[PATCH 00/18] scripts: enable compilation on macOS

[PATCH 00/18] scripts: enable compilation on macOS

[Ahmad Fatoum]

With this patch series applied, imx_v7_defconfig along with the host
tools can be built on macOS. It should also pave the way for building
on the BSDs. Currently it still fails on NetBSD though:

	In file included from scripts/mod/modpost.c:15:0:
	scripts/mod/modpost.h:16:0: warning: "Elf_Ehdr" redefined
	 #define Elf_Ehdr    Elf32_Ehdr
	 ^
	In file included from scripts/mod/modpost.h:10:0,
			 from scripts/mod/modpost.c:15:
	/usr/include/elf.h:1093:0: note: this is the location of the previous definition
	 #define Elf_Ehdr Elf64_Ehdr
	 ^

Issues addressed in the series can be split into three groups:

- host tools depend on glibc or Linux headers
	replaced where appropriate with standard ISO/POSIX functionality
- build depends on GNU utilites:
	only stat(1). fixed by importing the linux patch that uses ls
	instead.
- endianness helpers
	These aren't (yet) POSIX-standardized, so I extended compiler.h
	to support macOS and the BSDs.

Only thing I can't yet compile is MXS_HOSTTOOLS, which depends on
pkg-config finding OpenSSL, but I guess that's something I need to fix
at my side.


Ahmad Fatoum (16):
  scripts: extract symbol offsets using target, not host, nm
  scripts: includes: restrict strlcpy prototype to glibc
  scripts: compiler.h: don't include <stdint.h> twice
  scripts: don't depend on system <asm/types.h>
  scripts: bareboxcrc32: remove usage of loff_t
  scripts: bareboximd: remove usage of loff_t
  scripts: compiler.h: add endianness helpers for macOS
  scripts: use "compiler.h" endianness helpers
  scripts: mkimage: s/fdatasync/fsync/ on macOS & OpenBSD
  scripts: removes uses of <asm*/errno.h> in favor of <errno.h>
  scripts: omap3-usb-loader: don't depend on unportable le32toh
  scripts: omap3-usb-loader: drop unneeded header
  scripts: imx-usb-loader: don't depend on unportable headers
  scripts: kwbimage: fix build with non-glibc systems
  scripts: compiler.h: use Linux <endian.h> as default
  scripts: compiler.h: support BSDs as well

Masahiro Yamada (1):
  kbuild: suppress warnings from 'getconf LFS_*'

Michael Forney (1):
  kbuild: Use ls(1) instead of stat(1) to obtain file size

 Makefile                       |  6 +++---
 defaultenv/Makefile            |  3 ++-
 scripts/Makefile.lib           |  4 ++--
 scripts/bareboxcrc32.c         |  2 +-
 scripts/bareboximd.c           |  3 +--
 scripts/check_size             |  2 +-
 scripts/compiler.h             | 36 ++++++++++++++++++++++++++--------
 scripts/extract_symbol_offset  |  2 +-
 scripts/file-size.sh           |  4 ++++
 scripts/fix_size.c             |  3 ++-
 scripts/gen-dtb-s              |  4 ++--
 scripts/imx/imx-image.c        |  2 +-
 scripts/imx/imx-usb-loader.c   |  3 +--
 scripts/include/linux/err.h    |  2 +-
 scripts/include/linux/string.h |  6 +++++-
 scripts/include/linux/types.h  | 14 +++++++++++--
 scripts/kwbimage.c             |  8 ++++++--
 scripts/mkimage.c              | 12 ++++++++++--
 scripts/omap3-usb-loader.c     |  6 +-----
 19 files changed, 84 insertions(+), 38 deletions(-)
 create mode 100755 scripts/file-size.sh

-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 01/18] kbuild: suppress warnings from 'getconf LFS_*'

[Ahmad Fatoum]

From: Masahiro Yamada <yamada.masahiro@socionext.com>

Suppress warnings for systems that do not recognize LFS_*.

