* [PATCH v2 1/7] xen: implement byteswap.h
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 14:00 ` Andrew Cooper
2021-11-02 14:23 ` Jan Beulich
2021-10-22 10:47 ` [PATCH v2 2/7] crypto/vmac: Simplify code with byteswap.h Lin Liu
` (6 subsequent siblings)
7 siblings, 2 replies; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel
Cc: Lin Liu, Andrew Cooper, George Dunlap, Ian Jackson, Jan Beulich,
Julien Grall, Stefano Stabellini, Wei Liu
swab() is massively over complicated and can be simplified by builtins.
The compilers provide builtin function to swap bytes.
* gcc: https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html
* clang: https://clang.llvm.org/docs/LanguageExtensions.html
This patch simplify swab() with builtins and fallback for old compilers.
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: George Dunlap <george.dunlap@citrix.com>
Cc: Ian Jackson <iwj@xenproject.org>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Julien Grall <julien@xen.org>
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Wei Liu <wl@xen.org>
Changes in v2:
- Add fallback for compilers without __builtin_bswap
- Implement with plain C instead of macros
---
xen/include/xen/byteswap.h | 93 ++++++++++++++++++++++++++++++++++++++
xen/include/xen/compiler.h | 12 +++++
2 files changed, 105 insertions(+)
create mode 100644 xen/include/xen/byteswap.h
diff --git a/xen/include/xen/byteswap.h b/xen/include/xen/byteswap.h
new file mode 100644
index 0000000000..848a4bbaee
--- /dev/null
+++ b/xen/include/xen/byteswap.h
@@ -0,0 +1,93 @@
+#ifndef _BYTESWAP_H
+#define _BYTESWAP_H
+
+#include <xen/types.h>
+
+#if !__has_builtin(__builtin_bswap16)
+static always_inline uint16_t __builtin_bswap16(uint16_t val)
+{
+ return ((val & 0x00FF) << 8) | ((val & 0xFF00) >> 8);
+}
+#endif
+
+#if !__has_builtin(__builtin_bswap32)
+static always_inline uint32_t __builtin_bswap32(uint32_t val)
+{
+ return ((val & 0x000000FF) << 24) |
+ ((val & 0x0000FF00) << 8) |
+ ((val & 0x00FF0000) >> 8) |
+ ((val & 0xFF000000) >> 24);
+}
+#endif
+
+#if !__has_builtin(__builtin_bswap64)
+static always_inline uint64_t __builtin_bswap64(uint64_t val)
+{
+ return ((val & 0x00000000000000FF) << 56) |
+ ((val & 0x000000000000FF00) << 40) |
+ ((val & 0x0000000000FF0000) << 24) |
+ ((val & 0x00000000FF000000) << 8) |
+ ((val & 0x000000FF00000000) >> 8) |
+ ((val & 0x0000FF0000000000) >> 24) |
+ ((val & 0x00FF000000000000) >> 40) |
+ ((val & 0xFF00000000000000) >> 56);
+}
+#endif
+
+#define bswap16(x) __builtin_bswap16(x)
+#define bswap32(x) __builtin_bswap32(x)
+#define bswap64(x) __builtin_bswap64(x)
+
+#define bswap_ul(x) bswap##BITS_PER_LONG(x)
+
+#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
+
+# ifndef __LITTLE_ENDIAN
+# define __LITTLE_ENDIAN 1234
+# endif
+
+# ifndef __LITTLE_ENDIAN_BITFIELD
+# define __LITTLE_ENDIAN_BITFIELD
+# endif
+
+# define cpu_to_le64(x) (x)
+# define le64_to_cpu(x) (x)
+# define cpu_to_le32(x) (x)
+# define le32_to_cpu(x) (x)
+# define cpu_to_le16(x) (x)
+# define le16_to_cpu(x) (x)
+# define cpu_to_be64(x) bswap64(x)
+# define be64_to_cpu(x) bswap64(x)
+# define cpu_to_be32(x) bswap32(x)
+# define be32_to_cpu(x) bswap32(x)
+# define cpu_to_be16(x) bswap16(x)
+# define be16_to_cpu(x) bswap16(x)
+
+#elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
+
+# ifndef __BIG_ENDIAN
+# define __BIG_ENDIAN 4321
+# endif
+
+# ifndef __BIG_ENDIAN_BITFIELD
+# define __BIG_ENDIAN_BITFIELD
+# endif
+
+# define cpu_to_le64(x) bswap64(x)
+# define le64_to_cpu(x) bswap64(x)
+# define cpu_to_le32(x) bswap32(x)
+# define le32_to_cpu(x) bswap32(x)
+# define cpu_to_le16(x) bswap16(x)
+# define le16_to_cpu(x) bswap16(x)
+# define cpu_to_be64(x) (x)
+# define be64_to_cpu(x) (x)
+# define cpu_to_be32(x) (x)
+# define be32_to_cpu(x) (x)
+# define cpu_to_be16(x) (x)
+# define be16_to_cpu(x) (x)
+
+#else
+# error "Unknown Endianness"
+#endif /* __BYTE_ORDER__ */
+
+#endif /* _LINUX_BITOPS_H */
diff --git a/xen/include/xen/compiler.h b/xen/include/xen/compiler.h
index 696c7eb89e..68f28082a5 100644
--- a/xen/include/xen/compiler.h
+++ b/xen/include/xen/compiler.h
@@ -179,4 +179,16 @@
# define CLANG_DISABLE_WARN_GCC_COMPAT_END
#endif
+#if (!defined(__clang__) && (__GNUC__ < 10))
+/*
+ * Backwards compatibility for GCC < 10.
+ * All supported versions of Clang support __has_builtin
+ * */
+#define __has_builtin(x) GCC_has ## x
+
+#define GCC_has__builtin_bswap16 (CONFIG_GCC_VERSION >= 40800)
+#define GCC_has__builtin_bswap32 (CONFIG_GCC_VERSION >= 40400)
+#define GCC_has__builtin_bswap64 (CONFIG_GCC_VERSION >= 40400)
+#endif
+
#endif /* __LINUX_COMPILER_H */
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v2 1/7] xen: implement byteswap.h
2021-10-22 10:47 ` [PATCH v2 1/7] xen: implement byteswap.h Lin Liu
@ 2021-10-22 14:00 ` Andrew Cooper
2021-11-02 14:23 ` Jan Beulich
1 sibling, 0 replies; 16+ messages in thread
From: Andrew Cooper @ 2021-10-22 14:00 UTC (permalink / raw)
To: Lin Liu, xen-devel
Cc: George Dunlap, Ian Jackson, Jan Beulich, Julien Grall,
Stefano Stabellini, Wei Liu
On 22/10/2021 11:47, Lin Liu wrote:
> diff --git a/xen/include/xen/compiler.h b/xen/include/xen/compiler.h
> index 696c7eb89e..68f28082a5 100644
> --- a/xen/include/xen/compiler.h
> +++ b/xen/include/xen/compiler.h
> @@ -179,4 +179,16 @@
> # define CLANG_DISABLE_WARN_GCC_COMPAT_END
> #endif
>
> +#if (!defined(__clang__) && (__GNUC__ < 10))
This too lost my feedback.
It needs to be #ifndef __has_builtin because otherwise a random GCC 10
build I have fails with:
/local/security/xen.git/xen/include/xen/byteswap.h:6:19: error: missing
binary operator before token "("
6 | #if !__has_builtin(__builtin_bswap16)
I suspect it was a build of GCC 10 before __has_builtin() support was
added, but either way, we should be predicating on __has_builtin itself,
not version guesswork.
~Andrew
^ permalink raw reply [flat|nested] 16+ messages in thread* Re: [PATCH v2 1/7] xen: implement byteswap.h
2021-10-22 10:47 ` [PATCH v2 1/7] xen: implement byteswap.h Lin Liu
2021-10-22 14:00 ` Andrew Cooper
@ 2021-11-02 14:23 ` Jan Beulich
1 sibling, 0 replies; 16+ messages in thread
From: Jan Beulich @ 2021-11-02 14:23 UTC (permalink / raw)
To: Lin Liu
Cc: Andrew Cooper, George Dunlap, Ian Jackson, Julien Grall,
Stefano Stabellini, Wei Liu, xen-devel
On 22.10.2021 12:47, Lin Liu wrote:
> --- /dev/null
> +++ b/xen/include/xen/byteswap.h
> @@ -0,0 +1,93 @@
> +#ifndef _BYTESWAP_H
> +#define _BYTESWAP_H
> +
> +#include <xen/types.h>
> +
> +#if !__has_builtin(__builtin_bswap16)
> +static always_inline uint16_t __builtin_bswap16(uint16_t val)
> +{
> + return ((val & 0x00FF) << 8) | ((val & 0xFF00) >> 8);
> +}
> +#endif
> +
> +#if !__has_builtin(__builtin_bswap32)
> +static always_inline uint32_t __builtin_bswap32(uint32_t val)
> +{
> + return ((val & 0x000000FF) << 24) |
> + ((val & 0x0000FF00) << 8) |
> + ((val & 0x00FF0000) >> 8) |
> + ((val & 0xFF000000) >> 24);
> +}
> +#endif
> +
> +#if !__has_builtin(__builtin_bswap64)
> +static always_inline uint64_t __builtin_bswap64(uint64_t val)
> +{
> + return ((val & 0x00000000000000FF) << 56) |
> + ((val & 0x000000000000FF00) << 40) |
> + ((val & 0x0000000000FF0000) << 24) |
> + ((val & 0x00000000FF000000) << 8) |
> + ((val & 0x000000FF00000000) >> 8) |
> + ((val & 0x0000FF0000000000) >> 24) |
> + ((val & 0x00FF000000000000) >> 40) |
> + ((val & 0xFF00000000000000) >> 56);
> +}
> +#endif
> +
> +#define bswap16(x) __builtin_bswap16(x)
> +#define bswap32(x) __builtin_bswap32(x)
> +#define bswap64(x) __builtin_bswap64(x)
> +
> +#define bswap_ul(x) bswap##BITS_PER_LONG(x)
I don't see how this is supposed to work - the compiler isn't going to
expand BITS_PER_LONG before the token concatenation. You'll need helper
macros to achieve that. Linux has __PASTE(); we may want to "steal" that
(albeit preferably without any leading underscores).
> +#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
While you've worked towards abstracting older vs newer compiler
versions, I'm afraid these constants aren't available in gcc 4.1 yet.
They look to have appeared in 4.6.
> +# ifndef __LITTLE_ENDIAN
> +# define __LITTLE_ENDIAN 1234
> +# endif
> +
> +# ifndef __LITTLE_ENDIAN_BITFIELD
> +# define __LITTLE_ENDIAN_BITFIELD
> +# endif
These are definitions which I don't think belong into a header of this
name. They're imo well placed in byteorder/*.h.
