All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] Fix building kernel under Solaris
@ 2007-03-06 18:09 Deepak Saxena
  2007-03-06 18:22 ` Deepak Saxena
                   ` (2 more replies)
  0 siblings, 3 replies; 14+ messages in thread
From: Deepak Saxena @ 2007-03-06 18:09 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linux-kernel


Fix varous build breakages that occur when building on a Solaris system
(SunOS 4.8 to be exact)

- No asm/types.h

- lpptest doesn't make sense on a Solaris host (this should really be
  cross-built..)

- Need to define __s32 and __s16

- Need to define Elf32_Section and Elf64_Section

- Solaris does not provide strsep

Signed-off-by: Deepak Saxena <dsaxena@mvista.com>


diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h linux-2.6.18_pro500/include/linux/input.h
--- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h	2007-03-05 21:54:45.000000000 -0800
+++ linux-2.6.18_pro500/include/linux/input.h	2007-03-05 19:19:01.000000000 -0800
@@ -16,8 +16,10 @@
 #include <sys/time.h>
 #include <sys/ioctl.h>
 #include <sys/types.h>
+#ifndef __sun__
 #include <asm/types.h>
 #endif
+#endif
 
 /*
  * The event structure itself
diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile linux-2.6.18_pro500/scripts/Makefile
--- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile	2007-03-05 21:54:43.000000000 -0800
+++ linux-2.6.18_pro500/scripts/Makefile	2007-03-05 19:54:52.000000000 -0800
@@ -12,10 +12,12 @@
 hostprogs-$(CONFIG_VT)           += conmakehash
 hostprogs-$(CONFIG_PROM_CONSOLE) += conmakehash
 hostprogs-$(CONFIG_IKCONFIG)     += bin2c
+HOST_OS := $(shell uname)
+ifneq ($(HOST_OS),SunOS)
 ifdef CONFIG_LPPTEST
 hostprogs-y      += testlpp
 endif
-
+endif
 always		:= $(hostprogs-y)
 
 subdir-$(CONFIG_MODVERSIONS) += genksyms
diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/file2alias.c linux-2.6.18_pro500/scripts/mod/file2alias.c
--- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/file2alias.c	2007-03-05 21:54:42.000000000 -0800
+++ linux-2.6.18_pro500/scripts/mod/file2alias.c	2007-03-05 19:19:26.000000000 -0800
@@ -29,6 +29,11 @@
 
 #include <ctype.h>
 
+#if defined(__sun__)
+typedef int32_t	__s32;
+typedef int16_t	__s16;
+#endif
+
 #if defined(__CYGWIN__)
 typedef __uint32_t	__u32;
 #else
diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h linux-2.6.18_pro500/scripts/mod/modpost.h
--- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h	2007-03-05 21:54:42.000000000 -0800
+++ linux-2.6.18_pro500/scripts/mod/modpost.h	2007-03-05 19:13:16.000000000 -0800
@@ -15,6 +15,11 @@
 
 #include "elfconfig.h"
 
+#if defined(__sun__)
+typedef uint16_t Elf32_Section;
+typedef uint16_t Elf64_Section;
+#endif
+
 #if KERNEL_ELFCLASS == ELFCLASS32
 
 #define Elf_Ehdr    Elf32_Ehdr
diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c linux-2.6.18_pro500/scripts/mod/sumversion.c
--- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c	2007-03-05 21:54:42.000000000 -0800
+++ linux-2.6.18_pro500/scripts/mod/sumversion.c	2007-03-05 19:33:55.000000000 -0800
@@ -1,6 +1,7 @@
 #if defined(__sun__)
 #include <netinet/in.h>
 #include <inttypes.h>
+#include <limits.h>
 #elif defined(__CYGWIN__)
 #include <asm/byteorder.h>	/* For ntohl/htonl */
 #include <limits.h>
@@ -421,9 +422,16 @@
 	*end = '\0';
 
 	md4_init(&md);
+/*
+ * Solaris does not strsep
+ */
+#ifndef __sun__
 	while ((fname = strsep(&sources, " ")) != NULL) {
 		if (!*fname)
 			continue;
+#else
+	for (fname = strtok(sources, " "); fname; fname = strtok(NULL, " ")) {
+#endif
 		if (!parse_source_files(fname, &md))
 			goto release;
 	}


-- 
Deepak Saxena - dsaxena@plexity.net - http://www.plexity.net

In the end, they will not say, "those were dark times,"  they will ask
"why were their poets silent?" - Bertolt Brecht

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-06 18:09 [PATCH] Fix building kernel under Solaris Deepak Saxena
@ 2007-03-06 18:22 ` Deepak Saxena
  2007-03-06 19:23 ` Sam Ravnborg
  2007-03-07  9:42 ` Christoph Hellwig
  2 siblings, 0 replies; 14+ messages in thread
From: Deepak Saxena @ 2007-03-06 18:22 UTC (permalink / raw)
  To: Andrew Morton; +Cc: linux-kernel


SIGRETRACT

I have -rt merged into my tree so this won't apply cleanly to mainline.

