redhat.com> Signed-off-by:
-
Eric Dumazet authored
bond_info_seq_start() uses a read_lock(&dev_base_lock) to make sure device doesn’t disappear. Same goal can be achieved using RCU. Signed-off-by:
-
Eric Dumazet authored
dev_base_lock is the legacy way to lock the device list, and is planned to disappear. (writers hold RTNL, readers hold RCU lock) Convert aoecmd_cfg_pkts() to RCU locking. Signed-off-by:
-
