From: Jakub Kicinski <jakub.kicin...@netronome.com>
Date: Tue, 27 Nov 2018 14:04:12 -0800
> From: John Hurley <john.hur...@netronome.com>
>
> For flow offload adds, if the rhash insert code fails, the flow will still
> have been offloaded but the reference to it in the driver freed.
>
> Re-order the offload setup calls to ensure that a flow will only be written
> to FW if a kernel reference is held and stored in the rhashtable. Remove
> this hashtable entry if the offload fails.
>
> Fixes: c01d0efa5136 ("nfp: flower: use rhashtable for flow caching")
> Signed-off-by: John Hurley <john.hur...@netronome.com>
> Reviewed-by: Pieter Jansen van Vuuren <pieter.jansenvanvuu...@netronome.com>
> Reviewed-by: Jakub Kicinski <jakub.kicin...@netronome.com>
> ---
> Merge note: there will be a slight merge conflict with net-next
> here:
> - the first argument of nfp_flower_xmit_flow() changed from 'netdev'
> to 'app';
> - we only bump the port offload cnt if (port).
>
> FWIW the net-next version of the patch can be found at:
>
> git://git.kernel.org/pub/scm/linux/kernel/git/kuba/linux.git
>
> 17ed95873d51 nfp: flower: prevent offload if rhashtable insert fails
> d857fc8f472b nfp: flower: release metadata on offload failure
>
> CC: Stephen Rothwell <s...@canb.auug.org.au> # for linux-next
Applied.
