[U-Boot-Users] RFC: make new image support mandatory
Hello everybody,
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
Best regards,
Wolfgang Denk
On Mar 26, 2008, at 9:16 AM, Wolfgang Denk wrote:
Hello everybody,
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
I haven't tried out the CONFIG_FIT stuff yet so some of these concerns might not be real issues: 1. Are their changes to linux kernel build system to create a FIT .uImage ? 2. if "#define CONFIG_FIT 1" can we use an "old" style uImage?
If the answer to 2. is no, I'm concerned about boards that are part of a family (like Freescale reference boards for PPC) and having two different image formats to deal with.
- k
Kumar Gala wrote: [...]
I haven't tried out the CONFIG_FIT stuff yet so some of these concerns might not be real issues:
- Are their changes to linux kernel build system to create a
FIT .uImage ?
No, new uImage support has not been integrated with the kernel build system.
- if "#define CONFIG_FIT 1" can we use an "old" style uImage?
Yes, both old and new formats are supported.
Regards, Bartlomiej
In message A3CF2817-786F-45C9-92BA-6DF8CE150C85@kernel.crashing.org you wrote:
Can we agree on this?
I haven't tried out the CONFIG_FIT stuff yet so some of these concerns might not be real issues:
- Are their changes to linux kernel build system to create a
FIT .uImage ?
Not yet.
- if "#define CONFIG_FIT 1" can we use an "old" style uImage?
Yes, of course.
If the answer to 2. is no, I'm concerned about boards that are part of a family (like Freescale reference boards for PPC) and having two different image formats to deal with.
No, I don't intend to force incompatible changes on innocent users :-)
Ideally, nobody has even to take notice of the new feature - unless they run "help" on some of the commands.
But then, they *can* use the new features if they want.
Best regards,
Wolfgang Denk
On Mar 26, 2008, at 9:57 AM, Wolfgang Denk wrote:
In message A3CF2817-786F-45C9-92BA-6DF8CE150C85@kernel.crashing.org you wrote:
Can we agree on this?
I haven't tried out the CONFIG_FIT stuff yet so some of these concerns might not be real issues:
- Are their changes to linux kernel build system to create a
FIT .uImage ?
Not yet.
- if "#define CONFIG_FIT 1" can we use an "old" style uImage?
Yes, of course.
If the answer to 2. is no, I'm concerned about boards that are part of a family (like Freescale reference boards for PPC) and having two different image formats to deal with.
No, I don't intend to force incompatible changes on innocent users :-)
Ideally, nobody has even to take notice of the new feature - unless they run "help" on some of the commands.
But then, they *can* use the new features if they want.
Than I'm all for.
- k
Wolfgang Denk wrote:
Hello everybody,
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
new uImage format also requires:
#define CONFIG_OF_LIBFDT 1
Can we agree on this?
I'm all for it :)
Regards, Bartlomiej
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
Can we agree on this?
Can we please un-break the existing tree first?
gcc -g -Wall -idirafter /home/hskinnemoen/git/u-boot/upstream/include -idirafter /home/hskinnemoen/git/u-boot/upstream/include2 -idirafter /home/hskinnemoen/git/u-boot/upstream/include -DTEXT_BASE=0x00000000 -DUSE_HOSTCC -O -c -o fdt.o fdt.c In file included from fdt_host.h:24, from fdt.c:57: ../include/fdt.h:7: error: expected specifier-qualifier-list before ‘uint32_t’ ../include/fdt.h:26: error: expected specifier-qualifier-list before ‘uint64_t’ ../include/fdt.h:31: error: expected specifier-qualifier-list before ‘uint32_t’ ../include/fdt.h:36: error: expected specifier-qualifier-list before ‘uint32_t’ In file included from fdt_host.h:25, from fdt.c:57: ../include/libfdt.h:131: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘fdt_next_tag’ ../include/libfdt.h:156: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_magic’: ../include/libfdt.h:156: error: ‘struct fdt_header’ has no member named ‘magic’ ../include/libfdt.h:156: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h:156: error: (Each undeclared identifier is reported only once ../include/libfdt.h:156: error: for each function it appears in.) ../include/libfdt.h: At top level: ../include/libfdt.h:157: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_totalsize’: ../include/libfdt.h:157: error: ‘struct fdt_header’ has no member named ‘totalsize’ ../include/libfdt.h:157: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:158: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_off_dt_struct’: ../include/libfdt.h:158: error: ‘struct fdt_header’ has no member named ‘off_dt_struct’ ../include/libfdt.h:158: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:159: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_off_dt_strings’: ../include/libfdt.h:159: error: ‘struct fdt_header’ has no member named ‘off_dt_strings’ ../include/libfdt.h:159: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:160: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_off_mem_rsvmap’: ../include/libfdt.h:160: error: ‘struct fdt_header’ has no member named ‘off_mem_rsvmap’ ../include/libfdt.h:160: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:161: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_version’: ../include/libfdt.h:161: error: ‘struct fdt_header’ has no member named ‘version’ ../include/libfdt.h:161: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:162: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_last_comp_version’: ../include/libfdt.h:162: error: ‘struct fdt_header’ has no member named ‘last_comp_version’ ../include/libfdt.h:162: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:163: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_boot_cpuid_phys’: ../include/libfdt.h:163: error: ‘struct fdt_header’ has no member named ‘boot_cpuid_phys’ ../include/libfdt.h:163: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:164: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_size_dt_strings’: ../include/libfdt.h:164: error: ‘struct fdt_header’ has no member named ‘size_dt_strings’ ../include/libfdt.h:164: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:165: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_size_dt_struct’: ../include/libfdt.h:165: error: ‘struct fdt_header’ has no member named ‘size_dt_struct’ ../include/libfdt.h:165: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:251: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:251: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:423: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘fdt_get_phandle’ ../include/libfdt.h:590: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h:718: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_setprop_inplace_cell’: ../include/libfdt.h:720: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:779: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:779: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:783: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_property_cell’: ../include/libfdt.h:785: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:822: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:822: error: expected declaration specifiers or ‘...’ before ‘uint64_t’ ../include/libfdt.h:908: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_setprop_cell’: ../include/libfdt.h:910: error: ‘val’ undeclared (first use in this function) In file included from fdt.c:60: libfdt_internal.h: At top level: libfdt_internal.h:61: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘_fdt_next_tag’ libfdt_internal.h: In function ‘_fdt_offset_ptr’: libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h:67: error: ‘const struct fdt_header’ has no member named ‘off_dt_struct’ libfdt_internal.h: In function ‘_fdt_mem_rsv’: libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ libfdt_internal.h:78: error: ‘const struct fdt_header’ has no member named ‘off_mem_rsvmap’ fdt.c: In function ‘fdt_check_header’: fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:64: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:66: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:68: error: ‘const struct fdt_header’ has no member named ‘last_comp_version’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:70: error: ‘const struct fdt_header’ has no member named ‘magic’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:72: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c: In function ‘fdt_offset_ptr’: fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:85: error: ‘const struct fdt_header’ has no member named ‘version’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c:87: error: ‘const struct fdt_header’ has no member named ‘size_dt_struct’ fdt.c: At top level: fdt.c:97: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘fdt_next_tag’ fdt.c: In function ‘fdt_move’: fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:155: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ make[1]: *** [fdt.o] Error 1 make[1]: Leaving directory `/home/hskinnemoen/git/u-boot/upstream/tools' make: *** [tools] Error 2
$ gcc -v Using built-in specs. Target: i486-linux-gnu Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.2 --program-suffix=-4.2 --enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-mpfr --disable-libmudflap --enable-targets=all --enable-checking=release --build=i486-linux-gnu --host=i486-linux-gnu --target=i486-linux-gnu Thread model: posix gcc version 4.2.3 (Debian 4.2.3-2)
Haavard
Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
Can we agree on this?
