[dpdk-dev] [PATCH v5 1/2] ethdev: add supported hash function check

[Thomas Monjalon]

01/05/2018 13:04, Ferruh Yigit:
> On 4/23/2018 7:14 PM, Thomas Monjalon wrote:
> > 23/04/2018 18:06, Ferruh Yigit:
> >> On 4/20/2018 3:30 PM, Xueming Li wrote:
> >>> Add supported RSS hash function check in device configuration to
> >>> have better error verbosity for application developers.
> >>>
> >>> Signed-off-by: Xueming Li <xuemingl at mellanox.com>
> >>> Acked-by: Adrien Mazarguil <adrien.mazarguil at 6wind.com>
> >>>  
> >>> +	/* Check that device supports requested rss hash functions. */
> >>> +	if ((dev_info.flow_type_rss_offloads |
> >>> +	     dev_conf->rx_adv_conf.rss_conf.rss_hf) !=
> >>> +	    dev_info.flow_type_rss_offloads) {
> >>> +		RTE_PMD_DEBUG_TRACE("ethdev port_id=%d invalid rss_hf: "
> >>> +				    "0x%"PRIx64", valid value: 0x%"PRIx64"\n",
> >>> +				    port_id,
> >>> +				    dev_conf->rx_adv_conf.rss_conf.rss_hf,
> >>> +				    dev_info.flow_type_rss_offloads);
> >>> +		return -EINVAL;
> >>> +	}
> >>
> >> Hi Thomas,
> >>
> >> This can break the PMDs that are not setting flow_type_rss_offloads properly.
> >> How can we highlight this so that PMD owners can double check?
> > 
> > Can we have a check-list in the RC1 announce email?
> 
> Hi Thomas, Xueming,
> 
> This change is breaking multiple sample applications, testpmd was also broken
> but already fixed by Qi [1].
> 
> Indeed this patch should update sample applications and testpmd as well when
> doing an ethdev API update, also should update release notes "API Changes" section.
> 
> We can fix sample applications for rc2, but same thing also can hit users.
> 
> Or for this release we can remote returning error, instead update log message to
> error. Next release add the return and change log message back to debug.
> 
> What do you think?

Yes:
1/ update the API doc and sample apps in 18.05
2/ send a deprecation notice
3/ add error return in 18.08

I've  replied to your patch too:
	http://dpdk.org/ml/archives/dev/2018-May/099865.html