 getconf: no such configuration parameter `LFS_CFLAGS'
 getconf: no such configuration parameter `LFS_LDFLAGS'
 getconf: no such configuration parameter `LFS_LIBS'

Fixes: d7f14c66c273 ("kbuild: Enable Large File Support for hostprogs")
Reported-by: Chen Feng <puck.chen@hisilicon.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 Makefile | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/Makefile b/Makefile
index 29347b67803c..c09a58026af6 100644
--- a/Makefile
+++ b/Makefile
@@ -177,9 +177,9 @@ CONFIG_SHELL := $(shell if [ -x "$$BASH" ]; then echo $$BASH; \
 	  else if [ -x /bin/bash ]; then echo /bin/bash; \
 	  else echo sh; fi ; fi)
 
-HOST_LFS_CFLAGS := $(shell getconf LFS_CFLAGS)
-HOST_LFS_LDFLAGS := $(shell getconf LFS_LDFLAGS)
-HOST_LFS_LIBS := $(shell getconf LFS_LIBS)
+HOST_LFS_CFLAGS := $(shell getconf LFS_CFLAGS 2>/dev/null)
+HOST_LFS_LDFLAGS := $(shell getconf LFS_LDFLAGS 2>/dev/null)
+HOST_LFS_LIBS := $(shell getconf LFS_LIBS 2>/dev/null)
 
 HOSTCC       = gcc
 HOSTCXX      = g++
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 02/18] kbuild: Use ls(1) instead of stat(1) to obtain file size

[Ahmad Fatoum]

From: Michael Forney <forney@google.com>

stat(1) is not standardized and different implementations have their own
(conflicting) flags for querying the size of a file.

ls(1) provides the same information (value of st.st_size) in the 5th
column, except when the file is a character or block device. This output
is standardized[0]. The -n option turns on -l, which writes lines
formatted like

  "%s %u %s %s %u %s %s\n", <file mode>, <number of links>,
      <owner name>, <group name>, <size>, <date and time>,
      <pathname>

but instead of writing the <owner name> and <group name>, it writes the
numeric owner and group IDs (this avoids /etc/passwd and /etc/group
lookups as well as potential field splitting issues).

The <size> field is specified as "the value that would be returned for
the file in the st_size field of struct stat".

To avoid duplicating logic in several locations in the tree, create
scripts/file-size.sh and update callers to use that instead of stat(1).

[0] http://pubs.opengroup.org/onlinepubs/9699919799/utilities/ls.html#tag_20_73_10

Signed-off-by: Michael Forney <forney@google.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
[afa: imported script and adjusted barebox stat(1) callsites]
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 defaultenv/Makefile  | 3 ++-
 scripts/Makefile.lib | 4 ++--
 scripts/file-size.sh | 4 ++++
 scripts/gen-dtb-s    | 4 ++--
 4 files changed, 10 insertions(+), 5 deletions(-)
 create mode 100755 scripts/file-size.sh

diff --git a/defaultenv/Makefile b/defaultenv/Makefile
index 950ac29a3cee..e030355a4052 100644
--- a/defaultenv/Makefile
+++ b/defaultenv/Makefile
@@ -20,7 +20,8 @@ $(obj)/barebox_default_env: FORCE
 
 quiet_cmd_env_h = ENVH    $@
 cmd_env_h = cat $< | (cd $(obj) && $(objtree)/scripts/bin2c "__aligned(4) default_environment") > $@; \
-	echo "static const int default_environment_uncompress_size=`stat -c%s $(obj)/barebox_default_env`;" >> $@
+	echo "static const int default_environment_uncompress_size=`${CONFIG_SHELL} \"${srctree}/scripts/file-size.sh\" $(obj)/barebox_default_env`;" >> $@
+
 
