Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.misc > #5339 > unrolled thread

Can LILO do this too?

Back to article view | Back to comp.os.linux.misc

Contents

  Can LILO do this too? no.top.post@gmail.com - 2012-05-19 14:23 +0000
    Re: Can LILO do this too? Robert Heller <heller@deepsoft.com> - 2012-05-19 11:17 -0500
    Re: Can LILO do this too? Dan Espen <despen@verizon.net> - 2012-05-19 12:31 -0400
      Re: Can LILO do this too? SM <kasmra@ne-spamon.gmail.com> - 2012-05-19 17:41 +0000
        Re: Can LILO do this too? Dan Espen <despen@verizon.net> - 2012-05-19 14:36 -0400
          Re: Can LILO do this too? SM <kasmra@ne-spamon.gmail.com> - 2012-05-19 18:58 +0000
            Re: Can LILO do this too? Dan Espen <despen@verizon.net> - 2012-05-19 15:39 -0400
          Re: Can LILO do this too? John Hasler <jhasler@newsguy.com> - 2012-05-19 13:59 -0500
          Re: Can LILO do this too? "Chris F.A. Johnson" <cfajohnson@gmail.com> - 2012-05-19 14:59 -0400
        Re: Can LILO do this too? Robert Heller <heller@deepsoft.com> - 2012-05-19 17:27 -0500
    Re: Can LILO do this too? Chick Tower <c.tower@deadspam.com> - 2012-05-20 18:57 +0000
      Re: Can LILO do this too? Michael Black <et472@ncf.ca> - 2012-05-20 18:50 -0400
        Re: Can LILO do this too? Chick Tower <c.tower@deadspam.com> - 2012-05-21 21:53 +0000

#5339 — Can LILO do this too?

Having grub setup to boot from a CompactFlash,
and then plugging eg. a USB disk before you boot,
means that the CF may change from being /dev/sdb,
to being /dev/sdc,
because the USB disk gets 'allocated in front of'
the CF -- apparently.

With grub, you can just 'live edit' the immediately
visible menu, from sdb to sdc.

Besides grub can access a device via its LABEL;
so it doesn't matter <where the device is located>.

What do you do if you've got LILO instead?!

How is it possible that a 'knl2.4 grub' can see
a knl2.4 originated ext2, but can't see a
knl2.6 originated ext2 ?

Since grub has got the various FS-drivers: ext2, ext3, js*,
one can understand that it couldn't read an ext3 if the
driver was missing; but aren't ALL ext2 equal, independant
of the kernel-version that created them?
What about different inode sizes, can grub handle all?

Of course, it's nice with grub: it shows you what IT can see,
because even before the OS has booted, grub can effectively
do a <fdisk -l>. But with LILO, you are in the dark.

The <old>-grub has some quirks.
Is it worth getting grub-2 -- for distributions designed for LILO?

I've got a little-known-slak-based distribution CD, which allows
<editing of the knl, initrd, vga ..etc> before booting, which seems
nice; but I'm confused, because I thought this distribution used LILO.
 What does the <syslinux> system do?

== TIA

[toc] | [next] | [standalone]

#5340

At Sat, 19 May 2012 14:23:38 +0000 (UTC) no.top.post@gmail.com wrote:

> 
> Having grub setup to boot from a CompactFlash,
> and then plugging eg. a USB disk before you boot,
> means that the CF may change from being /dev/sdb,
> to being /dev/sdc,
> because the USB disk gets 'allocated in front of'
> the CF -- apparently.
> 
> With grub, you can just 'live edit' the immediately
> visible menu, from sdb to sdc.
> 
> Besides grub can access a device via its LABEL;
> so it doesn't matter <where the device is located>.
> 
> What do you do if you've got LILO instead?!

