Re: [PATCH 10/14] SIWv2: Transmit path: siw_qp_tx.c

On Thu, Jun 16, 2011 at 2:42 PM, Bernard Metzler [off-list ref] wrote:

[ ... ]

+/*
+ * Write out iov referencing hdr, data and trailer of current FPDU.
+ * Update transmit state dependent on write return status
+ */
+static int siw_tx_hdt(struct siw_iwarp_tx *c_tx, struct socket *s)
+{
+       struct siw_wqe          *wqe = c_tx->wqe;
+       struct siw_sge          *sge = &wqe->wr.sgl.sge[c_tx->sge_idx],
+                               *first_sge = sge;
+       struct siw_mr           *mr = NULL;
+       struct ib_umem_chunk    *chunk = c_tx->umem_chunk;
+
+       struct kvec             iov[MAX_ARRAY];
+       struct page             *page_array[MAX_ARRAY];
+       struct msghdr           msg = {.msg_flags = MSG_DONTWAIT};
+
+       int                     seg = 0, do_crc = c_tx->do_crc, is_kva = 0, rv;
+       unsigned int            data_len = c_tx->bytes_unsent,
+                               hdr_len = 0,
+                               trl_len = 0,
+                               sge_off = c_tx->sge_off,
+                               sge_idx = c_tx->sge_idx,
+                               pg_idx = c_tx->pg_idx;

Have you run the siw source code through sparse ? Sparse reports the
following for the above:

drivers/infiniband/hw/siw/siw_qp_tx.c:694:1: warning: the frame size
of 1120 bytes is larger than 1024 bytes

I assume that you know that one should be careful with stack
allocations in the kernel ?

+               if (tx_type == RDMAP_RDMA_READ_REQ) {

This is what the compiler says about the above statement:

drivers/infiniband/hw/siw/siw_qp_tx.c:1145:15: warning: comparison
between enum siw_wr_opcode and enum rdma_opcode

So how can that statement be correct ?

Bart.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html