 $(obj)/barebox_default_env.h: $(obj)/barebox_default_env$(DEFAULT_COMPRESSION_SUFFIX) FORCE
 	$(call if_changed,env_h)
diff --git a/scripts/Makefile.lib b/scripts/Makefile.lib
index 95eaf522abc9..87bff2d296e3 100644
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -341,7 +341,7 @@ cmd_env=$(srctree)/scripts/genenv $(srctree) $(objtree) $@ $<
 size_append = printf $(shell						\
 dec_size=0;								\
 for F in $1; do								\
-	fsize=$$(stat -c "%s" $$F);					\
+	fsize=$$(${CONFIG_SHELL} "${srctree}/scripts/file-size.sh" $$F);\
 	dec_size=$$(expr $$dec_size + $$fsize);				\
 done;									\
 printf "%08x\n" $$dec_size |						\
@@ -446,7 +446,7 @@ quiet_cmd_check_size = CHKSIZE $2
 # Check size of a file
 quiet_cmd_check_file_size = CHKFILESIZE $2
       cmd_check_file_size = set -e;					\
-	size=`stat -c%s $2`;						\
+	size=`${CONFIG_SHELL} "${srctree}/scripts/file-size.sh" $2`;	\
 	max_size=`printf "%d" $3`;					\
 	if [ $$size -gt $$max_size ] ;					\
 	then								\
diff --git a/scripts/file-size.sh b/scripts/file-size.sh
new file mode 100755
index 000000000000..7eb7423416b5
--- /dev/null
+++ b/scripts/file-size.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+set -- $(ls -dn "$1")
+printf '%s\n' "$5"
diff --git a/scripts/gen-dtb-s b/scripts/gen-dtb-s
index 307b1f68667e..b2dd253c274f 100755
--- a/scripts/gen-dtb-s
+++ b/scripts/gen-dtb-s
@@ -55,8 +55,8 @@ lzop -f -9 $dtb -o $dtb.lzo
 if [ $? != 0 ]; then
 	exit 1
 fi
-compressed=$(stat $dtb.lzo -c "%s")
-uncompressed=$(stat $dtb -c "%s")
+compressed=$(${CONFIG_SHELL} "${srctree}/scripts/file-size.sh" $dtb.lzo)
+uncompressed=$(${CONFIG_SHELL} "${srctree}/scripts/file-size.sh" $dtb)
 
 echo ".section .dtbz.rodata.${name},\"a\""
 echo ".balign STRUCT_ALIGNMENT"
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 03/18] scripts: extract symbol offsets using target, not host, nm

[Ahmad Fatoum]

The nm(1) on other platforms may be compiled without ELF support,
thus use the cross toolchain's nm instead if possible.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/check_size            | 2 +-
 scripts/extract_symbol_offset | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/scripts/check_size b/scripts/check_size
index 54f02a63c5ef..8530435d3a1b 100755
--- a/scripts/check_size
+++ b/scripts/check_size
@@ -5,7 +5,7 @@ file="$2"
 max="$3"
 
 # extract symbol offset from file, remove leading zeros
-ofs=$(nm -t d $file | grep "$symbol" | cut -d ' ' -f1 | sed "s/^[0]*//")
+ofs=$(${CROSS_COMPILE}nm -t d $file | grep "$symbol" | cut -d ' ' -f1 | sed "s/^[0]*//")
 
 if [ -z "${ofs}" ]; then
 	echo "symbol $symbol not found in $file"
diff --git a/scripts/extract_symbol_offset b/scripts/extract_symbol_offset
index 1a1260f526e0..78b866830e56 100755
--- a/scripts/extract_symbol_offset
+++ b/scripts/extract_symbol_offset
@@ -4,7 +4,7 @@ symbol="$1"
 file="$2"
 
 # extract symbol offset from file, remove leading zeros
-ofs=$(nm -t d $file | grep "$symbol" | cut -d ' ' -f1 | sed "s/^[0]*//")
+ofs=$(${CROSS_COMPILE}nm -t d $file | grep "$symbol" | cut -d ' ' -f1 | sed "s/^[0]*//")
 
 if [ -z "${ofs}" ]; then
 	echo "symbol $symbol not found in $file"
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 04/18] scripts: includes: restrict strlcpy prototype to glibc

[Ahmad Fatoum]

Defining strlcpy here clashes on macOS, which has some compiler magic
around their strlcpy declaration. Fix this by inverting the check
to declare the prototype only if we are on __GLIBC__. As uClibc may
define __GLIBC__ as well, check for absence of its __UCLIBC__.