Jan
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v2 2/7] crypto/vmac: Simplify code with byteswap.h
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
2021-10-22 10:47 ` [PATCH v2 1/7] xen: implement byteswap.h Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-11-02 15:00 ` Jan Beulich
2021-10-22 10:47 ` [PATCH v2 3/7] arm64/find_next_bit: Remove ext2_swab() Lin Liu
` (5 subsequent siblings)
7 siblings, 1 reply; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel
Cc: Lin Liu, Andrew Cooper, George Dunlap, Ian Jackson, Jan Beulich,
Julien Grall, Stefano Stabellini, Wei Liu
This file has its own implementation of swap bytes. Clean up
the code with xen/byteswap.h.
No functional change.
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: George Dunlap <george.dunlap@citrix.com>
Cc: Ian Jackson <iwj@xenproject.org>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Julien Grall <julien@xen.org>
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Wei Liu <wl@xen.org>
---
xen/crypto/vmac.c | 76 ++---------------------------------------------
1 file changed, 3 insertions(+), 73 deletions(-)
diff --git a/xen/crypto/vmac.c b/xen/crypto/vmac.c
index 294dd16a52..acb4e015f5 100644
--- a/xen/crypto/vmac.c
+++ b/xen/crypto/vmac.c
@@ -8,6 +8,7 @@
/* start for Xen */
#include <xen/init.h>
+#include <xen/byteswap.h>
#include <xen/types.h>
#include <xen/lib.h>
#include <crypto/vmac.h>
@@ -50,7 +51,6 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
* MUL64: 64x64->128-bit multiplication
* PMUL64: assumes top bits cleared on inputs
* ADD128: 128x128->128-bit addition
- * GET_REVERSED_64: load and byte-reverse 64-bit word
* ----------------------------------------------------------------------- */
/* ----------------------------------------------------------------------- */
@@ -68,22 +68,6 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
#define PMUL64 MUL64
-#define GET_REVERSED_64(p) \
- ({uint64_t x; \
- asm ("bswapq %0" : "=r" (x) : "0"(*(uint64_t *)(p))); x;})
-
-/* ----------------------------------------------------------------------- */
-#elif (__GNUC__ && __i386__)
-/* ----------------------------------------------------------------------- */
-
-#define GET_REVERSED_64(p) \
- ({ uint64_t x; \
- uint32_t *tp = (uint32_t *)(p); \
- asm ("bswap %%edx\n\t" \
- "bswap %%eax" \
- : "=A"(x) \
- : "a"(tp[1]), "d"(tp[0])); \
- x; })
/* ----------------------------------------------------------------------- */
#elif (__GNUC__ && __ppc64__)
@@ -103,37 +87,6 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
#define PMUL64 MUL64
-#define GET_REVERSED_64(p) \
- ({ uint32_t hi, lo, *_p = (uint32_t *)(p); \
- asm volatile ("lwbrx %0, %1, %2" : "=r"(lo) : "b%"(0), "r"(_p) ); \
- asm volatile ("lwbrx %0, %1, %2" : "=r"(hi) : "b%"(4), "r"(_p) ); \
- ((uint64_t)hi << 32) | (uint64_t)lo; } )
-
-/* ----------------------------------------------------------------------- */
-#elif (__GNUC__ && (__ppc__ || __PPC__))
-/* ----------------------------------------------------------------------- */
-
-#define GET_REVERSED_64(p) \
- ({ uint32_t hi, lo, *_p = (uint32_t *)(p); \
- asm volatile ("lwbrx %0, %1, %2" : "=r"(lo) : "b%"(0), "r"(_p) ); \
- asm volatile ("lwbrx %0, %1, %2" : "=r"(hi) : "b%"(4), "r"(_p) ); \
- ((uint64_t)hi << 32) | (uint64_t)lo; } )
-
-/* ----------------------------------------------------------------------- */
-#elif (__GNUC__ && (__ARMEL__ || __ARM__))
-/* ----------------------------------------------------------------------- */
-
-#define bswap32(v) \
-({ uint32_t tmp,out; \
- asm volatile( \
- "eor %1, %2, %2, ror #16\n" \
- "bic %1, %1, #0x00ff0000\n" \
- "mov %0, %2, ror #8\n" \
- "eor %0, %0, %1, lsr #8" \
- : "=r" (out), "=&r" (tmp) \
- : "r" (v)); \
- out;})
-
/* ----------------------------------------------------------------------- */
#elif _MSC_VER
/* ----------------------------------------------------------------------- */
@@ -154,11 +107,6 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
(rh) += (ih) + ((rl) < (_il)); \
}
-#if _MSC_VER >= 1300
-#define GET_REVERSED_64(p) _byteswap_uint64(*(uint64_t *)(p))
-#pragma intrinsic(_byteswap_uint64)
-#endif
-
#if _MSC_VER >= 1400 && \
(!defined(__INTEL_COMPILER) || __INTEL_COMPILER >= 1000)
#define MUL32(i1,i2) (__emulu((uint32_t)(i1),(uint32_t)(i2)))
@@ -219,24 +167,6 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
}
#endif
-#ifndef GET_REVERSED_64
-#ifndef bswap64
-#ifndef bswap32
-#define bswap32(x) \
- ({ uint32_t bsx = (x); \
- ((((bsx) & 0xff000000u) >> 24) | (((bsx) & 0x00ff0000u) >> 8) | \
- (((bsx) & 0x0000ff00u) << 8) | (((bsx) & 0x000000ffu) << 24)); })
-#endif
-#define bswap64(x) \
- ({ union { uint64_t ll; uint32_t l[2]; } w, r; \
- w.ll = (x); \
- r.l[0] = bswap32 (w.l[1]); \
- r.l[1] = bswap32 (w.l[0]); \
- r.ll; })
-#endif
-#define GET_REVERSED_64(p) bswap64(*(uint64_t *)(p))
-#endif
-
/* ----------------------------------------------------------------------- */
#if (VMAC_PREFER_BIG_ENDIAN)
@@ -247,9 +177,9 @@ const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
#if (VMAC_ARCH_BIG_ENDIAN)
# define get64BE(ptr) (*(uint64_t *)(ptr))
-# define get64LE(ptr) GET_REVERSED_64(ptr)
+# define get64LE(ptr) bswap64(*(uint64_t *)(ptr))
#else /* assume little-endian */
-# define get64BE(ptr) GET_REVERSED_64(ptr)
+# define get64BE(ptr) bswap64(*(uint64_t *)(ptr))
# define get64LE(ptr) (*(uint64_t *)(ptr))
#endif
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* [PATCH v2 3/7] arm64/find_next_bit: Remove ext2_swab()
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
2021-10-22 10:47 ` [PATCH v2 1/7] xen: implement byteswap.h Lin Liu
2021-10-22 10:47 ` [PATCH v2 2/7] crypto/vmac: Simplify code with byteswap.h Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 10:47 ` [PATCH v2 4/7] arm: Switch to byteswap.h Lin Liu
` (4 subsequent siblings)
7 siblings, 0 replies; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel; +Cc: Lin Liu, Stefano Stabellini, Julien Grall, Volodymyr Babchuk
ext2 has nothing to do with this logic. Clean up the code with
xen/byteswap.h which now has an unsigned long helper.
No functional change.
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Julien Grall <julien@xen.org>
Cc: Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
The patches are not well tested without running environment
---
xen/arch/arm/arm64/lib/find_next_bit.c | 40 ++++++--------------------
1 file changed, 8 insertions(+), 32 deletions(-)
diff --git a/xen/arch/arm/arm64/lib/find_next_bit.c b/xen/arch/arm/arm64/lib/find_next_bit.c
index 8ebf8bfe97..8a7cfc0949 100644
--- a/xen/arch/arm/arm64/lib/find_next_bit.c
+++ b/xen/arch/arm/arm64/lib/find_next_bit.c
@@ -8,9 +8,9 @@
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
-#include <xen/bitops.h>
#include <asm/types.h>
-#include <asm/byteorder.h>
+#include <xen/bitops.h>
+#include <xen/byteswap.h>
#ifndef find_next_bit
/*
@@ -161,30 +161,6 @@ EXPORT_SYMBOL(find_first_zero_bit);
#ifdef __BIG_ENDIAN
-/* include/linux/byteorder does not support "unsigned long" type */
-static inline unsigned long ext2_swabp(const unsigned long * x)
-{
-#if BITS_PER_LONG == 64
- return (unsigned long) __swab64p((u64 *) x);
-#elif BITS_PER_LONG == 32
- return (unsigned long) __swab32p((u32 *) x);
-#else
-#error BITS_PER_LONG not defined
-#endif
-}
-
-/* include/linux/byteorder doesn't support "unsigned long" type */
-static inline unsigned long ext2_swab(const unsigned long y)
-{
-#if BITS_PER_LONG == 64
- return (unsigned long) __swab64((u64) y);
-#elif BITS_PER_LONG == 32
- return (unsigned long) __swab32((u32) y);
-#else
-#error BITS_PER_LONG not defined
-#endif
-}
-
#ifndef find_next_zero_bit_le
unsigned long find_next_zero_bit_le(const void *addr, unsigned
long size, unsigned long offset)
@@ -199,7 +175,7 @@ unsigned long find_next_zero_bit_le(const void *addr, unsigned
size -= result;
offset &= (BITS_PER_LONG - 1UL);
if (offset) {
- tmp = ext2_swabp(p++);
+ tmp = bswap_ul(*p++);
tmp |= (~0UL >> (BITS_PER_LONG - offset));
if (size < BITS_PER_LONG)
goto found_first;
@@ -217,7 +193,7 @@ unsigned long find_next_zero_bit_le(const void *addr, unsigned
}
if (!size)
return result;
- tmp = ext2_swabp(p);
+ tmp = bswap_ul(*p);
found_first:
tmp |= ~0UL << size;
if (tmp == ~0UL) /* Are any bits zero? */
@@ -226,7 +202,7 @@ found_middle:
return result + ffz(tmp);
found_middle_swap:
- return result + ffz(ext2_swab(tmp));
+ return result + ffz(bswap_ul(tmp));
}
EXPORT_SYMBOL(find_next_zero_bit_le);
#endif
@@ -245,7 +221,7 @@ unsigned long find_next_bit_le(const void *addr, unsigned
size -= result;
offset &= (BITS_PER_LONG - 1UL);
if (offset) {
- tmp = ext2_swabp(p++);
+ tmp = bswap_ul(*p++);
tmp &= (~0UL << offset);
if (size < BITS_PER_LONG)
goto found_first;
@@ -264,7 +240,7 @@ unsigned long find_next_bit_le(const void *addr, unsigned
}
if (!size)
return result;
- tmp = ext2_swabp(p);
+ tmp = bswap_ul(*p);
found_first:
tmp &= (~0UL >> (BITS_PER_LONG - size));
if (tmp == 0UL) /* Are any bits set? */
@@ -273,7 +249,7 @@ found_middle:
return result + __ffs(tmp);
found_middle_swap:
- return result + __ffs(ext2_swab(tmp));
+ return result + __ffs(bswap_ul(tmp));
}
EXPORT_SYMBOL(find_next_bit_le);
#endif
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* [PATCH v2 4/7] arm: Switch to byteswap.h
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
` (2 preceding siblings ...)
