diff options
| author | Linus Torvalds <torvalds@linux-foundation.org> | 2022-01-06 18:35:17 -0800 |
|---|---|---|
| committer | Linus Torvalds <torvalds@linux-foundation.org> | 2022-01-06 18:35:17 -0800 |
| commit | ddec8ed2d4905d0967ce2ec432e440e582aa52c6 (patch) | |
| tree | 3fab263b764d86e4ea2cfd56cdd171332ddf4965 /drivers/infiniband/core/uverbs_marshall.c | |
| parent | b2b436ec0205abde78ef8fd438758125ffbb0fec (diff) | |
| parent | b35a0f4dd544eaa6162b6d2f13a2557a121ae5fd (diff) | |
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
"Last pull for 5.16, the reversion has been known for a while now but
didn't get a proper fix in time. Looks like we will have several
info-leak bugs to take care of going foward.
- Revert the patch fixing the DM related crash causing a widespread
regression for kernel ULPs. A proper fix just didn't appear this
cycle due to the holidays
- Missing NULL check on alloc in uverbs
- Double free in rxe error paths
- Fix a new kernel-infoleak report when forming ah_attr's without
GRH's in ucma"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/core: Don't infoleak GRH fields
RDMA/uverbs: Check for null return of kmalloc_array
Revert "RDMA/mlx5: Fix releasing unallocated memory in dereg MR flow"
RDMA/rxe: Prevent double freeing rxe_map_set()
Diffstat (limited to 'drivers/infiniband/core/uverbs_marshall.c')
| -rw-r--r-- | drivers/infiniband/core/uverbs_marshall.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c index b8d715c68ca4..11a080646916 100644 --- a/drivers/infiniband/core/uverbs_marshall.c +++ b/drivers/infiniband/core/uverbs_marshall.c @@ -66,7 +66,7 @@ void ib_copy_ah_attr_to_user(struct ib_device *device, struct rdma_ah_attr *src = ah_attr; struct rdma_ah_attr conv_ah; - memset(&dst->grh.reserved, 0, sizeof(dst->grh.reserved)); + memset(&dst->grh, 0, sizeof(dst->grh)); if ((ah_attr->type == RDMA_AH_ATTR_TYPE_OPA) && (rdma_ah_get_dlid(ah_attr) > be16_to_cpu(IB_LID_PERMISSIVE)) && |