On Tue, 13 Mar 2018 16:01:23 -0600 Jason Gunthorpe <jgg@xxxxxxxxxxxx> wrote:
> > diff -puN drivers/infiniband/core/verbs.c~drivers-infiniband-core-verbsc-fix-build-with-gcc-444 drivers/infiniband/core/verbs.c
> > +++ a/drivers/infiniband/core/verbs.c
> > @@ -2200,8 +2200,9 @@ static void __ib_drain_sq(struct ib_qp *
> > struct ib_send_wr *bad_swr;
> > struct ib_rdma_wr swr = {
> > .wr = {
> > + .next = NULL,
> > + { .wr_cqe = &sdrain.cqe, },
> > .opcode = IB_WR_RDMA_WRITE,
> > - .wr_cqe = &sdrain.cqe,
> > },
> > };
> > int ret;
> > _
> >
>
> There were two patches in the series that introduced this, and the
> same construct was also added in 2a78cb4db487 ("IB/srpt: Fix an
> out-of-bounds stack access in srpt_zerolength_write()")
>
> We should probably fix both??
yup. Sent.
> Should this go through the rdma -rc tree?
Via whichever tree contained the Fixes: commits, I guess.