On Fri, Jun 27, 2008 at 08:36:38PM +0200, Carl-Daniel Hailfinger wrote:
And I was convinced that it is ok in case it never worked correctly.
It did work correctly, but the "multiple flash chip support" patch made it a lot less useful.
It's not less useful, but it has way more false positives.
So what do you suggest is the right fix? Dummy entries is not it.
Reverting multiple flash chip support would be one of the possible solutions.
I disagree.
The preferred way to fix this would be a per-bus probe list, decoding the vendor ID if the device ID is unknown.
I agree this is something we will have in flashrom, but it will be a big change, long after 1.0.
It will also not neccessarily fix all false positives.
By the way, that would also be the only correct solution for multiple flash chip recognition.
Not neccessarily. (e.g. ID strapping on FWH)
Short term I prefer removing the catchall entries. They will be replaced with code to do the same thing, but in a different way, before 1.0, so rather soon.
The only OK kludge is to not match the catchall entries for anything but the first flash chip, but we might as well just remove them now and not type more text in email but type code instead. :)
//Peter