Merge branch 'master' into openslxopenslx

9 files changed, 204 insertions, 287 deletions

diff --git a/src/arch/arm64/Makefile b/src/arch/arm64/Makefile
index 9b9dd5ec8..719a1e61c 100644
--- a/src/arch/arm64/Makefile
+++ b/src/arch/arm64/Makefile
@@ -1,3 +1,7 @@
+# Specify compressor
+#
+ZBIN		= $(ZBIN64)
+
 # ARM64-specific directories containing source files
 #
 SRCDIRS		+= arch/arm64/core
diff --git a/src/arch/arm64/Makefile.efi b/src/arch/arm64/Makefile.efi
index 998a64d03..96f2953f4 100644
--- a/src/arch/arm64/Makefile.efi
+++ b/src/arch/arm64/Makefile.efi
@@ -4,10 +4,6 @@
 #
 ELF2EFI		= $(ELF2EFI64)
 
-# Specify EFI boot file
-#
-EFI_BOOT_FILE	= bootaa64.efi
-
 # Include generic EFI Makefile
 #
 MAKEDEPS	+= arch/arm/Makefile.efi
diff --git a/src/arch/arm64/core/arm64_bigint.c b/src/arch/arm64/core/arm64_bigint.c
deleted file mode 100644
index 7740f1aef..000000000
--- a/src/arch/arm64/core/arm64_bigint.c
+++ /dev/null
@@ -1,107 +0,0 @@
-/*
- * Copyright (C) 2016 Michael Brown <mbrown@fensystems.co.uk>.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation; either version 2 of the
- * License, or any later version.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
- * 02110-1301, USA.
- *
- * You can also choose to distribute this program under the terms of
- * the Unmodified Binary Distribution Licence (as given in the file
- * COPYING.UBDL), provided that you have satisfied its requirements.
- */
-
-FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
-
-#include <stdint.h>
-#include <string.h>
-#include <ipxe/bigint.h>
-
-/** @file
- *
- * Big integer support
- */
-
-/**
- * Multiply big integers
- *
- * @v multiplicand0	Element 0 of big integer to be multiplied
- * @v multiplicand_size	Number of elements in multiplicand
- * @v multiplier0	Element 0 of big integer to be multiplied
- * @v multiplier_size	Number of elements in multiplier
- * @v result0		Element 0 of big integer to hold result
- */
-void bigint_multiply_raw ( const uint64_t *multiplicand0,
-			   unsigned int multiplicand_size,
-			   const uint64_t *multiplier0,
-			   unsigned int multiplier_size,
-			   uint64_t *result0 ) {
-	unsigned int result_size = ( multiplicand_size + multiplier_size );
-	const bigint_t ( multiplicand_size ) __attribute__ (( may_alias ))
-		*multiplicand = ( ( const void * ) multiplicand0 );
-	const bigint_t ( multiplier_size ) __attribute__ (( may_alias ))
-		*multiplier = ( ( const void * ) multiplier0 );
-	bigint_t ( result_size ) __attribute__ (( may_alias ))
-		*result = ( ( void * ) result0 );
-	unsigned int i;
-	unsigned int j;
-	uint64_t multiplicand_element;
-	uint64_t multiplier_element;
-	uint64_t *result_elements;
-	uint64_t discard_low;
-	uint64_t discard_high;
-	uint64_t discard_temp_low;
-	uint64_t discard_temp_high;
-
-	/* Zero result */
-	memset ( result, 0, sizeof ( *result ) );
-
-	/* Multiply integers one element at a time */
-	for ( i = 0 ; i < multiplicand_size ; i++ ) {
-		multiplicand_element = multiplicand->element[i];
-		for ( j = 0 ; j < multiplier_size ; j++ ) {
-			multiplier_element = multiplier->element[j];
-			result_elements = &result->element[ i + j ];
-			/* Perform a single multiply, and add the
-			 * resulting double-element into the result,
-			 * carrying as necessary.  The carry can
-			 * never overflow beyond the end of the
-			 * result, since:
-			 *
-			 *     a < 2^{n}, b < 2^{m} => ab < 2^{n+m}
-			 */
-			__asm__ __volatile__ ( "mul %1, %6, %7\n\t"
-					       "umulh %2, %6, %7\n\t"
-					       "ldp %3, %4, [%0]\n\t"
-					       "adds %3, %3, %1\n\t"
-					       "adcs %4, %4, %2\n\t"
-					       "stp %3, %4, [%0], #16\n\t"
-					       "bcc 2f\n\t"
-					       "\n1:\n\t"
-					       "ldr %3, [%0]\n\t"
-					       "adcs %3, %3, xzr\n\t"
-					       "str %3, [%0], #8\n\t"
-					       "bcs 1b\n\t"
-					       "\n2:\n\t"
-					       : "+r" ( result_elements ),
-						 "=&r" ( discard_low ),
-						 "=&r" ( discard_high ),
-						 "=r" ( discard_temp_low ),
-						 "=r" ( discard_temp_high ),
-						 "+m" ( *result )
-					       : "r" ( multiplicand_element ),
-						 "r" ( multiplier_element )
-					       : "cc" );
-		}
-	}
-}
diff --git a/src/arch/arm64/core/arm64_string.c b/src/arch/arm64/core/arm64_string.c
index 28a2b73bc..07a7eefdf 100644
--- a/src/arch/arm64/core/arm64_string.c
+++ b/src/arch/arm64/core/arm64_string.c
@@ -31,6 +31,9 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
 
