Re: [PATCH] Fix building kernel under Solaris 11_snv
On Fri, Mar 09, 2007 at 11:23:32PM +0100, Sam Ravnborg wrote: > > >> === > > >> --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 > > >> 05:41:20.0 +0100 > > >> +++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000 > > >> +0100 > > >> @@ -378,6 +378,40 @@ > > >> table_cnt = pos; > > >> } > > >> > > >> +#ifdef __sun__ > > >> +/* Return the first occurrence of NEEDLE in HAYSTACK. */ > > >> +void * > > >> +memmem (haystack, haystack_len, needle, needle_len) > > >> + const void *haystack; > > >> + return (void *) begin; > > >> + > > >> + return NULL; > > >> +} > > >> +#endif > > >> + > > >> /* replace a given token in all the valid symbols. Use the sampled > > >> symbols > > >> * to update the counts */ > > >> static void compress_symbols(unsigned char *str, int idx) > > > > This one, I am just waiting for someone to object to the extra #if-#endif. > I was planning to ask Paulo if strstr could not be used - Paulo? Yeah. And if we can't use something else all the missing functions should be in a compat-solaris.c files, similarly to how sparse handles missing functions on various systems instead of polluting implementation files with ifdefs and duplicates of library code. > > >> -HOST_EXTRACFLAGS += -DLOCALE > > >> +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__ > > >> > > >> PHONY += $(obj)/dochecklxdialog > > >> $(obj)/dochecklxdialog: > > > > The error message for this one was: only valid in C99 mode. > > Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not > > know offhand who is right. > The -std= looks safe. It's not though. It's a gccisms, and I'm not sure we want to mandate gcc for the host copiler, given that the only way you'd build a working kernel on a foreign system would be a cross compilation anyway. > > >> #define M_EVENT (KEY_MAX+1) > > >> + > > >> +#ifndef KEY_RESIZE > > >> +# define KEY_RESIZE 0632 > > >> +#endif > > > > Solaris only has curses, not ncurses. Consider this a supreme hack. > > In fact, menuconfig has some weird display errors still. > This hack looks OK to me. Except that solaris curses doesn't actually understand this random value for KEY_RESIZE. Either we mandate that ncurses is installed on the host for menuconfig to work (and ncurses at least used to work on most unix variants), or some needs to do a proper curses port. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Fri, Mar 09, 2007 at 11:23:32PM +0100, Sam Ravnborg wrote: === --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000 +0100 @@ -378,6 +378,40 @@ table_cnt = pos; } +#ifdef __sun__ +/* Return the first occurrence of NEEDLE in HAYSTACK. */ +void * +memmem (haystack, haystack_len, needle, needle_len) + const void *haystack; + return (void *) begin; + + return NULL; +} +#endif + /* replace a given token in all the valid symbols. Use the sampled symbols * to update the counts */ static void compress_symbols(unsigned char *str, int idx) This one, I am just waiting for someone to object to the extra #if-#endif. I was planning to ask Paulo if strstr could not be used - Paulo? Yeah. And if we can't use something else all the missing functions should be in a compat-solaris.c files, similarly to how sparse handles missing functions on various systems instead of polluting implementation files with ifdefs and duplicates of library code. -HOST_EXTRACFLAGS += -DLOCALE +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__ PHONY += $(obj)/dochecklxdialog $(obj)/dochecklxdialog: The error message for this one was: stdbool.h only valid in C99 mode. Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not know offhand who is right. The -std= looks safe. It's not though. It's a gccisms, and I'm not sure we want to mandate gcc for the host copiler, given that the only way you'd build a working kernel on a foreign system would be a cross compilation anyway. #define M_EVENT (KEY_MAX+1) + +#ifndef KEY_RESIZE +# define KEY_RESIZE 0632 +#endif Solaris only has curses, not ncurses. Consider this a supreme hack. In fact, menuconfig has some weird display errors still. This hack looks OK to me. Except that solaris curses doesn't actually understand this random value for KEY_RESIZE. Either we mandate that ncurses is installed on the host for menuconfig to work (and ncurses at least used to work on most unix variants), or some needs to do a proper curses port. - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 9 2007 23:23, Sam Ravnborg wrote:
>Date: Fri, 9 Mar 2007 23:23:32 +0100
>From: Sam Ravnborg <[EMAIL PROTECTED]>
>To: Jan Engelhardt <[EMAIL PROTECTED]>,
>Paulo Marques <[EMAIL PROTECTED]>
>Cc: [EMAIL PROTECTED], Christoph Hellwig <[EMAIL PROTECTED]>,
>Deepak Saxena <[EMAIL PROTECTED]>,
>Andrew Morton <[EMAIL PROTECTED]>, linux-kernel@vger.kernel.org
>Subject: Re: [PATCH] Fix building kernel under Solaris 11_snv
>
Reference: http://lkml.org/lkml/2007/3/8/368
Signed-off-by: Jan Engelhardt <[EMAIL PROTECTED]>
Happy cherrypicking.
More comments below.
>> >> Index: linux-2.6.21-rc3/scripts/kallsyms.c
>> >> ===
>> >> --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07
>> >> 05:41:20.0 +0100
>> >> +++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
>> >> +0100
>> >> @@ -378,6 +378,40 @@
>> >> table_cnt = pos;
>> >> }
>> >>
>> >> +#ifdef __sun__
>> >> +/* Return the first occurrence of NEEDLE in HAYSTACK. */
>> >> +void *
>> >> +memmem (haystack, haystack_len, needle, needle_len)
>> >> + const void *haystack;
>> >> + return (void *) begin;
>> >> +
>> >> + return NULL;
>> >> +}
>> >> +#endif
>> >> +
>> >> /* replace a given token in all the valid symbols. Use the sampled
>> >> symbols
>> >> * to update the counts */
>> >> static void compress_symbols(unsigned char *str, int idx)
>>
>> This one, I am just waiting for someone to object to the extra #if-#endif.
