On Tue, 09 Mar 2010 09:07:34 -0600 "Steven J. Magnani" <steve@xxxxxxxxxxxxxxx> wrote: > On Tue, 2010-03-09 at 14:29 +0000, David Howells wrote: > > > However, it does occur to me that if it's not pinning the last page correctly, > > shouldn't that cause a usage counter underflow error on that page eventually > > when it's returned to the system allocator? > > Not in my scenario. All pages but the last get pinned; > drivers/dma/iovlock.c::dma_pin_iovec_pages() detects the mismatch > between what was requested and what was pinned, and unpins everything in > the pinned list. So AFAICT the net effect is that the network stack > falls back to plain old memcpy. > Guys, what do we need to do to get this issue wrapped up? Thanks. -- To unsubscribe from this list: send the line "unsubscribe mm-commits" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
