1. 21 Nov, 2008 4 commits
    • Ivo van Doorn's avatar
      rt2x00: Fix race condition when using inderect registers · 8ff48a8b
      Ivo van Doorn authored
      
      
      Indirect registers require multiple calls to the CSR
      register in order to access the indirect registers.
      This must be protected under a lock to prevent race
      conditions which could cause invalid data to
      be returned when reading from the indirect register or silent
      failures when writing data to the indirect register.
      
      USB drivers where already protected under a mutex,
      so rename the mutex and make PCI drivers use the mutex
      as well.
      This now means that BBP and RF registers are no longer
      accessible in interrupt context. That is not a bad
      situation since the slow behavior of accessing
      those registers means we don't _want_ to access them
      in interrupt context either.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      8ff48a8b
    • Ivo van Doorn's avatar
      rt2x00: Cleanup TX/RX entry handling · 798b7adb
      Ivo van Doorn authored
      
      
      Merge the callback functions init_txentry() and
      init_rxentry(). This makes life in rt2x00lib a
      lot simpler and we can cleanup several functions.
      
      rt2x00pci contained "fake" FIELD definitions for
      descriptor words. This is not flexible since it
      assumes the driver will always have the same field
      to indicate if a driver is available or not.
      This should be dependent on the driver, and we
      should add a callback function for this.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      798b7adb
    • Ivo van Doorn's avatar
      rt2x00: Fix BUG_ON() with antenna handling · 6d64360a
      Ivo van Doorn authored
      
      
      With the new configuration handling, and more specifically
      splitting the configuration of the antenna from the normal
      configuration steps allowed a BUG_ON() to be triggered
      in the driver because the SW_DIVERSITY was send to the
      driver. This fixes that by catching the value early in
      rt2x00config.c and replacing it with a sensible value.
      
      This also fixes a problem where the antenna is not being
      initialized at all when the radio is enabled. Since it
      no longer is part of the mac80211 configuration the
      only place where rt2x00 configured it was the SW diversity
      handler. Obviously this is broken for all non-diversity
      hardware and breaks SW diversity due to a broken initialization.
      
      When the radio is enabled the antenna will be configured
      once as soon as the config() callback function is called.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      6d64360a
    • Ivo van Doorn's avatar
      rt2x00: Remove RATE_BASIC flag · 3d8606a6
      Ivo van Doorn authored
      
      
      mac80211 is in charge of determining the basic rates,
      so we are not using the RATE_BASIC flag anymore.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      3d8606a6
  2. 10 Nov, 2008 2 commits
  3. 31 Oct, 2008 3 commits
  4. 15 Sep, 2008 1 commit
  5. 05 Sep, 2008 1 commit
  6. 29 Aug, 2008 3 commits
    • Ivo van Doorn's avatar
      rt2x00: Add Signal type flag · 6c6aa3c0
      Ivo van Doorn authored
      
      
      Instead of using the PLCP flag to indicate if the
      signal value is plcp or the bitrate we should add
      a new flag to mark the bitrate type explicitely.
      This is usefull when new types are added later for
      rt2800.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      6c6aa3c0
    • Mattias Nissler's avatar
      rt2x00: Only configure hardware when radio is enabled · 2575c11d
      Mattias Nissler authored
      
      
      Some hardware configuration registers such as antenna and channel configuration
      can only be written when the radio is enabled. Previously, we didn't consider
      this, so some configuration items could be set inconsistently after reenabling
      the radio. This patch changes the config() handler to only reprogram the
      hardware when the radio is enabled. Configuration changes that are made while
      the radio is off are postponed until the radio is switched back on. We also
      leave the radio turned off during initialization and only enable it when
      requested by mac80211. This allows us to get rid of the DIRTY_CONFIG flag,
      because the device is now guaranteed to be completely initialized when brought
      up by mac80211.
      Signed-off-by: default avatarMattias Nissler <mattias.nissler@gmx.de>
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      2575c11d
    • Ivo van Doorn's avatar
      rt2x00: Fix race conditions in flag handling · 0262ab0d
      Ivo van Doorn authored
      
      
      Some of the flags should be accessed atomically to
      prevent race conditions. The flags that are most important
      are those that can change often and indicate the actual
      state of the device, queue or queue entry.
      
      The big flag rename was done to move all state flags to
      the same naming type as the other rt2x00dev flags and
      made sure all places where the flags were used were changed. ;)
      
      Thanks to Stephen for most of the queue flags updates,
      which fixes some of the most obvious consequences of the
      race conditions. Among those the notorious:
      
      rt2x00queue_write_tx_frame: Error - Arrived at non-free entry in the non-full queue 0.
      rt2x00queue_write_tx_frame: Error - Arrived at non-free entry in the non-full queue 0.
      rt2x00queue_write_tx_frame: Error - Arrived at non-free entry in the non-full queue 0.
      Signed-off-by: default avatarStephen Blackheath <tramp.enshrine.stephen@blacksapphire.com>
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      0262ab0d
  7. 22 Aug, 2008 2 commits
  8. 29 Jul, 2008 1 commit
    • Ivo van Doorn's avatar
      rt2x00: Force full register config after start() · 74c0ee9b
      Ivo van Doorn authored
      
      
      rt2x00 will only perform configuration changes from
      mac80211 when the configuration option has changed.
      This means it keeps track of the current active configuration
      and will check these values when the config() callback function
      is used.
      
      However this causes breakage when the interface has been
      brought down and up again, since all stored active values
      aren't reset while the registers might have.
      This is for example the case with rt61pci antenna registers which
      will jump to invalid values when the interface has been started.
      
      To make sure a full configuration takes place after the start()
      callback function, a new flag is added which will be checked
      during config() and skips the "what's changed" phase.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      74c0ee9b
  9. 14 Jul, 2008 2 commits
  10. 08 Jul, 2008 3 commits
  11. 27 Jun, 2008 1 commit
    • Ivo van Doorn's avatar
      rt2x00: Fix lock dependency errror · 980dfcb9
      Ivo van Doorn authored
      
      
      This fixes a circular locking dependency in the workqueue handling.
      The interface work task uses the mac80211 function
      ieee80211_iterate_active_interfaces() which grabs the RTNL lock.
      
      However when the interface is brough down, this happens under the RTNL
      lock as well, this causes problems because mac80211 will flush the workqueue
      during the ifdown event. This causes mac80211 to wait until the driver has
      completed all work which can't finish because it is waiting on the RTNL lock.
      
      This is fixed by moving rt2x00 workqueue tasks on a different workqueue,
      this workqueue can be flushed when the ieee80211_hw structure is removed
      by the driver (when the driver is unloaded) which does not happen under the
      RTNL lock.
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      980dfcb9
  12. 26 Jun, 2008 4 commits
  13. 14 Jun, 2008 3 commits
  14. 10 Jun, 2008 1 commit
  15. 28 May, 2008 2 commits
  16. 21 May, 2008 7 commits