Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-20 Thread Edward Cree
On 20/05/2020 15:33, Pablo Neira Ayuso wrote: > #1 Drivers calling flow_action_hw_stats_check() fall within the > second branch (check_allow_bit is set on). > > } else if (check_allow_bit && <-- HERE > > These drivers are not honoring the _DONT_CARE bit, > __flow_action_hw_stats

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-20 Thread Pablo Neira Ayuso
On Tue, May 19, 2020 at 07:26:42PM +0100, Edward Cree wrote: > On 19/05/2020 18:35, Pablo Neira Ayuso wrote: [...] > > Netfilter is a client of this flow offload API, you have to test that > > your core updates do not break any of existing clients. > > Okay, but can we distinguish between "this nee

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread David Miller
From: Pablo Neira Ayuso Date: Tue, 19 May 2020 19:19:23 +0200 > This is breaking netfilter again. This is vague and not useful feedback. Please be specific about how the change breaks netfilter. What breaks exactly, and how?

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread Edward Cree
On 19/05/2020 18:35, Pablo Neira Ayuso wrote: > Did you test your patch with netfilter? I don't think. As I mentioned in v1 (and should have repeated on v2, sorry) this is  compile tested only, as I don't have the hardware to test it.  (I have  done some testing with the not-yet-upstream sfc_ef100

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread Pablo Neira Ayuso
On Tue, May 19, 2020 at 06:23:35PM +0100, Edward Cree wrote: > On 19/05/2020 18:19, Pablo Neira Ayuso wrote: > > This is breaking netfilter again. > > Still waiting for you to explain what this "breaks".  AFAICT the > new DONT_CARE has exactly the same effect that the old DONT_CARE > did, so as lo

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread Edward Cree
On 19/05/2020 18:19, Pablo Neira Ayuso wrote: > This is breaking netfilter again. Still waiting for you to explain what this "breaks".  AFAICT the  new DONT_CARE has exactly the same effect that the old DONT_CARE  did, so as long as netfilter is using DONT_CARE rather than (say)  a hard-coded 0, i

Re: [PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread Pablo Neira Ayuso
On Tue, May 19, 2020 at 06:02:02PM +0100, Edward Cree wrote: > Make FLOW_ACTION_HW_STATS_DONT_CARE be all bits, rather than none, so that > drivers and __flow_action_hw_stats_check can use simple bitwise checks. > > Only the kernel's internal API semantics change; the TC uAPI is unaffected. This

[PATCH net-next v2] net: flow_offload: simplify hw stats check handling

2020-05-19 Thread Edward Cree
Make FLOW_ACTION_HW_STATS_DONT_CARE be all bits, rather than none, so that drivers and __flow_action_hw_stats_check can use simple bitwise checks. Only the kernel's internal API semantics change; the TC uAPI is unaffected. v2: rebased on net-next, removed RFC tags. Signed-off-by: Edward Cree -