> From: Peter Chen <peter.chen@xxxxxxxxxxxxx> > Sent: Monday, March 25, 2013 7:58 PM > > On Fri, Mar 22, 2013 at 02:20:23PM +0100, Michael Grzeschik wrote: > > The udc uses an shared dma memory space between hard and software. This > > memory layout is described in ci13xxx_qh and ci13xxx_td which are marked > > with the attribute ((packed)). > > > > The packed attribute leads the compiler to generate one byte operations > > for addressing the mapped memory as it believes this memory has no > > alignement issues as common streaming data. This appeares on armv5 > > machines where the hardware does not support unaligned 32bit operations. > > Compilers for newer ARMs will probably still generate 32bit operations, > > as the hardware supports it. > > > > The Datasheet of the synopsys core describes, that some operations on > > the mapped memory need to be atomic double word operations. I.e. the > > next pointer addressing in the qhead, as otherwise the hardware could > > read wrong data and totally stuck. > > > > This patch fixes that issue by addressing every mapped area operation > > with explicit readl and writel operations where needed. It also adds an > > wmb() for the prepared TD data before it gets enqueued into the qhead. > > The writel includes wmb(), the wmb() isn't needed if the last instruction > is writel. > > Besides, please make sure all dQH/dTD mapped area access by readl/writel in > this patch, we don't want to add missing in future. Is 'sparse' happy with this? AFAIR it will complain if you use readl/writel on memory that is not marked with __iomem. Is removing the __attribute__ ((packed)) annotation not enough to fix the problem? -- Paul -- 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