~Deepak

On Mar 06 2007, at 10:09, Deepak Saxena was caught saying:
> 
> Fix varous build breakages that occur when building on a Solaris system
> (SunOS 4.8 to be exact)
> 
> - No asm/types.h
> 
> - lpptest doesn't make sense on a Solaris host (this should really be
>   cross-built..)
> 
> - Need to define __s32 and __s16
> 
> - Need to define Elf32_Section and Elf64_Section
> 
> - Solaris does not provide strsep
> 
> Signed-off-by: Deepak Saxena <dsaxena@mvista.com>
> 
> 
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h linux-2.6.18_pro500/include/linux/input.h
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h	2007-03-05 21:54:45.000000000 -0800
> +++ linux-2.6.18_pro500/include/linux/input.h	2007-03-05 19:19:01.000000000 -0800
> @@ -16,8 +16,10 @@
>  #include <sys/time.h>
>  #include <sys/ioctl.h>
>  #include <sys/types.h>
> +#ifndef __sun__
>  #include <asm/types.h>
>  #endif
> +#endif
>  
>  /*
>   * The event structure itself
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile linux-2.6.18_pro500/scripts/Makefile
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile	2007-03-05 21:54:43.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/Makefile	2007-03-05 19:54:52.000000000 -0800
> @@ -12,10 +12,12 @@
>  hostprogs-$(CONFIG_VT)           += conmakehash
>  hostprogs-$(CONFIG_PROM_CONSOLE) += conmakehash
>  hostprogs-$(CONFIG_IKCONFIG)     += bin2c
> +HOST_OS := $(shell uname)
> +ifneq ($(HOST_OS),SunOS)
>  ifdef CONFIG_LPPTEST
>  hostprogs-y      += testlpp
>  endif
> -
> +endif
>  always		:= $(hostprogs-y)
>  
>  subdir-$(CONFIG_MODVERSIONS) += genksyms
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/file2alias.c linux-2.6.18_pro500/scripts/mod/file2alias.c
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/file2alias.c	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/file2alias.c	2007-03-05 19:19:26.000000000 -0800
> @@ -29,6 +29,11 @@
>  
>  #include <ctype.h>
>  
> +#if defined(__sun__)
> +typedef int32_t	__s32;
> +typedef int16_t	__s16;
> +#endif
> +
>  #if defined(__CYGWIN__)
>  typedef __uint32_t	__u32;
>  #else
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h linux-2.6.18_pro500/scripts/mod/modpost.h
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/modpost.h	2007-03-05 19:13:16.000000000 -0800
> @@ -15,6 +15,11 @@
>  
>  #include "elfconfig.h"
>  
> +#if defined(__sun__)
> +typedef uint16_t Elf32_Section;
> +typedef uint16_t Elf64_Section;
> +#endif
> +
>  #if KERNEL_ELFCLASS == ELFCLASS32
>  
>  #define Elf_Ehdr    Elf32_Ehdr
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c linux-2.6.18_pro500/scripts/mod/sumversion.c
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/sumversion.c	2007-03-05 19:33:55.000000000 -0800
> @@ -1,6 +1,7 @@
>  #if defined(__sun__)
>  #include <netinet/in.h>
>  #include <inttypes.h>
> +#include <limits.h>
>  #elif defined(__CYGWIN__)
>  #include <asm/byteorder.h>	/* For ntohl/htonl */
>  #include <limits.h>
> @@ -421,9 +422,16 @@
>  	*end = '\0';
>  
>  	md4_init(&md);
> +/*
> + * Solaris does not strsep
> + */
> +#ifndef __sun__
>  	while ((fname = strsep(&sources, " ")) != NULL) {
>  		if (!*fname)
>  			continue;
> +#else
> +	for (fname = strtok(sources, " "); fname; fname = strtok(NULL, " ")) {
> +#endif
>  		if (!parse_source_files(fname, &md))
>  			goto release;
>  	}
> 
> 
> -- 
> Deepak Saxena - dsaxena@plexity.net - http://www.plexity.net
> 
> In the end, they will not say, "those were dark times,"  they will ask
> "why were their poets silent?" - Bertolt Brecht
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
Deepak Saxena - dsaxena@plexity.net - http://www.plexity.net

In the end, they will not say, "those were dark times,"  they will ask
"why were their poets silent?" - Bertolt Brecht

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-06 18:09 [PATCH] Fix building kernel under Solaris Deepak Saxena
  2007-03-06 18:22 ` Deepak Saxena
@ 2007-03-06 19:23 ` Sam Ravnborg
  2007-03-07  9:42 ` Christoph Hellwig
  2 siblings, 0 replies; 14+ messages in thread
From: Sam Ravnborg @ 2007-03-06 19:23 UTC (permalink / raw)
  To: Deepak Saxena; +Cc: Andrew Morton, linux-kernel