 #include <string.h>
 
+/** Block size (for "ldp"/"stp") */
+#define ARM64_STRING_BLKSZ 16
+
 /**
  * Copy memory area
  *
@@ -40,59 +43,70 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
  * @ret dest		Destination address
  */
 void arm64_memcpy ( void *dest, const void *src, size_t len ) {
-	void *discard_dest;
-	void *discard_end;
-	const void *discard_src;
-	size_t discard_offset;
+	size_t len_pre;
+	size_t len_mid;
+	size_t len_post;
 	unsigned long discard_data;
 	unsigned long discard_low;
 	unsigned long discard_high;
+	unsigned long discard_len;
 
-	/* If length is too short for an "ldp"/"stp" instruction pair,
-	 * then just copy individual bytes.
+	/* Calculate pre-aligned, aligned, and post-aligned lengths.
+	 * (Align on the destination address, on the assumption that
+	 * misaligned stores are likely to be more expensive than
+	 * misaligned loads.)
 	 */
-	if ( len < 16 ) {
-		__asm__ __volatile__ ( "cbz %0, 2f\n\t"
-				       "\n1:\n\t"
-				       "sub %0, %0, #1\n\t"
-				       "ldrb %w1, [%3, %0]\n\t"
-				       "strb %w1, [%2, %0]\n\t"
-				       "cbnz %0, 1b\n\t"
-				       "\n2:\n\t"
-				       : "=&r" ( discard_offset ),
-					 "=&r" ( discard_data )
-				       : "r" ( dest ), "r" ( src ), "0" ( len )
-				       : "memory" );
-		return;
-	}
+	len_pre = ( ( ARM64_STRING_BLKSZ - ( ( intptr_t ) dest ) ) &
+		    ( ARM64_STRING_BLKSZ - 1 ) );
+	if ( len_pre > len )
+		len_pre = len;
+	len -= len_pre;
+	len_mid = ( len & ~( ARM64_STRING_BLKSZ - 1 ) );
+	len -= len_mid;
+	len_post = len;
 
