On Sun, 20 Dec 2009, Filip Zyzniewski wrote: > In the meantime I have followed a suggestion from mkl lin > <mkl0301@xxxxxxxxxxx> to disable d-cache: > > > I did it. I prepared a minimal kernel with this config: > > http://filip.math.uni.lodz.pl/jornada/usb_host/dmesgs/kernelconfig > > I ran two test sessions. One with d-cache enabled, and one with > > d-cache disabled. > > > > Tested device was connected before booting. > > Testing with d-cache disables was a royal PITA, because it tends to > > freeze in random moments and > > reboot procedure of the jornada is a bit painful (running through > > wince dialogs). > > > > I have placed logs from various tests here: > > http://filip.math.uni.lodz.pl/jornada/usb_host/dmesgs/ > > > > It seems to go further with device initalisation without d-cache. It > > can even sometimes bind a driver > > to a device: > > http://filip.math.uni.lodz.pl/jornada/usb_host/dmesgs/without_d_cache/mouse.txt > > http://filip.math.uni.lodz.pl/jornada/usb_host/dmesgs/without_d_cache/gps-5.txt > > > > and works a bit with a hub: > > http://filip.math.uni.lodz.pl/jornada/usb_host/dmesgs/without_d_cache/hub-pendrive-mouse-3.txt > > > > I am not sure this is a problem with cache handling or maybe it does > > better because of a massive slowdown > > (4 bogomips instead of 140). > > Is this a valuable trace? I looked at one of these logs. Apparently there's a problem with the memory caching and DMA. With some of the commands, the debugging message shows incorrect data in the buffer. But later on, when the data gets used, it obviously is correct. This implies that something is wrong with the implementation of dma_map_single() or dma_unmap_unsingle(). And it might explain why changing the system's cache settings affects the behavior. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
