[v2,00/13] x86: assorted shadow mode adjustments

Message

Jan Beulich March 30, 2023, 11:22 a.m. UTC

This is kind of fallout from XSA-427 investigations, partly related to
there having been a more intrusive first approach.

Most patches aren't really dependent upon one another, so can probably
go in independently (as they get acked).

A few patches from v1 went in, but there are also three new patches in
v2. See individual patches for what has changed (in response to review
comments).

01: rename SHADOW_FOREACH_L<N>E() to FOREACH_PRESENT_L<N>E()
02: drop redundant present bit checks from SHADOW_FOREACH_L<N>E() "bodies"
03: reduce explicit log-dirty recording for HVM
04: call sh_update_cr3() directly from sh_page_fault()
05: don't generate bogus "domain dying" trace entry from sh_page_fault()
06: use lighter weight mode checks
07: move OOS functions to their own file
08: sh_rm_write_access_from_sl1p() is HVM-only
09: drop is_hvm_...() where easily possible
10: make monitor table create/destroy more consistent
11: vCPU-s never have "no mode"
12: "monitor table" is a HVM-only concept
13: adjust monitor table prealloc amount

Jan