-	/* Use "ldp"/"stp" to copy 16 bytes at a time: one initial
-	 * potentially unaligned access, multiple destination-aligned
-	 * accesses, one final potentially unaligned access.
-	 */
-	__asm__ __volatile__ ( "ldp %3, %4, [%1], #16\n\t"
-			       "stp %3, %4, [%0], #16\n\t"
-			       "and %3, %0, #15\n\t"
-			       "sub %0, %0, %3\n\t"
-			       "sub %1, %1, %3\n\t"
-			       "bic %2, %5, #15\n\t"
-			       "b 2f\n\t"
+	/* Copy pre-aligned section */
+	__asm__ __volatile__ ( "cbz %2, 2f\n\t"
+			       "\n1:\n\t"
+			       "ldrb %w3, [%1], #1\n\t"
+			       "strb %w3, [%0], #1\n\t"
+			       "sub %2, %2, #1\n\t"
+			       "cbnz %2, 1b\n\t"
+			       "\n2:\n\t"
+			       : "+r" ( dest ), "+r" ( src ),
+				 "=&r" ( discard_len ),
+				 "=&r" ( discard_data )
+			       : "2" ( len_pre )
+			       : "memory" );
+
+	/* Copy aligned section */
+	__asm__ __volatile__ ( "cbz %2, 2f\n\t"
 			       "\n1:\n\t"
 			       "ldp %3, %4, [%1], #16\n\t"
 			       "stp %3, %4, [%0], #16\n\t"
+			       "sub %2, %2, #16\n\t"
+			       "cbnz %2, 1b\n\t"
 			       "\n2:\n\t"
-			       "cmp %0, %2\n\t"
-			       "bne 1b\n\t"
-			       "ldp %3, %4, [%6, #-16]\n\t"
-			       "stp %3, %4, [%5, #-16]\n\t"
-			       : "=&r" ( discard_dest ),
-				 "=&r" ( discard_src ),
-				 "=&r" ( discard_end ),
+			       : "+r" ( dest ), "+r" ( src ),
+				 "=&r" ( discard_len ),
 				 "=&r" ( discard_low ),
 				 "=&r" ( discard_high )
-			       : "r" ( dest + len ), "r" ( src + len ),
-				 "0" ( dest ), "1" ( src )
-			       : "memory", "cc" );
+			       : "2" ( len_mid )
+			       : "memory" );
+
+	/* Copy post-aligned section */
+	__asm__ __volatile__ ( "cbz %2, 2f\n\t"
+			       "\n1:\n\t"
+			       "ldrb %w3, [%1], #1\n\t"
+			       "strb %w3, [%0], #1\n\t"
+			       "sub %2, %2, #1\n\t"
+			       "cbnz %2, 1b\n\t"
+			       "\n2:\n\t"
+			       : "+r" ( dest ), "+r" ( src ),
+				 "=&r" ( discard_len ),
+				 "=&r" ( discard_data )
+			       : "2" ( len_post )
+			       : "memory" );
 }
 
 /**
@@ -102,44 +116,56 @@ void arm64_memcpy ( void *dest, const void *src, size_t len ) {
  * @v len		Length
  */
 void arm64_bzero ( void *dest, size_t len ) {
-	size_t discard_offset;
-	void *discard_dest;
-	void *discard_end;
+	size_t len_pre;
+	size_t len_mid;
+	size_t len_post;
+	unsigned long discard_len;
 
-	/* If length is too short for an "stp" instruction, then just
-	 * zero individual bytes.
-	 */
-	if ( len < 16 ) {
-		__asm__ __volatile__ ( "cbz %0, 2f\n\t"
-				       "\n1:\n\t"
-				       "sub %0, %0, #1\n\t"
-				       "strb wzr, [%1, %0]\n\t"
-				       "cbnz %0, 1b\n\t"
-				       "\n2:\n\t"
-				       : "=&r" ( discard_offset )
-				       : "r" ( dest ), "0" ( len )
-				       : "memory" );
-		return;
-	}
+	/* Calculate pre-aligned, aligned, and post-aligned lengths */
+	len_pre = ( ( ARM64_STRING_BLKSZ - ( ( intptr_t ) dest ) ) &
+		    ( ARM64_STRING_BLKSZ - 1 ) );
+	if ( len_pre > len )
+		len_pre = len;
+	len -= len_pre;
+	len_mid = ( len & ~( ARM64_STRING_BLKSZ - 1 ) );
+	len -= len_mid;
+	len_post = len;
 
