| Message ID | 20200519120256.GC42765@mwanda (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | RDMA/rtrs: Fix some signedness bugs in error handling | expand |
On Tue, May 19, 2020 at 2:05 PM Dan Carpenter <dan.carpenter@oracle.com> wrote: > > The problem is that "req->sg_cnt" is an unsigned int so if "nr" is > negative, it gets type promoted to a high positive value and the > condition is false. This patch fixes it by handling negatives separately. > > Fixes: 6a98d71daea1 ("RDMA/rtrs: client: main functionality") > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Thanks Dan, fix looks correct, just one comment inline. > --- > drivers/infiniband/ulp/rtrs/rtrs-clt.c | 7 +++---- > drivers/infiniband/ulp/rtrs/rtrs-srv.c | 2 +- > 2 files changed, 4 insertions(+), 5 deletions(-) > > diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c > index 468fdd0d8713..17f99f0962d0 100644 > --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c > +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c > @@ -1047,11 +1047,10 @@ static int rtrs_map_sg_fr(struct rtrs_clt_io_req *req, size_t count) > > /* Align the MR to a 4K page size to match the block virt boundary */ > nr = ib_map_mr_sg(req->mr, req->sglist, count, NULL, SZ_4K); > - if (unlikely(nr < req->sg_cnt)) { > - if (nr < 0) > - return nr; > + if (nr < 0) > + return -EINVAL; Why not just return nr here? > + if (unlikely(nr < req->sg_cnt)) > return -EINVAL; > - } > ib_update_fast_reg_key(req->mr, ib_inc_rkey(req->mr->rkey)); > > return nr; > diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv.c b/drivers/infiniband/ulp/rtrs/rtrs-srv.c > index ba8ab33b94a2..eefd149ce7a4 100644 > --- a/drivers/infiniband/ulp/rtrs/rtrs-srv.c > +++ b/drivers/infiniband/ulp/rtrs/rtrs-srv.c > @@ -649,7 +649,7 @@ static int map_cont_bufs(struct rtrs_srv_sess *sess) > } > nr = ib_map_mr_sg(mr, sgt->sgl, sgt->nents, > NULL, max_chunk_size); > - if (nr < sgt->nents) { > + if (nr < 0 || nr < sgt->nents) { > err = nr < 0 ? nr : -EINVAL; > goto dereg_mr; > } > -- > 2.26.2 >
On Tue, May 19, 2020 at 02:40:41PM +0200, Jinpu Wang wrote: > On Tue, May 19, 2020 at 2:05 PM Dan Carpenter <dan.carpenter@oracle.com> wrote: > > > > The problem is that "req->sg_cnt" is an unsigned int so if "nr" is > > negative, it gets type promoted to a high positive value and the > > condition is false. This patch fixes it by handling negatives separately. > > > > Fixes: 6a98d71daea1 ("RDMA/rtrs: client: main functionality") > > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> > Thanks Dan, fix looks correct, just one comment inline. > > --- > > drivers/infiniband/ulp/rtrs/rtrs-clt.c | 7 +++---- > > drivers/infiniband/ulp/rtrs/rtrs-srv.c | 2 +- > > 2 files changed, 4 insertions(+), 5 deletions(-) > > > > diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c > > index 468fdd0d8713..17f99f0962d0 100644 > > --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c > > +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c > > @@ -1047,11 +1047,10 @@ static int rtrs_map_sg_fr(struct rtrs_clt_io_req *req, size_t count) > > > > /* Align the MR to a 4K page size to match the block virt boundary */ > > nr = ib_map_mr_sg(req->mr, req->sglist, count, NULL, SZ_4K); > > - if (unlikely(nr < req->sg_cnt)) { > > - if (nr < 0) > > - return nr; > > + if (nr < 0) > > + return -EINVAL; > Why not just return nr here? Sorry. I thought I did but I made a typo. Let me resend. regards, dan carpenter
diff --git a/drivers/infiniband/ulp/rtrs/rtrs-clt.c b/drivers/infiniband/ulp/rtrs/rtrs-clt.c index 468fdd0d8713..17f99f0962d0 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-clt.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-clt.c @@ -1047,11 +1047,10 @@ static int rtrs_map_sg_fr(struct rtrs_clt_io_req *req, size_t count) /* Align the MR to a 4K page size to match the block virt boundary */ nr = ib_map_mr_sg(req->mr, req->sglist, count, NULL, SZ_4K); - if (unlikely(nr < req->sg_cnt)) { - if (nr < 0) - return nr; + if (nr < 0) + return -EINVAL; + if (unlikely(nr < req->sg_cnt)) return -EINVAL; - } ib_update_fast_reg_key(req->mr, ib_inc_rkey(req->mr->rkey)); return nr; diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv.c b/drivers/infiniband/ulp/rtrs/rtrs-srv.c index ba8ab33b94a2..eefd149ce7a4 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-srv.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-srv.c @@ -649,7 +649,7 @@ static int map_cont_bufs(struct rtrs_srv_sess *sess) } nr = ib_map_mr_sg(mr, sgt->sgl, sgt->nents, NULL, max_chunk_size); - if (nr < sgt->nents) { + if (nr < 0 || nr < sgt->nents) { err = nr < 0 ? nr : -EINVAL; goto dereg_mr; }
The problem is that "req->sg_cnt" is an unsigned int so if "nr" is negative, it gets type promoted to a high positive value and the condition is false. This patch fixes it by handling negatives separately. Fixes: 6a98d71daea1 ("RDMA/rtrs: client: main functionality") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> --- drivers/infiniband/ulp/rtrs/rtrs-clt.c | 7 +++---- drivers/infiniband/ulp/rtrs/rtrs-srv.c | 2 +- 2 files changed, 4 insertions(+), 5 deletions(-)