- 29 Feb, 2008 40 commits
-
-
Michael Buesch authored
This adds support for 8bit wide register reads/writes. This is needed in order to support the gigabit ethernet core. Signed-off-by:
Michael Buesch <mb@bu3sch.de> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Johannes Berg authored
This consolidates all TIM handling code to avoid re-introducing errors with the bitmap/set_tim order and to reduce code. While reading the code I noticed a possible problem so I also added a comment about that. Signed-off-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
-
Johannes Berg authored
The TIM flag that is kept in each station's info is completely useless, there's no code (aside from the debugfs display code) checking it, hence it can be removed. While doing that, I noticed that the TIM handling is broken when buffered frames expire, so fix that. Signed-off-by:
-
Johannes Berg authored
Drivers should be allowed to simply get a complete new beacon when set_tim() is invoked (and set_tim() is required for drivers that just want a beacon template!), so we need to update our own TIM bitmap before calling set_tim() so that getting the beacon will now get an already updated beacon. Signed-off-by:
-
Roel Kluin authored
Convert list_for_each_rcu() to list_for_each_entry_rcu() Signed-off-by:
Roel Kluin <12o3l@tiscali.nl> Signed-off-by:
-
Tomas Winkler authored
This patch adds channels to US regulatory domain Signed-off-by:
Tomas Winkler <tomas.winkler@intel.com> Signed-off-by:
-
Ivo van Doorn authored
Signed-off-by:
Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by:
-
Ivo van Doorn authored
The MAC address offset defines were incorrect because the byte offset was used instead of word index. This bug had no affect on normal operations since these defines weren't used. (EEPROM_MAC_ADDR_0 was used to read 6 bytes from). Signed-off-by:
-
Gertjan van Wingerde authored
The supported_bands field of struct hw_mode_spec now represents a bitfield, so bitfield operators need to be tested with when setting the band data. The current code generates the following warning: [176624.986244] WARNING: at /usr/local/src/incoming/compat-wireless-2.6/net/wireless/core.c:269 wiphy_register() [176624.986249] Pid: 12548, comm: modprobe Tainted: P 2.6.24.2#4 [176624.986251] [176624.986251] Call Trace: [176624.986277] [<ffffffff881c56bf>] :cfg80211:wiphy_register+0x17f/0x1a0 [176624.986282] [<ffffffff881ddf80>] :rt61pci:rt61pci_eepromregister_write+0x0/0x80 [176624.986302] [<ffffffff88b7e4bc>] :mac80211:ieee80211_register_hw+0x2c/0x2b0 [176624.986310] [<ffffffff881cdc80>] :rt2x00lib:rt2x00lib_probe_dev+0x350/0x3f0 [176624.986318] [<ffffffff881d74b9>] :rt2x00pci:rt2x00pci_probe+0x149/0x200 [176624.986325] [<ffffffff8030c858>] pci_device_probe+0xf8/0x170 [176624.986331] [<ffffffff803594fc>] driver_probe_device+0x9c/0x1c0 [176624.986335] [<ffffffff80359700>] __driver_attach+0x0/0xb0 [176624.986337] [<ffffffff803597a5>] __driver_attach+0xa5/0xb0 [176624.986341] [<ffffffff8035877d>] bus_for_each_dev+0x4d/0x80 [176624.986347] [<ffffffff80358b8c>] bus_add_driver+0xac/0x210 [176624.986351] [<ffffffff8030cad3>] __pci_register_driver+0x73/0xc0 [176624.986357] [<ffffffff8025689e>] sys_init_module+0x18e/0x1a20 [176624.986374] [<ffffffff8020c42e>] system_call+0x7e/0x83 Signed-off-by:
Gertjan van Wingerde <gwingerde@kpnplanet.nl> Signed-off-by:
-
Ivo van Doorn authored
The async vendor requests are a ugly hack which is not working correctly. The proper fix for the scheduling while atomic issue is finding out why we can't use led classes for USB drivers and fix that. Just replace all async calls with the regular ones and print an error for the disallowed LED configuration attempts. That will help in determining which led class is causing the problem. Signed-off-by:
-
Ivo van Doorn authored
