- 11 Nov, 2009 40 commits
-
-
Ben Cahill authored
Follow-up to "update write pointers for all tx queues after wakeup"; that patch changed iwl_irq_tasklet_legacy(), but not iwl_irq_tasklet(), so newer devices were not covered. Comments from original patch: Wakeup interrupt has been updating write pointers (indexes, actually) only for tx queues 0-5. This is adequate just for 3945, but inadequate for other devices, all of which have more tx queues. Now updating all tx/command queues, so device can be aware of all new tx and host commands enqueued while device was asleep. This can potentially improve data traffic bandwidth and/or latency. Signed-off-by:
Ben Cahill <ben.m.cahill@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Ben Cahill authored
Signed-off-by:
-
Ben Cahill authored
Signed-off-by:
-
Ben Cahill authored
Take advantage of device's auto-increment for SRAM reads to eliminate extra write address accesses. Grab/release NIC access before/after entire read sequence, rather than for each read individually. After a quick check of dmesg logs, this seems to double Event Log dump speed, reducing from about 20 milliseconds to about 10 milliseconds for 512 entries using 3945. Signed-off-by:
-
Ben Cahill authored
If device provides bad values for Event Log parameters (due to being asleep or SRAM corruption, etc.), the size can be very, very large (e.g. 0xa5a5a5a5), which can flood system log. Sanity-check capacity and next_entry values and limit to reasonable size dump. Signed-off-by:
-
Wey-Yi Guy authored
Do not use AP's SM_PS setting for our own SM_PS setting. Reported-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
Wey-Yi Guy <wey-yi.w.guy@intel.com> Signed-off-by:
-
Wey-Yi Guy authored
Remove the support for all the PCI_IDs never make into production Signed-off-by:
-
Wey-Yi Guy authored
Different channel has different configuration, need to pass correct configuration to uCode when send "channel switch" command to uCode. Invalid configuration will cause sysassert in uCode and produce un-expected result. Even it is a very small windows, but we also need to consider and handle the case if commit_rxon occurred before the "channel switch announcement" notification received from uCode. Signed-off-by:
-
Johannes Berg authored
This is part of the code, but the comment doesn't have it, add pointers to the code and the FIFO usage for 5000 and up. Signed-off-by:
-
Wey-Yi Guy authored
Remove un-used parameter "recovery_rxon" from "priv" data structure Signed-off-by:
-
Johannes Berg authored
When I added that code setting the swq_id, I evidently did not understand the distinction between FIFO and TX queue yet and added code to compare a queue ID and a FIFO number, which is bogus. However, the code there need not be this specific, it can just set all queues to the identity mapping which will be overwritten by the aggregation queue code. As a bit of defensive coding, don't assign an swq_id to the command queue so that if we ever use it for frames we notice quickly. Signed-off-by:
-
Johannes Berg authored
The fourth queue (command queue) is already activated in the loop above that also maps it to the command FIFO and therefore doesn't need to be marked as activated again. Also change the TODO comment to be accurate -- we need to initialise the _queues_, not FIFOs, and map them to device FIFOs. Signed-off-by:
-
Wey-Yi Guy authored
For both 6x00 and 6x50 series devices, periodic Tx IQ calibration is disabled in uCode, driver do not need to set the periodic Tx/IQ calibration bit in calibration command. Signed-off-by:
-
Shanyu Zhao authored
Use configured valid rx chains in scan command instead of ANT_ABC, correcting valid rx chain configuration of 4965, should be ANT_ABC instead of ANT_AB. Signed-off-by:
Shanyu Zhao <shanyu.zhao@intel.com> Signed-off-by:
-
Shanyu Zhao authored
When doing scan in power saving mode, choose only 1 valid RX chain instead of turning all chains on. Signed-off-by:
-
Wey-Yi Guy authored
To help debug rx related issues, if IWL_DEBUG_RADIO flag is set, print the rxon configuration when rxon host command send to uCode. Signed-off-by:
-
Zhu Yi authored
Allocate 128 bytes linear buffer for rx skb. The first 64 bytes is reserved for mac80211 usage (for radiotap header expansion, etc). The frame header starts from the second 64 bytes. Cc: Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
Zhu Yi <yi.zhu@intel.com> Signed-off-by:
-
Reinette Chatre authored
Since these messages indicate failure we would be interested in seeing them always. Signed-off-by:
-
Michael Buesch authored
Do not mess with the original skb, but allocate an independent bouncebuffer. This protects against bad interference with mac80211's assumptions about the skb (which already caused bugs). Signed-off-by:
Michael Buesch <mb@bu3sch.de> Signed-off-by:
-
David S. Miller authored
Conflicts: drivers/net/wireless/iwlwifi/iwl-1000.c drivers/net/wireless/iwlwifi/iwl-6000.c drivers/net/wireless/iwlwifi/iwl-core.h
-
stephen hemminger authored
Use new function to avoid doing read_lock(). Signed-off-by:
Stephen Hemminger <shemminger@vyatta.com> Acked-by:
Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by:
Oliver Hartkopp <oliver@hartkopp.net> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
stephen hemminger authored
This also needs to be optimized for large number of devices. Signed-off-by:
-
stephen hemminger authored
When showing device statistics use RCU rather than read_lock(&dev_base_lock) Compile tested only. Signed-off-by:
-
stephen hemminger authored
This is similar to other cases where for_each_netdev_rcu can be used when gathering information. By inspection, don't have platform or cross-build environment to validate. Signed-off-by:
-
stephen hemminger authored
Use RCU to walk list of network devices in qdisc dump. This could be optimized for large number of devices. Signed-off-by:
-
stephen hemminger authored
Do not need to use read_lock(&dev_base_lock), use RCU instead. Signed-off-by:
-
stephen hemminger authored
This adds an RCU macro for continuing search, useful for some network devices like vlan. Signed-off-by:
-
Brian Haley authored
Change udp6_portaddr_hash() to use ipv6_addr_v4mapped() inline instead of ipv6_addr_type(). Signed-off-by:
Brian Haley <brian.haley@hp.com> Acked-by:
-
Eric Dumazet authored
Use for_each_netdev_rcu() and dont lock dev_base_lock anymore Signed-off-by:
-
Herbert Xu authored
This patch rearranges the SIT DF bit handling using the new IPIP DF code. The only externally visible effect should be the case where PMTU is enabled and the MTU is exactly 1280 bytes. In this case the previous code would send packets out with DF off while the new code would set the DF bit. This is inline with RFC 4213. Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au> Thanks, Signed-off-by:
-
Eric Dumazet authored
Apparently, inet6_dump_addr() is not able to handle more than 64 ipv6 addresses per device. We must break from inner loops in case skb is full, or else cursor is put at the end of list. Signed-off-by:
-
Eric Dumazet authored
When handling large number of netdevice, inet6_dump_ifinfo() is very slow because it has O(N^2) complexity. Instead of scanning one single list, we can use the 256 sub lists of the dev_index hash table, and RCU lookups. Signed-off-by:
-
Cyrill Gorcunov authored
Use guard DECLARE_SOCKADDR in a few more places which allow us to catch if the structure copied back is too big. Signed-off-by:
Cyrill Gorcunov <gorcunov@openvz.org> Signed-off-by:
-
Ben Hutchings authored
Some usbnet drivers update link state while others do not due to hardware limitations. Add a flag to distinguish those that do, and set the link down initially for their devices. This is intended to fix this bug: http://bugs.debian.org/444043Signed-off-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
-
Marin Mitov authored
Use DMA_BIT_MASK(44) instead of deprecated DMA_44BIT_MASK Signed-off-by:
Marin Mitov <mitov@issp.bas.bg> Signed-off-by:
-
Eric Dumazet authored
UDP bind() can be O(N^2) in some pathological cases. Thanks to secondary hash tables, we can make it O(N) Signed-off-by:
-
Eilon Greenstein authored
Signed-off-by:
Eilon Greenstein <eilong@broadcom.com> Signed-off-by:
-
Eilon Greenstein authored
Signed-off-by:
-
Eilon Greenstein authored
Failing to allocate MSI-X vectors is not an error and should not be printed as such Signed-off-by:
-
Eilon Greenstein authored
Making sure that whenever the FW/HW is configured for GSO, it is also configured to CSUM offload Signed-off-by:
-
