[PATCH v4.9.y 00/42] arm64 spectre patches

[PATCH v4.9.y 00/42] arm64 spectre patches

[Mark Rutland]

Hi Greg,

These patches backport arm64 spectre mitigations to v4.9.y (based on 4.9.93).

I've done this backport from scratch, following the patches in the arm64 kpti
branch [1] (which contrary to its name, also has spectre patches intended for
backporting).

To the best of my understanding the code is correct, in the context of the
v4.9.y kernel, and I've tested the series on arm64 hardware available to me,
using firmware with SMCCC 1.1 and branch predictor mitigations.

I've Cc'd the usual suspects for testing. For anyone willing to test this, I've
pushed this out to the stable/4.9.y/spectre branch of my kernel.org repo [2].

Note that I've left off the Qualcomm Falkor patches, as it was pointed out that
necessary errata workarounds for Falkor were never backported, and thus I
assume that no-one is using v4.9.y on Falkor parts. Shanker, do you have any
idea whether people are using v4.9.y on Falkor?

Thanks,
Mark.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=kpti
[2] git://git.kernel.org/pub/scm/linux/kernel/git/mark/linux.git stable/4.9.y/spectre

Catalin Marinas (1):
  arm64: Factor out TTBR0_EL1 post-update workaround into a specific asm
    macro

James Morse (1):
  arm64: cpufeature: __this_cpu_has_cap() shouldn't stop early

Jayachandran C (1):
  arm64: Branch predictor hardening for Cavium ThunderX2

Laura Abbott (1):
  mm: Introduce lm_alias

Marc Zyngier (20):
  arm64: Move post_ttbr_update_workaround to C code
  arm64: Move BP hardening to check_and_switch_context
  arm64: KVM: Use per-CPU vector when BP hardening is enabled
  arm64: cpu_errata: Allow an erratum to be match for all revisions of a
    core
  arm64: KVM: Increment PC after handling an SMC trap
  arm/arm64: KVM: Consolidate the PSCI include files
  arm/arm64: KVM: Add PSCI_VERSION helper
  arm/arm64: KVM: Add smccc accessors to PSCI code
  arm/arm64: KVM: Implement PSCI 1.0 support
  arm/arm64: KVM: Advertise SMCCC v1.1
  arm64: KVM: Make PSCI_VERSION a fast path
  arm/arm64: KVM: Turn kvm_psci_version into a static inline
  arm64: KVM: Report SMCCC_ARCH_WORKAROUND_1 BP hardening support
  arm64: KVM: Add SMCCC_ARCH_WORKAROUND_1 fast handling
  firmware/psci: Expose PSCI conduit
  firmware/psci: Expose SMCCC version through psci_ops
  arm/arm64: smccc: Make function identifiers an unsigned quantity
  arm/arm64: smccc: Implement SMCCC v1.1 inline primitive
  arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support
  arm64: Kill PSCI_GET_VERSION as a variant-2 workaround

Robin Murphy (3):
  arm64: Implement array_index_mask_nospec()
  arm64: Make USER_DS an inclusive limit
  arm64: Use pointer masking to limit uaccess speculation

Suzuki K Poulose (1):
  arm64: Run enable method for errata work arounds on late CPUs

Will Deacon (13):
  arm64: barrier: Add CSDB macros to control data-value prediction
  arm64: entry: Ensure branch through syscall table is bounded under
    speculation
  arm64: uaccess: Prevent speculative use of the current addr_limit
  arm64: uaccess: Don't bother eliding access_ok checks in __{get,
    put}_user
  arm64: uaccess: Mask __user pointers for __arch_{clear, copy_*}_user
  arm64: futex: Mask __user pointers prior to dereference
  arm64: cpufeature: Pass capability structure to ->enable callback
  drivers/firmware: Expose psci_get_version through psci_ops structure
  arm64: Add skeleton to harden the branch predictor against aliasing
    attacks
  arm64: entry: Apply BP hardening for high-priority synchronous
    exceptions
  arm64: entry: Apply BP hardening for suspicious interrupts from EL0
  arm64: cputype: Add missing MIDR values for Cortex-A72 and Cortex-A75
  arm64: Implement branch predictor hardening for affected Cortex-A CPUs

Yury Norov (1):
  arm64: move TASK_* definitions to <asm/processor.h>

 arch/arm/include/asm/kvm_host.h    |    6 +
 arch/arm/include/asm/kvm_mmu.h     |   10 +
 arch/arm/include/asm/kvm_psci.h    |   27 -
 arch/arm/kvm/arm.c                 |   11 +-
 arch/arm/kvm/handle_exit.c         |    4 +-
 arch/arm/kvm/psci.c                |  143 ++-
 arch/arm64/Kconfig                 |   17 +
 arch/arm64/crypto/sha256-core.S    | 2061 ++++++++++++++++++++++++++++++++++++
 arch/arm64/crypto/sha512-core.S    | 1085 +++++++++++++++++++
 arch/arm64/include/asm/assembler.h |   19 +
 arch/arm64/include/asm/barrier.h   |   23 +
 arch/arm64/include/asm/cpucaps.h   |    3 +-
 arch/arm64/include/asm/cputype.h   |    6 +
 arch/arm64/include/asm/futex.h     |    9 +-
 arch/arm64/include/asm/kvm_host.h  |    5 +
 arch/arm64/include/asm/kvm_mmu.h   |   38 +
 arch/arm64/include/asm/kvm_psci.h  |   27 -
 arch/arm64/include/asm/memory.h    |   15 -
 arch/arm64/include/asm/mmu.h       |   39 +
 arch/arm64/include/asm/processor.h |   24 +
 arch/arm64/include/asm/sysreg.h    |    2 +
 arch/arm64/include/asm/uaccess.h   |  153 ++-
 arch/arm64/kernel/Makefile         |    4 +
 arch/arm64/kernel/arm64ksyms.c     |    4 +-
 arch/arm64/kernel/bpi.S            |   75 ++
 arch/arm64/kernel/cpu_errata.c     |  189 +++-
 arch/arm64/kernel/cpufeature.c     |   10 +-
 arch/arm64/kernel/entry.S          |   25 +-
 arch/arm64/kvm/handle_exit.c       |   16 +-
 arch/arm64/kvm/hyp/hyp-entry.S     |   20 +-
 arch/arm64/kvm/hyp/switch.c        |    5 +-
 arch/arm64/lib/clear_user.S        |    6 +-
 arch/arm64/lib/copy_in_user.S      |    4 +-
 arch/arm64/mm/context.c            |   12 +
 arch/arm64/mm/fault.c              |   34 +-
 arch/arm64/mm/proc.S               |    7 +-
 drivers/firmware/psci.c            |   57 +-
 include/kvm/arm_psci.h             |   51 +
 include/linux/arm-smccc.h          |  165 ++-
 include/linux/mm.h                 |    4 +
 include/linux/psci.h               |   14 +
 include/uapi/linux/psci.h          |    3 +
 42 files changed, 4237 insertions(+), 195 deletions(-)
 delete mode 100644 arch/arm/include/asm/kvm_psci.h
 create mode 100644 arch/arm64/crypto/sha256-core.S
 create mode 100644 arch/arm64/crypto/sha512-core.S
 delete mode 100644 arch/arm64/include/asm/kvm_psci.h
 create mode 100644 arch/arm64/kernel/bpi.S
 create mode 100644 include/kvm/arm_psci.h

-- 
2.11.0

[PATCH v4.9.y 01/42] arm64: barrier: Add CSDB macros to control data-value prediction

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 669474e772b952b14f4de4845a1558fd4c0414a4 upstream.

For CPUs capable of data value prediction, CSDB waits for any outstanding
predictions to architecturally resolve before allowing speculative execution
to continue. Provide macros to expose it to the arch code.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com [v4.9 backport]
---
 arch/arm64/include/asm/assembler.h | 7 +++++++
 arch/arm64/include/asm/barrier.h   | 2 ++
 2 files changed, 9 insertions(+)

diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h
index 7193bf97b8da..b223b1b4d5cd 100644
--- a/arch/arm64/include/asm/assembler.h
+++ b/arch/arm64/include/asm/assembler.h
@@ -87,6 +87,13 @@
 	.endm
 
 /*
+ * Value prediction barrier
+ */
+	.macro	csdb
+	hint	#20
+	.endm
+
+/*
  * NOP sequence
  */
 	.macro	nops, num
diff --git a/arch/arm64/include/asm/barrier.h b/arch/arm64/include/asm/barrier.h
index 0fe7e43b7fbc..c68fdc5707ed 100644
--- a/arch/arm64/include/asm/barrier.h
+++ b/arch/arm64/include/asm/barrier.h
@@ -31,6 +31,8 @@
 #define dmb(opt)	asm volatile("dmb " #opt : : : "memory")
 #define dsb(opt)	asm volatile("dsb " #opt : : : "memory")
 
+#define csdb()		asm volatile("hint #20" : : : "memory")
+
 #define mb()		dsb(sy)
 #define rmb()		dsb(ld)
 #define wmb()		dsb(st)
-- 
2.11.0

[PATCH v4.9.y 02/42] arm64: Implement array_index_mask_nospec()

[Mark Rutland]

From: Robin Murphy <robin.murphy@arm.com>

commit 022620eed3d0bc4bf2027326f599f5ad71c2ea3f upstream.

Provide an optimised, assembly implementation of array_index_mask_nospec()
for arm64 so that the compiler is not in a position to transform the code
in ways which affect its ability to inhibit speculation (e.g. by introducing
conditional branches).

This is similar to the sequence used by x86, modulo architectural differences
in the carry/borrow flags.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/barrier.h | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/arch/arm64/include/asm/barrier.h b/arch/arm64/include/asm/barrier.h
index c68fdc5707ed..0b0755c961ac 100644
--- a/arch/arm64/include/asm/barrier.h
+++ b/arch/arm64/include/asm/barrier.h
@@ -40,6 +40,27 @@
 #define dma_rmb()	dmb(oshld)
 #define dma_wmb()	dmb(oshst)
 
+/*
+ * Generate a mask for array_index__nospec() that is ~0UL when 0 <= idx < sz
+ * and 0 otherwise.
+ */
+#define array_index_mask_nospec array_index_mask_nospec
+static inline unsigned long array_index_mask_nospec(unsigned long idx,
+						    unsigned long sz)
+{
+	unsigned long mask;
+
+	asm volatile(
+	"	cmp	%1, %2\n"
+	"	sbc	%0, xzr, xzr\n"
+	: "=r" (mask)
+	: "r" (idx), "Ir" (sz)
+	: "cc");
+
+	csdb();
+	return mask;
+}
+
 #define __smp_mb()	dmb(ish)
 #define __smp_rmb()	dmb(ishld)
 #define __smp_wmb()	dmb(ishst)
-- 
2.11.0

[PATCH v4.9.y 03/42] arm64: move TASK_* definitions to <asm/processor.h>

[Mark Rutland]

From: Yury Norov <ynorov@caviumnetworks.com>

commit eef94a3d09aab437c8c254de942d8b1aa76455e2 upstream.

ILP32 series [1] introduces the dependency on <asm/is_compat.h> for
TASK_SIZE macro. Which in turn requires <asm/thread_info.h>, and
<asm/thread_info.h> include <asm/memory.h>, giving a circular dependency,
because TASK_SIZE is currently located in <asm/memory.h>.

In other architectures, TASK_SIZE is defined in <asm/processor.h>, and
moving TASK_SIZE there fixes the problem.

Discussion: https://patchwork.kernel.org/patch/9929107/

[1] https://github.com/norov/linux/tree/ilp32-next

CC: Will Deacon <will.deacon@arm.com>
CC: Laura Abbott <labbott@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
[v4.9: necessary for making USER_DS an inclusive limit]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/memory.h    | 15 ---------------
 arch/arm64/include/asm/processor.h | 21 +++++++++++++++++++++
 arch/arm64/kernel/entry.S          |  1 +
 3 files changed, 22 insertions(+), 15 deletions(-)

diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
index 5e3faba689e0..ba917be5565a 100644
--- a/arch/arm64/include/asm/memory.h
+++ b/arch/arm64/include/asm/memory.h
@@ -60,8 +60,6 @@
  * KIMAGE_VADDR - the virtual address of the start of the kernel image
  * VA_BITS - the maximum number of bits for virtual addresses.
  * VA_START - the first kernel virtual address.
- * TASK_SIZE - the maximum size of a user space task.
- * TASK_UNMAPPED_BASE - the lower boundary of the mmap VM area.
  */
 #define VA_BITS			(CONFIG_ARM64_VA_BITS)
 #define VA_START		(UL(0xffffffffffffffff) - \
@@ -76,19 +74,6 @@
 #define PCI_IO_END		(VMEMMAP_START - SZ_2M)
 #define PCI_IO_START		(PCI_IO_END - PCI_IO_SIZE)
 #define FIXADDR_TOP		(PCI_IO_START - SZ_2M)
-#define TASK_SIZE_64		(UL(1) << VA_BITS)
-
-#ifdef CONFIG_COMPAT
-#define TASK_SIZE_32		UL(0x100000000)
-#define TASK_SIZE		(test_thread_flag(TIF_32BIT) ? \
-				TASK_SIZE_32 : TASK_SIZE_64)
-#define TASK_SIZE_OF(tsk)	(test_tsk_thread_flag(tsk, TIF_32BIT) ? \
-				TASK_SIZE_32 : TASK_SIZE_64)
-#else
-#define TASK_SIZE		TASK_SIZE_64
-#endif /* CONFIG_COMPAT */
-
-#define TASK_UNMAPPED_BASE	(PAGE_ALIGN(TASK_SIZE / 4))
 
 #define KERNEL_START      _text
 #define KERNEL_END        _end
diff --git a/arch/arm64/include/asm/processor.h b/arch/arm64/include/asm/processor.h
index 60e34824e18c..4258f4d594b5 100644
--- a/arch/arm64/include/asm/processor.h
+++ b/arch/arm64/include/asm/processor.h
@@ -19,6 +19,10 @@
 #ifndef __ASM_PROCESSOR_H
 #define __ASM_PROCESSOR_H
 
+#define TASK_SIZE_64		(UL(1) << VA_BITS)
+
+#ifndef __ASSEMBLY__
+
 /*
  * Default implementation of macro that returns current
  * instruction pointer ("program counter").
@@ -37,6 +41,22 @@
 #include <asm/ptrace.h>
 #include <asm/types.h>
 
+/*
+ * TASK_SIZE - the maximum size of a user space task.
+ * TASK_UNMAPPED_BASE - the lower boundary of the mmap VM area.
+ */
+#ifdef CONFIG_COMPAT
+#define TASK_SIZE_32		UL(0x100000000)
+#define TASK_SIZE		(test_thread_flag(TIF_32BIT) ? \
+				TASK_SIZE_32 : TASK_SIZE_64)
+#define TASK_SIZE_OF(tsk)	(test_tsk_thread_flag(tsk, TIF_32BIT) ? \
+				TASK_SIZE_32 : TASK_SIZE_64)
+#else
+#define TASK_SIZE		TASK_SIZE_64
+#endif /* CONFIG_COMPAT */
+
+#define TASK_UNMAPPED_BASE	(PAGE_ALIGN(TASK_SIZE / 4))
+
 #define STACK_TOP_MAX		TASK_SIZE_64
 #ifdef CONFIG_COMPAT
 #define AARCH32_VECTORS_BASE	0xffff0000
@@ -192,4 +212,5 @@ int cpu_enable_pan(void *__unused);
 int cpu_enable_uao(void *__unused);
 int cpu_enable_cache_maint_trap(void *__unused);
 
+#endif /* __ASSEMBLY__ */
 #endif /* __ASM_PROCESSOR_H */
diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 8d1600b18562..e10e9f2e48ad 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -30,6 +30,7 @@
 #include <asm/irq.h>
 #include <asm/memory.h>
 #include <asm/mmu.h>
+#include <asm/processor.h>
 #include <asm/thread_info.h>
 #include <asm/asm-uaccess.h>
 #include <asm/unistd.h>
-- 
2.11.0

[PATCH v4.9.y 04/42] arm64: Make USER_DS an inclusive limit

[Mark Rutland]

From: Robin Murphy <robin.murphy@arm.com>

commit 51369e398d0d33e8f524314e672b07e8cf870e79 upstream.

Currently, USER_DS represents an exclusive limit while KERNEL_DS is
inclusive. In order to do some clever trickery for speculation-safe
masking, we need them both to behave equivalently - there aren't enough
bits to make KERNEL_DS exclusive, so we have precisely one option. This
also happens to correct a longstanding false negative for a range
ending on the very top byte of kernel memory.

Mark Rutland points out that we've actually got the semantics of
addresses vs. segments muddled up in most of the places we need to
amend, so shuffle the {USER,KERNEL}_DS definitions around such that we
can correct those properly instead of just pasting "-1"s everywhere.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: avoid dependence on TTBR0 SW PAN and THREAD_INFO_IN_TASK_STRUCT]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/processor.h |  3 +++
 arch/arm64/include/asm/uaccess.h   | 46 ++++++++++++++++++++++----------------
 arch/arm64/kernel/entry.S          |  4 ++--
 arch/arm64/mm/fault.c              |  2 +-
 4 files changed, 33 insertions(+), 22 deletions(-)

diff --git a/arch/arm64/include/asm/processor.h b/arch/arm64/include/asm/processor.h
index 4258f4d594b5..5917147af0c4 100644
--- a/arch/arm64/include/asm/processor.h
+++ b/arch/arm64/include/asm/processor.h
@@ -21,6 +21,9 @@
 
 #define TASK_SIZE_64		(UL(1) << VA_BITS)
 
+#define KERNEL_DS	UL(-1)
+#define USER_DS		(TASK_SIZE_64 - 1)
+
 #ifndef __ASSEMBLY__
 
 /*
diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 811cf16a65f9..0f05ba51531c 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h
@@ -28,6 +28,7 @@
 
 #include <asm/alternative.h>
 #include <asm/cpufeature.h>
+#include <asm/processor.h>
 #include <asm/ptrace.h>
 #include <asm/sysreg.h>
 #include <asm/errno.h>
@@ -59,10 +60,7 @@ struct exception_table_entry
 
 extern int fixup_exception(struct pt_regs *regs);
 
-#define KERNEL_DS	(-1UL)
 #define get_ds()	(KERNEL_DS)
-
-#define USER_DS		TASK_SIZE_64
 #define get_fs()	(current_thread_info()->addr_limit)
 
 static inline void set_fs(mm_segment_t fs)
@@ -87,22 +85,32 @@ static inline void set_fs(mm_segment_t fs)
  * Returns 1 if the range is valid, 0 otherwise.
  *
  * This is equivalent to the following test:
- * (u65)addr + (u65)size <= current->addr_limit
- *
- * This needs 65-bit arithmetic.
+ * (u65)addr + (u65)size <= (u65)current->addr_limit + 1
  */
-#define __range_ok(addr, size)						\
-({									\
-	unsigned long __addr = (unsigned long __force)(addr);		\
-	unsigned long flag, roksum;					\
-	__chk_user_ptr(addr);						\
-	asm("adds %1, %1, %3; ccmp %1, %4, #2, cc; cset %0, ls"		\
-		: "=&r" (flag), "=&r" (roksum)				\
-		: "1" (__addr), "Ir" (size),				\
-		  "r" (current_thread_info()->addr_limit)		\
-		: "cc");						\
-	flag;								\
-})
+static inline unsigned long __range_ok(unsigned long addr, unsigned long size)
+{
+	unsigned long limit = current_thread_info()->addr_limit;
+
+	__chk_user_ptr(addr);
+	asm volatile(
+	// A + B <= C + 1 for all A,B,C, in four easy steps:
+	// 1: X = A + B; X' = X % 2^64
+	"	adds	%0, %0, %2\n"
+	// 2: Set C = 0 if X > 2^64, to guarantee X' > C in step 4
+	"	csel	%1, xzr, %1, hi\n"
+	// 3: Set X' = ~0 if X >= 2^64. For X == 2^64, this decrements X'
+	//    to compensate for the carry flag being set in step 4. For
+	//    X > 2^64, X' merely has to remain nonzero, which it does.
+	"	csinv	%0, %0, xzr, cc\n"
+	// 4: For X < 2^64, this gives us X' - C - 1 <= 0, where the -1
+	//    comes from the carry in being clear. Otherwise, we are
+	//    testing X' - C == 0, subject to the previous adjustments.
+	"	sbcs	xzr, %0, %1\n"
+	"	cset	%0, ls\n"
+	: "+r" (addr), "+r" (limit) : "Ir" (size) : "cc");
+
+	return addr;
+}
 
 /*
  * When dealing with data aborts, watchpoints, or instruction traps we may end
@@ -111,7 +119,7 @@ static inline void set_fs(mm_segment_t fs)
  */
 #define untagged_addr(addr)		sign_extend64(addr, 55)
 
-#define access_ok(type, addr, size)	__range_ok(addr, size)
+#define access_ok(type, addr, size)	__range_ok((unsigned long)(addr), size)
 #define user_addr_max			get_fs
 
 #define _ASM_EXTABLE(from, to)						\
diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index e10e9f2e48ad..3f1cd7a795bb 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -126,10 +126,10 @@ alternative_else_nop_endif
 	.else
 	add	x21, sp, #S_FRAME_SIZE
 	get_thread_info tsk
-	/* Save the task's original addr_limit and set USER_DS (TASK_SIZE_64) */
+	/* Save the task's original addr_limit and set USER_DS */
 	ldr	x20, [tsk, #TI_ADDR_LIMIT]
 	str	x20, [sp, #S_ORIG_ADDR_LIMIT]
-	mov	x20, #TASK_SIZE_64
+	mov	x20, #USER_DS
 	str	x20, [tsk, #TI_ADDR_LIMIT]
 	/* No need to reset PSTATE.UAO, hardware's already set it to 0 for us */
 	.endif /* \el == 0 */
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index 403fe9e57135..4df70c9fd762 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -332,7 +332,7 @@ static int __kprobes do_page_fault(unsigned long addr, unsigned int esr,
 		mm_flags |= FAULT_FLAG_WRITE;
 	}
 
-	if (is_permission_fault(esr) && (addr < USER_DS)) {
+	if (is_permission_fault(esr) && (addr < TASK_SIZE)) {
 		/* regs->orig_addr_limit may be 0 if we entered from EL0 */
 		if (regs->orig_addr_limit == KERNEL_DS)
 			die("Accessing user space memory with fs=KERNEL_DS", regs, esr);
-- 
2.11.0

[PATCH v4.9.y 05/42] arm64: Use pointer masking to limit uaccess speculation

[Mark Rutland]

From: Robin Murphy <robin.murphy@arm.com>

commit 4d8efc2d5ee4c9ccfeb29ee8afd47a8660d0c0ce upstream.

Similarly to x86, mitigate speculation past an access_ok() check by
masking the pointer against the address limit before use.

Even if we don't expect speculative writes per se, it is plausible that
a CPU may still speculate at least as far as fetching a cache line for
writing, hence we also harden put_user() and clear_user() for peace of
mind.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/uaccess.h | 26 +++++++++++++++++++++++---
 1 file changed, 23 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 0f05ba51531c..a9db051b967f 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h
@@ -129,6 +129,26 @@ static inline unsigned long __range_ok(unsigned long addr, unsigned long size)
 	"	.popsection\n"
 
 /*
+ * Sanitise a uaccess pointer such that it becomes NULL if above the
+ * current addr_limit.
+ */
+#define uaccess_mask_ptr(ptr) (__typeof__(ptr))__uaccess_mask_ptr(ptr)
+static inline void __user *__uaccess_mask_ptr(const void __user *ptr)
+{
+	void __user *safe_ptr;
+
+	asm volatile(
+	"	bics	xzr, %1, %2\n"
+	"	csel	%0, %1, xzr, eq\n"
+	: "=&r" (safe_ptr)
+	: "r" (ptr), "r" (current_thread_info()->addr_limit)
+	: "cc");
+
+	csdb();
+	return safe_ptr;
+}
+
+/*
  * The "__xxx" versions of the user access functions do not verify the address
  * space - it must have been done previously with a separate "access_ok()"
  * call.
@@ -202,7 +222,7 @@ do {									\
 	__typeof__(*(ptr)) __user *__p = (ptr);				\
 	might_fault();							\
 	access_ok(VERIFY_READ, __p, sizeof(*__p)) ?			\
-		__get_user((x), __p) :					\
+		__p = uaccess_mask_ptr(__p), __get_user((x), __p) :	\
 		((x) = 0, -EFAULT);					\
 })
 
@@ -270,7 +290,7 @@ do {									\
 	__typeof__(*(ptr)) __user *__p = (ptr);				\
 	might_fault();							\
 	access_ok(VERIFY_WRITE, __p, sizeof(*__p)) ?			\
-		__put_user((x), __p) :					\
+		__p = uaccess_mask_ptr(__p), __put_user((x), __p) :	\
 		-EFAULT;						\
 })
 
@@ -331,7 +351,7 @@ static inline unsigned long __must_check copy_in_user(void __user *to, const voi
 static inline unsigned long __must_check clear_user(void __user *to, unsigned long n)
 {
 	if (access_ok(VERIFY_WRITE, to, n))
-		n = __clear_user(to, n);
+		n = __clear_user(__uaccess_mask_ptr(to), n);
 	return n;
 }
 
-- 
2.11.0

[PATCH v4.9.y 06/42] arm64: entry: Ensure branch through syscall table is bounded under speculation

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 6314d90e64936c584f300a52ef173603fb2461b5 upstream.

In a similar manner to array_index_mask_nospec, this patch introduces an
assembly macro (mask_nospec64) which can be used to bound a value under
speculation. This macro is then used to ensure that the indirect branch
through the syscall table is bounded under speculation, with out-of-range
addresses speculating as calls to sys_io_setup (0).

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: use existing scno & sc_nr definitions]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/assembler.h | 11 +++++++++++
 arch/arm64/kernel/entry.S          |  1 +
 2 files changed, 12 insertions(+)

diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h
index b223b1b4d5cd..540c24f74837 100644
--- a/arch/arm64/include/asm/assembler.h
+++ b/arch/arm64/include/asm/assembler.h
@@ -94,6 +94,17 @@
 	.endm
 
 /*
+ * Sanitise a 64-bit bounded index wrt speculation, returning zero if out
+ * of bounds.
+ */
+	.macro	mask_nospec64, idx, limit, tmp
+	sub	\tmp, \idx, \limit
+	bic	\tmp, \tmp, \idx
+	and	\idx, \idx, \tmp, asr #63
+	csdb
+	.endm
+
+/*
  * NOP sequence
  */
 	.macro	nops, num
diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 3f1cd7a795bb..c154b1608405 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -795,6 +795,7 @@ el0_svc_naked:					// compat entry point
 	b.ne	__sys_trace
 	cmp     scno, sc_nr                     // check upper syscall limit
 	b.hs	ni_sys
+	mask_nospec64 scno, sc_nr, x19	// enforce bounds for syscall number
 	ldr	x16, [stbl, scno, lsl #3]	// address in the syscall table
 	blr	x16				// call sys_* routine
 	b	ret_fast_syscall
-- 
2.11.0

[PATCH v4.9.y 07/42] arm64: uaccess: Prevent speculative use of the current addr_limit

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit c2f0ad4fc089cff81cef6a13d04b399980ecbfcc upstream.

A mispredicted conditional call to set_fs could result in the wrong
addr_limit being forwarded under speculation to a subsequent access_ok
check, potentially forming part of a spectre-v1 attack using uaccess
routines.

This patch prevents this forwarding from taking place, but putting heavy
barriers in set_fs after writing the addr_limit.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/uaccess.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index a9db051b967f..900397e73fa6 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h
@@ -68,6 +68,13 @@ static inline void set_fs(mm_segment_t fs)
 	current_thread_info()->addr_limit = fs;
 
 	/*
+	 * Prevent a mispredicted conditional call to set_fs from forwarding
+	 * the wrong address limit to access_ok under speculation.
+	 */
+	dsb(nsh);
+	isb();
+
+	/*
 	 * Enable/disable UAO so that copy_to_user() etc can access
 	 * kernel memory with the unprivileged instructions.
 	 */
-- 
2.11.0

[PATCH v4.9.y 08/42] arm64: uaccess: Don't bother eliding access_ok checks in __{get, put}_user

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 84624087dd7e3b482b7b11c170ebc1f329b3a218 upstream.

access_ok isn't an expensive operation once the addr_limit for the current
thread has been loaded into the cache. Given that the initial access_ok
check preceding a sequence of __{get,put}_user operations will take
the brunt of the miss, we can make the __* variants identical to the
full-fat versions, which brings with it the benefits of address masking.

The likely cost in these sequences will be from toggling PAN/UAO, which
we can address later by implementing the *_unsafe versions.

Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/uaccess.h | 62 +++++++++++++++++++++++-----------------
 1 file changed, 36 insertions(+), 26 deletions(-)

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 900397e73fa6..0d4330c7d58d 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h
@@ -209,30 +209,35 @@ do {									\
 			CONFIG_ARM64_PAN));				\
 } while (0)
 
-#define __get_user(x, ptr)						\
+#define __get_user_check(x, ptr, err)					\
 ({									\
-	int __gu_err = 0;						\
-	__get_user_err((x), (ptr), __gu_err);				\
-	__gu_err;							\
+	__typeof__(*(ptr)) __user *__p = (ptr);				\
+	might_fault();							\
+	if (access_ok(VERIFY_READ, __p, sizeof(*__p))) {		\
+		__p = uaccess_mask_ptr(__p);				\
+		__get_user_err((x), __p, (err));			\
+	} else {							\
+		(x) = 0; (err) = -EFAULT;				\
+	}								\
 })
 
 #define __get_user_error(x, ptr, err)					\
 ({									\
-	__get_user_err((x), (ptr), (err));				\
+	__get_user_check((x), (ptr), (err));				\
 	(void)0;							\
 })
 
-#define __get_user_unaligned __get_user
-
-#define get_user(x, ptr)						\
+#define __get_user(x, ptr)						\
 ({									\
-	__typeof__(*(ptr)) __user *__p = (ptr);				\
-	might_fault();							\
-	access_ok(VERIFY_READ, __p, sizeof(*__p)) ?			\
-		__p = uaccess_mask_ptr(__p), __get_user((x), __p) :	\
-		((x) = 0, -EFAULT);					\
+	int __gu_err = 0;						\
+	__get_user_check((x), (ptr), __gu_err);				\
+	__gu_err;							\
 })
 
+#define __get_user_unaligned __get_user
+
+#define get_user	__get_user
+
 #define __put_user_asm(instr, alt_instr, reg, x, addr, err, feature)	\
 	asm volatile(							\
 	"1:"ALTERNATIVE(instr "     " reg "1, [%2]\n",			\
@@ -277,30 +282,35 @@ do {									\
 			CONFIG_ARM64_PAN));				\
 } while (0)
 
-#define __put_user(x, ptr)						\
+#define __put_user_check(x, ptr, err)					\
 ({									\
-	int __pu_err = 0;						\
-	__put_user_err((x), (ptr), __pu_err);				\
-	__pu_err;							\
+	__typeof__(*(ptr)) __user *__p = (ptr);				\
+	might_fault();							\
+	if (access_ok(VERIFY_WRITE, __p, sizeof(*__p))) {		\
+		__p = uaccess_mask_ptr(__p);				\
+		__put_user_err((x), __p, (err));			\
+	} else	{							\
+		(err) = -EFAULT;					\
+	}								\
 })
 
 #define __put_user_error(x, ptr, err)					\
 ({									\
-	__put_user_err((x), (ptr), (err));				\
+	__put_user_check((x), (ptr), (err));				\
 	(void)0;							\
 })
 
-#define __put_user_unaligned __put_user
-
-#define put_user(x, ptr)						\
+#define __put_user(x, ptr)						\
 ({									\
-	__typeof__(*(ptr)) __user *__p = (ptr);				\
-	might_fault();							\
-	access_ok(VERIFY_WRITE, __p, sizeof(*__p)) ?			\
-		__p = uaccess_mask_ptr(__p), __put_user((x), __p) :	\
-		-EFAULT;						\
+	int __pu_err = 0;						\
+	__put_user_check((x), (ptr), __pu_err);				\
+	__pu_err;							\
 })
 
+#define __put_user_unaligned __put_user
+
+#define put_user	__put_user
+
 extern unsigned long __must_check __arch_copy_from_user(void *to, const void __user *from, unsigned long n);
 extern unsigned long __must_check __arch_copy_to_user(void __user *to, const void *from, unsigned long n);
 extern unsigned long __must_check __copy_in_user(void __user *to, const void __user *from, unsigned long n);
-- 
2.11.0

[PATCH v4.9.y 09/42] arm64: uaccess: Mask __user pointers for __arch_{clear, copy_*}_user

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit f71c2ffcb20dd8626880747557014bb9a61eb90e upstream.

