Re: Bug#282763: Installation report amd64 (gcc-3.4)

2004-11-24 Thread Jim Lieb 
On Wednesday 24 November 2004 05:21, Joey Hess wrote:
> Harald Dunkel wrote:
> > After manually partition the disk there was a popup saying
> >
> >  Partition disks:
> >  No EFI partition was found
> >
> > Ignored. The rest of the installation went fine.
>
> I thought EFI was only an ia64 thing. Strange. Could you send a tarball
> of /var/log/debian-installer/ from the installed system so I can try to
> see why it was doing EFI stuff?
I set the arch to be ia64 for partman-efi where this lives.  I noticed that
there were hppa and other recipes in the /var/lib tree in the installer.
These should not even be present although it seemed other logic kept
them from being used.  If I understand this stuff correctly, this udeb
shouldn't even be built for non-ia64.
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#281569: de4x5 doesn't work on ia64 system

2004-11-19 Thread Jim Lieb 
As I remember, the de4x5 is an older chipset.  The 21142 should use the
tulip driver.  The only boards I ever saw that actually used this chip
were EISA boards for the DEC PCs.

On Tuesday 16 November 2004 10:30, Alex Williamson wrote:
> Package: discover1-data
> Version: 1.2004.11.09
> Severity: Normal
>
> Bug #267302 switched the default for "DECchip 21142/43" from tulip to
> de4x5.  Unfortunately, this breaks my system which only works with the
> tulip.  Device in question:
>
> # lspci -vvv -s 00:01.0
> 0001:00:01.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21142/43 (rev 41) Subsystem: Hewlett-Packard Company: Unknown device 104f
> Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
> Stepping- SERR- FastB2B- Status: Cap- 66MHz- UDF- FastB2B+ ParErr-
> DEVSEL=medium >TAbort- SERR-  min, 1ns max), Cache Line Size: 0x20 (128 bytes) Interrupt: pin A
> routed to IRQ 49
> Region 0: I/O ports at 0d00 [size=128]
> Region 1: Memory at 0f0100041000 (32-bit, non-prefetchable)
> [size=1K] Expansion ROM at c000 [disabled] [size=256K]
>
> # xxd /sys/devices/pci0001:00/0001:00:01.0/config
> 000: 1110 1900 0700 8002 4100 0002 2050   A... P..
> 010: 010d  0010 04c0      
> 020:       3c10 4f10  <.O.
> 030:  00c0     cf01 1428  ...(
> 040: 00bc         
> 050:          
> 060:          
> 070:          
> 080:          
> 090:          
> 0a0:          
> 0b0:          
> 0c0:          
> 0d0:          
> 0e0:          
> 0f0:          
>
> This occurs on an ia64 HP Superdome system.  I also see that the
> standard 4-port tulip card offering from HP has the same device/vendor:
>
> :41:04.0 Ethernet controller: Digital Equipment Corporation DECchip
> 21142/43 (rev 41) Subsystem: Hewlett-Packard Company: Unknown device 125a
> Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+
> Stepping- SERR+ FastB2B- Status: Cap- 66MHz- UDF- FastB2B+ ParErr-
> DEVSEL=medium >TAbort- SERR-  min, 1ns max), Cache Line Size: 0x20 (128 bytes) Interrupt: pin A
> routed to IRQ 58
> Region 0: I/O ports at 3180 [size=128]
> Region 1: Memory at 98103000 (32-bit, non-prefetchable)
> [size=1K] Expansion ROM at 980c [disabled] [size=256K]
>
> The de4x5 module loads, then promptly oopes when the installer tries to
> ifconfig it.  The tulip driver works just fine.  Given the previous bug
> #267302 there's obviously some disagreement on which module to use.  I
> know that nobody else (RH/SuSE) is using the de4x5 module on ia64.  Is
> there some way we can key on architecture?  Can we add Subsystem
> Device/Vendor ID checking?  Thanks,
>
>   Alex

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: adding an "outdated" warning to the installation manual for some arches

2004-10-07 Thread Jim Lieb 
On Wednesday 06 October 2004 16:43, Joey Hess wrote:
> I guess the installation manual is still not fact checked or up to date
>
>
> What other architectures besides i386 and powerpc are updated well
> enough to avoid such a warning?
ia-64
-- 
*****
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268554: weirdness trying to use existing efi partitions on ia64

2004-10-01 Thread Jim Lieb 
Closes: #268554

Found problem as reported.  Committed fix with revision r22446.

Jim

On Thursday 30 September 2004 21:54, Joey Hess wrote:
> Jim Lieb wrote:
> > Close: 268554
> >
> > This bug has been fixed by a combination of various partman core changes
> > and the latest patch to partman-efi.  I could not reproduce this symptom
> > in the current version.  If no one else can reproduce it, please close
> > this bug.
>
> I can still reproduce this:
>
> 1. Install debian using automatic partitioning.
> 2. Reinstall on top of that, manual partitioning, tell it to reformat
>the root partition. Since the EFI partition exists already, leave it
>as is
> 3. partman-efi complains that there is no EFI partition.

-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268554: weirdness trying to use existing efi partitions on ia64

2004-10-01 Thread Jim Lieb 
On Thursday 30 September 2004 21:54, Joey Hess wrote:
> Jim Lieb wrote:
> > Close: 268554
> >
> > This bug has been fixed by a combination of various partman core changes
> > and the latest patch to partman-efi.  I could not reproduce this symptom
> > in the current version.  If no one else can reproduce it, please close
> > this bug.
>
> I can still reproduce this:
>
> 1. Install debian using automatic partitioning.
> 2. Reinstall on top of that, manual partitioning, tell it to reformat
>the root partition. Since the EFI partition exists already, leave it
>as is
> 3. partman-efi complains that there is no EFI partition.
I just looked at the code and I'll get a patch commited today.
As a matter of process, do you want me to just add my comment
to the top of the current changelog or do you want me to start
a new entry?  I notice you have combined multiple comments
in the past.

Jim
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#264947: Acknowledgement (Busybox chroot crash/errors on Sarge 08/06 for ia64.)

2004-09-30 Thread Jim Lieb 
This problem no longer occurs because the missing kernel module
problem in the kernel package has been fixed which eliminates
the elilo error on installs.  Therefore, this is no longer a show stopper.

There seems to be a minor issue remaining on 2.6.8 kernels.
Although the elilo-installer runs correctly, the console displays
unaligned access fixup messages with this kernel.  These do not
show up with 2.4.27.

This bug can be either closed or reduced in priority since the 
error message is only a warning, not an error.

JIm
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#270746: the display of the efi boot partition is uninformative

2004-09-30 Thread Jim Lieb 
Package: partman-efi
Close: 270746

This bug is fixed with the same patch as fixes #268555.
Please close.

Jim
On Wednesday 08 September 2004 18:21, Joey Hess wrote:
> Package: partman-efi
> Severity: normal
> Tags: d-i
>
> This is how partman displays my partition table after automatic
> partitioning:
>
>   x >  #1 100.0 MB B F   a 
> x x >  #2  34.8 GB   F ext3/ a 
> x x >  #3   1.5 GB   F swapswap  a 
> x
>
> The efi boot partition does not have a type listed, which makes it not very
> clear what this partition is for.
>
> -- System Information:
> Debian Release: 3.1
>   APT prefers unstable
>   APT policy: (500, 'unstable')
> Architecture: i386 (i686)
> Kernel: Linux 2.4.27
> Locale: LANG=en_US, LC_CTYPE=en_US

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268554: weirdness trying to use existing efi partitions on ia64

2004-09-30 Thread Jim Lieb 
On Monday 30 August 2004 12:23, Jim Lieb wrote:
> > Package: partman-efi
> > Severity: normal
> > Tags: d-i

Close: 268554

This bug has been fixed by a combination of various partman core changes
and the latest patch to partman-efi.  I could not reproduce this symptom
in the current version.  If no one else can reproduce it, please close this
bug.

