In D54641#1258945, @markj wrote:

Have you tested with PCI passthru configured? With just this patch applied, I get an assertion failure in update_bar_address(). Same result if I test your github branch.

LGTM.

Address @markj 's comment - switch to INTERNALLIB

Address @markj 's comments.

Address @markj 's comments.

In D54641#1258945, @markj wrote:

Have you tested with PCI passthru configured? With just this patch applied, I get an assertion failure in update_bar_address(). Same result if I test your github branch.

Address @markj 's comments.

Address @markj 's comments.

Use procdesc(4) to enable slirp backend hotplugging.

Fix slirp_init_fds.

Fix the LIBBHYVEDIR variable.

Account for latest changes in D54650.

• constify command nvlist_t argument
• adapt vm_do_checkpoint to use the new return type

Account for latest changes in D54650.

Address @rew 's comments - change IPC command return type to nvlist_t.

Address @rew 's comments.

Address @rew 's comments.

Address @markj 's comments.

In D27220#1233531, @kib wrote:

Any progress with testing?

In D53728#1231626, @olce wrote:

There are already inline comments asking for a comment on the current limit. If possible, I think it would be great also to explain the point of this resource limit. In particular, which resources a VMM consumes that is not already accounted into existing resource limits, such as for memory: RLIMIT_AS, RLIMIT_RSS, RLIMIT_STACK, or CPU: RLIMIT_CPU? I suspect some kernel memory? Are there hardware limitations to the number of VMs that can be created?

Address @olce 's comments.

Defer incrementing the vmm counter.

Address @markj 's comments.

Address @markj 's comments.

Address @markj 's comments

Address @markj 's comments.

Address @markj 's comments.

Address @emaste 's comment.

Rebase.