[Linux-ia64] make xconfig

[Linux-ia64] make xconfig

[Jason Baron]

[-- Attachment #1: Type: text/plain, Size: 998 bytes --]

If I enable xconfig, I get these errors from the 
linux-2.4.19-ia64-020821.diff:

ERROR - Attempting to write value for unconfigured
variable (CONFIG_BLK_DEV_LOOP).
ERROR - Attempting to write value for unconfigured
variable (CONFIG_BLK_DEV_NBD).
ERROR - Attempting to write value for unconfigured
variable (CONFIG_BLK_DEV_RAM).

The consequence is that these parameters do not get
written to the appropriate files.  Also, there are two
top level "Block devices" menues.

This issue has been discussed a bit and I agree that
the right solution is probably to fix the way that
xconfig works.  Short of this, we could choose not to
source drivers/block/Config.in, and instead include
the drivers/block/Config.in parameters in
arch/ia64/config.in, as in the attached patch.  This
bloats the file a bit, but solves the issue in an arch
independent way.

-Jason Baron

__________________________________________________
Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes
http://finance.yahoo.com

[-- Attachment #2: config.in.patch --]
[-- Type: application/octet-stream, Size: 3385 bytes --]

--- arch/ia64/config.in.bak	Wed Aug 28 10:25:07 2002
+++ arch/ia64/config.in	Wed Aug 28 10:20:29 2002
@@ -136,7 +136,64 @@
 if [ "$CONFIG_IA64_HP_SIM" = "n" ]; then
   source drivers/mtd/Config.in
   source drivers/pnp/Config.in
-  source drivers/block/Config.in
+fi
+
+#
+# Block device driver configuration
+#
+mainmenu_option next_comment
+comment 'Block devices'
+
+if [ "$CONFIG_IA64_HP_SIM" = "n" ]; then
+  tristate 'Normal PC floppy disk support' CONFIG_BLK_DEV_FD
+  if [ "$CONFIG_AMIGA" = "y" ]; then
+    tristate 'Amiga floppy support' CONFIG_AMIGA_FLOPPY
+  fi
+  if [ "$CONFIG_ATARI" = "y" ]; then
+     tristate 'Atari floppy support' CONFIG_ATARI_FLOPPY
+  fi
+  if [ "$CONFIG_MAC" = "y" ]; then
+     dep_bool 'Macintosh IIfx/Quadra 900/Quadra 950 floppy support (EXPERIMENTAL)' CONFIG_BLK_DEV_SWIM_IOP $CONFIG_EXPERIMENTAL
+  fi
+  if [ "$CONFIG_MCA" = "y" ]; then
+     tristate 'PS/2 ESDI hard disk support' CONFIG_BLK_DEV_PS2
+  fi
+  if [ "$CONFIG_ZORRO" = "y" ]; then
+     tristate 'Amiga Zorro II ramdisk support' CONFIG_AMIGA_Z2RAM
+  fi
+  if [ "$CONFIG_ATARI" = "y" ]; then
+     tristate 'Atari ACSI support' CONFIG_ATARI_ACSI
+     if [ "$CONFIG_ATARI_ACSI" != "n" ]; then
+        comment 'Some devices (e.g. CD jukebox) support multiple LUNs'
+        bool '  Probe all LUNs on each ACSI device' CONFIG_ACSI_MULTI_LUN
+        tristate '  Atari SLM laser printer support' CONFIG_ATARI_SLM
+     fi
+  fi
+  dep_tristate 'XT hard disk support' CONFIG_BLK_DEV_XD $CONFIG_ISA
+  dep_tristate 'Parallel port IDE device support' CONFIG_PARIDE $CONFIG_PARPORT
+  if [ "$CONFIG_PARIDE" = "y" -o "$CONFIG_PARIDE" = "m" ]; then
+     source drivers/block/paride/Config.in
+  fi
+  dep_tristate 'Compaq SMART2 support' CONFIG_BLK_CPQ_DA $CONFIG_PCI
+  dep_tristate 'Compaq Smart Array 5xxx support' CONFIG_BLK_CPQ_CISS_DA $CONFIG_PCI 
+  dep_mbool '       SCSI tape drive support for Smart Array 5xxx' CONFIG_CISS_SCSI_TAPE $CONFIG_BLK_CPQ_CISS_DA $CONFIG_SCSI
+  dep_tristate 'Mylex DAC960/DAC1100 PCI RAID Controller support' CONFIG_BLK_DEV_DAC960 $CONFIG_PCI
+  dep_tristate 'Micro Memory MM5415 Battery Backed RAM support' CONFIG_BLK_DEV_UMEM $CONFIG_PCI $CONFIG_EXPERIMENTAL
+fi # !HP_SIM
+
+  tristate 'Loopback device support' CONFIG_BLK_DEV_LOOP
+  dep_tristate 'Network block device support' CONFIG_BLK_DEV_NBD $CONFIG_NET
+
+  tristate 'RAM disk support' CONFIG_BLK_DEV_RAM
+  if [ "$CONFIG_BLK_DEV_RAM" = "y" -o "$CONFIG_BLK_DEV_RAM" = "m" ]; then
+     int '  Default RAM disk size' CONFIG_BLK_DEV_RAM_SIZE 4096
+  fi
+  dep_bool '  Initial RAM disk (initrd) support' CONFIG_BLK_DEV_INITRD $CONFIG_BLK_DEV_RAM
+
+endmenu
+
+
+if [ "$CONFIG_IA64_HP_SIM" = "n" ]; then
   source drivers/ieee1394/Config.in
   source drivers/message/i2o/Config.in
   source drivers/md/Config.in
@@ -154,18 +211,7 @@
     define_bool CONFIG_BLK_DEV_HD n
   fi
   endmenu
-else # HP_SIM
-  mainmenu_option next_comment
-  comment 'Block devices'
-  tristate 'Loopback device support' CONFIG_BLK_DEV_LOOP
-  dep_tristate 'Network block device support' CONFIG_BLK_DEV_NBD $CONFIG_NET
-
-  tristate 'RAM disk support' CONFIG_BLK_DEV_RAM
-  if [ "$CONFIG_BLK_DEV_RAM" = "y" -o "$CONFIG_BLK_DEV_RAM" = "m" ]; then
-    int '  Default RAM disk size' CONFIG_BLK_DEV_RAM_SIZE 4096
-  fi
-  endmenu
-fi # HP_SIM
+fi # !HP_SIM
 
