Ok. Will do. Junling
> On Aug 2, 2020, at 4:52 PM, Samuel Thibault <[email protected]> wrote: > > Junling Ma, le dim. 02 août 2020 16:45:22 -0700, a ecrit: >> I guess you misunderstood what I mean. I mean that there seems to be no >> interest in the discussion on IRC yesterday that we introduce a new scheme. > > That's because you asked on IRC at a random time. > > That's what IRC is not good at: collecting the community opinion on > something. If you ask on IRC at a random time, you'll only get the opion > of the people who actually look at IRC at that random time. > > I (and I'm maintainer here...) do believe that the way you propose makes > sense, for the reason you mentioned (not having to implement the methods > for all device interface implementations while it's only useful for one > of them). Introducing another interface (like the i386 perm) could make > sense, but since device_read/write can offer exactly the same semantic, > there is no reason to introduce one. > >> If so, I will not work on that part. > > Please do. That part does make sense, is an improvement, and does not > require to revamp the whole thing, just the interface. Please however > also fix libddekit so we don't completely break netdde along the way > (and so that we actually have a well-tested testcase for the interface). > >> For part I, ie., moving the irq management to mach side, it people >> think it is useful, I will prepare a series of revised patch for that. > > I'm not sure what feature exactly you mean, I'll just remind the > (current) requirements ahead of time: > > - we still need the linux/ part to be working. > - we want to get rid of the whole linux/ directory at some point > > so working on moving pieces of the IRQ handling to mach does make a lot > of sense since that's what we'll want in the end. Just make sure to keep > the linux/ part working for now. > > Samuel