You can't do this.  LILO only uses low-level BIOS sector I/O calls
using BIOS defined device IDs.  It is unfortunate that some BIOS will
'move' hard disks #'s around when a *removable* device is plugged in. 
*I* would consider this a BIOS design defect -- complain to the BIOS
vendor.  (I'm assuming that the CF device is a 'hard connected' device,
in the sense that the CF interface is 'soldered' onto the motherboard
and is not itself a USB pluggable device.)  Note that what the kernel
driver scan does later is a completely separate issue.  One should NOT
depend on using '/dev/sdXY' in /etc/fstab or as an argument to the root=
kernel parameter.  It is always going to be better to use LABEL=<mumble> or
UUID=<mumble> in both of these cases. 

> 
> How is it possible that a 'knl2.4 grub' can see
> a knl2.4 originated ext2, but can't see a
> knl2.6 originated ext2 ?
> 
> Since grub has got the various FS-drivers: ext2, ext3, js*,
> one can understand that it couldn't read an ext3 if the
> driver was missing; but aren't ALL ext2 equal, independant
> of the kernel-version that created them?

Possibly not.  I guess there are really something like ext2.1, ext2.2,
etc. I suspect that the *real* kernel drivers are 'smart enough' to
deal with the differences (eg a kernel 2.4 ext2 driver can handle a FS
built under kernel 2.6), but the grub 'drivers' are not as clever -- eg
they don't have the sorts of fall-back handling for 'extensions' they
don't understand.

> What about different inode sizes, can grub handle all?
> 
> Of course, it's nice with grub: it shows you what IT can see,
> because even before the OS has booted, grub can effectively
> do a <fdisk -l>. But with LILO, you are in the dark.
> 
> The <old>-grub has some quirks.
> Is it worth getting grub-2 -- for distributions designed for LILO?
> 
> I've got a little-known-slak-based distribution CD, which allows
> <editing of the knl, initrd, vga ..etc> before booting, which seems
> nice; but I'm confused, because I thought this distribution used LILO.
>  What does the <syslinux> system do?
> 
> == TIA
> 
>                                                          

-- 
Robert Heller             -- 978-544-6933 / heller@deepsoft.com
Deepwoods Software        -- http://www.deepsoft.com/
()  ascii ribbon campaign -- against html e-mail
/\  www.asciiribbon.org   -- against proprietary attachments


                                                                                                                              

[toc] | [prev] | [next] | [standalone]

#5341

no.top.post@gmail.com writes:

> Having grub setup to boot from a CompactFlash,
> and then plugging eg. a USB disk before you boot,
> means that the CF may change from being /dev/sdb,
> to being /dev/sdc,
> because the USB disk gets 'allocated in front of'
> the CF -- apparently.
>
> With grub, you can just 'live edit' the immediately
> visible menu, from sdb to sdc.
>
> Besides grub can access a device via its LABEL;
> so it doesn't matter <where the device is located>.
>
> What do you do if you've got LILO instead?!

You start using a piece of software that is pretty
much obsolete.

I see that someone restarted development in 2010.
Given that the documentation is still not updated,
I'd say it's dead.

> How is it possible that a 'knl2.4 grub' can see
> a knl2.4 originated ext2, but can't see a
> knl2.6 originated ext2 ?

Read the source or ask on the mailing list.

> Since grub has got the various FS-drivers: ext2, ext3, js*,
> one can understand that it couldn't read an ext3 if the
> driver was missing; but aren't ALL ext2 equal, independant
> of the kernel-version that created them?
> What about different inode sizes, can grub handle all?

Read the source or ask on the mailing list.

> Of course, it's nice with grub: it shows you what IT can see,
> because even before the OS has booted, grub can effectively
> do a <fdisk -l>. But with LILO, you are in the dark.

Which leaves one wondering why LILO in this day and age.

> The <old>-grub has some quirks.
> Is it worth getting grub-2 -- for distributions designed for LILO?

Using old software when there is new and better is a fools game.

> I've got a little-known-slak-based distribution CD, which allows
> <editing of the knl, initrd, vga ..etc> before booting, which seems
> nice; but I'm confused, because I thought this distribution used LILO.
>  What does the <syslinux> system do?

http://www.syslinux.org/wiki/index.php/The_Syslinux_Project

-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#5342

2012-05-19, Dan Espen wrote:
> no.top.post@gmail.com writes:
>
> Which leaves one wondering why LILO in this day and age.