Jim
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-30 Thread Jim Lieb 
On Wednesday 29 September 2004 19:39, Joey Hess wrote:
> All the efi labels are now reasonable, although I do notice that in the
> list of partition types, it has "Use the partition as a EFI boot
> partition". It would be more consistent with the rest of the list to
> just have "EFI boot partition", I think. Since that string is the
> description of another template (partman-efi/text/efi) we could even
> make this change despite being in the string freeze.

I have no preferences.  I copied it from one of the other
templates and just did an s/foo/efi/g.  If it needs to change
to be consistant with other menus or fit usability, do it.
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-29 Thread Jim Lieb 
On Wednesday 29 September 2004 16:10, Joey Hess wrote:
> Jim Lieb wrote:
> > I just did an svn --show-updates -verbose status from
> > packages/partman/partman-efi in my working copy and got the following:
> >
> > * an "svn: subversion/libsvn_wc/status.c:899: tweak_statushash:Assertion
> >   repos_text_status == svn_wc_status_added failed" error.  This was
> >   at the end of scanning the update.d  directory.  The command
> >   aborted at that point so I can't see if valid_filesystems dir is still
> >   in the repository. (It shouldn't).
>
> It was there but with both files empty since I'd applied the patch but
> not noticed it tried to delete them. Now deleted.
>
> FWIW, your svn error seems to be a problem with your working copy, and
> not the repo.
>
> > * The file efi_visuals is empty or weird or ??? as it shows '?' for
> >revision and author.
>
> efi_visuals is present, I added it when applying the patch. I notice
> that its _numbers does not list it, that's probably why it's not
> working.
The relevant portion of the patch if it got lost in the shuffle:

Index: update.d/_numbers
===
--- update.d/_numbers (revision 21963)
+++ update.d/_numbers (working copy)
@@ -1 +1,2 @@
-60 efi_sync_flag
+21 efi_sync_flag
+60 efi_visuals

This got missed on the submit?  This would break it.
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-29 Thread Jim Lieb 
On Wednesday 29 September 2004 12:55, Joey Hess wrote:
> Jim Lieb wrote:
> > > I noticed that with this patch, there is still no type given for efi
> > > partitions in the main partman menu. With the last version of the
> > > patch, a type was given, though the menu item was doubled.
> >
> > I'm not sure what you mean.
>
> When partman displays the efi partition in its partition table, I see
> something like this:
>
>   x >  #1 100.0 MB B F   a 
> x x >  #2  34.8 GB   F ext3/ a 
> x x >  #3   1.5 GB   F swapswap  a 
> x
>
> Note that there's still no type given for the EFI partition. The type is
> however given everywhere else that it was not before.
I just did an svn --show-updates -verbose status from
packages/partman/partman-efi in my working copy and got the following:

* an "svn: subversion/libsvn_wc/status.c:899: tweak_statushash:Assertion
  repos_text_status == svn_wc_status_added failed" error.  This was
  at the end of scanning the update.d  directory.  The command
  aborted at that point so I can't see if valid_filesystems dir is still
  in the repository. (It shouldn't).

* The file efi_visuals is empty or weird or ??? as it shows '?' for
   revision and author.

This file fixes the problem you mention above.  You  should see
'EFIboot' following the 'F' and nothing in the mount point column
for these displays.

As for what is going on in svn, I don't like the looks of it.  If
the file efi_visuals is empty or missing, you will see the error
you report.

This appears to be a commit of the most recent patch since
the report does show debian/rules as being modified but what
is going on does not look good to me. 
  
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-29 Thread Jim Lieb 
On Tuesday 28 September 2004 23:21, Joey Hess wrote:
> Jim Lieb wrote:
> > As promised.  Here is the revised patch.  This has been tested
> > and I had my glasses on at the time so I could see that the 'efi'
> > was indeed gone.  As with the previous patch this has changes
> > to the templates but since they are deletions, they require no
> > translator work.
>
> I noticed that with this patch, there is still no type given for efi
> partitions in the main partman menu. With the last version of the patch,
> a type was given, though the menu item was doubled.

I'm not sure what you mean.  There were two menu entries in
the active_partition menu.  One was with (and generated by) the
code that was based on filesystem type.  That's where the 'efi'
still sneaked in with my earlier patch (right past my bifocals...).
The other is in the methods section along with swap and
lvm.  I removed the the valid_filesystems line from debian/rules,
leaving the 'method' entry so efi would be treated like swap.  The other
bug I fixed was the missing 'type' in the various partition state
displays.  That is fixed by the addition of the script to set
visual_filesystem in update.d/efi_visuals.  This means that any
partition that is formatted fat16 and bootable is displayed/handled
as an efi boot partition candidate.

Is there something else that I missed?  This is a funny duck in
that it has a vfs usable filesystem on it (fat16 'type') but the
kernel doesn't need/want it for mounting (although RH ES
does mount on /boot/efi).  I intended it to be handled like swap
(no mount point etc) for that reason.  The filesystem 'type' is
handled under the covers with parted with the type being set
to fat16 *and* the boot flag being set.

Jim
-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-28 Thread Jim Lieb 
As promised.  Here is the revised patch.  This has been tested
and I had my glasses on at the time so I could see that the 'efi'
was indeed gone.  As with the previous patch this has changes
to the templates but since they are deletions, they require no
translator work.

Jim
On Monday 27 September 2004 18:02, Jim Lieb wrote:
> On Monday 27 September 2004 11:55, Jim Lieb wrote:
> > On Monday 27 September 2004 09:54, Anton Zinoviev wrote:
> > > On Mon, Sep 27, 2004 at 07:15:42AM -0700, Jim Lieb wrote:
> > > > Since there is no template item, it doesn't process that line or
> > > > display it
> > >
> > > If partman doesn't process that line then this is a bug and I will
> > > have to fix it.  Have you tested that partman doesn't display this
> > > line?
> >
> > yes I have.  I noted a comment in changelog that indicated how to
> > handle  this template not being there, which is to cleanly skip it.
> >
> > > Anton Zinoviev
>
> An Oops on my part.  I tested again and saw the efi.  I must have
> just overlooked it.  Please drop this patch even though it does fix
> some other display issues.  I'll issue a new one shortly.
>
> Sorry,
>
> Jim
>
> > --
> > *
> > Jim Lieb   Wild Open Source Inc.
> > [EMAIL PROTECTED]  Cell: 831.295.9317
> > Office: 831.421.0883        Fax:  831.421.0885
>
> --
> *
> Jim Lieb   Wild Open Source Inc.
> [EMAIL PROTECTED]  Cell: 831.295.9317
> Office: 831.421.0883Fax:  831.421.0885

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: choose_method/efi/do_option
===
--- choose_method/efi/do_option	(revision 21963)
+++ choose_method/efi/do_option	(working copy)
@@ -6,7 +6,6 @@
 [ -d $dev/$id ] || mkdir $dev/$id
 
 echo efi >$dev/$id/method
-echo efi >$dev/$id/filesystem
 >$dev/$id/format
 >$dev/$id/bootable
 >$dev/$id/efi_bootable
@@ -14,5 +13,8 @@
 if [ -f $dev/$id/use_filesystem ]; then
 rm $dev/$id/use_filesystem
 fi
+if [ -f $dev/$id/filesystem ]; then
+rm $dev/$id/filesystem
+fi
 
 exit 0
Index: debian/partman-efi.templates
===
--- debian/partman-efi.templates	(revision 21963)
+++ debian/partman-efi.templates	(working copy)
@@ -25,9 +25,9 @@
 # Up to 10 character positions
 _Description: EFIboot
 
-Template: partman/filesystem_long/efi
-Type: text
-_Description: EFI boot partition (FAT16)
+#Template: partman/filesystem_long/efi
+#Type: text
+#_Description: EFI boot partition (FAT16)
 