This is in-line with what Linux has been doing since 0215d59b15
("tools lib: Reinstate strlcpy() header guard with __UCLIBC__").

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/include/linux/string.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/scripts/include/linux/string.h b/scripts/include/linux/string.h
index e26223f1f287..649287b80a8c 100644
--- a/scripts/include/linux/string.h
+++ b/scripts/include/linux/string.h
@@ -8,7 +8,11 @@ void *memdup(const void *src, size_t len);
 
 int strtobool(const char *s, bool *res);
 
-#ifndef __UCLIBC__
+/*
+ * glibc based builds needs the extern while uClibc doesn't.
+ * However uClibc headers also define __GLIBC__ hence the hack below
+ */
+#if defined(__GLIBC) && !defined(__UCLIBC__)
 extern size_t strlcpy(char *dest, const char *src, size_t size);
 #endif
 
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 05/18] scripts: compiler.h: don't include <stdint.h> twice

[Ahmad Fatoum]

We already include it a few lines further down, so drop it at this
location.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/compiler.h | 2 --
 1 file changed, 2 deletions(-)

diff --git a/scripts/compiler.h b/scripts/compiler.h
index 4cf179bb9b13..bd73c7a9a8d3 100644
--- a/scripts/compiler.h
+++ b/scripts/compiler.h
@@ -13,8 +13,6 @@
     defined(__sun__)	 || \
     defined(__APPLE__)
 # include <inttypes.h>
-#elif defined(__linux__) || defined(__WIN32__) || defined(__MINGW32__)
-# include <stdint.h>
 #endif
 
 #include <errno.h>
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 06/18] scripts: don't depend on system <asm/types.h>

[Ahmad Fatoum]

If we compile on non-Linux hosts, we can't assume this header to be
available. Thus typedef loff_t and [su](8|16|32|64) ourselves.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/include/linux/types.h | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/scripts/include/linux/types.h b/scripts/include/linux/types.h
index 8ebf6278b2ef..e81d7e810126 100644
--- a/scripts/include/linux/types.h
+++ b/scripts/include/linux/types.h
@@ -5,8 +5,18 @@
 #include <stddef.h>
 #include <stdint.h>
 
-#define __SANE_USERSPACE_TYPES__	/* For PPC64, to get LL64 types */
-#include <asm/types.h>
+typedef uint64_t __u64;
+typedef int64_t __s64;
+typedef uint32_t __u32;
+typedef int32_t __s32;
+typedef uint16_t __u16;
+typedef int16_t __s16;
+typedef uint8_t __u8;
+typedef int8_t __s8;
+
+#ifndef __linux__
+typedef long long loff_t;
+#endif
 
 struct page;
 struct kmem_cache;
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 07/18] scripts: bareboxcrc32: remove usage of loff_t

[Ahmad Fatoum]

loff_t is Linux-specific and unneeded at this place, as it's directly
passed into a parameter that's a ulong anyway. Thus use a ulong.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/bareboxcrc32.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/bareboxcrc32.c b/scripts/bareboxcrc32.c
index e00ffafeb7b5..1b105a452bd2 100644
--- a/scripts/bareboxcrc32.c
+++ b/scripts/bareboxcrc32.c
@@ -38,7 +38,7 @@
 
 int main(int argc, char *argv[])
 {
-	loff_t start = 0, size = ~0;
+	ulong start = 0, size = ~0;
 	ulong crc = 0, total = 0;
 	char *filename = NULL;
 	int i;
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 08/18] scripts: bareboximd: remove usage of loff_t