On Tue, Mar 06, 2007 at 10:09:40AM -0800, Deepak Saxena wrote:
> 
> Fix varous build breakages that occur when building on a Solaris system
> (SunOS 4.8 to be exact)
> 
> - No asm/types.h
> 
> - lpptest doesn't make sense on a Solaris host (this should really be
>   cross-built..)
> 
> - Need to define __s32 and __s16
> 
> - Need to define Elf32_Section and Elf64_Section
> 
> - Solaris does not provide strsep
> 
> Signed-off-by: Deepak Saxena <dsaxena@mvista.com>
> 
> 
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h linux-2.6.18_pro500/include/linux/input.h
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/include/linux/input.h	2007-03-05 21:54:45.000000000 -0800
> +++ linux-2.6.18_pro500/include/linux/input.h	2007-03-05 19:19:01.000000000 -0800
> @@ -16,8 +16,10 @@
>  #include <sys/time.h>
>  #include <sys/ioctl.h>
>  #include <sys/types.h>
> +#ifndef __sun__
>  #include <asm/types.h>
>  #endif
> +#endif
Either we need it - or we do not need it - no?
So we can just kill the include.

>  
>  /*
>   * The event structure itself
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile linux-2.6.18_pro500/scripts/Makefile
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/Makefile	2007-03-05 21:54:43.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/Makefile	2007-03-05 19:54:52.000000000 -0800
> @@ -12,10 +12,12 @@
>  hostprogs-$(CONFIG_VT)           += conmakehash
>  hostprogs-$(CONFIG_PROM_CONSOLE) += conmakehash
>  hostprogs-$(CONFIG_IKCONFIG)     += bin2c
> +HOST_OS := $(shell uname)
> +ifneq ($(HOST_OS),SunOS)
>  ifdef CONFIG_LPPTEST
>  hostprogs-y      += testlpp
>  endif
> -
> +endif
>  always		:= $(hostprogs-y)
testlpp is not in mainline - what is it?
[downloading -mm now]

> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/file2alias.c	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/file2alias.c	2007-03-05 19:19:26.000000000 -0800
> @@ -29,6 +29,11 @@
>  
>  #include <ctype.h>
>  
> +#if defined(__sun__)
> +typedef int32_t	__s32;
> +typedef int16_t	__s16;
> +#endif
> +
>  #if defined(__CYGWIN__)
>  typedef __uint32_t	__u32;
>  #else
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h linux-2.6.18_pro500/scripts/mod/modpost.h
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/modpost.h	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/modpost.h	2007-03-05 19:13:16.000000000 -0800
> @@ -15,6 +15,11 @@
>  
>  #include "elfconfig.h"
>  
> +#if defined(__sun__)
> +typedef uint16_t Elf32_Section;
> +typedef uint16_t Elf64_Section;
> +#endif
> +
>  #if KERNEL_ELFCLASS == ELFCLASS32
>  
>  #define Elf_Ehdr    Elf32_Ehdr
> diff -uNr -X /home/src/dontdiff /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c linux-2.6.18_pro500/scripts/mod/sumversion.c
> --- /home/src/mvista/kernel/PRO5-CLEAN/linux-2.6.18/scripts/mod/sumversion.c	2007-03-05 21:54:42.000000000 -0800
> +++ linux-2.6.18_pro500/scripts/mod/sumversion.c	2007-03-05 19:33:55.000000000 -0800
> @@ -1,6 +1,7 @@
>  #if defined(__sun__)
>  #include <netinet/in.h>
>  #include <inttypes.h>
> +#include <limits.h>
>  #elif defined(__CYGWIN__)
>  #include <asm/byteorder.h>	/* For ntohl/htonl */
>  #include <limits.h>
> @@ -421,9 +422,16 @@
>  	*end = '\0';
>  
>  	md4_init(&md);
> +/*
> + * Solaris does not strsep
> + */
> +#ifndef __sun__
>  	while ((fname = strsep(&sources, " ")) != NULL) {
>  		if (!*fname)
>  			continue;
> +#else
> +	for (fname = strtok(sources, " "); fname; fname = strtok(NULL, " ")) {
> +#endif
>  		if (!parse_source_files(fname, &md))
>  			goto release;
>  	}
Why the ifdef - I assume it is always OK to use the latter version?
Please send next diff cc: to me (at least the parts that touches scripts/

	Sam

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-06 18:09 [PATCH] Fix building kernel under Solaris Deepak Saxena
  2007-03-06 18:22 ` Deepak Saxena
  2007-03-06 19:23 ` Sam Ravnborg
@ 2007-03-07  9:42 ` Christoph Hellwig
  2007-03-07 22:45   ` Jan Engelhardt
  2 siblings, 1 reply; 14+ messages in thread
From: Christoph Hellwig @ 2007-03-07  9:42 UTC (permalink / raw)
  To: Deepak Saxena; +Cc: Andrew Morton, linux-kernel

On Tue, Mar 06, 2007 at 10:09:40AM -0800, Deepak Saxena wrote:
> @@ -16,8 +16,10 @@
>  #include <sys/time.h>
>  #include <sys/ioctl.h>
>  #include <sys/types.h>
> +#ifndef __sun__
>  #include <asm/types.h>
>  #endif
> +#endif

So if solaris doesn't need it, why do we need it on Linux?

