Hello Thomas.
Grub is currently very PC-specific code. When the project was born the
main goal was implementing a boot loader that can work with the several
OS's you find in the PC arena, while offering a standardized interface so
every OS could follow the specs and be able to be booted by
Title: Bug in stage2 loader (stage2/start.S)
To the grub team!
If one boots with lba mode enabled, than the routine copy_buffer
in stage2/start.S alters the cx register. But the ecx register
holds the total number of sectors on disc and the geometry check
fails! A bugfix to solve this
It seems that no one must have tried to use the 'setkey' command in GRUB
before today. I finally got around to starting to put together my dvorak
keymap for it[1], I found a nasty bug.
The issues is the table that GRUB uses internally to store the key
names, combined with the nature of the
I am planning to include the following changes in 0.5.96.1 (which we
agree is necessary because of the severity of the write_to_partition
bug):
* Fix for write_to_partition
* Fix for grub-install when using a separate /boot partition
* Nachtnebel's LBA fix for NO_BUGGY_BIOS_IN_THE_WORLD
Do only
Previously Gordon Matzigkeit wrote:
I'm about to release 0.5.96... perhaps you could try again with that
version and see if it does better?
Still doesn't work, but it shows more information:
grub find /boot/grub/stage1
(hd0,3)
grub root (hd0,3)
Filesystem type is ext2fs, partition type
From: Herbert Nachtnebel [EMAIL PROTECTED]
Subject: Bug in stage2 loader (stage2/start.S)
Date: Mon, 9 Oct 2000 12:24:08 +0200
If one boots with lba mode enabled, than the routine copy_buffer
in stage2/start.S alters the cx register. But the ecx register
holds the total number of sectors on
From: Michael Sullivan [EMAIL PROTECTED]
Subject: Initrd diskless netboot problems
Date: Mon, 9 Oct 2000 01:10:38 -0400
rdev -R vmlinuz.rd 0
rdev vmlinuz.rd /dev/initrd
IIRC, you must set the root device to /dev/ram instead of
/dev/initrd.
GRUB loads everything fine - both the kernel and
From: Thierry Laronde [EMAIL PROTECTED]
Subject: Re: [0.5.96] "Error 29: Disk write error", when installing GRUB
Date: Sun, 8 Oct 2000 17:58:16 +0200
Success, but not straightforward. There is a distinction that I do not
clearly understand. When one creates a floppy (with a ext2fs), and makes
A good idea to see in GRUB a protable boot loader. But we have to
consider, that workstations for example already have boot monitors
with complex functionality, for example handling the system environment.
So my idea is, that GRUB should also learn all the addition features
of the different
I comment on the issue a bit. As Christoph pointed out, one of the
features in GRUB is that it is (relatively) user-friendly. And, we
say, "GRUB is useful for experts like OS developers
simultaneously". But is the latter really true?
The strategy used in GRUB is to _pass information_ about
I think there is another problem with the line shown by Thierry:
grubdir_device=`df ${grubdir} | grep /dev/ \
| sed 's%.*\(/dev/[a-z0-9]*\).*%\1%'`
The sed expression is wrong as it restricts device names to be lowercase
alphanumeric chars. This is a bad assumption. I'd prefer to make
no
On Tue, Oct 10, 2000 at 12:56:06AM +0900, OKUJI Yoshinori wrote:
From: Thierry Laronde [EMAIL PROTECTED]
Subject: Re: [0.5.96] "Error 29: Disk write error", when installing GRUB
Date: Sun, 8 Oct 2000 17:58:16 +0200
Success, but not straightforward. There is a distinction that I do not
I have to say, that I don't know very much of Sun boot monitor, except
for that, the I called help and read about the commands on my
workstation (Sun SparcStation 5). But what I could see, that it is
possible to setup enviornmental settings. This is the only thing I
miss under GRUB. On the target
OKUJI Yoshinori wrote:
...
The strategy used in GRUB is to _pass information_ about hardware
and user input data to an OS image, while the strategy used in both
Open Firmware and EFI is to _give interfaces_ to access hardware. In
other words, GRUB provides higher level interfaces to an
Christoph Plattner writes:
CP But it would be a nice enhancement to support also an environment
CP block additional to the command line parameters
CP This only should be a "kick" for our design brain
CP stroming... this is no requirement !
Figure's O namesystem addresses this issue
Hi!
Is there a way to pass the information retrieved via bootp as linux kernel
commandline?
This would eliminate the need to have linux do bootp itself. Why do I want
to avoid this? there is no usable userspace bootp client for linux 2.2.x
and there is no way to use the kernel bootp client when
From: Thierry Laronde [EMAIL PROTECTED]
Subject: Re: [0.5.96] "Error 29: Disk write error", when installing GRUB
Date: Mon, 9 Oct 2000 23:34:04 +0200
For the installation on the MBR, root (hd0,0), setup (hd0) has _not_
taken care of the menu.lst. But the main difference here is that, with
the
From: "Thomas T. Soares" [EMAIL PROTECTED]
Subject: Re: About grub feature
Date: Mon, 09 Oct 2000 20:57:21 -0200
It would be nice to know what is EFI... could someone point me a URL
about
it?
http://developer.intel.com/technology/efi/efi.htm
Okuji
18 matches
Mail list logo