 Template: partman/filesystem_short/efi
 Type: text
Index: debian/changelog
===
--- debian/changelog	(revision 21963)
+++ debian/changelog	(working copy)
@@ -1,3 +1,12 @@
+partman-efi (4) UNRELEASED; urgency=low
+
+  * Jim Lieb
+- Get file system type displayed properly
+- remove extra menu label
+- Closes: #268555
+
+ -- Jim Lieb <[EMAIL PROTECTED]>  Fri,  24 Sep 2004 12:00:00 -0700
+
 partman-efi (3) UNRELEASED; urgency=low
 
   * Joey Hess
Index: debian/rules
===
--- debian/rules	(revision 21963)
+++ debian/rules	(working copy)
@@ -25,7 +25,6 @@
 	debian/install-rc commit.d
 	debian/install-rc init.d
 	debian/install-rc update.d
-	debian/install-rc valid_filesystems
 	dh_install parted_names lib/partman
 	rm -rf `find debian/$(PACKAGE) -name CVS`
 	rm -rf `find debian/$(PACKAGE) -name .svn`
Index: update.d/efi_sync_flag
===
--- update.d/efi_sync_flag	(revision 21963)
+++ update.d/efi_sync_flag	(working copy)
@@ -23,22 +23,7 @@
 if [ -f $id/method ]; then
 method=$(cat $id/method)
 fi
-filesys=
-if [ -f $id/filesystem ]; then
-filesys=$(cat $id/filesystem)
-fi
 
-if [ "$method" = keep -a "$filesys" = efi -a "$filesys" != "$fs" ]; then
-method=format
-fi
-
-if [ "$method" = format -a "$filesys" = efi ]; then
-echo efi >$id/method
-method=efi
-rm $id/mountpoint
-rm $id/visual_mountpoint
-fi
-
 flags=''
 open_dialog GET_FLAGS $id
 while { read_line flag; [ "$flag" ]; }; do
Index: update.d/efi_visuals
===
--- update.d/efi_visuals	(revision 0)
+++ update.d/efi_visuals	(revision 0)
@@ -0,0 +1,25 @@
+#!/bin/sh
+
+. /usr/share/debconf/confmodule
+
+dev=$1
+num=$2
+id=$3
+size=$4
+type=$5
+fs=$6
+path=$7
+name=$8
+
+cd $dev
+
+[ -f $id/method ] || exit 0
+method=$(cat $id/method)
+
+case "$me

Re: Bug#268555: two partition types for efi on ia64

2004-09-27 Thread Jim Lieb 
On Monday 27 September 2004 11:55, Jim Lieb wrote:
> On Monday 27 September 2004 09:54, Anton Zinoviev wrote:
> > On Mon, Sep 27, 2004 at 07:15:42AM -0700, Jim Lieb wrote:
> > > Since there is no template item, it doesn't process that line or
> > > display it
> >
> > If partman doesn't process that line then this is a bug and I will
> > have to fix it.  Have you tested that partman doesn't display this
> > line?
>
> yes I have.  I noted a comment in changelog that indicated how to
> handle  this template not being there, which is to cleanly skip it.
>
> > Anton Zinoviev
>
An Oops on my part.  I tested again and saw the efi.  I must have
just overlooked it.  Please drop this patch even though it does fix
some other display issues.  I'll issue a new one shortly.

Sorry,

Jim
> --
> *
> Jim Lieb   Wild Open Source Inc.
> [EMAIL PROTECTED]  Cell: 831.295.9317
> Office: 831.421.0883Fax:  831.421.0885

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#268555: two partition types for efi on ia64

2004-09-27 Thread Jim Lieb 
On Monday 27 September 2004 09:54, Anton Zinoviev wrote:
> On Mon, Sep 27, 2004 at 07:15:42AM -0700, Jim Lieb wrote:
> > Since there is no template item, it doesn't process that line or display
> > it
>
> If partman doesn't process that line then this is a bug and I will
> have to fix it.  Have you tested that partman doesn't display this
> line?
yes I have.  I noted a comment in changelog that indicated how to
handle  this template not being there, which is to cleanly skip it.
>
> Anton Zinoviev

-- 
*
Jim Lieb   Wild Open Source Inc.
[EMAIL PROTECTED]  Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#268555: two partition types for efi on ia64

2004-09-25 Thread Jim Lieb 
Package: partman-efi
Closes: 268555

The following patch closes this bug.  It also cleans up the missiing
display of file system type in the various menus.

Note:
  This does change the template file but it deletes a string does not
add or change any text which will not require any translation when it
will trigger the same for the various .po files.

On Friday 27 August 2004 19:35, Joey Hess wrote:
> Package: partman-efi
> Severity: normal
> Tags: d-i
>
> I am installing from sid, with the sid udebs, and in partman, the
> list of possible partition types includes two for efi:
>
> EFI boot partition (FAT16)
> Use the partition as an EFI boot partition
>
> The second of these should be removed. Also, the first on comes
> before ext3 in the menu for some reason, which is fine if there is no
> existing efi partition, but sorta annyoing once the efi partition is
> set up and many other partitions are being made.

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: debian/partman-efi.templates
===
--- debian/partman-efi.templates	(revision 21963)
+++ debian/partman-efi.templates	(working copy)
@@ -25,9 +25,9 @@
 # Up to 10 character positions
 _Description: EFIboot
 
-Template: partman/filesystem_long/efi
-Type: text
-_Description: EFI boot partition (FAT16)
+#Template: partman/filesystem_long/efi
+#Type: text
+#_Description: EFI boot partition (FAT16)
 
 Template: partman/filesystem_short/efi
 Type: text
Index: debian/changelog
===
--- debian/changelog	(revision 21963)
+++ debian/changelog	(working copy)
@@ -1,3 +1,12 @@
+partman-efi (4) UNRELEASED; urgency=low
+
+  * Jim Lieb
+- Get file system type displayed properly
+- remove extra menu label
+- Closes: #268555
+
+ -- Jim Lieb <[EMAIL PROTECTED]>  Fri,  24 Sep 2004 12:00:00 -0700
+
 partman-efi (3) UNRELEASED; urgency=low
 
   * Joey Hess
Index: update.d/efi_visuals
===
--- update.d/efi_visuals	(revision 0)
+++ update.d/efi_visuals	(revision 0)
@@ -0,0 +1,25 @@
+#!/bin/sh
+
+. /usr/share/debconf/confmodule
+
+dev=$1
+num=$2
+id=$3
+size=$4
+type=$5
+fs=$6
+path=$7
+name=$8
+
+cd $dev
+
+[ -f $id/method ] || exit 0
+method=$(cat $id/method)
+
+case "$method" in
+efi)
+	db_metaget partman/method_short/efi description || RET=''
+	printf "${RET:-efi}" >$id/visual_filesystem
+	>$id/visual_mountpoint
+	;;
+esac

Property changes on: update.d/efi_visuals
___
Name: svn:executable
   + *

Index: update.d/_numbers
===
--- update.d/_numbers	(revision 21963)
+++ update.d/_numbers	(working copy)
@@ -1 +1,2 @@
-60 efi_sync_flag
+21 efi_sync_flag
+60 efi_visuals

Bug#271907: Partman doesn't display cciss driver names properly

2004-09-17 Thread Jim Lieb 
On Friday 17 September 2004 10:53, Joey Hess wrote:
> I haven't gotten my cciss hooked up to any disks yet, but this looks
> obviously either correct, or unlikely to break anything else, so
> applied.. If you could use some more testing of this, I can try to
> get my cciss working, just lemme know.
I do have one on my rx2600 and have tested against it.  some testing
on other archs would be useful.
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#271907: Partman doesn't display cciss driver names properly

2004-09-15 Thread Jim Lieb 
Package: partman

This is a re-submit to add the package header line.  Disreard
the previous email w/ this same subject line.  Sorry.

Partman doesn't know how to display the info for various integrated
raid controllers.  In particular, it cannot display for the HP SMART
controller (cciss).  This patch properly parses and displays the
controller.

