Messages in this thread | | | Date | Wed, 15 Mar 2023 16:57:29 +0100 | From | Simon Horman <> | Subject | Re: [PATCH net-next v3 1/9] net: sunhme: Just restart autonegotiation if we can't bring the link up |
| |
On Wed, Mar 15, 2023 at 11:35:19AM -0400, Sean Anderson wrote: > On 3/15/23 04:20, Simon Horman wrote: > > On Mon, Mar 13, 2023 at 08:36:05PM -0400, Sean Anderson wrote: > > > If we've tried regular autonegotiation and forcing the link mode, just > > > restart autonegotiation instead of reinitializing the whole NIC. > > > > > > Signed-off-by: Sean Anderson <seanga2@gmail.com> > > > > Hi Sean, > > > > This patch looks fine to me, as do patches 3 - 4, which is as far as I have > > got with my review. > > > > I do, however, have a general question regarding most of the patches in this > > series: to what extent have they been tested on HW? > > I have tested them with some PCI cards, mostly with the other end > autonegotiating 100M. This series doesn't really touch the phy state > machines, so I think it is fine to just make sure the link comes up (and > things work after bringing the interface down and up).
Understood, I'll proceed with my review on that basis.
> > And my follow-up question is: to what extent should we consider removing > > support for hardware that isn't being tested and therefore has/will likely > > have become broken break at some point? Quattro, the subject of a latter > > patch in this series, seems to be a case in point. > > Well, I ordered a quattro card (this hardware is quite cheap on ebay) so > hopefully I can test that.
Yes, for some reason I was searching on eBay too :)
> The real question is whether there's anyone using this on sparc. I tried > CCing some sparc users mailing lists in the cover letter, but no luck so > far.
Yes, that is the question.
| |