Message ID | 20190318132340.16892-4-shiraz.saleem@intel.com (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | Use correct sizing on buffers holding page DMA addresses | expand |
diff --git a/drivers/infiniband/hw/mthca/mthca_provider.c b/drivers/infiniband/hw/mthca/mthca_provider.c index d063d7a..e32902e 100644 --- a/drivers/infiniband/hw/mthca/mthca_provider.c +++ b/drivers/infiniband/hw/mthca/mthca_provider.c @@ -914,7 +914,7 @@ static struct ib_mr *mthca_reg_user_mr(struct ib_pd *pd, u64 start, u64 length, goto err; } - n = mr->umem->nmap; + n = mr->umem->npages; mr->mtt = mthca_alloc_mtt(dev, n); if (IS_ERR(mr->mtt)) {
The buffer that holds the page DMA addresses is sized off umem->nmap. This can potentially cause out of bound accesses on the PBL array when iterating the umem DMA-mapped SGL. This is because if umem pages are combined, umem->nmap can be much lower than the number of system pages in umem. Use umem->npages to size this buffer. Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com> --- drivers/infiniband/hw/mthca/mthca_provider.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)