KVM: ppc: PowerPC 440 KVM implementation
This functionality is definitely experimental, but is capable of running
unmodified PowerPC 440 Linux kernels as guests on a PowerPC 440 host. (Only
tested with 440EP "Bamboo" guests so far, but with appropriate userspace
support other SoC/board combinations should work.)
See Documentation/powerpc/kvm_440.txt for technical details.
[stephen: build fix]
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Avi Kivity <avi@qumranet.com>
diff --git a/arch/powerpc/kvm/booke_interrupts.S b/arch/powerpc/kvm/booke_interrupts.S
new file mode 100644
index 0000000..3b653b5
--- /dev/null
+++ b/arch/powerpc/kvm/booke_interrupts.S
@@ -0,0 +1,436 @@
+/*
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ *
+ * Copyright IBM Corp. 2007
+ *
+ * Authors: Hollis Blanchard <hollisb@us.ibm.com>
+ */
+
+#include <asm/ppc_asm.h>
+#include <asm/kvm_asm.h>
+#include <asm/reg.h>
+#include <asm/mmu-44x.h>
+#include <asm/page.h>
+#include <asm/asm-offsets.h>
+
+#define KVMPPC_MSR_MASK (MSR_CE|MSR_EE|MSR_PR|MSR_DE|MSR_ME|MSR_IS|MSR_DS)
+
+#define VCPU_GPR(n) (VCPU_GPRS + (n * 4))
+
+/* The host stack layout: */
+#define HOST_R1 0 /* Implied by stwu. */
+#define HOST_CALLEE_LR 4
+#define HOST_RUN 8
+/* r2 is special: it holds 'current', and it made nonvolatile in the
+ * kernel with the -ffixed-r2 gcc option. */
+#define HOST_R2 12
+#define HOST_NV_GPRS 16
+#define HOST_NV_GPR(n) (HOST_NV_GPRS + ((n - 14) * 4))
+#define HOST_MIN_STACK_SIZE (HOST_NV_GPR(31) + 4)
+#define HOST_STACK_SIZE (((HOST_MIN_STACK_SIZE + 15) / 16) * 16) /* Align. */
+#define HOST_STACK_LR (HOST_STACK_SIZE + 4) /* In caller stack frame. */
+
+#define NEED_INST_MASK ((1<<BOOKE_INTERRUPT_PROGRAM) | \
+ (1<<BOOKE_INTERRUPT_DTLB_MISS))
+
+#define NEED_DEAR_MASK ((1<<BOOKE_INTERRUPT_DATA_STORAGE) | \
+ (1<<BOOKE_INTERRUPT_DTLB_MISS))
+
+#define NEED_ESR_MASK ((1<<BOOKE_INTERRUPT_DATA_STORAGE) | \
+ (1<<BOOKE_INTERRUPT_INST_STORAGE) | \
+ (1<<BOOKE_INTERRUPT_PROGRAM) | \
+ (1<<BOOKE_INTERRUPT_DTLB_MISS))
+
+.macro KVM_HANDLER ivor_nr
+_GLOBAL(kvmppc_handler_\ivor_nr)
+ /* Get pointer to vcpu and record exit number. */
+ mtspr SPRN_SPRG0, r4
+ mfspr r4, SPRN_SPRG1
+ stw r5, VCPU_GPR(r5)(r4)
+ stw r6, VCPU_GPR(r6)(r4)
+ mfctr r5
+ lis r6, kvmppc_resume_host@h
+ stw r5, VCPU_CTR(r4)
+ li r5, \ivor_nr
+ ori r6, r6, kvmppc_resume_host@l
+ mtctr r6
+ bctr
+.endm
+
+_GLOBAL(kvmppc_handlers_start)
+KVM_HANDLER BOOKE_INTERRUPT_CRITICAL
