On Tue, 17 Feb 2015 20:14:06 +0100 "Michael S. Tsirkin" <m...@redhat.com> wrote:
> On Tue, Feb 17, 2015 at 06:51:16PM +0100, Igor Mammedov wrote: > > On Tue, 17 Feb 2015 17:47:03 +0100 > > "Michael S. Tsirkin" <m...@redhat.com> wrote: > > > > > This seems to change the contents of the tables in too many ways. So of > > > it needs changing since PCI are somewhat broken by your recent patch. > > In which way is it broken? > > Try make check and look at the diffs. > They shouldn't be so big. > I want differences to be trivial, e.g. > 0->Zero would be ok. > Refactoring code generation should be separate from > changing generated code. if you run make check on whole series you will of cause se a lot of changes however if you run make check on every patch that switches current code to AML API you'll see small incremental changes, that's basically every patch with prefix: "pc: acpi-build: " > > > > > > > > So let's do this: please prepare a minimal patchset that just generates > > > PCI description dynamically. Drop everything that's not necessary, and > > > verify that make check produces trivial diffs only. > > > > > > We'll go from there. > > >
