Re: [PATCH rdma-next v7 6/9] RDMA/mana_ib: Post send and memory-management WRs on RC QPs

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] RDMA/mana_ib: __le32 to u32 Sparse type mismatch in mana_ib_post_sq()
--

commit 8ccaec72f638cf8641d6fe111aeda1bd2485a4f2
Author: Konstantin Taranov <kotaranov@microsoft.com>

RDMA/mana_ib: Post send and memory-management WRs on RC QPs

This commit implements send work request posting for RC QPs, routing
communication WRs to the send queue and memory management WRs to the memory
management queue. It also advertises IB_DEVICE_MEM_MGT_EXTENSIONS on
adapters supporting extended WQEs.

> diff --git a/drivers/infiniband/hw/mana/wr.c b/drivers/infiniband/hw/mana/wr.c
> index 5d42af91fce4a..d5dec02aa70a7 100644
> --- a/drivers/infiniband/hw/mana/wr.c
> +++ b/drivers/infiniband/hw/mana/wr.c

[ ... ]

> @@ -224,6 +289,294 @@ static int mana_ib_post_send_ud(struct mana_ib_qp *qp, const struct ib_send_wr *
>  	return err;
>  }
>  
> +static int mana_ib_post_sq(struct mana_ib_qp *qp, const struct ib_send_wr *wr)
> +{

[ ... ]

> +	switch (wr->opcode) {
> +	case IB_WR_SEND_WITH_INV:
> +		send_oob.rc_send.invalidate_key = wr->ex.invalidate_rkey;
> +		wqe_req.flags |= GDMA_WR_CHECK_SN;
> +		break;
> +	case IB_WR_SEND_WITH_IMM:
> +		send_oob.rc_send.immediate = cpu_to_le32(be32_to_cpu(wr->ex.imm_data));
> +		fallthrough;
> +	case IB_WR_SEND:
> +		wqe_req.flags |= GDMA_WR_CHECK_SN;
> +		break;
> +	case IB_WR_RDMA_WRITE_WITH_IMM:
> +		send_oob.immediate_ext = cpu_to_le32(be32_to_cpu(wr->ex.imm_data));

[Severity: Low]
In mana_ib_post_sq(), both IB_WR_SEND_WITH_IMM and IB_WR_RDMA_WRITE_WITH_IMM
assign the restricted __le32 value returned by cpu_to_le32() to
send_oob.rc_send.immediate and send_oob.immediate_ext, which are declared as
plain u32 in struct rdma_send_oob in drivers/infiniband/hw/mana/mana_ib.h:

	union {
		u32 immediate;
		u32 invalidate_key;
	} rc_send;
	...
	union {
		u32 immediate_ext;

When building with C=1, does this trigger Sparse endianness type-mismatch
warnings? Should those fields be annotated as __le32, or assigned using
be32_to_cpu(wr->ex.imm_data) to match the host-endian u32 fields in
struct rdma_send_oob?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261102162505.501853-1-kotaranov@linux.microsoft.com?part=6