Don't wildly pass any number for num_rates to rt2x00lib, instead pass which type of rates are supported (CCK, OFDM). Same for num_modes but then for the 2GHZ and 5GHZ band. This makes the interface look much nicer and makes extending it later easier. Signed-off-by:
-
Ivo van Doorn authored
The dscape stack was renamed to mac80211 a long time ago, we are long overdue with fixing all comments to reflect this. Signed-off-by:
-
Ivo van Doorn authored
Specifications indicate the TSF registers are read-only, so there is no point in writing 0 to those registers. As far as I know there isn't another way to reset the TSF registers. So removing these callbacks will notify mac80211 about the lack of support. Signed-off-by:
-
Ivo van Doorn authored
As Adam Baker reported the DMA address for the descriptor base was incorrectly initialized in the PCI drivers. Instead of the DMA base for the descriptor, the DMA base for the data was passed resulting in a broken TX/RX state for PCI drivers. Signed-off-by:
-
Ivo van Doorn authored
The ACK_CTS frame is a control frame, this means dropping the frame depends on the FIF_CONTROL flag for filtering. This also fixes an obvious typo in register definition. Signed-off-by:
-
Ivo van Doorn authored
This adds a new flag for the skb_frame_desc structure which is used to tag rts/cts frames that are generated by the driver. Through the tag we can recognize frames we have generated ourselves, so we don't report their tx status to mac80211. This patch is based on the original patch by Mattias Nissler <mattias.nissler@gmx.de>. Signed-off-by:
-
Ivo van Doorn authored
mac80211 sends the txpower to use during config(), we already store it in the rt2x00_dev structure. When writing the descriptor correctly initialize the txpower field with this value to make sure all frames are send out with the correct tx power. Signed-off-by:
-
Ivo van Doorn authored
Signed-off-by:
-
Ivo van Doorn authored
When the radio is being disabled we should also kill the guardian urb which could still be pending in the device. Signed-off-by:
-
Ivo van Doorn authored
Simplify the way rt2x00 assigns new objects to the rt2x00lib module. This saves a few if statements and overall does this looks much nicer. Signed-off-by:
-
Ivo van Doorn authored
As Adam Baker reported the queue->qid was not initialized correctly. The QID_AC_BE was assigned to the RX ring. This will move the queue initialization into a seperate function and makes sure that all queues are initialized directly with the correct qids. Signed-off-by:
-
Ivo van Doorn authored
skbdesc->data_len was not initialized correctly in rt2x00pci, rt2x00usb, rt2500usb and rt73usb. The value was set to queue->data_size which means that the incorrect frame size was pased to the upper layers. Correctly base the value on either the skb->len, or the rx frame size passed to the driver by the device. Signed-off-by:
-
Ivo van Doorn authored
The second eeprom recovery message is about the RSSI offset for ieee802.11 A. Signed-off-by:
-
Ivo van Doorn authored
Rename "frame" folder to "queue" folder, add extra file to this folder which contains statistics about all hardware queues. This will help debugging and spotting problems in the queue indexing system. Signed-off-by:
-
Ivo van Doorn authored
Call rt2x00_config_intf() outside of the spinlock context since the call will sleep for USB drivers. By using the ieee80211_if_conf values as arguments we make keep access tp rt2x00_intf thread safe even without the lock. Signed-off-by:
-
Ivo van Doorn authored
The csr_cache and csr_addr pointers are both the same size and they are never used both by the same driver. This makes them a nice candidate for an union. We could merge into 1 pointer, but that would either upset sparse, or require a lot of __force casts. Signed-off-by:
-
Ivo van Doorn authored
Select CONFIG_NEW_LEDS before selecting the other LED config options. This fixes a link error when NEW_LEDS was disabled. Signed-off-by:
