Message ID | 1665165447-1802-1-git-send-email-jdamato@fastly.com |
---|---|
Headers | show |
Series | i40e: Add a non-XDP i40e_napi_poll tracepoint | expand |
On 10/7/2022 10:57 AM, Joe Damato wrote: > Greetings: > > This is an RFC which is similar to the series up for review, except that > this implementation does not touch XDP at all and adds a conditional in > i40e_napi_poll to only fire the tracepoint when XDP is not enabled. > > This should avoid the issues that Maciej has with the naming of out > parameters (since none of that code is touched in this series) and it > clears the way for Maciej, Sridhar, et al to implement the XDP tracepoint. > > I am submitting this an alternative to what's already up for review. > > If you prefer to accept this code, please let me know that you want the > non-XDP version and I'll submit it as the 'v4'. Given the discussion, this is the series I prefer. I'm very happy to see some more debugging helpers coming into the driver so thanks for your work on this Joe! As for the rest of the team they seem to be fine speaking for themselves, so I imagine they'll let you know :-) For the series: Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com> Best regards, Jesse
On Fri, Oct 07, 2022 at 01:36:20PM -0700, Jesse Brandeburg wrote: > On 10/7/2022 10:57 AM, Joe Damato wrote: > >Greetings: > > > >This is an RFC which is similar to the series up for review, except that > >this implementation does not touch XDP at all and adds a conditional in > >i40e_napi_poll to only fire the tracepoint when XDP is not enabled. > > > >This should avoid the issues that Maciej has with the naming of out > >parameters (since none of that code is touched in this series) and it > >clears the way for Maciej, Sridhar, et al to implement the XDP tracepoint. > > > >I am submitting this an alternative to what's already up for review. > > > >If you prefer to accept this code, please let me know that you want the > >non-XDP version and I'll submit it as the 'v4'. > > Given the discussion, this is the series I prefer. I'm very happy to see > some more debugging helpers coming into the driver so thanks for your work > on this Joe! As for the rest of the team they seem to be fine speaking for > themselves, so I imagine they'll let you know :-) > > For the series: > Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com> OK, thanks! There's a minor build failure for a format string (lu should be u) in the tracepoint. I'll fix that now in this series and re-send it as a v4 with a proper cover letter. Thanks again for your detailed feedback and review; I appreciate your time and energy on this. Thanks, Joe
On 10/7/2022 3:55 PM, Joe Damato wrote: > On Fri, Oct 07, 2022 at 01:36:20PM -0700, Jesse Brandeburg wrote: >> On 10/7/2022 10:57 AM, Joe Damato wrote: >>> Greetings: >>> >>> This is an RFC which is similar to the series up for review, except that >>> this implementation does not touch XDP at all and adds a conditional in >>> i40e_napi_poll to only fire the tracepoint when XDP is not enabled. >>> >>> This should avoid the issues that Maciej has with the naming of out >>> parameters (since none of that code is touched in this series) and it >>> clears the way for Maciej, Sridhar, et al to implement the XDP tracepoint. >>> >>> I am submitting this an alternative to what's already up for review. >>> >>> If you prefer to accept this code, please let me know that you want the >>> non-XDP version and I'll submit it as the 'v4'. >> Given the discussion, this is the series I prefer. I'm very happy to see >> some more debugging helpers coming into the driver so thanks for your work >> on this Joe! As for the rest of the team they seem to be fine speaking for >> themselves, so I imagine they'll let you know :-) >> >> For the series: >> Acked-by: Jesse Brandeburg <jesse.brandeburg@intel.com> > OK, thanks! > > There's a minor build failure for a format string (lu should be u) in the > tracepoint. > > I'll fix that now in this series and re-send it as a v4 with a proper > cover letter. > > Thanks again for your detailed feedback and review; I appreciate your time > and energy on this. > > Thanks, > Joe Sorry for all the back and forth on this series. The tracepoint itself is definitely useful and this series looks good. Acked-by: Sridhar Samudrala <sridhar.samudrala@intel.com>