This series refactors the mshv memory region subsystem in preparation
for mapping populated pages into the hypervisor at movable region
creation time, rather than relying solely on demand faulting.
The primary motivation is to ensure that when userspace passes a
pre-populated mapping for a movable memory region, those pages are
immediately visible to the hypervisor. Previously, all movable regions
were created with HV_MAP_GPA_NO_ACCESS on every page regardless of
whether the backing pages were already present, deferring all mapping
to the fault handler. This added unnecessary fault overhead and
complicated the initial setup of child partitions with pre-populated
memory.
v3:
- Fixed a few issues.
- Reworked the code flow to simplify the logic.
v2:
- Rebased on top of latest mainline, simplified the check for valid PFNs,
added other minor cleanups and improvements.
---
Stanislav Kinsburskii (11):
mshv: Don't reject huge-page stride on unaligned region length
mshv: Don't request HMM write fault for read-only regions
mshv: Convert region storage from page pointers to PFNs
mshv: Refactor region segmentation into a dedicated helper
mshv: Support address range holes in remapping
mshv: Iterate VMAs when faulting in region pages
mshv: Scale fault granularity for non-4 KiB host pages
mshv: Move pinned region setup to mshv_regions.c
mshv: Map populated pages on movable region creation
mshv: Extract MMIO region mapping into separate function
mshv: Add tracepoint for map GPA hypercall
drivers/hv/mshv_regions.c | 636 ++++++++++++++++++++++++++++------------
drivers/hv/mshv_root.h | 30 +-
drivers/hv/mshv_root_hv_call.c | 52 ++-
drivers/hv/mshv_root_main.c | 105 +------
drivers/hv/mshv_trace.h | 36 ++
5 files changed, 530 insertions(+), 329 deletions(-)
