Re: [net PATCH] octeontx2-af: Enable hardware timestamping for VFs

[Jacob Keller]

On 9/29/2023 12:16 AM, Sai Krishna wrote:
> Currently for VFs, mailbox returns error when hardware timestamping enable
> is requested. This patch fixes this issue.
> 

The subject title implies that this is implementing a new feature (and
thus not really a good candidate for net), but the content implies this
is a fix for an existing feature thats not working properly.

It could use a slightly improved commit message.

> Fixes: 421572175ba5 ("octeontx2-af: Support to enable/disable HW timestamping")
> Signed-off-by: Subbaraya Sundeep <sbhatta@xxxxxxxxxxx>
> Signed-off-by: Sunil Kovvuri Goutham <sgoutham@xxxxxxxxxxx>
> Signed-off-by: Sai Krishna <saikrishnag@xxxxxxxxxxx>

Typically the author signed-off-by should go first, and then other
signed-off-by are for people in the chain of patch delivery. If the
other two co-authored the patch they should have Co-developed-by tag or
something. Otherwise I would expect that Subbaraya Sundeepd would be the
patch author since that signed-off-by is first.

> ---
>  drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 

The fix itself seems straight forward, though the comment is confusing
to me. You also changed the return from -ENODEV to -EPERM which makes
sense enough I suppose, but thats not called out in the change.

> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> index f2b1edf1bb43..aba0c530160c 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> @@ -756,12 +756,11 @@ static int rvu_cgx_ptp_rx_cfg(struct rvu *rvu, u16 pcifunc, bool enable)
>  	if (!is_mac_feature_supported(rvu, pf, RVU_LMAC_FEAT_PTP))
>  		return 0;
>  
> -	/* This msg is expected only from PFs that are mapped to CGX LMACs,
> +	/* This msg is expected only from PFs that are mapped to CGX/RPM LMACs,
>  	 * if received from other PF/VF simply ACK, nothing to do.
>  	 */

This comment implies to me that we wouldn't expect this message from a VF?

> -	if ((pcifunc & RVU_PFVF_FUNC_MASK) ||
> -	    !is_pf_cgxmapped(rvu, pf))
> -		return -ENODEV;
> +	if (!is_pf_cgxmapped(rvu, rvu_get_pf(pcifunc)))
> +		return -EPERM;
>  
>  	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
>  	cgxd = rvu_cgx_pdata(cgx_id, rvu);