Can we please un-break the existing tree first?
gcc -g -Wall -idirafter /home/hskinnemoen/git/u-boot/upstream/include -idirafter /home/hskinnemoen/git/u-boot/upstream/include2 -idirafter /home/hskinnemoen/git/u-boot/upstream/include -DTEXT_BASE=0x00000000 -DUSE_HOSTCC -O -c -o fdt.o fdt.c In file included from fdt_host.h:24, from fdt.c:57: ../include/fdt.h:7: error: expected specifier-qualifier-list before ‘uint32_t’
[...] er named ‘totalsize’
fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ make[1]: *** [fdt.o] Error 1 make[1]: Leaving directory `/home/hskinnemoen/git/u-boot/upstream/tools' make: *** [tools] Error 2
$ gcc -v Using built-in specs. Target: i486-linux-gnu Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.2 --program-suffix=-4.2 --enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-mpfr --disable-libmudflap --enable-targets=all --enable-checking=release --build=i486-linux-gnu --host=i486-linux-gnu --target=i486-linux-gnu Thread model: posix gcc version 4.2.3 (Debian 4.2.3-2)
Haavard,
What were the make invocations that let the above errors? I'll try to reproduce this and have a look.
Regards, Bartlomiej
Bartlomiej Sieka wrote:
Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
Can we agree on this?
Can we please un-break the existing tree first?
gcc -g -Wall -idirafter /home/hskinnemoen/git/u-boot/upstream/include -idirafter /home/hskinnemoen/git/u-boot/upstream/include2 -idirafter /home/hskinnemoen/git/u-boot/upstream/include -DTEXT_BASE=0x00000000 -DUSE_HOSTCC -O -c -o fdt.o fdt.c In file included from fdt_host.h:24, from fdt.c:57: ../include/fdt.h:7: error: expected specifier-qualifier-list before ‘uint32_t’
[...] er named ‘totalsize’
fdt.c:158: error: ‘const struct fdt_header’ has no member named ‘totalsize’ make[1]: *** [fdt.o] Error 1 make[1]: Leaving directory `/home/hskinnemoen/git/u-boot/upstream/tools' make: *** [tools] Error 2
$ gcc -v Using built-in specs. Target: i486-linux-gnu Configured with: ../src/configure -v --enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.2 --program-suffix=-4.2 --enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-mpfr --disable-libmudflap --enable-targets=all --enable-checking=release --build=i486-linux-gnu --host=i486-linux-gnu --target=i486-linux-gnu Thread model: posix gcc version 4.2.3 (Debian 4.2.3-2)
Haavard,
What were the make invocations that let the above errors? I'll try to reproduce this and have a look.
Regards, Bartlomiej
Hi Bartlomiej,
I also am running debian and am getting the error. Apparently our /usr/include/linux/types.h files are different from yours, or we have something set up differently/wrongly.
Adding -isystem ../include (to override the host's include files) to FIT_FLAGS "fixes" the problem.
-------------------------------------------------------------------- diff --git a/tools/Makefile b/tools/Makefile index 8784a6d..9b2f6fa 100644 --- a/tools/Makefile +++ b/tools/Makefile @@ -124,7 +124,7 @@ CPPFLAGS = -idirafter $(SRCTREE)/include \ CFLAGS = $(HOST_CFLAGS) $(CPPFLAGS) -O
# No -pedantic switch to avoid libfdt compilation warnings -FIT_CFLAGS = -Wall $(CPPFLAGS) -O +FIT_CFLAGS = -Wall -isystem ../include $(CPPFLAGS) -O
AFLAGS = -D__ASSEMBLY__ $(CPPFLAGS) CC = $(HOSTCC) --------------------------------------------------------------------
I've also attached the diff between /usr/include/linux/types.h and u-boot's include/linux/types.h in case it is helpful. Nothing jumps out at me. :-/
Best regards, gvb
On Thursday 27 March 2008, Jerry Van Baren wrote:
I also am running debian and am getting the error. Apparently our /usr/include/linux/types.h files are different from yours, or we have something set up differently/wrongly.
Adding -isystem ../include (to override the host's include files) to FIT_FLAGS "fixes" the problem.
