Re: x86 pcitag_t change proposal/patch

[Matt Thomas <matt%3am-software.com@localhost>]

On Dec 4, 2012, at 6:59 PM, Jonathan A. Kollasch wrote:

> 
> I'd like feedback on the attached patch.
> 
> Commit message would be approximately as follows:
> 
> Separate the implementation of PCI Configuration Access Mode 1 and
> Mode 2 from the x86 pcitag_t.  This is desirable so that any tags
> possibly created prior to parsing of the ACPI MCFG table(s) can be
> used with a memory-mapped configuration space mechanism afterwards.
> 
>       Jonathan Kollasch
> <pcitag-diff.txt>

Never ever use bitfields when talking to hardware.  They should only be used as 
a software abstraction.