Update from the ML:
Quote:
|
Originally Posted by Jerome
Hi,
Alex found that UART2 irqs (defined in irqs.h) where in a different order for uart1dm and uart2dm. We had to patch the msm_serial_hs.c driver so that uart2 could be used but the real problem was that the irqs were reversed.
Thus, this was causing arm9 to wake up arm11 when BT was on, because the wrong irq was selected in arm9 to wake up arm11.
With the attached patch, the irqs are now in the correct order (this has been fixed on irqs8xxx.h files in newer trees. In fact, no devices using 7xxx processors are using uart2 so they probably never encountered this bug). The patch also revert the change made to the msm_serial_hs.c driver.
Also, I made a small change in the debug outputs from smsm interrupts so that thing are displayed correctly now (typo + reversed variables order).
I've tested this on both diamond and rhodium device so I can confirm BT is still working on both TI and Broadcom bluetooth devices. The phone now successfully stay in sleep mode even if BT is on. It wakes up when BT is used (on BT events).
Regards, Jerome
|
Soon guys!
Edit - he still hasn't posted his libhtcacoustic patches, which is the userland gold... Hopefully we'll see that surface, he makes it sound like its working for him...