I'm running Ubuntu and am getting this errors too. Even without enabling CONFIG_FIT:
[stefan@ubuntu u-boot-ppc4xx (master)]$ ./MAKEALL sequoia Configuring for sequoia board... In file included from fdt_host.h:24, from fdt.c:57: ../include/fdt.h:7: error: expected specifier-qualifier-list before ‘uint32_t’ ../include/fdt.h:26: error: expected specifier-qualifier-list before ‘uint64_t’ ../include/fdt.h:31: error: expected specifier-qualifier-list before ‘uint32_t’ ../include/fdt.h:36: error: expected specifier-qualifier-list before ‘uint32_t’ In file included from fdt_host.h:25, from fdt.c:57: ../include/libfdt.h:131: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘fdt_next_tag’ ../include/libfdt.h:162: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_magic’: ../include/libfdt.h:162: error: ‘struct fdt_header’ has no member named ‘magic’ ../include/libfdt.h:162: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h:162: error: (Each undeclared identifier is reported only once ../include/libfdt.h:162: error: for each function it appears in.) ../include/libfdt.h: At top level: ../include/libfdt.h:163: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_totalsize’: ../include/libfdt.h:163: error: ‘struct fdt_header’ has no member named ‘totalsize’ ../include/libfdt.h:163: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:164: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_off_dt_struct’: ../include/libfdt.h:164: error: ‘struct fdt_header’ has no member named ‘off_dt_struct’ ../include/libfdt.h:164: error: ‘val’ undeclared (first use in this function) ../include/libfdt.h: At top level: ../include/libfdt.h:165: error: expected declaration specifiers or ‘...’ before ‘uint32_t’ ../include/libfdt.h: In function ‘fdt_set_off_dt_strings’: ../include/libfdt.h:165: error: ‘struct fdt_header’ has no member named ‘off_dt_strings’
....
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
In message 58A20A281BAF1047B4EAE68DE5C0BDC2DD833B@zch01exm21.fsl.freescale.net you wrote:
PiANCj4gSSdtIHJ1bm5pbmcgVWJ1bnR1IGFuZCBhbSBnZXR0aW5nIHRoaXMgZXJyb3JzIHRvby4g RXZlbiB3aXRob3V0DQo+IGVuYWJsaW5nIENPTkZJR19GSVQ6DQoNCldoZW4gSSBidWlsZCBNUEM4 MzYwRU1EUy9NUEM4MzdYRU1EUyBvbiBVYnVudHUsIHRoZSBzYW1lIGVycm9yDQpoYXBwZW4gdG8g bWUuIA0KDQpUaGFua3MsDQpEYXZlDQo=
Please do not send base 64 encoded messages.
Please send plain text only.
Message unreadable, ignored. Sorry.
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
Oh my...
Configuring for atstk1002 board... text data bss dec hex filename 112936 8032 216860 337828 527a4 ./u-boot
with CONFIG_FIT and CONFIG_OF_LIBFDT enabled, compared to
text data bss dec hex filename 87724 6840 215848 310412 4bc8c ./u-boot
on v1.3.2. That's 26KB...do we really want to make something _that_ expensive mandatory? Or can I somehow disable the old code to reclaim some of it?
Haavard
Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
Oh my...
Configuring for atstk1002 board... text data bss dec hex filename 112936 8032 216860 337828 527a4 ./u-boot
with CONFIG_FIT and CONFIG_OF_LIBFDT enabled, compared to
text data bss dec hex filename 87724 6840 215848 310412 4bc8c ./u-boot
BTW: were you able to compile the tools/ directory for the above board? If yes, what were the differences versus the failing build you posted in your other email?
on v1.3.2. That's 26KB...do we really want to make something _that_ expensive mandatory? Or can I somehow disable the old code to reclaim some of it?
There's no easy way to disable support for the old format currently.
Regards, Bartlomiej
On Wed, 26 Mar 2008 17:31:40 +0100 Bartlomiej Sieka tur@semihalf.com wrote:
BTW: were you able to compile the tools/ directory for the above board? If yes, what were the differences versus the failing build you posted in your other email?
Ah, sorry. I sent a patch, but didn't realize it went only to Wolfgang. Here it is.
Haavard
From b69cca7525e66808fed7fcd04e74900a2fb6db91 Mon Sep 17 00:00:00 2001 From: Haavard Skinnemoen haavard.skinnemoen@atmel.com Date: Wed, 26 Mar 2008 16:56:12 +0100 Subject: [PATCH] Fix host tool build breakage
Signed-off-by: Haavard Skinnemoen haavard.skinnemoen@atmel.com --- include/fdt.h | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-)
diff --git a/include/fdt.h b/include/fdt.h index 48ccfd9..39beada 100644 --- a/include/fdt.h +++ b/include/fdt.h @@ -3,6 +3,10 @@
#ifndef __ASSEMBLY__
+#ifdef USE_HOSTCC +#include <stdint.h> +#endif + struct fdt_header { uint32_t magic; /* magic word FDT_MAGIC */ uint32_t totalsize; /* total size of DT block */
On Thursday 27 March 2008, Haavard Skinnemoen wrote:
BTW: were you able to compile the tools/ directory for the above board? If yes, what were the differences versus the failing build you posted in your other email?
