On Fri, 13 Oct 2006, Matthew Wilcox wrote: > On Fri, Oct 13, 2006 at 01:57:48PM -0400, Alan Stern wrote: > > Would it be okay for pci_block_user_cfg_access() to use its own cache, so > > it doesn't interfere with data previously cached by pci_save_state()? > > My suggestion is just to require that the callers have previously called > pci_save_state(). The PCI PM stack already has, and it's a one-line > change to the IPR driver. Okay. Would you like to write a patch with that fix? Be sure to add a comment explaining the need for a previous call to pci_save_state(). At least it will get things going for now, even if it isn't perfectly correct in the long run. Alan Stern
