cvs commit: src/sys/dev/pci pci.c

John E Hein jhein at timing.com
Mon Feb 4 08:02:57 PST 2008


John Baldwin wrote at 09:44 -0500 on Feb  4, 2008:
 > We can blacklist those bridges just like we blacklist the AMD HT8131 bridge
 > which is broken.

That's okay with me.  We might use a PCIe switch to split an x16 lane
(used normally for PEG) into two x8 lanes (used for our own nefarious
purposes including MSI for sure).  These should all pass
PCIe-deviceness through the switch to a root complex (in all the
targets we've looked at using), however, and so should be unaffected
by your new PCIe detection, regardless of blacklisting a PCI or PCI-X
to PCIe bridge.


More information about the cvs-all mailing list