Once the order is the same on all architectures, we could change it to match Linux at any point. I think its important to stop holding up architectural unification in the name of the perfect solution.
The Coccinelle to switch the ordering is much safer to do after you have the ordering consistent and both arguments to the write functions are types which are incompatible with each other.
Personally I am in the camp of having the same order as Linux, but there is a lot of ARM work held up because they are *different*. This is the problem I am trying to solve. Lets make the best community decision regardless of our personal opinions.
Kevin
On Feb 18, 2015, at 8:40 AM, Aaron Durbin via coreboot coreboot@coreboot.org wrote:
On Wed, Feb 18, 2015 at 10:35 AM, Patrick Georgi <pgeorgi@google.com mailto:pgeorgi@google.com> wrote:
2015-02-18 17:23 GMT+01:00 Vadim Bendebury vbendeb@chromium.org:
kernel, u-boot, etc. They all have the write(val, addr) semantics. I see no good reason to artificially erect an ever present barrier for integrating code into a coreboot system.
Since all imported code requires some rework before it works for us, and redoing that particular issue is a matter of a simple spatch (http://coccinelle.lip6.fr/), I'm not sure that's actually that big a concern.
@@ expression a,v; @@ -writel(v,a) +write32(a, v)
Can you provide the spatch utility and script as well as the documentation so that every potential contributor can perform the necessary steps? Or we could be done w/ it once and for all and fix the existing coreboot semantics. Then there's not another step for anyone in perpetuity. Also keep in mind it's not just software going into coreboot but reusing software within coreboot as well. It's a 2 way street.
I did volunteer to do the big change, and if people want I will. Let me know.
-Aaron
-- coreboot mailing list: coreboot@coreboot.org mailto:coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot http://www.coreboot.org/mailman/listinfo/coreboot
-
Kevin Herbert