> +/*
> + * Solaris does not strsep
> + */
> +#ifndef __sun__
>  	while ((fname = strsep(&sources, " ")) != NULL) {
>  		if (!*fname)
>  			continue;
> +#else
> +	for (fname = strtok(sources, " "); fname; fname = strtok(NULL, " ")) {
> +#endif
>  		if (!parse_source_files(fname, &md))
>  			goto release;
>  	}

Please either provide a strsep for solaris, or use strtok unconditionally.
ut this ifdef mess is not acceptable.

>

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-07  9:42 ` Christoph Hellwig
@ 2007-03-07 22:45   ` Jan Engelhardt
  2007-03-08  8:35     ` Christoph Hellwig
  0 siblings, 1 reply; 14+ messages in thread
From: Jan Engelhardt @ 2007-03-07 22:45 UTC (permalink / raw)
  To: Christoph Hellwig; +Cc: Deepak Saxena, Andrew Morton, linux-kernel


On Mar 7 2007 09:42, Christoph Hellwig wrote:
>>  #include <sys/time.h>
>>  #include <sys/ioctl.h>
>>  #include <sys/types.h>
>> +#ifndef __sun__
>>  #include <asm/types.h>
>>  #endif
>> +#endif
>
>So if solaris doesn't need it, why do we need it on Linux?

For __s32 and __s16.


Jan
-- 

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-07 22:45   ` Jan Engelhardt
@ 2007-03-08  8:35     ` Christoph Hellwig
  2007-03-08 20:43       ` Jan Engelhardt
  0 siblings, 1 reply; 14+ messages in thread
From: Christoph Hellwig @ 2007-03-08  8:35 UTC (permalink / raw)
  To: Jan Engelhardt
  Cc: Christoph Hellwig, Deepak Saxena, Andrew Morton, linux-kernel

On Wed, Mar 07, 2007 at 11:45:11PM +0100, Jan Engelhardt wrote:
> 
> On Mar 7 2007 09:42, Christoph Hellwig wrote:
> >>  #include <sys/time.h>
> >>  #include <sys/ioctl.h>
> >>  #include <sys/types.h>
> >> +#ifndef __sun__
> >>  #include <asm/types.h>
> >>  #endif
> >> +#endif
> >
> >So if solaris doesn't need it, why do we need it on Linux?
> 
> For __s32 and __s16.

We shouldn't use those from asm/types.h on Linux either, but
get them consitantly by defining them to the C99 or BSD types
from the libc headers.

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-08  8:35     ` Christoph Hellwig
@ 2007-03-08 20:43       ` Jan Engelhardt
  2007-03-08 21:25         ` Sam Ravnborg
  0 siblings, 1 reply; 14+ messages in thread
From: Jan Engelhardt @ 2007-03-08 20:43 UTC (permalink / raw)
  To: Christoph Hellwig; +Cc: Deepak Saxena, Andrew Morton, linux-kernel


On Mar 8 2007 08:35, Christoph Hellwig wrote:
>On Wed, Mar 07, 2007 at 11:45:11PM +0100, Jan Engelhardt wrote:
>> 
>> On Mar 7 2007 09:42, Christoph Hellwig wrote:
>> >>  #include <sys/time.h>
>> >>  #include <sys/ioctl.h>
>> >>  #include <sys/types.h>
>> >> +#ifndef __sun__
>> >>  #include <asm/types.h>
>> >>  #endif
>> >> +#endif
>> >
>> >So if solaris doesn't need it, why do we need it on Linux?
>> 
>> For __s32 and __s16.
>
>We shouldn't use those from asm/types.h on Linux either, but
>get them consitantly by defining them to the C99 or BSD types
>from the libc headers.

It's a bit icky. file2alias.c includes "../../include/linux/input.h" and it is
_input.h_ that requires __s16/__s32 and hence includes <asm/types.h>.
If you ask me, it's file2alias.c that's broken.


Jan
-- 

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris
  2007-03-08 20:43       ` Jan Engelhardt
@ 2007-03-08 21:25         ` Sam Ravnborg
  2007-03-08 22:01           ` [PATCH] Fix building kernel under Solaris 11_snv Jan Engelhardt
  0 siblings, 1 reply; 14+ messages in thread
From: Sam Ravnborg @ 2007-03-08 21:25 UTC (permalink / raw)
  To: Jan Engelhardt, dmitry.torokhov
  Cc: Christoph Hellwig, Deepak Saxena, Andrew Morton, linux-kernel