Patch set follows as attachment

Jim

-- 
*****
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: partman/debian/changelog
===
--- partman/debian/changelog	(revision 21745)
+++ partman/debian/changelog	(working copy)
@@ -1,3 +1,10 @@
+partman (57) unstable; urgency=HIGH
+
+  * Jim Lieb
+- add correct name handling of cciss raid controllers
+
+ -- Jim Lieb <[EMAIL PROTECTED]>  Wed,  8 Sep 2004 16:00:00 -0700
+
 partman (56) unstable; urgency=low
 
   * Joey Hess
Index: partman/definitions.sh
===
--- partman/definitions.sh	(revision 21745)
+++ partman/definitions.sh	(working copy)
@@ -501,6 +501,40 @@
 	db_metaget partman/text/scsi_partition description
 	printf "$RET" ${scsinum} ${bus} ${target} ${lun} ${part} ${linux}
 	;;
+	/dev/cciss/*)
+	# /dev/cciss/hostN/targetM/disc is 2.6 form
+	# /dev/cciss/discM/disk seems to be 2.4 form
+	line=`echo $1 | sed 's,/dev/cciss/\([a-z]*\)\([0-9]*\)/\(.*\),\1 \2 \3,'`
+	cont=`echo "$line" | cut -d" " -f2`
+	host=`echo "$line" | cut -d" " -f1`
+	line=`echo "$line" | cut -d" " -f3`
+	if [ "$host" = host ] ; then
+	   line=`echo "$line" | sed 's,target\([0-9]*\)/\([a-z]*\)\(.*\),\1 \2 \3,'`
+	   lun=`echo  "$line" | cut -d" " -f1`
+	   disc=`echo "$line" | cut -d" " -f2`
+	   part=`echo "$line" | cut -d" " -f3`
+	else
+	   line=`echo "$line" | sed 's,disc\([0-9]*\)/\([a-z]*\)\(.*\),\1 \2 \3,'`
+	   lun=`echo  "$line" | cut -d" " -f1`
+	   if [ "$lun" > 15 ] ; then
+	  cont=$(($lun / 16))
+		  lun=$(($lun % 16))
+	   else
+		  cont=0
+	   fi
+	   disc=`echo "$line" | cut -d" " -f2`
+	   part=`echo "$line" | cut -d" " -f3`
+	fi
+	linux=$(mapdevfs $1)
+	linux=${linux#/dev/}
+	if [ "$disc" = disc ] ; then
+	   db_metaget partman/text/scsi_disk description
+	   printf "$RET" ".CCISS" "-" ${cont} ${lun} ${linux}
+	else
+	   db_metaget partman/text/scsi_partition description
+	   printf "$RET" ".CCISS" "-" ${cont} ${lun} ${part} ${linux}
+	fi
+	;;
 	/dev/md/*)
 	device=`echo "$1" | sed -e "s/.*md\/\?\(.*\)/\1/"`
 	type=`grep "^md${device}[ :]" /proc/mdstat | sed -e "s/^.* : active raid\([[:alnum:]]\).*/\1/"`

Bug#268491: mapdevfs disk names before showing to user

2004-09-15 Thread Jim Lieb 
On Friday 27 August 2004 19:03, Joey Hess wrote:
> Package: elilo-installer
> Severity: wishlist
> Tags: d-i
>
> It would be nicer if elilo-installer could run mapdevfs on the
> partition name(s) before displaying them in elilo-installer/bootpart.
> Since bootpart is mapdevfsed right after it's taken from the answer
> to that question, no necessary info will be lost, and users find
> devfs names offputting.

The following patch re-maps partition file names for the select
menu.

This patch also re-positions the error code capture for the elilo
command to immediately follow the command.

This patch closes bug 268491.

Jim
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: debian/elilo-installer.postinst
===
--- debian/elilo-installer.postinst	(revision 21745)
+++ debian/elilo-installer.postinst	(working copy)
@@ -57,6 +57,7 @@
 done	
 if [ -s /tmp/efi_boot.list ]; then
 for d in `cat /tmp/efi_boot.list`; do
+	d=`mapdevfs $d`
 	if [ -n "$BOOTPARTS" ]; then
 	BOOTPARTS="$BOOTPARTS, $d"
 	else
@@ -76,8 +77,6 @@
 
 db_get elilo-installer/bootpart
 bootpart=$RET
-
-bootpart=`mapdevfs $bootpart`
 rootfs=`mapdevfs $rootfs`
 
 # Write out elilo.conf
@@ -135,6 +134,7 @@
 
 chroot /target /usr/sbin/elilo --autoconf --boot $bootpart \
   --root $rootfs --efiboot
+errcode=$?
 
 ## umount sysfs if we mounted it above
 if [ $sysmount -eq 1 ]; then
@@ -143,7 +143,6 @@
   fi
 fi
 
-errcode=$?
 if [ "$errcode" = 0 ] ; then
 	info "Boot loader successfully installed into EFI partition"
 else
Index: debian/changelog
===
--- debian/changelog	(revision 21745)
+++ debian/changelog	(working copy)
@@ -18,8 +18,9 @@
 
 elilo-installer (0.0.7) unstable; urgency=low
 
-  * dann frazier
-- add dependency on fat-modules
+  * Jim Lieb
+- make partition names more people readable
+- Closes: #268491
 - add dependency on firmware-modules (for efivars)
 - modprobe efivars
 - mount sysfs for efibootmgr, if necessary.  Closes: #270934
@@ -27,7 +28,7 @@
 - Remove unsetting of seen flag, that is not necessary in the d-i
   environment and it makes preseeding difficult.
 
- -- dann frazier <[EMAIL PROTECTED]>  Fri, 10 Sep 2004 00:13:26 -0600
+ -- Jim Lieb <[EMAIL PROTECTED]>  Fri, 10 Sep 2004 14:38:00 -0700
 
 elilo-installer (0.0.6) unstable; urgency=low

Bug#268491: mapdevfs disk names before showing to user

2004-08-30 Thread Jim Lieb 
On Monday 30 August 2004 12:57, Joey Hess wrote:
> Jim Lieb wrote:
> > On Friday 27 August 2004 19:03, Joey Hess wrote:
> > > Package: elilo-installer
> > > Severity: wishlist
> > > Tags: d-i
> > >
> > > It would be nicer if elilo-installer could run mapdevfs on the
> > > partition name(s) before displaying them in
> > > elilo-installer/bootpart. Since bootpart is mapdevfsed right
> > > after it's taken from the answer to that question, no necessary
> > > info will be lost, and users find devfs names offputting.
> >
> > I agree.  But given that devfs is being deprecated in 2.6, sysfs is
> > a more reliable method of discovering devices, and devfs is so
> > deeply embedded in partman and friends, I suggest a decision on how
> > to go forward without devfs should be decided prior to making
> > changes to elilo-installer because it is not the only package that
> > needs work in this area.
>
> I don't understand your reasoning. Every other bit of d-i that
> presents a device name to the user runs it through mapdevfs. It seems
> strange to block making a change in elilo-installer on such a
> decision, epecially when such a decision is very much post-sarge.
I'll look at the script and see if this can be done.  What is there came
with the original script.  I was just pointing out that the various 
translations have added complexity and this would be one more
given that the future of devfs is by no means assured.

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#268554: weirdness trying to use existing efi partitions on ia64

2004-08-30 Thread Jim Lieb 
On Friday 27 August 2004 19:31, Joey Hess wrote:
> Package: partman-efi
> Severity: normal
> Tags: d-i
>
> I have an efi partition already from a previous, failed install.
> Partman sees it and shows the type as "EFI boot partition". It seems
> to be activated in partman, but on exit partman-efi complains that I
> have no efi boot partition.
>
> I've tried deleting the partition and re-making it, which works, but
> only if I slightly change the size. Apparently otherwise partman
> noticed I've not really changed anything, and does not format it.
> Unlike when it's dealing with other pre-existing partitions, partman
> does not add a "Format the partition: yes/no" to the partition edit
> menu, so I cannot force it to format it. Anyway, I'd rather not have
> to, as I hope to have multiple debian installs on this machine
> eventually.
>
> So to sum up, partman should let me control whether to format an
> existing efi partition or not, and partman-efi should not complain if
> the efi partition already exists and was previously formatted, or
> should perhaps complain less strenuously, since I suppose using an
> efi partition that has other cruft on it might be unsafe.
This seems to be caused by the core of partman that becomes convinced
that it should "keep" the partition.  At that point, none of the methods 
in partman-efi get called, making it somewhat difficult to add code 
there to do otherwise.  There must be some magic in the partman core
that I just don't grok yet but when partman gets convinced that either
it doesn't have to format because you aren't going mount it (which
we don't) or it hasn't been changed by your choices at some point,
it sets the method to "keep" and that is that.  Suggestions or
review by others would be welcome.