+KVM_HANDLER BOOKE_INTERRUPT_MACHINE_CHECK
+KVM_HANDLER BOOKE_INTERRUPT_DATA_STORAGE
+KVM_HANDLER BOOKE_INTERRUPT_INST_STORAGE
+KVM_HANDLER BOOKE_INTERRUPT_EXTERNAL
+KVM_HANDLER BOOKE_INTERRUPT_ALIGNMENT
+KVM_HANDLER BOOKE_INTERRUPT_PROGRAM
+KVM_HANDLER BOOKE_INTERRUPT_FP_UNAVAIL
+KVM_HANDLER BOOKE_INTERRUPT_SYSCALL
+KVM_HANDLER BOOKE_INTERRUPT_AP_UNAVAIL
+KVM_HANDLER BOOKE_INTERRUPT_DECREMENTER
+KVM_HANDLER BOOKE_INTERRUPT_FIT
+KVM_HANDLER BOOKE_INTERRUPT_WATCHDOG
+KVM_HANDLER BOOKE_INTERRUPT_DTLB_MISS
+KVM_HANDLER BOOKE_INTERRUPT_ITLB_MISS
+KVM_HANDLER BOOKE_INTERRUPT_DEBUG
+
+_GLOBAL(kvmppc_handler_len)
+ .long kvmppc_handler_1 - kvmppc_handler_0
+
+
+/* Registers:
+ * SPRG0: guest r4
+ * r4: vcpu pointer
+ * r5: KVM exit number
+ */
+_GLOBAL(kvmppc_resume_host)
+ stw r3, VCPU_GPR(r3)(r4)
+ mfcr r3
+ stw r3, VCPU_CR(r4)
+ stw r7, VCPU_GPR(r7)(r4)
+ stw r8, VCPU_GPR(r8)(r4)
+ stw r9, VCPU_GPR(r9)(r4)
+
+ li r6, 1
+ slw r6, r6, r5
+
+ /* Save the faulting instruction and all GPRs for emulation. */
+ andi. r7, r6, NEED_INST_MASK
+ beq ..skip_inst_copy
+ mfspr r9, SPRN_SRR0
+ mfmsr r8
+ ori r7, r8, MSR_DS
+ mtmsr r7
+ isync
+ lwz r9, 0(r9)
+ mtmsr r8
+ isync
+ stw r9, VCPU_LAST_INST(r4)
+
+ stw r15, VCPU_GPR(r15)(r4)
+ stw r16, VCPU_GPR(r16)(r4)
+ stw r17, VCPU_GPR(r17)(r4)
+ stw r18, VCPU_GPR(r18)(r4)
+ stw r19, VCPU_GPR(r19)(r4)
+ stw r20, VCPU_GPR(r20)(r4)
+ stw r21, VCPU_GPR(r21)(r4)
+ stw r22, VCPU_GPR(r22)(r4)
+ stw r23, VCPU_GPR(r23)(r4)
+ stw r24, VCPU_GPR(r24)(r4)
+ stw r25, VCPU_GPR(r25)(r4)
+ stw r26, VCPU_GPR(r26)(r4)
+ stw r27, VCPU_GPR(r27)(r4)
+ stw r28, VCPU_GPR(r28)(r4)
+ stw r29, VCPU_GPR(r29)(r4)
+ stw r30, VCPU_GPR(r30)(r4)
+ stw r31, VCPU_GPR(r31)(r4)
+..skip_inst_copy:
+
+ /* Also grab DEAR and ESR before the host can clobber them. */
+
+ andi. r7, r6, NEED_DEAR_MASK
+ beq ..skip_dear
+ mfspr r9, SPRN_DEAR
+ stw r9, VCPU_FAULT_DEAR(r4)
+..skip_dear:
+
+ andi. r7, r6, NEED_ESR_MASK
+ beq ..skip_esr
+ mfspr r9, SPRN_ESR
+ stw r9, VCPU_FAULT_ESR(r4)
+..skip_esr:
+
+ /* Save remaining volatile guest register state to vcpu. */
+ stw r0, VCPU_GPR(r0)(r4)
+ stw r1, VCPU_GPR(r1)(r4)
+ stw r2, VCPU_GPR(r2)(r4)
+ stw r10, VCPU_GPR(r10)(r4)
+ stw r11, VCPU_GPR(r11)(r4)
+ stw r12, VCPU_GPR(r12)(r4)
+ stw r13, VCPU_GPR(r13)(r4)
+ stw r14, VCPU_GPR(r14)(r4) /* We need a NV GPR below. */
+ mflr r3
+ stw r3, VCPU_LR(r4)
+ mfxer r3
+ stw r3, VCPU_XER(r4)
+ mfspr r3, SPRN_SPRG0