[Ahmad Fatoum]

loff_t is Linux-specific and unneeded at this place, as sizes are usually
denoted with a size_t. As the signedness of max_size isn't relied on
anywhere, replace it with a size_t.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/bareboximd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/bareboximd.c b/scripts/bareboximd.c
index 7d4cbeb9a83f..81a59ec63c23 100644
--- a/scripts/bareboximd.c
+++ b/scripts/bareboximd.c
@@ -58,7 +58,7 @@ int imd_command_setenv(const char *variable_name, const char *value)
 	return -EINVAL;
 }
 
-static int read_file_2(const char *filename, size_t *size, void **outbuf, loff_t max_size)
+static int read_file_2(const char *filename, size_t *size, void **outbuf, size_t max_size)
 {
 	off_t fsize;
 	ssize_t rsize;
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 09/18] scripts: compiler.h: add endianness helpers for macOS

[Ahmad Fatoum]

These are taken from Mathias Panzenböck's portable_endian.h[1]
which he had released it into the public domain.

[1]: https://gist.github.com/panzi/6856583

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/compiler.h | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

diff --git a/scripts/compiler.h b/scripts/compiler.h
index bd73c7a9a8d3..ee2097aa3691 100644
--- a/scripts/compiler.h
+++ b/scripts/compiler.h
@@ -42,7 +42,23 @@
 # include <endian.h>
 # include <byteswap.h>
 #elif defined(__MACH__)
-# include <machine/endian.h>
+# ifdef __APPLE__
+#  include <libkern/OSByteOrder.h>
+#  define htobe16(x) OSSwapHostToBigInt16(x)
+#  define htole16(x) OSSwapHostToLittleInt16(x)
+#  define be16toh(x) OSSwapBigToHostInt16(x)
+#  define le16toh(x) OSSwapLittleToHostInt16(x)
+#  define htobe32(x) OSSwapHostToBigInt32(x)
+#  define htole32(x) OSSwapHostToLittleInt32(x)
+#  define be32toh(x) OSSwapBigToHostInt32(x)
+#  define le32toh(x) OSSwapLittleToHostInt32(x)
+#  define htobe64(x) OSSwapHostToBigInt64(x)
+#  define htole64(x) OSSwapHostToLittleInt64(x)
+#  define be64toh(x) OSSwapBigToHostInt64(x)
+#  define le64toh(x) OSSwapLittleToHostInt64(x)
+# else /* non apple __MACH__ */
+#  include <machine/endian.h>
+# endif /* __APPLE__ */
 # define __BYTE_ORDER    BYTE_ORDER
 # define __LITTLE_ENDIAN LITTLE_ENDIAN
 # define __BIG_ENDIAN    BIG_ENDIAN
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 10/18] scripts: use "compiler.h" endianness helpers

[Ahmad Fatoum]

<endian.h> may not exist on other systems, don't include it
directly, but include "compiler.h", which either includes
it or provides a suitable alternative if possible.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/fix_size.c      | 3 ++-
 scripts/imx/imx-image.c | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/scripts/fix_size.c b/scripts/fix_size.c
index cb074e3cbc85..e33edecf6591 100644
--- a/scripts/fix_size.c
+++ b/scripts/fix_size.c
@@ -9,7 +9,8 @@
 #include <unistd.h>
 #include <stdint.h>
 #include <fcntl.h>
-#include <endian.h>
+
+#include "compiler.h"
 
 int main(int argc, char**argv)
 {
diff --git a/scripts/imx/imx-image.c b/scripts/imx/imx-image.c
index a7f1421fa371..3a7b1c0ab2e1 100644
--- a/scripts/imx/imx-image.c
+++ b/scripts/imx/imx-image.c
@@ -26,10 +26,10 @@
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <fcntl.h>
-#include <endian.h>
 #include <linux/kernel.h>
 #include <sys/file.h>
 #include <mach/imx_cpu_types.h>
+#include "../compiler.h"
 
 #include "imx.h"
 
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 11/18] scripts: mkimage: s/fdatasync/fsync/ on macOS & OpenBSD

[Ahmad Fatoum]

fdatasync is an optional part of POSIX and mkimage uses the more general
fsync on Sun and FreeBSD as an alternative. Add macOS[1] and OpenBSD to the
list as well.