As for an efi partition with cruft in it, elilo-installer cleans that 
out so it is not unsafe.  According to elilo docs, the options used
won't carry across any previous state.  In other words, the "problem"
would not be carrying forward old cruft but potentially removing
good stuff you may want.  Safest path: expect a re-format which


-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#267689: debian installer IA64 usb keyboard

2004-08-27 Thread Jim Lieb 
On Friday 27 August 2004 12:05, Joey Hess wrote:
> Jim Lieb wrote:
> > The problem is in the way the initrd was built.  If you look
> > carefully, you will find that there is no modules.dep which crashes
> > modprobe. This *should* have been done by the initrd builder but it
> > failed because the kernel version and the udeb versions did not
> > match.  Check the /config/ia64.cfg for the correct
> > KERNEL*VERSION values.
>
> Is this fixed in current svn?
Yes, Dann committed it yesterday.  The change makes handling the
versions easier but they still need to track, esp now as the versions
are moving up a notch...
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#267689: debian installer IA64 usb keyboard

2004-08-27 Thread Jim Lieb 
On Friday 27 August 2004 11:12, Joey Hess wrote:
> Alban Crequy wrote:
> > So I tested the network pxe installation but the keyboard still not
> > work.
> >
> > So I tried to modify the initrd in order to load the usb-uhci
> > kernel module: I mount -o loop the initrd file, and add these ugly
> > lines on /sbin/debian-installer-startup:
> >
> >   depmod -a
> >   modprobe usb-uhci
> >
> > With these 2 lines, the keyboard works.
> >
> > dmesg, cpuinfo, etc. can be found at:
> > http://linux.ensimag.fr/~acrequy/bazard/Debian-IA64-usbkeybord-bugr
> >eport/
> >
> > I suggest:
> > - either detecting correctly the USB keyboard
>
> As far as I can tell the current installer should detect your usb
> controller anbd load usb-uhci for it. Both of the PCI ids (808624c2,
> 808624c4) of your usb controller are in the discover database.
> usb-discover also has these values, and it is included on the initrd.
> It's possible that this was not the case in rc1, you could try a
> current daily build to be sure.
The problem is in the way the initrd was built.  If you look carefully,
you will find that there is no modules.dep which crashes modprobe.
This *should* have been done by the initrd builder but it failed because
the kernel version and the udeb versions did not match.  Check
the /config/ia64.cfg for the correct KERNEL*VERSION values.
Once the makefile can find the correct version, it does depmod
before cramfs'ing the initrd.  With the rollup to a new kernel version,
this needs to be checked (again).  btw, usb isn't the only one.  Other
modules that are needed prior to loading of some of the various kernel
udeb's which do a (new) depmod in their postinst also break.  Usb
is critical (and noticed) simply because that forces the choice of
serial or kb/video console at the very beginning.
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#259296: ia64 menus patch fixes 259296 and 267689

2004-08-26 Thread Jim Lieb 
Here is a corrected patch.  The CD menu patch was already uploaded so
please close that bug (#259296) using the previous patch (already 
committed).

This patch fixes the usb keyboard issue (#267689).

It seems that my orphaned workspace (anon from :3691) didn't 
update properly.  This diff is based on a new workspace from
the "main" url.

Note:
I have not changed the kernel version numbers for either 2.4 or 2.6.
The reason for this patch is it fixes a problem where the udeb
version tag and the actual kernel version tag are not the same.
Since the decision on 2.4 versions is not set yet  I have left them
alone.  I'll leave it to the actual submitter/uploader to update
ia64.cfg appropriately to be in sync with what is current in the
archive and svn at commit time.

Jim

On Tuesday 24 August 2004 23:30, dann frazier wrote:
> On Tue, Aug 24, 2004 at 04:27:50PM -0700, Jim Lieb wrote:
> > This patch adds menu support to the CD install for ia64.  It
> > supplies the following:
>
> This patch doesn't apply - it tries to modify
> build/boot/ia64/elilo-cd.conf, which doesn't exist - can you double
> check & post another?

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
--- current/config/ia64.cfg	2004-08-24 08:52:55.0 -0700
+++ new/config/ia64.cfg	2004-08-17 15:02:25.0 -0700
@@ -6,11 +6,12 @@
 
 # The version of the kernel to use.
 KERNELMAJOR = 2.4
-KERNELVERSION = 2.4.26-itanium-smp
+KERNELVERSION = 2.4.26-1-itanium-smp
 KERNELVERSION_2.6 = 2.6.7-1-itanium-smp
 KERNEL_FLAVOUR = di
 KERNELNAME = vmlinuz
-KERNELIMAGEVERSION = $(KERNELVERSION)
+KERNELIMAGEVERSION = 2.4.26-itanium-smp
+KERNELIMAGEVERSION_2.6 = $(KERNELVERSION_2.6)
 
 # The DOS volume id to use for DOS floppies. This is a 32 bit hexidecimal
 # number.
diff -Naur current/config/ia64/cdrom/2.6.cfg new/config/ia64/cdrom/2.6.cfg
--- current/config/ia64/cdrom/2.6.cfg	2004-08-24 08:52:56.0 -0700
+++ new/config/ia64/cdrom/2.6.cfg	2004-08-17 14:45:10.0 -0700
@@ -1,4 +1,5 @@
 KERNELVERSION = $(KERNELVERSION_2.6)
+KERNELIMAGEVERSION = $(KERNELIMAGEVERSION_2.6)
 KERNELMAJOR = 2.6
 EXTRANAME = 2.6/
 INITRD_FS = cramfs
diff -Naur current/config/ia64/miniiso/2.6.cfg new/config/ia64/miniiso/2.6.cfg
--- current/config/ia64/miniiso/2.6.cfg	2004-08-24 08:52:56.0 -0700
+++ new/config/ia64/miniiso/2.6.cfg	2004-08-17 15:03:09.0 -0700
@@ -1,4 +1,5 @@
 KERNELVERSION = $(KERNELVERSION_2.6)
+KERNELIMAGEVERSION = $(KERNELIMAGEVERSION_2.6)
 KERNELMAJOR = 2.6
 EXTRANAME = 2.6/
 INITRD_FS = cramfs
diff -Naur current/config/ia64/netboot/2.6.cfg new/config/ia64/netboot/2.6.cfg
--- current/config/ia64/netboot/2.6.cfg	2004-08-24 08:52:56.0 -0700
+++ new/config/ia64/netboot/2.6.cfg	2004-08-17 15:03:32.0 -0700
@@ -1,4 +1,5 @@
 KERNELVERSION = $(KERNELVERSION_2.6)
+KERNELIMAGEVERSION = $(KERNELIMAGEVERSION_2.6)
 KERNELMAJOR = 2.6
 EXTRANAME = 2.6/
 INITRD_FS = cramfs

Bug#262182: Patch for elilo-installer to call apt-install elilo

2004-08-24 Thread Jim Lieb 
Please apply the following patch and close this bug.

Changes:
1. Calls apt-install elilo per request in the bug report.

2. Adds error handling for elilo failures.

Notes:
* There are template changes (added messages) that
  require translation.

* A bug report (Bug#264947: Acknowledgement (Busybox chroot
  crash/errors on Sarge 08/06 for ia64.)) was filed because of 
  failures in this code.  That bug has not been fixed and the
  failure case cannot be tested until that bug is resolved.