-	/* Use "stp" to zero 16 bytes at a time: one initial
-	 * potentially unaligned access, multiple aligned accesses,
-	 * one final potentially unaligned access.
-	 */
-	__asm__ __volatile__ ( "stp xzr, xzr, [%0], #16\n\t"
-			       "bic %0, %0, #15\n\t"
-			       "bic %1, %2, #15\n\t"
-			       "b 2f\n\t"
+	/* Zero pre-aligned section */
+	__asm__ __volatile__ ( "cbz %1, 2f\n\t"
+			       "\n1:\n\t"
+			       "strb wzr, [%0], #1\n\t"
+			       "sub %1, %1, #1\n\t"
+			       "cbnz %1, 1b\n\t"
+			       "\n2:\n\t"
+			       : "+r" ( dest ),
+				 "=&r" ( discard_len )
+			       : "1" ( len_pre )
+			       : "memory" );
+
+	/* Zero aligned section */
+	__asm__ __volatile__ ( "cbz %1, 2f\n\t"
 			       "\n1:\n\t"
 			       "stp xzr, xzr, [%0], #16\n\t"
+			       "sub %1, %1, #16\n\t"
+			       "cbnz %1, 1b\n\t"
 			       "\n2:\n\t"
-			       "cmp %0, %1\n\t"
-			       "bne 1b\n\t"
-			       "stp xzr, xzr, [%2, #-16]\n\t"
-			       : "=&r" ( discard_dest ),
-				 "=&r" ( discard_end )
-			       : "r" ( dest + len ), "0" ( dest )
-			       : "memory", "cc" );
+			       : "+r" ( dest ),
+				 "=&r" ( discard_len )
+			       : "1" ( len_mid )
+			       : "memory" );
+
+	/* Zero post-aligned section */
+	__asm__ __volatile__ ( "cbz %1, 2f\n\t"
+			       "\n1:\n\t"
+			       "strb wzr, [%0], #1\n\t"
+			       "sub %1, %1, #1\n\t"
+			       "cbnz %1, 1b\n\t"
+			       "\n2:\n\t"
+			       : "+r" ( dest ),
+				 "=&r" ( discard_len )
+			       : "1" ( len_post )
+			       : "memory" );
 }
 
 /**
diff --git a/src/arch/arm64/include/bits/bigint.h b/src/arch/arm64/include/bits/bigint.h
index 0d08bbd65..f4032e335 100644
--- a/src/arch/arm64/include/bits/bigint.h
+++ b/src/arch/arm64/include/bits/bigint.h
@@ -43,8 +43,9 @@ bigint_init_raw ( uint64_t *value0, unsigned int size,
  * @v addend0		Element 0 of big integer to add
  * @v value0		Element 0 of big integer to be added to
  * @v size		Number of elements
+ * @ret carry		Carry out
  */
-static inline __attribute__ (( always_inline )) void
+static inline __attribute__ (( always_inline )) int
 bigint_add_raw ( const uint64_t *addend0, uint64_t *value0,
 		 unsigned int size ) {
 	bigint_t ( size ) __attribute__ (( may_alias )) *value =
@@ -54,6 +55,7 @@ bigint_add_raw ( const uint64_t *addend0, uint64_t *value0,
 	uint64_t discard_addend_i;
 	uint64_t discard_value_i;
 	unsigned int discard_size;
+	int carry;
 
 	__asm__ __volatile__ ( "cmn xzr, xzr\n\t" /* clear CF */
 			       "\n1:\n\t"
@@ -68,9 +70,11 @@ bigint_add_raw ( const uint64_t *addend0, uint64_t *value0,
 				 "=r" ( discard_size ),
 				 "=r" ( discard_addend_i ),
 				 "=r" ( discard_value_i ),
+				 "=@cccs" ( carry ),
 				 "+m" ( *value )
-			       : "0" ( addend0 ), "1" ( value0 ), "2" ( size )
-			       : "cc" );
+			       : "0" ( addend0 ), "1" ( value0 ),
+				 "2" ( size ) );
+	return carry;
 }
 
