Re: RFC: New userspace fetch/store API

[Jason Thorpe <thorpej%me.com@localhost>]

> On Feb 23, 2019, at 6:35 PM, Robert Elz <kre%munnari.OZ.AU@localhost> wrote:
> 
> fuword/fuiword/suword/suiword (etc) are all ancient, and I would
> have expected, largely deprecated these days - but what's wrong
> with copyin/copyout ?    Those could be optimised for particular
> data sizes if there was some real need for that.

Hey Robert —

The copyin / copyout API contract is like bcopy(); arbitrary byte alignment and no guarantees that the load / store to the specified memory location will be of any specific size, or will even be a single load or store operation.  I thought there would value in having something like the old fetch / store API, which did provide those sorts of API guarantees, just freshened up a little.

Furthermore, there isn’t a provision for using copyin / copyout in interrupt context (e.g. fuswintr()) … Now, it may be that is’s time to abandon that… but I’ll stick with the above argument.

-- thorpej