I believe that you need to update sys/modules/ixl/Makefile to document that the driver now depends on pci_iov_if.h as well

Please update the PCI_INIT_IOV(9), PCI_ADD_VF(9) and PCI_UNINIT_IOV(9) manpages to document the new header file and reflect the new method names.

I am having a horrific time trying to get HEAD installed on one of these systems. Unfortunately support for i386-only systems in our test lab is rotting and I haven't had time to debug the issue yet. :(

This is an Intel Core processor? I believe that I have access to one of these. Any idea what the symptoms would be?

I think that it would be sufficient to test that this still allows ppt to attach to a device dynamically:

Closed by commit rS280957 (authored by @rstone).

this was just a test run of my arc wrapper script

Closed by commit rS279939 (authored by @rstone).

Closed by commit rS279868 (authored by @rstone).

committed as r279836

committed as r279835

committed as r279834

committed as r279833

committed as r279832

committed as r279831

committed as r279830

Yes, it works now. Also, independent of this (but with the vmm.ko fix that already went in), devctl does work to force the ppt driver to attach to a device:

generated by test of script; ignore

generated by test of script; ignore

Closed by commit rS279829 (authored by @rstone).

Fix typo in event number

Thanks for doing this.

Add missing blank line and give proper credit.

RLOCK only gets a read lock. You want WLOCK to get a write lock to ensure serialization.

Closed by commit rS279439 (authored by @rstone).

Closed by commit rS279438 (authored by @rstone).

Closed by commit rS279436 (authored by @rstone).

Closed by commit rS279435 (authored by @rstone).

Closed by commit rS279425 (authored by @rstone).

Closed by commit rS279410 (authored by @rstone).

Err, that was supposed to be "installed it into a couple of *VMs*"

Ok, I did a full buildworld/buildkernel (with device zfs in my kernel config) and installed it into a couple of zfs. I tested creating a zpool out of a mfs disk, creating filesystems, setting some properties, doing a git clone into a file system, and then doing a zfs send of the result from one vm to the problem. No problems cropped up.

Fix the test case to actually test the clone

The primary testing that I did was to ensure that a kernel with "device zfs" would still link. However, I suppose that doesn't exclude the possibility of their being an zfs source file that is built wrong and doesn't include this header. Is there a zfs test suite that I could run?

Already fixed in r277925