[GIT PULL] KVM: x86 pull requests for 7.2

[Sean Christopherson]

Sorry for not getting these to you yesterday.  I was all prepared to work on
them before meetings hit, and then SQUIRREL!


Outside of these pull requests, there's a set of five selftests patches[1] that
I have sitting in "selftests_l2_stacks" and fed into linux-next.  I'm not
including a pull request for that branch, as it's deliberately built on top of
all the other local merges to avoid a pile of annoying conflicts (the changes
allocate a proper stack for all L2 vCPUs, and so touches every test that does
anything nested related).  If you can apply the patches directly after merging
the other pull requests, that would make me quite happy.  Alternatively, I'll
send a separate pull request once this batch is merged.

[1] https://lore.kernel.org/all/20260610003030.2957261-1-seanjc@xxxxxxxxxx


There are two conflicts I am aware of; thankfully both are fairly trivial.

Internal to these pull requests, "misc" and "svm" conflict on header includes
in svm.h: take the incoming headers and the kvm_cache_regs.h => regs.h rename.

diff --cc arch/x86/kvm/svm/svm.h
index d013a5ceb62e,87c6b105deef..000000000000
--- a/arch/x86/kvm/svm/svm.h
+++ b/arch/x86/kvm/svm/svm.h
@@@ -23,7 -23,9 +23,9 @@@
  #include <asm/sev-common.h>
  
  #include "cpuid.h"
 -#include "kvm_cache_regs.h"
 +#include "regs.h"
+ #include "x86.h"
+ #include "pmu.h"
  
  /*
   * Helpers to convert to/from physical addresses for pages whose address is

Externally, the "mmu" branch will conflict with the tip tree on
arch/x86/include/asm/tdx.h.  Mark's fixup in -next[2] is what I've been using
in my local resolutions.  Basically take the KVM changes, but keep the
void tdx_sys_disable(void); declaration.

[2] https://lore.kernel.org/all/ahb-PZwR5mEd8A9W@xxxxxxxxxxxxx