Jim
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: elilo-installer/debian/elilo-installer.postinst
===
--- elilo-installer/debian/elilo-installer.postinst	(revision 18778)
+++ elilo-installer/debian/elilo-installer.postinst	(working copy)
@@ -3,6 +3,22 @@
 
 . /usr/share/debconf/confmodule
 
+db_capb backup
+
+log=/var/log/messages
+
+log() {
+	logger -t elilo-installer "$@"
+}
+
+error() {
+	log "error: $@"
+}
+
+info() {
+log "info: $@"
+}
+
 if [ -f /etc/mtab ]; then
 MTAB=/etc/mtab
 else
@@ -52,6 +68,7 @@
 	db_go
 	exit 1
 fi
+rm -f /tmp/efi_boot.list
 db_subst elilo-installer/bootpart BOOTPARTS $BOOTPARTS
 
 db_fset elilo-installer/bootpart seen false
@@ -66,5 +83,45 @@
 
 # Write out elilo.conf
 
-chroot /target /usr/sbin/elilo --autoconf --boot $bootpart --root $rootfs --efiboot
+db_progress START 0 2 elilo-installer/title
 
+db_progress INFO elilo-installer/progress_installing
+
+if ! apt-install elilo ; then
+	info "Calling 'apt-install elilo' failed"
+	# Hm, unable to install elilo into /target/, what should we do?
+	db_input critical elilo-installer/apt-install-failed || [ $? -eq 30 ]
+	if ! db_go; then
+		db_progress STOP
+		exit 10 # back up to menu
+	fi
+	db_get elilo-installer/apt-install-failed
+	if [ true != "$RET" ] ; then
+		db_progress STOP
+		exit 1
+	fi
+fi
+
+db_progress STEP 1
+
+db_subst elilo-installer/progress_running bootdev "$bootpart"
+db_progress INFO elilo-installer/progress_running
+
+chroot /target /usr/sbin/elilo --autoconf --boot $bootpart \
+  --root $rootfs --efiboot
+errcode=$?
+if [ "$errcode" = 0 ] ; then
+	info "Boot loader successfully installed into EFI partition"
+else
+	db_subst elilo-installer/failed ERRCODE "$errcode"
+	db_input critical elilo-installer/failed || [ $? -eq 30 ]
+	db_progress STOP
+	db_go || exit 10 # back up
+	exit 1
+fi
+
+db_progress STEP 1
+
+db_progress STOP
+
+exit 0
Index: elilo-installer/debian/changelog
===
--- elilo-installer/debian/changelog	(revision 18778)
+++ elilo-installer/debian/changelog	(working copy)
@@ -1,3 +1,12 @@
+elilo-installer (0.0.6) unstable; urgency=low
+
+  * Jim Lieb
+- add apt-install to fix dependency on debootstrap to
+-  install bootloader.
+- Closes: #262182
+
+ -- Jim Lieb <[EMAIL PROTECTED]>  Sat, 31 Jun 2004 14:55:00 -0700
+
 elilo-installer (0.0.5) unstable; urgency=low
 
   * Colin Watson
Index: elilo-installer/debian/elilo-installer.templates
===
--- elilo-installer/debian/elilo-installer.templates	(revision 18778)
+++ elilo-installer/debian/elilo-installer.templates	(working copy)
@@ -11,7 +11,30 @@
  There were no suitable partitions found for elilo to use.  Elilo needs
  a partition with a FAT filesystem, and the boot flag set.
 
-Template: debian-installer/elilo-installer/title
+Template: elilo-installer/title
 Type: text
 #  Main menu item
 _Description: Install the elilo boot loader on a hard disk
+
+Template: elilo-installer/progress_installing
+Type: text
+_Description: Installing the ELILO package
+
+Template: elilo-installer/progress_running
+Type: text
+_Description: Running ELILO for ${bootdev}
+
+Template: elilo-installer/apt-install-failed
+Type: boolean
+Default: yes
+_Description: ELILO installation failed.  Continue anyway?
+ The elilo package failed to install into /target/.  Installing ELILO
+ as a boot loader is a required step.  The install problem might however be
+ unrelated to ELILO, so continuing the installation may be possible.
+
+Template: elilo-installer/failed
+Text: error
+_Description: ELILO installation failed
+ Running "/usr/sbin/elilo" failed with error code "${ERRCODE}".
+
+

Bug#259296: ia64 menus patch fixes 259296 and 267689

2004-08-24 Thread Jim Lieb 
This patch adds menu support to the CD install for ia64.  It supplies 
the following:

1. autoboot of CDs

2. manual updates reflecting the new functionality

3. support for booting a 2.6 kernel.  (full 2.6 support
  is another matter and another patch)

4. fixes bug 267689 (kernel does not find usb keyboard)

Issues to be resolved:
-

* bug 267689 was caused by the version on the udeb being out of
  sync with the kernel it packages.  The config file now handles this
  but could again be a problem.  The version mismatch triggers
  a makefile rule to fail to depmod on the initrd modules directory.
  This in turn causes modprobe (busybox) to fail to load any
  modules... 

* The kernels called out in the config are possibly out of date.  If
   so, please fix *all* KERNEL*VERSION* references accordingly.

* The manual update needs to be translated.
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885
Index: build/boot/ia64/general.msg
===
--- build/boot/ia64/general.msg	(revision 0)
+++ build/boot/ia64/general.msg	(revision 0)
@@ -0,0 +1,25 @@
+ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß 
+10 
+ 7fÚ´ 74General Screen7f ÃÄÄ70¿10
+ 7f³70³10
+ 7f³70There are four ways to install Debian from this CD. You can install ³10
+ 7f³70with either a 2.4 series kernel or the new 2.6 kernel.  Expert mode ³10
+ 7f³70gives you more control over the configuration of the system by  ³10
+ 7f³70giving you the option to change configuration parameters and control³10
+ 7f³70which packages are loaded.  Normal mode installs the default set of ³10
+ 7f³70packages and automatically configures them with the default settings.   ³10
+ 7f³70³10
+ 7f³70The 2.4 series Linux kernels are appropriate in existing environments   ³10
+ 7f³70where Linux 2.4 is already in production on other systems.  The 2.6 ³10
+ 7f³70kernel series is more appropriate for environments where the new³10
+ 7f³70functions and performance improvements are more important than  ³10
+ 7f³70compatibility with other systems.   ³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70  Press any key to return to main screen³10
+ 7f³70³10
+ 7f³70 Help: [71F170-General]  [71F270-Params]³10
+ 7fÀ70Ù10
Index: build/boot/ia64/params.msg
===
--- build/boot/ia64/params.msg	(revision 0)
+++ build/boot/ia64/params.msg	(revision 0)
@@ -0,0 +1,25 @@
+ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß 
+10 
+ 7fÚÄ´ 74Params Screen7f ÃÄÄ70¿10
+ 7f³70³10
+ 7f³70   You can enter boot parameters on the Boot: line. ³10
+ 7f³70³10
+ 7f³70   For the serial console:console=ttyS0,9600n8  ³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70³10
+ 7f³70

Bug#264947: Busybox chroot crash/errors on Sarge 08/06 for ia64.

2004-08-10 Thread Jim Lieb 
Package: busybox, partman

This bug is detected on sarge-netinst.iso for ia64 20040806  (RC1).
The system is an HP rx2600 dual Itanium II.  This error
occurs on both a pristine iso and an iso with my new boot.img
and an updated elilo-installer udeb installed.

The busybox shell running elilo-installer's postinst terminates 
pre-maturely, generating an incorrect error.  The upper levels
interpret this error as a failure of postinst rather than postinst
catching the error and reporting its own (more exact) error.