This is in-line with what U-Boot does since 31cbe80c33
("mkimage: fix compilation issues on OpenBSD").

[1]: macOS >=10.7 does indeed define (but not declare anywhere) a seemingly
     working fdatasync, but we won't consider this here.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/mkimage.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/scripts/mkimage.c b/scripts/mkimage.c
index 891d7b6de71c..7d283c550930 100644
--- a/scripts/mkimage.c
+++ b/scripts/mkimage.c
@@ -572,7 +572,11 @@ NXTARG:		;
 	}
 
 	/* We're a bit of paranoid */
-#if defined(_POSIX_SYNCHRONIZED_IO) && !defined(__sun__) && !defined(__FreeBSD__)
+#if defined(_POSIX_SYNCHRONIZED_IO) && \
+   !defined(__sun__) && \
+   !defined(__FreeBSD__) && \
+   !defined(__OpenBSD__) && \
+   !defined(__APPLE__)
 	(void) fdatasync (ifd);
 #else
 	(void) fsync (ifd);
@@ -622,7 +626,11 @@ NXTARG:		;
 	(void) munmap((void *)ptr, sbuf.st_size);
 
 	/* We're a bit of paranoid */
-#if defined(_POSIX_SYNCHRONIZED_IO) && !defined(__sun__) && !defined(__FreeBSD__)
+#if defined(_POSIX_SYNCHRONIZED_IO) && \
+   !defined(__sun__) && \
+   !defined(__FreeBSD__) && \
+   !defined(__OpenBSD__) && \
+   !defined(__APPLE__)
 	(void) fdatasync (ifd);
 #else
 	(void) fsync (ifd);
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 12/18] scripts: removes uses of <asm*/errno.h> in favor of <errno.h>

[Ahmad Fatoum]

We can't assume <asm/errno.h> to be available on non-Linux hosts.
Furthermore, errno constants values aren't fixed and passing them
to libc perror(3)/strerror (3) would yield wrong results.
Fix this by using the standard <errno.h> header instead.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/bareboximd.c        | 1 -
 scripts/include/linux/err.h | 2 +-
 2 files changed, 1 insertion(+), 2 deletions(-)

diff --git a/scripts/bareboximd.c b/scripts/bareboximd.c
index 81a59ec63c23..5ef91831c457 100644
--- a/scripts/bareboximd.c
+++ b/scripts/bareboximd.c
@@ -23,7 +23,6 @@
 #include <stdio.h>
 #include <sys/types.h>
 #include <stdint.h>
-#include <asm-generic/errno.h>
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <fcntl.h>
diff --git a/scripts/include/linux/err.h b/scripts/include/linux/err.h
index bdc3dd8131d4..9982ab17babe 100644
--- a/scripts/include/linux/err.h
+++ b/scripts/include/linux/err.h
@@ -4,7 +4,7 @@
 #include <linux/compiler.h>
 #include <linux/types.h>
 
-#include <asm/errno.h>
+#include <errno.h>
 
 /*
  * Original kernel header comment:
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 13/18] scripts: omap3-usb-loader: don't depend on unportable le32toh

[Ahmad Fatoum]

The file already defines le32_to_cpu(), which does the same, so
use it.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/omap3-usb-loader.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/omap3-usb-loader.c b/scripts/omap3-usb-loader.c
index 3ba9af06527c..73c5a6d97a53 100644
--- a/scripts/omap3-usb-loader.c
+++ b/scripts/omap3-usb-loader.c
@@ -403,7 +403,7 @@ static int transfer_first_stage(libusb_device_handle * handle, struct arg_state
 	data = file->data;
 	dbuf = data;
 
