UefiCpuPkg: Allow AP booting under SEV-ES

BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=2198

Typically, an AP is booted using the INIT-SIPI-SIPI sequence. This
sequence is intercepted by the hypervisor, which sets the AP's registers
to the values requested by the sequence. At that point, the hypervisor can
start the AP, which will then begin execution at the appropriate location.

Under SEV-ES, AP booting presents some challenges since the hypervisor is
not allowed to alter the AP's register state. In this situation, we have
to distinguish between the AP's first boot and AP's subsequent boots.

First boot:
 Once the AP's register state has been defined (which is before the guest
 is first booted) it cannot be altered. Should the hypervisor attempt to
 alter the register state, the change would be detected by the hardware
 and the VMRUN instruction would fail. Given this, the first boot for the
 AP is required to begin execution with this initial register state, which
 is typically the reset vector. This prevents the BSP from directing the
 AP startup location through the INIT-SIPI-SIPI sequence.

 To work around this, the firmware will provide a build time reserved area
 that can be used as the initial IP value. The hypervisor can extract this
 location value by checking for the SEV-ES reset block GUID that must be
 located 48-bytes from the end of the firmware. The format of the SEV-ES
 reset block area is:

   0x00 - 0x01 - SEV-ES Reset IP
   0x02 - 0x03 - SEV-ES Reset CS Segment Base[31:16]
   0x04 - 0x05 - Size of the SEV-ES reset block
   0x06 - 0x15 - SEV-ES Reset Block GUID
                   (00f771de-1a7e-4fcb-890e-68c77e2fb44e)

   The total size is 22 bytes. Any expansion to this block must be done
   by adding new values before existing values.

 The hypervisor will use the IP and CS values obtained from the SEV-ES
 reset block to set as the AP's initial values. The CS Segment Base
 represents the upper 16 bits of the CS segment base and must be left
 shifted by 16 bits to form the complete CS segment base value.

 Before booting the AP for the first time, the BSP must initialize the
 SEV-ES reset area. This consists of programming a FAR JMP instruction
 to the contents of a memory location that is also located in the SEV-ES
 reset area. The BSP must program the IP and CS values for the FAR JMP
 based on values drived from the INIT-SIPI-SIPI sequence.

Subsequent boots:
 Again, the hypervisor cannot alter the AP register state, so a method is
 required to take the AP out of halt state and redirect it to the desired
 IP location. If it is determined that the AP is running in an SEV-ES
 guest, then instead of calling CpuSleep(), a VMGEXIT is issued with the
 AP Reset Hold exit code (0x80000004). The hypervisor will put the AP in
 a halt state, waiting for an INIT-SIPI-SIPI sequence. Once the sequence
 is recognized, the hypervisor will resume the AP. At this point the AP
 must transition from the current 64-bit long mode down to 16-bit real
 mode and begin executing at the derived location from the INIT-SIPI-SIPI
 sequence.

 Another change is around the area of obtaining the (x2)APIC ID during AP
 startup. During AP startup, the AP can't take a #VC exception before the
 AP has established a stack. However, the AP stack is set by using the
 (x2)APIC ID, which is obtained through CPUID instructions. A CPUID
 instruction will cause a #VC, so a different method must be used. The
 GHCB protocol supports a method to obtain CPUID information from the
 hypervisor through the GHCB MSR. This method does not require a stack,
 so it is used to obtain the necessary CPUID information to determine the
 (x2)APIC ID.

The new 16-bit protected mode GDT entry is used in order to transition
from 64-bit long mode down to 16-bit real mode.

A new assembler routine is created that takes the AP from 64-bit long mode
to 16-bit real mode.  This is located under 1MB in memory and transitions
from 64-bit long mode to 32-bit compatibility mode to 16-bit protected
mode and finally 16-bit real mode.

Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
This commit is contained in:
Tom Lendacky
2020-08-12 15:21:42 -05:00
committed by mergify[bot]
parent e88a5b9833
commit 7b7508ad78
11 changed files with 738 additions and 15 deletions

View File

@@ -184,9 +184,97 @@ Releaselock:
add edi, StackStartAddressLocation
add rax, qword [edi]
mov rsp, rax
lea edi, [esi + SevEsIsEnabledLocation]
cmp byte [edi], 1 ; SevEsIsEnabled
jne CProcedureInvoke
;
; program GHCB
; Each page after the GHCB is a per-CPU page, so the calculation programs
; a GHCB to be every 8KB.
;
mov eax, SIZE_4KB
shl eax, 1 ; EAX = SIZE_4K * 2
mov ecx, ebx
mul ecx ; EAX = SIZE_4K * 2 * CpuNumber
mov edi, esi
add edi, GhcbBaseLocation
add rax, qword [edi]
mov rdx, rax
shr rdx, 32
mov rcx, 0xc0010130
wrmsr
jmp CProcedureInvoke
GetApicId:
lea edi, [esi + SevEsIsEnabledLocation]
cmp byte [edi], 1 ; SevEsIsEnabled
jne DoCpuid
;
; Since we don't have a stack yet, we can't take a #VC
; exception. Use the GHCB protocol to perform the CPUID
; calls.
;
mov rcx, 0xc0010130
rdmsr
shl rdx, 32
or rax, rdx
mov rdi, rax ; RDI now holds the original GHCB GPA
mov rdx, 0 ; CPUID function 0
mov rax, 0 ; RAX register requested
or rax, 4
wrmsr
rep vmmcall
rdmsr
cmp edx, 0bh
jb NoX2ApicSevEs ; CPUID level below CPUID_EXTENDED_TOPOLOGY
mov rdx, 0bh ; CPUID function 0x0b
mov rax, 040000000h ; RBX register requested
or rax, 4
wrmsr
rep vmmcall
rdmsr
test edx, 0ffffh
jz NoX2ApicSevEs ; CPUID.0BH:EBX[15:0] is zero
mov rdx, 0bh ; CPUID function 0x0b
mov rax, 0c0000000h ; RDX register requested
or rax, 4
wrmsr
rep vmmcall
rdmsr
; Processor is x2APIC capable; 32-bit x2APIC ID is now in EDX
jmp RestoreGhcb
NoX2ApicSevEs:
; Processor is not x2APIC capable, so get 8-bit APIC ID
mov rdx, 1 ; CPUID function 1
mov rax, 040000000h ; RBX register requested
or rax, 4
wrmsr
rep vmmcall
rdmsr
shr edx, 24
RestoreGhcb:
mov rbx, rdx ; Save x2APIC/APIC ID
mov rdx, rdi ; RDI holds the saved GHCB GPA
shr rdx, 32
mov eax, edi
wrmsr
mov rdx, rbx
; x2APIC ID or APIC ID is in EDX
jmp GetProcessorNumber
DoCpuid:
mov eax, 0
cpuid
cmp eax, 0bh
@@ -253,12 +341,158 @@ CProcedureInvoke:
RendezvousFunnelProcEnd:
;-------------------------------------------------------------------------------------
;SwitchToRealProc procedure follows.
;ALSO THIS PROCEDURE IS EXECUTED BY APs TRANSITIONING TO 16 BIT MODE. HENCE THIS PROC
;IS IN MACHINE CODE.
; SwitchToRealProc (UINTN BufferStart, UINT16 Code16, UINT16 Code32, UINTN StackStart)
; rcx - Buffer Start
; rdx - Code16 Selector Offset
; r8 - Code32 Selector Offset
; r9 - Stack Start
;-------------------------------------------------------------------------------------
global ASM_PFX(SwitchToRealProc)
ASM_PFX(SwitchToRealProc):
SwitchToRealProcStart:
BITS 64
cli
;
; Get RDX reset value before changing stacks since the
; new stack won't be able to accomodate a #VC exception.
;
push rax
push rbx
push rcx
push rdx
mov rax, 1
cpuid