 /**
@@ -79,8 +83,9 @@ bigint_add_raw ( const uint64_t *addend0, uint64_t *value0,
  * @v subtrahend0	Element 0 of big integer to subtract
  * @v value0		Element 0 of big integer to be subtracted from
  * @v size		Number of elements
+ * @ret borrow		Borrow out
  */
-static inline __attribute__ (( always_inline )) void
+static inline __attribute__ (( always_inline )) int
 bigint_subtract_raw ( const uint64_t *subtrahend0, uint64_t *value0,
 		      unsigned int size ) {
 	bigint_t ( size ) __attribute__ (( may_alias )) *value =
@@ -90,6 +95,7 @@ bigint_subtract_raw ( const uint64_t *subtrahend0, uint64_t *value0,
 	uint64_t discard_subtrahend_i;
 	uint64_t discard_value_i;
 	unsigned int discard_size;
+	int borrow;
 
 	__asm__ __volatile__ ( "cmp xzr, xzr\n\t" /* set CF */
 			       "\n1:\n\t"
@@ -104,25 +110,28 @@ bigint_subtract_raw ( const uint64_t *subtrahend0, uint64_t *value0,
 				 "=r" ( discard_size ),
 				 "=r" ( discard_subtrahend_i ),
 				 "=r" ( discard_value_i ),
+				 "=@cccc" ( borrow ),
 				 "+m" ( *value )
 			       : "0" ( subtrahend0 ), "1" ( value0 ),
-				 "2" ( size )
-			       : "cc" );
+				 "2" ( size ) );
+	return borrow;
 }
 
 /**
- * Rotate big integer left
+ * Shift big integer left
  *
  * @v value0		Element 0 of big integer
  * @v size		Number of elements
+ * @ret out		Bit shifted out
  */
-static inline __attribute__ (( always_inline )) void
-bigint_rol_raw ( uint64_t *value0, unsigned int size ) {
+static inline __attribute__ (( always_inline )) int
+bigint_shl_raw ( uint64_t *value0, unsigned int size ) {
 	bigint_t ( size ) __attribute__ (( may_alias )) *value =
 		( ( void * ) value0 );
 	uint64_t *discard_value;
 	uint64_t discard_value_i;
 	unsigned int discard_size;
+	int carry;
 
 	__asm__ __volatile__ ( "cmn xzr, xzr\n\t" /* clear CF */
 			       "\n1:\n\t"
@@ -134,40 +143,43 @@ bigint_rol_raw ( uint64_t *value0, unsigned int size ) {
 			       : "=r" ( discard_value ),
 				 "=r" ( discard_size ),
 				 "=r" ( discard_value_i ),
+				 "=@cccs" ( carry ),
 				 "+m" ( *value )
-			       : "0" ( value0 ), "1" ( size )
-			       : "cc" );
+			       : "0" ( value0 ), "1" ( size ) );
+	return carry;
 }
 
 /**
- * Rotate big integer right
+ * Shift big integer right
  *
  * @v value0		Element 0 of big integer
  * @v size		Number of elements
+ * @ret out		Bit shifted out
  */
-static inline __attribute__ (( always_inline )) void
-bigint_ror_raw ( uint64_t *value0, unsigned int size ) {
+static inline __attribute__ (( always_inline )) int
+bigint_shr_raw ( uint64_t *value0, unsigned int size ) {
 	bigint_t ( size ) __attribute__ (( may_alias )) *value =
 		( ( void * ) value0 );
 	uint64_t *discard_value;
-	uint64_t discard_value_i;
-	uint64_t discard_value_j;
+	uint64_t discard_high;
 	unsigned int discard_size;
+	uint64_t low;
 