2021-10-22 10:47 ` [PATCH v2 3/7] arm64/find_next_bit: Remove ext2_swab() Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 10:47 ` [PATCH v2 5/7] xen/xsm: " Lin Liu
` (3 subsequent siblings)
7 siblings, 0 replies; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel
Cc: Lin Liu, Stefano Stabellini, Julien Grall, Volodymyr Babchuk,
Konrad Rzeszutek Wilk, Ross Lagerwall
Update to use byteswap.h to swap bytes.
No functional change.
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Julien Grall <julien@xen.org>
Cc: Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Ross Lagerwall <ross.lagerwall@citrix.com>
---
xen/arch/arm/alternative.c | 2 +-
xen/arch/arm/arm64/livepatch.c | 2 +-
xen/arch/arm/kernel.c | 2 +-
xen/arch/arm/vgic/vgic-mmio.c | 2 +-
xen/include/asm-arm/arm32/io.h | 2 +-
xen/include/asm-arm/arm64/io.h | 2 +-
xen/include/xen/libfdt/libfdt_env.h | 2 +-
7 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/xen/arch/arm/alternative.c b/xen/arch/arm/alternative.c
index 237c4e5642..0f84260ac9 100644
--- a/xen/arch/arm/alternative.c
+++ b/xen/arch/arm/alternative.c
@@ -17,6 +17,7 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
+#include <xen/byteswap.h>
#include <xen/init.h>
#include <xen/types.h>
#include <xen/kernel.h>
@@ -27,7 +28,6 @@
#include <xen/virtual_region.h>
#include <asm/alternative.h>
#include <asm/atomic.h>
-#include <asm/byteorder.h>
#include <asm/cpufeature.h>
#include <asm/insn.h>
#include <asm/page.h>
diff --git a/xen/arch/arm/arm64/livepatch.c b/xen/arch/arm/arm64/livepatch.c
index 6ec8dc60f0..26b5729edf 100644
--- a/xen/arch/arm/arm64/livepatch.c
+++ b/xen/arch/arm/arm64/livepatch.c
@@ -3,6 +3,7 @@
*/
#include <xen/bitops.h>
+#include <xen/byteswap.h>
#include <xen/errno.h>
#include <xen/lib.h>
#include <xen/livepatch_elf.h>
@@ -11,7 +12,6 @@
#include <xen/vmap.h>
#include <asm/bitops.h>
-#include <asm/byteorder.h>
#include <asm/insn.h>
#include <asm/livepatch.h>
diff --git a/xen/arch/arm/kernel.c b/xen/arch/arm/kernel.c
index 8f43caa186..e89bb9cef8 100644
--- a/xen/arch/arm/kernel.c
+++ b/xen/arch/arm/kernel.c
@@ -3,6 +3,7 @@
*
* Copyright (C) 2011 Citrix Systems, Inc.
*/
+#include <xen/byteswap.h>
#include <xen/domain_page.h>
#include <xen/errno.h>
#include <xen/guest_access.h>
@@ -14,7 +15,6 @@
#include <xen/sched.h>
#include <xen/vmap.h>
-#include <asm/byteorder.h>
#include <asm/kernel.h>
#include <asm/setup.h>
diff --git a/xen/arch/arm/vgic/vgic-mmio.c b/xen/arch/arm/vgic/vgic-mmio.c
index 5d935a7301..d9c5066246 100644
--- a/xen/arch/arm/vgic/vgic-mmio.c
+++ b/xen/arch/arm/vgic/vgic-mmio.c
@@ -13,10 +13,10 @@
*/
#include <xen/bitops.h>
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/sched.h>
#include <asm/new_vgic.h>
-#include <asm/byteorder.h>
#include "vgic.h"
#include "vgic-mmio.h"
diff --git a/xen/include/asm-arm/arm32/io.h b/xen/include/asm-arm/arm32/io.h
index 73a879e9fb..df8547403c 100644
--- a/xen/include/asm-arm/arm32/io.h
+++ b/xen/include/asm-arm/arm32/io.h
@@ -21,8 +21,8 @@
#ifndef _ARM_ARM32_IO_H
#define _ARM_ARM32_IO_H
+#include <xen/byteswap.h>
#include <asm/system.h>
-#include <asm/byteorder.h>
static inline void __raw_writeb(u8 val, volatile void __iomem *addr)
{
diff --git a/xen/include/asm-arm/arm64/io.h b/xen/include/asm-arm/arm64/io.h
index 30bfc78d9e..db328d9e3c 100644
--- a/xen/include/asm-arm/arm64/io.h
+++ b/xen/include/asm-arm/arm64/io.h
@@ -20,8 +20,8 @@
#ifndef _ARM_ARM64_IO_H
#define _ARM_ARM64_IO_H
+#include <xen/byteswap.h>
#include <asm/system.h>
-#include <asm/byteorder.h>
#include <asm/alternative.h>
/*
diff --git a/xen/include/xen/libfdt/libfdt_env.h b/xen/include/xen/libfdt/libfdt_env.h
index 035bf754d2..f8ea1ea07a 100644
--- a/xen/include/xen/libfdt/libfdt_env.h
+++ b/xen/include/xen/libfdt/libfdt_env.h
@@ -1,9 +1,9 @@
#ifndef _LIBFDT_ENV_H
#define _LIBFDT_ENV_H
+#include <xen/byteswap.h>
#include <xen/types.h>
#include <xen/string.h>
-#include <asm/byteorder.h>
typedef uint16_t fdt16_t;
typedef uint32_t fdt32_t;
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* [PATCH v2 5/7] xen/xsm: Switch to byteswap.h
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
` (3 preceding siblings ...)
2021-10-22 10:47 ` [PATCH v2 4/7] arm: Switch to byteswap.h Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 15:46 ` Daniel P. Smith
2021-10-22 10:47 ` [PATCH v2 6/7] xen: " Lin Liu
` (2 subsequent siblings)
7 siblings, 1 reply; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel; +Cc: Lin Liu, Daniel De Graaf, Daniel P. Smith
Update to use byteswap.h to swap bytes
No functional change
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Cc: "Daniel P. Smith" <dpsmith@apertussolutions.com>
---
xen/xsm/flask/ss/avtab.c | 2 +-
xen/xsm/flask/ss/conditional.c | 2 +-
xen/xsm/flask/ss/ebitmap.c | 2 +-
xen/xsm/flask/ss/policydb.c | 2 +-
4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/xen/xsm/flask/ss/avtab.c b/xen/xsm/flask/ss/avtab.c
index bfc91c8b0c..1fa796f625 100644
--- a/xen/xsm/flask/ss/avtab.c
+++ b/xen/xsm/flask/ss/avtab.c
@@ -19,8 +19,8 @@
/* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
+#include <xen/byteswap.h>
#include <xen/lib.h>
-#include <asm/byteorder.h>
#include <xen/types.h>
#include <xen/xmalloc.h>
#include <xen/errno.h>
diff --git a/xen/xsm/flask/ss/conditional.c b/xen/xsm/flask/ss/conditional.c
index 3e58aea551..059f6e07e5 100644
--- a/xen/xsm/flask/ss/conditional.c
+++ b/xen/xsm/flask/ss/conditional.c
@@ -9,7 +9,7 @@
/* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/types.h>
#include <xen/errno.h>
diff --git a/xen/xsm/flask/ss/ebitmap.c b/xen/xsm/flask/ss/ebitmap.c
index e1d0a586a7..1550437c6f 100644
--- a/xen/xsm/flask/ss/ebitmap.c
+++ b/xen/xsm/flask/ss/ebitmap.c
@@ -10,7 +10,7 @@
/* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/xmalloc.h>
#include <xen/errno.h>
diff --git a/xen/xsm/flask/ss/policydb.c b/xen/xsm/flask/ss/policydb.c
index 9426164353..595005c3b7 100644
--- a/xen/xsm/flask/ss/policydb.c
+++ b/xen/xsm/flask/ss/policydb.c
@@ -22,7 +22,7 @@
/* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/types.h>
#include <xen/xmalloc.h>
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v2 5/7] xen/xsm: Switch to byteswap.h
2021-10-22 10:47 ` [PATCH v2 5/7] xen/xsm: " Lin Liu
@ 2021-10-22 15:46 ` Daniel P. Smith
2021-11-02 15:18 ` Jan Beulich
0 siblings, 1 reply; 16+ messages in thread
From: Daniel P. Smith @ 2021-10-22 15:46 UTC (permalink / raw)
To: Lin Liu, xen-devel; +Cc: Daniel De Graaf
On 10/22/21 6:47 AM, Lin Liu wrote:
> Update to use byteswap.h to swap bytes
>
> No functional change
>
> Signed-off-by: Lin Liu <lin.liu@citrix.com>
> ---
> Cc: Daniel De Graaf <dgdegra@tycho.nsa.gov>
> Cc: "Daniel P. Smith" <dpsmith@apertussolutions.com>
> ---
> xen/xsm/flask/ss/avtab.c | 2 +-
> xen/xsm/flask/ss/conditional.c | 2 +-
> xen/xsm/flask/ss/ebitmap.c | 2 +-
> xen/xsm/flask/ss/policydb.c | 2 +-
> 4 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/xen/xsm/flask/ss/avtab.c b/xen/xsm/flask/ss/avtab.c
> index bfc91c8b0c..1fa796f625 100644
> --- a/xen/xsm/flask/ss/avtab.c
> +++ b/xen/xsm/flask/ss/avtab.c
> @@ -19,8 +19,8 @@
>
> /* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> -#include <asm/byteorder.h>
> #include <xen/types.h>
> #include <xen/xmalloc.h>
> #include <xen/errno.h>
> diff --git a/xen/xsm/flask/ss/conditional.c b/xen/xsm/flask/ss/conditional.c
> index 3e58aea551..059f6e07e5 100644
> --- a/xen/xsm/flask/ss/conditional.c
> +++ b/xen/xsm/flask/ss/conditional.c
> @@ -9,7 +9,7 @@
>
> /* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> #include <xen/types.h>
> #include <xen/errno.h>
> diff --git a/xen/xsm/flask/ss/ebitmap.c b/xen/xsm/flask/ss/ebitmap.c
> index e1d0a586a7..1550437c6f 100644
> --- a/xen/xsm/flask/ss/ebitmap.c
> +++ b/xen/xsm/flask/ss/ebitmap.c
> @@ -10,7 +10,7 @@
>
> /* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> #include <xen/xmalloc.h>
> #include <xen/errno.h>
> diff --git a/xen/xsm/flask/ss/policydb.c b/xen/xsm/flask/ss/policydb.c
> index 9426164353..595005c3b7 100644
> --- a/xen/xsm/flask/ss/policydb.c
> +++ b/xen/xsm/flask/ss/policydb.c
> @@ -22,7 +22,7 @@
>
> /* Ported to Xen 3.0, George Coker, <gscoker@alpha.ncsc.mil> */
>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> #include <xen/types.h>
> #include <xen/xmalloc.h>
>
Reviewed by: Daniel P. Smith <dpsmith@apertussolutions.com>
^ permalink raw reply [flat|nested] 16+ messages in thread
* Re: [PATCH v2 5/7] xen/xsm: Switch to byteswap.h
2021-10-22 15:46 ` Daniel P. Smith
@ 2021-11-02 15:18 ` Jan Beulich
0 siblings, 0 replies; 16+ messages in thread
From: Jan Beulich @ 2021-11-02 15:18 UTC (permalink / raw)
To: Lin Liu; +Cc: Daniel De Graaf, Daniel P. Smith, xen-devel
On 22.10.2021 17:46, Daniel P. Smith wrote:
> On 10/22/21 6:47 AM, Lin Liu wrote:
>> Update to use byteswap.h to swap bytes
>>
>> No functional change
>>
>> Signed-off-by: Lin Liu <lin.liu@citrix.com>
>
> Reviewed by: Daniel P. Smith <dpsmith@apertussolutions.com>
Acked-by: Jan Beulich <jbeulich@suse.com>
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v2 6/7] xen: Switch to byteswap.h
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
` (4 preceding siblings ...)