This is a bit of an unusual bug in that it is triggered by another
regression bug in either partman or the kernel packaging for
ia64.  The following code fragment from elilo-installer.postinst
from my working copy with debugging lines illustrates the problem:

--- snip ---
db_progress INFO elilo-installer/progress_running
echo "about to elilo bootpart=$bootpart, rootfs=$rootfs" \
   >> /tmp/elilo-installer
chroot /target /usr/sbin/elilo --autoconf --boot $bootpart \
  --root $rootfs --efiboot > /tmp/errmsg 2>&1

ERRCODE=$?
echo "done elilo" >> /tmp/elilo-installer
echo "error code=$ERRCODE." >> /tmp/elilo-installer

--- snip ---

The trigger of this event is the failure of partman or ??? to
install fat-modules causing elilo to error return because it
cannot mount the vfat f/s in $bootpart (/dev/sda1).
The trace file /tmp/elilo-installer contains "about to ..."
as its last line with the shell vars properly expanded,
in this case, bootpart=/dev/sda1 and rootfs=/dev/sda2.

The echo "done elilo" is never reached.  I have tried
a number of alternatives including "err=$()".
The file /tmp/errmsg records the error message from elilo.
In addition, the command run from the install shell
returns $? of 0.  The same command run from multi-user
(bash) returns $? of 1.

a snippit of install syslog follows with inline comments:

Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (libblkid1-udeb): mark, 
dependency from e2fsprogs-udeb
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (e2fsprogs-udeb): mark, 
dependency from partman-basicfilesystems
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (ext2-modules): search, 
dependency from partman-basicfilesystems
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver 
(kernel-image-2.4.26-itanium-smp-di): mark, dependency from 
partman-basicfilesystems
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (fat-modules): package 
doesn't exist (ignored)
  --- this occurs earlier for fat, xfs, ext3, and reiserfs as well...
  --- the udebs appear to be on the netinst iso however...
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver 
(partman-basicfilesystems): mark, dependency from partman-target

Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (bugreporter-udeb): mark
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (cdrom-checker): mark
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (di-utils-shell): mark
Aug  9 22:58:05 main-menu[238]: DEBUG: resolver (di-utils-reboot): mark
Aug  9 22:58:05 main-menu[238]: DEBUG: Menu item 'elilo-installer' 
selected
Aug  9 22:58:05 main-menu[238]: DEBUG: configure elilo-installer, 
status: 2
Aug  9 22:58:12 main-menu[238]: (process:8790): mount: fs type vfat not 
supported by kernel
Aug  9 22:58:12 main-menu[238]: (process:8790): elilo: An error occured 
mounting /dev/sda1
   -- This is where elilo dies.  This run may not have had the redirect.
Aug  9 22:58:12 main-menu[238]: WARNING **: Configuring 
'elilo-installer' failed with error code 1
  -- postinst should have handled this.
Aug  9 22:58:12 main-menu[238]: WARNING **: Menu item 'elilo-installer' 
failed.
Aug  9 22:58:58 main-menu[238]: INFO: Modifying debconf priority limit 
from 'high' to 'medium'

There are additional issues caused by the module load failure(s) that 
render the RC1 iso unusable.  I am currently investigating those as 
well.

Regards,

Jim
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Bug#257910: Installation Report ia64 netinst beta4 : no fat16 /boot partition created

2004-08-07 Thread Jim Lieb 
The partman on both the sarge daily and sid daily I picked up last week 
sometime is seriously broken since it was re-worked.  I will test 
against the "official" iso noted in the release email.  If it is still
broken, we can mark ia64 as failed regressions because not only
does elilo-installer fail but partman doesn't even have the option
to create any f/s other than ext2.

Note: this may only be an anomaly w/ that particular day's iso builds.
I will know more tomorrow or Monday after I test the new iso.

Jim
On Saturday 07 August 2004 20:59, Florence Servant wrote:
> Hi Florent,
> I had the same problem as you with my itanium i2000 and finally
> solve the problem, in weird way, but that works. I created a first
> partition on my disk and marked it as unused and bootable. I created
> my usual linux partitions on the rest of the disk, making the /
> partition bootable and installed debian with debian-installer. Of
> course, elilo could not be installed because of the missing FAT16
> partition. After the installation is finished, I rebooted the
> computer, started
> debian-installer again just enough so my scsi disk was detected. Then
> I switched to console mode using  ctrl+alt+F2, mounted my / partition
> and used /sbin/mkfs.msdos to format my first partition in FAT16. Then
> I went back to the installation (ctrl+alt+F1) and elilo could be
> installed. Cheers,
> Florence

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: sarge-ia64-netinst (July 24) installation attempt and failure on DL590/64

2004-08-05 Thread Jim Lieb 
Zan,
  Did you file a bug for this?  I can't find anything further in bug 
reports or emails to this list.

It appears that the partitioner is seriously broken in version 48 for 
ia64.  I too see this error.  The partitioner (partman) only seems to
know about ext2 and swap partitions even though others are loaded.

Jim

On Monday 02 August 2004 19:41, Zan Lynx wrote:
> Hi!  I am not sure if this is the proper list or the correct format.
> Any tips on doing this right will be appreciated!
>
> I finally got around to trying the Sarge netinst ISO that I
> downloaded July 24.  I made a list of the problems I discovered:
>
> partitioner created ext3 partitions but kernel only supports ext2.
> partitioner did not create a FAT partition for elilo.
> partitioner cannot create a FAT partition?
> base system setup fails because kernel package can't find
> /dev/ida/c0d0p1.
> rerunning base setup fails because /usr/bin/awk exists and because
> /lib/modules/kernel-blah exists.

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: partman-efi

2004-07-20 Thread Jim Lieb 
On Tuesday 20 July 2004 12:53, Joey Hess wrote:
> Jim Lieb wrote:
> > ?? I have no idea where this comes from.  The only depends I can
> > see is for partman.  Is this a dependency in partman-auto?  The
> > recipes call for a 'method{ efi }' to trigger the formatting stuff
> > in partman-efi.  If it is in partman-auto, it should be for
> > partman-efi. My read of the partman-auto build indicates that only
> > the architecture specific recipes will be picked up, creating a
> > dependency.  If they all get picked up, that is a bug.  If this is
> > a dependency param that I missed or mis-configured, my apologies.
>
> It's in your patch in the dependencies for partman-efi.
>
> +++ partman-efi/debian/control  (revision 0)
> +Package: partman-efi
> +XC-Package-Type: udeb
> +Architecture: ia64
> +Depends: efi-udeb, partman
> +Description: Add to partman support for EFI boot partitions
Let me check this out, a possible oversight on my part.  If so, I'll
send a new patch for it.
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: partman-efi

2004-07-20 Thread Jim Lieb 
On Tuesday 20 July 2004 12:12, Joey Hess wrote:
> Jim Lieb wrote:
> > the efi udeb is in the patch set along with the partman-auto
> > additions.
>
> Not in the patch I'm looking at. This has:
>
> +Depends: efi-udeb, partman
>
> And no mention of where efi-udeb is supposed to come from.

?? I have no idea where this comes from.  The only depends I can see
is for partman.  Is this a dependency in partman-auto?  The recipes
call for a 'method{ efi }' to trigger the formatting stuff in 
partman-efi.  If it is in partman-auto, it should be for partman-efi.  
My read of the partman-auto build indicates that only the architecture
specific recipes will be picked up, creating a dependency.  If they
all get picked up, that is a bug.  If this is a dependency param that
I missed or mis-configured, my apologies.

Jim

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: partman-efi

2004-07-20 Thread Jim Lieb 
On Tuesday 20 July 2004 10:53, Joey Hess wrote:
> Jim Lieb wrote:
> > The strings in the template have already been "translated" for the
> > most part.  I picked the strings from the hppa/palo boot and did an
> > s/palo/efi/g on them. The big translation part is patch set I made
> > for the manual.
>
> As far as I can tell, this string is not translated at all:
>
> +"In order to start your new Debian system, the firmware on your
> Itanium " +"system loads the boot loader from its private EFI
> partition on the hard " +"disk.  The boot loader then loads the
> operating system from that same " +"partition.  An EFI partition has
> a FAT16 filesystem formatted on it and the " +"bootable flag set.
> Most installations place the EFI partition on the first " +"primary
> partition of the same hard disk that holds the root filesystem."
>
> The "EFIboot" string is marked as fuzzy, so is "EFI boot partition
> (FAT16)". "EFI-fat16" is also untranslated.
>
I'm not sure what you mean by the details here but yes, there are some
untranslated strings, these probably being the biggest.