Because it Just Works™.

-- 
:wq

[toc] | [prev] | [next] | [standalone]

#5343

SM <kasmra@ne-spamon.gmail.com> writes:

> 2012-05-19, Dan Espen wrote:
>> no.top.post@gmail.com writes:
>>
>> Which leaves one wondering why LILO in this day and age.
>
> Because it Just Works™.

Obviously for a seriously limited range of values for "works".

Besides you just cut out my point about active development.

-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#5344

2012-05-19, Dan Espen wrote:
> SM <kasmra@ne-spamon.gmail.com> writes:
>
>> 2012-05-19, Dan Espen wrote:
>>> no.top.post@gmail.com writes:
>>>
>>> Which leaves one wondering why LILO in this day and age.
>>
>> Because it Just Works™.
>
> Obviously for a seriously limited range of values for "works".

Yes. For more general cases, it works.

> Besides you just cut out my point about active development.

The point was noted. I'm not sure if LILO is actively developed or not.
In any case I haven't heard any news that suggested so. In my personal
experience LILO has always worked where GRUB has failed. But then again,
my use cases have not been this advanced.

-- 
:wq

[toc] | [prev] | [next] | [standalone]

#5347

SM <kasmra@ne-spamon.gmail.com> writes:

> 2012-05-19, Dan Espen wrote:
>> SM <kasmra@ne-spamon.gmail.com> writes:
>>
>>> 2012-05-19, Dan Espen wrote:
>>>> no.top.post@gmail.com writes:
>>>>
>>>> Which leaves one wondering why LILO in this day and age.
>>>
>>> Because it Just Works™.
>>
>> Obviously for a seriously limited range of values for "works".
>
> Yes. For more general cases, it works.
>
>> Besides you just cut out my point about active development.
>
> The point was noted. I'm not sure if LILO is actively developed or not.
> In any case I haven't heard any news that suggested so. In my personal
> experience LILO has always worked where GRUB has failed. But then again,
> my use cases have not been this advanced.

This appears to be the LILO home page:

http://lilo.alioth.debian.org/

They say development stopped and then started.
But when I went to read the documentation it's all
marked "old".

I'd say a project without documentation is no project at all.


-- 
Dan Espen

[toc] | [prev] | [next] | [standalone]

#5345

Dan Espen writes:
> Which leaves one wondering why LILO in this day and age.

SM writes:
> Because it Just Works™.

Dan Espen writes:
> Obviously for a seriously limited range of values for "works".

A range which includes everything I need.

Dan Espen writes:
> Besides you just cut out my point about active development.

LILO doesn't need "active development": just maintenance.  The fact that
"features" are not being added is a plus.
-- 
John Hasler 
jhasler@newsguy.com
Dancing Horse Hill
Elmwood, WI USA

[toc] | [prev] | [next] | [standalone]

#5346

On 2012-05-19, Dan Espen wrote:
> SM <kasmra@ne-spamon.gmail.com> writes:
>
>> 2012-05-19, Dan Espen wrote:
>>> no.top.post@gmail.com writes:
>>>
>>> Which leaves one wondering why LILO in this day and age.
>>
>> Because it Just Works?.
>
> Obviously for a seriously limited range of values for "works".

    Who cares how "limited" is it if it does what is needed?

    (I prefer lilo to grub, though not enough to go to the bother of
    switching.)

> Besides you just cut out my point about active development.

    If it does what is needed, why does it need further development?


