Messages in this thread | | | Date | Sun, 23 Dec 2007 00:09:49 -0500 | From | Jeff Garzik <> | Subject | Re: [patch] Make MMCONFIG space (extended PCI config space) a driver opt-in issue |
| |
Linus Torvalds wrote: > > On Sat, 22 Dec 2007, Jeff Garzik wrote: >> But regardless of problems, enabling should be done globally, not per >> device... > > I'm ok with trying the "globally" idea, but it has to be "globally but > only if absolutely required". > > And quite frankly, how do you tell whether it's absolutely required or > not? > > I have an idea: the drivers that really need it will do a "please enable > MMCONFIG, because I will need it" thing? > > Ok? > > And then, since we *need* such a "pci_enable_mmconfig()" call anyway, why > not let the driver give which device it controls too, so that we can print > out the information (in case the machine then hangs immediately > afterwards), and perhaps - if it is shown to help - only do the MMCONFIG > cycles for that particular device? > > Sounds like a plan?
As long as pci_enable_ext_cfg_space(pdev) enables extended accesses for -all- devices, the plan is mostly sound.
That largely eliminates the inconsistency issue.
The only thing I would worry about is whether "config space suddenly grew larger" condition will confuse userspace -- but that is NOT an objection, just a worry.
Jeff
| |