Ah, sorry. I sent a patch, but didn't realize it went only to Wolfgang. Here it is.
Haavard
From b69cca7525e66808fed7fcd04e74900a2fb6db91 Mon Sep 17 00:00:00 2001
From: Haavard Skinnemoen haavard.skinnemoen@atmel.com Date: Wed, 26 Mar 2008 16:56:12 +0100 Subject: [PATCH] Fix host tool build breakage
Signed-off-by: Haavard Skinnemoen haavard.skinnemoen@atmel.com
This fixes the problem for me too. Thanks.
Acked-by: Stefan Roese sr@denx.de
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
Before applying this patch, I encounter the same issue as Stefan on mpc8544 and mpc8548 boards. After applying this patch, they are OK. This patch should be applied ASAP, which affetcs boundle of boards. Roy On Thu, 2008-03-27 at 17:01 +0800, Stefan Roese wrote:
On Thursday 27 March 2008, Haavard Skinnemoen wrote:
BTW: were you able to compile the tools/ directory for the above
board?
If yes, what were the differences versus the failing build you
posted in
your other email?
Ah, sorry. I sent a patch, but didn't realize it went only to
Wolfgang.
Here it is.
Haavard
From b69cca7525e66808fed7fcd04e74900a2fb6db91 Mon Sep 17 00:00:00
2001
From: Haavard Skinnemoen haavard.skinnemoen@atmel.com Date: Wed, 26 Mar 2008 16:56:12 +0100 Subject: [PATCH] Fix host tool build breakage
Signed-off-by: Haavard Skinnemoen haavard.skinnemoen@atmel.com
This fixes the problem for me too. Thanks.
Acked-by: Stefan Roese sr@denx.de
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================
Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ U-Boot-Users mailing list U-Boot-Users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/u-boot-users
Stefan Roese sr@denx.de writes:
On Thursday 27 March 2008, Haavard Skinnemoen wrote:
BTW: were you able to compile the tools/ directory for the above board? If yes, what were the differences versus the failing build you posted in your other email?
Ah, sorry. I sent a patch, but didn't realize it went only to Wolfgang. Here it is.
Haavard
From b69cca7525e66808fed7fcd04e74900a2fb6db91 Mon Sep 17 00:00:00 2001
From: Haavard Skinnemoen haavard.skinnemoen@atmel.com Date: Wed, 26 Mar 2008 16:56:12 +0100 Subject: [PATCH] Fix host tool build breakage
Signed-off-by: Haavard Skinnemoen haavard.skinnemoen@atmel.com
This fixes the problem for me too. Thanks.
Acked-by: Stefan Roese sr@denx.de
Me too.
Acked-by: Markus Klotzbuecher mk@denx.de
Best regards
Markus Klotzbuecher
-- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de
Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 17:31:40 +0100 Bartlomiej Sieka tur@semihalf.com wrote:
BTW: were you able to compile the tools/ directory for the above board? If yes, what were the differences versus the failing build you posted in your other email?
Ah, sorry. I sent a patch, but didn't realize it went only to Wolfgang. Here it is.
Haavard,
This patch breaks build on FC. I'm about to send another patch that will hopefully restore the build for everyone.
Regards, Bartlomiej
Hi Bartlomiej,
I posted a week ago some considerations about the new image format (with subject "[U-Boot-Users] New Image format: headers hashes."). Please, can you give an your opinion about these consideration? I'm really interested to support new image in my firmwares, but I need to understand if the new format is "robust" respect to the accidental errors.
Thanks a lot for your work and your answers on the ml.
Ciao
luigi
Luigi 'Comio' Mantellini wrote:
Hi Bartlomiej,
I posted a week ago some considerations about the new image format (with subject "[U-Boot-Users] New Image format: headers hashes."). Please, can you give an your opinion about these consideration? I'm really interested to support new image in my firmwares, but I need to understand if the new format is "robust" respect to the accidental errors.