Like we've done for get_user and put_user, ensure that user pointers
are masked before invoking the underlying __arch_{clear,copy_*}_user
operations.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: fixup for v4.9-style uaccess primitives]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/uaccess.h | 18 ++++++++++--------
 arch/arm64/kernel/arm64ksyms.c   |  4 ++--
 arch/arm64/lib/clear_user.S      |  6 +++---
 arch/arm64/lib/copy_in_user.S    |  4 ++--
 4 files changed, 17 insertions(+), 15 deletions(-)

diff --git a/arch/arm64/include/asm/uaccess.h b/arch/arm64/include/asm/uaccess.h
index 0d4330c7d58d..1d047d6c421b 100644
--- a/arch/arm64/include/asm/uaccess.h
+++ b/arch/arm64/include/asm/uaccess.h
@@ -313,21 +313,20 @@ do {									\
 
 extern unsigned long __must_check __arch_copy_from_user(void *to, const void __user *from, unsigned long n);
 extern unsigned long __must_check __arch_copy_to_user(void __user *to, const void *from, unsigned long n);
-extern unsigned long __must_check __copy_in_user(void __user *to, const void __user *from, unsigned long n);
-extern unsigned long __must_check __clear_user(void __user *addr, unsigned long n);
+extern unsigned long __must_check __arch_copy_in_user(void __user *to, const void __user *from, unsigned long n);
 
 static inline unsigned long __must_check __copy_from_user(void *to, const void __user *from, unsigned long n)
 {
 	kasan_check_write(to, n);
 	check_object_size(to, n, false);
-	return __arch_copy_from_user(to, from, n);
+	return __arch_copy_from_user(to, __uaccess_mask_ptr(from), n);
 }
 
 static inline unsigned long __must_check __copy_to_user(void __user *to, const void *from, unsigned long n)
 {
 	kasan_check_read(from, n);
 	check_object_size(from, n, true);
-	return __arch_copy_to_user(to, from, n);
+	return __arch_copy_to_user(__uaccess_mask_ptr(to), from, n);
 }
 
 static inline unsigned long __must_check copy_from_user(void *to, const void __user *from, unsigned long n)
@@ -355,22 +354,25 @@ static inline unsigned long __must_check copy_to_user(void __user *to, const voi
 	return n;
 }
 
-static inline unsigned long __must_check copy_in_user(void __user *to, const void __user *from, unsigned long n)
+static inline unsigned long __must_check __copy_in_user(void __user *to, const void __user *from, unsigned long n)
 {
 	if (access_ok(VERIFY_READ, from, n) && access_ok(VERIFY_WRITE, to, n))
-		n = __copy_in_user(to, from, n);
+		n = __arch_copy_in_user(__uaccess_mask_ptr(to), __uaccess_mask_ptr(from), n);
 	return n;
 }
+#define copy_in_user __copy_in_user
 
 #define __copy_to_user_inatomic __copy_to_user
 #define __copy_from_user_inatomic __copy_from_user
 
-static inline unsigned long __must_check clear_user(void __user *to, unsigned long n)
+extern unsigned long __must_check __arch_clear_user(void __user *to, unsigned long n);
+static inline unsigned long __must_check __clear_user(void __user *to, unsigned long n)
 {
 	if (access_ok(VERIFY_WRITE, to, n))
-		n = __clear_user(__uaccess_mask_ptr(to), n);
+		n = __arch_clear_user(__uaccess_mask_ptr(to), n);
 	return n;
 }
+#define clear_user	__clear_user
 
 extern long strncpy_from_user(char *dest, const char __user *src, long count);
 
diff --git a/arch/arm64/kernel/arm64ksyms.c b/arch/arm64/kernel/arm64ksyms.c
index e9c4dc9e0ada..66be504edb6c 100644
--- a/arch/arm64/kernel/arm64ksyms.c
+++ b/arch/arm64/kernel/arm64ksyms.c
@@ -37,8 +37,8 @@ EXPORT_SYMBOL(clear_page);
 	/* user mem (segment) */
 EXPORT_SYMBOL(__arch_copy_from_user);
 EXPORT_SYMBOL(__arch_copy_to_user);
-EXPORT_SYMBOL(__clear_user);
-EXPORT_SYMBOL(__copy_in_user);
+EXPORT_SYMBOL(__arch_clear_user);
+EXPORT_SYMBOL(__arch_copy_in_user);
 
 	/* physical memory */
 EXPORT_SYMBOL(memstart_addr);
diff --git a/arch/arm64/lib/clear_user.S b/arch/arm64/lib/clear_user.S
index 5d1cad3ce6d6..efbf610eaf4e 100644
--- a/arch/arm64/lib/clear_user.S
+++ b/arch/arm64/lib/clear_user.S
@@ -24,7 +24,7 @@
 
 	.text
 
-/* Prototype: int __clear_user(void *addr, size_t sz)
+/* Prototype: int __arch_clear_user(void *addr, size_t sz)
  * Purpose  : clear some user memory
  * Params   : addr - user memory address to clear
  *          : sz   - number of bytes to clear
@@ -32,7 +32,7 @@
  *
  * Alignment fixed up by hardware.
  */
-ENTRY(__clear_user)
+ENTRY(__arch_clear_user)
 ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_ALT_PAN_NOT_UAO, \
 	    CONFIG_ARM64_PAN)
 	mov	x2, x1			// save the size for fixup return
@@ -57,7 +57,7 @@ uao_user_alternative 9f, strb, sttrb, wzr, x0, 0
 ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_ALT_PAN_NOT_UAO, \
 	    CONFIG_ARM64_PAN)
 	ret
-ENDPROC(__clear_user)
+ENDPROC(__arch_clear_user)
 
 	.section .fixup,"ax"
 	.align	2
diff --git a/arch/arm64/lib/copy_in_user.S b/arch/arm64/lib/copy_in_user.S
index f7292dd08c84..841bf8f7fab7 100644
--- a/arch/arm64/lib/copy_in_user.S
+++ b/arch/arm64/lib/copy_in_user.S
@@ -67,7 +67,7 @@
 	.endm
 
 end	.req	x5
-ENTRY(__copy_in_user)
+ENTRY(__arch_copy_in_user)
 ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(0)), ARM64_ALT_PAN_NOT_UAO, \
 	    CONFIG_ARM64_PAN)
 	add	end, x0, x2
@@ -76,7 +76,7 @@ ALTERNATIVE("nop", __stringify(SET_PSTATE_PAN(1)), ARM64_ALT_PAN_NOT_UAO, \
 	    CONFIG_ARM64_PAN)
 	mov	x0, #0
 	ret
-ENDPROC(__copy_in_user)
+ENDPROC(__arch_copy_in_user)
 
 	.section .fixup,"ax"
 	.align	2
-- 
2.11.0

[PATCH v4.9.y 10/42] arm64: futex: Mask __user pointers prior to dereference

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 91b2d3442f6a44dce875670d702af22737ad5eff upstream.

The arm64 futex code has some explicit dereferencing of user pointers
where performing atomic operations in response to a futex command. This
patch uses masking to limit any speculative futex operations to within
the user address space.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/futex.h | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/include/asm/futex.h b/arch/arm64/include/asm/futex.h
index f2585cdd32c2..1d123dd01ee0 100644
--- a/arch/arm64/include/asm/futex.h
+++ b/arch/arm64/include/asm/futex.h
@@ -51,13 +51,14 @@
 	: "memory")
 
 static inline int
-futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
+futex_atomic_op_inuser (int encoded_op, u32 __user *_uaddr)
 {
 	int op = (encoded_op >> 28) & 7;
 	int cmp = (encoded_op >> 24) & 15;
 	int oparg = (encoded_op << 8) >> 20;
 	int cmparg = (encoded_op << 20) >> 20;
 	int oldval = 0, ret, tmp;
+	u32 __user *uaddr = __uaccess_mask_ptr(_uaddr);
 
 	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
 		oparg = 1 << oparg;
@@ -109,15 +110,17 @@ futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
 }
 
 static inline int
-futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
+futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *_uaddr,
 			      u32 oldval, u32 newval)
 {
 	int ret = 0;
 	u32 val, tmp;
+	u32 __user *uaddr;
 
-	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
+	if (!access_ok(VERIFY_WRITE, _uaddr, sizeof(u32)))
 		return -EFAULT;
 
+	uaddr = __uaccess_mask_ptr(_uaddr);
 	asm volatile("// futex_atomic_cmpxchg_inatomic\n"
 ALTERNATIVE("nop", SET_PSTATE_PAN(0), ARM64_HAS_PAN, CONFIG_ARM64_PAN)
 "	prfm	pstl1strm, %2\n"
-- 
2.11.0

Re: [PATCH v4.9.y 10/42] arm64: futex: Mask __user pointers prior to dereference

[Greg KH]

On Thu, Apr 12, 2018 at 12:11:06PM +0100, Mark Rutland wrote:
> From: Will Deacon <will.deacon@arm.com>
> 
> commit 91b2d3442f6a44dce875670d702af22737ad5eff upstream.
> 
> The arm64 futex code has some explicit dereferencing of user pointers
> where performing atomic operations in response to a futex command. This
> patch uses masking to limit any speculative futex operations to within
> the user address space.
> 
> Signed-off-by: Will Deacon <will.deacon@arm.com>
> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
> Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
> Tested-by: Greg Hackmann <ghackmann@google.com>
> ---
>  arch/arm64/include/asm/futex.h | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/futex.h b/arch/arm64/include/asm/futex.h
> index f2585cdd32c2..1d123dd01ee0 100644
> --- a/arch/arm64/include/asm/futex.h
> +++ b/arch/arm64/include/asm/futex.h
> @@ -51,13 +51,14 @@
>  	: "memory")
>  
>  static inline int
> -futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
> +futex_atomic_op_inuser (int encoded_op, u32 __user *_uaddr)
>  {
>  	int op = (encoded_op >> 28) & 7;
>  	int cmp = (encoded_op >> 24) & 15;
>  	int oparg = (encoded_op << 8) >> 20;
>  	int cmparg = (encoded_op << 20) >> 20;
>  	int oldval = 0, ret, tmp;
> +	u32 __user *uaddr = __uaccess_mask_ptr(_uaddr);
>  
>  	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
>  		oparg = 1 << oparg;
> @@ -109,15 +110,17 @@ futex_atomic_op_inuser (int encoded_op, u32 __user *uaddr)
>  }
>  
>  static inline int
> -futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
> +futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *_uaddr,
>  			      u32 oldval, u32 newval)
>  {
>  	int ret = 0;
>  	u32 val, tmp;
> +	u32 __user *uaddr;
>  
> -	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
> +	if (!access_ok(VERIFY_WRITE, _uaddr, sizeof(u32)))
>  		return -EFAULT;
>  
> +	uaddr = __uaccess_mask_ptr(_uaddr);
>  	asm volatile("// futex_atomic_cmpxchg_inatomic\n"
>  ALTERNATIVE("nop", SET_PSTATE_PAN(0), ARM64_HAS_PAN, CONFIG_ARM64_PAN)
>  "	prfm	pstl1strm, %2\n"

This patch doesn't apply at all as it conflicts with commit
d7c5f8c815466fc00785bbff20f25b39643abe01 which was commit 5f16a046f8e1
("arm64: futex: Fix undefined behaviour with FUTEX_OP_OPARG_SHIFT
usage") upstream.

Any chance you can provide a correct backport of this?

thanks,

greg k-h

Re: [PATCH v4.9.y 10/42] arm64: futex: Mask __user pointers prior to dereference

[Mark Rutland]

On Tue, Apr 17, 2018 at 02:10:03PM +0200, Greg KH wrote:
> This patch doesn't apply at all as it conflicts with commit
> d7c5f8c815466fc00785bbff20f25b39643abe01 which was commit 5f16a046f8e1
> ("arm64: futex: Fix undefined behaviour with FUTEX_OP_OPARG_SHIFT
> usage") upstream.
> 
> Any chance you can provide a correct backport of this?

The below is rebased atop of the conflicting patch (I based it on
v4.9.94). Luckily it's just a trivial conflict in the function
prototype.

Is this the right way to resend this?

Thanks,
Mark.

---->8----
>From 4ff73a6f286a8438529462391eca262b6772e9c1 Mon Sep 17 00:00:00 2001
From: Will Deacon <will.deacon@arm.com>
Date: Mon, 5 Feb 2018 15:34:24 +0000
Subject: [PATCH] arm64: futex: Mask __user pointers prior to dereference

commit 91b2d3442f6a44dce875670d702af22737ad5eff upstream.

The arm64 futex code has some explicit dereferencing of user pointers
where performing atomic operations in response to a futex command. This
patch uses masking to limit any speculative futex operations to within
the user address space.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/futex.h | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/include/asm/futex.h b/arch/arm64/include/asm/futex.h
index 20dcb196b240..4e5f36a804b4 100644
--- a/arch/arm64/include/asm/futex.h
+++ b/arch/arm64/include/asm/futex.h
@@ -51,13 +51,14 @@
 	: "memory")
 
 static inline int
-futex_atomic_op_inuser(unsigned int encoded_op, u32 __user *uaddr)
+futex_atomic_op_inuser(unsigned int encoded_op, u32 __user *_uaddr)
 {
 	int op = (encoded_op >> 28) & 7;
 	int cmp = (encoded_op >> 24) & 15;
 	int oparg = (int)(encoded_op << 8) >> 20;
 	int cmparg = (int)(encoded_op << 20) >> 20;
 	int oldval = 0, ret, tmp;
+	u32 __user *uaddr = __uaccess_mask_ptr(_uaddr);
 
 	if (encoded_op & (FUTEX_OP_OPARG_SHIFT << 28))
 		oparg = 1U << (oparg & 0x1f);
@@ -109,15 +110,17 @@ futex_atomic_op_inuser(unsigned int encoded_op, u32 __user *uaddr)
 }
 
 static inline int
-futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
+futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *_uaddr,
 			      u32 oldval, u32 newval)
 {
 	int ret = 0;
 	u32 val, tmp;
+	u32 __user *uaddr;
 
-	if (!access_ok(VERIFY_WRITE, uaddr, sizeof(u32)))
+	if (!access_ok(VERIFY_WRITE, _uaddr, sizeof(u32)))
 		return -EFAULT;
 
+	uaddr = __uaccess_mask_ptr(_uaddr);
 	asm volatile("// futex_atomic_cmpxchg_inatomic\n"
 ALTERNATIVE("nop", SET_PSTATE_PAN(0), ARM64_HAS_PAN, CONFIG_ARM64_PAN)
 "	prfm	pstl1strm, %2\n"
-- 
2.11.0

Re: [PATCH v4.9.y 10/42] arm64: futex: Mask __user pointers prior to dereference

[Greg KH]

On Wed, Apr 18, 2018 at 11:56:36AM +0100, Mark Rutland wrote:
> On Tue, Apr 17, 2018 at 02:10:03PM +0200, Greg KH wrote:
> > This patch doesn't apply at all as it conflicts with commit
> > d7c5f8c815466fc00785bbff20f25b39643abe01 which was commit 5f16a046f8e1
> > ("arm64: futex: Fix undefined behaviour with FUTEX_OP_OPARG_SHIFT
> > usage") upstream.
> > 
> > Any chance you can provide a correct backport of this?
> 
> The below is rebased atop of the conflicting patch (I based it on
> v4.9.94). Luckily it's just a trivial conflict in the function
> prototype.
> 
> Is this the right way to resend this?

That works, now queued up, thanks.

greg k-h

[PATCH v4.9.y 11/42] arm64: cpufeature: __this_cpu_has_cap() shouldn't stop early

[Mark Rutland]

From: James Morse <james.morse@arm.com>

commit edf298cfce47ab7279d03b5203ae2ef3a58e49db upstream.

this_cpu_has_cap() tests caps->desc not caps->matches, so it stops
walking the list when it finds a 'silent' feature, instead of
walking to the end of the list.

Prior to v4.6's 644c2ae198412 ("arm64: cpufeature: Test 'matches' pointer
to find the end of the list") we always tested desc to find the end of
a capability list. This was changed for dubious things like PAN_NOT_UAO.
v4.7's e3661b128e53e ("arm64: Allow a capability to be checked on
single CPU") added this_cpu_has_cap() using the old desc style test.

CC: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/cpufeature.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 5056fc597ae9..cf000fd694fa 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1024,9 +1024,8 @@ static bool __this_cpu_has_cap(const struct arm64_cpu_capabilities *cap_array,
 	if (WARN_ON(preemptible()))
 		return false;
 
-	for (caps = cap_array; caps->desc; caps++)
+	for (caps = cap_array; caps->matches; caps++)
 		if (caps->capability == cap &&
-		    caps->matches &&
 		    caps->matches(caps, SCOPE_LOCAL_CPU))
 			return true;
 	return false;
-- 
2.11.0

[PATCH v4.9.y 12/42] arm64: Run enable method for errata work arounds on late CPUs

[Mark Rutland]

From: Suzuki K Poulose <suzuki.poulose@arm.com>

commit 55b35d070c2534dfb714b883f3c3ae05d02032da upstream.

When a CPU is brought up after we have finalised the system
wide capabilities (i.e, features and errata), we make sure the
new CPU doesn't need a new errata work around which has not been
detected already. However we don't run enable() method on the new
CPU for the errata work arounds already detected. This could
cause the new CPU running without potential work arounds.
It is upto the "enable()" method to decide if this CPU should
do something about the errata.

Fixes: commit 6a6efbb45b7d95c84 ("arm64: Verify CPU errata work arounds on hotplugged CPU")
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Andre Przywara <andre.przywara@arm.com>
Cc: Dave Martin <dave.martin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/cpu_errata.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index b75e917aac46..b866df79bb83 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -143,15 +143,18 @@ void verify_local_cpu_errata_workarounds(void)
 {
 	const struct arm64_cpu_capabilities *caps = arm64_errata;
 
-	for (; caps->matches; caps++)
-		if (!cpus_have_cap(caps->capability) &&
-			caps->matches(caps, SCOPE_LOCAL_CPU)) {
+	for (; caps->matches; caps++) {
+		if (cpus_have_cap(caps->capability)) {
+			if (caps->enable)
+				caps->enable((void *)caps);
+		} else if (caps->matches(caps, SCOPE_LOCAL_CPU)) {
 			pr_crit("CPU%d: Requires work around for %s, not detected"
 					" at boot time\n",
 				smp_processor_id(),
 				caps->desc ? : "an erratum");
 			cpu_die_early();
 		}
+	}
 }
 
 void update_cpu_errata_workarounds(void)
-- 
2.11.0

[PATCH v4.9.y 13/42] arm64: cpufeature: Pass capability structure to ->enable callback

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 0a0d111d40fd1dc588cc590fab6b55d86ddc71d3 upstream.

In order to invoke the CPU capability ->matches callback from the ->enable
callback for applying local-CPU workarounds, we need a handle on the
capability structure.

This patch passes a pointer to the capability structure to the ->enable
callback.

Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/cpufeature.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index cf000fd694fa..692f107bb46e 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1058,7 +1058,7 @@ void __init enable_cpu_capabilities(const struct arm64_cpu_capabilities *caps)
 			 * uses an IPI, giving us a PSTATE that disappears when
 			 * we return.
 			 */
-			stop_machine(caps->enable, NULL, cpu_online_mask);
+			stop_machine(caps->enable, (void *)caps, cpu_online_mask);
 }
 
 /*
@@ -1115,7 +1115,7 @@ verify_local_cpu_features(const struct arm64_cpu_capabilities *caps_list)
 			cpu_die_early();
 		}
 		if (caps->enable)
-			caps->enable(NULL);
+			caps->enable((void *)caps);
 	}
 }
 
-- 
2.11.0

[PATCH v4.9.y 14/42] drivers/firmware: Expose psci_get_version through psci_ops structure

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit d68e3ba5303f7e1099f51fdcd155f5263da8569b upstream.

Entry into recent versions of ARM Trusted Firmware will invalidate the CPU
branch predictor state in order to protect against aliasing attacks.

This patch exposes the PSCI "VERSION" function via psci_ops, so that it
can be invoked outside of the PSCI driver where necessary.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 drivers/firmware/psci.c | 2 ++
 include/linux/psci.h    | 1 +
 2 files changed, 3 insertions(+)

diff --git a/drivers/firmware/psci.c b/drivers/firmware/psci.c
index 8263429e21b8..9a3ce76304b6 100644
--- a/drivers/firmware/psci.c
+++ b/drivers/firmware/psci.c
@@ -496,6 +496,8 @@ static void __init psci_init_migrate(void)
 static void __init psci_0_2_set_functions(void)
 {
 	pr_info("Using standard PSCI v0.2 function IDs\n");
+	psci_ops.get_version = psci_get_version;
+
 	psci_function_id[PSCI_FN_CPU_SUSPEND] =
 					PSCI_FN_NATIVE(0_2, CPU_SUSPEND);
 	psci_ops.cpu_suspend = psci_cpu_suspend;
diff --git a/include/linux/psci.h b/include/linux/psci.h
index bdea1cb5e1db..6306ab10af18 100644
--- a/include/linux/psci.h
+++ b/include/linux/psci.h
@@ -26,6 +26,7 @@ int psci_cpu_init_idle(unsigned int cpu);
 int psci_cpu_suspend_enter(unsigned long index);
 
 struct psci_operations {
+	u32 (*get_version)(void);
 	int (*cpu_suspend)(u32 state, unsigned long entry_point);
 	int (*cpu_off)(u32 state);
 	int (*cpu_on)(unsigned long cpuid, unsigned long entry_point);
-- 
2.11.0

[PATCH v4.9.y 15/42] arm64: Factor out TTBR0_EL1 post-update workaround into a specific asm macro

[Mark Rutland]

From: Catalin Marinas <catalin.marinas@arm.com>

commit f33bcf03e6079668da6bf4eec4a7dcf9289131d0 upstream.

This patch takes the errata workaround code out of cpu_do_switch_mm into
a dedicated post_ttbr0_update_workaround macro which will be reused in a
subsequent patch.

Cc: Will Deacon <will.deacon@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Kees Cook <keescook@chromium.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/assembler.h | 14 ++++++++++++++
 arch/arm64/mm/proc.S               |  6 +-----
 2 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h
index 540c24f74837..c9bd4ad4db82 100644
--- a/arch/arm64/include/asm/assembler.h
+++ b/arch/arm64/include/asm/assembler.h
@@ -434,4 +434,18 @@ alternative_endif
 	.macro	pte_to_phys, phys, pte
 	and	\phys, \pte, #(((1 << (48 - PAGE_SHIFT)) - 1) << PAGE_SHIFT)
 	.endm
+
+/*
+ * Errata workaround post TTBR0_EL1 update.
+ */
+	.macro	post_ttbr0_update_workaround
+#ifdef CONFIG_CAVIUM_ERRATUM_27456
+alternative_if ARM64_WORKAROUND_CAVIUM_27456
+	ic	iallu
+	dsb	nsh
+	isb
+alternative_else_nop_endif
+#endif
+	.endm
+
 #endif	/* __ASM_ASSEMBLER_H */
diff --git a/arch/arm64/mm/proc.S b/arch/arm64/mm/proc.S
index c07d9cc057e6..135a698ce946 100644
--- a/arch/arm64/mm/proc.S
+++ b/arch/arm64/mm/proc.S
@@ -139,11 +139,7 @@ ENTRY(cpu_do_switch_mm)
 	isb
 	msr	ttbr0_el1, x0			// now update TTBR0
 	isb
-alternative_if ARM64_WORKAROUND_CAVIUM_27456
-	ic	iallu
-	dsb	nsh
-	isb
-alternative_else_nop_endif
+	post_ttbr0_update_workaround
 	ret
 ENDPROC(cpu_do_switch_mm)
 
-- 
2.11.0

[PATCH v4.9.y 16/42] arm64: Move post_ttbr_update_workaround to C code

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 95e3de3590e3f2358bb13f013911bc1bfa5d3f53 upstream.

We will soon need to invoke a CPU-specific function pointer after changing
page tables, so move post_ttbr_update_workaround out into C code to make
this possible.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/assembler.h | 13 -------------
 arch/arm64/mm/context.c            |  9 +++++++++
 arch/arm64/mm/proc.S               |  3 +--
 3 files changed, 10 insertions(+), 15 deletions(-)

diff --git a/arch/arm64/include/asm/assembler.h b/arch/arm64/include/asm/assembler.h
index c9bd4ad4db82..e60375ce0dd2 100644
--- a/arch/arm64/include/asm/assembler.h
+++ b/arch/arm64/include/asm/assembler.h
@@ -435,17 +435,4 @@ alternative_endif
 	and	\phys, \pte, #(((1 << (48 - PAGE_SHIFT)) - 1) << PAGE_SHIFT)
 	.endm
 
-/*
- * Errata workaround post TTBR0_EL1 update.
- */
-	.macro	post_ttbr0_update_workaround
-#ifdef CONFIG_CAVIUM_ERRATUM_27456
-alternative_if ARM64_WORKAROUND_CAVIUM_27456
-	ic	iallu
-	dsb	nsh
-	isb
-alternative_else_nop_endif
-#endif
-	.endm
-
 #endif	/* __ASM_ASSEMBLER_H */
diff --git a/arch/arm64/mm/context.c b/arch/arm64/mm/context.c
index f00f5eeb556f..b9b087564bac 100644
--- a/arch/arm64/mm/context.c
+++ b/arch/arm64/mm/context.c
@@ -233,6 +233,15 @@ void check_and_switch_context(struct mm_struct *mm, unsigned int cpu)
 	cpu_switch_mm(mm->pgd, mm);
 }
 
+/* Errata workaround post TTBRx_EL1 update. */
+asmlinkage void post_ttbr_update_workaround(void)
+{
+	asm(ALTERNATIVE("nop; nop; nop",
+			"ic iallu; dsb nsh; isb",
+			ARM64_WORKAROUND_CAVIUM_27456,
+			CONFIG_CAVIUM_ERRATUM_27456));
+}
+
 static int asids_init(void)
 {
 	asid_bits = get_cpu_asid_bits();
diff --git a/arch/arm64/mm/proc.S b/arch/arm64/mm/proc.S
index 135a698ce946..619da1cbd32b 100644
--- a/arch/arm64/mm/proc.S
+++ b/arch/arm64/mm/proc.S
@@ -139,8 +139,7 @@ ENTRY(cpu_do_switch_mm)
 	isb
 	msr	ttbr0_el1, x0			// now update TTBR0
 	isb
-	post_ttbr0_update_workaround
-	ret
+	b	post_ttbr_update_workaround	// Back to C code...
 ENDPROC(cpu_do_switch_mm)
 
 	.pushsection ".idmap.text", "awx"
-- 
2.11.0

[PATCH v4.9.y 17/42] arm64: Add skeleton to harden the branch predictor against aliasing attacks

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 0f15adbb2861ce6f75ccfc5a92b19eae0ef327d0 upstream.

Aliasing attacks against CPU branch predictors can allow an attacker to
redirect speculative control flow on some CPUs and potentially divulge
information from one context to another.

This patch adds initial skeleton code behind a new Kconfig option to
enable implementation-specific mitigations against these attacks for
CPUs that are affected.

Co-developed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: copy bp hardening cb via text mapping]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/Kconfig               | 17 +++++++++
 arch/arm64/include/asm/cpucaps.h |  3 +-
 arch/arm64/include/asm/mmu.h     | 39 +++++++++++++++++++++
 arch/arm64/include/asm/sysreg.h  |  2 ++
 arch/arm64/kernel/Makefile       |  4 +++
 arch/arm64/kernel/bpi.S          | 55 +++++++++++++++++++++++++++++
 arch/arm64/kernel/cpu_errata.c   | 74 ++++++++++++++++++++++++++++++++++++++++
 arch/arm64/kernel/cpufeature.c   |  3 +-
 arch/arm64/kernel/entry.S        |  8 +++--
 arch/arm64/mm/context.c          |  2 ++
 arch/arm64/mm/fault.c            | 17 +++++++++
 11 files changed, 219 insertions(+), 5 deletions(-)
 create mode 100644 arch/arm64/kernel/bpi.S

diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index c8471cf46cbb..90e58bbbd858 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -745,6 +745,23 @@ config UNMAP_KERNEL_AT_EL0
 
 	  If unsure, say Y.
 
+config HARDEN_BRANCH_PREDICTOR
+	bool "Harden the branch predictor against aliasing attacks" if EXPERT
+	default y
+	help
+	  Speculation attacks against some high-performance processors rely on
+	  being able to manipulate the branch predictor for a victim context by
+	  executing aliasing branches in the attacker context.  Such attacks
+	  can be partially mitigated against by clearing internal branch
+	  predictor state and limiting the prediction logic in some situations.
+
+	  This config option will take CPU-specific actions to harden the
+	  branch predictor against aliasing attacks and may rely on specific
+	  instruction sequences or control bits being set by the system
+	  firmware.
+
+	  If unsure, say Y.
+
 menuconfig ARMV8_DEPRECATED
 	bool "Emulate deprecated/obsolete ARMv8 instructions"
 	depends on COMPAT
diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
index 7ddf233f05bd..ce67bf6a0886 100644
--- a/arch/arm64/include/asm/cpucaps.h
+++ b/arch/arm64/include/asm/cpucaps.h
@@ -35,7 +35,8 @@
 #define ARM64_HYP_OFFSET_LOW			14
 #define ARM64_MISMATCHED_CACHE_LINE_SIZE	15
 #define ARM64_UNMAP_KERNEL_AT_EL0		16
+#define ARM64_HARDEN_BRANCH_PREDICTOR		17
 
-#define ARM64_NCAPS				17
+#define ARM64_NCAPS				18
 
 #endif /* __ASM_CPUCAPS_H */
diff --git a/arch/arm64/include/asm/mmu.h b/arch/arm64/include/asm/mmu.h
index a813edf28737..d51158a61892 100644
--- a/arch/arm64/include/asm/mmu.h
+++ b/arch/arm64/include/asm/mmu.h
@@ -20,6 +20,8 @@
 
 #ifndef __ASSEMBLY__
 
+#include <linux/percpu.h>
+
 typedef struct {
 	atomic64_t	id;
 	void		*vdso;
@@ -38,6 +40,43 @@ static inline bool arm64_kernel_unmapped_at_el0(void)
 	       cpus_have_cap(ARM64_UNMAP_KERNEL_AT_EL0);
 }
 
+typedef void (*bp_hardening_cb_t)(void);
+
+struct bp_hardening_data {
+	int			hyp_vectors_slot;
+	bp_hardening_cb_t	fn;
+};
+
+#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
+extern char __bp_harden_hyp_vecs_start[], __bp_harden_hyp_vecs_end[];
+
+DECLARE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
+
+static inline struct bp_hardening_data *arm64_get_bp_hardening_data(void)
+{
+	return this_cpu_ptr(&bp_hardening_data);
+}
+
+static inline void arm64_apply_bp_hardening(void)
+{
+	struct bp_hardening_data *d;
+
+	if (!cpus_have_cap(ARM64_HARDEN_BRANCH_PREDICTOR))
+		return;
+
+	d = arm64_get_bp_hardening_data();
+	if (d->fn)
+		d->fn();
+}
+#else
+static inline struct bp_hardening_data *arm64_get_bp_hardening_data(void)
+{
+	return NULL;
+}
+
+static inline void arm64_apply_bp_hardening(void)	{ }
+#endif	/* CONFIG_HARDEN_BRANCH_PREDICTOR */
+
 extern void paging_init(void);
 extern void bootmem_init(void);
 extern void __iomem *early_io_map(phys_addr_t phys, unsigned long virt);
diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h
index 7cb7f7cdcfbc..88bbe364b6ae 100644
--- a/arch/arm64/include/asm/sysreg.h
+++ b/arch/arm64/include/asm/sysreg.h
@@ -118,6 +118,8 @@
 
 /* id_aa64pfr0 */
 #define ID_AA64PFR0_CSV3_SHIFT		60
+#define ID_AA64PFR0_CSV2_SHIFT		56
+#define ID_AA64PFR0_SVE_SHIFT		32
 #define ID_AA64PFR0_GIC_SHIFT		24
 #define ID_AA64PFR0_ASIMD_SHIFT		20
 #define ID_AA64PFR0_FP_SHIFT		16
diff --git a/arch/arm64/kernel/Makefile b/arch/arm64/kernel/Makefile
index 7d66bbaafc0c..74b8fd860714 100644
--- a/arch/arm64/kernel/Makefile
+++ b/arch/arm64/kernel/Makefile
@@ -51,6 +51,10 @@ arm64-obj-$(CONFIG_HIBERNATION)		+= hibernate.o hibernate-asm.o
 arm64-obj-$(CONFIG_KEXEC)		+= machine_kexec.o relocate_kernel.o	\
 					   cpu-reset.o
 
+ifeq ($(CONFIG_KVM),y)
+arm64-obj-$(CONFIG_HARDEN_BRANCH_PREDICTOR)	+= bpi.o
+endif
+
 obj-y					+= $(arm64-obj-y) vdso/ probes/
 obj-m					+= $(arm64-obj-m)
 head-y					:= head.o
diff --git a/arch/arm64/kernel/bpi.S b/arch/arm64/kernel/bpi.S
new file mode 100644
index 000000000000..06a931eb2673
--- /dev/null
+++ b/arch/arm64/kernel/bpi.S
@@ -0,0 +1,55 @@
+/*
+ * Contains CPU specific branch predictor invalidation sequences
+ *
+ * Copyright (C) 2018 ARM Ltd.
+ *
+ * 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, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/linkage.h>
+
+.macro ventry target
+	.rept 31
+	nop
+	.endr
+	b	\target
+.endm
+
+.macro vectors target
+	ventry \target + 0x000
+	ventry \target + 0x080
+	ventry \target + 0x100
+	ventry \target + 0x180
+
+	ventry \target + 0x200
+	ventry \target + 0x280
+	ventry \target + 0x300
+	ventry \target + 0x380
+
+	ventry \target + 0x400
+	ventry \target + 0x480
+	ventry \target + 0x500
+	ventry \target + 0x580
+
+	ventry \target + 0x600
+	ventry \target + 0x680
+	ventry \target + 0x700
+	ventry \target + 0x780
+.endm
+
+	.align	11
+ENTRY(__bp_harden_hyp_vecs_start)
+	.rept 4
+	vectors __kvm_hyp_vector
+	.endr
+ENTRY(__bp_harden_hyp_vecs_end)
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index b866df79bb83..3884c4398597 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -46,6 +46,80 @@ static int cpu_enable_trap_ctr_access(void *__unused)
 	return 0;
 }
 
+#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
+#include <asm/mmu_context.h>
+#include <asm/cacheflush.h>
+
+DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
+
+#ifdef CONFIG_KVM
+static void __copy_hyp_vect_bpi(int slot, const char *hyp_vecs_start,
+				const char *hyp_vecs_end)
+{
+	void *dst = __bp_harden_hyp_vecs_start + slot * SZ_2K;
+	int i;
+
+	for (i = 0; i < SZ_2K; i += 0x80)
+		memcpy(dst + i, hyp_vecs_start, hyp_vecs_end - hyp_vecs_start);
+
+	flush_icache_range((uintptr_t)dst, (uintptr_t)dst + SZ_2K);
+}
+
+static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
+				      const char *hyp_vecs_start,
+				      const char *hyp_vecs_end)
+{
+	static int last_slot = -1;
+	static DEFINE_SPINLOCK(bp_lock);
+	int cpu, slot = -1;
+
+	spin_lock(&bp_lock);
+	for_each_possible_cpu(cpu) {
+		if (per_cpu(bp_hardening_data.fn, cpu) == fn) {
+			slot = per_cpu(bp_hardening_data.hyp_vectors_slot, cpu);
+			break;
+		}
+	}
+
+	if (slot == -1) {
+		last_slot++;
+		BUG_ON(((__bp_harden_hyp_vecs_end - __bp_harden_hyp_vecs_start)
+			/ SZ_2K) <= last_slot);
+		slot = last_slot;
+		__copy_hyp_vect_bpi(slot, hyp_vecs_start, hyp_vecs_end);
+	}
+
+	__this_cpu_write(bp_hardening_data.hyp_vectors_slot, slot);
+	__this_cpu_write(bp_hardening_data.fn, fn);
+	spin_unlock(&bp_lock);
+}
+#else
+static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
+				      const char *hyp_vecs_start,
+				      const char *hyp_vecs_end)
+{
+	__this_cpu_write(bp_hardening_data.fn, fn);
+}
+#endif	/* CONFIG_KVM */
+
+static void  install_bp_hardening_cb(const struct arm64_cpu_capabilities *entry,
+				     bp_hardening_cb_t fn,
+				     const char *hyp_vecs_start,
+				     const char *hyp_vecs_end)
+{
+	u64 pfr0;
+
+	if (!entry->matches(entry, SCOPE_LOCAL_CPU))
+		return;
+
+	pfr0 = read_cpuid(ID_AA64PFR0_EL1);
+	if (cpuid_feature_extract_unsigned_field(pfr0, ID_AA64PFR0_CSV2_SHIFT))
+		return;
+
+	__install_bp_hardening_cb(fn, hyp_vecs_start, hyp_vecs_end);
+}
+#endif	/* CONFIG_HARDEN_BRANCH_PREDICTOR */
+
 #define MIDR_RANGE(model, min, max) \
 	.def_scope = SCOPE_LOCAL_CPU, \
 	.matches = is_affected_midr_range, \
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 692f107bb46e..a0ee01202503 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -94,7 +94,8 @@ static const struct arm64_ftr_bits ftr_id_aa64isar0[] = {
 
 static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = {
 	ARM64_FTR_BITS(FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_CSV3_SHIFT, 4, 0),
-	ARM64_FTR_BITS(FTR_STRICT, FTR_EXACT, 32, 28, 0),
+	ARM64_FTR_BITS(FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_CSV2_SHIFT, 4, 0),
+	ARM64_FTR_BITS(FTR_STRICT, FTR_EXACT, 32, 24, 0),
 	ARM64_FTR_BITS(FTR_STRICT, FTR_EXACT, 28, 4, 0),
 	ARM64_FTR_BITS(FTR_STRICT, FTR_EXACT, ID_AA64PFR0_GIC_SHIFT, 4, 0),
 	S_ARM64_FTR_BITS(FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_ASIMD_SHIFT, 4, ID_AA64PFR0_ASIMD_NI),
diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index c154b1608405..fd6879d98bd7 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -589,13 +589,15 @@ el0_ia:
 	 * Instruction abort handling
 	 */
 	mrs	x26, far_el1
-	// enable interrupts before calling the main handler
-	enable_dbg_and_irq
+	msr     daifclr, #(8 | 4 | 1)
+#ifdef CONFIG_TRACE_IRQFLAGS
+	bl	trace_hardirqs_off
+#endif
 	ct_user_exit
 	mov	x0, x26
 	mov	x1, x25
 	mov	x2, sp
-	bl	do_mem_abort
+	bl	do_el0_ia_bp_hardening
 	b	ret_to_user
 el0_fpsimd_acc:
 	/*
diff --git a/arch/arm64/mm/context.c b/arch/arm64/mm/context.c
index b9b087564bac..accf7ead3945 100644
--- a/arch/arm64/mm/context.c
+++ b/arch/arm64/mm/context.c
@@ -240,6 +240,8 @@ asmlinkage void post_ttbr_update_workaround(void)
 			"ic iallu; dsb nsh; isb",
 			ARM64_WORKAROUND_CAVIUM_27456,
 			CONFIG_CAVIUM_ERRATUM_27456));
+
+	arm64_apply_bp_hardening();
 }
 
 static int asids_init(void)
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index 4df70c9fd762..c95b194d1c6e 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -590,6 +590,23 @@ asmlinkage void __exception do_mem_abort(unsigned long addr, unsigned int esr,
 	arm64_notify_die("", regs, &info, esr);
 }
 
+asmlinkage void __exception do_el0_ia_bp_hardening(unsigned long addr,
+						   unsigned int esr,
+						   struct pt_regs *regs)
+{
+	/*
+	 * We've taken an instruction abort from userspace and not yet
+	 * re-enabled IRQs. If the address is a kernel address, apply
+	 * BP hardening prior to enabling IRQs and pre-emption.
+	 */
+	if (addr > TASK_SIZE)
+		arm64_apply_bp_hardening();
+
+	local_irq_enable();
+	do_mem_abort(addr, esr, regs);
+}
+
+
 /*
  * Handle stack alignment exceptions.
  */
-- 
2.11.0

[PATCH v4.9.y 18/42] arm64: Move BP hardening to check_and_switch_context

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit a8e4c0a919ae310944ed2c9ace11cf3ccd8a609b upstream.