On Thu, Mar 08, 2007 at 09:43:46PM +0100, Jan Engelhardt wrote:
> 
> On Mar 8 2007 08:35, Christoph Hellwig wrote:
> >On Wed, Mar 07, 2007 at 11:45:11PM +0100, Jan Engelhardt wrote:
> >> 
> >> On Mar 7 2007 09:42, Christoph Hellwig wrote:
> >> >>  #include <sys/time.h>
> >> >>  #include <sys/ioctl.h>
> >> >>  #include <sys/types.h>
> >> >> +#ifndef __sun__
> >> >>  #include <asm/types.h>
> >> >>  #endif
> >> >> +#endif
> >> >
> >> >So if solaris doesn't need it, why do we need it on Linux?
> >> 
> >> For __s32 and __s16.
> >
> >We shouldn't use those from asm/types.h on Linux either, but
> >get them consitantly by defining them to the C99 or BSD types
> >from the libc headers.
> 
> It's a bit icky. file2alias.c includes "../../include/linux/input.h" and it is
> _input.h_ that requires __s16/__s32 and hence includes <asm/types.h>.
> If you ask me, it's file2alias.c that's broken.
Took a closer look and agree on that.
I have queued up below patch to fix it.
I expect to see an updated patch on the other issues from Deepak.

	Sam

diff --git a/include/linux/input.h b/include/linux/input.h
index bde65c8..13d510c 100644
--- a/include/linux/input.h
+++ b/include/linux/input.h
@@ -1010,6 +1010,10 @@ struct input_dev {
 #error "EV_MAX and INPUT_DEVICE_ID_EV_MAX do not match"
 #endif
 
+#if KEY_MIN_INTERESTING != INPUT_DEVICE_ID_KEY_MIN_INTERESTING
+#error "KEY_MIN_INTERESTING and INPUT_DEVICE_ID_KEY_MIN_INTERESTING do not match"
+#endif
+
 #if KEY_MAX != INPUT_DEVICE_ID_KEY_MAX
 #error "KEY_MAX and INPUT_DEVICE_ID_KEY_MAX do not match"
 #endif
diff --git a/include/linux/mod_devicetable.h b/include/linux/mod_devicetable.h
index e0c393c..5e89204 100644
--- a/include/linux/mod_devicetable.h
+++ b/include/linux/mod_devicetable.h
@@ -262,6 +262,7 @@ struct i2c_device_id {
 
 /* Input */
 #define INPUT_DEVICE_ID_EV_MAX		0x1f
+#define INPUT_DEVICE_ID_KEY_MIN_INTERESTING	0x71
 #define INPUT_DEVICE_ID_KEY_MAX		0x1ff
 #define INPUT_DEVICE_ID_REL_MAX		0x0f
 #define INPUT_DEVICE_ID_ABS_MAX		0x3f
diff --git a/scripts/mod/file2alias.c b/scripts/mod/file2alias.c
index f61c9cc..96319eb 100644
--- a/scripts/mod/file2alias.c
+++ b/scripts/mod/file2alias.c
@@ -37,7 +37,6 @@ typedef unsigned char	__u8;
  * even potentially has different endianness and word sizes, since
  * we handle those differences explicitly below */
 #include "../../include/linux/mod_devicetable.h"
-#include "../../include/linux/input.h"
 
 #define ADD(str, sep, cond, field)                              \
 do {                                                            \
@@ -416,31 +415,33 @@ static int do_input_entry(const char *filename, struct input_device_id *id,
 
 	sprintf(alias + strlen(alias), "-e*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_EVBIT)
-		do_input(alias, id->evbit, 0, EV_MAX);
+		do_input(alias, id->evbit, 0, INPUT_DEVICE_ID_EV_MAX);
 	sprintf(alias + strlen(alias), "k*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_KEYBIT)
-		do_input(alias, id->keybit, KEY_MIN_INTERESTING, KEY_MAX);
+		do_input(alias, id->keybit,
+			 INPUT_DEVICE_ID_KEY_MIN_INTERESTING,
+			 INPUT_DEVICE_ID_KEY_MAX);
 	sprintf(alias + strlen(alias), "r*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_RELBIT)
-		do_input(alias, id->relbit, 0, REL_MAX);
+		do_input(alias, id->relbit, 0, INPUT_DEVICE_ID_REL_MAX);
 	sprintf(alias + strlen(alias), "a*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_ABSBIT)
-		do_input(alias, id->absbit, 0, ABS_MAX);
+		do_input(alias, id->absbit, 0, INPUT_DEVICE_ID_ABS_MAX);
 	sprintf(alias + strlen(alias), "m*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_MSCIT)
-		do_input(alias, id->mscbit, 0, MSC_MAX);
+		do_input(alias, id->mscbit, 0, INPUT_DEVICE_ID_MSC_MAX);
 	sprintf(alias + strlen(alias), "l*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_LEDBIT)
-		do_input(alias, id->ledbit, 0, LED_MAX);
+		do_input(alias, id->ledbit, 0, INPUT_DEVICE_ID_LED_MAX);
 	sprintf(alias + strlen(alias), "s*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_SNDBIT)
-		do_input(alias, id->sndbit, 0, SND_MAX);
+		do_input(alias, id->sndbit, 0, INPUT_DEVICE_ID_SND_MAX);
 	sprintf(alias + strlen(alias), "f*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_FFBIT)
-		do_input(alias, id->ffbit, 0, FF_MAX);
+		do_input(alias, id->ffbit, 0, INPUT_DEVICE_ID_FF_MAX);
 	sprintf(alias + strlen(alias), "w*");
 	if (id->flags & INPUT_DEVICE_ID_MATCH_SWBIT)
-		do_input(alias, id->swbit, 0, SW_MAX);
+		do_input(alias, id->swbit, 0, INPUT_DEVICE_ID_SW_MAX);
 	return 1;
 }
 

