... shadow adjustments towards not building 2- and 3-level code
when !HVM. While the latter isn't functionally related to the
former, it depends on some of the rearrangements done there.

01: shadow: use __put_user() instead of __copy_to_user()
02: split __{get,put}_user() into "guest" and "unsafe" variants
03: split __copy_{from,to}_user() into "guest" and "unsafe" variants
04: PV: harden guest memory accesses against speculative abuse
05: rename {get,put}_user() to {get,put}_guest()
06: gdbsx: convert "user" to "guest" accesses
07: rename copy_{from,to}_user() to copy_{from,to}_guest_pv()
08: move stac()/clac() from {get,put}_unsafe_asm() ...
09: PV: use get_unsafe() instead of copy_from_unsafe()
10: shadow: use get_unsafe() instead of copy_from_unsafe()
11: shadow: polish shadow_write_entries()
12: shadow: move shadow_set_l<N>e() to their own source file
13: shadow: don't open-code SHF_* shorthands
14: shadow: SH_type_l2h_shadow is PV-only
15: shadow: drop SH_type_l2h_pae_shadow
16: shadow: only 4-level guest code needs building when !HVM
17: shadow: adjust is_pv_*() checks

Jan

At 16:01 +0100 on 14 Jan (1610640090), Jan Beulich wrote:
> ... shadow adjustments towards not building 2- and 3-level code
> when !HVM. While the latter isn't functionally related to the
> former, it depends on some of the rearrangements done there.

The shadow changes look good, thank you!
Reviewed-by: Tim Deegan <tim@xen.org>

I have read the uaccess stuff in passing and it looks OK to me too,
but I didn't review it in detail.

Cheers,

Tim.