We call arm64_apply_bp_hardening() from post_ttbr_update_workaround,
which has the unexpected consequence of being triggered on every
exception return to userspace when ARM64_SW_TTBR0_PAN is selected,
even if no context switch actually occured.

This is a bit suboptimal, and it would be more logical to only
invalidate the branch predictor when we actually switch to
a different mm.

In order to solve this, move the call to arm64_apply_bp_hardening()
into check_and_switch_context(), where we're guaranteed to pick
a different mm context.

Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/mm/context.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/mm/context.c b/arch/arm64/mm/context.c
index accf7ead3945..62d976e843fc 100644
--- a/arch/arm64/mm/context.c
+++ b/arch/arm64/mm/context.c
@@ -230,6 +230,9 @@ void check_and_switch_context(struct mm_struct *mm, unsigned int cpu)
 	raw_spin_unlock_irqrestore(&cpu_asid_lock, flags);
 
 switch_mm_fastpath:
+
+	arm64_apply_bp_hardening();
+
 	cpu_switch_mm(mm->pgd, mm);
 }
 
@@ -240,8 +243,6 @@ asmlinkage void post_ttbr_update_workaround(void)
 			"ic iallu; dsb nsh; isb",
 			ARM64_WORKAROUND_CAVIUM_27456,
 			CONFIG_CAVIUM_ERRATUM_27456));
-
-	arm64_apply_bp_hardening();
 }
 
 static int asids_init(void)
-- 
2.11.0

[PATCH v4.9.y 19/42] mm: Introduce lm_alias

[Mark Rutland]

From: Laura Abbott <labbott@redhat.com>

commit 568c5fe5a54f2654f5a4c599c45b8a62ed9a2013 upstream.

Certain architectures may have the kernel image mapped separately to
alias the linear map. Introduce a macro lm_alias to translate a kernel
image symbol into its linear alias. This is used in part with work to
add CONFIG_DEBUG_VIRTUAL support for arm64.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 include/linux/mm.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 8e506783631b..4a07ff4f38e1 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -76,6 +76,10 @@ extern int mmap_rnd_compat_bits __read_mostly;
 #define page_to_virt(x)	__va(PFN_PHYS(page_to_pfn(x)))
 #endif
 
+#ifndef lm_alias
+#define lm_alias(x)	__va(__pa_symbol(x))
+#endif
+
 /*
  * To prevent common memory management code establishing
  * a zero page mapping on a read fault.
-- 
2.11.0

[PATCH v4.9.y 20/42] arm64: KVM: Use per-CPU vector when BP hardening is enabled

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 6840bdd73d07216ab4bc46f5a8768c37ea519038 upstream.

Now that we have per-CPU vectors, let's plug then in the KVM/arm64 code.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream, use cpus_have_cap()]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/include/asm/kvm_mmu.h   | 10 ++++++++++
 arch/arm/kvm/arm.c               |  9 ++++++++-
 arch/arm64/include/asm/kvm_mmu.h | 38 ++++++++++++++++++++++++++++++++++++++
 arch/arm64/kvm/hyp/switch.c      |  2 +-
 4 files changed, 57 insertions(+), 2 deletions(-)

diff --git a/arch/arm/include/asm/kvm_mmu.h b/arch/arm/include/asm/kvm_mmu.h
index a58bbaa3ec60..d10e36235438 100644
--- a/arch/arm/include/asm/kvm_mmu.h
+++ b/arch/arm/include/asm/kvm_mmu.h
@@ -223,6 +223,16 @@ static inline unsigned int kvm_get_vmid_bits(void)
 	return 8;
 }
 
+static inline void *kvm_get_hyp_vector(void)
+{
+	return kvm_ksym_ref(__kvm_hyp_vector);
+}
+
+static inline int kvm_map_vectors(void)
+{
+	return 0;
+}
+
 #endif	/* !__ASSEMBLY__ */
 
 #endif /* __ARM_KVM_MMU_H__ */
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index c38bfbeec306..5e666b624dc1 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -1088,7 +1088,7 @@ static void cpu_init_hyp_mode(void *dummy)
 	pgd_ptr = kvm_mmu_get_httbr();
 	stack_page = __this_cpu_read(kvm_arm_hyp_stack_page);
 	hyp_stack_ptr = stack_page + PAGE_SIZE;
-	vector_ptr = (unsigned long)kvm_ksym_ref(__kvm_hyp_vector);
+	vector_ptr = (unsigned long)kvm_get_hyp_vector();
 
 	__cpu_init_hyp_mode(pgd_ptr, hyp_stack_ptr, vector_ptr);
 	__cpu_init_stage2();
@@ -1345,6 +1345,13 @@ static int init_hyp_mode(void)
 		goto out_err;
 	}
 
+
+	err = kvm_map_vectors();
+	if (err) {
+		kvm_err("Cannot map vectors\n");
+		goto out_err;
+	}
+
 	/*
 	 * Map the Hyp stack pages
 	 */
diff --git a/arch/arm64/include/asm/kvm_mmu.h b/arch/arm64/include/asm/kvm_mmu.h
index 6d22017ebbad..80bf33715ecb 100644
--- a/arch/arm64/include/asm/kvm_mmu.h
+++ b/arch/arm64/include/asm/kvm_mmu.h
@@ -313,5 +313,43 @@ static inline unsigned int kvm_get_vmid_bits(void)
 	return (cpuid_feature_extract_unsigned_field(reg, ID_AA64MMFR1_VMIDBITS_SHIFT) == 2) ? 16 : 8;
 }
 
+#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
+#include <asm/mmu.h>
+
+static inline void *kvm_get_hyp_vector(void)
+{
+	struct bp_hardening_data *data = arm64_get_bp_hardening_data();
+	void *vect = kvm_ksym_ref(__kvm_hyp_vector);
+
+	if (data->fn) {
+		vect = __bp_harden_hyp_vecs_start +
+		       data->hyp_vectors_slot * SZ_2K;
+
+		if (!cpus_have_cap(ARM64_HAS_VIRT_HOST_EXTN))
+			vect = lm_alias(vect);
+	}
+
+	return vect;
+}
+
+static inline int kvm_map_vectors(void)
+{
+	return create_hyp_mappings(kvm_ksym_ref(__bp_harden_hyp_vecs_start),
+				   kvm_ksym_ref(__bp_harden_hyp_vecs_end),
+				   PAGE_HYP_EXEC);
+}
+
+#else
+static inline void *kvm_get_hyp_vector(void)
+{
+	return kvm_ksym_ref(__kvm_hyp_vector);
+}
+
+static inline int kvm_map_vectors(void)
+{
+	return 0;
+}
+#endif
+
 #endif /* __ASSEMBLY__ */
 #endif /* __ARM64_KVM_MMU_H__ */
diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
index 0c848c18ca44..cf6d962f9639 100644
--- a/arch/arm64/kvm/hyp/switch.c
+++ b/arch/arm64/kvm/hyp/switch.c
@@ -50,7 +50,7 @@ static void __hyp_text __activate_traps_vhe(void)
 	val &= ~CPACR_EL1_FPEN;
 	write_sysreg(val, cpacr_el1);
 
-	write_sysreg(__kvm_hyp_vector, vbar_el1);
+	write_sysreg(kvm_get_hyp_vector(), vbar_el1);
 }
 
 static void __hyp_text __activate_traps_nvhe(void)
-- 
2.11.0

[PATCH v4.9.y 21/42] arm64: entry: Apply BP hardening for high-priority synchronous exceptions

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 5dfc6ed27710c42cbc15db5c0d4475699991da0a upstream.

Software-step and PC alignment fault exceptions have higher priority than
instruction abort exceptions, so apply the BP hardening hooks there too
if the user PC appears to reside in kernel space.

Reported-by: Dan Hettena <dhettena@nvidia.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/entry.S | 6 ++++--
 arch/arm64/mm/fault.c     | 9 +++++++++
 2 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index fd6879d98bd7..30f4331d477a 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -624,8 +624,10 @@ el0_sp_pc:
 	 * Stack or PC alignment exception handling
 	 */
 	mrs	x26, far_el1
-	// enable interrupts before calling the main handler
-	enable_dbg_and_irq
+	enable_dbg
+#ifdef CONFIG_TRACE_IRQFLAGS
+	bl	trace_hardirqs_off
+#endif
 	ct_user_exit
 	mov	x0, x26
 	mov	x1, x25
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index c95b194d1c6e..6120a1486054 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -617,6 +617,12 @@ asmlinkage void __exception do_sp_pc_abort(unsigned long addr,
 	struct siginfo info;
 	struct task_struct *tsk = current;
 
+	if (user_mode(regs)) {
+		if (instruction_pointer(regs) > TASK_SIZE)
+			arm64_apply_bp_hardening();
+		local_irq_enable();
+	}
+
 	if (show_unhandled_signals && unhandled_signal(tsk, SIGBUS))
 		pr_info_ratelimited("%s[%d]: %s exception: pc=%p sp=%p\n",
 				    tsk->comm, task_pid_nr(tsk),
@@ -676,6 +682,9 @@ asmlinkage int __exception do_debug_exception(unsigned long addr,
 	if (interrupts_enabled(regs))
 		trace_hardirqs_off();
 
+	if (user_mode(regs) && instruction_pointer(regs) > TASK_SIZE)
+		arm64_apply_bp_hardening();
+
 	if (!inf->fn(addr, esr, regs)) {
 		rv = 1;
 	} else {
-- 
2.11.0

[PATCH v4.9.y 22/42] arm64: entry: Apply BP hardening for suspicious interrupts from EL0

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit 30d88c0e3ace625a92eead9ca0ad94093a8f59fe upstream.

It is possible to take an IRQ from EL0 following a branch to a kernel
address in such a way that the IRQ is prioritised over the instruction
abort. Whilst an attacker would need to get the stars to align here,
it might be sufficient with enough calibration so perform BP hardening
in the rare case that we see a kernel address in the ELR when handling
an IRQ from EL0.

Reported-by: Dan Hettena <dhettena@nvidia.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/entry.S | 5 +++++
 arch/arm64/mm/fault.c     | 6 ++++++
 2 files changed, 11 insertions(+)

diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
index 30f4331d477a..b79e302d2a3e 100644
--- a/arch/arm64/kernel/entry.S
+++ b/arch/arm64/kernel/entry.S
@@ -686,6 +686,11 @@ el0_irq_naked:
 #endif
 
 	ct_user_exit
+#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
+	tbz	x22, #55, 1f
+	bl	do_el0_irq_bp_hardening
+1:
+#endif
 	irq_handler
 
 #ifdef CONFIG_TRACE_IRQFLAGS
diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
index 6120a1486054..ad49ae8f3967 100644
--- a/arch/arm64/mm/fault.c
+++ b/arch/arm64/mm/fault.c
@@ -590,6 +590,12 @@ asmlinkage void __exception do_mem_abort(unsigned long addr, unsigned int esr,
 	arm64_notify_die("", regs, &info, esr);
 }
 
+asmlinkage void __exception do_el0_irq_bp_hardening(void)
+{
+	/* PC has already been checked in entry.S */
+	arm64_apply_bp_hardening();
+}
+
 asmlinkage void __exception do_el0_ia_bp_hardening(unsigned long addr,
 						   unsigned int esr,
 						   struct pt_regs *regs)
-- 
2.11.0

[PATCH v4.9.y 23/42] arm64: cputype: Add missing MIDR values for Cortex-A72 and Cortex-A75

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit a65d219fe5dc7887fd5ca04c2ac3e9a34feb8dfc upstream.

Hook up MIDR values for the Cortex-A72 and Cortex-A75 CPUs, since they
will soon need MIDR matches for hardening the branch predictor.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/include/asm/cputype.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/arm64/include/asm/cputype.h b/arch/arm64/include/asm/cputype.h
index 1d47930c30dc..9ee3038a6b98 100644
--- a/arch/arm64/include/asm/cputype.h
+++ b/arch/arm64/include/asm/cputype.h
@@ -75,7 +75,10 @@
 #define ARM_CPU_PART_AEM_V8		0xD0F
 #define ARM_CPU_PART_FOUNDATION		0xD00
 #define ARM_CPU_PART_CORTEX_A57		0xD07
+#define ARM_CPU_PART_CORTEX_A72		0xD08
 #define ARM_CPU_PART_CORTEX_A53		0xD03
+#define ARM_CPU_PART_CORTEX_A73		0xD09
+#define ARM_CPU_PART_CORTEX_A75		0xD0A
 
 #define APM_CPU_PART_POTENZA		0x000
 
@@ -87,6 +90,9 @@
 
 #define MIDR_CORTEX_A53 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A53)
 #define MIDR_CORTEX_A57 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A57)
+#define MIDR_CORTEX_A72 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A72)
+#define MIDR_CORTEX_A73 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A73)
+#define MIDR_CORTEX_A75 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A75)
 #define MIDR_THUNDERX	MIDR_CPU_MODEL(ARM_CPU_IMP_CAVIUM, CAVIUM_CPU_PART_THUNDERX)
 #define MIDR_THUNDERX_81XX MIDR_CPU_MODEL(ARM_CPU_IMP_CAVIUM, CAVIUM_CPU_PART_THUNDERX_81XX)
 #define MIDR_CAVIUM_THUNDERX2 MIDR_CPU_MODEL(ARM_CPU_IMP_CAVIUM, CAVIUM_CPU_PART_THUNDERX2)
-- 
2.11.0

[PATCH v4.9.y 24/42] arm64: cpu_errata: Allow an erratum to be match for all revisions of a core

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 06f1494f837da8997d670a1ba87add7963b08922 upstream.

Some minor erratum may not be fixed in further revisions of a core,
leading to a situation where the workaround needs to be updated each
time an updated core is released.

Introduce a MIDR_ALL_VERSIONS match helper that will work for all
versions of that MIDR, once and for all.

Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/cpu_errata.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index 3884c4398597..b9fcbd155df2 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -127,6 +127,13 @@ static void  install_bp_hardening_cb(const struct arm64_cpu_capabilities *entry,
 	.midr_range_min = min, \
 	.midr_range_max = max
 
+#define MIDR_ALL_VERSIONS(model) \
+	.def_scope = SCOPE_LOCAL_CPU, \
+	.matches = is_affected_midr_range, \
+	.midr_model = model, \
+	.midr_range_min = 0, \
+	.midr_range_max = (MIDR_VARIANT_MASK | MIDR_REVISION_MASK)
+
 const struct arm64_cpu_capabilities arm64_errata[] = {
 #if	defined(CONFIG_ARM64_ERRATUM_826319) || \
 	defined(CONFIG_ARM64_ERRATUM_827319) || \
-- 
2.11.0

[PATCH v4.9.y 25/42] arm64: Implement branch predictor hardening for affected Cortex-A CPUs

[Mark Rutland]

From: Will Deacon <will.deacon@arm.com>

commit aa6acde65e03186b5add8151e1ffe36c3c62639b upstream.

Cortex-A57, A72, A73 and A75 are susceptible to branch predictor aliasing
and can theoretically be attacked by malicious code.

This patch implements a PSCI-based mitigation for these CPUs when available.
The call into firmware will invalidate the branch predictor state, preventing
any malicious entries from affecting other victim contexts.

Co-developed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/bpi.S        | 24 ++++++++++++++++++++++++
 arch/arm64/kernel/cpu_errata.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 66 insertions(+)

diff --git a/arch/arm64/kernel/bpi.S b/arch/arm64/kernel/bpi.S
index 06a931eb2673..dec95bd82e31 100644
--- a/arch/arm64/kernel/bpi.S
+++ b/arch/arm64/kernel/bpi.S
@@ -53,3 +53,27 @@ ENTRY(__bp_harden_hyp_vecs_start)
 	vectors __kvm_hyp_vector
 	.endr
 ENTRY(__bp_harden_hyp_vecs_end)
+ENTRY(__psci_hyp_bp_inval_start)
+	sub	sp, sp, #(8 * 18)
+	stp	x16, x17, [sp, #(16 * 0)]
+	stp	x14, x15, [sp, #(16 * 1)]
+	stp	x12, x13, [sp, #(16 * 2)]
+	stp	x10, x11, [sp, #(16 * 3)]
+	stp	x8, x9, [sp, #(16 * 4)]
+	stp	x6, x7, [sp, #(16 * 5)]
+	stp	x4, x5, [sp, #(16 * 6)]
+	stp	x2, x3, [sp, #(16 * 7)]
+	stp	x0, x1, [sp, #(16 * 8)]
+	mov	x0, #0x84000000
+	smc	#0
+	ldp	x16, x17, [sp, #(16 * 0)]
+	ldp	x14, x15, [sp, #(16 * 1)]
+	ldp	x12, x13, [sp, #(16 * 2)]
+	ldp	x10, x11, [sp, #(16 * 3)]
+	ldp	x8, x9, [sp, #(16 * 4)]
+	ldp	x6, x7, [sp, #(16 * 5)]
+	ldp	x4, x5, [sp, #(16 * 6)]
+	ldp	x2, x3, [sp, #(16 * 7)]
+	ldp	x0, x1, [sp, #(16 * 8)]
+	add	sp, sp, #(8 * 18)
+ENTRY(__psci_hyp_bp_inval_end)
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index b9fcbd155df2..5550d18b8180 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -53,6 +53,8 @@ static int cpu_enable_trap_ctr_access(void *__unused)
 DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
 
 #ifdef CONFIG_KVM
+extern char __psci_hyp_bp_inval_start[], __psci_hyp_bp_inval_end[];
+
 static void __copy_hyp_vect_bpi(int slot, const char *hyp_vecs_start,
 				const char *hyp_vecs_end)
 {
@@ -94,6 +96,9 @@ static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
 	spin_unlock(&bp_lock);
 }
 #else
+#define __psci_hyp_bp_inval_start	NULL
+#define __psci_hyp_bp_inval_end		NULL
+
 static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
 				      const char *hyp_vecs_start,
 				      const char *hyp_vecs_end)
@@ -118,6 +123,21 @@ static void  install_bp_hardening_cb(const struct arm64_cpu_capabilities *entry,
 
 	__install_bp_hardening_cb(fn, hyp_vecs_start, hyp_vecs_end);
 }
+
+#include <linux/psci.h>
+
+static int enable_psci_bp_hardening(void *data)
+{
+	const struct arm64_cpu_capabilities *entry = data;
+
+	if (psci_ops.get_version)
+		install_bp_hardening_cb(entry,
+				       (bp_hardening_cb_t)psci_ops.get_version,
+				       __psci_hyp_bp_inval_start,
+				       __psci_hyp_bp_inval_end);
+
+	return 0;
+}
 #endif	/* CONFIG_HARDEN_BRANCH_PREDICTOR */
 
 #define MIDR_RANGE(model, min, max) \
@@ -211,6 +231,28 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 		.def_scope = SCOPE_LOCAL_CPU,
 		.enable = cpu_enable_trap_ctr_access,
 	},
+#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A57),
+		.enable = enable_psci_bp_hardening,
+	},
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A72),
+		.enable = enable_psci_bp_hardening,
+	},
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A73),
+		.enable = enable_psci_bp_hardening,
+	},
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A75),
+		.enable = enable_psci_bp_hardening,
+	},
+#endif
 	{
 	}
 };
-- 
2.11.0

[PATCH v4.9.y 26/42] arm64: Branch predictor hardening for Cavium ThunderX2

[Mark Rutland]

From: Jayachandran C <jnair@caviumnetworks.com>

commit f3d795d9b360523beca6d13ba64c2c532f601149 upstream.

Use PSCI based mitigation for speculative execution attacks targeting
the branch predictor. We use the same mechanism as the one used for
Cortex-A CPUs, we expect the PSCI version call to have a side effect
of clearing the BTBs.

Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Jayachandran C <jnair@caviumnetworks.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/cpu_errata.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index 5550d18b8180..67cdfc6a458e 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -252,6 +252,16 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A75),
 		.enable = enable_psci_bp_hardening,
 	},
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_BRCM_VULCAN),
+		.enable = enable_psci_bp_hardening,
+	},
+	{
+		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		MIDR_ALL_VERSIONS(MIDR_CAVIUM_THUNDERX2),
+		.enable = enable_psci_bp_hardening,
+	},
 #endif
 	{
 	}
-- 
2.11.0

[PATCH v4.9.y 27/42] arm64: KVM: Increment PC after handling an SMC trap

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit f5115e8869e1dfafac0e414b4f1664f3a84a4683 upstream.

When handling an SMC trap, the "preferred return address" is set
to that of the SMC, and not the next PC (which is a departure from
the behaviour of an SMC that isn't trapped).

Increment PC in the handler, as the guest is otherwise forever
stuck...

Cc: stable@vger.kernel.org
Fixes: acfb3b883f6d ("arm64: KVM: Fix SMCCC handling of unimplemented SMC/HVC calls")
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kvm/handle_exit.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
index 2e6e9e99977b..5b56b09d317b 100644
--- a/arch/arm64/kvm/handle_exit.c
+++ b/arch/arm64/kvm/handle_exit.c
@@ -53,7 +53,16 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run)
 
 static int handle_smc(struct kvm_vcpu *vcpu, struct kvm_run *run)
 {
+	/*
+	 * "If an SMC instruction executed at Non-secure EL1 is
+	 * trapped to EL2 because HCR_EL2.TSC is 1, the exception is a
+	 * Trap exception, not a Secure Monitor Call exception [...]"
+	 *
+	 * We need to advance the PC after the trap, as it would
+	 * otherwise return to the same address...
+	 */
 	vcpu_set_reg(vcpu, 0, ~0UL);
+	kvm_skip_instr(vcpu, kvm_vcpu_trap_il_is32bit(vcpu));
 	return 1;
 }
 
-- 
2.11.0

[PATCH v4.9.y 28/42] arm/arm64: KVM: Consolidate the PSCI include files

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 1a2fb94e6a771ff94f4afa22497a4695187b820c upstream.

As we're about to update the PSCI support, and because I'm lazy,
let's move the PSCI include file to include/kvm so that both
ARM architectures can find it.

Acked-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/include/asm/kvm_psci.h                    | 27 ----------------------
 arch/arm/kvm/arm.c                                 |  2 +-
 arch/arm/kvm/handle_exit.c                         |  2 +-
 arch/arm/kvm/psci.c                                |  3 ++-
 arch/arm64/kvm/handle_exit.c                       |  5 +++-
 .../asm/kvm_psci.h => include/kvm/arm_psci.h       |  6 ++---
 6 files changed, 11 insertions(+), 34 deletions(-)
 delete mode 100644 arch/arm/include/asm/kvm_psci.h
 rename arch/arm64/include/asm/kvm_psci.h => include/kvm/arm_psci.h (89%)

diff --git a/arch/arm/include/asm/kvm_psci.h b/arch/arm/include/asm/kvm_psci.h
deleted file mode 100644
index 6bda945d31fa..000000000000
--- a/arch/arm/include/asm/kvm_psci.h
+++ /dev/null
@@ -1,27 +0,0 @@
-/*
- * Copyright (C) 2012 - ARM Ltd
- * Author: Marc Zyngier <marc.zyngier@arm.com>
- *
- * 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, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef __ARM_KVM_PSCI_H__
-#define __ARM_KVM_PSCI_H__
-
-#define KVM_ARM_PSCI_0_1	1
-#define KVM_ARM_PSCI_0_2	2
-
-int kvm_psci_version(struct kvm_vcpu *vcpu);
-int kvm_psci_call(struct kvm_vcpu *vcpu);
-
-#endif /* __ARM_KVM_PSCI_H__ */
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index 5e666b624dc1..ef6595c7d697 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -29,6 +29,7 @@
 #include <linux/kvm.h>
 #include <trace/events/kvm.h>
 #include <kvm/arm_pmu.h>
+#include <kvm/arm_psci.h>
 
 #define CREATE_TRACE_POINTS
 #include "trace.h"
@@ -44,7 +45,6 @@
 #include <asm/kvm_mmu.h>
 #include <asm/kvm_emulate.h>
 #include <asm/kvm_coproc.h>
-#include <asm/kvm_psci.h>
 #include <asm/sections.h>
 
 #ifdef REQUIRES_VIRT
diff --git a/arch/arm/kvm/handle_exit.c b/arch/arm/kvm/handle_exit.c
index 4e57ebca6e69..0a2b75882ca8 100644
--- a/arch/arm/kvm/handle_exit.c
+++ b/arch/arm/kvm/handle_exit.c
@@ -21,7 +21,7 @@
 #include <asm/kvm_emulate.h>
 #include <asm/kvm_coproc.h>
 #include <asm/kvm_mmu.h>
-#include <asm/kvm_psci.h>
+#include <kvm/arm_psci.h>
 #include <trace/events/kvm.h>
 
 #include "trace.h"
diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index a08d7a93aebb..f44ecf677414 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -21,9 +21,10 @@
 
 #include <asm/cputype.h>
 #include <asm/kvm_emulate.h>
-#include <asm/kvm_psci.h>
 #include <asm/kvm_host.h>
 
+#include <kvm/arm_psci.h>
+
 #include <uapi/linux/psci.h>
 
 /*
diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
index 5b56b09d317b..ff7c18b89e5e 100644
--- a/arch/arm64/kvm/handle_exit.c
+++ b/arch/arm64/kvm/handle_exit.c
@@ -22,12 +22,15 @@
 #include <linux/kvm.h>
 #include <linux/kvm_host.h>
 
+#include <kvm/arm_psci.h>
+
 #include <asm/esr.h>
 #include <asm/kvm_asm.h>
 #include <asm/kvm_coproc.h>
 #include <asm/kvm_emulate.h>
 #include <asm/kvm_mmu.h>
-#include <asm/kvm_psci.h>
+#include <asm/debug-monitors.h>
+#include <asm/traps.h>
 
 #define CREATE_TRACE_POINTS
 #include "trace.h"
diff --git a/arch/arm64/include/asm/kvm_psci.h b/include/kvm/arm_psci.h
similarity index 89%
rename from arch/arm64/include/asm/kvm_psci.h
rename to include/kvm/arm_psci.h
index bc39e557c56c..2042bb909474 100644
--- a/arch/arm64/include/asm/kvm_psci.h
+++ b/include/kvm/arm_psci.h
@@ -15,8 +15,8 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
-#ifndef __ARM64_KVM_PSCI_H__
-#define __ARM64_KVM_PSCI_H__
+#ifndef __KVM_ARM_PSCI_H__
+#define __KVM_ARM_PSCI_H__
 
 #define KVM_ARM_PSCI_0_1	1
 #define KVM_ARM_PSCI_0_2	2
@@ -24,4 +24,4 @@
 int kvm_psci_version(struct kvm_vcpu *vcpu);
 int kvm_psci_call(struct kvm_vcpu *vcpu);
 
-#endif /* __ARM64_KVM_PSCI_H__ */
+#endif /* __KVM_ARM_PSCI_H__ */
-- 
2.11.0

[PATCH v4.9.y 29/42] arm/arm64: KVM: Add PSCI_VERSION helper

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit d0a144f12a7ca8368933eae6583c096c363ec506 upstream.

As we're about to trigger a PSCI version explosion, it doesn't
hurt to introduce a PSCI_VERSION helper that is going to be
used everywhere.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/kvm/psci.c       | 4 +---
 include/kvm/arm_psci.h    | 6 ++++--
 include/uapi/linux/psci.h | 3 +++
 3 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index f44ecf677414..5d8ad0e6661c 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -25,8 +25,6 @@
 
 #include <kvm/arm_psci.h>
 
-#include <uapi/linux/psci.h>
-
 /*
  * This is an implementation of the Power State Coordination Interface
  * as described in ARM document number ARM DEN 0022A.
@@ -220,7 +218,7 @@ static int kvm_psci_0_2_call(struct kvm_vcpu *vcpu)
 		 * Bits[31:16] = Major Version = 0
 		 * Bits[15:0] = Minor Version = 2
 		 */
-		val = 2;
+		val = KVM_ARM_PSCI_0_2;
 		break;
 	case PSCI_0_2_FN_CPU_SUSPEND:
 	case PSCI_0_2_FN64_CPU_SUSPEND:
diff --git a/include/kvm/arm_psci.h b/include/kvm/arm_psci.h
index 2042bb909474..5659343580a3 100644
--- a/include/kvm/arm_psci.h
+++ b/include/kvm/arm_psci.h
@@ -18,8 +18,10 @@
 #ifndef __KVM_ARM_PSCI_H__
 #define __KVM_ARM_PSCI_H__
 
-#define KVM_ARM_PSCI_0_1	1
-#define KVM_ARM_PSCI_0_2	2
+#include <uapi/linux/psci.h>
+
+#define KVM_ARM_PSCI_0_1	PSCI_VERSION(0, 1)
+#define KVM_ARM_PSCI_0_2	PSCI_VERSION(0, 2)
 
 int kvm_psci_version(struct kvm_vcpu *vcpu);
 int kvm_psci_call(struct kvm_vcpu *vcpu);
diff --git a/include/uapi/linux/psci.h b/include/uapi/linux/psci.h
index 3d7a0fc021a7..39930ca998cd 100644
--- a/include/uapi/linux/psci.h
+++ b/include/uapi/linux/psci.h
@@ -87,6 +87,9 @@
 		(((ver) & PSCI_VERSION_MAJOR_MASK) >> PSCI_VERSION_MAJOR_SHIFT)
 #define PSCI_VERSION_MINOR(ver)			\
 		((ver) & PSCI_VERSION_MINOR_MASK)
+#define PSCI_VERSION(maj, min)						\
+	((((maj) << PSCI_VERSION_MAJOR_SHIFT) & PSCI_VERSION_MAJOR_MASK) | \
+	 ((min) & PSCI_VERSION_MINOR_MASK))
 
 /* PSCI features decoding (>=1.0) */
 #define PSCI_1_0_FEATURES_CPU_SUSPEND_PF_SHIFT	1
-- 
2.11.0

[PATCH v4.9.y 30/42] arm/arm64: KVM: Add smccc accessors to PSCI code

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 84684fecd7ea381824a96634a027b7719587fb77 upstream.