2021-10-22 10:47 ` [PATCH v2 5/7] xen/xsm: " Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 11:25 ` Andrew Cooper
2021-11-02 15:23 ` Jan Beulich
2021-10-22 10:47 ` [PATCH v2 7/7] byteorder: Remove byteorder Lin Liu
2021-10-22 11:50 ` [PATCH v2 0/7] Implement byteswap and update references Andrew Cooper
7 siblings, 2 replies; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel
Cc: Lin Liu, Andrew Cooper, George Dunlap, Ian Jackson, Jan Beulich,
Julien Grall, Stefano Stabellini, Wei Liu, Roger Pau Monné
Update to use byteswap.h to swap bytes.
No functional chagne.
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: George Dunlap <george.dunlap@citrix.com>
Cc: Ian Jackson <iwj@xenproject.org>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Julien Grall <julien@xen.org>
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Wei Liu <wl@xen.org>
Cc: "Roger Pau Monné" <roger.pau@citrix.com>
---
xen/common/bitmap.c | 2 +-
xen/common/gdbstub.c | 2 +-
xen/common/libelf/libelf-private.h | 8 ++++----
xen/common/lz4/defs.h | 2 +-
xen/common/lzo.c | 2 +-
xen/common/unlzo.c | 2 +-
xen/common/xz/private.h | 4 ++--
xen/drivers/char/ehci-dbgp.c | 2 +-
xen/include/asm-x86/msi.h | 2 +-
xen/include/xen/bitmap.h | 2 +-
xen/include/xen/device_tree.h | 2 +-
xen/include/xen/unaligned.h | 14 +++++++-------
xen/lib/divmod.c | 2 +-
13 files changed, 23 insertions(+), 23 deletions(-)
diff --git a/xen/common/bitmap.c b/xen/common/bitmap.c
index 7d4551f782..be274ca04a 100644
--- a/xen/common/bitmap.c
+++ b/xen/common/bitmap.c
@@ -9,10 +9,10 @@
#include <xen/errno.h>
#include <xen/bitmap.h>
#include <xen/bitops.h>
+#include <xen/byteswap.h>
#include <xen/cpumask.h>
#include <xen/guest_access.h>
#include <xen/lib.h>
-#include <asm/byteorder.h>
/*
* bitmaps provide an array of bits, implemented using an an
diff --git a/xen/common/gdbstub.c b/xen/common/gdbstub.c
index 848c1f4327..3c8ed52d6b 100644
--- a/xen/common/gdbstub.c
+++ b/xen/common/gdbstub.c
@@ -33,6 +33,7 @@
/* Resuming after we've stopped used to work, but more through luck
than any actual intention. It doesn't at the moment. */
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/spinlock.h>
#include <xen/serial.h>
@@ -45,7 +46,6 @@
#include <xen/console.h>
#include <xen/errno.h>
#include <xen/delay.h>
-#include <asm/byteorder.h>
/* Printk isn't particularly safe just after we've trapped to the
debugger. so avoid it. */
diff --git a/xen/common/libelf/libelf-private.h b/xen/common/libelf/libelf-private.h
index 47db679966..b7089cb31b 100644
--- a/xen/common/libelf/libelf-private.h
+++ b/xen/common/libelf/libelf-private.h
@@ -17,10 +17,10 @@
#ifdef __XEN__
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/libelf.h>
#include <xen/softirq.h>
-#include <asm/byteorder.h>
#include <public/elfnote.h>
/* we would like to use elf->log_callback but we can't because
@@ -31,9 +31,9 @@
printk(fmt, ## args )
#define strtoull(str, end, base) simple_strtoull(str, end, base)
-#define bswap_16(x) swab16(x)
-#define bswap_32(x) swab32(x)
-#define bswap_64(x) swab64(x)
+#define bswap_16(x) bswap16(x)
+#define bswap_32(x) bswap32(x)
+#define bswap_64(x) bswap64(x)
#else /* !__XEN__ */
diff --git a/xen/common/lz4/defs.h b/xen/common/lz4/defs.h
index 10609f5a53..1ce4476478 100644
--- a/xen/common/lz4/defs.h
+++ b/xen/common/lz4/defs.h
@@ -9,7 +9,7 @@
*/
#ifdef __XEN__
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <asm/unaligned.h>
#else
diff --git a/xen/common/lzo.c b/xen/common/lzo.c
index a87c76dded..17be9675f4 100644
--- a/xen/common/lzo.c
+++ b/xen/common/lzo.c
@@ -96,7 +96,7 @@
#ifdef __XEN__
#include <xen/lib.h>
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <asm/unaligned.h>
#else
#define get_unaligned_le16(_p) (*(u16 *)(_p))
diff --git a/xen/common/unlzo.c b/xen/common/unlzo.c
index 74056778eb..f908d2a61f 100644
--- a/xen/common/unlzo.c
+++ b/xen/common/unlzo.c
@@ -33,7 +33,7 @@
#include <xen/lzo.h>
#ifdef __XEN__
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <asm/unaligned.h>
#else
diff --git a/xen/common/xz/private.h b/xen/common/xz/private.h
index 511343fcc2..647f9699a7 100644
--- a/xen/common/xz/private.h
+++ b/xen/common/xz/private.h
@@ -12,7 +12,7 @@
#ifdef __XEN__
#include <xen/kernel.h>
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#include <asm/unaligned.h>
#else
@@ -28,7 +28,7 @@ static inline void put_unaligned_le32(u32 val, void *p)
#endif
-#define get_le32(p) le32_to_cpup((const uint32_t *)(p))
+#define get_le32(p) le32_to_cpu(*(const uint32_t *)(p))
#define false 0
#define true 1
diff --git a/xen/drivers/char/ehci-dbgp.c b/xen/drivers/char/ehci-dbgp.c
index c893d246de..8412da1b11 100644
--- a/xen/drivers/char/ehci-dbgp.c
+++ b/xen/drivers/char/ehci-dbgp.c
@@ -5,13 +5,13 @@
* Linux; see the Linux source for authorship and copyrights.
*/
+#include <xen/byteswap.h>
#include <xen/console.h>
#include <xen/delay.h>
#include <xen/errno.h>
#include <xen/param.h>
#include <xen/pci.h>
#include <xen/serial.h>
-#include <asm/byteorder.h>
#include <asm/io.h>
#include <asm/fixmap.h>
#include <public/physdev.h>
diff --git a/xen/include/asm-x86/msi.h b/xen/include/asm-x86/msi.h
index e228b0f3f3..277375183c 100644
--- a/xen/include/asm-x86/msi.h
+++ b/xen/include/asm-x86/msi.h
@@ -1,9 +1,9 @@
#ifndef __ASM_MSI_H
#define __ASM_MSI_H
+#include <xen/byteswap.h>
#include <xen/cpumask.h>
#include <xen/pci.h>
-#include <asm/byteorder.h>
#include <asm/hvm/vmx/vmcs.h>
/*
diff --git a/xen/include/xen/bitmap.h b/xen/include/xen/bitmap.h
index e9175ab54a..c44a1cb63c 100644
--- a/xen/include/xen/bitmap.h
+++ b/xen/include/xen/bitmap.h
@@ -229,7 +229,7 @@ static inline int bitmap_weight(const unsigned long *src, int nbits)
return __bitmap_weight(src, nbits);
}
-#include <asm/byteorder.h>
+#include <xen/byteswap.h>
#ifdef __LITTLE_ENDIAN
#define BITMAP_MEM_ALIGNMENT 8
diff --git a/xen/include/xen/device_tree.h b/xen/include/xen/device_tree.h
index fd6cd00b43..4921e6b142 100644
--- a/xen/include/xen/device_tree.h
+++ b/xen/include/xen/device_tree.h
@@ -10,10 +10,10 @@
#ifndef __XEN_DEVICE_TREE_H__
#define __XEN_DEVICE_TREE_H__
-#include <asm/byteorder.h>
#include <asm/device.h>
#include <public/xen.h>
#include <public/device_tree_defs.h>
+#include <xen/byteswap.h>
#include <xen/kernel.h>
#include <xen/string.h>
#include <xen/types.h>
diff --git a/xen/include/xen/unaligned.h b/xen/include/xen/unaligned.h
index 0a2b16d05d..8a9ec8a0ac 100644
--- a/xen/include/xen/unaligned.h
+++ b/xen/include/xen/unaligned.h
@@ -11,8 +11,8 @@
#define __XEN_UNALIGNED_H__
#ifdef __XEN__
+#include <xen/byteswap.h>
#include <xen/types.h>
-#include <asm/byteorder.h>
#endif
#define get_unaligned(p) (*(p))
@@ -20,7 +20,7 @@
static inline uint16_t get_unaligned_be16(const void *p)
{
- return be16_to_cpup(p);
+ return be16_to_cpu(*(uint16_t*)p);
}
static inline void put_unaligned_be16(uint16_t val, void *p)
@@ -30,7 +30,7 @@ static inline void put_unaligned_be16(uint16_t val, void *p)
static inline uint32_t get_unaligned_be32(const void *p)
{
- return be32_to_cpup(p);
+ return be32_to_cpu(*(uint32_t*)p);
}
static inline void put_unaligned_be32(uint32_t val, void *p)
@@ -40,7 +40,7 @@ static inline void put_unaligned_be32(uint32_t val, void *p)
static inline uint64_t get_unaligned_be64(const void *p)
{
- return be64_to_cpup(p);
+ return be64_to_cpu(*(uint64_t*)p);
}
static inline void put_unaligned_be64(uint64_t val, void *p)
@@ -50,7 +50,7 @@ static inline void put_unaligned_be64(uint64_t val, void *p)
static inline uint16_t get_unaligned_le16(const void *p)
{
- return le16_to_cpup(p);
+ return le16_to_cpu(*(uint16_t*)p);
}
static inline void put_unaligned_le16(uint16_t val, void *p)
@@ -60,7 +60,7 @@ static inline void put_unaligned_le16(uint16_t val, void *p)
static inline uint32_t get_unaligned_le32(const void *p)
{
- return le32_to_cpup(p);
+ return le32_to_cpu(*(uint32_t*)p);
}
static inline void put_unaligned_le32(uint32_t val, void *p)
@@ -70,7 +70,7 @@ static inline void put_unaligned_le32(uint32_t val, void *p)
static inline uint64_t get_unaligned_le64(const void *p)
{
- return le64_to_cpup(p);
+ return le64_to_cpu(*(uint64_t*)p);
}
static inline void put_unaligned_le64(uint64_t val, void *p)
diff --git a/xen/lib/divmod.c b/xen/lib/divmod.c
index 0be6ccc700..dfc1129cfe 100644
--- a/xen/lib/divmod.c
+++ b/xen/lib/divmod.c
@@ -1,6 +1,6 @@
+#include <xen/byteswap.h>
#include <xen/lib.h>
#include <xen/types.h>
-#include <asm/byteorder.h>
/*
* A couple of 64 bit operations ported from FreeBSD.