-- 
   Chris F.A. Johnson, <http://cfajohnson.com>
   Author:
   Pro Bash Programming: Scripting the GNU/Linux Shell (2009, Apress)
   Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)

[toc] | [prev] | [next] | [standalone]

#5348

At Sat, 19 May 2012 17:41:16 +0000 (UTC) SM <kasmra@ne-spamon.gmail.com> wrote:

> 
> 2012-05-19, Dan Espen wrote:
> > no.top.post@gmail.com writes:
> >
> > Which leaves one wondering why LILO in this day and age.
> 
> Because it Just Works™.
> 

+1


-- 
Robert Heller             -- 978-544-6933 / heller@deepsoft.com
Deepwoods Software        -- http://www.deepsoft.com/
()  ascii ribbon campaign -- against html e-mail
/\  www.asciiribbon.org   -- against proprietary attachments


                                                             

[toc] | [prev] | [next] | [standalone]

#5363

On 2012-05-19, no.top.post@gmail.com <no.top.post@gmail.com> wrote:
> Since grub has got the various FS-drivers: ext2, ext3, js*,
> one can understand that it couldn't read an ext3 if the
> driver was missing; but aren't ALL ext2 equal, independant
> of the kernel-version that created them?
> What about different inode sizes, can grub handle all?

I don't know about your ext2 question, n.t.p., but it may be analogous
to how Slackware changed how it creates an ext3 filesystem during
installation.  I believe they changed the inode size, but I don't
recall; it was in the changelog.  They said it was in preparation for
using ext4, when it was new.  When I installed Slackware 13.0, and
formatted my root partition in ext3, legacy grub could no longer
recognize the filesystem, so I had to install lilo in Slackware's root
partition and modify my menu.lst to just chainload Slackware.  So it's
possible that an ext2 partition with a "non-standard" inode size would 
confound legacy grub.
-- 
                                 Chick Tower

For e-mail:  colm DOT sent DOT towerboy AT xoxy DOT net

[toc] | [prev] | [next] | [standalone]

#5365

On Sun, 20 May 2012, Chick Tower wrote:

> On 2012-05-19, no.top.post@gmail.com <no.top.post@gmail.com> wrote:
>> Since grub has got the various FS-drivers: ext2, ext3, js*,
>> one can understand that it couldn't read an ext3 if the
>> driver was missing; but aren't ALL ext2 equal, independant
>> of the kernel-version that created them?
>> What about different inode sizes, can grub handle all?
>
> I don't know about your ext2 question, n.t.p., but it may be analogous
> to how Slackware changed how it creates an ext3 filesystem during
> installation.  I believe they changed the inode size, but I don't
> recall; it was in the changelog.  They said it was in preparation for
> using ext4, when it was new.  When I installed Slackware 13.0, and
> formatted my root partition in ext3, legacy grub could no longer
> recognize the filesystem, so I had to install lilo in Slackware's root
> partition and modify my menu.lst to just chainload Slackware.  So it's
> possible that an ext2 partition with a "non-standard" inode size would
> confound legacy grub.

But does Slackware allow one to define the inode size anymore?

I installed 12.2 on secondary computer a couple of weeks ago, and when I 
specified ext2, there was no further box to define the inode size. And 
since that seemed familiar, I think the same was the case with 12.0.

I have a relatively small /home partition, and since most of it would be 
text, I had been formatting it with smaller inodes, whereas a partition 
with larger files would be better off (I think) with larger inode size.

So I noticed the absence, when the option disappeared.  Either that, or 
I've forgotten some stage.

    Michael

[toc] | [prev] | [next] | [standalone]

#5391

On 2012-05-20, Michael Black <et472@ncf.ca> wrote:
> But does Slackware allow one to define the inode size anymore?

I don't know, Michael, but I suspect you could format your partitions as
you like them before installing Slackware.

Here's something I found in the 12.1 ChangeLog.txt.  I guess I need to
use a newer version of legacy grub to boot Slackware now, rather than
rely upon lilo for it.

"Thu Mar  6 03:13:15 CST 2008
 
extra/grub/grub-0.97-i486-4.tgz:  Patched to handle the new 256 byte
ext2/ext3 inode size.  Thanks to Michael Wagner for the patch."
-- 
                                 Chick Tower

For e-mail:  aols2 DOT sent DOT towerboy AT xoxy DOT net

[toc] | [prev] | [standalone]

Back to top | Article view | comp.os.linux.misc

csiph-web