"ext Frantisek Dufka" <[EMAIL PROTECTED]> writes: >> [50622.038146] We haven't got a READY interrupt from WAKEUP. (firmware >> crashed?) >> [50622.038269] Try again... >> [50622.038330] succeeded!!! >> >> I'm attaching the same patch here. It is not very clean, but it does >> the job (for Nokia 770). > > Not seeing whole code now but I wondered before - isn't it possible > just to increase timeout inside 'if (time_after(jiffies, timeout))' > instead of two passes in loop?
I also haven't seen the full code, only took a quick look at the patch. But what I saw is that the whole frame is retransmitted, instead of just waiting for the WR_READY interrupt from the firmware. In other words, if the frame is lost during the tranmission to firmware, the firmware won't receive the frame and it will not assert the WR_READY interrupt. So it doesn't help even if the driver waits longer, only frame restransmission helps. > The N800 code looks nicer, where the CPU is burnt there? The N800/N810 driver uses omap2_mcspi and SPI framework for accessing the SPI bus. Basically omap2_mcpi waits for DMA interrupts to know when the transfer has finished. But there are lots context switching between cx3110x and omap2_mspi threads, that might consume some extra CPU cycles. -- Kalle Valo _______________________________________________ maemo-developers mailing list maemo-developers@maemo.org https://lists.maemo.org/mailman/listinfo/maemo-developers