>
>I was planning to ask Paulo if strstr could not be used - Paulo?
I am not sure, but I would tend to say "no".
kallsyms.c:
p1 = table[i].sym;
/* find the token on the symbol */
p2 = memmem(p1, len, str, 2);
if (!p2) continue;
My first impression would be that 'p1' is a multi-nul-terminated string
array ("foo\0bar\0next\0symbol\0..."), much like the entire .strtab
section (I have not actually bothered to check, so it is a raw guess), and
I do not think strstr would replace memmem here if that was the case.
>> >> Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
>> >> ===
>> >> --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile2007-03-07
>> >> 05:41:20.0 +0100
>> >> +++ linux-2.6.21-rc3/scripts/kconfig/Makefile 2007-03-07
>> >> 23:21:19.730679000 +0100
>> >> @@ -88,7 +88,7 @@
>> >> HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags)
>> >> HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags
>> >> $(HOSTCC))
>> >>
>> >> -HOST_EXTRACFLAGS += -DLOCALE
>> >> +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__
>> >>
>> >> PHONY += $(obj)/dochecklxdialog
>> >> $(obj)/dochecklxdialog:
>>
>> The error message for this one was: only valid in C99 mode.
>> Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not
>> know offhand who is right.
>
>The -std= looks safe. The __EXTENSIONS__ part seems safe and needed after a
>bit googling.
Humm humm. Linux's /usr/include does not have any (glibc) file with the
string __EXTENSIONS__ in it. Though, the linux manpage for fileno() does
not mention fileno() conforming to anything. Under Linux, fileno() is
wrapped inside __USE_POSIX, and the comments in Solaris's headers also
indicate it is something posixy.
Perhaps have something like
#ifdef __sun__
# define __EXTENSIONS__ 1
#endif
>> >> Index: linux-2.6.21-rc3/scripts/mod/file2alias.c
>> >> ===
>> >> --- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c2007-03-07
>> >> 05:41:20.0 +0100
>> >> +++ linux-2.6.21-rc3/scripts/mod/file2alias.c 2007-03-07
>> >> 23:41:23.772026000 +0100
>> >> @@ -32,6 +32,8 @@
>> >> typedef uint32_t __u32;
>> >> typedef uint16_t __u16;
>> >> typedef unsigned char__u8;
>> >> +typedef int32_t __s32;
>> >> +typedef int16_t
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Fri, Mar 09, 2007 at 09:16:35PM +0100, Jan Engelhardt wrote: > > On Mar 9 2007 20:00, Sam Ravnborg wrote: > >On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote: > >> > >> Since Solaris seems to be on the run, I did myself try compile it. > >> However, unlike the original poster who said he did so on SunOS 4.8, I > >> did it on 5.11_snv39, yielding a bigger changeset. I thought I just > >> share the diff that piled up so far. It needs a lot of hacks on the > >> Solaris side - prioritizing GNU names, then, second, gnu ld has a > >> glitch, then, gcc has a missing file... it's fun fun fun! > > > >Can I please have a signed-off version of this patch. > > _Are you sure_ you want all these hacks without further > review from other people? Also note the patch is incomplete, > for example I could not compile the acpi pieces because > acsolaris.h -- which is referenced in the acpi includes -- > does not exist. (Yet another piece of software that has > crossplatform compatibilty stuff, like XFS.) The Signed-off-by: document the origin of the path. I'm planning only to take the sensible bits of the patch anyway. > >> --- linux-2.6.21-rc3.orig/include/linux/input.h2007-03-07 > >> 05:41:20.0 +0100 > >> +++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000 > >> +0100 > >> @@ -16,7 +16,9 @@ > >> #include > >> #include > >> #include > >> -#include > >> +#ifndef __sun__ > >> +# include > >> +#endif > >> #endif > > This is not a proper fix for sure. The problem lies in > file2alias.c, see (your own) http://lkml.org/lkml/2007/3/8/339 I already committed my own fix - so this chunk can be ignored. > > >> Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c > >> === > >> --- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07 > >> 05:41:20.0 +0100 > >> +++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07 > >> 23:28:35.659555000 +0100 > >> @@ -21,6 +21,7 @@ > >> along with this program; if not, write to the Free Software Foundation, > >> Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ > >> > >> +#include > >> #include > >> #include > >> #include > > This is however, is valid. Can I gave sign-offs for single hunks? Agree and yes. > > >> Index: linux-2.6.21-rc3/scripts/kallsyms.c > >> === > >> --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 > >> 05:41:20.0 +0100 > >> +++ linux-2.6.21-rc3/scripts/kallsyms.c2007-03-07 23:46:46.249005000 > >> +0100 > >> @@ -378,6 +378,40 @@ > >>table_cnt = pos; > >> } > >> > >> +#ifdef __sun__ > >> +/* Return the first occurrence of NEEDLE in HAYSTACK. */ > >> +void * > >> +memmem (haystack, haystack_len, needle, needle_len) > >> + const void *haystack; > >> + return (void *) begin; > >> + > >> + return NULL; > >> +} > >> +#endif > >> + > >> /* replace a given token in all the valid symbols. Use the sampled symbols > >> * to update the counts */ > >> static void compress_symbols(unsigned char *str, int idx) > > This one, I am just waiting for someone to object to the extra #if-#endif. I was planning to ask Paulo if strstr could not be used - Paulo? > > >> Index: linux-2.6.21-rc3/scripts/kconfig/Makefile > >> === > >> --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile 2007-03-07 > >> 05:41:20.0 +0100 > >> +++ linux-2.6.21-rc3/scripts/kconfig/Makefile 2007-03-07 > >> 23:21:19.730679000 +0100 > >> @@ -88,7 +88,7 @@ > >> HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags) > >> HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags > >> $(HOSTCC)) > >> > >> -HOST_EXTRACFLAGS += -DLOCALE > >> +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__ > >> > >> PHONY += $(obj)/dochecklxdialog > >> $(obj)/dochecklxdialog: > > The error message for this one was: only valid in C99 mode. > Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not > know offhand who is right. The -std= looks safe. The __EXTENSIONS__ part seems safe and needed after a bit googling. > > >> Index: linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h > >> === > >> --- linux-2.6.21-rc3.orig/scripts/kconfig/lxdialog/dialog.h > >> 2007-03-07 05:41:20.0 +0100 > >> +++ linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h 2007-03-07 > >> 23:14:48.462956000 +0100 > >> @@ -222,3 +222,7 @@ > >> * -- uppercase chars are used to invoke the button (M_EVENT + 'O') > >> */ > >> #define M_EVENT (KEY_MAX+1) > >> + > >> +#ifndef KEY_RESIZE > >> +# define KEY_RESIZE 0632 > >> +#endif > > Solaris only has curses, not ncurses. Consider this a supreme hack. > In fact, menuconfig has some weird display errors still.
