Actions
Bug #14251
openoperationalize bhyve dirty page tracking
Start date:
Due date:
% Done:
0%
Estimated time:
Difficulty:
Medium
Tags:
Gerrit CR:
Description
The change for #13896 / #13981 added a kernel variable gpt_track_dirty
, defaulting to false
, which enables the tracking of dirtied pages for the guest-physical (EPT/NPT) address space. It is not on by default due to cost (tracking those pages, especially from device emulation, is not free). The bhyve ioctl interface should have a means of enabling that for a VM, probably at the time of its creation. Userspace consumers of the kernel vmm interfaces should have a straightforward way of checking for proper hardware support (Intel doesn't have hardware A/D tracking until at least Haswell, it would seem), in addition to manipulating its value on guests.
Related issues
Updated by Patrick Mooney 6 months ago
- Related to Bug #13896: bhyve VM interfaces should be better fit added
Actions