xfs
[Top] [All Lists]

Re: CML2

To: Keith Owens <kaos@xxxxxxx>
Subject: Re: CML2
From: Austin Gonyou <austin@xxxxxxxxxxxxxxx>
Date: 27 Dec 2001 12:18:46 -0600
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <11061.1009421410@ocs3.intra.ocs.com.au>
References: <11061.1009421410@ocs3.intra.ocs.com.au>
Sender: owner-linux-xfs@xxxxxxxxxxx
BTW, here's the info I got from ESR concerning my dilemma. I should have
a patch up sometime today. 

Austin Gonyou <austin@xxxxxxxxxxxxxxx>:
> Eric,
>   Good show on CML2. I like it very much. There's one thing missing
I've
> found though. That would be an easy way to integrate new code into
CML2.
> At least, it's not easy for me to do.  At any rate, here's my problem,
> and I'd just like to know where, in the READ.ME, cml2-paper, or
> cml2-reference, does it say how to add options to the config? 
> 
> The reason I'm asking, is because I'm trying to use CML2 with the SGI
> XFS kernel tree. After installing CML2, all XFS references in the
> config, menuconfig, etc, are missing. I'm not exactly sure how to add
> them back. The files themselves are still there, but when I add the
> references as I think they should be, fs/rules.cml:
> 
> <snip>
> ...
> # CONFIG_JBD could be its own option (even modular), but until there
are
> # other users than ext3, we will simply make it be the same as
> CONFIG_EXT3_FS
> derive JBD from EXT3_FS
> 
> menu ufs # Linux file system options
>         FS_POSIX_ACL!
>         EXT2_FS! EXT3_FS! {JBD_DEBUG {BUFFER_DEBUG}}
>         QUOTA AUTOFS_FS! AUTOFS4_FS! ROMFS_FS!
>         REISERFS_FS! {REISERFS_CHECK REISERFS_PROC_INFO}
>         PROC_FS {PROC_HARDWARE PROC_DEVICETREE SUN_OPENPROMFS?}
>         DEVFS_FS        {DEVFS_MOUNT DEVFS_DEBUG}
>         DEVPTS_FS
>         ADFS_FS!        {ADFS_FS_RW}
>         JFFS_FS!        {JFFS_FS_VERBOSE% JFFS_PROC_FS}
>         JFFS2_FS!       {JFFS2_FS_DEBUG%}
>         PAGE_BUF! {XFS_FS XFS_RT XFS_QUOTA HAVE_ATTRCTL XFS_DMAPI
> HAVE_XFS_DMAPI}
> 
> menu lfs # Legacy Unix file systems
> ...
> </snip>
> 
> I get the following output:
> 
> <error>
> [austin@UberGeek linux]$ make menuconfig
> rm -f include/asm
> ( cd include ; ln -sf asm-i386 asm)
> python2 -O scripts/cmlcompile.py rules.cml 
> Compiling rules, please wait...\
> "fs/rules.cml", line 28:  unknown symbol word='FS_POSIX_ACL' in
wordlist
> "fs/rules.cml", line 38:  unknown symbol word='PAGE_BUF' in wordlist
> "fs/rules.cml", line 38:  unknown symbol word='XFS_FS' in wordlist
> "fs/rules.cml", line 38:  unknown symbol word='XFS_RT' in wordlist
> "fs/rules.cml", line 38:  unknown symbol word='XFS_QUOTA' in wordlist
> "fs/rules.cml", line 38:  unknown symbol word='HAVE_ATTRCTL' in
wordlist
> "fs/rules.cml", line 38:  unknown symbol word='XFS_DMAPI' in wordlist
> "fs/rules.cml", line 38:  unknown symbol word='HAVE_XFS_DMAPI' in
> wordlist
> cmlcompile: JFFS2_FS_DEBUG is not of numeric type but has numeric
> constant default
> cmlcompile: rulebase write suppressed due to errors.
> make: *** [rules.out] Error 1
> </error>
> 
> Where do I define the symbols that is being complained about here?
I'll
> continue to dig, but after looking most of today, I got a little
> frustrated and thought I'd ask you right out. Sorry for the
> inconvenience, but thanks if you could point me in the right
direction. 

OK, what you need to do is add some declarations to the symbols.cml
file.
One for each new symbol, specifying prompt text for it.

There is a temporary issue with which file(s) should be patched.
Presently the actual symbols.cml file in the kernel tree is generated
at CML2 installation time from a symbols.cml template and the
Configure.help file, both supplied by the CML2 distribution.  After
CML2 is merged into the 2.5 tree this will no longer be the case; you'll
patch the kernel-tree symbols.cml file directly.  

Right now, you have a choice of patching the either the distributed CML2
files 
or the kernel-tree files after CML2 installation.
-- 
                <a href="http://www.tuxedo.org/~esr/";>Eric S.
Raymond</a>

"Are we to understand," asked the judge, "that you hold your own
interests
above the interests of the public?"