+ stw r3, VCPU_GPR(r4)(r4)
+ mfspr r3, SPRN_SRR0
+ stw r3, VCPU_PC(r4)
+
+ /* Restore host stack pointer and PID before IVPR, since the host
+ * exception handlers use them. */
+ lwz r1, VCPU_HOST_STACK(r4)
+ lwz r3, VCPU_HOST_PID(r4)
+ mtspr SPRN_PID, r3
+
+ /* Restore host IVPR before re-enabling interrupts. We cheat and know
+ * that Linux IVPR is always 0xc0000000. */
+ lis r3, 0xc000
+ mtspr SPRN_IVPR, r3
+
+ /* Switch to kernel stack and jump to handler. */
+ LOAD_REG_ADDR(r3, kvmppc_handle_exit)
+ mtctr r3
+ lwz r3, HOST_RUN(r1)
+ lwz r2, HOST_R2(r1)
+ mr r14, r4 /* Save vcpu pointer. */
+
+ bctrl /* kvmppc_handle_exit() */
+
+ /* Restore vcpu pointer and the nonvolatiles we used. */
+ mr r4, r14
+ lwz r14, VCPU_GPR(r14)(r4)
+
+ /* Sometimes instruction emulation must restore complete GPR state. */
+ andi. r5, r3, RESUME_FLAG_NV
+ beq ..skip_nv_load
+ lwz r15, VCPU_GPR(r15)(r4)
+ lwz r16, VCPU_GPR(r16)(r4)
+ lwz r17, VCPU_GPR(r17)(r4)
+ lwz r18, VCPU_GPR(r18)(r4)
+ lwz r19, VCPU_GPR(r19)(r4)
+ lwz r20, VCPU_GPR(r20)(r4)
+ lwz r21, VCPU_GPR(r21)(r4)
+ lwz r22, VCPU_GPR(r22)(r4)
+ lwz r23, VCPU_GPR(r23)(r4)
+ lwz r24, VCPU_GPR(r24)(r4)
+ lwz r25, VCPU_GPR(r25)(r4)
+ lwz r26, VCPU_GPR(r26)(r4)
+ lwz r27, VCPU_GPR(r27)(r4)
+ lwz r28, VCPU_GPR(r28)(r4)
+ lwz r29, VCPU_GPR(r29)(r4)
+ lwz r30, VCPU_GPR(r30)(r4)
+ lwz r31, VCPU_GPR(r31)(r4)
+..skip_nv_load:
+
+ /* Should we return to the guest? */
+ andi. r5, r3, RESUME_FLAG_HOST
+ beq lightweight_exit
+
+ srawi r3, r3, 2 /* Shift -ERR back down. */
+
+heavyweight_exit:
+ /* Not returning to guest. */
+
+ /* We already saved guest volatile register state; now save the
+ * non-volatiles. */
+ stw r15, VCPU_GPR(r15)(r4)
+ stw r16, VCPU_GPR(r16)(r4)
+ stw r17, VCPU_GPR(r17)(r4)
+ stw r18, VCPU_GPR(r18)(r4)
+ stw r19, VCPU_GPR(r19)(r4)
+ stw r20, VCPU_GPR(r20)(r4)
+ stw r21, VCPU_GPR(r21)(r4)
+ stw r22, VCPU_GPR(r22)(r4)
+ stw r23, VCPU_GPR(r23)(r4)
+ stw r24, VCPU_GPR(r24)(r4)
+ stw r25, VCPU_GPR(r25)(r4)
+ stw r26, VCPU_GPR(r26)(r4)
+ stw r27, VCPU_GPR(r27)(r4)
+ stw r28, VCPU_GPR(r28)(r4)
+ stw r29, VCPU_GPR(r29)(r4)
+ stw r30, VCPU_GPR(r30)(r4)
+ stw r31, VCPU_GPR(r31)(r4)
+
+ /* Load host non-volatile register state from host stack. */
+ lwz r14, HOST_NV_GPR(r14)(r1)
+ lwz r15, HOST_NV_GPR(r15)(r1)
+ lwz r16, HOST_NV_GPR(r16)(r1)
+ lwz r17, HOST_NV_GPR(r17)(r1)
+ lwz r18, HOST_NV_GPR(r18)(r1)
+ lwz r19, HOST_NV_GPR(r19)(r1)
+ lwz r20, HOST_NV_GPR(r20)(r1)
