[PATCH v2 0/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE fixes

STALE3525d

Revision v2 of 2 in this series.

11 messages, 3 authors, 2017-01-31 · open the first message on its own page

[PATCH v2 0/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE fixes

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2016-12-21 18:29:43

Hi,

I took into account the feedback from last time. I was sceptical at
first about moving the store to save r9 from the KVM test to the
KVM handler. After implementing it, it actually worked well.

Thanks,
Nick

Nicholas Piggin (3):
  KVM: PPC: Book3S: Change interrupt call to reduce scratch space use on HV
  KVM: PPC: Book3S: Move 64-bit KVM interrupt handler out from alt section
  KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

 arch/powerpc/include/asm/exception-64s.h | 63 ++++++++++++++++++++++++--------
 arch/powerpc/include/asm/head-64.h       |  2 +-
 arch/powerpc/kernel/exceptions-64s.S     | 10 ++---
 arch/powerpc/kvm/book3s_hv_rmhandlers.S  | 15 ++++++--
 arch/powerpc/kvm/book3s_segment.S        | 32 ++++++++++++----
 5 files changed, 89 insertions(+), 33 deletions(-)

-- 
2.11.0

[PATCH 1/3] KVM: PPC: Book3S: Change interrupt call to reduce scratch space use on HV

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2016-12-21 18:29:52

Change the calling convention to put the trap number together with
CR in two halves of r12, which frees up HSTATE_SCRATCH2 in the HV
handler.

The 64-bit PR handler entry translates the calling convention back
to match the previous call convention (i.e., shared with 32-bit), for
simplicity.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/exception-64s.h | 24 +++++++++++-------------
 arch/powerpc/kvm/book3s_hv_rmhandlers.S  | 16 +++++++++-------
 arch/powerpc/kvm/book3s_segment.S        | 25 ++++++++++++++++++-------
 3 files changed, 38 insertions(+), 27 deletions(-)

diff --git a/arch/powerpc/include/asm/exception-64s.h b/arch/powerpc/include/asm/exception-64s.h
index 9a3eee661297..a02a268bde6b 100644
--- a/arch/powerpc/include/asm/exception-64s.h
+++ b/arch/powerpc/include/asm/exception-64s.h

@@ -233,7 +233,7 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 
 #endif
 
-#define __KVM_HANDLER_PROLOG(area, n)					\
+#define __KVM_HANDLER(area, h, n)					\
 	BEGIN_FTR_SECTION_NESTED(947)					\
 	ld	r10,area+EX_CFAR(r13);					\
 	std	r10,HSTATE_CFAR(r13);					\

@@ -243,30 +243,28 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r10,HSTATE_PPR(r13);					\
 	END_FTR_SECTION_NESTED(CPU_FTR_HAS_PPR,CPU_FTR_HAS_PPR,948);	\
 	ld	r10,area+EX_R10(r13);					\
-	stw	r9,HSTATE_SCRATCH1(r13);				\
-	ld	r9,area+EX_R9(r13);					\
 	std	r12,HSTATE_SCRATCH0(r13);				\
-
-#define __KVM_HANDLER(area, h, n)					\
-	__KVM_HANDLER_PROLOG(area, n)					\
-	li	r12,n;							\
+	sldi	r12,r9,32;						\
+	ori	r12,r12,(n);						\
+	ld	r9,area+EX_R9(r13);					\
 	b	kvmppc_interrupt
 
 #define __KVM_HANDLER_SKIP(area, h, n)					\
 	cmpwi	r10,KVM_GUEST_MODE_SKIP;				\
-	ld	r10,area+EX_R10(r13);					\
 	beq	89f;							\
-	stw	r9,HSTATE_SCRATCH1(r13);				\
 	BEGIN_FTR_SECTION_NESTED(948)					\
-	ld	r9,area+EX_PPR(r13);					\
-	std	r9,HSTATE_PPR(r13);					\
+	ld	r10,area+EX_PPR(r13);					\
+	std	r10,HSTATE_PPR(r13);					\
 	END_FTR_SECTION_NESTED(CPU_FTR_HAS_PPR,CPU_FTR_HAS_PPR,948);	\
-	ld	r9,area+EX_R9(r13);					\
+	ld	r10,area+EX_R10(r13);					\
 	std	r12,HSTATE_SCRATCH0(r13);				\
-	li	r12,n;							\
+	sldi	r12,r9,32;						\
+	ori	r12,r12,(n);						\
+	ld	r9,area+EX_R9(r13);					\
 	b	kvmppc_interrupt;					\
 89:	mtocrf	0x80,r9;						\
 	ld	r9,area+EX_R9(r13);					\
+	ld	r10,area+EX_R10(r13);					\
 	b	kvmppc_skip_##h##interrupt
 
 #ifdef CONFIG_KVM_BOOK3S_64_HANDLER

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 9338a818e05c..11882aac8216 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1057,19 +1057,18 @@ hdec_soon:
 kvmppc_interrupt_hv:
 	/*
 	 * Register contents:
-	 * R12		= interrupt vector
+	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
-	 * guest CR, R12 saved in shadow VCPU SCRATCH1/0
+	 * guest R12 saved in shadow VCPU SCRATCH0
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
-	std	r9, HSTATE_SCRATCH2(r13)
-
+	std	r9, HSTATE_SCRATCH1(r13)
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
 #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE
 	cmpwi	r9, KVM_GUEST_MODE_GUEST
-	ld	r9, HSTATE_SCRATCH2(r13)
+	ld	r9, HSTATE_SCRATCH1(r13)
 	beq	kvmppc_interrupt_pr
 #endif
 	/* We're now back in the host but in guest MMU context */

@@ -1089,13 +1088,14 @@ kvmppc_interrupt_hv:
 	std	r6, VCPU_GPR(R6)(r9)
 	std	r7, VCPU_GPR(R7)(r9)
 	std	r8, VCPU_GPR(R8)(r9)
-	ld	r0, HSTATE_SCRATCH2(r13)
+	ld	r0, HSTATE_SCRATCH1(r13)
 	std	r0, VCPU_GPR(R9)(r9)
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)
 	ld	r3, HSTATE_SCRATCH0(r13)
-	lwz	r4, HSTATE_SCRATCH1(r13)
 	std	r3, VCPU_GPR(R12)(r9)
+	/* CR is in the high half of r12 */
+	srdi	r4, r12, 32
 	stw	r4, VCPU_CR(r9)
 BEGIN_FTR_SECTION
 	ld	r3, HSTATE_CFAR(r13)

@@ -1114,6 +1114,8 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 	mfspr	r11, SPRN_SRR1
 	std	r10, VCPU_SRR0(r9)
 	std	r11, VCPU_SRR1(r9)
+	/* trap is in the low half of r12, clear CR from the high half */
+	clrldi	r12, r12, 32
 	andi.	r0, r12, 2		/* need to read HSRR0/1? */
 	beq	1f
 	mfspr	r10, SPRN_HSRR0

diff --git a/arch/powerpc/kvm/book3s_segment.S b/arch/powerpc/kvm/book3s_segment.S
index ca8f174289bb..68e45080cf93 100644
--- a/arch/powerpc/kvm/book3s_segment.S
+++ b/arch/powerpc/kvm/book3s_segment.S

@@ -167,20 +167,31 @@ kvmppc_handler_trampoline_enter_end:
  *                                                                            *
  *****************************************************************************/
 
-.global kvmppc_handler_trampoline_exit
-kvmppc_handler_trampoline_exit:
-
 .global kvmppc_interrupt_pr
 kvmppc_interrupt_pr:
+	/* 64-bit entry. Register usage at this point:
+	 *
+	 * SPRG_SCRATCH0   = guest R13
+	 * R12             = (guest CR << 32) | exit handler id
+	 * R13             = PACA
+	 * HSTATE.SCRATCH0 = guest R12
+	 */
+#ifdef CONFIG_PPC64
+	/* Match 32-bit entry */
+	rotldi	r12, r12, 32		  /* Flip R12 halves for stw */
+	stw	r12, HSTATE_SCRATCH1(r13) /* CR is now in the low half */
+	srdi	r12, r12, 32		  /* shift trap into low half */
+#endif
 
+.global kvmppc_handler_trampoline_exit
+kvmppc_handler_trampoline_exit:
 	/* Register usage at this point:
 	 *
-	 * SPRG_SCRATCH0  = guest R13
-	 * R12            = exit handler id
-	 * R13            = shadow vcpu (32-bit) or PACA (64-bit)
+	 * SPRG_SCRATCH0   = guest R13
+	 * R12             = exit handler id
+	 * R13             = shadow vcpu (32-bit) or PACA (64-bit)
 	 * HSTATE.SCRATCH0 = guest R12
 	 * HSTATE.SCRATCH1 = guest CR
-	 *
 	 */
 
