On Wed, Dec 09, 2009 at 11:43:57AM +0000, Hugh Dickins wrote: > ... my hair ;-) > I have to say, Dr Frankenstein, that this idea fills me with dread. > > I'm not saying it's impossible, but the resulting creature sounds like > it's going to be special in several easily-buggy hard-to-maintain ways. > > I think you already realize that shmem file pages (shared) live by > different rules from anonymous pages (COWed): they're both swappable, > but switching a group of pages from one to the other is going to be > weird new territory. (In fairness, my suggestion involves some > weird new territory too, but considerably less scary to me.) Umm... Note that these guys are considerably simpler than shmem; _nothing_ will have them mapped anywhere until after that eviction and nothing will modify their address_space in any way (no truncation, etc.). And no, I'm not particulary happy about the picture - the sight of, ahem, locking hierarcy rules in mm/filemap.c alone is a cause for dread. It had been a while since I'd done any serious RTFS on mm/*, so... Anyway, it's obviously *not* -rc1 fodder at the moment. So what I'm going to do is to leave that one as it is in mainline, push the rest of mmap/mremap stuff to Linus later today and get to VFS and audit queues. After the -rc1, OTOH... -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