> > Given that without this patch, people will end up doing a complete
> > base install and *then* find out that elilo-installer can't find a
> > partition will make for bug reports like the one filed.  I'm no
> > language expert but the strings are passable and the manual is
> > necessary to keep happy clients even if it is only in english at
> > present.  I believe that is preferable to having bug reports from
> > users after they see a message that they should have seen 20 mins
> > earlier indicating that they have to repeat that 30 min exercise
> > with little enlightement as to what the problem entails.
>
> I could possibly pass on the translations, if the translation
> coordinators were ok with that, but it does seem to need a
> nonexistant efi-udeb too.
the efi udeb is in the patch set along with the partman-auto additions.
In fact, the strings are from partman-efi.  You are probably referring 
to some admin work wrt the package inventory lists for the kit.  Yes,
whatever that is, it needs to be done...

As for motivation, IA64 installers are innocently lead down a blind 
alley, only to find out at the end of the install that elilo can't be
installed, making all that work unusable.  User frustration is
directly proportional to how late in the install you tell them to
start over.  In this case, the partitioner changes catch the problem
in the very beginning whereas the existing state trips them up
at the very last step; not conducive to happy campers.  We overcome
this in the Woody based HP distro by way of a step-by-step install
manual with lots of cautions.  You see some of that reflected
in the changes I made to the manual.  btw, I still boot the sarge
on my dev disk because I spent a lot some time in that blind alley
myself.  Fortunately (for me) I get paid to wander around in alleys,
our Debian customers are not...

If there is any way we can expedite this, or, if unable, make sure it
gets in at the beginning of the next cycle we would appreciate it.
This is a show-stopper bug for the final release and I'd like to verify
that it is fixed so we can concentrate on to other issues.

Thanks,
Jim
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: partman-efi

2004-07-20 Thread Jim Lieb 
On Tuesday 20 July 2004 10:05, Joey Hess wrote:
> dann frazier wrote:
> >   Jim Lieb filed a patch for #257910 which adds the partman-efi
> > package and adds efi support to partman-auto.  I've tested an image
> > he made for us, and it worked fine on my system.
> >
> >   I'd appreciate if someone more partman knowledgable than I could
> > take a look at the patch; if it looks reasonable to such a person,
> > I'll be happy to commit these changes and do an upload.
>
> I looked at it mainly to see if we should consider it for this
> month's release. It adds new strings, and depends on a efi-udeb that
> does not seem to exist yet, so I think we should hold off on it until
> after the release, as we're currently in a string freeze, and have
> only a few days to get final udebs into the archive. After that, it
> looks ok to me, of course I don't know a lot about either partman or
> efi so am not the most qualified to look at it.

The strings in the template have already been "translated" for the 
most part.  I picked the strings from the hppa/palo boot and did an
s/palo/efi/g on them. The big translation part is patch set I made for 
the manual.

Given that without this patch, people will end up doing a complete base
install and *then* find out that elilo-installer can't find a partition 
will make for bug reports like the one filed.  I'm no language expert
but the strings are passable and the manual is necessary to keep happy
clients even if it is only in english at present.  I believe that is
preferable to having bug reports from users after they see a message 
that they should have seen 20 mins earlier indicating that they have
to repeat that 30 min exercise with little enlightement as to what the
problem entails.

Jim
-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Partman

2004-07-20 Thread Jim Lieb 
I'm the implementor of partman.efi.

Let's get this down to specifics.  All IA64 platforms and
newer server class Xeon and bigger systems use EFI instead
of the traditional BIOS that PC users are expecting.

This patch set is IA64 architecture and IA32+EFI specific.
Other partitioning methods are not affected by this change.

EFI expects a bootable FAT filesystem for the boot image, not
an MBR or partition boot block.  That is why partman-efi and
friends were written.  The primary reason for partman-efi is its
check for the bootable partition *before* the base install begins.
The elilo installer does indeed ask the question but at the end
of 10-15 mins of package install.  It asks it then because it
needs the infrastructure in /target to make things happen.  Not
having a bootable FAT available at that point requires a complete
installation re-start because the disk has to be re-partitioned and
the root is now partially loaded.

For further detail, I refer you to both the info messages in the 
template file(s) and the install manual additions for ia64 in an
accompanying patch which explain the issues in greater detail.

Our primary issue right now is to get these patches into svn and
the daily builds.  Does anyone have an indication when that  will
happen so Dann F and I finish validation on this portion of the
ia64 port?

Thanks,

Jim

On Monday 19 July 2004 20:51, John Summerfield wrote:
> Margarita Manterola wrote:
> >-= This is a code fix.  No string change needed =-
> >
> >It's kind of dumb to show the "Bootable flag" for certain partition
> >types (like swap), I think it would be better to not show it when it
> >makes no difference, so as not to confuse the user unnecesarily.
>
> To the best of my knowledge, the bootable flag has no significance
> unless you have a DOS-family MBR.
>
> Why ask at all if there is no choice to be made?
>
> If you're installing GRUB or LILO to the boot sector, there is no
> choice If you do _not_ have GRUB or LILO on the boot sector,
> and there is only one possible bootable partition there is no
> choice and there are two or three bootable partitions, one containing
> grub is a safe default.
>
> Getting the right answer without asking the user (who might not know
> anyway) might require reading some boot sectors and inspecting their
> contents.
>
>
> This, of course, points to it being a question to be answered at the
> time of installing (or not) a boot loader, and not when partitioning.
>
>
> --
>
> Cheers
> John
>
> -- spambait
> [EMAIL PROTECTED]  [EMAIL PROTECTED]
> Tourist pics
> http://portgeographe.environmentaldisasters.cds.merseine.nu/

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#257910: Installation Report ia64 netinst beta4 : no fat16 /boot partition created

2004-07-06 Thread Jim Lieb 
I am working on fixing this problem for my company and will submit
a bug report and patch very shortly.

Jim
On Tuesday 06 July 2004 09:11, Florent de Dinechin wrote:
> Package: installation-reports
> Debian-installer-version:sarge-ia64-netinst.iso beta4
>
> Machine : HP i2000 (bios 117c) but the problem is probably relevant
> to all ia64 systems.
>
> The installation of the kernel and elilo on a an ia64 system
> requires a FAT16 partition,
> (also known as an EFI partition to sound more modern)
> otherwise the installation fails during "installation of the base
> system".
>
> However the (previous) partitionning step gives no hint to that
> respect. What's worse, the suggested automatic-for-dummies options
>(create one big partition, or create two for / and /home)
> do not create this FAT16 /boot partition, leading to failure of the
> kernel installation.
>
> Besides, I did not see any hint in the error messages that
> the failure is due to a missing FAT16 partition.
>
>
> Best regards, and sorry if there is something wrong in this bug
> report, it is my first one.
>
>   Florent

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Bug#254935: The way the swap size is calculated seems not correct

2004-06-17 Thread Jim Lieb 
On Thursday 17 June 2004 13:45, Thiemo Seufer wrote:
> Margarita Manterola wrote:
> [snip]
>
> > I think it would be nicer if the parameters for swap size could be
> > taken from the memory size, the minimum being the memory size, and
> > the maximum the double of the memory size... Or something like
> > that.
>
> This might give weird results for a machine with plenty of RAM but
> small hard disk. Adding a maximum cutoff at about 3* memory size is
> probably better.
Agreed.  We have 12GB in our ia64 systems and don't need to eat
a whole 36G disk just for swap.  Just an upper limit and keep it
advisory w/ appropriate defaults.
>
>
> Thiemo

-- 
*
Jim LiebWild Open Source Inc.
[EMAIL PROTECTED]   Cell: 831.295.9317
Office: 831.421.0883Fax:  831.421.0885



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]