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-08 21:25         ` Sam Ravnborg
@ 2007-03-08 22:01           ` Jan Engelhardt
  2007-03-09 19:00             ` Sam Ravnborg
  0 siblings, 1 reply; 14+ messages in thread
From: Jan Engelhardt @ 2007-03-08 22:01 UTC (permalink / raw)
  To: Sam Ravnborg
  Cc: dmitry.torokhov, Christoph Hellwig, Deepak Saxena, Andrew Morton,
	linux-kernel


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 -m	uname -p
	SOL	i86pc		i386
	LINUX	i686		athlon


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 05:41:20.000000000 +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 05:41:20.000000000 +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 05:41:20.000000000 +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
===================================================================
--- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile	2007-03-07 05:41:20.000000000 +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:
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 05:41:20.000000000 +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
Index: linux-2.6.21-rc3/scripts/mod/file2alias.c
===================================================================
--- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c	2007-03-07 05:41:20.000000000 +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
Index: linux-2.6.21-rc3/scripts/mod/modpost.h
===================================================================
--- linux-2.6.21-rc3.orig/scripts/mod/modpost.h	2007-03-07 05:41:20.000000000 +0100
+++ linux-2.6.21-rc3/scripts/mod/modpost.h	2007-03-07 23:37:01.315290000 +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
 {
Index: linux-2.6.21-rc3/scripts/mod/sumversion.c
===================================================================
--- linux-2.6.21-rc3.orig/scripts/mod/sumversion.c	2007-03-07 05:41:20.000000000 +0100
+++ linux-2.6.21-rc3/scripts/mod/sumversion.c	2007-03-07 23:43:55.668334000 +0100
@@ -6,6 +6,7 @@
 #endif
 #include <ctype.h>
 #include <errno.h>
+#include <limits.h>
 #include <string.h>
 #include "modpost.h"
 
@@ -417,7 +418,8 @@
 	*end = '\0';
 
 	md4_init(&md);
-	while ((fname = strsep(&sources, " ")) != NULL) {
+	for(fname = strtok(sources, " "); fname != NULL;
+	  fname = strtok(NULL, " ")) {
 		if (!*fname)
 			continue;
 		if (!parse_source_files(fname, &md))
#<EOF>

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-08 22:01           ` [PATCH] Fix building kernel under Solaris 11_snv Jan Engelhardt
@ 2007-03-09 19:00             ` Sam Ravnborg
  2007-03-09 20:16               ` Jan Engelhardt
  0 siblings, 1 reply; 14+ messages in thread
From: Sam Ravnborg @ 2007-03-09 19:00 UTC (permalink / raw)
  To: Jan Engelhardt
  Cc: dmitry.torokhov, Christoph Hellwig, Deepak Saxena, Andrew Morton,
	linux-kernel

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 -m	uname -p
> 	SOL	i86pc		i386
> 	LINUX	i686		athlon
> 
> 
> 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 05:41:20.000000000 +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 05:41:20.000000000 +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 05:41:20.000000000 +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
> ===================================================================
> --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile	2007-03-07 05:41:20.000000000 +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:
> 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 05:41:20.000000000 +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
> Index: linux-2.6.21-rc3/scripts/mod/file2alias.c
> ===================================================================
> --- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c	2007-03-07 05:41:20.000000000 +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
> Index: linux-2.6.21-rc3/scripts/mod/modpost.h
> ===================================================================
> --- linux-2.6.21-rc3.orig/scripts/mod/modpost.h	2007-03-07 05:41:20.000000000 +0100
> +++ linux-2.6.21-rc3/scripts/mod/modpost.h	2007-03-07 23:37:01.315290000 +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
>  {
> Index: linux-2.6.21-rc3/scripts/mod/sumversion.c
> ===================================================================
> --- linux-2.6.21-rc3.orig/scripts/mod/sumversion.c	2007-03-07 05:41:20.000000000 +0100
> +++ linux-2.6.21-rc3/scripts/mod/sumversion.c	2007-03-07 23:43:55.668334000 +0100
> @@ -6,6 +6,7 @@
>  #endif
>  #include <ctype.h>
>  #include <errno.h>
> +#include <limits.h>
>  #include <string.h>
>  #include "modpost.h"
>  
> @@ -417,7 +418,8 @@
>  	*end = '\0';
>  
>  	md4_init(&md);
> -	while ((fname = strsep(&sources, " ")) != NULL) {
> +	for(fname = strtok(sources, " "); fname != NULL;
> +	  fname = strtok(NULL, " ")) {
>  		if (!*fname)
>  			continue;
>  		if (!parse_source_files(fname, &md))
> #<EOF>

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-09 19:00             ` Sam Ravnborg
@ 2007-03-09 20:16               ` Jan Engelhardt
  2007-03-09 22:23                 ` Sam Ravnborg
  0 siblings, 1 reply; 14+ messages in thread
From: Jan Engelhardt @ 2007-03-09 20:16 UTC (permalink / raw)
  To: Sam Ravnborg
  Cc: dmitry.torokhov, Christoph Hellwig, Deepak Saxena, Andrew Morton,
	linux-kernel


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 05:41:20.000000000 +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.c	2007-03-07 05:41:20.000000000 +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 05:41:20.000000000 +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 05:41:20.000000000 +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.

