On 04.08.2008 19:56, Peter Stuge wrote:
On Mon, Aug 04, 2008 at 07:50:32PM +0200, Carl-Daniel Hailfinger wrote:
a patch to apply on top of Ron's commit.
If I'm successful, I'll be undoing 50% of Ron's patch.
That doesn't matter. This is early work in progress. Key word: progress
If something has to be rolled back, it's the opposite of progress.
- pci_write_config32(a, b, c);
- pci_cf8_conf1.write32(NULL, 0, a, b, c);
Can't we just use a macro for that stuff? #define PCI_WRITE_CONFIG32_EARLY(a,b,c) pci_cf8_conf1.write32(NULL, 0, a, b, c);
Sorry, I don't see the point. I expect pci_cf8_* to change at least once anyway. I think time is better spent elsewhere.
That's why I don't want it committed in the first place. The least intrusive change would be the following #define at the beginning of each affected C file:
#define pci_write_config32(a,b,c) pci_cf8_conf1.write32(NULL, 0, a, b, c);
No other code changes needed. It reduces code churn and keeps the familiar interface. With that change, I see no reason to hold the commit back.
Regards, Carl-Daniel