#11050 NORM Future : sd8686 firmware often crashes soon after wake-on-WLAN

[Zarro Boogs per Child]

#11050: sd8686 firmware often crashes soon after wake-on-WLAN
----------------------+-----------------------------------------------------
 Reporter:  dsd       |                 Owner:  dsd           
     Type:  defect    |                Status:  new           
 Priority:  normal    |             Milestone:  Future Release
Component:  wireless  |               Version:  not specified 
 Keywords:            |           Next_action:  never set     
 Verified:  0         |   Deployment_affected:                
Blockedby:            |              Blocking:                
----------------------+-----------------------------------------------------
 In #10748 we saw that soon after resuming the system, where WLAN was the
 wakeup source, the libertas hardware often fails to respond to commands.
 More specifically, it doesn't even send the 0x2 interrupt
 (IF_SDIO_H_INT_DNLD) to confirm that it is processing the command. We
 worked around this by fixing up command timeout detection logic, and
 implementing the ability for libertas to reset the SD card.

 After fixing that, in #10969 we then saw a similar issue where libertas
 would freeze up during system resume. In this case, it was not commands
 but transmitted data frames that were holding things up. As the hardware
 can only handle one host-to-card communication at a time, libertas
 correctly was holding everything up pending acknowledgement that the data
 frame would be transmitted by the card. We worked around this by adding a
 timer that resets the card if a transmission is not acknowledged by the
 card within 5 seconds.

 Even though workarounds are in place, the approach of resetting the card
 is not great. It disrupts any ongoing network communications and software
 is not guaranteed to reestablish (or to be able to reestablish) the same
 connection after the reset.

 This ticket tracks the solution of the actual bug which resides in the
 sd8686 firmware or hardware. It is reported on marvell symplicity as
 #451148.

-- 
Ticket URL: <http://dev.laptop.org/ticket/11050>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system