https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=197076
--- Comment #11 from John Baldwin <j...@freebsd.org> --- Sorry I missed replying to this before. The patch referenced in comment 2 and 3 did indeed fix the original bug (in the updated dmesg an re1 device is now probed), but I think I know why it still didn't work. I think now if you were able to force a re-attach post-boot using devctl(8) then re1 would work, but the current bug is that we allocate resources for the parent PCI-PCI bridge of the re1 device that conflict with resources already assigned to another bridge, so trying to read the registers from re1 doesn't work. I think I know how to fix that in that the PCI bus driver should reserve resources initial I/O window regions when enumerating the bus and not just initial BARs. -- You are receiving this mail because: You are the assignee for the bug.