>> 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 05:41:20.000000000 +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.

>> Index: linux-2.6.21-rc3/scripts/mod/file2alias.c
>> ===================================================================
>> --- linux-2.6.21-rc3.orig/scripts/mod/file2alias.c	2007-03-07 05:41:20.000000000 +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.

>> Index: linux-2.6.21-rc3/scripts/mod/modpost.h
>> ===================================================================
>> --- linux-2.6.21-rc3.orig/scripts/mod/modpost.h	2007-03-07 05:41:20.000000000 +0100
>> +++ linux-2.6.21-rc3/scripts/mod/modpost.h	2007-03-07 23:37:01.315290000 +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 >:-(

>> Index: linux-2.6.21-rc3/scripts/mod/sumversion.c
>> ===================================================================
>> --- linux-2.6.21-rc3.orig/scripts/mod/sumversion.c	2007-03-07 05:41:20.000000000 +0100
>> +++ linux-2.6.21-rc3/scripts/mod/sumversion.c	2007-03-07 23:43:55.668334000 +0100
>> @@ -6,6 +6,7 @@
>>  #endif
>>  #include <ctype.h>
>>  #include <errno.h>
>> +#include <limits.h>
>>  #include <string.h>
>>  #include "modpost.h"
>>  

Valid hunk, should be applied (to get the PATH_MAX constant).

>> @@ -417,7 +418,8 @@
>>  	*end = '\0';
>>  
>>  	md4_init(&md);
>> -	while ((fname = strsep(&sources, " ")) != NULL) {
>> +	for(fname = strtok(sources, " "); fname != NULL;
>> +	  fname = strtok(NULL, " ")) {
>>  		if (!*fname)
>>  			continue;
>>  		if (!parse_source_files(fname, &md))
>> #<EOF>

See thread debate as to whether to keep strsep or use
strtok, since strtok is not thread-safe, and strtok_r
is a GNUism unlikely to be available in Solaris either.


Jan
-- 

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-09 20:16               ` Jan Engelhardt
@ 2007-03-09 22:23                 ` Sam Ravnborg
  2007-03-09 22:55                   ` Jan Engelhardt
  2007-03-10  9:37                   ` Christoph Hellwig
  0 siblings, 2 replies; 14+ messages in thread
From: Sam Ravnborg @ 2007-03-09 22:23 UTC (permalink / raw)
  To: Jan Engelhardt, Paulo Marques
  Cc: dmitry.torokhov, Christoph Hellwig, Deepak Saxena, Andrew Morton,
	linux-kernel

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.h	2007-03-07 05:41:20.000000000 +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 05:41:20.000000000 +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 05:41:20.000000000 +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?

> 
> >> Index: linux-2.6.21-rc3/scripts/kconfig/Makefile
> >> ===================================================================
> >> --- linux-2.6.21-rc3.orig/scripts/kconfig/Makefile	2007-03-07 05:41:20.000000000 +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 05:41:20.000000000 +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	2007-03-07 05:41:20.000000000 +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.

> 
> >> Index: linux-2.6.21-rc3/scripts/mod/modpost.h
> >> ===================================================================
> >> --- linux-2.6.21-rc3.orig/scripts/mod/modpost.h	2007-03-07 05:41:20.000000000 +0100
> >> +++ linux-2.6.21-rc3/scripts/mod/modpost.h	2007-03-07 23:37:01.315290000 +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.

> 
> >> Index: linux-2.6.21-rc3/scripts/mod/sumversion.c
> >> ===================================================================
> >> --- linux-2.6.21-rc3.orig/scripts/mod/sumversion.c	2007-03-07 05:41:20.000000000 +0100
> >> +++ linux-2.6.21-rc3/scripts/mod/sumversion.c	2007-03-07 23:43:55.668334000 +0100
> >> @@ -6,6 +6,7 @@
> >>  #endif
> >>  #include <ctype.h>
> >>  #include <errno.h>
> >> +#include <limits.h>
> >>  #include <string.h>
> >>  #include "modpost.h"
> >>  
> 
> Valid hunk, should be applied (to get the PATH_MAX constant).
Yes.

