On Thu, Apr 21, 2011 at 5:47 PM, Daniel GlÃckner <daniel-gl@xxxxxxx> wrote: > On Thu, Apr 21, 2011 at 04:20:36PM +0800, Bob Liu wrote: >> > on mmu systems do_mmap_pgoff contains a len = PAGE_ALIGN(len); line. >> > If we depend on this behavior, why not do it here as well and get rid >> > of the #ifdef? >> > >> >> If do it in do_mmap_pgoff() the whole system will be effected, I am >> not sure whether >> it's correct and needed for other subsystem. > > With "here" I was referring to uvc_queue_mmap. > I am sorry, I didn't get your idea. You mean using PAGE_ALIGN() here for both mmu and no-mmu arch ? >> >> + Â Â addr = (unsigned long)queue->mem + buffer->buf.m.offset; >> >> + Â Â ret = addr; >> > >> > Why the intermediate step using addr? >> > >> >> If don't return addr, do_mmap_pgoff() will return failure and we can't >> setup vma correctly. >> See mm/nommu.c line 1386(add = file->f_op->get_unmmapped_area() ). > > I know, but why not do > Â Â Â Âret = (unsigned long)queue->mem + buffer->buf.m.offset; > instead? > Okay. Thanks -- Regards, --Bob -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html