-	__asm__ __volatile__ ( "mov %3, #0\n\t"
+	__asm__ __volatile__ ( "mov %2, #0\n\t"
 			       "\n1:\n\t"
 			       "sub %w1, %w1, #1\n\t"
-			       "ldr %2, [%0, %1, lsl #3]\n\t"
-			       "extr %3, %3, %2, #1\n\t"
-			       "str %3, [%0, %1, lsl #3]\n\t"
-			       "mov %3, %2\n\t"
+			       "ldr %3, [%0, %1, lsl #3]\n\t"
+			       "extr %2, %2, %3, #1\n\t"
+			       "str %2, [%0, %1, lsl #3]\n\t"
+			       "mov %2, %3\n\t"
 			       "cbnz %w1, 1b\n\t"
 			       : "=r" ( discard_value ),
 				 "=r" ( discard_size ),
-				 "=r" ( discard_value_i ),
-				 "=r" ( discard_value_j ),
+				 "=r" ( discard_high ),
+				 "=r" ( low ),
 				 "+m" ( *value )
 			       : "0" ( value0 ), "1" ( size ) );
+	return ( low & 1 );
 }
 
 /**
@@ -218,25 +230,6 @@ bigint_is_geq_raw ( const uint64_t *value0, const uint64_t *reference0,
 }
 
 /**
- * Test if bit is set in big integer
- *
- * @v value0		Element 0 of big integer
- * @v size		Number of elements
- * @v bit		Bit to test
- * @ret is_set		Bit is set
- */
-static inline __attribute__ (( always_inline )) int
-bigint_bit_is_set_raw ( const uint64_t *value0, unsigned int size,
-			unsigned int bit ) {
-	const bigint_t ( size ) __attribute__ (( may_alias )) *value =
-		( ( const void * ) value0 );
-	unsigned int index = ( bit / ( 8 * sizeof ( value->element[0] ) ) );
-	unsigned int subindex = ( bit % ( 8 * sizeof ( value->element[0] ) ) );
-
-	return ( !! ( value->element[index] & ( 1UL << subindex ) ) );
-}
-
-/**
  * Find highest bit set in big integer
  *
  * @v value0		Element 0 of big integer
@@ -310,10 +303,36 @@ bigint_done_raw ( const uint64_t *value0, unsigned int size __unused,
 		*(--out_byte) = *(value_byte++);
 }
 
-extern void bigint_multiply_raw ( const uint64_t *multiplicand0,
-				  unsigned int multiplicand_size,
-				  const uint64_t *multiplier0,
-				  unsigned int multiplier_size,
-				  uint64_t *value0 );
+/**
+ * Multiply big integer elements
+ *
+ * @v multiplicand	Multiplicand element
+ * @v multiplier	Multiplier element
+ * @v result		Result element
+ * @v carry		Carry element
+ */
+static inline __attribute__ (( always_inline )) void
+bigint_multiply_one ( const uint64_t multiplicand, const uint64_t multiplier,
+		      uint64_t *result, uint64_t *carry ) {
+	uint64_t discard_low;
+	uint64_t discard_high;
+
+	__asm__ __volatile__ ( /* Perform multiplication */
+			       "mul %0, %4, %5\n\t"
+			       "umulh %1, %4, %5\n\t"
+			       /* Accumulate result */
+			       "adds %2, %2, %0\n\t"
+			       "adc %1, %1, xzr\n\t"
+			       /* Accumulate carry (cannot overflow) */
+			       "adds %2, %2, %3\n\t"
+			       "adc %3, %1, xzr\n\t"
+			       : "=&r" ( discard_low ),
+				 "=r" ( discard_high ),
+				 "+r" ( *result ),
+				 "+r" ( *carry )
+			       : "r" ( multiplicand ),
+				 "r" ( multiplier )
+			       : "cc" );
+}
 
 #endif /* _BITS_BIGINT_H */
diff --git a/src/arch/arm64/include/bits/lkrn.h b/src/arch/arm64/include/bits/lkrn.h
new file mode 100644
index 000000000..943464e9b
--- /dev/null
+++ b/src/arch/arm64/include/bits/lkrn.h
@@ -0,0 +1,30 @@
+#ifndef _BITS_LKRN_H
+#define _BITS_LKRN_H
+
+/** @file
+ *
+ * Linux kernel image invocation
+ *
+ */
+
+FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
+
+/** Header magic value */
+#define LKRN_MAGIC_ARCH LKRN_MAGIC_AARCH64
+
+/**
+ * Jump to kernel entry point
+ *
+ * @v entry		Kernel entry point
+ * @v fdt		Device tree
+ */
+static inline __attribute__ (( noreturn )) void
+lkrn_jump ( physaddr_t entry, physaddr_t fdt ) {
+	register unsigned long x0 asm ( "x0" ) = fdt;
+
+	__asm__ __volatile__ ( "br %1"
+			       : : "r" ( x0 ), "r" ( entry ) );
+	__builtin_unreachable();
+}
+
+#endif /* _BITS_LKRN_H */
diff --git a/src/arch/arm64/include/bits/profile.h b/src/arch/arm64/include/bits/profile.h
index 62ffa3772..4a5b3f7a1 100644
--- a/src/arch/arm64/include/bits/profile.h
+++ b/src/arch/arm64/include/bits/profile.h
@@ -16,7 +16,7 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
  *
  * @ret timestamp	Timestamp
  */
-static inline __attribute__ (( always_inline )) uint64_t
+static inline __attribute__ (( always_inline )) unsigned long
 profile_timestamp ( void ) {
 	uint64_t cycles;
 
diff --git a/src/arch/arm64/include/setjmp.h b/src/arch/arm64/include/bits/setjmp.h
index 85a7a9cad..6ffd2fb0a 100644
--- a/src/arch/arm64/include/setjmp.h
+++ b/src/arch/arm64/include/bits/setjmp.h
@@ -1,5 +1,5 @@
-#ifndef _SETJMP_H
-#define _SETJMP_H
+#ifndef _BITS_SETJMP_H
+#define _BITS_SETJMP_H
 
 FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
 
@@ -35,10 +35,4 @@ typedef struct {
 	uint64_t sp;
 } jmp_buf[1];
 
-extern int __asmcall __attribute__ (( returns_twice ))
-setjmp ( jmp_buf env );
-
-extern void __asmcall __attribute__ (( noreturn ))
-longjmp ( jmp_buf env, int val );
-
-#endif /* _SETJMP_H */
+#endif /* _BITS_SETJMP_H */
diff --git a/src/arch/arm64/include/gdbmach.h b/src/arch/arm64/include/gdbmach.h
deleted file mode 100644
index cd152eedd..000000000
--- a/src/arch/arm64/include/gdbmach.h
+++ /dev/null
@@ -1,45 +0,0 @@
-#ifndef GDBMACH_H
-#define GDBMACH_H
-
-/** @file
- *
- * GDB architecture specifics
- *
- * This file declares functions for manipulating the machine state and
- * debugging context.
- *
- */
-
-#include <stdint.h>
-
-typedef unsigned long gdbreg_t;
-
-/* Register snapshot */
-enum {
-	/* Not yet implemented */
-	GDBMACH_NREGS,
-};
-
-#define GDBMACH_SIZEOF_REGS ( GDBMACH_NREGS * sizeof ( gdbreg_t ) )
-
-static inline void gdbmach_set_pc ( gdbreg_t *regs, gdbreg_t pc ) {
-	/* Not yet implemented */
-	( void ) regs;
-	( void ) pc;
-}
-
-static inline void gdbmach_set_single_step ( gdbreg_t *regs, int step ) {
-	/* Not yet implemented */
-	( void ) regs;
-	( void ) step;
-}
-
-static inline void gdbmach_breakpoint ( void ) {
-	/* Not yet implemented */
-}
-
-extern int gdbmach_set_breakpoint ( int type, unsigned long addr, size_t len,
-				    int enable );
-extern void gdbmach_init ( void );
-
-#endif /* GDBMACH_H */