[OpenAFS] open AFS on Solaris 8 on x86

Fatima Streit fatima@ifh.de
Wed, 12 Sep 2001 13:33:04 +0200 (MET DST)


On Tue, 11 Sep 2001, John Tang Boyland wrote:

> Thanks for the patches. I'm working through them.
> Some, however, do not seem to be appropriate,
> For example, consider the following changes:
>
> --- /openafs-1.0.4a/src/venus/Makefile  Wed Jul 11 10:36:50 2001
> +++ /openafs-1.0.4a/src/venus/Makefile.orig     Mon Apr 23 09:53:14 2001
> @@ -84,7 +84,7 @@
>                 for f in kdump.IP??; \
>                 do ${INSTALL} -s $$f ${DESTDIR}etc/$$f || exit $$? ; \
>                 done ;; \
> -       sun4x_57|sunx86_57 ) \
> +       sun4x_57 ) \
>                 ${INSTALLex} -f kdump.sh.solaris7 $@; \
>                 ${INSTALL} -s -f $? ${DESTDIR}etc/kdump32;; \
>         *linux* ) \

Thanks for your inspection of the patches. For the above one you are in
principle right, but while the script is not required at
all kdump.sh.solaris7 does the proper thing also for intel solaris. I
simply added mechanically sunx86_57 where sun4x_57 ocurred in the scripts
and fixed then, what did not work.

> @@ -101,7 +101,7 @@
>  ${DESTDIR}etc/kdump64: kdump64
>         -set -x; \
>         case ${SYS_NAME} in \
> -       sun4x_57 | sunx86_57|hp_ux11* ) \
> +       sun4x_57 | hp_ux11* ) \
>                 ${INSTALL} -s $? $@ ;;\
>         * ) \
>                 echo skipping kdump64 for ${SYS_NAME} ;; \
>

This explains also the above change, which is really wrong, as there is no
such thing like solaris for intel64. Thanks for pointing this out.

> Finally the param.sunx86_57_usr.h
> file seems strange to me:
>
> ---------------------
> [...]
> #ifndef _PARAM_SUN4C_51_H_
> #define _PARAM_SUN4C_51_H_
>
> [...]
> #define AFS_USR_SUN6_ENV        1
> [...]
> ---------------------
>
That again was taken from the corresponding sparc solaris file for 5.7
While I would have expected there a line #define AFS_USR_SUN7_ENV        1
as well, it was missing in the original file. So I left it out as well.
A grep trough the sources showed, that only AFS_USR_SUN5_ENV seems to be
really used, therefore I guess, it is a flaw in the original file without
any consequences.

Best regards
-----------------------------------------------------------------------------
DESY Zeuthen * Fatima Streit *
Postfach * 15735 Zeuthen  * Germany * EMail fatima@ifh.de *
tel +49 (033762) 77361 * fax +49 (033762) 77216
-----------------------------------------------------------------------------