Thanks a lot for your work and your answers on the ml.
Hi Luigi,
Sorry for being slow to respond -- I'm processing ML comments on the new image, and will reply to your postings today/tomorrow.
Regards, Bartlomiej
On Wed, 26 Mar 2008, Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
Oh my...
Configuring for atstk1002 board... text data bss dec hex filename 112936 8032 216860 337828 527a4 ./u-boot
with CONFIG_FIT and CONFIG_OF_LIBFDT enabled, compared to
text data bss dec hex filename 87724 6840 215848 310412 4bc8c ./u-boot
on v1.3.2. That's 26KB...do we really want to make something _that_ expensive mandatory? Or can I somehow disable the old code to reclaim some of it?
I'm getting 34KB increase iin the resulting binary just by enabling the required for FIT config macros in an arm11 build with the current git HEAD. Which is almost 30% increase... Sorry, I didn't follow the "new format" discussion, is it also going to be useful on ARM and on other platforms not using FDT?
Thanks Guennadi --- Guennadi Liakhovetski, Ph.D.
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de
In message Pine.LNX.4.64.0803262132010.5050@axis700.grange you wrote:
I'm getting 34KB increase iin the resulting binary just by enabling the required for FIT config macros in an arm11 build with the current git
We need to check where this size penalty is coming from. It's not supposed top be that big.
HEAD. Which is almost 30% increase... Sorry, I didn't follow the "new format" discussion, is it also going to be useful on ARM and on other platforms not using FDT?
Yes, it is useful for all architectures.
Best regards,
Wolfgang Denk
Guennadi Liakhovetski wrote:
HEAD. Which is almost 30% increase... Sorry, I didn't follow the "new format" discussion, is it also going to be useful on ARM and on other platforms not using FDT?
As MIPS custodian I feel the same, too.
I'm just wondering, if this feature is useful for all architectures and you say it's mandaroty, why not defining it in <common.h> or so?
If you make it selectable, I hope we could build existing targets without CONFIG_FIT.
Shinya
Haavard Skinnemoen wrote:
On Wed, 26 Mar 2008 15:16:44 +0100 Wolfgang Denk wd@denx.de wrote:
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
#define CONFIG_FIT 1
in it's board config file; I recommend to add also
#define CONFIG_FIT_VERBOSE 1 /* enable fit_format_{error,warning}() */
Can we agree on this?
Oh my...
Configuring for atstk1002 board... text data bss dec hex filename 112936 8032 216860 337828 527a4 ./u-boot
with CONFIG_FIT and CONFIG_OF_LIBFDT enabled, compared to
text data bss dec hex filename 87724 6840 215848 310412 4bc8c ./u-boot
on v1.3.2. That's 26KB...do we really want to make something _that_ expensive mandatory? Or can I somehow disable the old code to reclaim some of it?
Haavard
On a PowerPC target that includes libfdt already...
Configuring for MPC8360EMDS board with CONFIG_FIT:
text data bss dec hex filename 231588 12072 26836 270496 420a0 ./u-boot
without CONFIG_FIT:
text data bss dec hex filename 211648 11604 26796 250048 3d0c0 ./u-boot
delta text = 19940 (19.5K). Still pretty heavy. :-/ I have not looked at where the weight is coming from, just thought it would be interesting to see what the delta looked like when libfdt was already included. I was expecting a bigger hit from libfdt and less from FIT. Hmmmm.
Best regards, gvb
On 15:16 Wed 26 Mar , Wolfgang Denk wrote:
Hello everybody,
now that the new image format has been merged into mainline, I propose that we make it mandatory for all newly added boards. That means, that each new board that gets added shall at least contain
I'll personnaly do it in the next merge windows. I'll prefer to have a better hindsight on the new image format before used as default.
Best Regards, J.
participants (14)
-
Bartlomiej Sieka -
Dave Liu -
Guennadi Liakhovetski -
Haavard Skinnemoen -
Jean-Christophe PLAGNIOL-VILLARD -
Jerry Van Baren -
Kumar Gala -
Liu Dave -
Luigi 'Comio' Mantellini -
Markus Klotzbücher -
Roy Zang -
Shinya Kuribayashi -
Stefan Roese -
Wolfgang Denk