Instead of open coding the accesses to the various registers,
let's add explicit SMCCC accessors.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/kvm/psci.c | 52 ++++++++++++++++++++++++++++++++++++++++++----------
 1 file changed, 42 insertions(+), 10 deletions(-)

diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index 5d8ad0e6661c..f579ed9d279b 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -32,6 +32,38 @@
 
 #define AFFINITY_MASK(level)	~((0x1UL << ((level) * MPIDR_LEVEL_BITS)) - 1)
 
+static u32 smccc_get_function(struct kvm_vcpu *vcpu)
+{
+	return vcpu_get_reg(vcpu, 0);
+}
+
+static unsigned long smccc_get_arg1(struct kvm_vcpu *vcpu)
+{
+	return vcpu_get_reg(vcpu, 1);
+}
+
+static unsigned long smccc_get_arg2(struct kvm_vcpu *vcpu)
+{
+	return vcpu_get_reg(vcpu, 2);
+}
+
+static unsigned long smccc_get_arg3(struct kvm_vcpu *vcpu)
+{
+	return vcpu_get_reg(vcpu, 3);
+}
+
+static void smccc_set_retval(struct kvm_vcpu *vcpu,
+			     unsigned long a0,
+			     unsigned long a1,
+			     unsigned long a2,
+			     unsigned long a3)
+{
+	vcpu_set_reg(vcpu, 0, a0);
+	vcpu_set_reg(vcpu, 1, a1);
+	vcpu_set_reg(vcpu, 2, a2);
+	vcpu_set_reg(vcpu, 3, a3);
+}
+
 static unsigned long psci_affinity_mask(unsigned long affinity_level)
 {
 	if (affinity_level <= 3)
@@ -74,7 +106,7 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu *source_vcpu)
 	unsigned long context_id;
 	phys_addr_t target_pc;
 
-	cpu_id = vcpu_get_reg(source_vcpu, 1) & MPIDR_HWID_BITMASK;
+	cpu_id = smccc_get_arg1(source_vcpu) & MPIDR_HWID_BITMASK;
 	if (vcpu_mode_is_32bit(source_vcpu))
 		cpu_id &= ~((u32) 0);
 
@@ -93,8 +125,8 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu *source_vcpu)
 			return PSCI_RET_INVALID_PARAMS;
 	}
 
-	target_pc = vcpu_get_reg(source_vcpu, 2);
-	context_id = vcpu_get_reg(source_vcpu, 3);
+	target_pc = smccc_get_arg2(source_vcpu);
+	context_id = smccc_get_arg3(source_vcpu);
 
 	kvm_reset_vcpu(vcpu);
 
@@ -113,7 +145,7 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu *source_vcpu)
 	 * NOTE: We always update r0 (or x0) because for PSCI v0.1
 	 * the general puspose registers are undefined upon CPU_ON.
 	 */
-	vcpu_set_reg(vcpu, 0, context_id);
+	smccc_set_retval(vcpu, context_id, 0, 0, 0);
 	vcpu->arch.power_off = false;
 	smp_mb();		/* Make sure the above is visible */
 
@@ -133,8 +165,8 @@ static unsigned long kvm_psci_vcpu_affinity_info(struct kvm_vcpu *vcpu)
 	struct kvm *kvm = vcpu->kvm;
 	struct kvm_vcpu *tmp;
 
-	target_affinity = vcpu_get_reg(vcpu, 1);
-	lowest_affinity_level = vcpu_get_reg(vcpu, 2);
+	target_affinity = smccc_get_arg1(vcpu);
+	lowest_affinity_level = smccc_get_arg2(vcpu);
 
 	/* Determine target affinity mask */
 	target_affinity_mask = psci_affinity_mask(lowest_affinity_level);
@@ -208,7 +240,7 @@ int kvm_psci_version(struct kvm_vcpu *vcpu)
 static int kvm_psci_0_2_call(struct kvm_vcpu *vcpu)
 {
 	struct kvm *kvm = vcpu->kvm;
-	unsigned long psci_fn = vcpu_get_reg(vcpu, 0) & ~((u32) 0);
+	unsigned long psci_fn = smccc_get_function(vcpu);
 	unsigned long val;
 	int ret = 1;
 
@@ -275,14 +307,14 @@ static int kvm_psci_0_2_call(struct kvm_vcpu *vcpu)
 		break;
 	}
 
-	vcpu_set_reg(vcpu, 0, val);
+	smccc_set_retval(vcpu, val, 0, 0, 0);
 	return ret;
 }
 
 static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
 {
 	struct kvm *kvm = vcpu->kvm;
-	unsigned long psci_fn = vcpu_get_reg(vcpu, 0) & ~((u32) 0);
+	unsigned long psci_fn = smccc_get_function(vcpu);
 	unsigned long val;
 
 	switch (psci_fn) {
@@ -300,7 +332,7 @@ static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
 		break;
 	}
 
-	vcpu_set_reg(vcpu, 0, val);
+	smccc_set_retval(vcpu, val, 0, 0, 0);
 	return 1;
 }
 
-- 
2.11.0

[PATCH v4.9.y 31/42] arm/arm64: KVM: Implement PSCI 1.0 support

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 58e0b2239a4d997094ba63986ef4de29ddc91d87 upstream.

PSCI 1.0 can be trivially implemented by providing the FEATURES
call on top of PSCI 0.2 and returning 1.0 as the PSCI version.

We happily ignore everything else, as they are either optional or
are clarifications that do not require any additional change.

PSCI 1.0 is now the default until we decide to add a userspace
selection API.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/kvm/psci.c    | 45 ++++++++++++++++++++++++++++++++++++++++++++-
 include/kvm/arm_psci.h |  3 +++
 2 files changed, 47 insertions(+), 1 deletion(-)

diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index f579ed9d279b..220eaa04b1c4 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -232,7 +232,7 @@ static void kvm_psci_system_reset(struct kvm_vcpu *vcpu)
 int kvm_psci_version(struct kvm_vcpu *vcpu)
 {
 	if (test_bit(KVM_ARM_VCPU_PSCI_0_2, vcpu->arch.features))
-		return KVM_ARM_PSCI_0_2;
+		return KVM_ARM_PSCI_LATEST;
 
 	return KVM_ARM_PSCI_0_1;
 }
@@ -311,6 +311,47 @@ static int kvm_psci_0_2_call(struct kvm_vcpu *vcpu)
 	return ret;
 }
 
+static int kvm_psci_1_0_call(struct kvm_vcpu *vcpu)
+{
+	u32 psci_fn = smccc_get_function(vcpu);
+	u32 feature;
+	unsigned long val;
+	int ret = 1;
+
+	switch(psci_fn) {
+	case PSCI_0_2_FN_PSCI_VERSION:
+		val = KVM_ARM_PSCI_1_0;
+		break;
+	case PSCI_1_0_FN_PSCI_FEATURES:
+		feature = smccc_get_arg1(vcpu);
+		switch(feature) {
+		case PSCI_0_2_FN_PSCI_VERSION:
+		case PSCI_0_2_FN_CPU_SUSPEND:
+		case PSCI_0_2_FN64_CPU_SUSPEND:
+		case PSCI_0_2_FN_CPU_OFF:
+		case PSCI_0_2_FN_CPU_ON:
+		case PSCI_0_2_FN64_CPU_ON:
+		case PSCI_0_2_FN_AFFINITY_INFO:
+		case PSCI_0_2_FN64_AFFINITY_INFO:
+		case PSCI_0_2_FN_MIGRATE_INFO_TYPE:
+		case PSCI_0_2_FN_SYSTEM_OFF:
+		case PSCI_0_2_FN_SYSTEM_RESET:
+		case PSCI_1_0_FN_PSCI_FEATURES:
+			val = 0;
+			break;
+		default:
+			val = PSCI_RET_NOT_SUPPORTED;
+			break;
+		}
+		break;
+	default:
+		return kvm_psci_0_2_call(vcpu);
+	}
+
+	smccc_set_retval(vcpu, val, 0, 0, 0);
+	return ret;
+}
+
 static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
 {
 	struct kvm *kvm = vcpu->kvm;
@@ -353,6 +394,8 @@ static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
 int kvm_psci_call(struct kvm_vcpu *vcpu)
 {
 	switch (kvm_psci_version(vcpu)) {
+	case KVM_ARM_PSCI_1_0:
+		return kvm_psci_1_0_call(vcpu);
 	case KVM_ARM_PSCI_0_2:
 		return kvm_psci_0_2_call(vcpu);
 	case KVM_ARM_PSCI_0_1:
diff --git a/include/kvm/arm_psci.h b/include/kvm/arm_psci.h
index 5659343580a3..32360432cff5 100644
--- a/include/kvm/arm_psci.h
+++ b/include/kvm/arm_psci.h
@@ -22,6 +22,9 @@
 
 #define KVM_ARM_PSCI_0_1	PSCI_VERSION(0, 1)
 #define KVM_ARM_PSCI_0_2	PSCI_VERSION(0, 2)
+#define KVM_ARM_PSCI_1_0	PSCI_VERSION(1, 0)
+
+#define KVM_ARM_PSCI_LATEST	KVM_ARM_PSCI_1_0
 
 int kvm_psci_version(struct kvm_vcpu *vcpu);
 int kvm_psci_call(struct kvm_vcpu *vcpu);
-- 
2.11.0

[PATCH v4.9.y 32/42] arm/arm64: KVM: Advertise SMCCC v1.1

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 09e6be12effdb33bf7210c8867bbd213b66a499e upstream.

The new SMC Calling Convention (v1.1) allows for a reduced overhead
when calling into the firmware, and provides a new feature discovery
mechanism.

Make it visible to KVM guests.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/kvm/handle_exit.c   |  2 +-
 arch/arm/kvm/psci.c          | 24 +++++++++++++++++++++++-
 arch/arm64/kvm/handle_exit.c |  2 +-
 include/kvm/arm_psci.h       |  2 +-
 include/linux/arm-smccc.h    | 13 +++++++++++++
 5 files changed, 39 insertions(+), 4 deletions(-)

diff --git a/arch/arm/kvm/handle_exit.c b/arch/arm/kvm/handle_exit.c
index 0a2b75882ca8..de1aedce2a8b 100644
--- a/arch/arm/kvm/handle_exit.c
+++ b/arch/arm/kvm/handle_exit.c
@@ -36,7 +36,7 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run)
 		      kvm_vcpu_hvc_get_imm(vcpu));
 	vcpu->stat.hvc_exit_stat++;
 
-	ret = kvm_psci_call(vcpu);
+	ret = kvm_hvc_call_handler(vcpu);
 	if (ret < 0) {
 		vcpu_set_reg(vcpu, 0, ~0UL);
 		return 1;
diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index 220eaa04b1c4..02efb0d4ed4d 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -15,6 +15,7 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <linux/arm-smccc.h>
 #include <linux/preempt.h>
 #include <linux/kvm_host.h>
 #include <linux/wait.h>
@@ -337,6 +338,7 @@ static int kvm_psci_1_0_call(struct kvm_vcpu *vcpu)
 		case PSCI_0_2_FN_SYSTEM_OFF:
 		case PSCI_0_2_FN_SYSTEM_RESET:
 		case PSCI_1_0_FN_PSCI_FEATURES:
+		case ARM_SMCCC_VERSION_FUNC_ID:
 			val = 0;
 			break;
 		default:
@@ -391,7 +393,7 @@ static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
  * Errors:
  * -EINVAL: Unrecognized PSCI function
  */
-int kvm_psci_call(struct kvm_vcpu *vcpu)
+static int kvm_psci_call(struct kvm_vcpu *vcpu)
 {
 	switch (kvm_psci_version(vcpu)) {
 	case KVM_ARM_PSCI_1_0:
@@ -404,3 +406,23 @@ int kvm_psci_call(struct kvm_vcpu *vcpu)
 		return -EINVAL;
 	};
 }
+
+int kvm_hvc_call_handler(struct kvm_vcpu *vcpu)
+{
+	u32 func_id = smccc_get_function(vcpu);
+	u32 val = PSCI_RET_NOT_SUPPORTED;
+
+	switch (func_id) {
+	case ARM_SMCCC_VERSION_FUNC_ID:
+		val = ARM_SMCCC_VERSION_1_1;
+		break;
+	case ARM_SMCCC_ARCH_FEATURES_FUNC_ID:
+		/* Nothing supported yet */
+		break;
+	default:
+		return kvm_psci_call(vcpu);
+	}
+
+	smccc_set_retval(vcpu, val, 0, 0, 0);
+	return 1;
+}
diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
index ff7c18b89e5e..efe43c5f2dc1 100644
--- a/arch/arm64/kvm/handle_exit.c
+++ b/arch/arm64/kvm/handle_exit.c
@@ -45,7 +45,7 @@ static int handle_hvc(struct kvm_vcpu *vcpu, struct kvm_run *run)
 			    kvm_vcpu_hvc_get_imm(vcpu));
 	vcpu->stat.hvc_exit_stat++;
 
-	ret = kvm_psci_call(vcpu);
+	ret = kvm_hvc_call_handler(vcpu);
 	if (ret < 0) {
 		vcpu_set_reg(vcpu, 0, ~0UL);
 		return 1;
diff --git a/include/kvm/arm_psci.h b/include/kvm/arm_psci.h
index 32360432cff5..ed1dd8088f1c 100644
--- a/include/kvm/arm_psci.h
+++ b/include/kvm/arm_psci.h
@@ -27,6 +27,6 @@
 #define KVM_ARM_PSCI_LATEST	KVM_ARM_PSCI_1_0
 
 int kvm_psci_version(struct kvm_vcpu *vcpu);
-int kvm_psci_call(struct kvm_vcpu *vcpu);
+int kvm_hvc_call_handler(struct kvm_vcpu *vcpu);
 
 #endif /* __KVM_ARM_PSCI_H__ */
diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
index 4c5bca38c653..dc68aa5a7261 100644
--- a/include/linux/arm-smccc.h
+++ b/include/linux/arm-smccc.h
@@ -60,6 +60,19 @@
 #define ARM_SMCCC_QUIRK_NONE		0
 #define ARM_SMCCC_QUIRK_QCOM_A6		1 /* Save/restore register a6 */
 
+#define ARM_SMCCC_VERSION_1_0		0x10000
+#define ARM_SMCCC_VERSION_1_1		0x10001
+
+#define ARM_SMCCC_VERSION_FUNC_ID					\
+	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,				\
+			   ARM_SMCCC_SMC_32,				\
+			   0, 0)
+
+#define ARM_SMCCC_ARCH_FEATURES_FUNC_ID					\
+	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,				\
+			   ARM_SMCCC_SMC_32,				\
+			   0, 1)
+
 #ifndef __ASSEMBLY__
 
 #include <linux/linkage.h>
-- 
2.11.0

[PATCH v4.9.y 33/42] arm64: KVM: Make PSCI_VERSION a fast path

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 90348689d500410ca7a55624c667f956771dce7f upstream.

For those CPUs that require PSCI to perform a BP invalidation,
going all the way to the PSCI code for not much is a waste of
precious cycles. Let's terminate that call as early as possible.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kvm/hyp/switch.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
index cf6d962f9639..3eab6ac18d7d 100644
--- a/arch/arm64/kvm/hyp/switch.c
+++ b/arch/arm64/kvm/hyp/switch.c
@@ -17,6 +17,7 @@
 
 #include <linux/types.h>
 #include <linux/jump_label.h>
+#include <uapi/linux/psci.h>
 
 #include <asm/kvm_asm.h>
 #include <asm/kvm_emulate.h>
@@ -308,6 +309,18 @@ int __hyp_text __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 	if (exit_code == ARM_EXCEPTION_TRAP && !__populate_fault_info(vcpu))
 		goto again;
 
+	if (exit_code == ARM_EXCEPTION_TRAP &&
+	    (kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC64 ||
+	     kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC32) &&
+	    vcpu_get_reg(vcpu, 0) == PSCI_0_2_FN_PSCI_VERSION) {
+		u64 val = PSCI_RET_NOT_SUPPORTED;
+		if (test_bit(KVM_ARM_VCPU_PSCI_0_2, vcpu->arch.features))
+			val = 2;
+
+		vcpu_set_reg(vcpu, 0, val);
+		goto again;
+	}
+
 	if (static_branch_unlikely(&vgic_v2_cpuif_trap) &&
 	    exit_code == ARM_EXCEPTION_TRAP) {
 		bool valid;
-- 
2.11.0

[PATCH v4.9.y 34/42] arm/arm64: KVM: Turn kvm_psci_version into a static inline

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit a4097b351118e821841941a79ec77d3ce3f1c5d9 upstream.

We're about to need kvm_psci_version in HYP too. So let's turn it
into a static inline, and pass the kvm structure as a second
parameter (so that HYP can do a kern_hyp_va on it).

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/kvm/psci.c         | 12 ++----------
 arch/arm64/kvm/hyp/switch.c | 20 ++++++++++++--------
 include/kvm/arm_psci.h      | 21 ++++++++++++++++++++-
 3 files changed, 34 insertions(+), 19 deletions(-)

diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index 02efb0d4ed4d..1120ad0bd3d2 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -120,7 +120,7 @@ static unsigned long kvm_psci_vcpu_on(struct kvm_vcpu *source_vcpu)
 	if (!vcpu)
 		return PSCI_RET_INVALID_PARAMS;
 	if (!vcpu->arch.power_off) {
-		if (kvm_psci_version(source_vcpu) != KVM_ARM_PSCI_0_1)
+		if (kvm_psci_version(source_vcpu, kvm) != KVM_ARM_PSCI_0_1)
 			return PSCI_RET_ALREADY_ON;
 		else
 			return PSCI_RET_INVALID_PARAMS;
@@ -230,14 +230,6 @@ static void kvm_psci_system_reset(struct kvm_vcpu *vcpu)
 	kvm_prepare_system_event(vcpu, KVM_SYSTEM_EVENT_RESET);
 }
 
-int kvm_psci_version(struct kvm_vcpu *vcpu)
-{
-	if (test_bit(KVM_ARM_VCPU_PSCI_0_2, vcpu->arch.features))
-		return KVM_ARM_PSCI_LATEST;
-
-	return KVM_ARM_PSCI_0_1;
-}
-
 static int kvm_psci_0_2_call(struct kvm_vcpu *vcpu)
 {
 	struct kvm *kvm = vcpu->kvm;
@@ -395,7 +387,7 @@ static int kvm_psci_0_1_call(struct kvm_vcpu *vcpu)
  */
 static int kvm_psci_call(struct kvm_vcpu *vcpu)
 {
-	switch (kvm_psci_version(vcpu)) {
+	switch (kvm_psci_version(vcpu, vcpu->kvm)) {
 	case KVM_ARM_PSCI_1_0:
 		return kvm_psci_1_0_call(vcpu);
 	case KVM_ARM_PSCI_0_2:
diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
index 3eab6ac18d7d..996328edd233 100644
--- a/arch/arm64/kvm/hyp/switch.c
+++ b/arch/arm64/kvm/hyp/switch.c
@@ -19,6 +19,8 @@
 #include <linux/jump_label.h>
 #include <uapi/linux/psci.h>
 
+#include <kvm/arm_psci.h>
+
 #include <asm/kvm_asm.h>
 #include <asm/kvm_emulate.h>
 #include <asm/kvm_hyp.h>
@@ -311,14 +313,16 @@ int __hyp_text __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 
 	if (exit_code == ARM_EXCEPTION_TRAP &&
 	    (kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC64 ||
-	     kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC32) &&
-	    vcpu_get_reg(vcpu, 0) == PSCI_0_2_FN_PSCI_VERSION) {
-		u64 val = PSCI_RET_NOT_SUPPORTED;
-		if (test_bit(KVM_ARM_VCPU_PSCI_0_2, vcpu->arch.features))
-			val = 2;
-
-		vcpu_set_reg(vcpu, 0, val);
-		goto again;
+	     kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC32)) {
+		u32 val = vcpu_get_reg(vcpu, 0);
+
+		if (val == PSCI_0_2_FN_PSCI_VERSION) {
+			val = kvm_psci_version(vcpu, kern_hyp_va(vcpu->kvm));
+			if (unlikely(val == KVM_ARM_PSCI_0_1))
+				val = PSCI_RET_NOT_SUPPORTED;
+			vcpu_set_reg(vcpu, 0, val);
+			goto again;
+		}
 	}
 
 	if (static_branch_unlikely(&vgic_v2_cpuif_trap) &&
diff --git a/include/kvm/arm_psci.h b/include/kvm/arm_psci.h
index ed1dd8088f1c..e518e4e3dfb5 100644
--- a/include/kvm/arm_psci.h
+++ b/include/kvm/arm_psci.h
@@ -18,6 +18,7 @@
 #ifndef __KVM_ARM_PSCI_H__
 #define __KVM_ARM_PSCI_H__
 
+#include <linux/kvm_host.h>
 #include <uapi/linux/psci.h>
 
 #define KVM_ARM_PSCI_0_1	PSCI_VERSION(0, 1)
@@ -26,7 +27,25 @@
 
 #define KVM_ARM_PSCI_LATEST	KVM_ARM_PSCI_1_0
 
-int kvm_psci_version(struct kvm_vcpu *vcpu);
+/*
+ * We need the KVM pointer independently from the vcpu as we can call
+ * this from HYP, and need to apply kern_hyp_va on it...
+ */
+static inline int kvm_psci_version(struct kvm_vcpu *vcpu, struct kvm *kvm)
+{
+	/*
+	 * Our PSCI implementation stays the same across versions from
+	 * v0.2 onward, only adding the few mandatory functions (such
+	 * as FEATURES with 1.0) that are required by newer
+	 * revisions. It is thus safe to return the latest.
+	 */
+	if (test_bit(KVM_ARM_VCPU_PSCI_0_2, vcpu->arch.features))
+		return KVM_ARM_PSCI_LATEST;
+
+	return KVM_ARM_PSCI_0_1;
+}
+
+
 int kvm_hvc_call_handler(struct kvm_vcpu *vcpu);
 
 #endif /* __KVM_ARM_PSCI_H__ */
-- 
2.11.0

[PATCH v4.9.y 35/42] arm64: KVM: Report SMCCC_ARCH_WORKAROUND_1 BP hardening support

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 6167ec5c9145cdf493722dfd80a5d48bafc4a18a upstream.

A new feature of SMCCC 1.1 is that it offers firmware-based CPU
workarounds. In particular, SMCCC_ARCH_WORKAROUND_1 provides
BP hardening for CVE-2017-5715.

If the host has some mitigation for this issue, report that
we deal with it using SMCCC_ARCH_WORKAROUND_1, as we apply the
host workaround on every guest exit.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
[v4.9: account for files moved to virt/ upstream]
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm/include/asm/kvm_host.h   | 6 ++++++
 arch/arm/kvm/psci.c               | 9 ++++++++-
 arch/arm64/include/asm/kvm_host.h | 5 +++++
 include/linux/arm-smccc.h         | 5 +++++
 4 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/arch/arm/include/asm/kvm_host.h b/arch/arm/include/asm/kvm_host.h
index d5423ab15ed5..9fe1043e72d2 100644
--- a/arch/arm/include/asm/kvm_host.h
+++ b/arch/arm/include/asm/kvm_host.h
@@ -318,4 +318,10 @@ static inline int kvm_arm_vcpu_arch_has_attr(struct kvm_vcpu *vcpu,
 	return -ENXIO;
 }
 
+static inline bool kvm_arm_harden_branch_predictor(void)
+{
+	/* No way to detect it yet, pretend it is not there. */
+	return false;
+}
+
 #endif /* __ARM_KVM_HOST_H__ */
diff --git a/arch/arm/kvm/psci.c b/arch/arm/kvm/psci.c
index 1120ad0bd3d2..3d962257c166 100644
--- a/arch/arm/kvm/psci.c
+++ b/arch/arm/kvm/psci.c
@@ -403,13 +403,20 @@ int kvm_hvc_call_handler(struct kvm_vcpu *vcpu)
 {
 	u32 func_id = smccc_get_function(vcpu);
 	u32 val = PSCI_RET_NOT_SUPPORTED;
+	u32 feature;
 
 	switch (func_id) {
 	case ARM_SMCCC_VERSION_FUNC_ID:
 		val = ARM_SMCCC_VERSION_1_1;
 		break;
 	case ARM_SMCCC_ARCH_FEATURES_FUNC_ID:
-		/* Nothing supported yet */
+		feature = smccc_get_arg1(vcpu);
+		switch(feature) {
+		case ARM_SMCCC_ARCH_WORKAROUND_1:
+			if (kvm_arm_harden_branch_predictor())
+				val = 0;
+			break;
+		}
 		break;
 	default:
 		return kvm_psci_call(vcpu);
diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index e5050388e062..37d56e85036e 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -393,4 +393,9 @@ static inline void __cpu_init_stage2(void)
 		  "PARange is %d bits, unsupported configuration!", parange);
 }
 
+static inline bool kvm_arm_harden_branch_predictor(void)
+{
+	return cpus_have_cap(ARM64_HARDEN_BRANCH_PREDICTOR);
+}
+
 #endif /* __ARM64_KVM_HOST_H__ */
diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
index dc68aa5a7261..e1ef944ef1da 100644
--- a/include/linux/arm-smccc.h
+++ b/include/linux/arm-smccc.h
@@ -73,6 +73,11 @@
 			   ARM_SMCCC_SMC_32,				\
 			   0, 1)
 
+#define ARM_SMCCC_ARCH_WORKAROUND_1					\
+	ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL,				\
+			   ARM_SMCCC_SMC_32,				\
+			   0, 0x8000)
+
 #ifndef __ASSEMBLY__
 
 #include <linux/linkage.h>
-- 
2.11.0

[PATCH v4.9.y 36/42] arm64: KVM: Add SMCCC_ARCH_WORKAROUND_1 fast handling

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit f72af90c3783d924337624659b43e2d36f1b36b4 upstream.

We want SMCCC_ARCH_WORKAROUND_1 to be fast. As fast as possible.
So let's intercept it as early as we can by testing for the
function call number as soon as we've identified a HVC call
coming from the guest.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kvm/hyp/hyp-entry.S | 20 ++++++++++++++++++--
 1 file changed, 18 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/kvm/hyp/hyp-entry.S b/arch/arm64/kvm/hyp/hyp-entry.S
index 4e92399f7105..4e9d50c3e658 100644
--- a/arch/arm64/kvm/hyp/hyp-entry.S
+++ b/arch/arm64/kvm/hyp/hyp-entry.S
@@ -15,6 +15,7 @@
  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 
+#include <linux/arm-smccc.h>
 #include <linux/linkage.h>
 
 #include <asm/alternative.h>
@@ -79,10 +80,11 @@ alternative_endif
 	lsr	x0, x1, #ESR_ELx_EC_SHIFT
 
 	cmp	x0, #ESR_ELx_EC_HVC64
+	ccmp	x0, #ESR_ELx_EC_HVC32, #4, ne
 	b.ne	el1_trap
 
-	mrs	x1, vttbr_el2		// If vttbr is valid, the 64bit guest
-	cbnz	x1, el1_trap		// called HVC
+	mrs	x1, vttbr_el2		// If vttbr is valid, the guest
+	cbnz	x1, el1_hvc_guest	// called HVC
 
 	/* Here, we're pretty sure the host called HVC. */
 	ldp	x0, x1, [sp], #16
@@ -101,6 +103,20 @@ alternative_endif
 
 2:	eret
 
+el1_hvc_guest:
+	/*
+	 * Fastest possible path for ARM_SMCCC_ARCH_WORKAROUND_1.
+	 * The workaround has already been applied on the host,
+	 * so let's quickly get back to the guest. We don't bother
+	 * restoring x1, as it can be clobbered anyway.
+	 */
+	ldr	x1, [sp]				// Guest's x0
+	eor	w1, w1, #ARM_SMCCC_ARCH_WORKAROUND_1
+	cbnz	w1, el1_trap
+	mov	x0, x1
+	add	sp, sp, #16
+	eret
+
 el1_trap:
 	/*
 	 * x0: ESR_EC
-- 
2.11.0

[PATCH v4.9.y 37/42] firmware/psci: Expose PSCI conduit

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 09a8d6d48499f93e2abde691f5800081cd858726 upstream.

In order to call into the firmware to apply workarounds, it is
useful to find out whether we're using HVC or SMC. Let's expose
this through the psci_ops.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 drivers/firmware/psci.c | 28 +++++++++++++++++++++++-----
 include/linux/psci.h    |  7 +++++++
 2 files changed, 30 insertions(+), 5 deletions(-)

diff --git a/drivers/firmware/psci.c b/drivers/firmware/psci.c
index 9a3ce76304b6..a49196ae0840 100644
--- a/drivers/firmware/psci.c
+++ b/drivers/firmware/psci.c
@@ -59,7 +59,9 @@ bool psci_tos_resident_on(int cpu)
 	return cpu == resident_cpu;
 }
 
-struct psci_operations psci_ops;
+struct psci_operations psci_ops = {
+	.conduit = PSCI_CONDUIT_NONE,
+};
 
 typedef unsigned long (psci_fn)(unsigned long, unsigned long,
 				unsigned long, unsigned long);
@@ -210,6 +212,22 @@ static unsigned long psci_migrate_info_up_cpu(void)
 			      0, 0, 0);
 }
 
+static void set_conduit(enum psci_conduit conduit)
+{
+	switch (conduit) {
+	case PSCI_CONDUIT_HVC:
+		invoke_psci_fn = __invoke_psci_fn_hvc;
+		break;
+	case PSCI_CONDUIT_SMC:
+		invoke_psci_fn = __invoke_psci_fn_smc;
+		break;
+	default:
+		WARN(1, "Unexpected PSCI conduit %d\n", conduit);
+	}
+
+	psci_ops.conduit = conduit;
+}
+
 static int get_set_conduit_method(struct device_node *np)
 {
 	const char *method;
@@ -222,9 +240,9 @@ static int get_set_conduit_method(struct device_node *np)
 	}
 
 	if (!strcmp("hvc", method)) {
-		invoke_psci_fn = __invoke_psci_fn_hvc;
+		set_conduit(PSCI_CONDUIT_HVC);
 	} else if (!strcmp("smc", method)) {
-		invoke_psci_fn = __invoke_psci_fn_smc;
+		set_conduit(PSCI_CONDUIT_SMC);
 	} else {
 		pr_warn("invalid \"method\" property: %s\n", method);
 		return -EINVAL;
@@ -654,9 +672,9 @@ int __init psci_acpi_init(void)
 	pr_info("probing for conduit method from ACPI.\n");
 
 	if (acpi_psci_use_hvc())
-		invoke_psci_fn = __invoke_psci_fn_hvc;
+		set_conduit(PSCI_CONDUIT_HVC);
 	else
-		invoke_psci_fn = __invoke_psci_fn_smc;
+		set_conduit(PSCI_CONDUIT_SMC);
 
 	return psci_probe();
 }
diff --git a/include/linux/psci.h b/include/linux/psci.h
index 6306ab10af18..66ff54787d3e 100644
--- a/include/linux/psci.h
+++ b/include/linux/psci.h
@@ -25,6 +25,12 @@ bool psci_tos_resident_on(int cpu);
 int psci_cpu_init_idle(unsigned int cpu);
 int psci_cpu_suspend_enter(unsigned long index);
 
+enum psci_conduit {
+	PSCI_CONDUIT_NONE,
+	PSCI_CONDUIT_SMC,
+	PSCI_CONDUIT_HVC,
+};
+
 struct psci_operations {
 	u32 (*get_version)(void);
 	int (*cpu_suspend)(u32 state, unsigned long entry_point);
@@ -34,6 +40,7 @@ struct psci_operations {
 	int (*affinity_info)(unsigned long target_affinity,
 			unsigned long lowest_affinity_level);
 	int (*migrate_info_type)(void);
+	enum psci_conduit conduit;
 };
 
 extern struct psci_operations psci_ops;
-- 
2.11.0

[PATCH v4.9.y 38/42] firmware/psci: Expose SMCCC version through psci_ops

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit e78eef554a912ef6c1e0bbf97619dafbeae3339f upstream.

Since PSCI 1.0 allows the SMCCC version to be (indirectly) probed,
let's do that at boot time, and expose the version of the calling
convention as part of the psci_ops structure.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 drivers/firmware/psci.c | 27 +++++++++++++++++++++++++++
 include/linux/psci.h    |  6 ++++++
 2 files changed, 33 insertions(+)

diff --git a/drivers/firmware/psci.c b/drivers/firmware/psci.c
index a49196ae0840..79a48c37fb35 100644
--- a/drivers/firmware/psci.c
+++ b/drivers/firmware/psci.c
@@ -61,6 +61,7 @@ bool psci_tos_resident_on(int cpu)
 
 struct psci_operations psci_ops = {
 	.conduit = PSCI_CONDUIT_NONE,
+	.smccc_version = SMCCC_VERSION_1_0,
 };
 
 typedef unsigned long (psci_fn)(unsigned long, unsigned long,
@@ -511,6 +512,31 @@ static void __init psci_init_migrate(void)
 	pr_info("Trusted OS resident on physical CPU 0x%lx\n", cpuid);
 }
 
+static void __init psci_init_smccc(void)
+{
+	u32 ver = ARM_SMCCC_VERSION_1_0;
+	int feature;
+
+	feature = psci_features(ARM_SMCCC_VERSION_FUNC_ID);
+
+	if (feature != PSCI_RET_NOT_SUPPORTED) {
+		u32 ret;
+		ret = invoke_psci_fn(ARM_SMCCC_VERSION_FUNC_ID, 0, 0, 0);
+		if (ret == ARM_SMCCC_VERSION_1_1) {
+			psci_ops.smccc_version = SMCCC_VERSION_1_1;
+			ver = ret;
+		}
+	}
+
+	/*
+	 * Conveniently, the SMCCC and PSCI versions are encoded the
+	 * same way. No, this isn't accidental.
+	 */
+	pr_info("SMC Calling Convention v%d.%d\n",
+		PSCI_VERSION_MAJOR(ver), PSCI_VERSION_MINOR(ver));
+
+}
+
 static void __init psci_0_2_set_functions(void)
 {
 	pr_info("Using standard PSCI v0.2 function IDs\n");
@@ -559,6 +585,7 @@ static int __init psci_probe(void)
 	psci_init_migrate();
 
 	if (PSCI_VERSION_MAJOR(ver) >= 1) {
+		psci_init_smccc();
 		psci_init_cpu_suspend();
 		psci_init_system_suspend();
 	}
diff --git a/include/linux/psci.h b/include/linux/psci.h
index 66ff54787d3e..347077cf19c6 100644
--- a/include/linux/psci.h
+++ b/include/linux/psci.h
@@ -31,6 +31,11 @@ enum psci_conduit {
 	PSCI_CONDUIT_HVC,
 };
 
+enum smccc_version {
+	SMCCC_VERSION_1_0,
+	SMCCC_VERSION_1_1,
+};
+
 struct psci_operations {
 	u32 (*get_version)(void);
 	int (*cpu_suspend)(u32 state, unsigned long entry_point);
@@ -41,6 +46,7 @@ struct psci_operations {
 			unsigned long lowest_affinity_level);
 	int (*migrate_info_type)(void);
 	enum psci_conduit conduit;
+	enum smccc_version smccc_version;
 };
 
 extern struct psci_operations psci_ops;
-- 
2.11.0

[PATCH v4.9.y 39/42] arm/arm64: smccc: Make function identifiers an unsigned quantity

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit ded4c39e93f3b72968fdb79baba27f3b83dad34c upstream.

Function identifiers are a 32bit, unsigned quantity. But we never
tell so to the compiler, resulting in the following:

 4ac:   b26187e0        mov     x0, #0xffffffff80000001

We thus rely on the firmware narrowing it for us, which is not
always a reasonable expectation.

Cc: stable@vger.kernel.org
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 include/linux/arm-smccc.h | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
index e1ef944ef1da..dd44d8458c04 100644
--- a/include/linux/arm-smccc.h
+++ b/include/linux/arm-smccc.h
@@ -14,14 +14,16 @@
 #ifndef __LINUX_ARM_SMCCC_H
 #define __LINUX_ARM_SMCCC_H
 
+#include <uapi/linux/const.h>
+
 /*
  * This file provides common defines for ARM SMC Calling Convention as
  * specified in
  * http://infocenter.arm.com/help/topic/com.arm.doc.den0028a/index.html
  */
 
-#define ARM_SMCCC_STD_CALL		0
-#define ARM_SMCCC_FAST_CALL		1
+#define ARM_SMCCC_STD_CALL	        _AC(0,U)
+#define ARM_SMCCC_FAST_CALL	        _AC(1,U)
 #define ARM_SMCCC_TYPE_SHIFT		31
 
 #define ARM_SMCCC_SMC_32		0
-- 
2.11.0

[PATCH v4.9.y 40/42] arm/arm64: smccc: Implement SMCCC v1.1 inline primitive

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit f2d3b2e8759a5833df6f022e42df2d581e6d843c upstream.

One of the major improvement of SMCCC v1.1 is that it only clobbers
the first 4 registers, both on 32 and 64bit. This means that it
becomes very easy to provide an inline version of the SMC call
primitive, and avoid performing a function call to stash the
registers that would otherwise be clobbered by SMCCC v1.0.

Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 include/linux/arm-smccc.h | 141 ++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 141 insertions(+)

diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h
index dd44d8458c04..a031897fca76 100644
--- a/include/linux/arm-smccc.h
+++ b/include/linux/arm-smccc.h
@@ -150,5 +150,146 @@ asmlinkage void __arm_smccc_hvc(unsigned long a0, unsigned long a1,
 
 #define arm_smccc_hvc_quirk(...) __arm_smccc_hvc(__VA_ARGS__)
 
+/* SMCCC v1.1 implementation madness follows */
+#ifdef CONFIG_ARM64
+
+#define SMCCC_SMC_INST	"smc	#0"
+#define SMCCC_HVC_INST	"hvc	#0"
+
+#elif defined(CONFIG_ARM)
+#include <asm/opcodes-sec.h>
+#include <asm/opcodes-virt.h>
+
+#define SMCCC_SMC_INST	__SMC(0)
+#define SMCCC_HVC_INST	__HVC(0)
+
+#endif
+
+#define ___count_args(_0, _1, _2, _3, _4, _5, _6, _7, _8, x, ...) x
+
+#define __count_args(...)						\
+	___count_args(__VA_ARGS__, 7, 6, 5, 4, 3, 2, 1, 0)
+
+#define __constraint_write_0						\
+	"+r" (r0), "=&r" (r1), "=&r" (r2), "=&r" (r3)
+#define __constraint_write_1						\
+	"+r" (r0), "+r" (r1), "=&r" (r2), "=&r" (r3)
+#define __constraint_write_2						\
+	"+r" (r0), "+r" (r1), "+r" (r2), "=&r" (r3)
+#define __constraint_write_3						\
+	"+r" (r0), "+r" (r1), "+r" (r2), "+r" (r3)
+#define __constraint_write_4	__constraint_write_3
+#define __constraint_write_5	__constraint_write_4
+#define __constraint_write_6	__constraint_write_5
+#define __constraint_write_7	__constraint_write_6
+
+#define __constraint_read_0
+#define __constraint_read_1
+#define __constraint_read_2
+#define __constraint_read_3
+#define __constraint_read_4	"r" (r4)
+#define __constraint_read_5	__constraint_read_4, "r" (r5)
+#define __constraint_read_6	__constraint_read_5, "r" (r6)
+#define __constraint_read_7	__constraint_read_6, "r" (r7)
+
+#define __declare_arg_0(a0, res)					\
+	struct arm_smccc_res   *___res = res;				\
+	register u32           r0 asm("r0") = a0;			\
+	register unsigned long r1 asm("r1");				\
+	register unsigned long r2 asm("r2");				\
+	register unsigned long r3 asm("r3")
+
+#define __declare_arg_1(a0, a1, res)					\
+	struct arm_smccc_res   *___res = res;				\
+	register u32           r0 asm("r0") = a0;			\
+	register typeof(a1)    r1 asm("r1") = a1;			\
+	register unsigned long r2 asm("r2");				\
+	register unsigned long r3 asm("r3")
+
+#define __declare_arg_2(a0, a1, a2, res)				\
+	struct arm_smccc_res   *___res = res;				\
+	register u32           r0 asm("r0") = a0;			\
+	register typeof(a1)    r1 asm("r1") = a1;			\
+	register typeof(a2)    r2 asm("r2") = a2;			\
+	register unsigned long r3 asm("r3")
+
+#define __declare_arg_3(a0, a1, a2, a3, res)				\
+	struct arm_smccc_res   *___res = res;				\
+	register u32           r0 asm("r0") = a0;			\
+	register typeof(a1)    r1 asm("r1") = a1;			\
+	register typeof(a2)    r2 asm("r2") = a2;			\
+	register typeof(a3)    r3 asm("r3") = a3
+
+#define __declare_arg_4(a0, a1, a2, a3, a4, res)			\
+	__declare_arg_3(a0, a1, a2, a3, res);				\
+	register typeof(a4) r4 asm("r4") = a4
+
+#define __declare_arg_5(a0, a1, a2, a3, a4, a5, res)			\
+	__declare_arg_4(a0, a1, a2, a3, a4, res);			\
+	register typeof(a5) r5 asm("r5") = a5
+
+#define __declare_arg_6(a0, a1, a2, a3, a4, a5, a6, res)		\
+	__declare_arg_5(a0, a1, a2, a3, a4, a5, res);			\
+	register typeof(a6) r6 asm("r6") = a6
+
+#define __declare_arg_7(a0, a1, a2, a3, a4, a5, a6, a7, res)		\
+	__declare_arg_6(a0, a1, a2, a3, a4, a5, a6, res);		\
+	register typeof(a7) r7 asm("r7") = a7
+
+#define ___declare_args(count, ...) __declare_arg_ ## count(__VA_ARGS__)
+#define __declare_args(count, ...)  ___declare_args(count, __VA_ARGS__)
+
+#define ___constraints(count)						\
+	: __constraint_write_ ## count					\
+	: __constraint_read_ ## count					\
+	: "memory"
+#define __constraints(count)	___constraints(count)
+
+/*
+ * We have an output list that is not necessarily used, and GCC feels
+ * entitled to optimise the whole sequence away. "volatile" is what
+ * makes it stick.
+ */
+#define __arm_smccc_1_1(inst, ...)					\
+	do {								\
+		__declare_args(__count_args(__VA_ARGS__), __VA_ARGS__);	\
+		asm volatile(inst "\n"					\
+			     __constraints(__count_args(__VA_ARGS__)));	\
+		if (___res)						\
+			*___res = (typeof(*___res)){r0, r1, r2, r3};	\
+	} while (0)
+
+/*
+ * arm_smccc_1_1_smc() - make an SMCCC v1.1 compliant SMC call
+ *
+ * This is a variadic macro taking one to eight source arguments, and
+ * an optional return structure.
+ *
+ * @a0-a7: arguments passed in registers 0 to 7
+ * @res: result values from registers 0 to 3
+ *
+ * This macro is used to make SMC calls following SMC Calling Convention v1.1.
+ * The content of the supplied param are copied to registers 0 to 7 prior
+ * to the SMC instruction. The return values are updated with the content
+ * from register 0 to 3 on return from the SMC instruction if not NULL.
+ */
+#define arm_smccc_1_1_smc(...)	__arm_smccc_1_1(SMCCC_SMC_INST, __VA_ARGS__)
+
+/*
+ * arm_smccc_1_1_hvc() - make an SMCCC v1.1 compliant HVC call
+ *
+ * This is a variadic macro taking one to eight source arguments, and
+ * an optional return structure.
+ *
+ * @a0-a7: arguments passed in registers 0 to 7
+ * @res: result values from registers 0 to 3
+ *
+ * This macro is used to make HVC calls following SMC Calling Convention v1.1.
+ * The content of the supplied param are copied to registers 0 to 7 prior
+ * to the HVC instruction. The return values are updated with the content
+ * from register 0 to 3 on return from the HVC instruction if not NULL.
+ */
+#define arm_smccc_1_1_hvc(...)	__arm_smccc_1_1(SMCCC_HVC_INST, __VA_ARGS__)
+
 #endif /*__ASSEMBLY__*/
 #endif /*__LINUX_ARM_SMCCC_H*/
-- 
2.11.0

[PATCH v4.9.y 41/42] arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit b092201e0020614127f495c092e0a12d26a2116e upstream.

Add the detection and runtime code for ARM_SMCCC_ARCH_WORKAROUND_1.
It is lovely. Really.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/crypto/sha256-core.S | 2061 +++++++++++++++++++++++++++++++++++++++
 arch/arm64/crypto/sha512-core.S | 1085 +++++++++++++++++++++
 arch/arm64/kernel/bpi.S         |   20 +
 arch/arm64/kernel/cpu_errata.c  |   72 +-
 4 files changed, 3235 insertions(+), 3 deletions(-)
 create mode 100644 arch/arm64/crypto/sha256-core.S
 create mode 100644 arch/arm64/crypto/sha512-core.S

