
16 Nov
2011
16 Nov
'11
3:26 a.m.
On Tuesday 15 November 2011 13:01:52 Stefan Roese wrote:
--- a/arch/powerpc/cpu/ppc4xx/usbdev.c +++ b/arch/powerpc/cpu/ppc4xx/usbdev.c
- int temp, temp1;
...
/*copy packet */ setup_packet_pt[0] = *(unsigned int *)USB2D0_FIFO_0; setup_packet_pt[1] = *(unsigned int *)USB2D0_FIFO_0;
temp = *(unsigned int *)USB2D0_FIFO_0;
temp1 = *(unsigned int *)USB2D0_FIFO_0;
this seems to be treading into possible ugly volatile area ... perhaps best if this was acked/tested by someone with actual hardware first ... -mike
My words definitelly ... maybe replace with in_be32() ? It seems very suspicious, as if there was a reason for this access. Like you need to do more accesses to the fifo to clean up some trailing crud.
M