 	/* Save registers */

-- 
2.11.0

[PATCH 2/3] KVM: PPC: Book3S: Move 64-bit KVM interrupt handler out from alt section

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2016-12-21 18:29:57

A subsequent patch to make KVM handlers relocation-safe makes them
unusable from within alt section "else" cases (due to the way fixed
addresses are taken from within fixed section head code).

Stop open-coding the KVM handlers, and add them both as normal. A more
optimal fix may be to allow some level of alternate feature patching in
the exception macros themselves, but for now this will do.

The TRAMP_KVM handlers must be moved to the "virt" fixed section area
(name is arbitrary) in order to be closer to .text and avoid the dreaded
"relocation truncated to fit" error.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/head-64.h   | 2 +-
 arch/powerpc/kernel/exceptions-64s.S | 6 ++----
 2 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/head-64.h b/arch/powerpc/include/asm/head-64.h
index fca7033839a9..9bd81619d090 100644
--- a/arch/powerpc/include/asm/head-64.h
+++ b/arch/powerpc/include/asm/head-64.h

@@ -218,7 +218,7 @@ end_##sname:
 
 #ifdef CONFIG_KVM_BOOK3S_64_HANDLER
 #define TRAMP_KVM_BEGIN(name)						\
-	TRAMP_REAL_BEGIN(name)
+	TRAMP_VIRT_BEGIN(name)
 #else
 #define TRAMP_KVM_BEGIN(name)
 #endif

diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index d39d6118c6e9..89b4f122aec6 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S

@@ -717,13 +717,9 @@ hardware_interrupt_hv:
 	BEGIN_FTR_SECTION
 		_MASKABLE_EXCEPTION_PSERIES(0x500, hardware_interrupt_common,
 					    EXC_HV, SOFTEN_TEST_HV)
-do_kvm_H0x500:
-		KVM_HANDLER(PACA_EXGEN, EXC_HV, 0x502)
 	FTR_SECTION_ELSE
 		_MASKABLE_EXCEPTION_PSERIES(0x500, hardware_interrupt_common,
 					    EXC_STD, SOFTEN_TEST_PR)
-do_kvm_0x500:
-		KVM_HANDLER(PACA_EXGEN, EXC_STD, 0x500)
 	ALT_FTR_SECTION_END_IFSET(CPU_FTR_HVMODE | CPU_FTR_ARCH_206)
 EXC_REAL_END(hardware_interrupt, 0x500, 0x600)

@@ -737,6 +733,8 @@ hardware_interrupt_relon_hv:
 	ALT_FTR_SECTION_END_IFSET(CPU_FTR_HVMODE)
 EXC_VIRT_END(hardware_interrupt, 0x4500, 0x4600)
 
+TRAMP_KVM(PACA_EXGEN, 0x500)
+TRAMP_KVM_HV(PACA_EXGEN, 0x500)
 EXC_COMMON_ASYNC(hardware_interrupt_common, 0x500, do_IRQ)

-- 
2.11.0

[PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2016-12-21 18:30:00

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/exception-64s.h | 43 +++++++++++++++++++++++++++++---
 arch/powerpc/kernel/exceptions-64s.S     |  4 +--
 arch/powerpc/kvm/book3s_hv_rmhandlers.S  | 11 +++++---
 arch/powerpc/kvm/book3s_segment.S        |  7 ++++++
 4 files changed, 56 insertions(+), 9 deletions(-)

diff --git a/arch/powerpc/include/asm/exception-64s.h b/arch/powerpc/include/asm/exception-64s.h
index a02a268bde6b..e36c741e22d5 100644
--- a/arch/powerpc/include/asm/exception-64s.h
+++ b/arch/powerpc/include/asm/exception-64s.h

@@ -97,6 +97,11 @@
 	ld	reg,PACAKBASE(r13);					\
 	ori	reg,reg,(ABS_ADDR(label))@l;
 
+#define __LOAD_FAR_HANDLER(reg, label)					\
+	ld	reg,PACAKBASE(r13);					\
+	ori	reg,reg,(ABS_ADDR(label))@l;				\
+	addis	reg,reg,(ABS_ADDR(label))@h;
+
 /* Exception register prefixes */
 #define EXC_HV	H
 #define EXC_STD

@@ -227,12 +232,44 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	mtctr	reg;							\
 	bctr
 
+/*
+ * KVM requires >64K branches when branching from unrelocated code.
+ */
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	mfctr	r9;							\
+	std	r9,HSTATE_SCRATCH1(r13);				\
+	__LOAD_FAR_HANDLER(r9, label);					\
+	mtctr	r9;							\
+	ld	r9,area+EX_R9(r13);					\
+	bctr
+
+#define BRANCH_TO_KVM(reg, label)					\
+	__LOAD_FAR_HANDLER(reg, label);					\
+	mtctr	reg;							\
+	bctr
+
+#define BRANCH_LINK_TO_KVM(reg, label)					\
+	__LOAD_FAR_HANDLER(reg, label);					\
+	mtctr	reg;							\
+	bctrl
+
 #else
 #define BRANCH_TO_COMMON(reg, label)					\
 	b	label
 
+#define BRANCH_TO_KVM(reg, label)					\
+	b	label
+
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	ld	r9,area+EX_R9(r13);					\
+	b	label
+
+#define BRANCH_LINK_TO_KVM(reg, label)					\
+	bl	label
+
 #endif
 
+
 #define __KVM_HANDLER(area, h, n)					\
 	BEGIN_FTR_SECTION_NESTED(947)					\
 	ld	r10,area+EX_CFAR(r13);					\

@@ -246,8 +283,7 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r12,HSTATE_SCRATCH0(r13);				\
 	sldi	r12,r9,32;						\
 	ori	r12,r12,(n);						\
-	ld	r9,area+EX_R9(r13);					\
-	b	kvmppc_interrupt
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt)
 
 #define __KVM_HANDLER_SKIP(area, h, n)					\
 	cmpwi	r10,KVM_GUEST_MODE_SKIP;				\

@@ -260,8 +296,7 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r12,HSTATE_SCRATCH0(r13);				\
 	sldi	r12,r9,32;						\
 	ori	r12,r12,(n);						\
-	ld	r9,area+EX_R9(r13);					\
-	b	kvmppc_interrupt;					\
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt);			\
 89:	mtocrf	0x80,r9;						\
 	ld	r9,area+EX_R9(r13);					\
 	ld	r10,area+EX_R10(r13);					\

diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 89b4f122aec6..d40307cfeb39 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S

@@ -142,7 +142,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300)
 	lbz	r0,HSTATE_HWTHREAD_REQ(r13)
 	cmpwi	r0,0
 	beq	1f
-	b	kvm_start_guest
+	BRANCH_TO_KVM(r10, kvm_start_guest)
 1:
 #endif

@@ -977,7 +977,7 @@ TRAMP_REAL_BEGIN(hmi_exception_early)
 	EXCEPTION_PROLOG_COMMON_2(PACA_EXGEN)
 	EXCEPTION_PROLOG_COMMON_3(0xe60)
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	hmi_exception_realmode
+	BRANCH_LINK_TO_KVM(r4, hmi_exception_realmode)
 	/* Windup the stack. */
 	/* Move original HSRR0 and HSRR1 into the respective regs */
 	ld	r9,_MSR(r1)

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 11882aac8216..c18ce740452b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1060,15 +1060,16 @@ kvmppc_interrupt_hv:
 	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
 	 * guest R12 saved in shadow VCPU SCRATCH0
+	 * guest CTR saved in shadow VCPU SCRATCH1 if RELOCATABLE
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
-	std	r9, HSTATE_SCRATCH1(r13)
+	std	r9, HSTATE_SCRATCH2(r13)
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
 #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE
 	cmpwi	r9, KVM_GUEST_MODE_GUEST
-	ld	r9, HSTATE_SCRATCH1(r13)
+	ld	r9, HSTATE_SCRATCH2(r13)
 	beq	kvmppc_interrupt_pr
 #endif
 	/* We're now back in the host but in guest MMU context */

@@ -1088,7 +1089,7 @@ kvmppc_interrupt_hv:
 	std	r6, VCPU_GPR(R6)(r9)
 	std	r7, VCPU_GPR(R7)(r9)
 	std	r8, VCPU_GPR(R8)(r9)
-	ld	r0, HSTATE_SCRATCH1(r13)
+	ld	r0, HSTATE_SCRATCH2(r13)
 	std	r0, VCPU_GPR(R9)(r9)
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)

@@ -1151,7 +1152,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)
+#else
 	mfctr	r3
+#endif
 	mfxer	r4
 	std	r3, VCPU_CTR(r9)
 	std	r4, VCPU_XER(r9)

diff --git a/arch/powerpc/kvm/book3s_segment.S b/arch/powerpc/kvm/book3s_segment.S
index 68e45080cf93..2a2b96d53999 100644
--- a/arch/powerpc/kvm/book3s_segment.S
+++ b/arch/powerpc/kvm/book3s_segment.S

@@ -175,9 +175,16 @@ kvmppc_interrupt_pr:
 	 * R12             = (guest CR << 32) | exit handler id
 	 * R13             = PACA
 	 * HSTATE.SCRATCH0 = guest R12
+	 * HSTATE.SCRATCH1 = guest CTR if RELOCATABLE
 	 */
 #ifdef CONFIG_PPC64
 	/* Match 32-bit entry */
+#ifdef CONFIG_RELOCATABLE
+	std	r9, HSTATE_SCRATCH2(r13)
+	ld	r9, HSTATE_SCRATCH1(r13)
+	mtctr	r9
+	ld	r9, HSTATE_SCRATCH2(r13)
+#endif
 	rotldi	r12, r12, 32		  /* Flip R12 halves for stw */
 	stw	r12, HSTATE_SCRATCH1(r13) /* CR is now in the low half */
 	srdi	r12, r12, 32		  /* shift trap into low half */

-- 
2.11.0

Re: [PATCH 1/3] KVM: PPC: Book3S: Change interrupt call to reduce scratch space use on HV

From: Paul Mackerras <hidden>
Date: 2017-01-27 02:21:15

On Thu, Dec 22, 2016 at 04:29:25AM +1000, Nicholas Piggin wrote:

Change the calling convention to put the trap number together with
CR in two halves of r12, which frees up HSTATE_SCRATCH2 in the HV
handler.

The 64-bit PR handler entry translates the calling convention back
to match the previous call convention (i.e., shared with 32-bit), for
simplicity.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Acked-by: Paul Mackerras <redacted>

I notice that I forgot to add the code to save CFAR to the
__KVM_HANDLER_SKIP macro.  We should fix that.

Paul.

Re: [PATCH 2/3] KVM: PPC: Book3S: Move 64-bit KVM interrupt handler out from alt section

From: Paul Mackerras <hidden>
Date: 2017-01-27 02:50:23

On Thu, Dec 22, 2016 at 04:29:26AM +1000, Nicholas Piggin wrote:

A subsequent patch to make KVM handlers relocation-safe makes them
unusable from within alt section "else" cases (due to the way fixed
addresses are taken from within fixed section head code).

Stop open-coding the KVM handlers, and add them both as normal. A more
optimal fix may be to allow some level of alternate feature patching in
the exception macros themselves, but for now this will do.

The TRAMP_KVM handlers must be moved to the "virt" fixed section area
(name is arbitrary) in order to be closer to .text and avoid the dreaded
"relocation truncated to fit" error.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Acked-by: Paul Mackerras <redacted>

Re: [PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Paul Mackerras <hidden>
Date: 2017-01-27 02:50:23

On Thu, Dec 22, 2016 at 04:29:27AM +1000, Nicholas Piggin wrote:

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Looks OK to me.

I have a slight quibble about the naming of the "BRANCH_LINK_TO_KVM"
macro because neither its definition nor the place where it's used
have anything to do with KVM as far as I can see.  That needn't stop
the patch going in, though.

Acked-by: Paul Mackerras <redacted>

Re: [PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2017-01-27 04:00:48

On Fri, 27 Jan 2017 13:50:19 +1100
Paul Mackerras [off-list ref] wrote:

On Thu, Dec 22, 2016 at 04:29:27AM +1000, Nicholas Piggin wrote:

quoted

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Looks OK to me.

I have a slight quibble about the naming of the "BRANCH_LINK_TO_KVM"
macro because neither its definition nor the place where it's used
have anything to do with KVM as far as I can see.  That needn't stop
the patch going in, though.

Acked-by: Paul Mackerras <redacted>

No that makes sense, good point. Here's an updated patch 3 with the
hmi handler removed and some comments slightly updated (no code
changes otherwise).

I'll send the hmi relocation fix as another patch.

Thanks,
Nick

--

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/exception-64s.h | 45 +++++++++++++++++++++++++++++---
 arch/powerpc/kernel/exceptions-64s.S     |  2 +-
 arch/powerpc/kvm/book3s_hv_rmhandlers.S  | 11 +++++---
 arch/powerpc/kvm/book3s_segment.S        |  7 +++++
 4 files changed, 57 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/include/asm/exception-64s.h b/arch/powerpc/include/asm/exception-64s.h
index a02a268bde6b..9a5dbfb2d9f2 100644
--- a/arch/powerpc/include/asm/exception-64s.h
+++ b/arch/powerpc/include/asm/exception-64s.h

@@ -97,6 +97,15 @@
 	ld	reg,PACAKBASE(r13);					\
 	ori	reg,reg,(ABS_ADDR(label))@l;
 
+/*
+ * Branches from unrelocated code (e.g., interrupts) to labels outside
+ * head-y require >64K offsets.
+ */
+#define __LOAD_FAR_HANDLER(reg, label)					\
+	ld	reg,PACAKBASE(r13);					\
+	ori	reg,reg,(ABS_ADDR(label))@l;				\
+	addis	reg,reg,(ABS_ADDR(label))@h;
+
 /* Exception register prefixes */
 #define EXC_HV	H
 #define EXC_STD

@@ -227,12 +236,40 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	mtctr	reg;							\
 	bctr
 
+/*
+ * KVM requires __LOAD_FAR_HANDLER.
+ *
+ * __BRANCH_TO_KVM_EXIT branches are also a special case because they
+ * explicitly use r9 then reload it from PACA before branching. Hence
+ * the double-underscore.
+ */
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	mfctr	r9;							\
+	std	r9,HSTATE_SCRATCH1(r13);				\
+	__LOAD_FAR_HANDLER(r9, label);					\
+	mtctr	r9;							\
+	ld	r9,area+EX_R9(r13);					\
+	bctr
+
+#define BRANCH_TO_KVM(reg, label)					\
+	__LOAD_FAR_HANDLER(reg, label);					\
+	mtctr	reg;							\
+	bctr
+
 #else
 #define BRANCH_TO_COMMON(reg, label)					\
 	b	label
 
+#define BRANCH_TO_KVM(reg, label)					\
+	b	label
+
+#define __BRANCH_TO_KVM_EXIT(area, label)				\
+	ld	r9,area+EX_R9(r13);					\
+	b	label
+
 #endif
 
+
 #define __KVM_HANDLER(area, h, n)					\
 	BEGIN_FTR_SECTION_NESTED(947)					\
 	ld	r10,area+EX_CFAR(r13);					\

@@ -246,8 +283,8 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r12,HSTATE_SCRATCH0(r13);				\
 	sldi	r12,r9,32;						\
 	ori	r12,r12,(n);						\
-	ld	r9,area+EX_R9(r13);					\
-	b	kvmppc_interrupt
+	/* This reloads r9 before branching to kvmppc_interrupt */	\
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt)
 
 #define __KVM_HANDLER_SKIP(area, h, n)					\
 	cmpwi	r10,KVM_GUEST_MODE_SKIP;				\

@@ -260,8 +297,8 @@ END_FTR_SECTION_NESTED(ftr,ftr,943)
 	std	r12,HSTATE_SCRATCH0(r13);				\
 	sldi	r12,r9,32;						\
 	ori	r12,r12,(n);						\
-	ld	r9,area+EX_R9(r13);					\
-	b	kvmppc_interrupt;					\
+	/* This reloads r9 before branching to kvmppc_interrupt */	\
+	__BRANCH_TO_KVM_EXIT(area, kvmppc_interrupt);			\
 89:	mtocrf	0x80,r9;						\
 	ld	r9,area+EX_R9(r13);					\
 	ld	r10,area+EX_R10(r13);					\

diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 89b4f122aec6..65a2559eeb7f 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S

@@ -142,7 +142,7 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_300)
 	lbz	r0,HSTATE_HWTHREAD_REQ(r13)
 	cmpwi	r0,0
 	beq	1f
-	b	kvm_start_guest
+	BRANCH_TO_KVM(r10, kvm_start_guest)
 1:
 #endif

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 11882aac8216..c18ce740452b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1060,15 +1060,16 @@ kvmppc_interrupt_hv:
 	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
 	 * guest R12 saved in shadow VCPU SCRATCH0
+	 * guest CTR saved in shadow VCPU SCRATCH1 if RELOCATABLE
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
-	std	r9, HSTATE_SCRATCH1(r13)
+	std	r9, HSTATE_SCRATCH2(r13)
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
 #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE
 	cmpwi	r9, KVM_GUEST_MODE_GUEST
-	ld	r9, HSTATE_SCRATCH1(r13)
+	ld	r9, HSTATE_SCRATCH2(r13)
 	beq	kvmppc_interrupt_pr
 #endif
 	/* We're now back in the host but in guest MMU context */

@@ -1088,7 +1089,7 @@ kvmppc_interrupt_hv:
 	std	r6, VCPU_GPR(R6)(r9)
 	std	r7, VCPU_GPR(R7)(r9)
 	std	r8, VCPU_GPR(R8)(r9)
-	ld	r0, HSTATE_SCRATCH1(r13)
+	ld	r0, HSTATE_SCRATCH2(r13)
 	std	r0, VCPU_GPR(R9)(r9)
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)

@@ -1151,7 +1152,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)
+#else
 	mfctr	r3
+#endif
 	mfxer	r4
 	std	r3, VCPU_CTR(r9)
 	std	r4, VCPU_XER(r9)

diff --git a/arch/powerpc/kvm/book3s_segment.S b/arch/powerpc/kvm/book3s_segment.S
index 68e45080cf93..2a2b96d53999 100644
--- a/arch/powerpc/kvm/book3s_segment.S
+++ b/arch/powerpc/kvm/book3s_segment.S

@@ -175,9 +175,16 @@ kvmppc_interrupt_pr:
 	 * R12             = (guest CR << 32) | exit handler id
 	 * R13             = PACA
 	 * HSTATE.SCRATCH0 = guest R12
+	 * HSTATE.SCRATCH1 = guest CTR if RELOCATABLE
 	 */
 #ifdef CONFIG_PPC64
 	/* Match 32-bit entry */
+#ifdef CONFIG_RELOCATABLE
+	std	r9, HSTATE_SCRATCH2(r13)
+	ld	r9, HSTATE_SCRATCH1(r13)
+	mtctr	r9
+	ld	r9, HSTATE_SCRATCH2(r13)
+#endif
 	rotldi	r12, r12, 32		  /* Flip R12 halves for stw */
 	stw	r12, HSTATE_SCRATCH1(r13) /* CR is now in the low half */
 	srdi	r12, r12, 32		  /* shift trap into low half */

-- 
2.11.0

Re: [PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Paul Mackerras <hidden>
Date: 2017-01-31 08:02:01

On Thu, Dec 22, 2016 at 04:29:27AM +1000, Nicholas Piggin wrote:

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Unfortunately, this patch causes guests to crash in various ways when
running under a CONFIG_RELOCATABLE kernel.  See below...

quoted hunk

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 11882aac8216..c18ce740452b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1060,15 +1060,16 @@ kvmppc_interrupt_hv:
 	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
 	 * guest R12 saved in shadow VCPU SCRATCH0
+	 * guest CTR saved in shadow VCPU SCRATCH1 if RELOCATABLE
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
-	std	r9, HSTATE_SCRATCH1(r13)
+	std	r9, HSTATE_SCRATCH2(r13)
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
 #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE
 	cmpwi	r9, KVM_GUEST_MODE_GUEST
-	ld	r9, HSTATE_SCRATCH1(r13)
+	ld	r9, HSTATE_SCRATCH2(r13)
 	beq	kvmppc_interrupt_pr
 #endif
 	/* We're now back in the host but in guest MMU context */

@@ -1088,7 +1089,7 @@ kvmppc_interrupt_hv:
 	std	r6, VCPU_GPR(R6)(r9)
 	std	r7, VCPU_GPR(R7)(r9)
 	std	r8, VCPU_GPR(R8)(r9)
-	ld	r0, HSTATE_SCRATCH1(r13)
+	ld	r0, HSTATE_SCRATCH2(r13)
 	std	r0, VCPU_GPR(R9)(r9)
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)

@@ -1151,7 +1152,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)

The problem is here.  Depending on what sort of interrupt it is, we
can end up going back to the guest via fast_guest_return, which
doesn't reload CTR.  Adding

	mtctr	r3

at this point makes things work.

Paul.

Re: [PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Michael Ellerman <mpe@ellerman.id.au>
Date: 2017-01-31 08:18:02

Paul Mackerras [off-list ref] writes:

On Thu, Dec 22, 2016 at 04:29:27AM +1000, Nicholas Piggin wrote:

quoted

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 11882aac8216..c18ce740452b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1151,7 +1152,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)

The problem is here.  Depending on what sort of interrupt it is, we
can end up going back to the guest via fast_guest_return, which
doesn't reload CTR.  Adding

	mtctr	r3

at this point makes things work.

I folded this in.

cheers

Re: [PATCH 3/3] KVM: PPC: Book3S: 64-bit CONFIG_RELOCATABLE support for interrupts

From: Nicholas Piggin <npiggin@gmail.com>
Date: 2017-01-31 10:22:06

On Tue, 31 Jan 2017 19:01:52 +1100
Paul Mackerras [off-list ref] wrote:

On Thu, Dec 22, 2016 at 04:29:27AM +1000, Nicholas Piggin wrote:

quoted

64-bit Book3S exception handlers must find the dynamic kernel base
to add to the target address when branching beyond __end_interrupts,
in order to support kernel running at non-0 physical address.

Support this in KVM by branching with CTR, similarly to regular
interrupt handlers. The guest CTR saved in HSTATE_SCRATCH1 and
restored after the branch.

Without this, the host kernel hangs and crashes randomly when it is
running at a non-0 address and a KVM guest is started.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>

Unfortunately, this patch causes guests to crash in various ways when
running under a CONFIG_RELOCATABLE kernel.  See below...

quoted

diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 11882aac8216..c18ce740452b 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S

@@ -1060,15 +1060,16 @@ kvmppc_interrupt_hv:
 	 * R12		= (guest CR << 32) | interrupt vector
 	 * R13		= PACA
 	 * guest R12 saved in shadow VCPU SCRATCH0
+	 * guest CTR saved in shadow VCPU SCRATCH1 if RELOCATABLE
 	 * guest R13 saved in SPRN_SCRATCH0
 	 */
-	std	r9, HSTATE_SCRATCH1(r13)
+	std	r9, HSTATE_SCRATCH2(r13)
 	lbz	r9, HSTATE_IN_GUEST(r13)
 	cmpwi	r9, KVM_GUEST_MODE_HOST_HV
 	beq	kvmppc_bad_host_intr
 #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE
 	cmpwi	r9, KVM_GUEST_MODE_GUEST
-	ld	r9, HSTATE_SCRATCH1(r13)
+	ld	r9, HSTATE_SCRATCH2(r13)
 	beq	kvmppc_interrupt_pr
 #endif
 	/* We're now back in the host but in guest MMU context */

@@ -1088,7 +1089,7 @@ kvmppc_interrupt_hv:
 	std	r6, VCPU_GPR(R6)(r9)
 	std	r7, VCPU_GPR(R7)(r9)
 	std	r8, VCPU_GPR(R8)(r9)
-	ld	r0, HSTATE_SCRATCH1(r13)
+	ld	r0, HSTATE_SCRATCH2(r13)
 	std	r0, VCPU_GPR(R9)(r9)
 	std	r10, VCPU_GPR(R10)(r9)
 	std	r11, VCPU_GPR(R11)(r9)

@@ -1151,7 +1152,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 11:	stw	r3,VCPU_HEIR(r9)
 
 	/* these are volatile across C function calls */
+#ifdef CONFIG_RELOCATABLE
+	ld	r3, HSTATE_SCRATCH1(r13)

The problem is here.  Depending on what sort of interrupt it is, we
can end up going back to the guest via fast_guest_return, which
doesn't reload CTR.  Adding

	mtctr	r3

at this point makes things work.

Thanks for tracking it down. Not enough testing on my end, I'm
sorry.

Thanks,
Nick

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help