-
Adam Baker authored
When calculating the offset to add to the queue entry base to get the individual entry's private data area the base address must be treated as a char * not a struct queue_entry so we can do byte oriented pointer arithmetic with it. Signed-off-by:
Adam Baker <linux@baker-net.org.uk> Signed-off-by:
-
Ivo van Doorn authored
Remove the last remnants of the MGMT ring initialization from rt61pci.ko Signed-off-by:
-
Ivo van Doorn authored
Remove the debug messages regarding initialization from EEPROM. The values are vendor specific, and are not really needed for debug purposes. If they ever become usefull we still have access to them through debugfs which also prints the exact same values... Signed-off-by:
-
Ivo van Doorn authored
Be more strict when using the queue_entry_priv_pci_rx and queue_entry_priv_pci_tx structures. Only use a particular type that matches the queue type. When freeing the DMA the priv_tx->data and priv_tx->dma was used. This is incorrect since the start of the DMA was in fact the priv_tx->desc pointer. Instead of recalculating the dma_addr_t for the DMA start this patch will swap the data and descriptor part of the allocated memory. Signed-off-by:
-
Randy Dunlap authored
Michael Büker <m.bueker@berlin.de> reports that the RT2x00 drivers are not indented as they should be, so use proper dependencies to make them be indented as expected. Signed-off-by:
Randy Dunlap <randy.dunlap@oracle.com> Ack-by:
-
Matthias Kaehlcke authored
The semaphore wpa_sem is used as mutex, convert it to the mutex API Signed-off-by:
Matthias Kaehlcke <matthias@kaehlcke.net> Acked-by:
Luis R. Rodriguez <mcgrof@winlab.rutgers.edu> Signed-off-by:
-
Matthias Kaehlcke authored
The semaphore stats_sem is used as mutex, convert it to the mutex API Signed-off-by:
-
Matthias Kaehlcke authored
The semaphore acl->sem is used as mutex, convert it to the mutex API Signed-off-by:
-
Marcin Slusarz authored
replace all: little_endian_variable = cpu_to_leX(leX_to_cpu(little_endian_variable) + expression_in_cpu_byteorder); with: leX_add_cpu(&little_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by:
Marcin Slusarz <marcin.slusarz@gmail.com> Cc: Zhu Yi <yi.zhu@intel.com> Signed-off-by:
Reinette Chatre <reinette.chatre@intel.com> Signed-off-by:
-
Daniel Drake authored
Most wireless drivers load their firmware at interface open time, which generally occurs after the filesystem is available. However, the ipw drivers load their firmware at probe time because firmware is required to read the device MAC address. When built-in, probe happens before the filesystem is available, hence device init will only complete successfully if the user has made special arrangements (including firmware plus a loader in the initramfs). Note all this in the kconfig help text for both drivers. Signed-off-by:
Daniel Drake <dsd@gentoo.org> Signed-off-by:
-
Christian Lamparter authored
This trival one-liner changes the QoS initialization values to match IEEE 802.11e defaults. Signed-off-by:
Christian Lamparter <chunkeey@web.de> Acked-by:
-
Michael Buesch authored
We must not clear the FIF_BCN_PRBRESP_PROMISC bit in the new_flags. The zd-driver does support sending beacons and probe responses to the host. What the flag does is say "Send me all beacons and probe responses". And we actually do that. We always do that, so we ignore the case when the bit is disabled. But that is fine. But we must not clear the flag, as that tells mac80211 that we do not support passing beacons and probe responses to the stack. And that's not true. Signed-off-by:
-
Javier Cardona authored
Trial and error reveals that CR_ZD1211B_TX_PWR_CTL* do not affect the transmission power. Instead these registers seem to control the contention windows limits for different QoS access categories. Signed-off-by:
Javier Cardona <javier@cozybit.com> Signed-off-by:
-