 mainmenu_option next_comment
 comment 'SCSI support'

Re: [Linux-ia64] make xconfig

[Bjorn Helgaas]

> This issue has been discussed a bit and I agree that
> the right solution is probably to fix the way that
> xconfig works.  Short of this, we could choose not to
> source drivers/block/Config.in, and instead include
> the drivers/block/Config.in parameters in
> arch/ia64/config.in, as in the attached patch.  This
> bloats the file a bit, but solves the issue in an arch
> independent way.

Duplicating the contents of drivers/block/Config.in doesn't
feel like an improvement.  Doesn't it just add more opportunity
for me to make merge mistakes?  I don't need any more of
those :-)

Bjorn

Re: [Linux-ia64] make xconfig

[Jason Baron]

--- Bjorn Helgaas <bjorn_helgaas@hp.com> wrote:
> > This issue has been discussed a bit and I agree
> that
> > the right solution is probably to fix the way that
> > xconfig works.  Short of this, we could choose not
> to
> > source drivers/block/Config.in, and instead
> include
> > the drivers/block/Config.in parameters in
> > arch/ia64/config.in, as in the attached patch. 
> This
> > bloats the file a bit, but solves the issue in an
> arch
> > independent way.
> 
> Duplicating the contents of drivers/block/Config.in
> doesn't
> feel like an improvement.  

They are already duplicated.

-Jason



__________________________________________________
Do You Yahoo!?
Yahoo! Finance - Get real-time stock quotes
http://finance.yahoo.com

Re: [Linux-ia64] make xconfig

[Keith Owens]

On Wed, 28 Aug 2002 09:02:16 -0700 (PDT), 
Jason Baron <almostivan@yahoo.com> wrote:
>If I enable xconfig, I get these errors from the 
>linux-2.4.19-ia64-020821.diff:
>
>ERROR - Attempting to write value for unconfigured
>variable (CONFIG_BLK_DEV_LOOP).
>ERROR - Attempting to write value for unconfigured
>variable (CONFIG_BLK_DEV_NBD).
>ERROR - Attempting to write value for unconfigured
>variable (CONFIG_BLK_DEV_RAM).
>
>The consequence is that these parameters do not get
>written to the appropriate files.  Also, there are two
>top level "Block devices" menues.
>
>This issue has been discussed a bit and I agree that
>the right solution is probably to fix the way that
>xconfig works.

It is an undocumented restriction of xconfig that an option should only
be set in one menu.  xconfig can detect sets of options within a menu,
it sometimes detects sets of options across menus but do not rely on
that.  Nobody who has looked at the xconfig code is willing to remove
this restriction (they usually turn white and and run away).

Unfortunately some architectures break this rule and break xconfig.

>Short of this, we could choose not to
>source drivers/block/Config.in, and instead include
>the drivers/block/Config.in parameters in
>arch/ia64/config.in, as in the attached patch.  This
>bloats the file a bit, but solves the issue in an arch
>independent way.

Given the current menuconfig and xconfig code, the only workable fix is
to follow the config rules.  That means any variable should only be set
in one menu.  Do not duplicate variables in arch/ia64/config.in.
Instead add conditions around the variables in drivers/block/Config.in
and other Config.in files.

I sent DM a patch to do just this, but he rejected it.  Instead xconfig
was disabled.  Disabling xconfig is not the solution, multiple
definitions of a variable in multiple menus will silently break
menuconfig as well (think out of order menu selection).