From: David Woodhouse <[EMAIL PROTECTED]>
Subject: Re: [patch-2.4.0-test1-ac4]linux-2.4.0-test1-ac4+grub
Date: Tue, 30 May 2000 18:48:26 +0100
> The fact that Grub passes an incorrect mem= parameter in the first place,
> which needs to be overridden by manually adding the second mem= option, is
According to Christoph Plattner:
> This is correct. But my point of view does not violate yours. In
> the bootptab (or in the analog file in dhcp), you set up a 'rp=', a
> 'bf=' etc. for every computer (client).
I ask you to believe that if I'm going to support backwards
compatibility, I need to
> Gregg C Levine writes:
GCL> The machine who is running Intel based Red Hat Linux, currently
GCL> ver 5.1, boots from a disk. I elected not to mount the booting
GCL> mechanism on the MBR of the first disk drive. And my question is
GCL> what is being provided for these configurations in l
[EMAIL PROTECTED] said:
>
> Thanks for your report, but I don't think this problem can be solved
> in GRUB, since you may pass unusual parameters with the "mem" option
> to the kernel (such as "no-hlt") and so GRUB cannot interpret what you
> specified in a reliable way. IMO, the Linux kernel
Hello from Gregg C Levine usually with Jedi Knight Computers
That "This" is when you install Red Hat Linux, (Insert version, and
processor type), onto an appropriate grouping of hard drives. And then
instead of placing the boot records, onto the MBR, Master Boot Record, of
the first drive, it goes
From: Christoph Plattner <[EMAIL PROTECTED]>
Subject: Re: appropriate code for a GRUB-specific option
Date: Tue, 30 May 2000 14:54:01 +0200
> I never wanted and want to violate against standards. Standards are
> the most important thing in my opinion and the basic for the
> UNIX world.
I'm sor
> OKUJI Yoshinori writes:
OY> Please update the section "Stage2 errors" in the documentation
OY> as well. I think this kind of inconsistency is difficult to find
OY> afterward, so it is important to change both the code and the
OY> documentation at a time.
Alright, done.
--
Gordon M
> Chip Salzenberg writes:
CS> This silences two warnings when compiling eepro100.c.
Ok, done.
--
Gordon Matzigkeit <[EMAIL PROTECTED]> //\ I'm a FIG (http://fig.org/)
Committed to freedom and diversity \// I use GNU (http://www.gnu.org/)
I never wanted and want to violate against standards. Standards are
the most important thing in my opinion and the basic for the
UNIX world.
My idea - if good or not - was to say that the root path 'rp=' is
seen as 'GRUB root path' holding the 'menu.lst' (if we follow standards,
then only the pat
From: [EMAIL PROTECTED]
Subject: Re: [patch-2.4.0-test1-ac4]linux-2.4.0-test1-ac4+grub
Date: 30 May 2000 14:37:34 +0200
> a bigger problem is where the initrd is loaded. If you precise less memory than
> grub detected, grub loads the initrd too high in memory. The kernel can't access
> it.
Tha
From: "Gregg C Levine" <[EMAIL PROTECTED]>
Subject: RE: OT:What povisions are being made for machines who boot fromdisks?
Date: Tue, 30 May 2000 08:40:46 -0400
> correctly. It takes into account most things, but this is one eventuality
> that the writers missed.
What is "this"? I can't see wha
Hello from Gregg C Levine usually with Jedi Knight Computers
It just so happens that I read the documentation for GRUB, twice over, and
the installation documentation three times to make sure I interpeted it
correctly. It takes into account most things, but this is one eventuality
that the writers
OKUJI Yoshinori <[EMAIL PROTECTED]> writes:
> From: HIBINO Kei <[EMAIL PROTECTED]>
> Subject: [patch-2.4.0-test1-ac4]linux-2.4.0-test1-ac4+grub
> Date: Tue, 30 May 2000 19:44:14 +0900
>
> > This parameter is inconsistent with 'mem=' parameter
> > specified by an user.
> >
> > For example, when
From: "Gregg C Levine" <[EMAIL PROTECTED]>
Subject: OT:What povisions are being made for machines who boot from disks?
Date: Tue, 30 May 2000 08:05:49 -0400
> boots from a disk. I elected not to mount the booting mechanism on the MBR
> of the first disk drive. And my question is what is being pro
From: Christoph Plattner <[EMAIL PROTECTED]>
Subject: Re: appropriate code for a GRUB-specific option
Date: Mon, 29 May 2000 09:12:32 +0200
> We should consider, that a server admin has to define bootp (or dhcp)
> to allow diskless booting of grub.To define the boot file the
> 'bf=' option is use
Hello from Gregg C Levine usually with Jedi Knight Computers
The machine who is running Intel based Red Hat Linux, currently ver 5.1,
boots from a disk. I elected not to mount the booting mechanism on the MBR
of the first disk drive. And my question is what is being provided for these
configuratio
From: HIBINO Kei <[EMAIL PROTECTED]>
Subject: [patch-2.4.0-test1-ac4]linux-2.4.0-test1-ac4+grub
Date: Tue, 30 May 2000 19:44:14 +0900
> This parameter is inconsistent with 'mem=' parameter
> specified by an user.
>
> For example, when GRUB detects 65550K bytes memory size,
> and an user specifi
This is correct. But my point of view does not violate yours.
In the bootptab (or in the analog file in dhcp), you set up
a 'rp=', a 'bf=' etc. for every computer (client).
So machine A has to boot with GRUB ->
:
:bf=/tftpboot/grub/nbGrub:\
:rp=/tftpboot/machineA/menu.lst
Hello.
We found linux-2.4.0-test1-ac1 can't boot with GNU GRUB + User
'mem=' parameter on iX86 architecture
1. The problem of Linux kernel memory region setting code
(arch/i386/kernel/setup.c in Linux kernel tree)
On linux-2.2.XX, only one memory region is prepared at boot time.
For example,
According to Christoph Plattner:
> Only the 'menu.lst'-file should configure which OS, which root paths,
> etc. The bootptab or the dhcp configuration only should define
> which machine which ip-address and which 'menu.lst' file the machine
> gets.
Your point is well-taken. However, it applies o
You are right with your consideration. This could be a bad idea.
On the other hand, we violate an other idea !
Only the 'menu.lst'-file should configure which OS, which root paths,
etc. The bootptab or the dhcp configuration only should define
which machine which ip-address and which 'menu.lst' fi
According to OKUJI Yoshinori:
> > +#ifdef SUPPORT_DISKLESS
> > + if (boot_drive == NETWORK_DRIVE)
> > +{
> > + print_network_configuration ();
> > + printf ("\n");
> > +}
> > +#endif
>
> The idea is good, but why do you need to check if BOOT_DRIVE is
> NETWORK_DRIVE?
I've in
22 matches
Mail list logo
