Re: [PATCH v4 4/6] infiniband: cxgb4: Eliminate duplicate barriers on weakly-ordered archs

From: Jason Gunthorpe
Date: Thu Mar 22 2018 - 17:25:24 EST


On Thu, Mar 22, 2018 at 08:45:11PM +0000, Casey Leedom wrote:

> I'm guessing~ that this line in the documentation ~may~ imply the GCC
> ordering:
>
> ... Note that relaxed accesses to
> the same peripheral are guaranteed to be ordered with respect to each
> other. ...

An arch can't guarentee "ordered with respect to each other" without
preventing the compiler from re-ordering, so yes, any correct
implementation of writel_relaxed must prevent compiler re-ordering.

eg with volatile or a compiler barrier, or whatever.

> In any case, we really only have a few places where we (the various Chelsio
> drivers) need to worry about this: the "Fast Paths" where we have a lot of
> I/O to the device. I think we should leave everything else alone.

Yes.

Jason