diff options
| author | Bart Van Assche <bart.vanassche@sandisk.com> | 2016-09-26 22:58:49 +0300 |
|---|---|---|
| committer | Ben Hutchings <ben@decadent.org.uk> | 2017-02-23 06:54:10 +0300 |
| commit | 00a05de9ae402f3408fe7fb43fb5654c435831ed (patch) | |
| tree | 0de30e2a48c215d17cf3efac8f60f664e707b9e7 /drivers/infiniband/ulp | |
| parent | 3c54629c642f3b11dd33e412b9519091b5bb3ede (diff) | |
| download | linux-00a05de9ae402f3408fe7fb43fb5654c435831ed.tar.xz | |
IB/srp: Fix infinite loop when FMR sg[0].offset != 0
commit 681cc3608355737c1effebc8145f95c8c3344bc3 upstream.
Avoid that mapping an sg-list in which the first element has a
non-zero offset triggers an infinite loop when using FMR. This
patch makes the FMR mapping code similar to that of ib_sg_to_pages().
Note: older Mellanox HCAs do not support non-zero offsets for FMR.
See also commit 8c4037b501ac ("IB/srp: always avoid non-zero offsets
into an FMR").
Reported-by: Alex Estrin <alex.estrin@intel.com>
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Diffstat (limited to 'drivers/infiniband/ulp')
| -rw-r--r-- | drivers/infiniband/ulp/srp/ib_srp.c | 8 |
1 files changed, 5 insertions, 3 deletions
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c index 36876960f97e..6311fd448c36 100644 --- a/drivers/infiniband/ulp/srp/ib_srp.c +++ b/drivers/infiniband/ulp/srp/ib_srp.c @@ -1310,7 +1310,9 @@ static int srp_map_sg_entry(struct srp_map_state *state, while (dma_len) { unsigned offset = dma_addr & ~dev->mr_page_mask; - if (state->npages == dev->max_pages_per_mr || offset != 0) { + + if (state->npages == dev->max_pages_per_mr || + (state->npages > 0 && offset != 0)) { ret = srp_finish_mapping(state, target); if (ret) return ret; @@ -1329,12 +1331,12 @@ static int srp_map_sg_entry(struct srp_map_state *state, } /* - * If the last entry of the MR wasn't a full page, then we need to + * If the end of the MR is not on a page boundary then we need to * close it out and start a new one -- we can only merge at page * boundries. */ ret = 0; - if (len != dev->mr_page_size) { + if ((dma_addr & ~dev->mr_page_mask) != 0) { ret = srp_finish_mapping(state, target); if (!ret) srp_map_update_start(state, NULL, 0, 0); |