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v2 6/7] xen: Switch to byteswap.h
2021-10-22 10:47 ` [PATCH v2 6/7] xen: " Lin Liu
@ 2021-10-22 11:25 ` Andrew Cooper
2021-11-02 15:23 ` Jan Beulich
1 sibling, 0 replies; 16+ messages in thread
From: Andrew Cooper @ 2021-10-22 11:25 UTC (permalink / raw)
To: Lin Liu, xen-devel
Cc: George Dunlap, Ian Jackson, Jan Beulich, Julien Grall,
Stefano Stabellini, Wei Liu, Roger Pau Monné
On 22/10/2021 11:47, Lin Liu wrote:
> Update to use byteswap.h to swap bytes.
>
> No functional chagne.
>
> Signed-off-by: Lin Liu <lin.liu@citrix.com>
> ---
> Cc: Andrew Cooper <andrew.cooper3@citrix.com>
> Cc: George Dunlap <george.dunlap@citrix.com>
> Cc: Ian Jackson <iwj@xenproject.org>
> Cc: Jan Beulich <jbeulich@suse.com>
> Cc: Julien Grall <julien@xen.org>
> Cc: Stefano Stabellini <sstabellini@kernel.org>
> Cc: Wei Liu <wl@xen.org>
> Cc: "Roger Pau Monné" <roger.pau@citrix.com>
> ---
> xen/common/bitmap.c | 2 +-
> xen/common/gdbstub.c | 2 +-
> xen/common/libelf/libelf-private.h | 8 ++++----
> xen/common/lz4/defs.h | 2 +-
> xen/common/lzo.c | 2 +-
> xen/common/unlzo.c | 2 +-
> xen/common/xz/private.h | 4 ++--
> xen/drivers/char/ehci-dbgp.c | 2 +-
> xen/include/asm-x86/msi.h | 2 +-
> xen/include/xen/bitmap.h | 2 +-
> xen/include/xen/device_tree.h | 2 +-
> xen/include/xen/unaligned.h | 14 +++++++-------
> xen/lib/divmod.c | 2 +-
> 13 files changed, 23 insertions(+), 23 deletions(-)
>
> diff --git a/xen/common/bitmap.c b/xen/common/bitmap.c
> index 7d4551f782..be274ca04a 100644
> --- a/xen/common/bitmap.c
> +++ b/xen/common/bitmap.c
> @@ -9,10 +9,10 @@
> #include <xen/errno.h>
> #include <xen/bitmap.h>
> #include <xen/bitops.h>
> +#include <xen/byteswap.h>
> #include <xen/cpumask.h>
> #include <xen/guest_access.h>
> #include <xen/lib.h>
> -#include <asm/byteorder.h>
>
> /*
> * bitmaps provide an array of bits, implemented using an an
> diff --git a/xen/common/gdbstub.c b/xen/common/gdbstub.c
> index 848c1f4327..3c8ed52d6b 100644
> --- a/xen/common/gdbstub.c
> +++ b/xen/common/gdbstub.c
> @@ -33,6 +33,7 @@
> /* Resuming after we've stopped used to work, but more through luck
> than any actual intention. It doesn't at the moment. */
>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> #include <xen/spinlock.h>
> #include <xen/serial.h>
> @@ -45,7 +46,6 @@
> #include <xen/console.h>
> #include <xen/errno.h>
> #include <xen/delay.h>
> -#include <asm/byteorder.h>
>
> /* Printk isn't particularly safe just after we've trapped to the
> debugger. so avoid it. */
> diff --git a/xen/common/libelf/libelf-private.h b/xen/common/libelf/libelf-private.h
> index 47db679966..b7089cb31b 100644
> --- a/xen/common/libelf/libelf-private.h
> +++ b/xen/common/libelf/libelf-private.h
> @@ -17,10 +17,10 @@
>
> #ifdef __XEN__
>
> +#include <xen/byteswap.h>
> #include <xen/lib.h>
> #include <xen/libelf.h>
> #include <xen/softirq.h>
> -#include <asm/byteorder.h>
> #include <public/elfnote.h>
>
> /* we would like to use elf->log_callback but we can't because
> @@ -31,9 +31,9 @@
> printk(fmt, ## args )
>
> #define strtoull(str, end, base) simple_strtoull(str, end, base)
> -#define bswap_16(x) swab16(x)
> -#define bswap_32(x) swab32(x)
> -#define bswap_64(x) swab64(x)
> +#define bswap_16(x) bswap16(x)
> +#define bswap_32(x) bswap32(x)
> +#define bswap_64(x) bswap64(x)
>
> #else /* !__XEN__ */
>
> diff --git a/xen/common/lz4/defs.h b/xen/common/lz4/defs.h
> index 10609f5a53..1ce4476478 100644
> --- a/xen/common/lz4/defs.h
> +++ b/xen/common/lz4/defs.h
> @@ -9,7 +9,7 @@
> */
>
> #ifdef __XEN__
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <asm/unaligned.h>
> #else
>
> diff --git a/xen/common/lzo.c b/xen/common/lzo.c
> index a87c76dded..17be9675f4 100644
> --- a/xen/common/lzo.c
> +++ b/xen/common/lzo.c
> @@ -96,7 +96,7 @@
>
> #ifdef __XEN__
> #include <xen/lib.h>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <asm/unaligned.h>
> #else
> #define get_unaligned_le16(_p) (*(u16 *)(_p))
> diff --git a/xen/common/unlzo.c b/xen/common/unlzo.c
> index 74056778eb..f908d2a61f 100644
> --- a/xen/common/unlzo.c
> +++ b/xen/common/unlzo.c
> @@ -33,7 +33,7 @@
> #include <xen/lzo.h>
>
> #ifdef __XEN__
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <asm/unaligned.h>
> #else
>
> diff --git a/xen/common/xz/private.h b/xen/common/xz/private.h
> index 511343fcc2..647f9699a7 100644
> --- a/xen/common/xz/private.h
> +++ b/xen/common/xz/private.h
> @@ -12,7 +12,7 @@
>
> #ifdef __XEN__
> #include <xen/kernel.h>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
> #include <asm/unaligned.h>
> #else
>
> @@ -28,7 +28,7 @@ static inline void put_unaligned_le32(u32 val, void *p)
>
> #endif
>
> -#define get_le32(p) le32_to_cpup((const uint32_t *)(p))
> +#define get_le32(p) le32_to_cpu(*(const uint32_t *)(p))
>
> #define false 0
> #define true 1
> diff --git a/xen/drivers/char/ehci-dbgp.c b/xen/drivers/char/ehci-dbgp.c
> index c893d246de..8412da1b11 100644
> --- a/xen/drivers/char/ehci-dbgp.c
> +++ b/xen/drivers/char/ehci-dbgp.c
> @@ -5,13 +5,13 @@
> * Linux; see the Linux source for authorship and copyrights.
> */
>
> +#include <xen/byteswap.h>
> #include <xen/console.h>
> #include <xen/delay.h>
> #include <xen/errno.h>
> #include <xen/param.h>
> #include <xen/pci.h>
> #include <xen/serial.h>
> -#include <asm/byteorder.h>
> #include <asm/io.h>
> #include <asm/fixmap.h>
> #include <public/physdev.h>
> diff --git a/xen/include/asm-x86/msi.h b/xen/include/asm-x86/msi.h
> index e228b0f3f3..277375183c 100644
> --- a/xen/include/asm-x86/msi.h
> +++ b/xen/include/asm-x86/msi.h
> @@ -1,9 +1,9 @@
> #ifndef __ASM_MSI_H
> #define __ASM_MSI_H
>
> +#include <xen/byteswap.h>
> #include <xen/cpumask.h>
> #include <xen/pci.h>
> -#include <asm/byteorder.h>
> #include <asm/hvm/vmx/vmcs.h>
>
> /*
> diff --git a/xen/include/xen/bitmap.h b/xen/include/xen/bitmap.h
> index e9175ab54a..c44a1cb63c 100644
> --- a/xen/include/xen/bitmap.h
> +++ b/xen/include/xen/bitmap.h
> @@ -229,7 +229,7 @@ static inline int bitmap_weight(const unsigned long *src, int nbits)
> return __bitmap_weight(src, nbits);
> }
>
> -#include <asm/byteorder.h>
> +#include <xen/byteswap.h>
>
> #ifdef __LITTLE_ENDIAN
> #define BITMAP_MEM_ALIGNMENT 8
> diff --git a/xen/include/xen/device_tree.h b/xen/include/xen/device_tree.h
> index fd6cd00b43..4921e6b142 100644
> --- a/xen/include/xen/device_tree.h
> +++ b/xen/include/xen/device_tree.h
> @@ -10,10 +10,10 @@
> #ifndef __XEN_DEVICE_TREE_H__
> #define __XEN_DEVICE_TREE_H__
>
> -#include <asm/byteorder.h>
> #include <asm/device.h>
> #include <public/xen.h>
> #include <public/device_tree_defs.h>
> +#include <xen/byteswap.h>
> #include <xen/kernel.h>
> #include <xen/string.h>
> #include <xen/types.h>
> diff --git a/xen/include/xen/unaligned.h b/xen/include/xen/unaligned.h
> index 0a2b16d05d..8a9ec8a0ac 100644
> --- a/xen/include/xen/unaligned.h
> +++ b/xen/include/xen/unaligned.h
> @@ -11,8 +11,8 @@
> #define __XEN_UNALIGNED_H__
>
> #ifdef __XEN__
> +#include <xen/byteswap.h>
> #include <xen/types.h>
> -#include <asm/byteorder.h>
> #endif
>
> #define get_unaligned(p) (*(p))
> @@ -20,7 +20,7 @@
>
> static inline uint16_t get_unaligned_be16(const void *p)
> {
> - return be16_to_cpup(p);
> + return be16_to_cpu(*(uint16_t*)p);
Hmm - this missed one of my pieces of internal feedback.
*(const uint16_t *)p
to get correct style and avoid casting away const.
Also, the put functions what to drop the __force and use typecasting
like this.
I can fix up on commit if there are no other concerns with the series.
~Andrew
^ permalink raw reply [flat|nested] 16+ messages in thread* Re: [PATCH v2 6/7] xen: Switch to byteswap.h
2021-10-22 10:47 ` [PATCH v2 6/7] xen: " Lin Liu
2021-10-22 11:25 ` Andrew Cooper
@ 2021-11-02 15:23 ` Jan Beulich
1 sibling, 0 replies; 16+ messages in thread
From: Jan Beulich @ 2021-11-02 15:23 UTC (permalink / raw)
To: Lin Liu
Cc: Andrew Cooper, George Dunlap, Ian Jackson, Julien Grall,
Stefano Stabellini, Wei Liu, Roger Pau Monné,
xen-devel
On 22.10.2021 12:47, Lin Liu wrote:
> @@ -28,7 +28,7 @@ static inline void put_unaligned_le32(u32 val, void *p)
>
> #endif
>
> -#define get_le32(p) le32_to_cpup((const uint32_t *)(p))
> +#define get_le32(p) le32_to_cpu(*(const uint32_t *)(p))
While here the adjustment may be okay as there is already an
open-coded cast, ...
> @@ -20,7 +20,7 @@
>
> static inline uint16_t get_unaligned_be16(const void *p)
> {
> - return be16_to_cpup(p);
> + return be16_to_cpu(*(uint16_t*)p);
... here and below you add open-coded casts. Is there a reason you
don't retain the ..._to_cpup() macros?
Jan
^ permalink raw reply [flat|nested] 16+ messages in thread
* [PATCH v2 7/7] byteorder: Remove byteorder
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
` (5 preceding siblings ...)
2021-10-22 10:47 ` [PATCH v2 6/7] xen: " Lin Liu
@ 2021-10-22 10:47 ` Lin Liu
2021-10-22 11:50 ` [PATCH v2 0/7] Implement byteswap and update references Andrew Cooper
7 siblings, 0 replies; 16+ messages in thread
From: Lin Liu @ 2021-10-22 10:47 UTC (permalink / raw)
To: xen-devel
Cc: Lin Liu, Stefano Stabellini, Julien Grall, Volodymyr Babchuk,
Andrew Cooper, George Dunlap, Ian Jackson, Jan Beulich, Wei Liu,
Roger Pau Monné
include/xen/byteswap.h has simplify the interface, just clean
the old interface
No functional change
Signed-off-by: Lin Liu <lin.liu@citrix.com>
---
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Julien Grall <julien@xen.org>
Cc: Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: George Dunlap <george.dunlap@citrix.com>
Cc: Ian Jackson <iwj@xenproject.org>
Cc: Jan Beulich <jbeulich@suse.com>
Cc: Wei Liu <wl@xen.org>
Cc: "Roger Pau Monné" <roger.pau@citrix.com>
---
xen/include/asm-arm/byteorder.h | 16 --
xen/include/asm-x86/byteorder.h | 36 -----
xen/include/xen/byteorder/big_endian.h | 102 ------------
xen/include/xen/byteorder/generic.h | 68 --------
xen/include/xen/byteorder/little_endian.h | 102 ------------
xen/include/xen/byteorder/swab.h | 183 ----------------------
6 files changed, 507 deletions(-)
delete mode 100644 xen/include/asm-arm/byteorder.h
delete mode 100644 xen/include/asm-x86/byteorder.h
delete mode 100644 xen/include/xen/byteorder/big_endian.h
delete mode 100644 xen/include/xen/byteorder/generic.h
delete mode 100644 xen/include/xen/byteorder/little_endian.h
delete mode 100644 xen/include/xen/byteorder/swab.h
diff --git a/xen/include/asm-arm/byteorder.h b/xen/include/asm-arm/byteorder.h
deleted file mode 100644
index 9c712c4788..0000000000
--- a/xen/include/asm-arm/byteorder.h
+++ /dev/null
@@ -1,16 +0,0 @@
-#ifndef __ASM_ARM_BYTEORDER_H__
-#define __ASM_ARM_BYTEORDER_H__
-
-#define __BYTEORDER_HAS_U64__
-
-#include <xen/byteorder/little_endian.h>
-
-#endif /* __ASM_ARM_BYTEORDER_H__ */
-/*
- * Local variables:
- * mode: C
- * c-file-style: "BSD"
- * c-basic-offset: 4
- * indent-tabs-mode: nil
- * End:
- */
diff --git a/xen/include/asm-x86/byteorder.h b/xen/include/asm-x86/byteorder.h
deleted file mode 100644
index 1f77e502a5..0000000000
--- a/xen/include/asm-x86/byteorder.h
+++ /dev/null
@@ -1,36 +0,0 @@
-#ifndef __ASM_X86_BYTEORDER_H__
-#define __ASM_X86_BYTEORDER_H__
-
-#include <asm/types.h>
-#include <xen/compiler.h>
-
-static inline __attribute_const__ __u32 ___arch__swab32(__u32 x)
-{
- asm("bswap %0" : "=r" (x) : "0" (x));
- return x;
-}
-
-static inline __attribute_const__ __u64 ___arch__swab64(__u64 val)
-{
- union {
- struct { __u32 a,b; } s;
- __u64 u;
- } v;
- v.u = val;
- asm("bswapl %0 ; bswapl %1 ; xchgl %0,%1"
- : "=r" (v.s.a), "=r" (v.s.b)
- : "0" (v.s.a), "1" (v.s.b));
- return v.u;
-}
-
-/* Do not define swab16. Gcc is smart enough to recognize "C" version and
- convert it into rotation or exhange. */
-
-#define __arch__swab64(x) ___arch__swab64(x)
-#define __arch__swab32(x) ___arch__swab32(x)
-
-#define __BYTEORDER_HAS_U64__
-
-#include <xen/byteorder/little_endian.h>
-
-#endif /* __ASM_X86_BYTEORDER_H__ */
diff --git a/xen/include/xen/byteorder/big_endian.h b/xen/include/xen/byteorder/big_endian.h
deleted file mode 100644
index 40eb80a390..0000000000
--- a/xen/include/xen/byteorder/big_endian.h
+++ /dev/null
@@ -1,102 +0,0 @@
-#ifndef __XEN_BYTEORDER_BIG_ENDIAN_H__
-#define __XEN_BYTEORDER_BIG_ENDIAN_H__
-
-#ifndef __BIG_ENDIAN
-#define __BIG_ENDIAN 4321
-#endif
-#ifndef __BIG_ENDIAN_BITFIELD
-#define __BIG_ENDIAN_BITFIELD
-#endif
-
-#include <xen/types.h>
-#include <xen/byteorder/swab.h>
-
-#define __constant_cpu_to_le64(x) ((__force __le64)___constant_swab64((x)))
-#define __constant_le64_to_cpu(x) ___constant_swab64((__force __u64)(__le64)(x))
-#define __constant_cpu_to_le32(x) ((__force __le32)___constant_swab32((x)))
-#define __constant_le32_to_cpu(x) ___constant_swab32((__force __u32)(__le32)(x))
-#define __constant_cpu_to_le16(x) ((__force __le16)___constant_swab16((x)))
-#define __constant_le16_to_cpu(x) ___constant_swab16((__force __u16)(__le16)(x))
-#define __constant_cpu_to_be64(x) ((__force __be64)(__u64)(x))
-#define __constant_be64_to_cpu(x) ((__force __u64)(__be64)(x))
-#define __constant_cpu_to_be32(x) ((__force __be32)(__u32)(x))
-#define __constant_be32_to_cpu(x) ((__force __u32)(__be32)(x))
-#define __constant_cpu_to_be16(x) ((__force __be16)(__u16)(x))
-#define __constant_be16_to_cpu(x) ((__force __u16)(__be16)(x))
-#define __cpu_to_le64(x) ((__force __le64)__swab64((x)))
-#define __le64_to_cpu(x) __swab64((__force __u64)(__le64)(x))
-#define __cpu_to_le32(x) ((__force __le32)__swab32((x)))
-#define __le32_to_cpu(x) __swab32((__force __u32)(__le32)(x))
-#define __cpu_to_le16(x) ((__force __le16)__swab16((x)))
-#define __le16_to_cpu(x) __swab16((__force __u16)(__le16)(x))
-#define __cpu_to_be64(x) ((__force __be64)(__u64)(x))
-#define __be64_to_cpu(x) ((__force __u64)(__be64)(x))
-#define __cpu_to_be32(x) ((__force __be32)(__u32)(x))
-#define __be32_to_cpu(x) ((__force __u32)(__be32)(x))
-#define __cpu_to_be16(x) ((__force __be16)(__u16)(x))
-#define __be16_to_cpu(x) ((__force __u16)(__be16)(x))
-
-static inline __le64 __cpu_to_le64p(const __u64 *p)
-{
- return (__force __le64)__swab64p(p);
-}
-static inline __u64 __le64_to_cpup(const __le64 *p)
-{
- return __swab64p((__u64 *)p);
-}
-static inline __le32 __cpu_to_le32p(const __u32 *p)
-{
- return (__force __le32)__swab32p(p);
-}
-static inline __u32 __le32_to_cpup(const __le32 *p)
-{
- return __swab32p((__u32 *)p);
-}
-static inline __le16 __cpu_to_le16p(const __u16 *p)
-{
- return (__force __le16)__swab16p(p);
-}
-static inline __u16 __le16_to_cpup(const __le16 *p)
-{
- return __swab16p((__u16 *)p);
-}
-static inline __be64 __cpu_to_be64p(const __u64 *p)
-{
- return (__force __be64)*p;
-}
-static inline __u64 __be64_to_cpup(const __be64 *p)
-{
- return (__force __u64)*p;
-}
-static inline __be32 __cpu_to_be32p(const __u32 *p)
-{
- return (__force __be32)*p;
-}
-static inline __u32 __be32_to_cpup(const __be32 *p)
-{
- return (__force __u32)*p;
-}
-static inline __be16 __cpu_to_be16p(const __u16 *p)
-{
- return (__force __be16)*p;
-}
-static inline __u16 __be16_to_cpup(const __be16 *p)
-{
- return (__force __u16)*p;
-}
-#define __cpu_to_le64s(x) __swab64s((x))
-#define __le64_to_cpus(x) __swab64s((x))
-#define __cpu_to_le32s(x) __swab32s((x))
-#define __le32_to_cpus(x) __swab32s((x))
-#define __cpu_to_le16s(x) __swab16s((x))
-#define __le16_to_cpus(x) __swab16s((x))
-#define __cpu_to_be64s(x) do {} while (0)
-#define __be64_to_cpus(x) do {} while (0)
-#define __cpu_to_be32s(x) do {} while (0)
-#define __be32_to_cpus(x) do {} while (0)
-#define __cpu_to_be16s(x) do {} while (0)
-#define __be16_to_cpus(x) do {} while (0)
-
-#include <xen/byteorder/generic.h>
-
-#endif /* __XEN_BYTEORDER_BIG_ENDIAN_H__ */
diff --git a/xen/include/xen/byteorder/generic.h b/xen/include/xen/byteorder/generic.h
deleted file mode 100644
index 8a0006b755..0000000000
--- a/xen/include/xen/byteorder/generic.h
+++ /dev/null
@@ -1,68 +0,0 @@
-#ifndef __XEN_BYTEORDER_GENERIC_H__
-#define __XEN_BYTEORDER_GENERIC_H__
-
-/*
- * Generic Byte-reordering support
- *
- * The "... p" macros, like le64_to_cpup, can be used with pointers
- * to unaligned data, but there will be a performance penalty on
- * some architectures. Use get_unaligned for unaligned data.
- *
- * The following macros are to be defined by <asm/byteorder.h>:
- *
- * Conversion of XX-bit integers (16- 32- or 64-)
- * between native CPU format and little/big endian format
- * 64-bit stuff only defined for proper architectures
- * cpu_to_[bl]eXX(__uXX x)
- * [bl]eXX_to_cpu(__uXX x)
- *
- * The same, but takes a pointer to the value to convert
- * cpu_to_[bl]eXXp(__uXX x)
- * [bl]eXX_to_cpup(__uXX x)
- *
- * The same, but change in situ
- * cpu_to_[bl]eXXs(__uXX x)
- * [bl]eXX_to_cpus(__uXX x)
- *
- * See asm-foo/byteorder.h for examples of how to provide
- * architecture-optimized versions
- */
-
-#define cpu_to_le64 __cpu_to_le64
-#define le64_to_cpu __le64_to_cpu
-#define cpu_to_le32 __cpu_to_le32
-#define le32_to_cpu __le32_to_cpu
-#define cpu_to_le16 __cpu_to_le16
-#define le16_to_cpu __le16_to_cpu
-#define cpu_to_be64 __cpu_to_be64
-#define be64_to_cpu __be64_to_cpu
-#define cpu_to_be32 __cpu_to_be32
-#define be32_to_cpu __be32_to_cpu
-#define cpu_to_be16 __cpu_to_be16
-#define be16_to_cpu __be16_to_cpu
-#define cpu_to_le64p __cpu_to_le64p
-#define le64_to_cpup __le64_to_cpup
-#define cpu_to_le32p __cpu_to_le32p
-#define le32_to_cpup __le32_to_cpup
-#define cpu_to_le16p __cpu_to_le16p
-#define le16_to_cpup __le16_to_cpup
-#define cpu_to_be64p __cpu_to_be64p
-#define be64_to_cpup __be64_to_cpup
-#define cpu_to_be32p __cpu_to_be32p
-#define be32_to_cpup __be32_to_cpup
-#define cpu_to_be16p __cpu_to_be16p
-#define be16_to_cpup __be16_to_cpup
-#define cpu_to_le64s __cpu_to_le64s
-#define le64_to_cpus __le64_to_cpus
-#define cpu_to_le32s __cpu_to_le32s
-#define le32_to_cpus __le32_to_cpus
-#define cpu_to_le16s __cpu_to_le16s
-#define le16_to_cpus __le16_to_cpus
-#define cpu_to_be64s __cpu_to_be64s
-#define be64_to_cpus __be64_to_cpus
-#define cpu_to_be32s __cpu_to_be32s
-#define be32_to_cpus __be32_to_cpus
-#define cpu_to_be16s __cpu_to_be16s
-#define be16_to_cpus __be16_to_cpus
-
-#endif /* __XEN_BYTEORDER_GENERIC_H__ */
diff --git a/xen/include/xen/byteorder/little_endian.h b/xen/include/xen/byteorder/little_endian.h
deleted file mode 100644
index 4955632793..0000000000
--- a/xen/include/xen/byteorder/little_endian.h
+++ /dev/null
@@ -1,102 +0,0 @@
-#ifndef __XEN_BYTEORDER_LITTLE_ENDIAN_H__
-#define __XEN_BYTEORDER_LITTLE_ENDIAN_H__
-
-#ifndef __LITTLE_ENDIAN
-#define __LITTLE_ENDIAN 1234
-#endif
-#ifndef __LITTLE_ENDIAN_BITFIELD
-#define __LITTLE_ENDIAN_BITFIELD
-#endif
-
-#include <xen/types.h>
-#include <xen/byteorder/swab.h>
-
-#define __constant_cpu_to_le64(x) ((__force __le64)(__u64)(x))
-#define __constant_le64_to_cpu(x) ((__force __u64)(__le64)(x))
-#define __constant_cpu_to_le32(x) ((__force __le32)(__u32)(x))
-#define __constant_le32_to_cpu(x) ((__force __u32)(__le32)(x))
-#define __constant_cpu_to_le16(x) ((__force __le16)(__u16)(x))
-#define __constant_le16_to_cpu(x) ((__force __u16)(__le16)(x))
-#define __constant_cpu_to_be64(x) ((__force __be64)___constant_swab64((x)))
-#define __constant_be64_to_cpu(x) ___constant_swab64((__force __u64)(__be64)(x))
-#define __constant_cpu_to_be32(x) ((__force __be32)___constant_swab32((x)))
-#define __constant_be32_to_cpu(x) ___constant_swab32((__force __u32)(__be32)(x))
-#define __constant_cpu_to_be16(x) ((__force __be16)___constant_swab16((x)))
-#define __constant_be16_to_cpu(x) ___constant_swab16((__force __u16)(__be16)(x))
-#define __cpu_to_le64(x) ((__force __le64)(__u64)(x))
-#define __le64_to_cpu(x) ((__force __u64)(__le64)(x))
-#define __cpu_to_le32(x) ((__force __le32)(__u32)(x))
-#define __le32_to_cpu(x) ((__force __u32)(__le32)(x))
-#define __cpu_to_le16(x) ((__force __le16)(__u16)(x))
-#define __le16_to_cpu(x) ((__force __u16)(__le16)(x))
-#define __cpu_to_be64(x) ((__force __be64)__swab64((x)))
-#define __be64_to_cpu(x) __swab64((__force __u64)(__be64)(x))
-#define __cpu_to_be32(x) ((__force __be32)__swab32((x)))
-#define __be32_to_cpu(x) __swab32((__force __u32)(__be32)(x))
-#define __cpu_to_be16(x) ((__force __be16)__swab16((x)))
-#define __be16_to_cpu(x) __swab16((__force __u16)(__be16)(x))
-
-static inline __le64 __cpu_to_le64p(const __u64 *p)
-{
- return (__force __le64)*p;
-}
-static inline __u64 __le64_to_cpup(const __le64 *p)
-{
- return (__force __u64)*p;
-}
-static inline __le32 __cpu_to_le32p(const __u32 *p)
-{
- return (__force __le32)*p;
-}
-static inline __u32 __le32_to_cpup(const __le32 *p)
-{
- return (__force __u32)*p;
-}
-static inline __le16 __cpu_to_le16p(const __u16 *p)
-{
- return (__force __le16)*p;
-}
-static inline __u16 __le16_to_cpup(const __le16 *p)
-{
- return (__force __u16)*p;
-}
-static inline __be64 __cpu_to_be64p(const __u64 *p)
-{
- return (__force __be64)__swab64p(p);
-}
-static inline __u64 __be64_to_cpup(const __be64 *p)
-{
- return __swab64p((__u64 *)p);
-}
-static inline __be32 __cpu_to_be32p(const __u32 *p)
-{
- return (__force __be32)__swab32p(p);
-}
-static inline __u32 __be32_to_cpup(const __be32 *p)
-{
- return __swab32p((__u32 *)p);
-}
-static inline __be16 __cpu_to_be16p(const __u16 *p)
-{
- return (__force __be16)__swab16p(p);
-}
-static inline __u16 __be16_to_cpup(const __be16 *p)
-{
- return __swab16p((__u16 *)p);
-}
-#define __cpu_to_le64s(x) do {} while (0)
-#define __le64_to_cpus(x) do {} while (0)
-#define __cpu_to_le32s(x) do {} while (0)
-#define __le32_to_cpus(x) do {} while (0)
-#define __cpu_to_le16s(x) do {} while (0)
-#define __le16_to_cpus(x) do {} while (0)
-#define __cpu_to_be64s(x) __swab64s((x))
-#define __be64_to_cpus(x) __swab64s((x))
-#define __cpu_to_be32s(x) __swab32s((x))
-#define __be32_to_cpus(x) __swab32s((x))
-#define __cpu_to_be16s(x) __swab16s((x))
-#define __be16_to_cpus(x) __swab16s((x))
-
-#include <xen/byteorder/generic.h>
-
-#endif /* __XEN_BYTEORDER_LITTLE_ENDIAN_H__ */
diff --git a/xen/include/xen/byteorder/swab.h b/xen/include/xen/byteorder/swab.h
deleted file mode 100644
index b7e30f0503..0000000000
--- a/xen/include/xen/byteorder/swab.h
+++ /dev/null
@@ -1,183 +0,0 @@
-#ifndef __XEN_BYTEORDER_SWAB_H__
-#define __XEN_BYTEORDER_SWAB_H__
-
-/*
- * Byte-swapping, independently from CPU endianness
- * swabXX[ps]?(foo)
- *
- * Francois-Rene Rideau <fare@tunes.org> 19971205
- * separated swab functions from cpu_to_XX,
- * to clean up support for bizarre-endian architectures.
