x86/cpu/intel: Detect TME keyid bits before setting MTRR mask registers
authorPaolo Bonzini <pbonzini@redhat.com>
Wed, 31 Jan 2024 23:09:02 +0000 (00:09 +0100)
committerDave Hansen <dave.hansen@linux.intel.com>
Mon, 26 Feb 2024 16:16:16 +0000 (08:16 -0800)
commit6890cb1ace350b4386c8aee1343dc3b3ddd214da
treeb4845bc159648b883130430b349fc88104728dba
parent9a458198eba98b7207669a166e64d04b04cb651b
x86/cpu/intel: Detect TME keyid bits before setting MTRR mask registers

MKTME repurposes the high bit of physical address to key id for encryption
key and, even though MAXPHYADDR in CPUID[0x80000008] remains the same,
the valid bits in the MTRR mask register are based on the reduced number
of physical address bits.

detect_tme() in arch/x86/kernel/cpu/intel.c detects TME and subtracts
it from the total usable physical bits, but it is called too late.
Move the call to early_init_intel() so that it is called in setup_arch(),
before MTRRs are setup.

This fixes boot on TDX-enabled systems, which until now only worked with
"disable_mtrr_cleanup".  Without the patch, the values written to the
MTRRs mask registers were 52-bit wide (e.g. 0x000fffff_80000800) and
the writes failed; with the patch, the values are 46-bit wide, which
matches the reduced MAXPHYADDR that is shown in /proc/cpuinfo.

Reported-by: Zixi Chen <zixchen@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc:stable@vger.kernel.org
Link: https://lore.kernel.org/all/20240131230902.1867092-3-pbonzini%40redhat.com
arch/x86/kernel/cpu/intel.c