Re: [PATCH v3 2/2] virt-aa-helper: Allow SR-IOV VF PCI for hostdev networks
On 6/25/25 18:49, Tim Small wrote:
On 09/06/2025 13:51, Michal Prívozník wrote:
But this new IPC got me thinking - the XML we're parsing here in the aa helper is formatted by libvirtd who already know the actual type of each net. But looking into the code - this is lost when the daemon talks to the helper. I'm not on an AppArmor machine right now, so can you please check whether the following change (applied on the top of this very patch) works? It would save us from those IPCs.
Thanks Michal. The additional patches which you posted (below) did not work in isolation, but with an additional change to the src/qemu/ qemu_process.c to ensure that the network is resolved before the apparmor helper is called, it works.
I've updated my branch and tested, and will post the updated patch to the mailing list, but before I do-so...
What would you like me to do with regard to the changes with you proposed?
. Post with credit in the patch commentary. . Post with credit in the commit message.
If you want, then you can put Suggested-by: Michal Privoznik <mprivozn@redhat.com> line just above your SoB line. Thanks!
. Something else?
Thanks for the tip!
No problem. Introducing new RPC call always sets my alarm off :-D Michal
participants (1)
-
Michal Prívozník