- */
-
-/* casts are necessary for constants, because we never know how for sure
- * how U/UL/ULL map to __u16, __u32, __u64. At least not in a portable way.
- */
-#define ___swab16(x) \
-({ \
- __u16 __x = (x); \
- ((__u16)( \
- (((__u16)(__x) & (__u16)0x00ffU) << 8) | \
- (((__u16)(__x) & (__u16)0xff00U) >> 8) )); \
-})
-
-#define ___swab32(x) \
-({ \
- __u32 __x = (x); \
- ((__u32)( \
- (((__u32)(__x) & (__u32)0x000000ffUL) << 24) | \
- (((__u32)(__x) & (__u32)0x0000ff00UL) << 8) | \
- (((__u32)(__x) & (__u32)0x00ff0000UL) >> 8) | \
- (((__u32)(__x) & (__u32)0xff000000UL) >> 24) )); \
-})
-
-#define ___swab64(x) \
-({ \
- __u64 __x = (x); \
- ((__u64)( \
- (__u64)(((__u64)(__x) & (__u64)0x00000000000000ffULL) << 56) | \
- (__u64)(((__u64)(__x) & (__u64)0x000000000000ff00ULL) << 40) | \
- (__u64)(((__u64)(__x) & (__u64)0x0000000000ff0000ULL) << 24) | \
- (__u64)(((__u64)(__x) & (__u64)0x00000000ff000000ULL) << 8) | \
- (__u64)(((__u64)(__x) & (__u64)0x000000ff00000000ULL) >> 8) | \
- (__u64)(((__u64)(__x) & (__u64)0x0000ff0000000000ULL) >> 24) | \
- (__u64)(((__u64)(__x) & (__u64)0x00ff000000000000ULL) >> 40) | \
- (__u64)(((__u64)(__x) & (__u64)0xff00000000000000ULL) >> 56) )); \
-})
-
-#define ___constant_swab16(x) \
- ((__u16)( \
- (((__u16)(x) & (__u16)0x00ffU) << 8) | \
- (((__u16)(x) & (__u16)0xff00U) >> 8) ))
-#define ___constant_swab32(x) \
- ((__u32)( \
- (((__u32)(x) & (__u32)0x000000ffUL) << 24) | \
- (((__u32)(x) & (__u32)0x0000ff00UL) << 8) | \
- (((__u32)(x) & (__u32)0x00ff0000UL) >> 8) | \
- (((__u32)(x) & (__u32)0xff000000UL) >> 24) ))
-#define ___constant_swab64(x) \
- ((__u64)( \
- (__u64)(((__u64)(x) & (__u64)0x00000000000000ffULL) << 56) | \
- (__u64)(((__u64)(x) & (__u64)0x000000000000ff00ULL) << 40) | \
- (__u64)(((__u64)(x) & (__u64)0x0000000000ff0000ULL) << 24) | \
- (__u64)(((__u64)(x) & (__u64)0x00000000ff000000ULL) << 8) | \
- (__u64)(((__u64)(x) & (__u64)0x000000ff00000000ULL) >> 8) | \
- (__u64)(((__u64)(x) & (__u64)0x0000ff0000000000ULL) >> 24) | \
- (__u64)(((__u64)(x) & (__u64)0x00ff000000000000ULL) >> 40) | \
- (__u64)(((__u64)(x) & (__u64)0xff00000000000000ULL) >> 56) ))
-
-/*
- * provide defaults when no architecture-specific optimization is detected
- */
-#ifndef __arch__swab16
-# define __arch__swab16(x) ({ __u16 __tmp = (x) ; ___swab16(__tmp); })
-#endif
-#ifndef __arch__swab32
-# define __arch__swab32(x) ({ __u32 __tmp = (x) ; ___swab32(__tmp); })
-#endif
-#ifndef __arch__swab64
-# define __arch__swab64(x) ({ __u64 __tmp = (x) ; ___swab64(__tmp); })
-#endif
-
-#ifndef __arch__swab16p
-# define __arch__swab16p(x) __arch__swab16(*(x))
-#endif
-#ifndef __arch__swab32p
-# define __arch__swab32p(x) __arch__swab32(*(x))
-#endif
-#ifndef __arch__swab64p
-# define __arch__swab64p(x) __arch__swab64(*(x))
-#endif
-
-#ifndef __arch__swab16s
-# define __arch__swab16s(x) do { *(x) = __arch__swab16p((x)); } while (0)
-#endif
-#ifndef __arch__swab32s
-# define __arch__swab32s(x) do { *(x) = __arch__swab32p((x)); } while (0)
-#endif
-#ifndef __arch__swab64s
-# define __arch__swab64s(x) do { *(x) = __arch__swab64p((x)); } while (0)
-#endif
-
-
-/*
- * Allow constant folding
- */
-#if defined(__GNUC__) && defined(__OPTIMIZE__)
-# define __swab16(x) \
-(__builtin_constant_p((__u16)(x)) ? \
- ___swab16((x)) : \
- __fswab16((x)))
-# define __swab32(x) \
-(__builtin_constant_p((__u32)(x)) ? \
- ___swab32((x)) : \
- __fswab32((x)))
-# define __swab64(x) \
-(__builtin_constant_p((__u64)(x)) ? \
- ___swab64((x)) : \
- __fswab64((x)))
-#else
-# define __swab16(x) __fswab16(x)
-# define __swab32(x) __fswab32(x)
-# define __swab64(x) __fswab64(x)
-#endif /* OPTIMIZE */
-
-
-static inline __attribute_const__ __u16 __fswab16(__u16 x)
-{
- return __arch__swab16(x);
-}
-static inline __u16 __swab16p(const __u16 *x)
-{
- return __arch__swab16p(x);
-}
-static inline void __swab16s(__u16 *addr)
-{
- __arch__swab16s(addr);
-}
-
-static inline __attribute_const__ __u32 __fswab32(__u32 x)
-{
- return __arch__swab32(x);
-}
-static inline __u32 __swab32p(const __u32 *x)
-{
- return __arch__swab32p(x);
-}
-static inline void __swab32s(__u32 *addr)
-{
- __arch__swab32s(addr);
-}
-
-#ifdef __BYTEORDER_HAS_U64__
-static inline __attribute_const__ __u64 __fswab64(__u64 x)
-{
-# ifdef __SWAB_64_THRU_32__
- __u32 h = x >> 32;
- __u32 l = x & ((1ULL<<32)-1);
- return (((__u64)__swab32(l)) << 32) | ((__u64)(__swab32(h)));
-# else
- return __arch__swab64(x);
-# endif
-}
-static inline __u64 __swab64p(const __u64 *x)
-{
- return __arch__swab64p(x);
-}
-static inline void __swab64s(__u64 *addr)
-{
- __arch__swab64s(addr);
-}
-#endif /* __BYTEORDER_HAS_U64__ */
-
-#define swab16 __swab16
-#define swab32 __swab32
-#define swab64 __swab64
-#define swab16p __swab16p
-#define swab32p __swab32p
-#define swab64p __swab64p
-#define swab16s __swab16s
-#define swab32s __swab32s
-#define swab64s __swab64s
-
-#endif /* __XEN_BYTEORDER_SWAB_H__ */
--
2.27.0
^ permalink raw reply related [flat|nested] 16+ messages in thread* Re: [PATCH v2 0/7] Implement byteswap and update references
2021-10-22 10:47 [PATCH v2 0/7] Implement byteswap and update references Lin Liu
` (6 preceding siblings ...)
2021-10-22 10:47 ` [PATCH v2 7/7] byteorder: Remove byteorder Lin Liu
@ 2021-10-22 11:50 ` Andrew Cooper
7 siblings, 0 replies; 16+ messages in thread
From: Andrew Cooper @ 2021-10-22 11:50 UTC (permalink / raw)
To: Lin Liu, xen-devel
Cc: Daniel De Graaf, Daniel P. Smith, George Dunlap, Ian Jackson,
Jan Beulich, Julien Grall, Konrad Rzeszutek Wilk,
Roger Pau Monné,
Ross Lagerwall, Stefano Stabellini, Volodymyr Babchuk, Wei Liu
On 22/10/2021 11:47, Lin Liu wrote:
> The swab() is massively over complicated
> Simplify it with compiler builtins and fallback to plain C function
> if undefined.
> Update components to switch to this new swap bytes.
>
> <snip>
> 34 files changed, 150 insertions(+), 646 deletions(-)
It is worth saying a couple of things.
x86's ___arch__swab64 is wrong. Well - it was mostly ok for 32bit
builds of Xen, and is not ok for 64bit builds. As a consequence, this
series nets an improvement of:
$ ../scripts/bloat-o-meter xen-syms-before xen-syms-after
add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-54 (-54)
Function old new delta
elf_access_unsigned 173 151 -22
unlzo 1128 1096 -32
Total: Before=3803328, After=3803274, chg -0.00%
because the code generation for bswap64 goes from:
ffff82d0402059b0 <_bswap64>:
ffff82d0402059b0: 48 89 f8 mov %rdi,%rax
ffff82d0402059b3: 48 c1 e8 20 shr $0x20,%rax
ffff82d0402059b7: 0f cf bswap %edi
ffff82d0402059b9: 0f c8 bswap %eax
ffff82d0402059bb: 97 xchg %eax,%edi
ffff82d0402059bc: 48 89 c2 mov %rax,%rdx
ffff82d0402059bf: 89 f8 mov %edi,%eax
ffff82d0402059c1: 48 c1 e2 20 shl $0x20,%rdx
ffff82d0402059c5: 48 09 d0 or %rdx,%rax
ffff82d0402059c8: c3 retq
to
ffff82d0402059b0 <_bswap64>:
ffff82d0402059b0: 48 89 f8 mov %rdi,%rax
ffff82d0402059b3: 48 0f c8 bswap %rax
ffff82d0402059b6: c3 retq
Almost all byteswapping is done on 32bit quantities, not 64, which is
why the delta is so small.
However, it also drops 500 lines of code, which is a demonstration of
how silly the swab() infrastructure was. It also removes the need for
per-arch code to do any of this.
I'd say its safe to go into 4.16, but I'll understand if others want to
push back on that at this point in the release.
~Andrew
^ permalink raw reply [flat|nested] 16+ messages in thread