On 09.02.2019 20:50, Heiner Kallweit wrote: > On 09.02.2019 20:42, Andrew Lunn wrote: >>> I know, it's patch 15 in your series ;) And I'm aware that usually new >>> core functionality is acceptable only if it comes together with a user, >>> to avoid having billions of orphaned good ideas in the code. >>> I focused on the core here to not get lost in all the new stuff, and to >>> provide Maxime with some direction for adjusting his series. >> >> I'm just trying to avoid Maxime reimplementing something when we >> already have a patch: >> > Sure, I didn't mean Maxime should re-implement things we have in the pipe. > I meant it more in a way that he gets an idea in which direction we're moving. > >> https://github.com/lunn/linux/commit/07e3fa8f183f05a09d969a9378534da35238eeb9 >> >> Maxime, feel free to cherry-pick this into your series. >> > I'll submit this one too. > Just saw that this patch depends on patch 5 of Maxime's series. And it needs a small change because the generic function was renamed to genphy_c45_pma_read_abilities(). So indeed it may be better if Maxime adds this patch to his series.
One comment regarding the pause bits in Maxime's patch 5: If no pause bit is set the core automatically sets both. So we have to do this neither in the marvell10g driver nor in the generic read-abilities function. >> Andrew >> Heiner > Heiner > > Oh, and I just saw: When we talk about/with somebody, we should add him to the > mail addressee list. >