+ lwz r21, HOST_NV_GPR(r21)(r1)
+ lwz r22, HOST_NV_GPR(r22)(r1)
+ lwz r23, HOST_NV_GPR(r23)(r1)
+ lwz r24, HOST_NV_GPR(r24)(r1)
+ lwz r25, HOST_NV_GPR(r25)(r1)
+ lwz r26, HOST_NV_GPR(r26)(r1)
+ lwz r27, HOST_NV_GPR(r27)(r1)
+ lwz r28, HOST_NV_GPR(r28)(r1)
+ lwz r29, HOST_NV_GPR(r29)(r1)
+ lwz r30, HOST_NV_GPR(r30)(r1)
+ lwz r31, HOST_NV_GPR(r31)(r1)
+
+ /* Return to kvm_vcpu_run(). */
+ lwz r4, HOST_STACK_LR(r1)
+ addi r1, r1, HOST_STACK_SIZE
+ mtlr r4
+ /* r3 still contains the return code from kvmppc_handle_exit(). */
+ blr
+
+
+/* Registers:
+ * r3: kvm_run pointer
+ * r4: vcpu pointer
+ */
+_GLOBAL(__kvmppc_vcpu_run)
+ stwu r1, -HOST_STACK_SIZE(r1)
+ stw r1, VCPU_HOST_STACK(r4) /* Save stack pointer to vcpu. */
+
+ /* Save host state to stack. */
+ stw r3, HOST_RUN(r1)
+ mflr r3
+ stw r3, HOST_STACK_LR(r1)
+
+ /* Save host non-volatile register state to stack. */
+ stw r14, HOST_NV_GPR(r14)(r1)
+ stw r15, HOST_NV_GPR(r15)(r1)
+ stw r16, HOST_NV_GPR(r16)(r1)
+ stw r17, HOST_NV_GPR(r17)(r1)
+ stw r18, HOST_NV_GPR(r18)(r1)
+ stw r19, HOST_NV_GPR(r19)(r1)
+ stw r20, HOST_NV_GPR(r20)(r1)
+ stw r21, HOST_NV_GPR(r21)(r1)
+ stw r22, HOST_NV_GPR(r22)(r1)
+ stw r23, HOST_NV_GPR(r23)(r1)
+ stw r24, HOST_NV_GPR(r24)(r1)
+ stw r25, HOST_NV_GPR(r25)(r1)
+ stw r26, HOST_NV_GPR(r26)(r1)
+ stw r27, HOST_NV_GPR(r27)(r1)
+ stw r28, HOST_NV_GPR(r28)(r1)
+ stw r29, HOST_NV_GPR(r29)(r1)
+ stw r30, HOST_NV_GPR(r30)(r1)
+ stw r31, HOST_NV_GPR(r31)(r1)
+
+ /* Load guest non-volatiles. */
+ lwz r14, VCPU_GPR(r14)(r4)
+ lwz r15, VCPU_GPR(r15)(r4)
+ lwz r16, VCPU_GPR(r16)(r4)
+ lwz r17, VCPU_GPR(r17)(r4)
+ lwz r18, VCPU_GPR(r18)(r4)
+ lwz r19, VCPU_GPR(r19)(r4)
+ lwz r20, VCPU_GPR(r20)(r4)
+ lwz r21, VCPU_GPR(r21)(r4)
+ lwz r22, VCPU_GPR(r22)(r4)
+ lwz r23, VCPU_GPR(r23)(r4)
+ lwz r24, VCPU_GPR(r24)(r4)
+ lwz r25, VCPU_GPR(r25)(r4)
+ lwz r26, VCPU_GPR(r26)(r4)
+ lwz r27, VCPU_GPR(r27)(r4)
+ lwz r28, VCPU_GPR(r28)(r4)
+ lwz r29, VCPU_GPR(r29)(r4)
+ lwz r30, VCPU_GPR(r30)(r4)
+ lwz r31, VCPU_GPR(r31)(r4)
+
+lightweight_exit:
+ stw r2, HOST_R2(r1)
+
+ mfspr r3, SPRN_PID
+ stw r3, VCPU_HOST_PID(r4)
+ lwz r3, VCPU_PID(r4)
+ mtspr SPRN_PID, r3
+
+ /* Prevent all TLB updates. */
+ mfmsr r5
+ lis r6, (MSR_EE|MSR_CE|MSR_ME|MSR_DE)@h
+ ori r6, r6, (MSR_EE|MSR_CE|MSR_ME|MSR_DE)@l
+ andc r6, r5, r6
+ mtmsr r6
+
+ /* Save the host's non-pinned TLB mappings, and load the guest mappings