-	if (le32toh(dbuf[5]) == 0x45534843) {
+	if (le32_to_cpu(dbuf[5]) == 0x45534843) {
 		int chsettingssize = 512 + 2 * sizeof(uint32_t);
 
 		log_info("CHSETTINGS image detected. Skipping header\n");
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 14/18] scripts: omap3-usb-loader: drop unneeded header

[Ahmad Fatoum]

Declarations from the header aren't used anywhere, so drop it.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/omap3-usb-loader.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/scripts/omap3-usb-loader.c b/scripts/omap3-usb-loader.c
index 73c5a6d97a53..ae6f1258df05 100644
--- a/scripts/omap3-usb-loader.c
+++ b/scripts/omap3-usb-loader.c
@@ -33,10 +33,6 @@
 #define OMAP_IS_BIG_ENDIAN
 #endif
 
-#ifdef OMAP_IS_BIG_ENDIAN
-#include <arpa/inet.h>
-#endif
-
 #include <unistd.h>		/* for usleep and friends */
 #include <getopt.h>
 #include <errno.h>
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 15/18] scripts: imx-usb-loader: don't depend on unportable headers

[Ahmad Fatoum]

<sys/io.h> is non-standard and even unavailable on some Linux
systems, like arm64. Drop it as it's unused anyway.

<endian.h> is Linux-specific, so don't include it directly, but
let "compiler.h" figure out, whether it should be included.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/imx/imx-usb-loader.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/scripts/imx/imx-usb-loader.c b/scripts/imx/imx-usb-loader.c
index d4fba0ce54e8..cc26cdd8afb6 100644
--- a/scripts/imx/imx-usb-loader.c
+++ b/scripts/imx/imx-usb-loader.c
@@ -26,16 +26,15 @@
 
 #include <unistd.h>
 #include <ctype.h>
-#include <sys/io.h>
 #include <errno.h>
 #include <string.h>
 #include <stdlib.h>
 #include <libusb.h>
 #include <getopt.h>
-#include <endian.h>
 #include <arpa/inet.h>
 #include <linux/kernel.h>
 
+#include "../compiler.h"
 #include "imx.h"
 
 #define get_min(a, b) (((a) < (b)) ? (a) : (b))
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 16/18] scripts: kwbimage: fix build with non-glibc systems

[Ahmad Fatoum]

get_current_dir_name is a glibc extension, thus replace it
with a call to standard POSIX pathconf/malloc/gecwd.

The result slightly differs, because get_current_dir_name consults
the $PWD environment variable as as well, but that's ok, as it's
just an error message.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/kwbimage.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/scripts/kwbimage.c b/scripts/kwbimage.c
index 2a052a7ff385..6ba4abaa3076 100644
--- a/scripts/kwbimage.c
+++ b/scripts/kwbimage.c
@@ -860,12 +860,16 @@ static void *image_create_v1(struct image_cfg_element *image_cfg,
 
 		ret = stat(binarye->binary.file, &s);
 		if (ret < 0) {
-			char *cwd = get_current_dir_name();
+			char *buf, *cwd = NULL;
+			size_t size = (size_t)pathconf(".", _PC_PATH_MAX);
+			buf = malloc(size);
+			if (buf)
+				cwd = getcwd(buf, size);
 			fprintf(stderr,
 				"Didn't find the file '%s' in '%s' which is mandatory to generate the image\n"
 				"This file generally contains the DDR3 training code, and should be extracted from an existing bootable\n"
 				"image for your board. See 'kwbimage -x' to extract it from an existing image.\n",
-				binarye->binary.file, cwd);
+				binarye->binary.file, cwd ? cwd : "current working directory");
 			free(cwd);
 			return NULL;
 		}
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 17/18] scripts: compiler.h: use Linux <endian.h> as default

[Ahmad Fatoum]

Other systems mimic the GNU endianness headers, e.g. Cygwin. Instead of
explicitly listing them, just make the Linux case the default case by
moving it into the #else clause.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/compiler.h | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/scripts/compiler.h b/scripts/compiler.h
index ee2097aa3691..01b0de44f18a 100644
--- a/scripts/compiler.h
+++ b/scripts/compiler.h
@@ -38,10 +38,7 @@
 #define O_BINARY	0
 #endif
 