diff --git a/arch/arm64/crypto/sha256-core.S b/arch/arm64/crypto/sha256-core.S
new file mode 100644
index 000000000000..3ce82cc860bc
--- /dev/null
+++ b/arch/arm64/crypto/sha256-core.S
@@ -0,0 +1,2061 @@
+// Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
+//
+// Licensed under the OpenSSL license (the "License").  You may not use
+// this file except in compliance with the License.  You can obtain a copy
+// in the file LICENSE in the source distribution or at
+// https://www.openssl.org/source/license.html
+
+// ====================================================================
+// Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+// project. The module is, however, dual licensed under OpenSSL and
+// CRYPTOGAMS licenses depending on where you obtain it. For further
+// details see http://www.openssl.org/~appro/cryptogams/.
+//
+// Permission to use under GPLv2 terms is granted.
+// ====================================================================
+//
+// SHA256/512 for ARMv8.
+//
+// Performance in cycles per processed byte and improvement coefficient
+// over code generated with "default" compiler:
+//
+//		SHA256-hw	SHA256(*)	SHA512
+// Apple A7	1.97		10.5 (+33%)	6.73 (-1%(**))
+// Cortex-A53	2.38		15.5 (+115%)	10.0 (+150%(***))
+// Cortex-A57	2.31		11.6 (+86%)	7.51 (+260%(***))
+// Denver	2.01		10.5 (+26%)	6.70 (+8%)
+// X-Gene			20.0 (+100%)	12.8 (+300%(***))
+// Mongoose	2.36		13.0 (+50%)	8.36 (+33%)
+//
+// (*)	Software SHA256 results are of lesser relevance, presented
+//	mostly for informational purposes.
+// (**)	The result is a trade-off: it's possible to improve it by
+//	10% (or by 1 cycle per round), but at the cost of 20% loss
+//	on Cortex-A53 (or by 4 cycles per round).
+// (***)	Super-impressive coefficients over gcc-generated code are
+//	indication of some compiler "pathology", most notably code
+//	generated with -mgeneral-regs-only is significanty faster
+//	and the gap is only 40-90%.
+//
+// October 2016.
+//
+// Originally it was reckoned that it makes no sense to implement NEON
+// version of SHA256 for 64-bit processors. This is because performance
+// improvement on most wide-spread Cortex-A5x processors was observed
+// to be marginal, same on Cortex-A53 and ~10% on A57. But then it was
+// observed that 32-bit NEON SHA256 performs significantly better than
+// 64-bit scalar version on *some* of the more recent processors. As
+// result 64-bit NEON version of SHA256 was added to provide best
+// all-round performance. For example it executes ~30% faster on X-Gene
+// and Mongoose. [For reference, NEON version of SHA512 is bound to
+// deliver much less improvement, likely *negative* on Cortex-A5x.
+// Which is why NEON support is limited to SHA256.]
+
+#ifndef	__KERNEL__
+# include "arm_arch.h"
+#endif
+
+.text
+
+.extern	OPENSSL_armcap_P
+.globl	sha256_block_data_order
+.type	sha256_block_data_order,%function
+.align	6
+sha256_block_data_order:
+#ifndef	__KERNEL__
+# ifdef	__ILP32__
+	ldrsw	x16,.LOPENSSL_armcap_P
+# else
+	ldr	x16,.LOPENSSL_armcap_P
+# endif
+	adr	x17,.LOPENSSL_armcap_P
+	add	x16,x16,x17
+	ldr	w16,[x16]
+	tst	w16,#ARMV8_SHA256
+	b.ne	.Lv8_entry
+	tst	w16,#ARMV7_NEON
+	b.ne	.Lneon_entry
+#endif
+	stp	x29,x30,[sp,#-128]!
+	add	x29,sp,#0
+
+	stp	x19,x20,[sp,#16]
+	stp	x21,x22,[sp,#32]
+	stp	x23,x24,[sp,#48]
+	stp	x25,x26,[sp,#64]
+	stp	x27,x28,[sp,#80]
+	sub	sp,sp,#4*4
+
+	ldp	w20,w21,[x0]				// load context
+	ldp	w22,w23,[x0,#2*4]
+	ldp	w24,w25,[x0,#4*4]
+	add	x2,x1,x2,lsl#6	// end of input
+	ldp	w26,w27,[x0,#6*4]
+	adr	x30,.LK256
+	stp	x0,x2,[x29,#96]
+
+.Loop:
+	ldp	w3,w4,[x1],#2*4
+	ldr	w19,[x30],#4			// *K++
+	eor	w28,w21,w22				// magic seed
+	str	x1,[x29,#112]
+#ifndef	__AARCH64EB__
+	rev	w3,w3			// 0
+#endif
+	ror	w16,w24,#6
+	add	w27,w27,w19			// h+=K[i]
+	eor	w6,w24,w24,ror#14
+	and	w17,w25,w24
+	bic	w19,w26,w24
+	add	w27,w27,w3			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w20,w21			// a^b, b^c in next round
+	eor	w16,w16,w6,ror#11	// Sigma1(e)
+	ror	w6,w20,#2
+	add	w27,w27,w17			// h+=Ch(e,f,g)
+	eor	w17,w20,w20,ror#9
+	add	w27,w27,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w23,w23,w27			// d+=h
+	eor	w28,w28,w21			// Maj(a,b,c)
+	eor	w17,w6,w17,ror#13	// Sigma0(a)
+	add	w27,w27,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w27,w27,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w4,w4			// 1
+#endif
+	ldp	w5,w6,[x1],#2*4
+	add	w27,w27,w17			// h+=Sigma0(a)
+	ror	w16,w23,#6
+	add	w26,w26,w28			// h+=K[i]
+	eor	w7,w23,w23,ror#14
+	and	w17,w24,w23
+	bic	w28,w25,w23
+	add	w26,w26,w4			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w27,w20			// a^b, b^c in next round
+	eor	w16,w16,w7,ror#11	// Sigma1(e)
+	ror	w7,w27,#2
+	add	w26,w26,w17			// h+=Ch(e,f,g)
+	eor	w17,w27,w27,ror#9
+	add	w26,w26,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w22,w22,w26			// d+=h
+	eor	w19,w19,w20			// Maj(a,b,c)
+	eor	w17,w7,w17,ror#13	// Sigma0(a)
+	add	w26,w26,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w26,w26,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w5,w5			// 2
+#endif
+	add	w26,w26,w17			// h+=Sigma0(a)
+	ror	w16,w22,#6
+	add	w25,w25,w19			// h+=K[i]
+	eor	w8,w22,w22,ror#14
+	and	w17,w23,w22
+	bic	w19,w24,w22
+	add	w25,w25,w5			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w26,w27			// a^b, b^c in next round
+	eor	w16,w16,w8,ror#11	// Sigma1(e)
+	ror	w8,w26,#2
+	add	w25,w25,w17			// h+=Ch(e,f,g)
+	eor	w17,w26,w26,ror#9
+	add	w25,w25,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w21,w21,w25			// d+=h
+	eor	w28,w28,w27			// Maj(a,b,c)
+	eor	w17,w8,w17,ror#13	// Sigma0(a)
+	add	w25,w25,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w25,w25,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w6,w6			// 3
+#endif
+	ldp	w7,w8,[x1],#2*4
+	add	w25,w25,w17			// h+=Sigma0(a)
+	ror	w16,w21,#6
+	add	w24,w24,w28			// h+=K[i]
+	eor	w9,w21,w21,ror#14
+	and	w17,w22,w21
+	bic	w28,w23,w21
+	add	w24,w24,w6			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w25,w26			// a^b, b^c in next round
+	eor	w16,w16,w9,ror#11	// Sigma1(e)
+	ror	w9,w25,#2
+	add	w24,w24,w17			// h+=Ch(e,f,g)
+	eor	w17,w25,w25,ror#9
+	add	w24,w24,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w20,w20,w24			// d+=h
+	eor	w19,w19,w26			// Maj(a,b,c)
+	eor	w17,w9,w17,ror#13	// Sigma0(a)
+	add	w24,w24,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w24,w24,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w7,w7			// 4
+#endif
+	add	w24,w24,w17			// h+=Sigma0(a)
+	ror	w16,w20,#6
+	add	w23,w23,w19			// h+=K[i]
+	eor	w10,w20,w20,ror#14
+	and	w17,w21,w20
+	bic	w19,w22,w20
+	add	w23,w23,w7			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w24,w25			// a^b, b^c in next round
+	eor	w16,w16,w10,ror#11	// Sigma1(e)
+	ror	w10,w24,#2
+	add	w23,w23,w17			// h+=Ch(e,f,g)
+	eor	w17,w24,w24,ror#9
+	add	w23,w23,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w27,w27,w23			// d+=h
+	eor	w28,w28,w25			// Maj(a,b,c)
+	eor	w17,w10,w17,ror#13	// Sigma0(a)
+	add	w23,w23,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w23,w23,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w8,w8			// 5
+#endif
+	ldp	w9,w10,[x1],#2*4
+	add	w23,w23,w17			// h+=Sigma0(a)
+	ror	w16,w27,#6
+	add	w22,w22,w28			// h+=K[i]
+	eor	w11,w27,w27,ror#14
+	and	w17,w20,w27
+	bic	w28,w21,w27
+	add	w22,w22,w8			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w23,w24			// a^b, b^c in next round
+	eor	w16,w16,w11,ror#11	// Sigma1(e)
+	ror	w11,w23,#2
+	add	w22,w22,w17			// h+=Ch(e,f,g)
+	eor	w17,w23,w23,ror#9
+	add	w22,w22,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w26,w26,w22			// d+=h
+	eor	w19,w19,w24			// Maj(a,b,c)
+	eor	w17,w11,w17,ror#13	// Sigma0(a)
+	add	w22,w22,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w22,w22,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w9,w9			// 6
+#endif
+	add	w22,w22,w17			// h+=Sigma0(a)
+	ror	w16,w26,#6
+	add	w21,w21,w19			// h+=K[i]
+	eor	w12,w26,w26,ror#14
+	and	w17,w27,w26
+	bic	w19,w20,w26
+	add	w21,w21,w9			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w22,w23			// a^b, b^c in next round
+	eor	w16,w16,w12,ror#11	// Sigma1(e)
+	ror	w12,w22,#2
+	add	w21,w21,w17			// h+=Ch(e,f,g)
+	eor	w17,w22,w22,ror#9
+	add	w21,w21,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w25,w25,w21			// d+=h
+	eor	w28,w28,w23			// Maj(a,b,c)
+	eor	w17,w12,w17,ror#13	// Sigma0(a)
+	add	w21,w21,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w21,w21,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w10,w10			// 7
+#endif
+	ldp	w11,w12,[x1],#2*4
+	add	w21,w21,w17			// h+=Sigma0(a)
+	ror	w16,w25,#6
+	add	w20,w20,w28			// h+=K[i]
+	eor	w13,w25,w25,ror#14
+	and	w17,w26,w25
+	bic	w28,w27,w25
+	add	w20,w20,w10			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w21,w22			// a^b, b^c in next round
+	eor	w16,w16,w13,ror#11	// Sigma1(e)
+	ror	w13,w21,#2
+	add	w20,w20,w17			// h+=Ch(e,f,g)
+	eor	w17,w21,w21,ror#9
+	add	w20,w20,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w24,w24,w20			// d+=h
+	eor	w19,w19,w22			// Maj(a,b,c)
+	eor	w17,w13,w17,ror#13	// Sigma0(a)
+	add	w20,w20,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w20,w20,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w11,w11			// 8
+#endif
+	add	w20,w20,w17			// h+=Sigma0(a)
+	ror	w16,w24,#6
+	add	w27,w27,w19			// h+=K[i]
+	eor	w14,w24,w24,ror#14
+	and	w17,w25,w24
+	bic	w19,w26,w24
+	add	w27,w27,w11			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w20,w21			// a^b, b^c in next round
+	eor	w16,w16,w14,ror#11	// Sigma1(e)
+	ror	w14,w20,#2
+	add	w27,w27,w17			// h+=Ch(e,f,g)
+	eor	w17,w20,w20,ror#9
+	add	w27,w27,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w23,w23,w27			// d+=h
+	eor	w28,w28,w21			// Maj(a,b,c)
+	eor	w17,w14,w17,ror#13	// Sigma0(a)
+	add	w27,w27,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w27,w27,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w12,w12			// 9
+#endif
+	ldp	w13,w14,[x1],#2*4
+	add	w27,w27,w17			// h+=Sigma0(a)
+	ror	w16,w23,#6
+	add	w26,w26,w28			// h+=K[i]
+	eor	w15,w23,w23,ror#14
+	and	w17,w24,w23
+	bic	w28,w25,w23
+	add	w26,w26,w12			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w27,w20			// a^b, b^c in next round
+	eor	w16,w16,w15,ror#11	// Sigma1(e)
+	ror	w15,w27,#2
+	add	w26,w26,w17			// h+=Ch(e,f,g)
+	eor	w17,w27,w27,ror#9
+	add	w26,w26,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w22,w22,w26			// d+=h
+	eor	w19,w19,w20			// Maj(a,b,c)
+	eor	w17,w15,w17,ror#13	// Sigma0(a)
+	add	w26,w26,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w26,w26,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w13,w13			// 10
+#endif
+	add	w26,w26,w17			// h+=Sigma0(a)
+	ror	w16,w22,#6
+	add	w25,w25,w19			// h+=K[i]
+	eor	w0,w22,w22,ror#14
+	and	w17,w23,w22
+	bic	w19,w24,w22
+	add	w25,w25,w13			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w26,w27			// a^b, b^c in next round
+	eor	w16,w16,w0,ror#11	// Sigma1(e)
+	ror	w0,w26,#2
+	add	w25,w25,w17			// h+=Ch(e,f,g)
+	eor	w17,w26,w26,ror#9
+	add	w25,w25,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w21,w21,w25			// d+=h
+	eor	w28,w28,w27			// Maj(a,b,c)
+	eor	w17,w0,w17,ror#13	// Sigma0(a)
+	add	w25,w25,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w25,w25,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w14,w14			// 11
+#endif
+	ldp	w15,w0,[x1],#2*4
+	add	w25,w25,w17			// h+=Sigma0(a)
+	str	w6,[sp,#12]
+	ror	w16,w21,#6
+	add	w24,w24,w28			// h+=K[i]
+	eor	w6,w21,w21,ror#14
+	and	w17,w22,w21
+	bic	w28,w23,w21
+	add	w24,w24,w14			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w25,w26			// a^b, b^c in next round
+	eor	w16,w16,w6,ror#11	// Sigma1(e)
+	ror	w6,w25,#2
+	add	w24,w24,w17			// h+=Ch(e,f,g)
+	eor	w17,w25,w25,ror#9
+	add	w24,w24,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w20,w20,w24			// d+=h
+	eor	w19,w19,w26			// Maj(a,b,c)
+	eor	w17,w6,w17,ror#13	// Sigma0(a)
+	add	w24,w24,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w24,w24,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w15,w15			// 12
+#endif
+	add	w24,w24,w17			// h+=Sigma0(a)
+	str	w7,[sp,#0]
+	ror	w16,w20,#6
+	add	w23,w23,w19			// h+=K[i]
+	eor	w7,w20,w20,ror#14
+	and	w17,w21,w20
+	bic	w19,w22,w20
+	add	w23,w23,w15			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w24,w25			// a^b, b^c in next round
+	eor	w16,w16,w7,ror#11	// Sigma1(e)
+	ror	w7,w24,#2
+	add	w23,w23,w17			// h+=Ch(e,f,g)
+	eor	w17,w24,w24,ror#9
+	add	w23,w23,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w27,w27,w23			// d+=h
+	eor	w28,w28,w25			// Maj(a,b,c)
+	eor	w17,w7,w17,ror#13	// Sigma0(a)
+	add	w23,w23,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w23,w23,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w0,w0			// 13
+#endif
+	ldp	w1,w2,[x1]
+	add	w23,w23,w17			// h+=Sigma0(a)
+	str	w8,[sp,#4]
+	ror	w16,w27,#6
+	add	w22,w22,w28			// h+=K[i]
+	eor	w8,w27,w27,ror#14
+	and	w17,w20,w27
+	bic	w28,w21,w27
+	add	w22,w22,w0			// h+=X[i]
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w23,w24			// a^b, b^c in next round
+	eor	w16,w16,w8,ror#11	// Sigma1(e)
+	ror	w8,w23,#2
+	add	w22,w22,w17			// h+=Ch(e,f,g)
+	eor	w17,w23,w23,ror#9
+	add	w22,w22,w16			// h+=Sigma1(e)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	add	w26,w26,w22			// d+=h
+	eor	w19,w19,w24			// Maj(a,b,c)
+	eor	w17,w8,w17,ror#13	// Sigma0(a)
+	add	w22,w22,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	//add	w22,w22,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w1,w1			// 14
+#endif
+	ldr	w6,[sp,#12]
+	add	w22,w22,w17			// h+=Sigma0(a)
+	str	w9,[sp,#8]
+	ror	w16,w26,#6
+	add	w21,w21,w19			// h+=K[i]
+	eor	w9,w26,w26,ror#14
+	and	w17,w27,w26
+	bic	w19,w20,w26
+	add	w21,w21,w1			// h+=X[i]
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w22,w23			// a^b, b^c in next round
+	eor	w16,w16,w9,ror#11	// Sigma1(e)
+	ror	w9,w22,#2
+	add	w21,w21,w17			// h+=Ch(e,f,g)
+	eor	w17,w22,w22,ror#9
+	add	w21,w21,w16			// h+=Sigma1(e)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	add	w25,w25,w21			// d+=h
+	eor	w28,w28,w23			// Maj(a,b,c)
+	eor	w17,w9,w17,ror#13	// Sigma0(a)
+	add	w21,w21,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	//add	w21,w21,w17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	w2,w2			// 15
+#endif
+	ldr	w7,[sp,#0]
+	add	w21,w21,w17			// h+=Sigma0(a)
+	str	w10,[sp,#12]
+	ror	w16,w25,#6
+	add	w20,w20,w28			// h+=K[i]
+	ror	w9,w4,#7
+	and	w17,w26,w25
+	ror	w8,w1,#17
+	bic	w28,w27,w25
+	ror	w10,w21,#2
+	add	w20,w20,w2			// h+=X[i]
+	eor	w16,w16,w25,ror#11
+	eor	w9,w9,w4,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w21,w22			// a^b, b^c in next round
+	eor	w16,w16,w25,ror#25	// Sigma1(e)
+	eor	w10,w10,w21,ror#13
+	add	w20,w20,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w8,w8,w1,ror#19
+	eor	w9,w9,w4,lsr#3	// sigma0(X[i+1])
+	add	w20,w20,w16			// h+=Sigma1(e)
+	eor	w19,w19,w22			// Maj(a,b,c)
+	eor	w17,w10,w21,ror#22	// Sigma0(a)
+	eor	w8,w8,w1,lsr#10	// sigma1(X[i+14])
+	add	w3,w3,w12
+	add	w24,w24,w20			// d+=h
+	add	w20,w20,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w3,w3,w9
+	add	w20,w20,w17			// h+=Sigma0(a)
+	add	w3,w3,w8
+.Loop_16_xx:
+	ldr	w8,[sp,#4]
+	str	w11,[sp,#0]
+	ror	w16,w24,#6
+	add	w27,w27,w19			// h+=K[i]
+	ror	w10,w5,#7
+	and	w17,w25,w24
+	ror	w9,w2,#17
+	bic	w19,w26,w24
+	ror	w11,w20,#2
+	add	w27,w27,w3			// h+=X[i]
+	eor	w16,w16,w24,ror#11
+	eor	w10,w10,w5,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w20,w21			// a^b, b^c in next round
+	eor	w16,w16,w24,ror#25	// Sigma1(e)
+	eor	w11,w11,w20,ror#13
+	add	w27,w27,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w9,w9,w2,ror#19
+	eor	w10,w10,w5,lsr#3	// sigma0(X[i+1])
+	add	w27,w27,w16			// h+=Sigma1(e)
+	eor	w28,w28,w21			// Maj(a,b,c)
+	eor	w17,w11,w20,ror#22	// Sigma0(a)
+	eor	w9,w9,w2,lsr#10	// sigma1(X[i+14])
+	add	w4,w4,w13
+	add	w23,w23,w27			// d+=h
+	add	w27,w27,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w4,w4,w10
+	add	w27,w27,w17			// h+=Sigma0(a)
+	add	w4,w4,w9
+	ldr	w9,[sp,#8]
+	str	w12,[sp,#4]
+	ror	w16,w23,#6
+	add	w26,w26,w28			// h+=K[i]
+	ror	w11,w6,#7
+	and	w17,w24,w23
+	ror	w10,w3,#17
+	bic	w28,w25,w23
+	ror	w12,w27,#2
+	add	w26,w26,w4			// h+=X[i]
+	eor	w16,w16,w23,ror#11
+	eor	w11,w11,w6,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w27,w20			// a^b, b^c in next round
+	eor	w16,w16,w23,ror#25	// Sigma1(e)
+	eor	w12,w12,w27,ror#13
+	add	w26,w26,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w10,w10,w3,ror#19
+	eor	w11,w11,w6,lsr#3	// sigma0(X[i+1])
+	add	w26,w26,w16			// h+=Sigma1(e)
+	eor	w19,w19,w20			// Maj(a,b,c)
+	eor	w17,w12,w27,ror#22	// Sigma0(a)
+	eor	w10,w10,w3,lsr#10	// sigma1(X[i+14])
+	add	w5,w5,w14
+	add	w22,w22,w26			// d+=h
+	add	w26,w26,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w5,w5,w11
+	add	w26,w26,w17			// h+=Sigma0(a)
+	add	w5,w5,w10
+	ldr	w10,[sp,#12]
+	str	w13,[sp,#8]
+	ror	w16,w22,#6
+	add	w25,w25,w19			// h+=K[i]
+	ror	w12,w7,#7
+	and	w17,w23,w22
+	ror	w11,w4,#17
+	bic	w19,w24,w22
+	ror	w13,w26,#2
+	add	w25,w25,w5			// h+=X[i]
+	eor	w16,w16,w22,ror#11
+	eor	w12,w12,w7,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w26,w27			// a^b, b^c in next round
+	eor	w16,w16,w22,ror#25	// Sigma1(e)
+	eor	w13,w13,w26,ror#13
+	add	w25,w25,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w11,w11,w4,ror#19
+	eor	w12,w12,w7,lsr#3	// sigma0(X[i+1])
+	add	w25,w25,w16			// h+=Sigma1(e)
+	eor	w28,w28,w27			// Maj(a,b,c)
+	eor	w17,w13,w26,ror#22	// Sigma0(a)
+	eor	w11,w11,w4,lsr#10	// sigma1(X[i+14])
+	add	w6,w6,w15
+	add	w21,w21,w25			// d+=h
+	add	w25,w25,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w6,w6,w12
+	add	w25,w25,w17			// h+=Sigma0(a)
+	add	w6,w6,w11
+	ldr	w11,[sp,#0]
+	str	w14,[sp,#12]
+	ror	w16,w21,#6
+	add	w24,w24,w28			// h+=K[i]
+	ror	w13,w8,#7
+	and	w17,w22,w21
+	ror	w12,w5,#17
+	bic	w28,w23,w21
+	ror	w14,w25,#2
+	add	w24,w24,w6			// h+=X[i]
+	eor	w16,w16,w21,ror#11
+	eor	w13,w13,w8,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w25,w26			// a^b, b^c in next round
+	eor	w16,w16,w21,ror#25	// Sigma1(e)
+	eor	w14,w14,w25,ror#13
+	add	w24,w24,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w12,w12,w5,ror#19
+	eor	w13,w13,w8,lsr#3	// sigma0(X[i+1])
+	add	w24,w24,w16			// h+=Sigma1(e)
+	eor	w19,w19,w26			// Maj(a,b,c)
+	eor	w17,w14,w25,ror#22	// Sigma0(a)
+	eor	w12,w12,w5,lsr#10	// sigma1(X[i+14])
+	add	w7,w7,w0
+	add	w20,w20,w24			// d+=h
+	add	w24,w24,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w7,w7,w13
+	add	w24,w24,w17			// h+=Sigma0(a)
+	add	w7,w7,w12
+	ldr	w12,[sp,#4]
+	str	w15,[sp,#0]
+	ror	w16,w20,#6
+	add	w23,w23,w19			// h+=K[i]
+	ror	w14,w9,#7
+	and	w17,w21,w20
+	ror	w13,w6,#17
+	bic	w19,w22,w20
+	ror	w15,w24,#2
+	add	w23,w23,w7			// h+=X[i]
+	eor	w16,w16,w20,ror#11
+	eor	w14,w14,w9,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w24,w25			// a^b, b^c in next round
+	eor	w16,w16,w20,ror#25	// Sigma1(e)
+	eor	w15,w15,w24,ror#13
+	add	w23,w23,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w13,w13,w6,ror#19
+	eor	w14,w14,w9,lsr#3	// sigma0(X[i+1])
+	add	w23,w23,w16			// h+=Sigma1(e)
+	eor	w28,w28,w25			// Maj(a,b,c)
+	eor	w17,w15,w24,ror#22	// Sigma0(a)
+	eor	w13,w13,w6,lsr#10	// sigma1(X[i+14])
+	add	w8,w8,w1
+	add	w27,w27,w23			// d+=h
+	add	w23,w23,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w8,w8,w14
+	add	w23,w23,w17			// h+=Sigma0(a)
+	add	w8,w8,w13
+	ldr	w13,[sp,#8]
+	str	w0,[sp,#4]
+	ror	w16,w27,#6
+	add	w22,w22,w28			// h+=K[i]
+	ror	w15,w10,#7
+	and	w17,w20,w27
+	ror	w14,w7,#17
+	bic	w28,w21,w27
+	ror	w0,w23,#2
+	add	w22,w22,w8			// h+=X[i]
+	eor	w16,w16,w27,ror#11
+	eor	w15,w15,w10,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w23,w24			// a^b, b^c in next round
+	eor	w16,w16,w27,ror#25	// Sigma1(e)
+	eor	w0,w0,w23,ror#13
+	add	w22,w22,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w14,w14,w7,ror#19
+	eor	w15,w15,w10,lsr#3	// sigma0(X[i+1])
+	add	w22,w22,w16			// h+=Sigma1(e)
+	eor	w19,w19,w24			// Maj(a,b,c)
+	eor	w17,w0,w23,ror#22	// Sigma0(a)
+	eor	w14,w14,w7,lsr#10	// sigma1(X[i+14])
+	add	w9,w9,w2
+	add	w26,w26,w22			// d+=h
+	add	w22,w22,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w9,w9,w15
+	add	w22,w22,w17			// h+=Sigma0(a)
+	add	w9,w9,w14
+	ldr	w14,[sp,#12]
+	str	w1,[sp,#8]
+	ror	w16,w26,#6
+	add	w21,w21,w19			// h+=K[i]
+	ror	w0,w11,#7
+	and	w17,w27,w26
+	ror	w15,w8,#17
+	bic	w19,w20,w26
+	ror	w1,w22,#2
+	add	w21,w21,w9			// h+=X[i]
+	eor	w16,w16,w26,ror#11
+	eor	w0,w0,w11,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w22,w23			// a^b, b^c in next round
+	eor	w16,w16,w26,ror#25	// Sigma1(e)
+	eor	w1,w1,w22,ror#13
+	add	w21,w21,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w15,w15,w8,ror#19
+	eor	w0,w0,w11,lsr#3	// sigma0(X[i+1])
+	add	w21,w21,w16			// h+=Sigma1(e)
+	eor	w28,w28,w23			// Maj(a,b,c)
+	eor	w17,w1,w22,ror#22	// Sigma0(a)
+	eor	w15,w15,w8,lsr#10	// sigma1(X[i+14])
+	add	w10,w10,w3
+	add	w25,w25,w21			// d+=h
+	add	w21,w21,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w10,w10,w0
+	add	w21,w21,w17			// h+=Sigma0(a)
+	add	w10,w10,w15
+	ldr	w15,[sp,#0]
+	str	w2,[sp,#12]
+	ror	w16,w25,#6
+	add	w20,w20,w28			// h+=K[i]
+	ror	w1,w12,#7
+	and	w17,w26,w25
+	ror	w0,w9,#17
+	bic	w28,w27,w25
+	ror	w2,w21,#2
+	add	w20,w20,w10			// h+=X[i]
+	eor	w16,w16,w25,ror#11
+	eor	w1,w1,w12,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w21,w22			// a^b, b^c in next round
+	eor	w16,w16,w25,ror#25	// Sigma1(e)
+	eor	w2,w2,w21,ror#13
+	add	w20,w20,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w0,w0,w9,ror#19
+	eor	w1,w1,w12,lsr#3	// sigma0(X[i+1])
+	add	w20,w20,w16			// h+=Sigma1(e)
+	eor	w19,w19,w22			// Maj(a,b,c)
+	eor	w17,w2,w21,ror#22	// Sigma0(a)
+	eor	w0,w0,w9,lsr#10	// sigma1(X[i+14])
+	add	w11,w11,w4
+	add	w24,w24,w20			// d+=h
+	add	w20,w20,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w11,w11,w1
+	add	w20,w20,w17			// h+=Sigma0(a)
+	add	w11,w11,w0
+	ldr	w0,[sp,#4]
+	str	w3,[sp,#0]
+	ror	w16,w24,#6
+	add	w27,w27,w19			// h+=K[i]
+	ror	w2,w13,#7
+	and	w17,w25,w24
+	ror	w1,w10,#17
+	bic	w19,w26,w24
+	ror	w3,w20,#2
+	add	w27,w27,w11			// h+=X[i]
+	eor	w16,w16,w24,ror#11
+	eor	w2,w2,w13,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w20,w21			// a^b, b^c in next round
+	eor	w16,w16,w24,ror#25	// Sigma1(e)
+	eor	w3,w3,w20,ror#13
+	add	w27,w27,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w1,w1,w10,ror#19
+	eor	w2,w2,w13,lsr#3	// sigma0(X[i+1])
+	add	w27,w27,w16			// h+=Sigma1(e)
+	eor	w28,w28,w21			// Maj(a,b,c)
+	eor	w17,w3,w20,ror#22	// Sigma0(a)
+	eor	w1,w1,w10,lsr#10	// sigma1(X[i+14])
+	add	w12,w12,w5
+	add	w23,w23,w27			// d+=h
+	add	w27,w27,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w12,w12,w2
+	add	w27,w27,w17			// h+=Sigma0(a)
+	add	w12,w12,w1
+	ldr	w1,[sp,#8]
+	str	w4,[sp,#4]
+	ror	w16,w23,#6
+	add	w26,w26,w28			// h+=K[i]
+	ror	w3,w14,#7
+	and	w17,w24,w23
+	ror	w2,w11,#17
+	bic	w28,w25,w23
+	ror	w4,w27,#2
+	add	w26,w26,w12			// h+=X[i]
+	eor	w16,w16,w23,ror#11
+	eor	w3,w3,w14,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w27,w20			// a^b, b^c in next round
+	eor	w16,w16,w23,ror#25	// Sigma1(e)
+	eor	w4,w4,w27,ror#13
+	add	w26,w26,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w2,w2,w11,ror#19
+	eor	w3,w3,w14,lsr#3	// sigma0(X[i+1])
+	add	w26,w26,w16			// h+=Sigma1(e)
+	eor	w19,w19,w20			// Maj(a,b,c)
+	eor	w17,w4,w27,ror#22	// Sigma0(a)
+	eor	w2,w2,w11,lsr#10	// sigma1(X[i+14])
+	add	w13,w13,w6
+	add	w22,w22,w26			// d+=h
+	add	w26,w26,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w13,w13,w3
+	add	w26,w26,w17			// h+=Sigma0(a)
+	add	w13,w13,w2
+	ldr	w2,[sp,#12]
+	str	w5,[sp,#8]
+	ror	w16,w22,#6
+	add	w25,w25,w19			// h+=K[i]
+	ror	w4,w15,#7
+	and	w17,w23,w22
+	ror	w3,w12,#17
+	bic	w19,w24,w22
+	ror	w5,w26,#2
+	add	w25,w25,w13			// h+=X[i]
+	eor	w16,w16,w22,ror#11
+	eor	w4,w4,w15,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w26,w27			// a^b, b^c in next round
+	eor	w16,w16,w22,ror#25	// Sigma1(e)
+	eor	w5,w5,w26,ror#13
+	add	w25,w25,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w3,w3,w12,ror#19
+	eor	w4,w4,w15,lsr#3	// sigma0(X[i+1])
+	add	w25,w25,w16			// h+=Sigma1(e)
+	eor	w28,w28,w27			// Maj(a,b,c)
+	eor	w17,w5,w26,ror#22	// Sigma0(a)
+	eor	w3,w3,w12,lsr#10	// sigma1(X[i+14])
+	add	w14,w14,w7
+	add	w21,w21,w25			// d+=h
+	add	w25,w25,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w14,w14,w4
+	add	w25,w25,w17			// h+=Sigma0(a)
+	add	w14,w14,w3
+	ldr	w3,[sp,#0]
+	str	w6,[sp,#12]
+	ror	w16,w21,#6
+	add	w24,w24,w28			// h+=K[i]
+	ror	w5,w0,#7
+	and	w17,w22,w21
+	ror	w4,w13,#17
+	bic	w28,w23,w21
+	ror	w6,w25,#2
+	add	w24,w24,w14			// h+=X[i]
+	eor	w16,w16,w21,ror#11
+	eor	w5,w5,w0,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w25,w26			// a^b, b^c in next round
+	eor	w16,w16,w21,ror#25	// Sigma1(e)
+	eor	w6,w6,w25,ror#13
+	add	w24,w24,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w4,w4,w13,ror#19
+	eor	w5,w5,w0,lsr#3	// sigma0(X[i+1])
+	add	w24,w24,w16			// h+=Sigma1(e)
+	eor	w19,w19,w26			// Maj(a,b,c)
+	eor	w17,w6,w25,ror#22	// Sigma0(a)
+	eor	w4,w4,w13,lsr#10	// sigma1(X[i+14])
+	add	w15,w15,w8
+	add	w20,w20,w24			// d+=h
+	add	w24,w24,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w15,w15,w5
+	add	w24,w24,w17			// h+=Sigma0(a)
+	add	w15,w15,w4
+	ldr	w4,[sp,#4]
+	str	w7,[sp,#0]
+	ror	w16,w20,#6
+	add	w23,w23,w19			// h+=K[i]
+	ror	w6,w1,#7
+	and	w17,w21,w20
+	ror	w5,w14,#17
+	bic	w19,w22,w20
+	ror	w7,w24,#2
+	add	w23,w23,w15			// h+=X[i]
+	eor	w16,w16,w20,ror#11
+	eor	w6,w6,w1,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w24,w25			// a^b, b^c in next round
+	eor	w16,w16,w20,ror#25	// Sigma1(e)
+	eor	w7,w7,w24,ror#13
+	add	w23,w23,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w5,w5,w14,ror#19
+	eor	w6,w6,w1,lsr#3	// sigma0(X[i+1])
+	add	w23,w23,w16			// h+=Sigma1(e)
+	eor	w28,w28,w25			// Maj(a,b,c)
+	eor	w17,w7,w24,ror#22	// Sigma0(a)
+	eor	w5,w5,w14,lsr#10	// sigma1(X[i+14])
+	add	w0,w0,w9
+	add	w27,w27,w23			// d+=h
+	add	w23,w23,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w0,w0,w6
+	add	w23,w23,w17			// h+=Sigma0(a)
+	add	w0,w0,w5
+	ldr	w5,[sp,#8]
+	str	w8,[sp,#4]
+	ror	w16,w27,#6
+	add	w22,w22,w28			// h+=K[i]
+	ror	w7,w2,#7
+	and	w17,w20,w27
+	ror	w6,w15,#17
+	bic	w28,w21,w27
+	ror	w8,w23,#2
+	add	w22,w22,w0			// h+=X[i]
+	eor	w16,w16,w27,ror#11
+	eor	w7,w7,w2,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w23,w24			// a^b, b^c in next round
+	eor	w16,w16,w27,ror#25	// Sigma1(e)
+	eor	w8,w8,w23,ror#13
+	add	w22,w22,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w6,w6,w15,ror#19
+	eor	w7,w7,w2,lsr#3	// sigma0(X[i+1])
+	add	w22,w22,w16			// h+=Sigma1(e)
+	eor	w19,w19,w24			// Maj(a,b,c)
+	eor	w17,w8,w23,ror#22	// Sigma0(a)
+	eor	w6,w6,w15,lsr#10	// sigma1(X[i+14])
+	add	w1,w1,w10
+	add	w26,w26,w22			// d+=h
+	add	w22,w22,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w1,w1,w7
+	add	w22,w22,w17			// h+=Sigma0(a)
+	add	w1,w1,w6
+	ldr	w6,[sp,#12]
+	str	w9,[sp,#8]
+	ror	w16,w26,#6
+	add	w21,w21,w19			// h+=K[i]
+	ror	w8,w3,#7
+	and	w17,w27,w26
+	ror	w7,w0,#17
+	bic	w19,w20,w26
+	ror	w9,w22,#2
+	add	w21,w21,w1			// h+=X[i]
+	eor	w16,w16,w26,ror#11
+	eor	w8,w8,w3,ror#18
+	orr	w17,w17,w19			// Ch(e,f,g)
+	eor	w19,w22,w23			// a^b, b^c in next round
+	eor	w16,w16,w26,ror#25	// Sigma1(e)
+	eor	w9,w9,w22,ror#13
+	add	w21,w21,w17			// h+=Ch(e,f,g)
+	and	w28,w28,w19			// (b^c)&=(a^b)
+	eor	w7,w7,w0,ror#19
+	eor	w8,w8,w3,lsr#3	// sigma0(X[i+1])
+	add	w21,w21,w16			// h+=Sigma1(e)
+	eor	w28,w28,w23			// Maj(a,b,c)
+	eor	w17,w9,w22,ror#22	// Sigma0(a)
+	eor	w7,w7,w0,lsr#10	// sigma1(X[i+14])
+	add	w2,w2,w11
+	add	w25,w25,w21			// d+=h
+	add	w21,w21,w28			// h+=Maj(a,b,c)
+	ldr	w28,[x30],#4		// *K++, w19 in next round
+	add	w2,w2,w8
+	add	w21,w21,w17			// h+=Sigma0(a)
+	add	w2,w2,w7
+	ldr	w7,[sp,#0]
+	str	w10,[sp,#12]
+	ror	w16,w25,#6
+	add	w20,w20,w28			// h+=K[i]
+	ror	w9,w4,#7
+	and	w17,w26,w25
+	ror	w8,w1,#17
+	bic	w28,w27,w25
+	ror	w10,w21,#2
+	add	w20,w20,w2			// h+=X[i]
+	eor	w16,w16,w25,ror#11
+	eor	w9,w9,w4,ror#18
+	orr	w17,w17,w28			// Ch(e,f,g)
+	eor	w28,w21,w22			// a^b, b^c in next round
+	eor	w16,w16,w25,ror#25	// Sigma1(e)
+	eor	w10,w10,w21,ror#13
+	add	w20,w20,w17			// h+=Ch(e,f,g)
+	and	w19,w19,w28			// (b^c)&=(a^b)
+	eor	w8,w8,w1,ror#19
+	eor	w9,w9,w4,lsr#3	// sigma0(X[i+1])
+	add	w20,w20,w16			// h+=Sigma1(e)
+	eor	w19,w19,w22			// Maj(a,b,c)
+	eor	w17,w10,w21,ror#22	// Sigma0(a)
+	eor	w8,w8,w1,lsr#10	// sigma1(X[i+14])
+	add	w3,w3,w12
+	add	w24,w24,w20			// d+=h
+	add	w20,w20,w19			// h+=Maj(a,b,c)
+	ldr	w19,[x30],#4		// *K++, w28 in next round
+	add	w3,w3,w9
+	add	w20,w20,w17			// h+=Sigma0(a)
+	add	w3,w3,w8
+	cbnz	w19,.Loop_16_xx
+
+	ldp	x0,x2,[x29,#96]
+	ldr	x1,[x29,#112]
+	sub	x30,x30,#260		// rewind
+
+	ldp	w3,w4,[x0]
+	ldp	w5,w6,[x0,#2*4]
+	add	x1,x1,#14*4			// advance input pointer
+	ldp	w7,w8,[x0,#4*4]
+	add	w20,w20,w3
+	ldp	w9,w10,[x0,#6*4]
+	add	w21,w21,w4
+	add	w22,w22,w5
+	add	w23,w23,w6
+	stp	w20,w21,[x0]
+	add	w24,w24,w7
+	add	w25,w25,w8
+	stp	w22,w23,[x0,#2*4]
+	add	w26,w26,w9
+	add	w27,w27,w10
+	cmp	x1,x2
+	stp	w24,w25,[x0,#4*4]
+	stp	w26,w27,[x0,#6*4]
+	b.ne	.Loop
+
+	ldp	x19,x20,[x29,#16]
+	add	sp,sp,#4*4
+	ldp	x21,x22,[x29,#32]
+	ldp	x23,x24,[x29,#48]
+	ldp	x25,x26,[x29,#64]
+	ldp	x27,x28,[x29,#80]
+	ldp	x29,x30,[sp],#128
+	ret
+.size	sha256_block_data_order,.-sha256_block_data_order
+
+.align	6
+.type	.LK256,%object
+.LK256:
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+	.long	0	//terminator
+.size	.LK256,.-.LK256
+#ifndef	__KERNEL__
+.align	3
+.LOPENSSL_armcap_P:
+# ifdef	__ILP32__
+	.long	OPENSSL_armcap_P-.
+# else
+	.quad	OPENSSL_armcap_P-.
+# endif
+#endif
+.asciz	"SHA256 block transform for ARMv8, CRYPTOGAMS by <appro@openssl.org>"
+.align	2
+#ifndef	__KERNEL__
+.type	sha256_block_armv8,%function
+.align	6
+sha256_block_armv8:
+.Lv8_entry:
+	stp		x29,x30,[sp,#-16]!
+	add		x29,sp,#0
+
+	ld1		{v0.4s,v1.4s},[x0]
+	adr		x3,.LK256
+
+.Loop_hw:
+	ld1		{v4.16b-v7.16b},[x1],#64
+	sub		x2,x2,#1
+	ld1		{v16.4s},[x3],#16
+	rev32		v4.16b,v4.16b
+	rev32		v5.16b,v5.16b
+	rev32		v6.16b,v6.16b
+	rev32		v7.16b,v7.16b
+	orr		v18.16b,v0.16b,v0.16b		// offload
+	orr		v19.16b,v1.16b,v1.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v4.4s
+	.inst	0x5e2828a4	//sha256su0 v4.16b,v5.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e0760c4	//sha256su1 v4.16b,v6.16b,v7.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v5.4s
+	.inst	0x5e2828c5	//sha256su0 v5.16b,v6.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0460e5	//sha256su1 v5.16b,v7.16b,v4.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v6.4s
+	.inst	0x5e2828e6	//sha256su0 v6.16b,v7.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e056086	//sha256su1 v6.16b,v4.16b,v5.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v7.4s
+	.inst	0x5e282887	//sha256su0 v7.16b,v4.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0660a7	//sha256su1 v7.16b,v5.16b,v6.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v4.4s
+	.inst	0x5e2828a4	//sha256su0 v4.16b,v5.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e0760c4	//sha256su1 v4.16b,v6.16b,v7.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v5.4s
+	.inst	0x5e2828c5	//sha256su0 v5.16b,v6.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0460e5	//sha256su1 v5.16b,v7.16b,v4.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v6.4s
+	.inst	0x5e2828e6	//sha256su0 v6.16b,v7.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e056086	//sha256su1 v6.16b,v4.16b,v5.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v7.4s
+	.inst	0x5e282887	//sha256su0 v7.16b,v4.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0660a7	//sha256su1 v7.16b,v5.16b,v6.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v4.4s
+	.inst	0x5e2828a4	//sha256su0 v4.16b,v5.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e0760c4	//sha256su1 v4.16b,v6.16b,v7.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v5.4s
+	.inst	0x5e2828c5	//sha256su0 v5.16b,v6.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0460e5	//sha256su1 v5.16b,v7.16b,v4.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v6.4s
+	.inst	0x5e2828e6	//sha256su0 v6.16b,v7.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+	.inst	0x5e056086	//sha256su1 v6.16b,v4.16b,v5.16b
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v7.4s
+	.inst	0x5e282887	//sha256su0 v7.16b,v4.16b
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+	.inst	0x5e0660a7	//sha256su1 v7.16b,v5.16b,v6.16b
+	ld1		{v17.4s},[x3],#16
+	add		v16.4s,v16.4s,v4.4s
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+
+	ld1		{v16.4s},[x3],#16
+	add		v17.4s,v17.4s,v5.4s
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+
+	ld1		{v17.4s},[x3]
+	add		v16.4s,v16.4s,v6.4s
+	sub		x3,x3,#64*4-16	// rewind
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e104020	//sha256h v0.16b,v1.16b,v16.4s
+	.inst	0x5e105041	//sha256h2 v1.16b,v2.16b,v16.4s
+
+	add		v17.4s,v17.4s,v7.4s
+	orr		v2.16b,v0.16b,v0.16b
+	.inst	0x5e114020	//sha256h v0.16b,v1.16b,v17.4s
+	.inst	0x5e115041	//sha256h2 v1.16b,v2.16b,v17.4s
+
+	add		v0.4s,v0.4s,v18.4s
+	add		v1.4s,v1.4s,v19.4s
+
+	cbnz		x2,.Loop_hw
+
+	st1		{v0.4s,v1.4s},[x0]
+
+	ldr		x29,[sp],#16
+	ret
+.size	sha256_block_armv8,.-sha256_block_armv8
+#endif
+#ifdef	__KERNEL__
+.globl	sha256_block_neon
+#endif
+.type	sha256_block_neon,%function
+.align	4
+sha256_block_neon:
+.Lneon_entry:
+	stp	x29, x30, [sp, #-16]!
+	mov	x29, sp
+	sub	sp,sp,#16*4
+
+	adr	x16,.LK256
+	add	x2,x1,x2,lsl#6	// len to point at the end of inp
+
+	ld1	{v0.16b},[x1], #16
+	ld1	{v1.16b},[x1], #16
+	ld1	{v2.16b},[x1], #16
+	ld1	{v3.16b},[x1], #16
+	ld1	{v4.4s},[x16], #16
+	ld1	{v5.4s},[x16], #16
+	ld1	{v6.4s},[x16], #16
+	ld1	{v7.4s},[x16], #16
+	rev32	v0.16b,v0.16b		// yes, even on
+	rev32	v1.16b,v1.16b		// big-endian
+	rev32	v2.16b,v2.16b
+	rev32	v3.16b,v3.16b
+	mov	x17,sp
+	add	v4.4s,v4.4s,v0.4s
+	add	v5.4s,v5.4s,v1.4s
+	add	v6.4s,v6.4s,v2.4s
+	st1	{v4.4s-v5.4s},[x17], #32
+	add	v7.4s,v7.4s,v3.4s
+	st1	{v6.4s-v7.4s},[x17]
+	sub	x17,x17,#32
+
+	ldp	w3,w4,[x0]
+	ldp	w5,w6,[x0,#8]
+	ldp	w7,w8,[x0,#16]
+	ldp	w9,w10,[x0,#24]
+	ldr	w12,[sp,#0]
+	mov	w13,wzr
+	eor	w14,w4,w5
+	mov	w15,wzr
+	b	.L_00_48
+
+.align	4
+.L_00_48:
+	ext	v4.16b,v0.16b,v1.16b,#4
+	add	w10,w10,w12
+	add	w3,w3,w15
+	and	w12,w8,w7
+	bic	w15,w9,w7
+	ext	v7.16b,v2.16b,v3.16b,#4
+	eor	w11,w7,w7,ror#5
+	add	w3,w3,w13
+	mov	d19,v3.d[1]
+	orr	w12,w12,w15
+	eor	w11,w11,w7,ror#19
+	ushr	v6.4s,v4.4s,#7
+	eor	w15,w3,w3,ror#11
+	ushr	v5.4s,v4.4s,#3
+	add	w10,w10,w12
+	add	v0.4s,v0.4s,v7.4s
+	ror	w11,w11,#6
+	sli	v6.4s,v4.4s,#25
+	eor	w13,w3,w4
+	eor	w15,w15,w3,ror#20
+	ushr	v7.4s,v4.4s,#18
+	add	w10,w10,w11
+	ldr	w12,[sp,#4]
+	and	w14,w14,w13
+	eor	v5.16b,v5.16b,v6.16b
+	ror	w15,w15,#2
+	add	w6,w6,w10
+	sli	v7.4s,v4.4s,#14
+	eor	w14,w14,w4
+	ushr	v16.4s,v19.4s,#17
+	add	w9,w9,w12
+	add	w10,w10,w15
+	and	w12,w7,w6
+	eor	v5.16b,v5.16b,v7.16b
+	bic	w15,w8,w6
+	eor	w11,w6,w6,ror#5
+	sli	v16.4s,v19.4s,#15
+	add	w10,w10,w14
+	orr	w12,w12,w15
+	ushr	v17.4s,v19.4s,#10
+	eor	w11,w11,w6,ror#19
+	eor	w15,w10,w10,ror#11
+	ushr	v7.4s,v19.4s,#19
+	add	w9,w9,w12
+	ror	w11,w11,#6
+	add	v0.4s,v0.4s,v5.4s
+	eor	w14,w10,w3
+	eor	w15,w15,w10,ror#20
+	sli	v7.4s,v19.4s,#13
+	add	w9,w9,w11
+	ldr	w12,[sp,#8]
+	and	w13,w13,w14
+	eor	v17.16b,v17.16b,v16.16b
+	ror	w15,w15,#2
+	add	w5,w5,w9
+	eor	w13,w13,w3
+	eor	v17.16b,v17.16b,v7.16b
+	add	w8,w8,w12
+	add	w9,w9,w15
+	and	w12,w6,w5
+	add	v0.4s,v0.4s,v17.4s
+	bic	w15,w7,w5
+	eor	w11,w5,w5,ror#5
+	add	w9,w9,w13
+	ushr	v18.4s,v0.4s,#17
+	orr	w12,w12,w15
+	ushr	v19.4s,v0.4s,#10
+	eor	w11,w11,w5,ror#19
+	eor	w15,w9,w9,ror#11
+	sli	v18.4s,v0.4s,#15
+	add	w8,w8,w12
+	ushr	v17.4s,v0.4s,#19
+	ror	w11,w11,#6
+	eor	w13,w9,w10
+	eor	v19.16b,v19.16b,v18.16b
+	eor	w15,w15,w9,ror#20
+	add	w8,w8,w11
+	sli	v17.4s,v0.4s,#13
+	ldr	w12,[sp,#12]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	ld1	{v4.4s},[x16], #16
+	add	w4,w4,w8
+	eor	v19.16b,v19.16b,v17.16b
+	eor	w14,w14,w10
+	eor	v17.16b,v17.16b,v17.16b
+	add	w7,w7,w12
+	add	w8,w8,w15
+	and	w12,w5,w4
+	mov	v17.d[1],v19.d[0]
+	bic	w15,w6,w4
+	eor	w11,w4,w4,ror#5
+	add	w8,w8,w14
+	add	v0.4s,v0.4s,v17.4s
+	orr	w12,w12,w15
+	eor	w11,w11,w4,ror#19
+	eor	w15,w8,w8,ror#11
+	add	v4.4s,v4.4s,v0.4s
+	add	w7,w7,w12
+	ror	w11,w11,#6
+	eor	w14,w8,w9
+	eor	w15,w15,w8,ror#20
+	add	w7,w7,w11
+	ldr	w12,[sp,#16]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w3,w3,w7
+	eor	w13,w13,w9
+	st1	{v4.4s},[x17], #16
+	ext	v4.16b,v1.16b,v2.16b,#4
+	add	w6,w6,w12
+	add	w7,w7,w15
+	and	w12,w4,w3
+	bic	w15,w5,w3
+	ext	v7.16b,v3.16b,v0.16b,#4
+	eor	w11,w3,w3,ror#5
+	add	w7,w7,w13
+	mov	d19,v0.d[1]
+	orr	w12,w12,w15
+	eor	w11,w11,w3,ror#19
+	ushr	v6.4s,v4.4s,#7
+	eor	w15,w7,w7,ror#11
+	ushr	v5.4s,v4.4s,#3
+	add	w6,w6,w12
+	add	v1.4s,v1.4s,v7.4s
+	ror	w11,w11,#6
+	sli	v6.4s,v4.4s,#25
+	eor	w13,w7,w8
+	eor	w15,w15,w7,ror#20
+	ushr	v7.4s,v4.4s,#18
+	add	w6,w6,w11
+	ldr	w12,[sp,#20]
+	and	w14,w14,w13
+	eor	v5.16b,v5.16b,v6.16b
+	ror	w15,w15,#2
+	add	w10,w10,w6
+	sli	v7.4s,v4.4s,#14
+	eor	w14,w14,w8
+	ushr	v16.4s,v19.4s,#17
+	add	w5,w5,w12
+	add	w6,w6,w15
+	and	w12,w3,w10
+	eor	v5.16b,v5.16b,v7.16b
+	bic	w15,w4,w10
+	eor	w11,w10,w10,ror#5
+	sli	v16.4s,v19.4s,#15
+	add	w6,w6,w14
+	orr	w12,w12,w15
+	ushr	v17.4s,v19.4s,#10
+	eor	w11,w11,w10,ror#19
+	eor	w15,w6,w6,ror#11
+	ushr	v7.4s,v19.4s,#19
+	add	w5,w5,w12
+	ror	w11,w11,#6
+	add	v1.4s,v1.4s,v5.4s
+	eor	w14,w6,w7
+	eor	w15,w15,w6,ror#20
+	sli	v7.4s,v19.4s,#13
+	add	w5,w5,w11
+	ldr	w12,[sp,#24]
+	and	w13,w13,w14
+	eor	v17.16b,v17.16b,v16.16b
+	ror	w15,w15,#2
+	add	w9,w9,w5
+	eor	w13,w13,w7
+	eor	v17.16b,v17.16b,v7.16b
+	add	w4,w4,w12
+	add	w5,w5,w15
+	and	w12,w10,w9
+	add	v1.4s,v1.4s,v17.4s
+	bic	w15,w3,w9
+	eor	w11,w9,w9,ror#5
+	add	w5,w5,w13
+	ushr	v18.4s,v1.4s,#17
+	orr	w12,w12,w15
+	ushr	v19.4s,v1.4s,#10
+	eor	w11,w11,w9,ror#19
+	eor	w15,w5,w5,ror#11
+	sli	v18.4s,v1.4s,#15
+	add	w4,w4,w12
+	ushr	v17.4s,v1.4s,#19
+	ror	w11,w11,#6
+	eor	w13,w5,w6
+	eor	v19.16b,v19.16b,v18.16b
+	eor	w15,w15,w5,ror#20
+	add	w4,w4,w11
+	sli	v17.4s,v1.4s,#13
+	ldr	w12,[sp,#28]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	ld1	{v4.4s},[x16], #16
+	add	w8,w8,w4
+	eor	v19.16b,v19.16b,v17.16b
+	eor	w14,w14,w6
+	eor	v17.16b,v17.16b,v17.16b
+	add	w3,w3,w12
+	add	w4,w4,w15
+	and	w12,w9,w8
+	mov	v17.d[1],v19.d[0]
+	bic	w15,w10,w8
+	eor	w11,w8,w8,ror#5
+	add	w4,w4,w14
+	add	v1.4s,v1.4s,v17.4s
+	orr	w12,w12,w15
+	eor	w11,w11,w8,ror#19
+	eor	w15,w4,w4,ror#11
+	add	v4.4s,v4.4s,v1.4s
+	add	w3,w3,w12
+	ror	w11,w11,#6
+	eor	w14,w4,w5
+	eor	w15,w15,w4,ror#20
+	add	w3,w3,w11
+	ldr	w12,[sp,#32]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w7,w7,w3
+	eor	w13,w13,w5
+	st1	{v4.4s},[x17], #16
+	ext	v4.16b,v2.16b,v3.16b,#4
+	add	w10,w10,w12
+	add	w3,w3,w15
+	and	w12,w8,w7
+	bic	w15,w9,w7
+	ext	v7.16b,v0.16b,v1.16b,#4
+	eor	w11,w7,w7,ror#5
+	add	w3,w3,w13
+	mov	d19,v1.d[1]
+	orr	w12,w12,w15
+	eor	w11,w11,w7,ror#19
+	ushr	v6.4s,v4.4s,#7
+	eor	w15,w3,w3,ror#11
+	ushr	v5.4s,v4.4s,#3
+	add	w10,w10,w12
+	add	v2.4s,v2.4s,v7.4s
+	ror	w11,w11,#6
+	sli	v6.4s,v4.4s,#25
+	eor	w13,w3,w4
+	eor	w15,w15,w3,ror#20
+	ushr	v7.4s,v4.4s,#18
+	add	w10,w10,w11
+	ldr	w12,[sp,#36]
+	and	w14,w14,w13
+	eor	v5.16b,v5.16b,v6.16b
+	ror	w15,w15,#2
+	add	w6,w6,w10
+	sli	v7.4s,v4.4s,#14
+	eor	w14,w14,w4
+	ushr	v16.4s,v19.4s,#17
+	add	w9,w9,w12
+	add	w10,w10,w15
+	and	w12,w7,w6
+	eor	v5.16b,v5.16b,v7.16b
+	bic	w15,w8,w6
+	eor	w11,w6,w6,ror#5
+	sli	v16.4s,v19.4s,#15
+	add	w10,w10,w14
+	orr	w12,w12,w15
+	ushr	v17.4s,v19.4s,#10
+	eor	w11,w11,w6,ror#19
+	eor	w15,w10,w10,ror#11
+	ushr	v7.4s,v19.4s,#19
+	add	w9,w9,w12
+	ror	w11,w11,#6
+	add	v2.4s,v2.4s,v5.4s
+	eor	w14,w10,w3
+	eor	w15,w15,w10,ror#20
+	sli	v7.4s,v19.4s,#13
+	add	w9,w9,w11
+	ldr	w12,[sp,#40]
+	and	w13,w13,w14
+	eor	v17.16b,v17.16b,v16.16b
+	ror	w15,w15,#2
+	add	w5,w5,w9
+	eor	w13,w13,w3
+	eor	v17.16b,v17.16b,v7.16b
+	add	w8,w8,w12
+	add	w9,w9,w15
+	and	w12,w6,w5
+	add	v2.4s,v2.4s,v17.4s
+	bic	w15,w7,w5
+	eor	w11,w5,w5,ror#5
+	add	w9,w9,w13
+	ushr	v18.4s,v2.4s,#17
+	orr	w12,w12,w15
+	ushr	v19.4s,v2.4s,#10
+	eor	w11,w11,w5,ror#19
+	eor	w15,w9,w9,ror#11
+	sli	v18.4s,v2.4s,#15
+	add	w8,w8,w12
+	ushr	v17.4s,v2.4s,#19
+	ror	w11,w11,#6
+	eor	w13,w9,w10
+	eor	v19.16b,v19.16b,v18.16b
+	eor	w15,w15,w9,ror#20
+	add	w8,w8,w11
+	sli	v17.4s,v2.4s,#13
+	ldr	w12,[sp,#44]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	ld1	{v4.4s},[x16], #16
+	add	w4,w4,w8
+	eor	v19.16b,v19.16b,v17.16b
+	eor	w14,w14,w10
+	eor	v17.16b,v17.16b,v17.16b
+	add	w7,w7,w12
+	add	w8,w8,w15
+	and	w12,w5,w4
+	mov	v17.d[1],v19.d[0]
+	bic	w15,w6,w4
+	eor	w11,w4,w4,ror#5
+	add	w8,w8,w14
+	add	v2.4s,v2.4s,v17.4s
+	orr	w12,w12,w15
+	eor	w11,w11,w4,ror#19
+	eor	w15,w8,w8,ror#11
+	add	v4.4s,v4.4s,v2.4s
+	add	w7,w7,w12
+	ror	w11,w11,#6
+	eor	w14,w8,w9
+	eor	w15,w15,w8,ror#20
+	add	w7,w7,w11
+	ldr	w12,[sp,#48]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w3,w3,w7
+	eor	w13,w13,w9
+	st1	{v4.4s},[x17], #16
+	ext	v4.16b,v3.16b,v0.16b,#4
+	add	w6,w6,w12
+	add	w7,w7,w15
+	and	w12,w4,w3
+	bic	w15,w5,w3
+	ext	v7.16b,v1.16b,v2.16b,#4
+	eor	w11,w3,w3,ror#5
+	add	w7,w7,w13
+	mov	d19,v2.d[1]
+	orr	w12,w12,w15
+	eor	w11,w11,w3,ror#19
+	ushr	v6.4s,v4.4s,#7
+	eor	w15,w7,w7,ror#11
+	ushr	v5.4s,v4.4s,#3
+	add	w6,w6,w12
+	add	v3.4s,v3.4s,v7.4s
+	ror	w11,w11,#6
+	sli	v6.4s,v4.4s,#25
+	eor	w13,w7,w8
+	eor	w15,w15,w7,ror#20
+	ushr	v7.4s,v4.4s,#18
+	add	w6,w6,w11
+	ldr	w12,[sp,#52]
+	and	w14,w14,w13
+	eor	v5.16b,v5.16b,v6.16b
+	ror	w15,w15,#2
+	add	w10,w10,w6
+	sli	v7.4s,v4.4s,#14
+	eor	w14,w14,w8
+	ushr	v16.4s,v19.4s,#17
+	add	w5,w5,w12
+	add	w6,w6,w15
+	and	w12,w3,w10
+	eor	v5.16b,v5.16b,v7.16b
+	bic	w15,w4,w10
+	eor	w11,w10,w10,ror#5
+	sli	v16.4s,v19.4s,#15
+	add	w6,w6,w14
+	orr	w12,w12,w15
+	ushr	v17.4s,v19.4s,#10
+	eor	w11,w11,w10,ror#19
+	eor	w15,w6,w6,ror#11
+	ushr	v7.4s,v19.4s,#19
+	add	w5,w5,w12
+	ror	w11,w11,#6
+	add	v3.4s,v3.4s,v5.4s
+	eor	w14,w6,w7
+	eor	w15,w15,w6,ror#20
+	sli	v7.4s,v19.4s,#13
+	add	w5,w5,w11
+	ldr	w12,[sp,#56]
+	and	w13,w13,w14
+	eor	v17.16b,v17.16b,v16.16b
+	ror	w15,w15,#2
+	add	w9,w9,w5
+	eor	w13,w13,w7
+	eor	v17.16b,v17.16b,v7.16b
+	add	w4,w4,w12
+	add	w5,w5,w15
+	and	w12,w10,w9
+	add	v3.4s,v3.4s,v17.4s
+	bic	w15,w3,w9
+	eor	w11,w9,w9,ror#5
+	add	w5,w5,w13
+	ushr	v18.4s,v3.4s,#17
+	orr	w12,w12,w15
+	ushr	v19.4s,v3.4s,#10
+	eor	w11,w11,w9,ror#19
+	eor	w15,w5,w5,ror#11
+	sli	v18.4s,v3.4s,#15
+	add	w4,w4,w12
+	ushr	v17.4s,v3.4s,#19
+	ror	w11,w11,#6
+	eor	w13,w5,w6
+	eor	v19.16b,v19.16b,v18.16b
+	eor	w15,w15,w5,ror#20
+	add	w4,w4,w11
+	sli	v17.4s,v3.4s,#13
+	ldr	w12,[sp,#60]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	ld1	{v4.4s},[x16], #16
+	add	w8,w8,w4
+	eor	v19.16b,v19.16b,v17.16b
+	eor	w14,w14,w6
+	eor	v17.16b,v17.16b,v17.16b
+	add	w3,w3,w12
+	add	w4,w4,w15
+	and	w12,w9,w8
+	mov	v17.d[1],v19.d[0]
+	bic	w15,w10,w8
+	eor	w11,w8,w8,ror#5
+	add	w4,w4,w14
+	add	v3.4s,v3.4s,v17.4s
+	orr	w12,w12,w15
+	eor	w11,w11,w8,ror#19
+	eor	w15,w4,w4,ror#11
+	add	v4.4s,v4.4s,v3.4s
+	add	w3,w3,w12
+	ror	w11,w11,#6
+	eor	w14,w4,w5
+	eor	w15,w15,w4,ror#20
+	add	w3,w3,w11
+	ldr	w12,[x16]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w7,w7,w3
+	eor	w13,w13,w5
+	st1	{v4.4s},[x17], #16
+	cmp	w12,#0				// check for K256 terminator
+	ldr	w12,[sp,#0]
+	sub	x17,x17,#64
+	bne	.L_00_48
+
+	sub	x16,x16,#256		// rewind x16
+	cmp	x1,x2
+	mov	x17, #64
+	csel	x17, x17, xzr, eq
+	sub	x1,x1,x17			// avoid SEGV
+	mov	x17,sp
+	add	w10,w10,w12
+	add	w3,w3,w15
+	and	w12,w8,w7
+	ld1	{v0.16b},[x1],#16
+	bic	w15,w9,w7
+	eor	w11,w7,w7,ror#5
+	ld1	{v4.4s},[x16],#16
+	add	w3,w3,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w7,ror#19
+	eor	w15,w3,w3,ror#11
+	rev32	v0.16b,v0.16b
+	add	w10,w10,w12
+	ror	w11,w11,#6
+	eor	w13,w3,w4
+	eor	w15,w15,w3,ror#20
+	add	v4.4s,v4.4s,v0.4s
+	add	w10,w10,w11
+	ldr	w12,[sp,#4]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w6,w6,w10
+	eor	w14,w14,w4
+	add	w9,w9,w12
+	add	w10,w10,w15
+	and	w12,w7,w6
+	bic	w15,w8,w6
+	eor	w11,w6,w6,ror#5
+	add	w10,w10,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w6,ror#19
+	eor	w15,w10,w10,ror#11
+	add	w9,w9,w12
+	ror	w11,w11,#6
+	eor	w14,w10,w3
+	eor	w15,w15,w10,ror#20
+	add	w9,w9,w11
+	ldr	w12,[sp,#8]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w5,w5,w9
+	eor	w13,w13,w3
+	add	w8,w8,w12
+	add	w9,w9,w15
+	and	w12,w6,w5
+	bic	w15,w7,w5
+	eor	w11,w5,w5,ror#5
+	add	w9,w9,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w5,ror#19
+	eor	w15,w9,w9,ror#11
+	add	w8,w8,w12
+	ror	w11,w11,#6
+	eor	w13,w9,w10
+	eor	w15,w15,w9,ror#20
+	add	w8,w8,w11
+	ldr	w12,[sp,#12]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w4,w4,w8
+	eor	w14,w14,w10
+	add	w7,w7,w12
+	add	w8,w8,w15
+	and	w12,w5,w4
+	bic	w15,w6,w4
+	eor	w11,w4,w4,ror#5
+	add	w8,w8,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w4,ror#19
+	eor	w15,w8,w8,ror#11
+	add	w7,w7,w12
+	ror	w11,w11,#6
+	eor	w14,w8,w9
+	eor	w15,w15,w8,ror#20
+	add	w7,w7,w11
+	ldr	w12,[sp,#16]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w3,w3,w7
+	eor	w13,w13,w9
+	st1	{v4.4s},[x17], #16
+	add	w6,w6,w12
+	add	w7,w7,w15
+	and	w12,w4,w3
+	ld1	{v1.16b},[x1],#16
+	bic	w15,w5,w3
+	eor	w11,w3,w3,ror#5
+	ld1	{v4.4s},[x16],#16
+	add	w7,w7,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w3,ror#19
+	eor	w15,w7,w7,ror#11
+	rev32	v1.16b,v1.16b
+	add	w6,w6,w12
+	ror	w11,w11,#6
+	eor	w13,w7,w8
+	eor	w15,w15,w7,ror#20
+	add	v4.4s,v4.4s,v1.4s
+	add	w6,w6,w11
+	ldr	w12,[sp,#20]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w10,w10,w6
+	eor	w14,w14,w8
+	add	w5,w5,w12
+	add	w6,w6,w15
+	and	w12,w3,w10
+	bic	w15,w4,w10
+	eor	w11,w10,w10,ror#5
+	add	w6,w6,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w10,ror#19
+	eor	w15,w6,w6,ror#11
+	add	w5,w5,w12
+	ror	w11,w11,#6
+	eor	w14,w6,w7
+	eor	w15,w15,w6,ror#20
+	add	w5,w5,w11
+	ldr	w12,[sp,#24]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w9,w9,w5
+	eor	w13,w13,w7
+	add	w4,w4,w12
+	add	w5,w5,w15
+	and	w12,w10,w9
+	bic	w15,w3,w9
+	eor	w11,w9,w9,ror#5
+	add	w5,w5,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w9,ror#19
+	eor	w15,w5,w5,ror#11
+	add	w4,w4,w12
+	ror	w11,w11,#6
+	eor	w13,w5,w6
+	eor	w15,w15,w5,ror#20
+	add	w4,w4,w11
+	ldr	w12,[sp,#28]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w8,w8,w4
+	eor	w14,w14,w6
+	add	w3,w3,w12
+	add	w4,w4,w15
+	and	w12,w9,w8
+	bic	w15,w10,w8
+	eor	w11,w8,w8,ror#5
+	add	w4,w4,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w8,ror#19
+	eor	w15,w4,w4,ror#11
+	add	w3,w3,w12
+	ror	w11,w11,#6
+	eor	w14,w4,w5
+	eor	w15,w15,w4,ror#20
+	add	w3,w3,w11
+	ldr	w12,[sp,#32]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w7,w7,w3
+	eor	w13,w13,w5
+	st1	{v4.4s},[x17], #16
+	add	w10,w10,w12
+	add	w3,w3,w15
+	and	w12,w8,w7
+	ld1	{v2.16b},[x1],#16
+	bic	w15,w9,w7
+	eor	w11,w7,w7,ror#5
+	ld1	{v4.4s},[x16],#16
+	add	w3,w3,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w7,ror#19
+	eor	w15,w3,w3,ror#11
+	rev32	v2.16b,v2.16b
+	add	w10,w10,w12
+	ror	w11,w11,#6
+	eor	w13,w3,w4
+	eor	w15,w15,w3,ror#20
+	add	v4.4s,v4.4s,v2.4s
+	add	w10,w10,w11
+	ldr	w12,[sp,#36]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w6,w6,w10
+	eor	w14,w14,w4
+	add	w9,w9,w12
+	add	w10,w10,w15
+	and	w12,w7,w6
+	bic	w15,w8,w6
+	eor	w11,w6,w6,ror#5
+	add	w10,w10,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w6,ror#19
+	eor	w15,w10,w10,ror#11
+	add	w9,w9,w12
+	ror	w11,w11,#6
+	eor	w14,w10,w3
+	eor	w15,w15,w10,ror#20
+	add	w9,w9,w11
+	ldr	w12,[sp,#40]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w5,w5,w9
+	eor	w13,w13,w3
+	add	w8,w8,w12
+	add	w9,w9,w15
+	and	w12,w6,w5
+	bic	w15,w7,w5
+	eor	w11,w5,w5,ror#5
+	add	w9,w9,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w5,ror#19
+	eor	w15,w9,w9,ror#11
+	add	w8,w8,w12
+	ror	w11,w11,#6
+	eor	w13,w9,w10
+	eor	w15,w15,w9,ror#20
+	add	w8,w8,w11
+	ldr	w12,[sp,#44]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w4,w4,w8
+	eor	w14,w14,w10
+	add	w7,w7,w12
+	add	w8,w8,w15
+	and	w12,w5,w4
+	bic	w15,w6,w4
+	eor	w11,w4,w4,ror#5
+	add	w8,w8,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w4,ror#19
+	eor	w15,w8,w8,ror#11
+	add	w7,w7,w12
+	ror	w11,w11,#6
+	eor	w14,w8,w9
+	eor	w15,w15,w8,ror#20
+	add	w7,w7,w11
+	ldr	w12,[sp,#48]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w3,w3,w7
+	eor	w13,w13,w9
+	st1	{v4.4s},[x17], #16
+	add	w6,w6,w12
+	add	w7,w7,w15
+	and	w12,w4,w3
+	ld1	{v3.16b},[x1],#16
+	bic	w15,w5,w3
+	eor	w11,w3,w3,ror#5
+	ld1	{v4.4s},[x16],#16
+	add	w7,w7,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w3,ror#19
+	eor	w15,w7,w7,ror#11
+	rev32	v3.16b,v3.16b
+	add	w6,w6,w12
+	ror	w11,w11,#6
+	eor	w13,w7,w8
+	eor	w15,w15,w7,ror#20
+	add	v4.4s,v4.4s,v3.4s
+	add	w6,w6,w11
+	ldr	w12,[sp,#52]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w10,w10,w6
+	eor	w14,w14,w8
+	add	w5,w5,w12
+	add	w6,w6,w15
+	and	w12,w3,w10
+	bic	w15,w4,w10
+	eor	w11,w10,w10,ror#5
+	add	w6,w6,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w10,ror#19
+	eor	w15,w6,w6,ror#11
+	add	w5,w5,w12
+	ror	w11,w11,#6
+	eor	w14,w6,w7
+	eor	w15,w15,w6,ror#20
+	add	w5,w5,w11
+	ldr	w12,[sp,#56]
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w9,w9,w5
+	eor	w13,w13,w7
+	add	w4,w4,w12
+	add	w5,w5,w15
+	and	w12,w10,w9
+	bic	w15,w3,w9
+	eor	w11,w9,w9,ror#5
+	add	w5,w5,w13
+	orr	w12,w12,w15
+	eor	w11,w11,w9,ror#19
+	eor	w15,w5,w5,ror#11
+	add	w4,w4,w12
+	ror	w11,w11,#6
+	eor	w13,w5,w6
+	eor	w15,w15,w5,ror#20
+	add	w4,w4,w11
+	ldr	w12,[sp,#60]
+	and	w14,w14,w13
+	ror	w15,w15,#2
+	add	w8,w8,w4
+	eor	w14,w14,w6
+	add	w3,w3,w12
+	add	w4,w4,w15
+	and	w12,w9,w8
+	bic	w15,w10,w8
+	eor	w11,w8,w8,ror#5
+	add	w4,w4,w14
+	orr	w12,w12,w15
+	eor	w11,w11,w8,ror#19
+	eor	w15,w4,w4,ror#11
+	add	w3,w3,w12
+	ror	w11,w11,#6
+	eor	w14,w4,w5
+	eor	w15,w15,w4,ror#20
+	add	w3,w3,w11
+	and	w13,w13,w14
+	ror	w15,w15,#2
+	add	w7,w7,w3
+	eor	w13,w13,w5
+	st1	{v4.4s},[x17], #16
+	add	w3,w3,w15			// h+=Sigma0(a) from the past
+	ldp	w11,w12,[x0,#0]
+	add	w3,w3,w13			// h+=Maj(a,b,c) from the past
+	ldp	w13,w14,[x0,#8]
+	add	w3,w3,w11			// accumulate
+	add	w4,w4,w12
+	ldp	w11,w12,[x0,#16]
+	add	w5,w5,w13
+	add	w6,w6,w14
+	ldp	w13,w14,[x0,#24]
+	add	w7,w7,w11
+	add	w8,w8,w12
+	 ldr	w12,[sp,#0]
+	stp	w3,w4,[x0,#0]
+	add	w9,w9,w13
+	 mov	w13,wzr
+	stp	w5,w6,[x0,#8]
+	add	w10,w10,w14
+	stp	w7,w8,[x0,#16]
+	 eor	w14,w4,w5
+	stp	w9,w10,[x0,#24]
+	 mov	w15,wzr
+	 mov	x17,sp
+	b.ne	.L_00_48
+
+	ldr	x29,[x29]
+	add	sp,sp,#16*4+16
+	ret
+.size	sha256_block_neon,.-sha256_block_neon
+#ifndef	__KERNEL__
+.comm	OPENSSL_armcap_P,4,4
+#endif
diff --git a/arch/arm64/crypto/sha512-core.S b/arch/arm64/crypto/sha512-core.S
new file mode 100644
index 000000000000..bd0f59f06c9d
--- /dev/null
+++ b/arch/arm64/crypto/sha512-core.S
@@ -0,0 +1,1085 @@
+// Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
+//
+// Licensed under the OpenSSL license (the "License").  You may not use
+// this file except in compliance with the License.  You can obtain a copy
+// in the file LICENSE in the source distribution or at
+// https://www.openssl.org/source/license.html
+
+// ====================================================================
+// Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+// project. The module is, however, dual licensed under OpenSSL and
+// CRYPTOGAMS licenses depending on where you obtain it. For further
+// details see http://www.openssl.org/~appro/cryptogams/.
+//
+// Permission to use under GPLv2 terms is granted.
+// ====================================================================
+//
+// SHA256/512 for ARMv8.
+//
+// Performance in cycles per processed byte and improvement coefficient
+// over code generated with "default" compiler:
+//
+//		SHA256-hw	SHA256(*)	SHA512
+// Apple A7	1.97		10.5 (+33%)	6.73 (-1%(**))
+// Cortex-A53	2.38		15.5 (+115%)	10.0 (+150%(***))
+// Cortex-A57	2.31		11.6 (+86%)	7.51 (+260%(***))
+// Denver	2.01		10.5 (+26%)	6.70 (+8%)
+// X-Gene			20.0 (+100%)	12.8 (+300%(***))
+// Mongoose	2.36		13.0 (+50%)	8.36 (+33%)
+//
+// (*)	Software SHA256 results are of lesser relevance, presented
+//	mostly for informational purposes.
+// (**)	The result is a trade-off: it's possible to improve it by
+//	10% (or by 1 cycle per round), but at the cost of 20% loss
+//	on Cortex-A53 (or by 4 cycles per round).
+// (***)	Super-impressive coefficients over gcc-generated code are
+//	indication of some compiler "pathology", most notably code
+//	generated with -mgeneral-regs-only is significanty faster
+//	and the gap is only 40-90%.
+//
+// October 2016.
+//
+// Originally it was reckoned that it makes no sense to implement NEON
+// version of SHA256 for 64-bit processors. This is because performance
+// improvement on most wide-spread Cortex-A5x processors was observed
+// to be marginal, same on Cortex-A53 and ~10% on A57. But then it was
+// observed that 32-bit NEON SHA256 performs significantly better than
+// 64-bit scalar version on *some* of the more recent processors. As
+// result 64-bit NEON version of SHA256 was added to provide best
+// all-round performance. For example it executes ~30% faster on X-Gene
+// and Mongoose. [For reference, NEON version of SHA512 is bound to
+// deliver much less improvement, likely *negative* on Cortex-A5x.
+// Which is why NEON support is limited to SHA256.]
+
+#ifndef	__KERNEL__
+# include "arm_arch.h"
+#endif
+
+.text
+
+.extern	OPENSSL_armcap_P
+.globl	sha512_block_data_order
+.type	sha512_block_data_order,%function
+.align	6
+sha512_block_data_order:
+	stp	x29,x30,[sp,#-128]!
+	add	x29,sp,#0
+
+	stp	x19,x20,[sp,#16]
+	stp	x21,x22,[sp,#32]
+	stp	x23,x24,[sp,#48]
+	stp	x25,x26,[sp,#64]
+	stp	x27,x28,[sp,#80]
+	sub	sp,sp,#4*8
+
+	ldp	x20,x21,[x0]				// load context
+	ldp	x22,x23,[x0,#2*8]
+	ldp	x24,x25,[x0,#4*8]
+	add	x2,x1,x2,lsl#7	// end of input
+	ldp	x26,x27,[x0,#6*8]
+	adr	x30,.LK512
+	stp	x0,x2,[x29,#96]
+
+.Loop:
+	ldp	x3,x4,[x1],#2*8
+	ldr	x19,[x30],#8			// *K++
+	eor	x28,x21,x22				// magic seed
+	str	x1,[x29,#112]
+#ifndef	__AARCH64EB__
+	rev	x3,x3			// 0
+#endif
+	ror	x16,x24,#14
+	add	x27,x27,x19			// h+=K[i]
+	eor	x6,x24,x24,ror#23
+	and	x17,x25,x24
+	bic	x19,x26,x24
+	add	x27,x27,x3			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x20,x21			// a^b, b^c in next round
+	eor	x16,x16,x6,ror#18	// Sigma1(e)
+	ror	x6,x20,#28
+	add	x27,x27,x17			// h+=Ch(e,f,g)
+	eor	x17,x20,x20,ror#5
+	add	x27,x27,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x23,x23,x27			// d+=h
+	eor	x28,x28,x21			// Maj(a,b,c)
+	eor	x17,x6,x17,ror#34	// Sigma0(a)
+	add	x27,x27,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x27,x27,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x4,x4			// 1
+#endif
+	ldp	x5,x6,[x1],#2*8
+	add	x27,x27,x17			// h+=Sigma0(a)
+	ror	x16,x23,#14
+	add	x26,x26,x28			// h+=K[i]
+	eor	x7,x23,x23,ror#23
+	and	x17,x24,x23
+	bic	x28,x25,x23
+	add	x26,x26,x4			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x27,x20			// a^b, b^c in next round
+	eor	x16,x16,x7,ror#18	// Sigma1(e)
+	ror	x7,x27,#28
+	add	x26,x26,x17			// h+=Ch(e,f,g)
+	eor	x17,x27,x27,ror#5
+	add	x26,x26,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x22,x22,x26			// d+=h
+	eor	x19,x19,x20			// Maj(a,b,c)
+	eor	x17,x7,x17,ror#34	// Sigma0(a)
+	add	x26,x26,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x26,x26,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x5,x5			// 2
+#endif
+	add	x26,x26,x17			// h+=Sigma0(a)
+	ror	x16,x22,#14
+	add	x25,x25,x19			// h+=K[i]
+	eor	x8,x22,x22,ror#23
+	and	x17,x23,x22
+	bic	x19,x24,x22
+	add	x25,x25,x5			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x26,x27			// a^b, b^c in next round
+	eor	x16,x16,x8,ror#18	// Sigma1(e)
+	ror	x8,x26,#28
+	add	x25,x25,x17			// h+=Ch(e,f,g)
+	eor	x17,x26,x26,ror#5
+	add	x25,x25,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x21,x21,x25			// d+=h
+	eor	x28,x28,x27			// Maj(a,b,c)
+	eor	x17,x8,x17,ror#34	// Sigma0(a)
+	add	x25,x25,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x25,x25,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x6,x6			// 3
+#endif
+	ldp	x7,x8,[x1],#2*8
+	add	x25,x25,x17			// h+=Sigma0(a)
+	ror	x16,x21,#14
+	add	x24,x24,x28			// h+=K[i]
+	eor	x9,x21,x21,ror#23
+	and	x17,x22,x21
+	bic	x28,x23,x21
+	add	x24,x24,x6			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x25,x26			// a^b, b^c in next round
+	eor	x16,x16,x9,ror#18	// Sigma1(e)
+	ror	x9,x25,#28
+	add	x24,x24,x17			// h+=Ch(e,f,g)
+	eor	x17,x25,x25,ror#5
+	add	x24,x24,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x20,x20,x24			// d+=h
+	eor	x19,x19,x26			// Maj(a,b,c)
+	eor	x17,x9,x17,ror#34	// Sigma0(a)
+	add	x24,x24,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x24,x24,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x7,x7			// 4
+#endif
+	add	x24,x24,x17			// h+=Sigma0(a)
+	ror	x16,x20,#14
+	add	x23,x23,x19			// h+=K[i]
+	eor	x10,x20,x20,ror#23
+	and	x17,x21,x20
+	bic	x19,x22,x20
+	add	x23,x23,x7			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x24,x25			// a^b, b^c in next round
+	eor	x16,x16,x10,ror#18	// Sigma1(e)
+	ror	x10,x24,#28
+	add	x23,x23,x17			// h+=Ch(e,f,g)
+	eor	x17,x24,x24,ror#5
+	add	x23,x23,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x27,x27,x23			// d+=h
+	eor	x28,x28,x25			// Maj(a,b,c)
+	eor	x17,x10,x17,ror#34	// Sigma0(a)
+	add	x23,x23,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x23,x23,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x8,x8			// 5
+#endif
+	ldp	x9,x10,[x1],#2*8
+	add	x23,x23,x17			// h+=Sigma0(a)
+	ror	x16,x27,#14
+	add	x22,x22,x28			// h+=K[i]
+	eor	x11,x27,x27,ror#23
+	and	x17,x20,x27
+	bic	x28,x21,x27
+	add	x22,x22,x8			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x23,x24			// a^b, b^c in next round
+	eor	x16,x16,x11,ror#18	// Sigma1(e)
+	ror	x11,x23,#28
+	add	x22,x22,x17			// h+=Ch(e,f,g)
+	eor	x17,x23,x23,ror#5
+	add	x22,x22,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x26,x26,x22			// d+=h
+	eor	x19,x19,x24			// Maj(a,b,c)
+	eor	x17,x11,x17,ror#34	// Sigma0(a)
+	add	x22,x22,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x22,x22,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x9,x9			// 6
+#endif
+	add	x22,x22,x17			// h+=Sigma0(a)
+	ror	x16,x26,#14
+	add	x21,x21,x19			// h+=K[i]
+	eor	x12,x26,x26,ror#23
+	and	x17,x27,x26
+	bic	x19,x20,x26
+	add	x21,x21,x9			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x22,x23			// a^b, b^c in next round
+	eor	x16,x16,x12,ror#18	// Sigma1(e)
+	ror	x12,x22,#28
+	add	x21,x21,x17			// h+=Ch(e,f,g)
+	eor	x17,x22,x22,ror#5
+	add	x21,x21,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x25,x25,x21			// d+=h
+	eor	x28,x28,x23			// Maj(a,b,c)
+	eor	x17,x12,x17,ror#34	// Sigma0(a)
+	add	x21,x21,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x21,x21,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x10,x10			// 7
+#endif
+	ldp	x11,x12,[x1],#2*8
+	add	x21,x21,x17			// h+=Sigma0(a)
+	ror	x16,x25,#14
+	add	x20,x20,x28			// h+=K[i]
+	eor	x13,x25,x25,ror#23
+	and	x17,x26,x25
+	bic	x28,x27,x25
+	add	x20,x20,x10			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x21,x22			// a^b, b^c in next round
+	eor	x16,x16,x13,ror#18	// Sigma1(e)
+	ror	x13,x21,#28
+	add	x20,x20,x17			// h+=Ch(e,f,g)
+	eor	x17,x21,x21,ror#5
+	add	x20,x20,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x24,x24,x20			// d+=h
+	eor	x19,x19,x22			// Maj(a,b,c)
+	eor	x17,x13,x17,ror#34	// Sigma0(a)
+	add	x20,x20,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x20,x20,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x11,x11			// 8
+#endif
+	add	x20,x20,x17			// h+=Sigma0(a)
+	ror	x16,x24,#14
+	add	x27,x27,x19			// h+=K[i]
+	eor	x14,x24,x24,ror#23
+	and	x17,x25,x24
+	bic	x19,x26,x24
+	add	x27,x27,x11			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x20,x21			// a^b, b^c in next round
+	eor	x16,x16,x14,ror#18	// Sigma1(e)
+	ror	x14,x20,#28
+	add	x27,x27,x17			// h+=Ch(e,f,g)
+	eor	x17,x20,x20,ror#5
+	add	x27,x27,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x23,x23,x27			// d+=h
+	eor	x28,x28,x21			// Maj(a,b,c)
+	eor	x17,x14,x17,ror#34	// Sigma0(a)
+	add	x27,x27,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x27,x27,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x12,x12			// 9
+#endif
+	ldp	x13,x14,[x1],#2*8
+	add	x27,x27,x17			// h+=Sigma0(a)
+	ror	x16,x23,#14
+	add	x26,x26,x28			// h+=K[i]
+	eor	x15,x23,x23,ror#23
+	and	x17,x24,x23
+	bic	x28,x25,x23
+	add	x26,x26,x12			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x27,x20			// a^b, b^c in next round
+	eor	x16,x16,x15,ror#18	// Sigma1(e)
+	ror	x15,x27,#28
+	add	x26,x26,x17			// h+=Ch(e,f,g)
+	eor	x17,x27,x27,ror#5
+	add	x26,x26,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x22,x22,x26			// d+=h
+	eor	x19,x19,x20			// Maj(a,b,c)
+	eor	x17,x15,x17,ror#34	// Sigma0(a)
+	add	x26,x26,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x26,x26,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x13,x13			// 10
+#endif
+	add	x26,x26,x17			// h+=Sigma0(a)
+	ror	x16,x22,#14
+	add	x25,x25,x19			// h+=K[i]
+	eor	x0,x22,x22,ror#23
+	and	x17,x23,x22
+	bic	x19,x24,x22
+	add	x25,x25,x13			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x26,x27			// a^b, b^c in next round
+	eor	x16,x16,x0,ror#18	// Sigma1(e)
+	ror	x0,x26,#28
+	add	x25,x25,x17			// h+=Ch(e,f,g)
+	eor	x17,x26,x26,ror#5
+	add	x25,x25,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x21,x21,x25			// d+=h
+	eor	x28,x28,x27			// Maj(a,b,c)
+	eor	x17,x0,x17,ror#34	// Sigma0(a)
+	add	x25,x25,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x25,x25,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x14,x14			// 11
+#endif
+	ldp	x15,x0,[x1],#2*8
+	add	x25,x25,x17			// h+=Sigma0(a)
+	str	x6,[sp,#24]
+	ror	x16,x21,#14
+	add	x24,x24,x28			// h+=K[i]
+	eor	x6,x21,x21,ror#23
+	and	x17,x22,x21
+	bic	x28,x23,x21
+	add	x24,x24,x14			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x25,x26			// a^b, b^c in next round
+	eor	x16,x16,x6,ror#18	// Sigma1(e)
+	ror	x6,x25,#28
+	add	x24,x24,x17			// h+=Ch(e,f,g)
+	eor	x17,x25,x25,ror#5
+	add	x24,x24,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x20,x20,x24			// d+=h
+	eor	x19,x19,x26			// Maj(a,b,c)
+	eor	x17,x6,x17,ror#34	// Sigma0(a)
+	add	x24,x24,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x24,x24,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x15,x15			// 12
+#endif
+	add	x24,x24,x17			// h+=Sigma0(a)
+	str	x7,[sp,#0]
+	ror	x16,x20,#14
+	add	x23,x23,x19			// h+=K[i]
+	eor	x7,x20,x20,ror#23
+	and	x17,x21,x20
+	bic	x19,x22,x20
+	add	x23,x23,x15			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x24,x25			// a^b, b^c in next round
+	eor	x16,x16,x7,ror#18	// Sigma1(e)
+	ror	x7,x24,#28
+	add	x23,x23,x17			// h+=Ch(e,f,g)
+	eor	x17,x24,x24,ror#5
+	add	x23,x23,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x27,x27,x23			// d+=h
+	eor	x28,x28,x25			// Maj(a,b,c)
+	eor	x17,x7,x17,ror#34	// Sigma0(a)
+	add	x23,x23,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x23,x23,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x0,x0			// 13
+#endif
+	ldp	x1,x2,[x1]
+	add	x23,x23,x17			// h+=Sigma0(a)
+	str	x8,[sp,#8]
+	ror	x16,x27,#14
+	add	x22,x22,x28			// h+=K[i]
+	eor	x8,x27,x27,ror#23
+	and	x17,x20,x27
+	bic	x28,x21,x27
+	add	x22,x22,x0			// h+=X[i]
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x23,x24			// a^b, b^c in next round
+	eor	x16,x16,x8,ror#18	// Sigma1(e)
+	ror	x8,x23,#28
+	add	x22,x22,x17			// h+=Ch(e,f,g)
+	eor	x17,x23,x23,ror#5
+	add	x22,x22,x16			// h+=Sigma1(e)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	add	x26,x26,x22			// d+=h
+	eor	x19,x19,x24			// Maj(a,b,c)
+	eor	x17,x8,x17,ror#34	// Sigma0(a)
+	add	x22,x22,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	//add	x22,x22,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x1,x1			// 14
+#endif
+	ldr	x6,[sp,#24]
+	add	x22,x22,x17			// h+=Sigma0(a)
+	str	x9,[sp,#16]
+	ror	x16,x26,#14
+	add	x21,x21,x19			// h+=K[i]
+	eor	x9,x26,x26,ror#23
+	and	x17,x27,x26
+	bic	x19,x20,x26
+	add	x21,x21,x1			// h+=X[i]
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x22,x23			// a^b, b^c in next round
+	eor	x16,x16,x9,ror#18	// Sigma1(e)
+	ror	x9,x22,#28
+	add	x21,x21,x17			// h+=Ch(e,f,g)
+	eor	x17,x22,x22,ror#5
+	add	x21,x21,x16			// h+=Sigma1(e)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	add	x25,x25,x21			// d+=h
+	eor	x28,x28,x23			// Maj(a,b,c)
+	eor	x17,x9,x17,ror#34	// Sigma0(a)
+	add	x21,x21,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	//add	x21,x21,x17			// h+=Sigma0(a)
+#ifndef	__AARCH64EB__
+	rev	x2,x2			// 15
+#endif
+	ldr	x7,[sp,#0]
+	add	x21,x21,x17			// h+=Sigma0(a)
+	str	x10,[sp,#24]
+	ror	x16,x25,#14
+	add	x20,x20,x28			// h+=K[i]
+	ror	x9,x4,#1
+	and	x17,x26,x25
+	ror	x8,x1,#19
+	bic	x28,x27,x25
+	ror	x10,x21,#28
+	add	x20,x20,x2			// h+=X[i]
+	eor	x16,x16,x25,ror#18
+	eor	x9,x9,x4,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x21,x22			// a^b, b^c in next round
+	eor	x16,x16,x25,ror#41	// Sigma1(e)
+	eor	x10,x10,x21,ror#34
+	add	x20,x20,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x8,x8,x1,ror#61
+	eor	x9,x9,x4,lsr#7	// sigma0(X[i+1])
+	add	x20,x20,x16			// h+=Sigma1(e)
+	eor	x19,x19,x22			// Maj(a,b,c)
+	eor	x17,x10,x21,ror#39	// Sigma0(a)
+	eor	x8,x8,x1,lsr#6	// sigma1(X[i+14])
+	add	x3,x3,x12
+	add	x24,x24,x20			// d+=h
+	add	x20,x20,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x3,x3,x9
+	add	x20,x20,x17			// h+=Sigma0(a)
+	add	x3,x3,x8
+.Loop_16_xx:
+	ldr	x8,[sp,#8]
+	str	x11,[sp,#0]
+	ror	x16,x24,#14
+	add	x27,x27,x19			// h+=K[i]
+	ror	x10,x5,#1
+	and	x17,x25,x24
+	ror	x9,x2,#19
+	bic	x19,x26,x24
+	ror	x11,x20,#28
+	add	x27,x27,x3			// h+=X[i]
+	eor	x16,x16,x24,ror#18
+	eor	x10,x10,x5,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x20,x21			// a^b, b^c in next round
+	eor	x16,x16,x24,ror#41	// Sigma1(e)
+	eor	x11,x11,x20,ror#34
+	add	x27,x27,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x9,x9,x2,ror#61
+	eor	x10,x10,x5,lsr#7	// sigma0(X[i+1])
+	add	x27,x27,x16			// h+=Sigma1(e)
+	eor	x28,x28,x21			// Maj(a,b,c)
+	eor	x17,x11,x20,ror#39	// Sigma0(a)
+	eor	x9,x9,x2,lsr#6	// sigma1(X[i+14])
+	add	x4,x4,x13
+	add	x23,x23,x27			// d+=h
+	add	x27,x27,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x4,x4,x10
+	add	x27,x27,x17			// h+=Sigma0(a)
+	add	x4,x4,x9
+	ldr	x9,[sp,#16]
+	str	x12,[sp,#8]
+	ror	x16,x23,#14
+	add	x26,x26,x28			// h+=K[i]
+	ror	x11,x6,#1
+	and	x17,x24,x23
+	ror	x10,x3,#19
+	bic	x28,x25,x23
+	ror	x12,x27,#28
+	add	x26,x26,x4			// h+=X[i]
+	eor	x16,x16,x23,ror#18
+	eor	x11,x11,x6,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x27,x20			// a^b, b^c in next round
+	eor	x16,x16,x23,ror#41	// Sigma1(e)
+	eor	x12,x12,x27,ror#34
+	add	x26,x26,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x10,x10,x3,ror#61
+	eor	x11,x11,x6,lsr#7	// sigma0(X[i+1])
+	add	x26,x26,x16			// h+=Sigma1(e)
+	eor	x19,x19,x20			// Maj(a,b,c)
+	eor	x17,x12,x27,ror#39	// Sigma0(a)
+	eor	x10,x10,x3,lsr#6	// sigma1(X[i+14])
+	add	x5,x5,x14
+	add	x22,x22,x26			// d+=h
+	add	x26,x26,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x5,x5,x11
+	add	x26,x26,x17			// h+=Sigma0(a)
+	add	x5,x5,x10
+	ldr	x10,[sp,#24]
+	str	x13,[sp,#16]
+	ror	x16,x22,#14
+	add	x25,x25,x19			// h+=K[i]
+	ror	x12,x7,#1
+	and	x17,x23,x22
+	ror	x11,x4,#19
+	bic	x19,x24,x22
+	ror	x13,x26,#28
+	add	x25,x25,x5			// h+=X[i]
+	eor	x16,x16,x22,ror#18
+	eor	x12,x12,x7,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x26,x27			// a^b, b^c in next round
+	eor	x16,x16,x22,ror#41	// Sigma1(e)
+	eor	x13,x13,x26,ror#34
+	add	x25,x25,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x11,x11,x4,ror#61
+	eor	x12,x12,x7,lsr#7	// sigma0(X[i+1])
+	add	x25,x25,x16			// h+=Sigma1(e)
+	eor	x28,x28,x27			// Maj(a,b,c)
+	eor	x17,x13,x26,ror#39	// Sigma0(a)
+	eor	x11,x11,x4,lsr#6	// sigma1(X[i+14])
+	add	x6,x6,x15
+	add	x21,x21,x25			// d+=h
+	add	x25,x25,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x6,x6,x12
+	add	x25,x25,x17			// h+=Sigma0(a)
+	add	x6,x6,x11
+	ldr	x11,[sp,#0]
+	str	x14,[sp,#24]
+	ror	x16,x21,#14
+	add	x24,x24,x28			// h+=K[i]
+	ror	x13,x8,#1
+	and	x17,x22,x21
+	ror	x12,x5,#19
+	bic	x28,x23,x21
+	ror	x14,x25,#28
+	add	x24,x24,x6			// h+=X[i]
+	eor	x16,x16,x21,ror#18
+	eor	x13,x13,x8,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x25,x26			// a^b, b^c in next round
+	eor	x16,x16,x21,ror#41	// Sigma1(e)
+	eor	x14,x14,x25,ror#34
+	add	x24,x24,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x12,x12,x5,ror#61
+	eor	x13,x13,x8,lsr#7	// sigma0(X[i+1])
+	add	x24,x24,x16			// h+=Sigma1(e)
+	eor	x19,x19,x26			// Maj(a,b,c)
+	eor	x17,x14,x25,ror#39	// Sigma0(a)
+	eor	x12,x12,x5,lsr#6	// sigma1(X[i+14])
+	add	x7,x7,x0
+	add	x20,x20,x24			// d+=h
+	add	x24,x24,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x7,x7,x13
+	add	x24,x24,x17			// h+=Sigma0(a)
+	add	x7,x7,x12
+	ldr	x12,[sp,#8]
+	str	x15,[sp,#0]
+	ror	x16,x20,#14
+	add	x23,x23,x19			// h+=K[i]
+	ror	x14,x9,#1
+	and	x17,x21,x20
+	ror	x13,x6,#19
+	bic	x19,x22,x20
+	ror	x15,x24,#28
+	add	x23,x23,x7			// h+=X[i]
+	eor	x16,x16,x20,ror#18
+	eor	x14,x14,x9,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x24,x25			// a^b, b^c in next round
+	eor	x16,x16,x20,ror#41	// Sigma1(e)
+	eor	x15,x15,x24,ror#34
+	add	x23,x23,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x13,x13,x6,ror#61
+	eor	x14,x14,x9,lsr#7	// sigma0(X[i+1])
+	add	x23,x23,x16			// h+=Sigma1(e)
+	eor	x28,x28,x25			// Maj(a,b,c)
+	eor	x17,x15,x24,ror#39	// Sigma0(a)
+	eor	x13,x13,x6,lsr#6	// sigma1(X[i+14])
+	add	x8,x8,x1
+	add	x27,x27,x23			// d+=h
+	add	x23,x23,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x8,x8,x14
+	add	x23,x23,x17			// h+=Sigma0(a)
+	add	x8,x8,x13
+	ldr	x13,[sp,#16]
+	str	x0,[sp,#8]
+	ror	x16,x27,#14
+	add	x22,x22,x28			// h+=K[i]
+	ror	x15,x10,#1
+	and	x17,x20,x27
+	ror	x14,x7,#19
+	bic	x28,x21,x27
+	ror	x0,x23,#28
+	add	x22,x22,x8			// h+=X[i]
+	eor	x16,x16,x27,ror#18
+	eor	x15,x15,x10,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x23,x24			// a^b, b^c in next round
+	eor	x16,x16,x27,ror#41	// Sigma1(e)
+	eor	x0,x0,x23,ror#34
+	add	x22,x22,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x14,x14,x7,ror#61
+	eor	x15,x15,x10,lsr#7	// sigma0(X[i+1])
+	add	x22,x22,x16			// h+=Sigma1(e)
+	eor	x19,x19,x24			// Maj(a,b,c)
+	eor	x17,x0,x23,ror#39	// Sigma0(a)
+	eor	x14,x14,x7,lsr#6	// sigma1(X[i+14])
+	add	x9,x9,x2
+	add	x26,x26,x22			// d+=h
+	add	x22,x22,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x9,x9,x15
+	add	x22,x22,x17			// h+=Sigma0(a)
+	add	x9,x9,x14
+	ldr	x14,[sp,#24]
+	str	x1,[sp,#16]
+	ror	x16,x26,#14
+	add	x21,x21,x19			// h+=K[i]
+	ror	x0,x11,#1
+	and	x17,x27,x26
+	ror	x15,x8,#19
+	bic	x19,x20,x26
+	ror	x1,x22,#28
+	add	x21,x21,x9			// h+=X[i]
+	eor	x16,x16,x26,ror#18
+	eor	x0,x0,x11,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x22,x23			// a^b, b^c in next round
+	eor	x16,x16,x26,ror#41	// Sigma1(e)
+	eor	x1,x1,x22,ror#34
+	add	x21,x21,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x15,x15,x8,ror#61
+	eor	x0,x0,x11,lsr#7	// sigma0(X[i+1])
+	add	x21,x21,x16			// h+=Sigma1(e)
+	eor	x28,x28,x23			// Maj(a,b,c)
+	eor	x17,x1,x22,ror#39	// Sigma0(a)
+	eor	x15,x15,x8,lsr#6	// sigma1(X[i+14])
+	add	x10,x10,x3
+	add	x25,x25,x21			// d+=h
+	add	x21,x21,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x10,x10,x0
+	add	x21,x21,x17			// h+=Sigma0(a)
+	add	x10,x10,x15
+	ldr	x15,[sp,#0]
+	str	x2,[sp,#24]
+	ror	x16,x25,#14
+	add	x20,x20,x28			// h+=K[i]
+	ror	x1,x12,#1
+	and	x17,x26,x25
+	ror	x0,x9,#19
+	bic	x28,x27,x25
+	ror	x2,x21,#28
+	add	x20,x20,x10			// h+=X[i]
+	eor	x16,x16,x25,ror#18
+	eor	x1,x1,x12,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x21,x22			// a^b, b^c in next round
+	eor	x16,x16,x25,ror#41	// Sigma1(e)
+	eor	x2,x2,x21,ror#34
+	add	x20,x20,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x0,x0,x9,ror#61
+	eor	x1,x1,x12,lsr#7	// sigma0(X[i+1])
+	add	x20,x20,x16			// h+=Sigma1(e)
+	eor	x19,x19,x22			// Maj(a,b,c)
+	eor	x17,x2,x21,ror#39	// Sigma0(a)
+	eor	x0,x0,x9,lsr#6	// sigma1(X[i+14])
+	add	x11,x11,x4
+	add	x24,x24,x20			// d+=h
+	add	x20,x20,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x11,x11,x1
+	add	x20,x20,x17			// h+=Sigma0(a)
+	add	x11,x11,x0
+	ldr	x0,[sp,#8]
+	str	x3,[sp,#0]
+	ror	x16,x24,#14
+	add	x27,x27,x19			// h+=K[i]
+	ror	x2,x13,#1
+	and	x17,x25,x24
+	ror	x1,x10,#19
+	bic	x19,x26,x24
+	ror	x3,x20,#28
+	add	x27,x27,x11			// h+=X[i]
+	eor	x16,x16,x24,ror#18
+	eor	x2,x2,x13,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x20,x21			// a^b, b^c in next round
+	eor	x16,x16,x24,ror#41	// Sigma1(e)
+	eor	x3,x3,x20,ror#34
+	add	x27,x27,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x1,x1,x10,ror#61
+	eor	x2,x2,x13,lsr#7	// sigma0(X[i+1])
+	add	x27,x27,x16			// h+=Sigma1(e)
+	eor	x28,x28,x21			// Maj(a,b,c)
+	eor	x17,x3,x20,ror#39	// Sigma0(a)
+	eor	x1,x1,x10,lsr#6	// sigma1(X[i+14])
+	add	x12,x12,x5
+	add	x23,x23,x27			// d+=h
+	add	x27,x27,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x12,x12,x2
+	add	x27,x27,x17			// h+=Sigma0(a)
+	add	x12,x12,x1
+	ldr	x1,[sp,#16]
+	str	x4,[sp,#8]
+	ror	x16,x23,#14
+	add	x26,x26,x28			// h+=K[i]
+	ror	x3,x14,#1
+	and	x17,x24,x23
+	ror	x2,x11,#19
+	bic	x28,x25,x23
+	ror	x4,x27,#28
+	add	x26,x26,x12			// h+=X[i]
+	eor	x16,x16,x23,ror#18
+	eor	x3,x3,x14,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x27,x20			// a^b, b^c in next round
+	eor	x16,x16,x23,ror#41	// Sigma1(e)
+	eor	x4,x4,x27,ror#34
+	add	x26,x26,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x2,x2,x11,ror#61
+	eor	x3,x3,x14,lsr#7	// sigma0(X[i+1])
+	add	x26,x26,x16			// h+=Sigma1(e)
+	eor	x19,x19,x20			// Maj(a,b,c)
+	eor	x17,x4,x27,ror#39	// Sigma0(a)
+	eor	x2,x2,x11,lsr#6	// sigma1(X[i+14])
+	add	x13,x13,x6
+	add	x22,x22,x26			// d+=h
+	add	x26,x26,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x13,x13,x3
+	add	x26,x26,x17			// h+=Sigma0(a)
+	add	x13,x13,x2
+	ldr	x2,[sp,#24]
+	str	x5,[sp,#16]
+	ror	x16,x22,#14
+	add	x25,x25,x19			// h+=K[i]
+	ror	x4,x15,#1
+	and	x17,x23,x22
+	ror	x3,x12,#19
+	bic	x19,x24,x22
+	ror	x5,x26,#28
+	add	x25,x25,x13			// h+=X[i]
+	eor	x16,x16,x22,ror#18
+	eor	x4,x4,x15,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x26,x27			// a^b, b^c in next round
+	eor	x16,x16,x22,ror#41	// Sigma1(e)
+	eor	x5,x5,x26,ror#34
+	add	x25,x25,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x3,x3,x12,ror#61
+	eor	x4,x4,x15,lsr#7	// sigma0(X[i+1])
+	add	x25,x25,x16			// h+=Sigma1(e)
+	eor	x28,x28,x27			// Maj(a,b,c)
+	eor	x17,x5,x26,ror#39	// Sigma0(a)
+	eor	x3,x3,x12,lsr#6	// sigma1(X[i+14])
+	add	x14,x14,x7
+	add	x21,x21,x25			// d+=h
+	add	x25,x25,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x14,x14,x4
+	add	x25,x25,x17			// h+=Sigma0(a)
+	add	x14,x14,x3
+	ldr	x3,[sp,#0]
+	str	x6,[sp,#24]
+	ror	x16,x21,#14
+	add	x24,x24,x28			// h+=K[i]
+	ror	x5,x0,#1
+	and	x17,x22,x21
+	ror	x4,x13,#19
+	bic	x28,x23,x21
+	ror	x6,x25,#28
+	add	x24,x24,x14			// h+=X[i]
+	eor	x16,x16,x21,ror#18
+	eor	x5,x5,x0,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x25,x26			// a^b, b^c in next round
+	eor	x16,x16,x21,ror#41	// Sigma1(e)
+	eor	x6,x6,x25,ror#34
+	add	x24,x24,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x4,x4,x13,ror#61
+	eor	x5,x5,x0,lsr#7	// sigma0(X[i+1])
+	add	x24,x24,x16			// h+=Sigma1(e)
+	eor	x19,x19,x26			// Maj(a,b,c)
+	eor	x17,x6,x25,ror#39	// Sigma0(a)
+	eor	x4,x4,x13,lsr#6	// sigma1(X[i+14])
+	add	x15,x15,x8
+	add	x20,x20,x24			// d+=h
+	add	x24,x24,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x15,x15,x5
+	add	x24,x24,x17			// h+=Sigma0(a)
+	add	x15,x15,x4
+	ldr	x4,[sp,#8]
+	str	x7,[sp,#0]
+	ror	x16,x20,#14
+	add	x23,x23,x19			// h+=K[i]
+	ror	x6,x1,#1
+	and	x17,x21,x20
+	ror	x5,x14,#19
+	bic	x19,x22,x20
+	ror	x7,x24,#28
+	add	x23,x23,x15			// h+=X[i]
+	eor	x16,x16,x20,ror#18
+	eor	x6,x6,x1,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x24,x25			// a^b, b^c in next round
+	eor	x16,x16,x20,ror#41	// Sigma1(e)
+	eor	x7,x7,x24,ror#34
+	add	x23,x23,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x5,x5,x14,ror#61
+	eor	x6,x6,x1,lsr#7	// sigma0(X[i+1])
+	add	x23,x23,x16			// h+=Sigma1(e)
+	eor	x28,x28,x25			// Maj(a,b,c)
+	eor	x17,x7,x24,ror#39	// Sigma0(a)
+	eor	x5,x5,x14,lsr#6	// sigma1(X[i+14])
+	add	x0,x0,x9
+	add	x27,x27,x23			// d+=h
+	add	x23,x23,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x0,x0,x6
+	add	x23,x23,x17			// h+=Sigma0(a)
+	add	x0,x0,x5
+	ldr	x5,[sp,#16]
+	str	x8,[sp,#8]
+	ror	x16,x27,#14
+	add	x22,x22,x28			// h+=K[i]
+	ror	x7,x2,#1
+	and	x17,x20,x27
+	ror	x6,x15,#19
+	bic	x28,x21,x27
+	ror	x8,x23,#28
+	add	x22,x22,x0			// h+=X[i]
+	eor	x16,x16,x27,ror#18
+	eor	x7,x7,x2,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x23,x24			// a^b, b^c in next round
+	eor	x16,x16,x27,ror#41	// Sigma1(e)
+	eor	x8,x8,x23,ror#34
+	add	x22,x22,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x6,x6,x15,ror#61
+	eor	x7,x7,x2,lsr#7	// sigma0(X[i+1])
+	add	x22,x22,x16			// h+=Sigma1(e)
+	eor	x19,x19,x24			// Maj(a,b,c)
+	eor	x17,x8,x23,ror#39	// Sigma0(a)
+	eor	x6,x6,x15,lsr#6	// sigma1(X[i+14])
+	add	x1,x1,x10
+	add	x26,x26,x22			// d+=h
+	add	x22,x22,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x1,x1,x7
+	add	x22,x22,x17			// h+=Sigma0(a)
+	add	x1,x1,x6
+	ldr	x6,[sp,#24]
+	str	x9,[sp,#16]
+	ror	x16,x26,#14
+	add	x21,x21,x19			// h+=K[i]
+	ror	x8,x3,#1
+	and	x17,x27,x26
+	ror	x7,x0,#19
+	bic	x19,x20,x26
+	ror	x9,x22,#28
+	add	x21,x21,x1			// h+=X[i]
+	eor	x16,x16,x26,ror#18
+	eor	x8,x8,x3,ror#8
+	orr	x17,x17,x19			// Ch(e,f,g)
+	eor	x19,x22,x23			// a^b, b^c in next round
+	eor	x16,x16,x26,ror#41	// Sigma1(e)
+	eor	x9,x9,x22,ror#34
+	add	x21,x21,x17			// h+=Ch(e,f,g)
+	and	x28,x28,x19			// (b^c)&=(a^b)
+	eor	x7,x7,x0,ror#61
+	eor	x8,x8,x3,lsr#7	// sigma0(X[i+1])
+	add	x21,x21,x16			// h+=Sigma1(e)
+	eor	x28,x28,x23			// Maj(a,b,c)
+	eor	x17,x9,x22,ror#39	// Sigma0(a)
+	eor	x7,x7,x0,lsr#6	// sigma1(X[i+14])
+	add	x2,x2,x11
+	add	x25,x25,x21			// d+=h
+	add	x21,x21,x28			// h+=Maj(a,b,c)
+	ldr	x28,[x30],#8		// *K++, x19 in next round
+	add	x2,x2,x8
+	add	x21,x21,x17			// h+=Sigma0(a)
+	add	x2,x2,x7
+	ldr	x7,[sp,#0]
+	str	x10,[sp,#24]
+	ror	x16,x25,#14
+	add	x20,x20,x28			// h+=K[i]
+	ror	x9,x4,#1
+	and	x17,x26,x25
+	ror	x8,x1,#19
+	bic	x28,x27,x25
+	ror	x10,x21,#28
+	add	x20,x20,x2			// h+=X[i]
+	eor	x16,x16,x25,ror#18
+	eor	x9,x9,x4,ror#8
+	orr	x17,x17,x28			// Ch(e,f,g)
+	eor	x28,x21,x22			// a^b, b^c in next round
+	eor	x16,x16,x25,ror#41	// Sigma1(e)
+	eor	x10,x10,x21,ror#34
+	add	x20,x20,x17			// h+=Ch(e,f,g)
+	and	x19,x19,x28			// (b^c)&=(a^b)
+	eor	x8,x8,x1,ror#61
+	eor	x9,x9,x4,lsr#7	// sigma0(X[i+1])
+	add	x20,x20,x16			// h+=Sigma1(e)
+	eor	x19,x19,x22			// Maj(a,b,c)
+	eor	x17,x10,x21,ror#39	// Sigma0(a)
+	eor	x8,x8,x1,lsr#6	// sigma1(X[i+14])
+	add	x3,x3,x12
+	add	x24,x24,x20			// d+=h
+	add	x20,x20,x19			// h+=Maj(a,b,c)
+	ldr	x19,[x30],#8		// *K++, x28 in next round
+	add	x3,x3,x9
+	add	x20,x20,x17			// h+=Sigma0(a)
+	add	x3,x3,x8
+	cbnz	x19,.Loop_16_xx
+
+	ldp	x0,x2,[x29,#96]
+	ldr	x1,[x29,#112]
+	sub	x30,x30,#648		// rewind
+
+	ldp	x3,x4,[x0]
+	ldp	x5,x6,[x0,#2*8]
+	add	x1,x1,#14*8			// advance input pointer
+	ldp	x7,x8,[x0,#4*8]
+	add	x20,x20,x3
+	ldp	x9,x10,[x0,#6*8]
+	add	x21,x21,x4
+	add	x22,x22,x5
+	add	x23,x23,x6
+	stp	x20,x21,[x0]
+	add	x24,x24,x7
+	add	x25,x25,x8
+	stp	x22,x23,[x0,#2*8]
+	add	x26,x26,x9
+	add	x27,x27,x10
+	cmp	x1,x2
+	stp	x24,x25,[x0,#4*8]
+	stp	x26,x27,[x0,#6*8]
+	b.ne	.Loop
+
+	ldp	x19,x20,[x29,#16]
+	add	sp,sp,#4*8
+	ldp	x21,x22,[x29,#32]
+	ldp	x23,x24,[x29,#48]
+	ldp	x25,x26,[x29,#64]
+	ldp	x27,x28,[x29,#80]
+	ldp	x29,x30,[sp],#128
+	ret
+.size	sha512_block_data_order,.-sha512_block_data_order
+
+.align	6
+.type	.LK512,%object
+.LK512:
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+	.quad	0	// terminator
+.size	.LK512,.-.LK512
+#ifndef	__KERNEL__
+.align	3
+.LOPENSSL_armcap_P:
+# ifdef	__ILP32__
+	.long	OPENSSL_armcap_P-.
+# else
+	.quad	OPENSSL_armcap_P-.
+# endif
+#endif
+.asciz	"SHA512 block transform for ARMv8, CRYPTOGAMS by <appro@openssl.org>"
+.align	2
+#ifndef	__KERNEL__
+.comm	OPENSSL_armcap_P,4,4
+#endif
diff --git a/arch/arm64/kernel/bpi.S b/arch/arm64/kernel/bpi.S
index dec95bd82e31..c72f261f4b64 100644
--- a/arch/arm64/kernel/bpi.S
+++ b/arch/arm64/kernel/bpi.S
@@ -17,6 +17,7 @@
  */
 
 #include <linux/linkage.h>
+#include <linux/arm-smccc.h>
 
 .macro ventry target
 	.rept 31
@@ -77,3 +78,22 @@ ENTRY(__psci_hyp_bp_inval_start)
 	ldp	x0, x1, [sp, #(16 * 8)]
 	add	sp, sp, #(8 * 18)
 ENTRY(__psci_hyp_bp_inval_end)
+
+.macro smccc_workaround_1 inst
+	sub	sp, sp, #(8 * 4)
+	stp	x2, x3, [sp, #(8 * 0)]
+	stp	x0, x1, [sp, #(8 * 2)]
+	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_1
+	\inst	#0
+	ldp	x2, x3, [sp, #(8 * 0)]
+	ldp	x0, x1, [sp, #(8 * 2)]
+	add	sp, sp, #(8 * 4)
+.endm
+
+ENTRY(__smccc_workaround_1_smc_start)
+	smccc_workaround_1	smc
+ENTRY(__smccc_workaround_1_smc_end)
+
+ENTRY(__smccc_workaround_1_hvc_start)
+	smccc_workaround_1	hvc
+ENTRY(__smccc_workaround_1_hvc_end)
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index 67cdfc6a458e..a91239bd8696 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -54,6 +54,10 @@ DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
 
 #ifdef CONFIG_KVM
 extern char __psci_hyp_bp_inval_start[], __psci_hyp_bp_inval_end[];
+extern char __smccc_workaround_1_smc_start[];
+extern char __smccc_workaround_1_smc_end[];
+extern char __smccc_workaround_1_hvc_start[];
+extern char __smccc_workaround_1_hvc_end[];
 
 static void __copy_hyp_vect_bpi(int slot, const char *hyp_vecs_start,
 				const char *hyp_vecs_end)
@@ -96,8 +100,12 @@ static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
 	spin_unlock(&bp_lock);
 }
 #else
-#define __psci_hyp_bp_inval_start	NULL
-#define __psci_hyp_bp_inval_end		NULL
+#define __psci_hyp_bp_inval_start		NULL
+#define __psci_hyp_bp_inval_end			NULL
+#define __smccc_workaround_1_smc_start		NULL
+#define __smccc_workaround_1_smc_end		NULL
+#define __smccc_workaround_1_hvc_start		NULL
+#define __smccc_workaround_1_hvc_end		NULL
 
 static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
 				      const char *hyp_vecs_start,
@@ -124,17 +132,75 @@ static void  install_bp_hardening_cb(const struct arm64_cpu_capabilities *entry,
 	__install_bp_hardening_cb(fn, hyp_vecs_start, hyp_vecs_end);
 }
 
+#include <uapi/linux/psci.h>
+#include <linux/arm-smccc.h>
 #include <linux/psci.h>
 
+static void call_smc_arch_workaround_1(void)
+{
+	arm_smccc_1_1_smc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
+}
+
+static void call_hvc_arch_workaround_1(void)
+{
+	arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
+}
+
+static bool check_smccc_arch_workaround_1(const struct arm64_cpu_capabilities *entry)
+{
+	bp_hardening_cb_t cb;
+	void *smccc_start, *smccc_end;
+	struct arm_smccc_res res;
+
+	if (!entry->matches(entry, SCOPE_LOCAL_CPU))
+		return false;
+
+	if (psci_ops.smccc_version == SMCCC_VERSION_1_0)
+		return false;
+
+	switch (psci_ops.conduit) {
+	case PSCI_CONDUIT_HVC:
+		arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
+				  ARM_SMCCC_ARCH_WORKAROUND_1, &res);
+		if (res.a0)
+			return false;
+		cb = call_hvc_arch_workaround_1;
+		smccc_start = __smccc_workaround_1_hvc_start;
+		smccc_end = __smccc_workaround_1_hvc_end;
+		break;
+
+	case PSCI_CONDUIT_SMC:
+		arm_smccc_1_1_smc(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
+				  ARM_SMCCC_ARCH_WORKAROUND_1, &res);
+		if (res.a0)
+			return false;
+		cb = call_smc_arch_workaround_1;
+		smccc_start = __smccc_workaround_1_smc_start;
+		smccc_end = __smccc_workaround_1_smc_end;
+		break;
+
+	default:
+		return false;
+	}
+
+	install_bp_hardening_cb(entry, cb, smccc_start, smccc_end);
+
+	return true;
+}
+
 static int enable_psci_bp_hardening(void *data)
 {
 	const struct arm64_cpu_capabilities *entry = data;
 
-	if (psci_ops.get_version)
+	if (psci_ops.get_version) {
+		if (check_smccc_arch_workaround_1(entry))
+			return 0;
+
 		install_bp_hardening_cb(entry,
 				       (bp_hardening_cb_t)psci_ops.get_version,
 				       __psci_hyp_bp_inval_start,
 				       __psci_hyp_bp_inval_end);
+	}
 
 	return 0;
 }
-- 
2.11.0

[PATCH v4.9.y 42/42] arm64: Kill PSCI_GET_VERSION as a variant-2 workaround

[Mark Rutland]

From: Marc Zyngier <marc.zyngier@arm.com>

commit 3a0a397ff5ff8b56ca9f7908b75dee6bf0b5fabb upstream.

Now that we've standardised on SMCCC v1.1 to perform the branch
prediction invalidation, let's drop the previous band-aid.
If vendors haven't updated their firmware to do SMCCC 1.1, they
haven't updated PSCI either, so we don't loose anything.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com> [v4.9 backport]
---
 arch/arm64/kernel/bpi.S        | 24 ----------------------
 arch/arm64/kernel/cpu_errata.c | 45 ++++++++++++------------------------------
 arch/arm64/kvm/hyp/switch.c    | 14 -------------
 3 files changed, 13 insertions(+), 70 deletions(-)

diff --git a/arch/arm64/kernel/bpi.S b/arch/arm64/kernel/bpi.S
index c72f261f4b64..dc4eb154e33b 100644
--- a/arch/arm64/kernel/bpi.S
+++ b/arch/arm64/kernel/bpi.S
@@ -54,30 +54,6 @@ ENTRY(__bp_harden_hyp_vecs_start)
 	vectors __kvm_hyp_vector
 	.endr
 ENTRY(__bp_harden_hyp_vecs_end)
-ENTRY(__psci_hyp_bp_inval_start)
-	sub	sp, sp, #(8 * 18)
-	stp	x16, x17, [sp, #(16 * 0)]
-	stp	x14, x15, [sp, #(16 * 1)]
-	stp	x12, x13, [sp, #(16 * 2)]
-	stp	x10, x11, [sp, #(16 * 3)]
-	stp	x8, x9, [sp, #(16 * 4)]
-	stp	x6, x7, [sp, #(16 * 5)]
-	stp	x4, x5, [sp, #(16 * 6)]
-	stp	x2, x3, [sp, #(16 * 7)]
-	stp	x0, x1, [sp, #(16 * 8)]
-	mov	x0, #0x84000000
-	smc	#0
-	ldp	x16, x17, [sp, #(16 * 0)]
-	ldp	x14, x15, [sp, #(16 * 1)]
-	ldp	x12, x13, [sp, #(16 * 2)]
-	ldp	x10, x11, [sp, #(16 * 3)]
-	ldp	x8, x9, [sp, #(16 * 4)]
-	ldp	x6, x7, [sp, #(16 * 5)]
-	ldp	x4, x5, [sp, #(16 * 6)]
-	ldp	x2, x3, [sp, #(16 * 7)]
-	ldp	x0, x1, [sp, #(16 * 8)]
-	add	sp, sp, #(8 * 18)
-ENTRY(__psci_hyp_bp_inval_end)
 
 .macro smccc_workaround_1 inst
 	sub	sp, sp, #(8 * 4)
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index a91239bd8696..74107134cc30 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -53,7 +53,6 @@ static int cpu_enable_trap_ctr_access(void *__unused)
 DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
 
 #ifdef CONFIG_KVM
-extern char __psci_hyp_bp_inval_start[], __psci_hyp_bp_inval_end[];
 extern char __smccc_workaround_1_smc_start[];
 extern char __smccc_workaround_1_smc_end[];
 extern char __smccc_workaround_1_hvc_start[];
@@ -100,8 +99,6 @@ static void __install_bp_hardening_cb(bp_hardening_cb_t fn,
 	spin_unlock(&bp_lock);
 }
 #else
-#define __psci_hyp_bp_inval_start		NULL
-#define __psci_hyp_bp_inval_end			NULL
 #define __smccc_workaround_1_smc_start		NULL
 #define __smccc_workaround_1_smc_end		NULL
 #define __smccc_workaround_1_hvc_start		NULL
@@ -146,24 +143,25 @@ static void call_hvc_arch_workaround_1(void)
 	arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
 }
 
-static bool check_smccc_arch_workaround_1(const struct arm64_cpu_capabilities *entry)
+static int enable_smccc_arch_workaround_1(void *data)
 {
+	const struct arm64_cpu_capabilities *entry = data;
 	bp_hardening_cb_t cb;
 	void *smccc_start, *smccc_end;
 	struct arm_smccc_res res;
 
 	if (!entry->matches(entry, SCOPE_LOCAL_CPU))
-		return false;
+		return 0;
 
 	if (psci_ops.smccc_version == SMCCC_VERSION_1_0)
-		return false;
+		return 0;
 
 	switch (psci_ops.conduit) {
 	case PSCI_CONDUIT_HVC:
 		arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
 				  ARM_SMCCC_ARCH_WORKAROUND_1, &res);
 		if (res.a0)
-			return false;
+			return 0;
 		cb = call_hvc_arch_workaround_1;
 		smccc_start = __smccc_workaround_1_hvc_start;
 		smccc_end = __smccc_workaround_1_hvc_end;
@@ -173,35 +171,18 @@ static bool check_smccc_arch_workaround_1(const struct arm64_cpu_capabilities *e
 		arm_smccc_1_1_smc(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
 				  ARM_SMCCC_ARCH_WORKAROUND_1, &res);
 		if (res.a0)
-			return false;
+			return 0;
 		cb = call_smc_arch_workaround_1;
 		smccc_start = __smccc_workaround_1_smc_start;
 		smccc_end = __smccc_workaround_1_smc_end;
 		break;
 
 	default:
-		return false;
+		return 0;
 	}
 
 	install_bp_hardening_cb(entry, cb, smccc_start, smccc_end);
 
-	return true;
-}
-
-static int enable_psci_bp_hardening(void *data)
-{
-	const struct arm64_cpu_capabilities *entry = data;
-
-	if (psci_ops.get_version) {
-		if (check_smccc_arch_workaround_1(entry))
-			return 0;
-
-		install_bp_hardening_cb(entry,
-				       (bp_hardening_cb_t)psci_ops.get_version,
-				       __psci_hyp_bp_inval_start,
-				       __psci_hyp_bp_inval_end);
-	}
-
 	return 0;
 }
 #endif	/* CONFIG_HARDEN_BRANCH_PREDICTOR */
@@ -301,32 +282,32 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A57),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A72),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A73),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_CORTEX_A75),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_BRCM_VULCAN),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 	{
 		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
 		MIDR_ALL_VERSIONS(MIDR_CAVIUM_THUNDERX2),
-		.enable = enable_psci_bp_hardening,
+		.enable = enable_smccc_arch_workaround_1,
 	},
 #endif
 	{
diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
index 996328edd233..154b47139362 100644
--- a/arch/arm64/kvm/hyp/switch.c
+++ b/arch/arm64/kvm/hyp/switch.c
@@ -311,20 +311,6 @@ int __hyp_text __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 	if (exit_code == ARM_EXCEPTION_TRAP && !__populate_fault_info(vcpu))
 		goto again;
 
-	if (exit_code == ARM_EXCEPTION_TRAP &&
-	    (kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC64 ||
-	     kvm_vcpu_trap_get_class(vcpu) == ESR_ELx_EC_HVC32)) {
-		u32 val = vcpu_get_reg(vcpu, 0);
-
-		if (val == PSCI_0_2_FN_PSCI_VERSION) {
-			val = kvm_psci_version(vcpu, kern_hyp_va(vcpu->kvm));
-			if (unlikely(val == KVM_ARM_PSCI_0_1))
-				val = PSCI_RET_NOT_SUPPORTED;
-			vcpu_set_reg(vcpu, 0, val);
-			goto again;
-		}
-	}
-
 	if (static_branch_unlikely(&vgic_v2_cpuif_trap) &&
 	    exit_code == ARM_EXCEPTION_TRAP) {
 		bool valid;
-- 
2.11.0

Re: [PATCH v4.9.y 00/42] arm64 spectre patches

[Greg Hackmann]

On 04/12/2018 04:10 AM, Mark Rutland wrote:
> Hi Greg,
> 
> These patches backport arm64 spectre mitigations to v4.9.y (based on 4.9.93).
> 
> I've done this backport from scratch, following the patches in the arm64 kpti
> branch [1] (which contrary to its name, also has spectre patches intended for
> backporting).
> 
> To the best of my understanding the code is correct, in the context of the
> v4.9.y kernel, and I've tested the series on arm64 hardware available to me,
> using firmware with SMCCC 1.1 and branch predictor mitigations.
> 
> I've Cc'd the usual suspects for testing. For anyone willing to test this, I've
> pushed this out to the stable/4.9.y/spectre branch of my kernel.org repo [2].
> 
> Note that I've left off the Qualcomm Falkor patches, as it was pointed out that
> necessary errata workarounds for Falkor were never backported, and thus I
> assume that no-one is using v4.9.y on Falkor parts. Shanker, do you have any
> idea whether people are using v4.9.y on Falkor?
> 
> Thanks,
> Mark.

I've run this series on hikey on top of vanilla 4.9.93 (boots AOSP to
serial shell) and android-linaro-hikey-4.9 (boots AOSP to home screen),
with both SMCCC 1.0 and SMCCC 1.1 firmware.

Tested-by: Greg Hackmann <ghackmann@google.com>

Re: [PATCH v4.9.y 00/42] arm64 spectre patches

[Greg KH]

On Thu, Apr 12, 2018 at 12:10:56PM +0100, Mark Rutland wrote:
> Hi Greg,
> 
> These patches backport arm64 spectre mitigations to v4.9.y (based on 4.9.93).
> 
> I've done this backport from scratch, following the patches in the arm64 kpti
> branch [1] (which contrary to its name, also has spectre patches intended for
> backporting).
> 
> To the best of my understanding the code is correct, in the context of the
> v4.9.y kernel, and I've tested the series on arm64 hardware available to me,
> using firmware with SMCCC 1.1 and branch predictor mitigations.
> 
> I've Cc'd the usual suspects for testing. For anyone willing to test this, I've
> pushed this out to the stable/4.9.y/spectre branch of my kernel.org repo [2].
> 
> Note that I've left off the Qualcomm Falkor patches, as it was pointed out that
> necessary errata workarounds for Falkor were never backported, and thus I
> assume that no-one is using v4.9.y on Falkor parts. Shanker, do you have any
> idea whether people are using v4.9.y on Falkor?

I've taken all of these now except the futex patch which did not apply.

thanks,

greg k-h