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 9 2007 20:00, Sam Ravnborg wrote:
>On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote:
>>
>> Since Solaris seems to be on the run, I did myself try compile it.
>> However, unlike the original poster who said he did so on SunOS 4.8, I
>> did it on 5.11_snv39, yielding a bigger changeset. I thought I just
>> share the diff that piled up so far. It needs a lot of hacks on the
>> Solaris side - prioritizing GNU names, then, second, gnu ld has a
>> glitch, then, gcc has a missing file... it's fun fun fun!
>
>Can I please have a signed-off version of this patch.
_Are you sure_ you want all these hacks without further
review from other people? Also note the patch is incomplete,
for example I could not compile the acpi pieces because
acsolaris.h -- which is referenced in the acpi includes --
does not exist. (Yet another piece of software that has
crossplatform compatibilty stuff, like XFS.)
>> --- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07
>> 05:41:20.0 +0100
>> +++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000
>> +0100
>> @@ -16,7 +16,9 @@
>> #include
>> #include
>> #include
>> -#include
>> +#ifndef __sun__
>> +# include
>> +#endif
>> #endif
This is not a proper fix for sure. The problem lies in
file2alias.c, see (your own) http://lkml.org/lkml/2007/3/8/339
>> Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
>> ===
>> --- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c2007-03-07
>> 05:41:20.0 +0100
>> +++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07
>> 23:28:35.659555000 +0100
>> @@ -21,6 +21,7 @@
>> along with this program; if not, write to the Free Software Foundation,
>> Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
>>
>> +#include
>> #include
>> #include
>> #include
This is however, is valid. Can I gave sign-offs for single hunks?
>> Index: linux-2.6.21-rc3/scripts/kallsyms.c
>> ===
>> --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0
>> +0100
>> +++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
>> +0100
>> @@ -378,6 +378,40 @@
>> table_cnt = pos;
>> }
>>
>> +#ifdef __sun__
>> +/* Return the first occurrence of NEEDLE in HAYSTACK. */
>> +void *
>> +memmem (haystack, haystack_len, needle, needle_len)
>> + const void *haystack;
>> + size_t haystack_len;
>> + const void *needle;
>> + size_t needle_len;
>> +{
>> + const char *begin;
>> + const char *const last_possible
>> += (const char *) haystack + haystack_len - needle_len;
>> +
>> + if (needle_len == 0)
>> +/* The first occurrence of the empty string is deemed to occur at
>> + the beginning of the string. */
>> +return (void *) haystack;
>> +
>> + /* Sanity check, otherwise the loop might search through the whole
>> + memory. */
>> + if (__builtin_expect (haystack_len < needle_len, 0))
>> +return NULL;
>> +
>> + for (begin = (const char *) haystack; begin <= last_possible; ++begin)
>> +if (begin[0] == ((const char *) needle)[0] &&
>> +!memcmp ((const void *) [1],
>> + (const void *) ((const char *) needle + 1),
>> + needle_len - 1))
>> + return (void *) begin;
>> +
>> + return NULL;
>> +}
>> +#endif
>> +
>> /* replace a given token in all the valid symbols. Use the sampled symbols
>> * to update the counts */
>> static void compress_symbols(unsigned char *str, int idx)
This one, I am just waiting for someone to object to the extra #if-#endif.
>> Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
>> ===
>> --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile 2007-03-07
>> 05:41:20.0 +0100
>> +++ linux-2.6.21-rc3/scripts/kconfig/Makefile2007-03-07
>> 23:21:19.730679000 +0100
>> @@ -88,7 +88,7 @@
>> HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags)
>> HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags
>> $(HOSTCC))
>>
>> -HOST_EXTRACFLAGS += -DLOCALE
>> +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__
>>
>> PHONY += $(obj)/dochecklxdialog
>> $(obj)/dochecklxdialog:
The error message for this one was: only valid in C99 mode.
Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not
know offhand who is right.
>> Index: linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h
>> ===
>> --- linux-2.6.21-rc3.orig/scripts/kconfig/lxdialog/dialog.h 2007-03-07
>> 05:41:20.0 +0100
>> +++ linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h 2007-03-07
>> 23:14:48.462956000 +0100
>> @@ -222,3 +222,7 @@
>> * -- uppercase chars are used to invoke the button (M_EVENT + 'O')
>> */
>> #define
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote:
>
> On Mar 8 2007 22:25, Sam Ravnborg wrote:
> >Subject: Re: [PATCH] Fix building kernel under Solaris
>
> Since Solaris seems to be on the run, I did myself try compile it.
> However, unlike the original poster who said he did so on SunOS 4.8, I
> did it on 5.11_snv39, yielding a bigger changeset. I thought I just
> share the diff that piled up so far. It needs a lot of hacks on the
> Solaris side - prioritizing GNU names, then, second, gnu ld has a
> glitch, then, gcc has a missing file... it's fun fun fun!
Can I please have a signed-off version of this patch.
Thanks,
Sam
>
> Well, I will iterate the key problem with the missing file:
>
> * include/linux/kernel.h (and many others) include
> BUT - since we are using -nostdinc, /usr/include/stdarg.h is not
> considered. And gcc's stdarg.h (which lives at
> /usr/lib/gcc/i586-suse-linux/4.1.2/include/stdarg.h in Linux land)
> is missing in Solaris' GCC (which is version 3.4.3).
>
> Hack #1:
> ln -s \
>
> /usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/install-tools/include/stdarg.h \
> /usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/include/stdarg.h
>
> Hack #2: GNU programs...
>
> mkdir -p ~/gnulink;
> for i in addr2line ar as egrep grep ld make nm objcopy objdump \
> ranlib readelf size string strip tar; do
> ln -s "/usr/sfw/bin/$i" "~/gnulink/$i";
> done;
>
> for i in cat chgrp chmod chown chroot cksum cmp cp cut date \
> dd df diff du echo env expand expr false fgrep find fold \
> getopt groups head hostid install join ln locate ls mkdir \
> mkfifo mknod mv nice nohup od pwd rm rmdir sed seq shred \
> sleep sort split stty tac tail tee touch tr true uname \
> uniq uptime wc who whoami xargs yes gawk; do
> ln -s "/opt/csw/bin/$i" "~/gnulink/$i";
> done;
>
> Hack #3: Diff file...
>
> Hack #4: GNU ld glitch workaround (GNU ld looks in the current dir...)
>
> cd linux-2.6.21-rc3
> ln -s /usr/sfw/i386-sun-solaris2.11/lib/ldscripts ldscripts
>
> Fun #1:
>
> export PATH="$HOME/gnulink:$PATH";
> make ARCH=i386
>
> Oddity #1:
>
> ARCH=i386 required because the Makefiles seem to use `uname -m`
> (which returns "i86pc") rather than `uname -p`. I think we are
> at odds here though...
>
> uname -muname -p
> SOL i86pc i386
> LINUX i686athlon
>
>
> Expect compiler failures, especially with assembler code.
>
>
> Jan
>
> <<< PATCH BELOW <<<
>
> Index: linux-2.6.21-rc3/include/linux/input.h
> ===
> --- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07
> 05:41:20.0 +0100
> +++ linux-2.6.21-rc3/include/linux/input.h2007-03-07 23:40:39.417339000
> +0100
> @@ -16,7 +16,9 @@
> #include
> #include
> #include
> -#include
> +#ifndef __sun__
> +#include
> +#endif
> #endif
>
> /*
> Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
> ===
> --- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07
> 05:41:20.0 +0100
> +++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07
> 23:28:35.659555000 +0100
> @@ -21,6 +21,7 @@
> along with this program; if not, write to the Free Software Foundation,
> Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
>
> +#include
> #include
> #include
> #include
> Index: linux-2.6.21-rc3/scripts/kallsyms.c
> ===
> --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0
> +0100
> +++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
> +0100
> @@ -378,6 +378,40 @@
> table_cnt = pos;
> }
>
> +#ifdef __sun__
> +/* Return the first occurrence of NEEDLE in HAYSTACK. */
> +void *
> +memmem (haystack, haystack_len, needle, needle_len)
> + const void *haystack;
> + size_t haystack_len;
> + const void *needle;
> + size_t needle_len;
> +{
> + const char *begin;
> + const char *const last_possible
> += (const char *) haystack + haystack_len - needle_len;
> +
> + if (needle_len == 0)
> +/* The first occurrence of the empty string is deemed to occur at
> + the beginning of the string. */
> +return (void *) haystack;
> +
> + /* Sanity check, otherwise the loop might search through the whole
> + memory. */
> + if (__builtin_expect (haystack_len < needle_len, 0))
> +return NULL;
> +
> + for (begin = (const char *) haystack; begin <= last_possible; ++begin)
> +if (begin[0] == ((const char *) needle)[0] &&
> +!memcmp ((const void *) [1],
> + (const void *) ((const char *) needle + 1),
> + needle_len -
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote:
On Mar 8 2007 22:25, Sam Ravnborg wrote:
Subject: Re: [PATCH] Fix building kernel under Solaris
Since Solaris seems to be on the run, I did myself try compile it.
However, unlike the original poster who said he did so on SunOS 4.8, I
did it on 5.11_snv39, yielding a bigger changeset. I thought I just
share the diff that piled up so far. It needs a lot of hacks on the
Solaris side - prioritizing GNU names, then, second, gnu ld has a
glitch, then, gcc has a missing file... it's fun fun fun!
Can I please have a signed-off version of this patch.
Thanks,
Sam
Well, I will iterate the key problem with the missing file:
* include/linux/kernel.h (and many others) include stdarg.h
BUT - since we are using -nostdinc, /usr/include/stdarg.h is not
considered. And gcc's stdarg.h (which lives at
/usr/lib/gcc/i586-suse-linux/4.1.2/include/stdarg.h in Linux land)
is missing in Solaris' GCC (which is version 3.4.3).
Hack #1:
ln -s \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/install-tools/include/stdarg.h \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/include/stdarg.h
Hack #2: GNU programs...
mkdir -p ~/gnulink;
for i in addr2line ar as egrep grep ld make nm objcopy objdump \
ranlib readelf size string strip tar; do
ln -s /usr/sfw/bin/$i ~/gnulink/$i;
done;
for i in cat chgrp chmod chown chroot cksum cmp cp cut date \
dd df diff du echo env expand expr false fgrep find fold \
getopt groups head hostid install join ln locate ls mkdir \
mkfifo mknod mv nice nohup od pwd rm rmdir sed seq shred \
sleep sort split stty tac tail tee touch tr true uname \
uniq uptime wc who whoami xargs yes gawk; do
ln -s /opt/csw/bin/$i ~/gnulink/$i;
done;
Hack #3: Diff file...
Hack #4: GNU ld glitch workaround (GNU ld looks in the current dir...)
cd linux-2.6.21-rc3
ln -s /usr/sfw/i386-sun-solaris2.11/lib/ldscripts ldscripts
Fun #1:
export PATH=$HOME/gnulink:$PATH;
make ARCH=i386
Oddity #1:
ARCH=i386 required because the Makefiles seem to use `uname -m`
(which returns i86pc) rather than `uname -p`. I think we are
at odds here though...
uname -muname -p
SOL i86pc i386
LINUX i686athlon
Expect compiler failures, especially with assembler code.
Jan
PATCH BELOW
Index: linux-2.6.21-rc3/include/linux/input.h
===
--- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/include/linux/input.h2007-03-07 23:40:39.417339000
+0100
@@ -16,7 +16,9 @@
#include sys/time.h
#include sys/ioctl.h
#include sys/types.h
-#include asm/types.h
+#ifndef __sun__
+#include asm/types.h
+#endif
#endif
/*
Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
===
--- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07
23:28:35.659555000 +0100
@@ -21,6 +21,7 @@
along with this program; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+#include alloca.h
#include stdio.h
#include string.h
#include stdlib.h
Index: linux-2.6.21-rc3/scripts/kallsyms.c
===
--- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
+0100
@@ -378,6 +378,40 @@
table_cnt = pos;
}
+#ifdef __sun__
+/* Return the first occurrence of NEEDLE in HAYSTACK. */
+void *
+memmem (haystack, haystack_len, needle, needle_len)
+ const void *haystack;
+ size_t haystack_len;
+ const void *needle;
+ size_t needle_len;
+{
+ const char *begin;
+ const char *const last_possible
+= (const char *) haystack + haystack_len - needle_len;
+
+ if (needle_len == 0)
+/* The first occurrence of the empty string is deemed to occur at
+ the beginning of the string. */
+return (void *) haystack;
+
+ /* Sanity check, otherwise the loop might search through the whole
+ memory. */
+ if (__builtin_expect (haystack_len needle_len, 0))
+return NULL;
+
+ for (begin = (const char *) haystack; begin = last_possible; ++begin)
+if (begin[0] == ((const char *) needle)[0]
+!memcmp ((const void *) begin[1],
+ (const void *) ((const char *) needle + 1),
+ needle_len - 1))
+ return (void *) begin;
+
+ return NULL;
+}
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 9 2007 20:00, Sam Ravnborg wrote:
On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote:
Since Solaris seems to be on the run, I did myself try compile it.
However, unlike the original poster who said he did so on SunOS 4.8, I
did it on 5.11_snv39, yielding a bigger changeset. I thought I just
share the diff that piled up so far. It needs a lot of hacks on the
Solaris side - prioritizing GNU names, then, second, gnu ld has a
glitch, then, gcc has a missing file... it's fun fun fun!
Can I please have a signed-off version of this patch.
_Are you sure_ you want all these hacks without further
review from other people? Also note the patch is incomplete,
for example I could not compile the acpi pieces because
acsolaris.h -- which is referenced in the acpi includes --
does not exist. (Yet another piece of software that has
crossplatform compatibilty stuff, like XFS.)
--- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000
+0100
@@ -16,7 +16,9 @@
#include sys/time.h
#include sys/ioctl.h
#include sys/types.h
-#include asm/types.h
+#ifndef __sun__
+# include asm/types.h
+#endif
#endif
This is not a proper fix for sure. The problem lies in
file2alias.c, see (your own) http://lkml.org/lkml/2007/3/8/339
Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
===
--- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07
23:28:35.659555000 +0100
@@ -21,6 +21,7 @@
along with this program; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+#include alloca.h
#include stdio.h
#include string.h
#include stdlib.h
This is however, is valid. Can I gave sign-offs for single hunks?
Index: linux-2.6.21-rc3/scripts/kallsyms.c
===
--- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
+0100
@@ -378,6 +378,40 @@
table_cnt = pos;
}
+#ifdef __sun__
+/* Return the first occurrence of NEEDLE in HAYSTACK. */
+void *
+memmem (haystack, haystack_len, needle, needle_len)
+ const void *haystack;
+ size_t haystack_len;
+ const void *needle;
+ size_t needle_len;
+{
+ const char *begin;
+ const char *const last_possible
+= (const char *) haystack + haystack_len - needle_len;
+
+ if (needle_len == 0)
+/* The first occurrence of the empty string is deemed to occur at
+ the beginning of the string. */
+return (void *) haystack;
+
+ /* Sanity check, otherwise the loop might search through the whole
+ memory. */
+ if (__builtin_expect (haystack_len needle_len, 0))
+return NULL;
+
+ for (begin = (const char *) haystack; begin = last_possible; ++begin)
+if (begin[0] == ((const char *) needle)[0]
+!memcmp ((const void *) begin[1],
+ (const void *) ((const char *) needle + 1),
+ needle_len - 1))
+ return (void *) begin;
+
+ return NULL;
+}
+#endif
+
/* replace a given token in all the valid symbols. Use the sampled symbols
* to update the counts */
static void compress_symbols(unsigned char *str, int idx)
This one, I am just waiting for someone to object to the extra #if-#endif.
Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
===
--- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/kconfig/Makefile2007-03-07
23:21:19.730679000 +0100
@@ -88,7 +88,7 @@
HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags)
HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags
$(HOSTCC))
-HOST_EXTRACFLAGS += -DLOCALE
+HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__
PHONY += $(obj)/dochecklxdialog
$(obj)/dochecklxdialog:
The error message for this one was: stdbool.h only valid in C99 mode.
Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not
know offhand who is right.
Index: linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h
===
--- linux-2.6.21-rc3.orig/scripts/kconfig/lxdialog/dialog.h 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h 2007-03-07
23:14:48.462956000 +0100
@@ -222,3 +222,7 @@
* -- uppercase chars are used to invoke the button (M_EVENT + 'O')
*/
#define M_EVENT (KEY_MAX+1)
+
+#ifndef KEY_RESIZE
+# define KEY_RESIZE 0632
+#endif
Solaris only has curses, not ncurses.
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Fri, Mar 09, 2007 at 09:16:35PM +0100, Jan Engelhardt wrote: On Mar 9 2007 20:00, Sam Ravnborg wrote: On Thu, Mar 08, 2007 at 11:01:57PM +0100, Jan Engelhardt wrote: Since Solaris seems to be on the run, I did myself try compile it. However, unlike the original poster who said he did so on SunOS 4.8, I did it on 5.11_snv39, yielding a bigger changeset. I thought I just share the diff that piled up so far. It needs a lot of hacks on the Solaris side - prioritizing GNU names, then, second, gnu ld has a glitch, then, gcc has a missing file... it's fun fun fun! Can I please have a signed-off version of this patch. _Are you sure_ you want all these hacks without further review from other people? Also note the patch is incomplete, for example I could not compile the acpi pieces because acsolaris.h -- which is referenced in the acpi includes -- does not exist. (Yet another piece of software that has crossplatform compatibilty stuff, like XFS.) The Signed-off-by: document the origin of the path. I'm planning only to take the sensible bits of the patch anyway. --- linux-2.6.21-rc3.orig/include/linux/input.h2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000 +0100 @@ -16,7 +16,9 @@ #include sys/time.h #include sys/ioctl.h #include sys/types.h -#include asm/types.h +#ifndef __sun__ +# include asm/types.h +#endif #endif This is not a proper fix for sure. The problem lies in file2alias.c, see (your own) http://lkml.org/lkml/2007/3/8/339 I already committed my own fix - so this chunk can be ignored. Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c === --- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c 2007-03-07 23:28:35.659555000 +0100 @@ -21,6 +21,7 @@ along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ +#include alloca.h #include stdio.h #include string.h #include stdlib.h This is however, is valid. Can I gave sign-offs for single hunks? Agree and yes. Index: linux-2.6.21-rc3/scripts/kallsyms.c === --- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/scripts/kallsyms.c2007-03-07 23:46:46.249005000 +0100 @@ -378,6 +378,40 @@ table_cnt = pos; } +#ifdef __sun__ +/* Return the first occurrence of NEEDLE in HAYSTACK. */ +void * +memmem (haystack, haystack_len, needle, needle_len) + const void *haystack; + return (void *) begin; + + return NULL; +} +#endif + /* replace a given token in all the valid symbols. Use the sampled symbols * to update the counts */ static void compress_symbols(unsigned char *str, int idx) This one, I am just waiting for someone to object to the extra #if-#endif. I was planning to ask Paulo if strstr could not be used - Paulo? Index: linux-2.6.21-rc3/scripts/kconfig/Makefile === --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile 2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/scripts/kconfig/Makefile 2007-03-07 23:21:19.730679000 +0100 @@ -88,7 +88,7 @@ HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags) HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags $(HOSTCC)) -HOST_EXTRACFLAGS += -DLOCALE +HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__ PHONY += $(obj)/dochecklxdialog $(obj)/dochecklxdialog: The error message for this one was: stdbool.h only valid in C99 mode. Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not know offhand who is right. The -std= looks safe. The __EXTENSIONS__ part seems safe and needed after a bit googling. Index: linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h === --- linux-2.6.21-rc3.orig/scripts/kconfig/lxdialog/dialog.h 2007-03-07 05:41:20.0 +0100 +++ linux-2.6.21-rc3/scripts/kconfig/lxdialog/dialog.h 2007-03-07 23:14:48.462956000 +0100 @@ -222,3 +222,7 @@ * -- uppercase chars are used to invoke the button (M_EVENT + 'O') */ #define M_EVENT (KEY_MAX+1) + +#ifndef KEY_RESIZE +# define KEY_RESIZE 0632 +#endif Solaris only has curses, not ncurses. Consider this a supreme hack. In fact, menuconfig has some weird display errors still. This hack looks OK to me. Index: linux-2.6.21-rc3/scripts/mod/file2alias.c === --- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 9 2007 23:23, Sam Ravnborg wrote:
Date: Fri, 9 Mar 2007 23:23:32 +0100
From: Sam Ravnborg [EMAIL PROTECTED]
To: Jan Engelhardt [EMAIL PROTECTED],
Paulo Marques [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED], Christoph Hellwig [EMAIL PROTECTED],
Deepak Saxena [EMAIL PROTECTED],
Andrew Morton [EMAIL PROTECTED], linux-kernel@vger.kernel.org
Subject: Re: [PATCH] Fix building kernel under Solaris 11_snv
Reference: http://lkml.org/lkml/2007/3/8/368
Signed-off-by: Jan Engelhardt [EMAIL PROTECTED]
Happy cherrypicking.
More comments below.
Index: linux-2.6.21-rc3/scripts/kallsyms.c
===
--- linux-2.6.21-rc3.orig/scripts/kallsyms.c 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000
+0100
@@ -378,6 +378,40 @@
table_cnt = pos;
}
+#ifdef __sun__
+/* Return the first occurrence of NEEDLE in HAYSTACK. */
+void *
+memmem (haystack, haystack_len, needle, needle_len)
+ const void *haystack;
+ return (void *) begin;
+
+ return NULL;
+}
+#endif
+
/* replace a given token in all the valid symbols. Use the sampled
symbols
* to update the counts */
static void compress_symbols(unsigned char *str, int idx)
This one, I am just waiting for someone to object to the extra #if-#endif.
I was planning to ask Paulo if strstr could not be used - Paulo?
I am not sure, but I would tend to say no.
kallsyms.c:
p1 = table[i].sym;
/* find the token on the symbol */
p2 = memmem(p1, len, str, 2);
if (!p2) continue;
My first impression would be that 'p1' is a multi-nul-terminated string
array (foo\0bar\0next\0symbol\0...), much like the entire .strtab
section (I have not actually bothered to check, so it is a raw guess), and
I do not think strstr would replace memmem here if that was the case.
Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
===
--- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/kconfig/Makefile 2007-03-07
23:21:19.730679000 +0100
@@ -88,7 +88,7 @@
HOST_EXTRACFLAGS = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ccflags)
HOST_LOADLIBES = $(shell $(CONFIG_SHELL) $(check-lxdialog) -ldflags
$(HOSTCC))
-HOST_EXTRACFLAGS += -DLOCALE
+HOST_EXTRACFLAGS += -DLOCALE -std=c99 -D__EXTENSIONS__
PHONY += $(obj)/dochecklxdialog
$(obj)/dochecklxdialog:
The error message for this one was: stdbool.h only valid in C99 mode.
Linux GCC 4.1.2 does not print that, Solaris GCC 3.4.3 does. I do not
know offhand who is right.
The -std= looks safe. The __EXTENSIONS__ part seems safe and needed after a
bit googling.
Humm humm. Linux's /usr/include does not have any (glibc) file with the
string __EXTENSIONS__ in it. Though, the linux manpage for fileno() does
not mention fileno() conforming to anything. Under Linux, fileno() is
wrapped inside __USE_POSIX, and the comments in Solaris's headers also
indicate it is something posixy.
Perhaps have something like
#ifdef __sun__
# define __EXTENSIONS__ 1
#endif
Index: linux-2.6.21-rc3/scripts/mod/file2alias.c
===
--- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/mod/file2alias.c 2007-03-07
23:41:23.772026000 +0100
@@ -32,6 +32,8 @@
typedef uint32_t __u32;
typedef uint16_t __u16;
typedef unsigned char__u8;
+typedef int32_t __s32;
+typedef int16_t __s16;
/* Big exception to the don't include kernel headers into userspace,
which
* even potentially has different endianness and word sizes, since
HAX.
This goes away when we do not include input.h I think.
Exactly.
Index: linux-2.6.21-rc3/scripts/mod/modpost.h
===
--- linux-2.6.21-rc3.orig/scripts/mod/modpost.h 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/mod/modpost.h2007-03-07
23:37:01.31529 +0100
@@ -41,6 +41,11 @@
#define ELF_R_TYPE ELF64_R_TYPE
#endif
+#ifdef __sun__
+typedef uint16_t Elf32_Section;
+typedef uint16_t Elf64_Section;
+#endif
+
/* The 64-bit MIPS ELF ABI uses an unusual reloc format. */
typedef struct
{
More HAX because Sol does not have them :-(
Yup - but needed.
How much do we want to support Solaris anyway? It's like trying to make
ICC a thoroughly supported compiler. Hey, do we want to support GCC
compilation under
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 8 2007 22:25, Sam Ravnborg wrote:
>Subject: Re: [PATCH] Fix building kernel under Solaris
Since Solaris seems to be on the run, I did myself try compile it.
However, unlike the original poster who said he did so on SunOS 4.8, I
did it on 5.11_snv39, yielding a bigger changeset. I thought I just
share the diff that piled up so far. It needs a lot of hacks on the
Solaris side - prioritizing GNU names, then, second, gnu ld has a
glitch, then, gcc has a missing file... it's fun fun fun!
Well, I will iterate the key problem with the missing file:
* include/linux/kernel.h (and many others) include
BUT - since we are using -nostdinc, /usr/include/stdarg.h is not
considered. And gcc's stdarg.h (which lives at
/usr/lib/gcc/i586-suse-linux/4.1.2/include/stdarg.h in Linux land)
is missing in Solaris' GCC (which is version 3.4.3).
Hack #1:
ln -s \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/install-tools/include/stdarg.h \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/include/stdarg.h
Hack #2: GNU programs...
mkdir -p ~/gnulink;
for i in addr2line ar as egrep grep ld make nm objcopy objdump \
ranlib readelf size string strip tar; do
ln -s "/usr/sfw/bin/$i" "~/gnulink/$i";
done;
for i in cat chgrp chmod chown chroot cksum cmp cp cut date \
dd df diff du echo env expand expr false fgrep find fold \
getopt groups head hostid install join ln locate ls mkdir \
mkfifo mknod mv nice nohup od pwd rm rmdir sed seq shred \
sleep sort split stty tac tail tee touch tr true uname \
uniq uptime wc who whoami xargs yes gawk; do
ln -s "/opt/csw/bin/$i" "~/gnulink/$i";
done;
Hack #3: Diff file...
Hack #4: GNU ld glitch workaround (GNU ld looks in the current dir...)
cd linux-2.6.21-rc3
ln -s /usr/sfw/i386-sun-solaris2.11/lib/ldscripts ldscripts
Fun #1:
export PATH="$HOME/gnulink:$PATH";
make ARCH=i386
Oddity #1:
ARCH=i386 required because the Makefiles seem to use `uname -m`
(which returns "i86pc") rather than `uname -p`. I think we are
at odds here though...
uname -muname -p
SOL i86pc i386
LINUX i686athlon
Expect compiler failures, especially with assembler code.
Jan
<<< PATCH BELOW <<<
Index: linux-2.6.21-rc3/include/linux/input.h
===
--- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000
+0100
@@ -16,7 +16,9 @@
#include
#include
#include
-#include
+#ifndef __sun__
+# include
+#endif
#endif
/*
Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
===
--- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c2007-03-07
23:28:35.659555000 +0100
@@ -21,6 +21,7 @@
along with this program; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+#include
#include
#include
#include
Index: linux-2.6.21-rc3/scripts/kallsyms.c
===
--- linux-2.6.21-rc3.orig/scripts/kallsyms.c2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000 +0100
@@ -378,6 +378,40 @@
table_cnt = pos;
}
+#ifdef __sun__
+/* Return the first occurrence of NEEDLE in HAYSTACK. */
+void *
+memmem (haystack, haystack_len, needle, needle_len)
+ const void *haystack;
+ size_t haystack_len;
+ const void *needle;
+ size_t needle_len;
+{
+ const char *begin;
+ const char *const last_possible
+= (const char *) haystack + haystack_len - needle_len;
+
+ if (needle_len == 0)
+/* The first occurrence of the empty string is deemed to occur at
+ the beginning of the string. */
+return (void *) haystack;
+
+ /* Sanity check, otherwise the loop might search through the whole
+ memory. */
+ if (__builtin_expect (haystack_len < needle_len, 0))
+return NULL;
+
+ for (begin = (const char *) haystack; begin <= last_possible; ++begin)
+if (begin[0] == ((const char *) needle)[0] &&
+!memcmp ((const void *) [1],
+ (const void *) ((const char *) needle + 1),
+ needle_len - 1))
+ return (void *) begin;
+
+ return NULL;
+}
+#endif
+
/* replace a given token in all the valid symbols. Use the sampled symbols
* to update the counts */
static void compress_symbols(unsigned char *str, int idx)
Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
===
---
Re: [PATCH] Fix building kernel under Solaris 11_snv
On Mar 8 2007 22:25, Sam Ravnborg wrote:
Subject: Re: [PATCH] Fix building kernel under Solaris
Since Solaris seems to be on the run, I did myself try compile it.
However, unlike the original poster who said he did so on SunOS 4.8, I
did it on 5.11_snv39, yielding a bigger changeset. I thought I just
share the diff that piled up so far. It needs a lot of hacks on the
Solaris side - prioritizing GNU names, then, second, gnu ld has a
glitch, then, gcc has a missing file... it's fun fun fun!
Well, I will iterate the key problem with the missing file:
* include/linux/kernel.h (and many others) include stdarg.h
BUT - since we are using -nostdinc, /usr/include/stdarg.h is not
considered. And gcc's stdarg.h (which lives at
/usr/lib/gcc/i586-suse-linux/4.1.2/include/stdarg.h in Linux land)
is missing in Solaris' GCC (which is version 3.4.3).
Hack #1:
ln -s \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/install-tools/include/stdarg.h \
/usr/sfw/lib/gcc/i386-pc-solaris2.11/3.4.3/include/stdarg.h
Hack #2: GNU programs...
mkdir -p ~/gnulink;
for i in addr2line ar as egrep grep ld make nm objcopy objdump \
ranlib readelf size string strip tar; do
ln -s /usr/sfw/bin/$i ~/gnulink/$i;
done;
for i in cat chgrp chmod chown chroot cksum cmp cp cut date \
dd df diff du echo env expand expr false fgrep find fold \
getopt groups head hostid install join ln locate ls mkdir \
mkfifo mknod mv nice nohup od pwd rm rmdir sed seq shred \
sleep sort split stty tac tail tee touch tr true uname \
uniq uptime wc who whoami xargs yes gawk; do
ln -s /opt/csw/bin/$i ~/gnulink/$i;
done;
Hack #3: Diff file...
Hack #4: GNU ld glitch workaround (GNU ld looks in the current dir...)
cd linux-2.6.21-rc3
ln -s /usr/sfw/i386-sun-solaris2.11/lib/ldscripts ldscripts
Fun #1:
export PATH=$HOME/gnulink:$PATH;
make ARCH=i386
Oddity #1:
ARCH=i386 required because the Makefiles seem to use `uname -m`
(which returns i86pc) rather than `uname -p`. I think we are
at odds here though...
uname -muname -p
SOL i86pc i386
LINUX i686athlon
Expect compiler failures, especially with assembler code.
Jan
PATCH BELOW
Index: linux-2.6.21-rc3/include/linux/input.h
===
--- linux-2.6.21-rc3.orig/include/linux/input.h 2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/include/linux/input.h 2007-03-07 23:40:39.417339000
+0100
@@ -16,7 +16,9 @@
#include sys/time.h
#include sys/ioctl.h
#include sys/types.h
-#include asm/types.h
+#ifndef __sun__
+# include asm/types.h
+#endif
#endif
/*
Index: linux-2.6.21-rc3/scripts/genksyms/genksyms.c
===
--- linux-2.6.21-rc3.orig/scripts/genksyms/genksyms.c 2007-03-07
05:41:20.0 +0100
+++ linux-2.6.21-rc3/scripts/genksyms/genksyms.c2007-03-07
23:28:35.659555000 +0100
@@ -21,6 +21,7 @@
along with this program; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+#include alloca.h
#include stdio.h
#include string.h
#include stdlib.h
Index: linux-2.6.21-rc3/scripts/kallsyms.c
===
--- linux-2.6.21-rc3.orig/scripts/kallsyms.c2007-03-07 05:41:20.0
+0100
+++ linux-2.6.21-rc3/scripts/kallsyms.c 2007-03-07 23:46:46.249005000 +0100
@@ -378,6 +378,40 @@
table_cnt = pos;
}
+#ifdef __sun__
+/* Return the first occurrence of NEEDLE in HAYSTACK. */
+void *
+memmem (haystack, haystack_len, needle, needle_len)
+ const void *haystack;
+ size_t haystack_len;
+ const void *needle;
+ size_t needle_len;
+{
+ const char *begin;
+ const char *const last_possible
+= (const char *) haystack + haystack_len - needle_len;
+
+ if (needle_len == 0)
+/* The first occurrence of the empty string is deemed to occur at
+ the beginning of the string. */
+return (void *) haystack;
+
+ /* Sanity check, otherwise the loop might search through the whole
+ memory. */
+ if (__builtin_expect (haystack_len needle_len, 0))
+return NULL;
+
+ for (begin = (const char *) haystack; begin = last_possible; ++begin)
+if (begin[0] == ((const char *) needle)[0]
+!memcmp ((const void *) begin[1],
+ (const void *) ((const char *) needle + 1),
+ needle_len - 1))
+ return (void *) begin;
+
+ return NULL;
+}
+#endif
+
/* replace a given token in all the valid symbols. Use the sampled symbols
* to update the counts */
static void compress_symbols(unsigned char *str, int idx)
Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
