IB/qib: Convert qib_user_sdma_pin_pages() to use get_user_pages_fast()

qib_user_sdma_queue_pkts() gets called with mmap_sem held for
writing. Except for get_user_pages() deep down in
qib_user_sdma_pin_pages() we don't seem to need mmap_sem at all.  Even
more interestingly the function qib_user_sdma_queue_pkts() (and also
qib_user_sdma_coalesce() called somewhat later) call copy_from_user()
which can hit a page fault and we deadlock on trying to get mmap_sem
when handling that fault.

So just make qib_user_sdma_pin_pages() use get_user_pages_fast() and
leave mmap_sem locking for mm.

This deadlock has actually been observed in the wild when the node
is under memory pressure.

Cc: <stable@vger.kernel.org>
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Roland Dreier <roland@purestorage.com>

diff --git a/drivers/infiniband/hw/qib/qib_user_sdma.c b/drivers/infiniband/hw/qib/qib_user_sdma.c
index d0a0ea0c14d6a965afe2bcf8d359052fdb946e26..165aee2ca8a0c38dfc1e00b1965172acf4f30193 100644 (file)
--- a/drivers/infiniband/hw/qib/qib_user_sdma.c
+++ b/drivers/infiniband/hw/qib/qib_user_sdma.c
@@ -594,8 +594,7 @@ static int qib_user_sdma_pin_pages(const struct qib_devdata *dd,
                else
                        j = npages;
 
-               ret = get_user_pages(current, current->mm, addr,
-                            j, 0, 1, pages, NULL);
+               ret = get_user_pages_fast(addr, j, 0, pages);
                if (ret != j) {
                        i = 0;
                        j = ret;
@@ -1294,11 +1293,8 @@ int qib_user_sdma_writev(struct qib_ctxtdata *rcd,
                int mxp = 8;
                int ndesc = 0;
 
-               down_write(&current->mm->mmap_sem);
                ret = qib_user_sdma_queue_pkts(dd, ppd, pq,
                                iov, dim, &list, &mxp, &ndesc);
-               up_write(&current->mm->mmap_sem);
-
                if (ret < 0)
                        goto done_unlock;
                else {