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