-#ifdef __linux__
-# include <endian.h>
-# include <byteswap.h>
-#elif defined(__MACH__)
+#if defined(__MACH__)
 # ifdef __APPLE__
 #  include <libkern/OSByteOrder.h>
 #  define htobe16(x) OSSwapHostToBigInt16(x)
@@ -64,6 +61,9 @@
 # define __BIG_ENDIAN    BIG_ENDIAN
 typedef unsigned long ulong;
 typedef unsigned int  uint;
+#else /* assume Linux */
+# include <endian.h>
+# include <byteswap.h>
 #endif
 
 typedef uint8_t __u8;
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

[PATCH 18/18] scripts: compiler.h: support BSDs as well

[Ahmad Fatoum]

The BSDs have their endianness changing functions in <sys/endian.h> and
define both _BYTE_ORDER as well as BYTE_ORDER (if strict POSIX conformance
isn't requested). Extend the header to compile, so it supports these
platforms as well.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
 scripts/compiler.h | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/scripts/compiler.h b/scripts/compiler.h
index 01b0de44f18a..0ad25f9e8da9 100644
--- a/scripts/compiler.h
+++ b/scripts/compiler.h
@@ -56,16 +56,22 @@
 # else /* non apple __MACH__ */
 #  include <machine/endian.h>
 # endif /* __APPLE__ */
-# define __BYTE_ORDER    BYTE_ORDER
-# define __LITTLE_ENDIAN LITTLE_ENDIAN
-# define __BIG_ENDIAN    BIG_ENDIAN
 typedef unsigned long ulong;
 typedef unsigned int  uint;
+#elif defined(__OpenBSD__) || defined(__FreeBSD__) || \
+      defined(__NetBSD__) || defined(__DragonFly__)
+# include <sys/endian.h>
 #else /* assume Linux */
 # include <endian.h>
 # include <byteswap.h>
 #endif
 
+#if defined(__BYTE_ORDER) && !defined(BYTE_ORDER)
+# define __BYTE_ORDER    BYTE_ORDER
+# define __BIG_ENDIAN    BIG_ENDIAN
+# define __LITTLE_ENDIAN LITTLE_ENDIAN
+#endif
+
 typedef uint8_t __u8;
 typedef uint16_t __u16;
 typedef uint32_t __u32;
-- 
2.20.1


_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

Re: [PATCH 00/18] scripts: enable compilation on macOS

[Sascha Hauer]

On Mon, May 27, 2019 at 11:57:26AM +0200, Ahmad Fatoum wrote:
> With this patch series applied, imx_v7_defconfig along with the host
> tools can be built on macOS. It should also pave the way for building
> on the BSDs. Currently it still fails on NetBSD though:
> 
> 	In file included from scripts/mod/modpost.c:15:0:
> 	scripts/mod/modpost.h:16:0: warning: "Elf_Ehdr" redefined
> 	 #define Elf_Ehdr    Elf32_Ehdr
> 	 ^
> 	In file included from scripts/mod/modpost.h:10:0,
> 			 from scripts/mod/modpost.c:15:
> 	/usr/include/elf.h:1093:0: note: this is the location of the previous definition
> 	 #define Elf_Ehdr Elf64_Ehdr
> 	 ^
> 
> Issues addressed in the series can be split into three groups:
> 
> - host tools depend on glibc or Linux headers
> 	replaced where appropriate with standard ISO/POSIX functionality
> - build depends on GNU utilites:
> 	only stat(1). fixed by importing the linux patch that uses ls
> 	instead.
> - endianness helpers
> 	These aren't (yet) POSIX-standardized, so I extended compiler.h
> 	to support macOS and the BSDs.
> 
> Only thing I can't yet compile is MXS_HOSTTOOLS, which depends on
> pkg-config finding OpenSSL, but I guess that's something I need to fix
> at my side.

Applied, thanks
 Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox