- 03 Feb, 2011 7 commits
-
-
David Gnedt authored
The queue stopping/waking functionality was broken in a way that could cause the TX to stall if the right circumstances are met. The problem was caused by tx_work, which is scheduled on each TX operation. If the firmware buffer is full, tx_work does nothing. In combinition with stopped queues or non-continues transfers, tx_work is never scheduled again. Moreover the low watermark introduced by 9df86e2e never takes effect because of some old code. Solve this by scheduling tx_work every time tx_queue is non-empty and firmware buffer is freed on tx_complete. This also solves a possible but unlikely case: If less frames than the high watermark are queued, but more than firmware buffer can hold. This results in queues staying awake but the only scheduled tx_work doesn't transfer all frames, so the remaining frames are stuck in the queue until more frames get queued and tx_work is scheduled again. Signed-off-by:
David Gnedt <david.gnedt@davizone.at> Acked-by:
Kalle Valo <kvalo@adurom.com> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Ben Greear authored
I have a netgear WNDR3700 that appears to have an off-by-four bug in how it fills out the hti->control_chan (I configure the AP to channel 11, it reports 15 as control_chan). Poke a message into the kernel logs to give users a clue as to why they are not getting the expected channel-type or rate. Signed-off-by:
Ben Greear <greearb@candelatech.com> Signed-off-by:
-
Ben Greear authored
Signed-off-by:
-
Ben Greear authored
If we cannot set the channel type, set the channel back to the original. Don't update the driver hardware if nothing actually changed. Signed-off-by:
-
Vasily Khoruzhick authored
To support suspend/resume in if_spi we need two things: - re-setup fw in lbs_resume(), because if_spi powercycles card; - don't touch hwaddr on second lbs_update_hw_spec() call for same reason; Signed-off-by:
Vasily Khoruzhick <anarsoul@gmail.com> Signed-off-by:
-
Vasily Khoruzhick authored
Use workqueue to perform SPI xfers, it's necessary to fix nasty "BUG: scheduling while atomic", because spu_write() calls spi_sync() and spi_sync() may sleep, but hw_host_to_card() callback can be called from atomic context. Remove kthread completely, workqueue now does its job. Restore intermediate buffers which were removed in commit 86c34fe8 that introduced mentioned bug. Signed-off-by:
-
John W. Linville authored
Merge branch 'wireless-next-2.6' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-2.6
-
- 01 Feb, 2011 2 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6John W. Linville authored
Conflicts: drivers/net/wireless/wl12xx/spi.c
-
Mohammed Shafi Shajakhan authored
free the skb's when the Tx of PAPRD frames fails and also add a debug message indicating that. Signed-off-by:
Mohammed Shafi Shajakhan <mshajakhan@atheros.com> Signed-off-by:
-
- 31 Jan, 2011 31 commits
-
-
Wey-Yi Guy authored
sco_cmd is not being used, remove it Signed-off-by:Wey-Yi Guy <wey-yi.w.guy@intel.com>
-
Wey-Yi Guy authored
During the period of BT coex changes, REPLY_BT_COEX_SCO host command is no longer needed to support SCO/eSCO type of traffic. delete it. Signed-off-by: -
Wey-Yi Guy authored
For 2030 series of devices, 2030 macro need to be used. Signed-off-by: -
Johannes Berg authored
The beacon notification changed between 4965 and agn because the embedded TX response changed, but iwlwifi was never updated to know about this. Update it now so the IBSS manager status will be tracked correctly. Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Johannes Berg authored
Use the values from the peer to set up the ucode for the right maximum number of subframes in an aggregate. Since the ucode only tracks this per station, use the minimum across all aggregation sessions with this peer. Signed-off-by:
-
Johannes Berg authored
Allow peers to size their reorder buffer more accurately by advertising that we'll never send aggregates longer than the default (31). Signed-off-by:
-
Johannes Berg authored
In order to support RSN IBSS, we need to (ok actually maybe it's just easiest to) disable group key programming so that any group-addressed frames will be decrypted in software which handles the per-station keys for this easily. We could keep the encryption in the device, but that takes more work and seems unnecessary. Signed-off-by:
-
Wey-Yi Guy authored
For agn devices, always support idle mode which help power consumption in idle unassociated state. Signed-off-by: -
Wey-Yi Guy authored
The I/Q swapping is extremely important and should be dealt with extra care. It will affects OFDM and CCK differently. For 6000/6005/6030 series devices, the I/Q were swapped, and for 2000 series devices, it is in non-swapped status (but its swapped with respected to 6000/6005/6030). so the CSR_GP_DRIVER_REG_BIT_RADIO_IQ_INVER register need to be set to support the correct behavior. Signed-off-by: -
Wey-Yi Guy authored
Minor adjustment for rate scale table Signed-off-by: -
Wey-Yi Guy authored
Driver check alive message from ucode, if it is not ok, then need to restart the loading process. instead of checking multiple places for failure, only need to check in once place when receive alive message from uCode. Signed-off-by: -
git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6John W. Linville authored
Conflicts: drivers/net/wireless/wl12xx/spi.c
-
Luciano Coelho authored
Commit 0b01f030 added a new argument to the ampdu_action operation. The ampdu_action operation in the wl12xx driver currently doesn't have that argument and this generates a warning. This happened during merging of the latest mac80211 patches with the wl12xx BA patches. CC [M] drivers/net/wireless/wl12xx/main.o drivers/net/wireless/wl12xx/main.c:3035: warning: initialization from incompatible pointer type The wl12xx driver doesn't need to do anything about the buf_size argument since the AMPDU TX is fully handled by the firmware. Signed-off-by:
Luciano Coelho <coelho@ti.com> Signed-off-by:
-
Rajkumar Manoharan authored
PS should not be enabled if an infra AP vif exists in the interface list. So while recalculating PS, AP vif type should be taken into account. Reviewed-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
Rajkumar Manoharan <rmanoharan@atheros.com> Signed-off-by:
-
Stanislaw Gruszka authored
After commit 59eb21a6 "cfg80211: Extend channel to frequency mapping for 802.11j" we use uninitialized sband->band when assign channel frequencies, what results that 5GHz channels have erroneous (zero) center_freq value. Patch fixes problem and simplifies code a bit. Signed-off-by:
Stanislaw Gruszka <sgruszka@redhat.com> Reviewed-by:
-
Ivo van Doorn authored
The TX/RX work structures must be able to run independently of other workqueues. This is because mac80211 might use the flush() callback function from various context, which depends on the TX/RX work to complete while the main thread is blocked (until the the TX queues are empty). This should reduce the number of 'Queue %d failed to flush' warnings. Signed-off-by:
Ivo van Doorn <IvDoorn@gmail.com> Acked-by:
Gertjan van Wingerde <gwingerde@gmail.com> Signed-off-by:
-
Ivo van Doorn authored
During device removal all pending work and tasklets must be guaranteed to be halted. So far only the txstatus_tasklet was killed. Signed-off-by:
Helmut Schaa <helmut.schaa@googlemail.com> Acked-by:
-
Gertjan van Wingerde authored
Both rt61pci and rt73usb check the Michael MIC in hardware and strip the Michael MIC from received frames. This is perfectly allowed by mac80211 as long as this is properly reported to mac80211. Both these drivers reported the Michael MIC handling properly to mac80211, but still contained a FIXME comment on this, which is not needed to be handled, since mac80211 doesn't really need the Michael MIC in this case. Signed-off-by:
-
Gertjan van Wingerde authored
Use the specific mac field of the wcid_entry structure to copy the MAC address to, instead of just overwriting the structure. Previous code resulted in the same, but this form is cleaner. Signed-off-by:
-
Gertjan van Wingerde authored
As reported and found by Johannes Stezenbach: rt2800{pci,usb} do not report the Michael MIC in RXed frames, but do check the Michael MIC in hardware. Therefore we have to report to mac80211 that the received frame does not include the Michael MIC. Signed-off-by: -
Johannes Stezenbach authored
Acked-by:
Johannes Stezenbach <js@sig21.net> Signed-off-by:
-
Mark Einon authored
Fixing a trivial comment typo. Signed-off-by:
Mark Einon <mark.einon@gmail.com> Acked-by:
-
Helmut Schaa authored
Signed-off-by:
-
RA-Jay Hung authored
CLOCK CYCLE: Clock cycle count in 1us PCI:0x21, PCIE:0x7d, USB:0x1e Signed-off-by:
RA-Jay Hung <jay_hung@ralinktech.com> Acked-by:
-
Helmut Schaa authored
Remove STATE_RADIO_IRQ_OFF_ISR and STATE_RADIO_IRQ_ON_ISR as they are not used anymore. Signed-off-by:
-
Helmut Schaa authored
No driver uses interrupt threads anymore. Remove the remaining interrupt thread artifacts. Signed-off-by:
-
Helmut Schaa authored
Fix interrupt processing on slow machines by using individual tasklets for each different device interrupt. This ensures that while a RX or TX status tasklet is scheduled only the according device interrupt is masked and other interrupts such as TBTT can still be processed. Also, this allows us to use tasklet_hi_schedule for TBTT processing which is required to not send out beacons with a wrong DTIM count (due to delayed periodic beacon updates). Furthermore, this improves the latency between the TBTT and sending out buffered multi- and broadcast traffic. As a nice bonus, the interrupt handling overhead should be much lower. Compile-tested only. Signed-off-by:
-
Helmut Schaa authored
Fix interrupt processing on slow machines by using individual tasklets for each different device interrupt. This ensures that while a RX or TX status tasklet is scheduled only the according device interrupt is masked and other interrupts such as TBTT can still be processed. Also, this allows us to use tasklet_hi_schedule for TBTT processing which is required to not send out beacons with a wrong DTIM count (due to delayed periodic beacon updates). Furthermore, this improves the latency between the TBTT and sending out buffered multi- and broadcast traffic. As a nice bonus, the interrupt handling overhead should be much lower. Compile-tested only. Signed-off-by:
-
Helmut Schaa authored
Fix interrupt processing on slow machines by using individual tasklets for each different device interrupt. This ensures that while a RX or TX status tasklet is scheduled only the according device interrupt is masked and other interrupts such as TBTT can still be processed. Also, this allows us to use tasklet_hi_schedule for TBTT processing which is required to not send out beacons with a wrong DTIM count (due to delayed periodic beacon updates). Furthermore, this improves the latency between the TBTT and sending out buffered multi- and broadcast traffic. As a nice bonus, the interrupt handling overhead should be much lower. Compile-tested only. Signed-off-by:
-
Helmut Schaa authored
Fix interrupt processing on slow machines by using individual tasklets for each different device interrupt. This ensures that while a RX or TX status tasklet is scheduled only the according device interrupt is masked and other interrupts such as TBTT can still be processed. Also, this allows us to use tasklet_hi_schedule for TBTT and PRETBTT processing which is required to not send out beacons with a wrong DTIM count (due to delayed periodic beacon updates). Furthermore, this improves the latency between the TBTT and sending out buffered multi- and broadcast traffic. As a nice bonus, the interrupt handling overhead is reduced such that rt2800pci gains around 25% more throuhput on a rt3052 MIPS board. Signed-off-by:
-
Helmut Schaa authored
Enable the txstatus tasklet when interrupts are enabled and disable it together with the interrupts. Also make the txstatus tasklet useful even without the tx status FIFO and make use of the generic rt2x00 tasklet initialization macro. Signed-off-by:
-
