David Marchand <[email protected]> writes:
> On Tue, 20 Jan 2026 at 19:02, Bruce Richardson
> <[email protected]> wrote:
>> > void
>> > iavf_set_no_poll(struct iavf_adapter *adapter, bool link_change)
>> > {
>> > diff --git a/drivers/net/intel/iavf/rte_pmd_iavf.h
>> > b/drivers/net/intel/iavf/rte_pmd_iavf.h
>> > index dea1bd2789..d13417eb08 100644
>> > --- a/drivers/net/intel/iavf/rte_pmd_iavf.h
>> > +++ b/drivers/net/intel/iavf/rte_pmd_iavf.h
>> > @@ -20,6 +20,8 @@
>> > #include <rte_mbuf.h>
>> > #include <rte_mbuf_dyn.h>
>> >
>> > +#include "iavf.h"
>> > +
>> This include is an issue, because it's an exported i.e. installed, header
>> including a private header that will not be installed as part of DPDK. This
>> gets picked up by running chkincs from test-meson-builds.sh script. [I'm
>> surprised this wasn't caught in the CI].
>
> Something looks wrong in the robot branch itself.
> I don't see the patch in the associated branch.
> https://github.com/ovsrobot/dpdk/commits/refs/heads/series_37048/
Weird... actually - not weird. I just checked a status report from that
week:
** There was another ozlabs outage, and that caused reports to be
malformed, leading to missing results. A recheck request should
work, but I need to get back to the python rework to fix this.
I think this probably had an issue applying and due to a database
corruption (which I fixed) had weird behavior when applying patches.
So it's possible that this patch got caught up in the bot breakages that
happened around the beginning of Dec.
> Otoh, pushing this patch in my own repo does raise a chkincs failure.
> https://github.com/david-marchand/dpdk/actions/runs/21203869950/job/60995562483#step:19:5413