+ * over them. Leave the host's "pinned" kernel mappings in place. */
+ /* XXX optimization: use generation count to avoid swapping unmodified
+ * entries. */
+ mfspr r10, SPRN_MMUCR /* Save host MMUCR. */
+ lis r8, tlb_44x_hwater@ha
+ lwz r8, tlb_44x_hwater@l(r8)
+ addi r3, r4, VCPU_HOST_TLB - 4
+ addi r9, r4, VCPU_SHADOW_TLB - 4
+ li r6, 0
+1:
+ /* Save host entry. */
+ tlbre r7, r6, PPC44x_TLB_PAGEID
+ mfspr r5, SPRN_MMUCR
+ stwu r5, 4(r3)
+ stwu r7, 4(r3)
+ tlbre r7, r6, PPC44x_TLB_XLAT
+ stwu r7, 4(r3)
+ tlbre r7, r6, PPC44x_TLB_ATTRIB
+ stwu r7, 4(r3)
+ /* Load guest entry. */
+ lwzu r7, 4(r9)
+ mtspr SPRN_MMUCR, r7
+ lwzu r7, 4(r9)
+ tlbwe r7, r6, PPC44x_TLB_PAGEID
+ lwzu r7, 4(r9)
+ tlbwe r7, r6, PPC44x_TLB_XLAT
+ lwzu r7, 4(r9)
+ tlbwe r7, r6, PPC44x_TLB_ATTRIB
+ /* Increment index. */
+ addi r6, r6, 1
+ cmpw r6, r8
+ blt 1b
+ mtspr SPRN_MMUCR, r10 /* Restore host MMUCR. */
+
+ iccci 0, 0 /* XXX hack */
+
+ /* Load some guest volatiles. */
+ lwz r0, VCPU_GPR(r0)(r4)
+ lwz r2, VCPU_GPR(r2)(r4)
+ lwz r9, VCPU_GPR(r9)(r4)
+ lwz r10, VCPU_GPR(r10)(r4)
+ lwz r11, VCPU_GPR(r11)(r4)
+ lwz r12, VCPU_GPR(r12)(r4)
+ lwz r13, VCPU_GPR(r13)(r4)
+ lwz r3, VCPU_LR(r4)
+ mtlr r3
+ lwz r3, VCPU_XER(r4)
+ mtxer r3
+
+ /* Switch the IVPR. XXX If we take a TLB miss after this we're screwed,
+ * so how do we make sure vcpu won't fault? */
+ lis r8, kvmppc_booke_handlers@ha
+ lwz r8, kvmppc_booke_handlers@l(r8)
+ mtspr SPRN_IVPR, r8
+
+ /* Save vcpu pointer for the exception handlers. */
+ mtspr SPRN_SPRG1, r4
+
+ /* Can't switch the stack pointer until after IVPR is switched,
+ * because host interrupt handlers would get confused. */
+ lwz r1, VCPU_GPR(r1)(r4)
+
+ /* XXX handle USPRG0 */
+ /* Host interrupt handlers may have clobbered these guest-readable
+ * SPRGs, so we need to reload them here with the guest's values. */
+ lwz r3, VCPU_SPRG4(r4)
+ mtspr SPRN_SPRG4, r3
+ lwz r3, VCPU_SPRG5(r4)
+ mtspr SPRN_SPRG5, r3
+ lwz r3, VCPU_SPRG6(r4)
+ mtspr SPRN_SPRG6, r3
+ lwz r3, VCPU_SPRG7(r4)
+ mtspr SPRN_SPRG7, r3
+
+ /* Finish loading guest volatiles and jump to guest. */
+ lwz r3, VCPU_CTR(r4)
+ mtctr r3
+ lwz r3, VCPU_CR(r4)
+ mtcr r3
+ lwz r5, VCPU_GPR(r5)(r4)
+ lwz r6, VCPU_GPR(r6)(r4)
+ lwz r7, VCPU_GPR(r7)(r4)
+ lwz r8, VCPU_GPR(r8)(r4)
+ lwz r3, VCPU_PC(r4)
+ mtsrr0 r3
+ lwz r3, VCPU_MSR(r4)
+ oris r3, r3, KVMPPC_MSR_MASK@h
+ ori r3, r3, KVMPPC_MSR_MASK@l
+ mtsrr1 r3
+ lwz r3, VCPU_GPR(r3)(r4)
+ lwz r4, VCPU_GPR(r4)(r4)
+ rfi
