The purpose of this series is to fix the fact that you cannot boot 32bit PV
guests older than Linux 3.8ish on Broadwell processors when Xen is using
SMAP.

It should be noted that the same problem exists with SMEP and 32bit PV guests,
but I have no sympathy for PV kernels which start executing code via user
mappings, or compunction to work around the issue.

This series is RFC because there are still outstanding tests to narrow down
whether a windows BSOD is caused by this series (unlikely I hope, as nothing
HVM-related has changed), or whether it is from recent changes in staging
(more likely, but still not obvious what).

Either way, the series is good enough for review so is presented here.

Andrew Cooper (8):
  common/vsprintf: Special-case DOMID_IDLE handling for %pv
  x86/traps: Avoid using current too early on boot
  x86/setup: Initialise CR4 before creating idle_vcpu[0]
  xen/x86: Clean up CR4 definitions
  xen/x86: Drop PSE from XEN_MINIMAL_CR4
  xen/x86: Calculate PV CR4 masks at boot
  xen/x86: Rework CR4 handling for PV guests
  xen/x86: Additional SMAP modes to work around buggy 32bit PV guests

 docs/misc/xen-command-line.markdown |   25 +++++++-
 xen/arch/x86/domain.c               |  119 +++++++++++++++++++++++++++++++----
 xen/arch/x86/setup.c                |   59 +++++++++++------
 xen/arch/x86/traps.c                |   47 +++++++++-----
 xen/arch/x86/x86_64/traps.c         |    2 +-
 xen/common/vsprintf.c               |    5 +-
 xen/include/asm-x86/domain.h        |   19 +-----
 xen/include/asm-x86/processor.h     |   49 +++++++++------
 8 files changed, 239 insertions(+), 86 deletions(-)

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

Reply via email to