So consider a set of usefull modifications to the current spec and
modify it. I don't mean to hijack the multiboot spec.
Thank you.
Nicolas
__ Reply Separator _
Subject: Re: ELF loading
Author: okuji ([
From: [EMAIL PROTECTED]
Subject: Re: ELF loading
Date: Fri, 25 Feb 2000 13:20:10 +0100
> And why not?
It is not reasonable at all to change the specification just due to
the GRUB-specific limitation. Changing a specification must be done
enough conservatively, otherwise it will easily spo
And why not?
Is it so painfull to sacrifice a few pages of memory?
Nicolas
__ Reply Separator _
Subject: Re: ELF loading
Author: okuji ([EMAIL PROTECTED]) at unix,mime
Date:25.02.00 13:03
Can't you us
> "OY" == OKUJI Yoshinori <[EMAIL PROTECTED]> writes:
OY> Now I can boot Linux by downloading the image from my
OY> network. Perhaps your problem is fixed as well.
Great, I'll check it out!
Can't you use this trick:
1. GRUB loads your OS at >1MB.
2. Your OS copies the necessary data of the Multiboot information to
a safe place (for example, in the heap).
3. Your OS relocates the SMP code.
4. Jump to the relocated address.
From: Stephen Early <[EMAIL PROTECTED]>
Subject: Re: serial console support
Date: Thu, 24 Feb 2000 17:30:40 + (GMT)
> A command line is best. If menus are required, you can simply ask for
> the appropriate item number or name to be entered.
Okay.
> You can assume CR, LF and BS. Don't assu
I wasn't clear on this. My OS is going to use some part of the memory
below 511Kb in the sense that I am going to link some part (the SMP
boot code) in a location which is below 511Kb (with p_vaddr ==
p_paddr). Which means that I will not be able to check at OS runtime
Now I can boot Linux by downloading the image from my
network. Perhaps your problem is fixed as well.
From: [EMAIL PROTECTED]
Subject: ELF loading
Date: Fri, 25 Feb 2000 10:47:03 +0100
> grub-0.5.93.1/stage2/boot.c lines 423:
>
> if (type == KERNEL_TYPE_FREEBSD)
> memaddr = RAW_ADDR (phdr->p_paddr & 0xFF);
> else
> memaddr = RAW_ADDR
grub-0.5.93.1/stage2/boot.c lines 423:
if (type == KERNEL_TYPE_FREEBSD)
memaddr = RAW_ADDR (phdr->p_paddr & 0xFF);
else
memaddr = RAW_ADDR (phdr->p_paddr);
Honestly I find these two lines really strange if you consider that
10 matches
Mail list logo