mov rsi, rax ; Save off the reset value for RDX
pop rdx
pop rcx
pop rbx
pop rax
;
; Establish stack below 1MB
;
mov rsp, r9
;
; Push ultimate Reset Vector onto the stack
;
mov rax, rcx
shr rax, 4
push word 0x0002 ; RFLAGS
push ax ; CS
push word 0x0000 ; RIP
push word 0x0000 ; For alignment, will be discarded
;
; Get address of "16-bit operand size" label
;
lea rbx, [PM16Mode]
;
; Push addresses used to change to compatibility mode
;
lea rax, [CompatMode]
push r8
push rax
;
; Clear R8 - R15, for reset, before going into 32-bit mode
;
xor r8, r8
xor r9, r9
xor r10, r10
xor r11, r11
xor r12, r12
xor r13, r13
xor r14, r14
xor r15, r15
;
; Far return into 32-bit mode
;
o64 retf
BITS 32
CompatMode:
;
; Set up stack to prepare for exiting protected mode
;
push edx ; Code16 CS
push ebx ; PM16Mode label address
;
; Disable paging
;
mov eax, cr0 ; Read CR0
btr eax, 31 ; Set PG=0
mov cr0, eax ; Write CR0
;
; Disable long mode
;
mov ecx, 0c0000080h ; EFER MSR number
rdmsr ; Read EFER
btr eax, 8 ; Set LME=0
wrmsr ; Write EFER
;
; Disable PAE
;
mov eax, cr4 ; Read CR4
btr eax, 5 ; Set PAE=0
mov cr4, eax ; Write CR4
mov edx, esi ; Restore RDX reset value
;
; Switch to 16-bit operand size
;
retf
BITS 16
;
; At entry to this label
; - RDX will have its reset value
; - On the top of the stack
; - Alignment data (two bytes) to be discarded
; - IP for Real Mode (two bytes)
; - CS for Real Mode (two bytes)
;
PM16Mode:
mov eax, cr0 ; Read CR0
btr eax, 0 ; Set PE=0
mov cr0, eax ; Write CR0
pop ax ; Discard alignment data
;
; Clear registers (except RDX and RSP) before going into 16-bit mode
;
xor eax, eax
xor ebx, ebx
xor ecx, ecx
xor esi, esi
xor edi, edi
xor ebp, ebp
iret
SwitchToRealProcEnd:
;-------------------------------------------------------------------------------------
; AsmRelocateApLoop (MwaitSupport, ApTargetCState, PmCodeSegment, TopOfApStack, CountTofinish);
;-------------------------------------------------------------------------------------
global ASM_PFX(AsmRelocateApLoop)
ASM_PFX(AsmRelocateApLoop):
AsmRelocateApLoopStart:
BITS 64
cli ; Disable interrupt before switching to 32-bit mode
mov rax, [rsp + 40] ; CountTofinish
lock dec dword [rax] ; (*CountTofinish)--
@@ -324,6 +558,11 @@ ASM_PFX(AsmGetAddressMap):
mov qword [rcx + 18h], rax
mov qword [rcx + 20h], AsmRelocateApLoopEnd - AsmRelocateApLoopStart
mov qword [rcx + 28h], Flat32Start - RendezvousFunnelProcStart
mov qword [rcx + 30h], SwitchToRealProcEnd - SwitchToRealProcStart ; SwitchToRealSize
mov qword [rcx + 38h], SwitchToRealProcStart - RendezvousFunnelProcStart ; SwitchToRealOffset
mov qword [rcx + 40h], SwitchToRealProcStart - Flat32Start ; SwitchToRealNoNxOffset
mov qword [rcx + 48h], PM16Mode - RendezvousFunnelProcStart ; SwitchToRealPM16ModeOffset
mov qword [rcx + 50h], SwitchToRealProcEnd - PM16Mode ; SwitchToRealPM16ModeSize
ret
;-------------------------------------------------------------------------------------