IrDA patches for 2.6.9
Hi David,
This are all the IrDA patches that were pending, ready for
inclusion in 2.6.9. Tested on my SMP box and laptop with 2.6.9, and by various IrDA users.
Most changes will be invisible to users, apart from the via
and stir changes that allow the drivers to work better.
Linus might be busy under the mountain of patches, but I'm
sure he need my patches ;-)
Jean
-------------------------------------------------------------
[FEATURE] : Add a new feature to the IrDA stack [CORRECT] : Fix to have the correct/expected behaviour [CRITICA] : Fix potential kernel crash
ir269_lmp_lsap_inuse-3.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~~~
o [CRITICA] Fix locking in error path in IrLMP (Stanford checker)
o [CORRECT] Don't reuse unconnected LSAPs (listening sockets)
o [CORRECT] Make sure the LSAP we are picking has just not been grabed
o [CORRECT] Wrap around the LSAP space properly back to 0x10
ir269_nsc_dongle-2.diff :
~~~~~~~~~~~~~~~~~~~~~~~
<Original patch from Maik Broemme>
o [CRITICA] Don't Oops on invalid dongle-id in nsc-ircc driver
ir269_irnet_alias.diff :
~~~~~~~~~~~~~~~~~~~~~~
o [FEATURE] Add module alias for IrNET char dev
ir269_ias_safety.diff :
~~~~~~~~~~~~~~~~~~~~~
o [FEATURE] Make optional the del of IAS object when del IAS attrib
o [FEATURE] Clarify when/why it's safe to to the above
ir269_adaptive_query_timer-2.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
o [FEATURE] Adapt to the rate of the peer discovery (passive discovery)
o [FEATURE] Add extra safety margin in passive discovery Allow to interoperate properly with device performing slow discovery
ir269_ircomm_ias_fix-1.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~~~
o [CORRECT] Restore properly the IAS object when IrCOMM disconnect. Allow "pppd passive persist" to work properly.
ir269_via_speed_fixes-2.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
o [CORRECT] Speed change fixes in via-ircc driver
o [FEATURE] Add new dongle ID in via-ircc driver
o [FEATURE] Various code cleanups in via-ircc driver
irXXX_debug_mod_parm.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~
<Patch from Stephen Hemminger>
o [CORRECT] irda 2.6 - fix module info
The module parameter info for irda is incorrect.
The debug parameter is named "debug", the variable is irda_debug.
irXXX_stir_reset.diff :
~~~~~~~~~~~~~~~~~~~~~
<Patch from Stephen Hemminger>
o [CORRECT] stir4200 - get rid of reset on speed change
The Sigmatel 4200 doesn't accept the address setting which gets done on USB reset. The USB core recently changed to resend address (or something like that), so usb_reset_device is failing.
The device works without doing the USB reset on speed change, it just will be less robust in recovering when things get wedged (like coming out of FIR mode).
irXXX_stir_suspend.diff :
~~~~~~~~~~~~~~~~~~~~~~~
<Patch from Stephen Hemminger>
o [FEATURE] stir4200: don't need suspend/resume if !CONFIG_PM
The suspend/resume code only needs to be compiled in if power management is enabled.
irXXX_stid_netdev_priv.diff :
~~~~~~~~~~~~~~~~~~~~~~~~~~~
<Patch from Stephen Hemminger>
o [FEATURE] stir4200: netdev_priv and message cleanup
Stir4200 driver cleanup's:
* use netdev_priv
* make sure messages identify the driver
* get rid of unneeded message