Hi Marc, Marc Kleine-Budde [mailto:mkl@xxxxxxxxxxxxxx] (On Thursday, March 28, 2013 4:16 PM) >On 03/28/2013 10:28 AM, Alexander Shishkin wrote: >> Svetoslav Neykov <svetoslav@xxxxxxxxxxx> writes: >> >>> Convert between big-endian and little-endian format when accessing >>> the usb controller structures which are little-endian by >>> specification. Fix cases where the little-endian memory layout is >>> taken for granted. The patch doesn't have any effect on the already >>> supported little-endian architectures. > >Has anyone tested how the cpu_to_le32 and vice versa effects the >load/store operations? Does the compiler generate full 32 bit accesses >on mips (and big endian arm) or is a byte-shift-or pattern used? Better late than never... I have checked your question, the value is loaded in a register and then swapped, so the read is performed only once. Regards, Svetoslav.