"I hold that such a question can never arise except in a society of
cannibals."
        -- Ayn Rand

On Wed, 2001-12-26 at 20:50, Keith Owens wrote:
> On 26 Dec 2001 13:51:20 -0600, 
> Austin Gonyou <austin@xxxxxxxxxxxxxxx> wrote:
> >Is there a patch for integrating CML2 make system with the XFS tree?
> Any
> >pointers on this would be great.
> 
> Are you sure that you mean CML2?  CML2 just defines the config symbols,
> their help text and generates the .config file.  There is only two
> config.in files affected, one for KDB and one for XFS.  Simply convert
> the KDB/XFS changes below to CML2, the help text is already in
> Configure.help and will be converted by the install-cml2 script.
> 
> If you really mean the new kernel build system (kbuild 2.5), that works
> with both CML1 and CML2.  XFS is already kbuild 2.5 ready, see
> http://marc.theaimsgroup.com/?l=linux-xfs&m=100231202715631&w=2
> 
> diff -urNp 2.4.16/arch/i386/config.in 2.4.16-xfs/arch/i386/config.in
> --- 2.4.16/arch/i386/config.in  Fri Nov 23 17:35:37 2001
> +++ 2.4.16-xfs/arch/i386/config.in      Tue Dec  4 17:29:36 2001
> @@ -407,6 +407,16 @@ if [ "$CONFIG_DEBUG_KERNEL" != "n" ]; th
>     bool '  Magic SysRq key' CONFIG_MAGIC_SYSRQ
>     bool '  Spinlock debugging' CONFIG_DEBUG_SPINLOCK
>     bool '  Verbose BUG() reporting (adds 70K)' CONFIG_DEBUG_BUGVERBOSE
> +   bool '  Built-in Kernel Debugger support' CONFIG_KDB
> +   dep_tristate '    KDB modules' CONFIG_KDB_MODULES $CONFIG_KDB
> +   if [ "$CONFIG_KDB" = "y" ]; then
> +      bool '    KDB off by default' CONFIG_KDB_OFF
> +      comment '  Load all symbols for debugging is required for KDB'
> +      define_bool CONFIG_KALLSYMS y
> +   else
> +      bool '  Load all symbols for debugging' CONFIG_KALLSYMS
> +   fi
> +   bool '  Compile the kernel with frame pointers' CONFIG_FRAME_POINTER
>  fi
>  
>  endmenu
> 
> diff -urNp 2.4.16/fs/Config.in 2.4.16-xfs/fs/Config.in
> --- 2.4.16/fs/Config.in Fri Nov 23 17:36:10 2001
> +++ 2.4.16-xfs/fs/Config.in     Tue Dec  4 17:29:36 2001
> @@ -5,6 +5,7 @@ mainmenu_option next_comment
>  comment 'File systems'
>  
>  bool 'Quota support' CONFIG_QUOTA
> +bool 'POSIX Access Control List support' CONFIG_FS_POSIX_ACL
>  tristate 'Kernel automounter support' CONFIG_AUTOFS_FS
>  tristate 'Kernel automounter version 4 support (also supports v3)'
> CONFIG_AUTOFS4_FS
>  
> @@ -85,6 +86,20 @@ dep_mbool '  UDF write support (DANGEROU
>  
>  tristate 'UFS file system support (read only)' CONFIG_UFS_FS
>  dep_mbool '  UFS file system write support (DANGEROUS)'
> CONFIG_UFS_FS_WRITE $CONFIG_UFS_FS $CONFIG_EXPERIMENTAL
> +
> +tristate 'Page Buffer support' CONFIG_PAGE_BUF
> +
> +if [ "$CONFIG_PAGE_BUF" = "n" ]; then
> +   comment '  Page Buffer support needed for XFS filesystem'
> +else
> +   dep_tristate 'SGI XFS filesystem support' CONFIG_XFS_FS
> $CONFIG_PAGE_BUF
> +   dep_mbool    '  Enable XFS Realtime support' CONFIG_XFS_RT
> $CONFIG_XFS_FS
> +   dep_mbool    '  Enable XFS Quota' CONFIG_XFS_QUOTA $CONFIG_XFS_FS
> +   if [ "$CONFIG_XFS_FS" != "n" ]; then
> +      define_bool CONFIG_HAVE_ATTRCTL y
> +      dep_mbool '  Enable XFS DMAPI' CONFIG_XFS_DMAPI $CONFIG_XFS_FS
> +   fi
> +fi
>  
>  if [ "$CONFIG_NET" = "y" ]; then
>  
> 
-- 
Austin Gonyou
Systems Architect, CCNA 
Coremetrics, Inc.
Phone: 512-698-7250
email: austin@xxxxxxxxxxxxxxx
 
"Have regard for your name, since it will remain for you longer than a
great store of gold."
Ecclesiastes, Aprocrypha

Attachment: signature.asc
Description: This is a digitally signed message part

<Prev in Thread] Current Thread [Next in Thread>