> 
> >> @@ -417,7 +418,8 @@
> >>  	*end = '\0';
> >>  
> >>  	md4_init(&md);
> >> -	while ((fname = strsep(&sources, " ")) != NULL) {
> >> +	for(fname = strtok(sources, " "); fname != NULL;
> >> +	  fname = strtok(NULL, " ")) {
> >>  		if (!*fname)
> >>  			continue;
> >>  		if (!parse_source_files(fname, &md))
> >> #<EOF>
> 
> See thread debate as to whether to keep strsep or use
> strtok, since strtok is not thread-safe, and strtok_r
> is a GNUism unlikely to be available in Solaris either.

For this usage the thread safety does not matter as I see it.
But we can just open code it:

diff --git a/scripts/mod/sumversion.c b/scripts/mod/sumversion.c
index 8a28756..d10a6a3 100644
--- a/scripts/mod/sumversion.c
+++ b/scripts/mod/sumversion.c
@@ -379,7 +379,7 @@ void get_src_version(const char *modname, char sum[], unsigned sumlen)
 	void *file;
 	unsigned long len;
 	struct md4_ctx md;
-	char *sources, *end, *fname;
+	char *sources, *end, *fname, *endname;
 	const char *basename;
 	char filelist[PATH_MAX + 1];
 	char *modverdir = getenv("MODVERDIR");
@@ -417,13 +417,16 @@ void get_src_version(const char *modname, char sum[], unsigned sumlen)
 	*end = '\0';
 
 	md4_init(&md);
-	while ((fname = strsep(&sources, " ")) != NULL) {
-		if (!*fname)
-			continue;
+	fname = endname = sources;
+	while ((endname != end) && *fname) {
+		char *endname = fname;
+		while (*endname && !isspace(*endname))
+			endname++;
+		*endname = '\0';
 		if (!parse_source_files(fname, &md))
 			goto release;
+		fname = ++endname;
 	}
-
 	md4_final_ascii(&md, sum, sumlen);
 release:
 	release_file(file, len);


	Sam

^ permalink raw reply related	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-09 22:23                 ` Sam Ravnborg
@ 2007-03-09 22:55                   ` Jan Engelhardt
  2007-03-10  9:37                   ` Christoph Hellwig
  1 sibling, 0 replies; 14+ messages in thread
From: Jan Engelhardt @ 2007-03-09 22:55 UTC (permalink / raw)
  To: Sam Ravnborg
  Cc: Paulo Marques, dmitry.torokhov, Christoph Hellwig, Deepak Saxena,
	Andrew Morton, linux-kernel


On Mar 9 2007 23:23, Sam Ravnborg wrote:

>Date: Fri, 9 Mar 2007 23:23:32 +0100
>From: Sam Ravnborg <sam@ravnborg.org>
>To: Jan Engelhardt <jengelh@linux01.gwdg.de>,
>    Paulo Marques <pmarques@grupopie.com>
>Cc: dmitry.torokhov@gmail.com, Christoph Hellwig <hch@infradead.org>,
>    Deepak Saxena <dsaxena@plexity.net>,
>    Andrew Morton <akpm@linux-foundation.org>, 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 <jengelh@gmx.de>

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 05:41:20.000000000 +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/Makefile	2007-03-07 05:41:20.000000000 +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.c	2007-03-07 05:41:20.000000000 +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 05:41:20.000000000 +0100
>> >> +++ linux-2.6.21-rc3/scripts/mod/modpost.h	2007-03-07 23:37:01.315290000 +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 BSD too? What about CYGWIN - there have been efforts
before! :p  Seriously, what build platforms do we want to support in
the near future?



Thanks,
Jan
-- 

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH] Fix building kernel under Solaris 11_snv
  2007-03-09 22:23                 ` Sam Ravnborg
  2007-03-09 22:55                   ` Jan Engelhardt
@ 2007-03-10  9:37                   ` Christoph Hellwig
  1 sibling, 0 replies; 14+ messages in thread
From: Christoph Hellwig @ 2007-03-10  9:37 UTC (permalink / raw)
  To: Sam Ravnborg
  Cc: Jan Engelhardt, Paulo Marques, dmitry.torokhov,
	Christoph Hellwig, Deepak Saxena, Andrew Morton, linux-kernel

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 05:41:20.000000000 +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.


^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2007-03-10  9:37 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-03-06 18:09 [PATCH] Fix building kernel under Solaris Deepak Saxena
2007-03-06 18:22 ` Deepak Saxena
2007-03-06 19:23 ` Sam Ravnborg
2007-03-07  9:42 ` Christoph Hellwig
2007-03-07 22:45   ` Jan Engelhardt
2007-03-08  8:35     ` Christoph Hellwig
2007-03-08 20:43       ` Jan Engelhardt
2007-03-08 21:25         ` Sam Ravnborg
2007-03-08 22:01           ` [PATCH] Fix building kernel under Solaris 11_snv Jan Engelhardt
2007-03-09 19:00             ` Sam Ravnborg
2007-03-09 20:16               ` Jan Engelhardt
2007-03-09 22:23                 ` Sam Ravnborg
2007-03-09 22:55                   ` Jan Engelhardt
2007-03-10  9:37                   ` Christoph Hellwig

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.