Re: debian-wheezy-amd64-efi-test3.iso does not boot from CD via BIOS
On Thu, Sep 06, 2012 at 08:54:35PM +0200, Thomas Schmitt wrote:
>i tested bootability of
> GIGABYTE GA-MA74GM-S2H AM2
> Athlon II X4 620 Quad-Core 4x 2.6GHz
>which is probably not UEFI-ready.
>At least it does not tell anything about "EFI" when booting.
>The image boots from USB stick but not from CD.
>Strange, because i see an El Torito ISOLINUX boot image for x86 BIOS
>before the GRUB boot image for EFI.
>A repacked image with
> -boot-load-size 4
> -boot-load-size 12
>boots from CD and USB stick.
>Is size 12 intentional ? Must it be ?
>My man page advises 4 as "safe default for non-emulating boot".
>(If i only could remember why exactly i wrote that ...)
Ah, bugger. I changed the code in debian-cd to recalculate the boot
load size value a while back, and I also did that for the x86 isolinux
image when I should have left it at 4. Thanks for the report, I'll fix
>Further i see that fdisk -lu is once again unhappy with a xorriso
>generated partition table. Probably because of -append_partition.
>I will have to investigate.
>isohybrid prescribes an alignment to 1 MiB, but i am not sure why
>and whether it applies to images with more than one partition.
>Shall i strive for alignment to 1 MiB with each partition if
>isohybrid is activated ?
I honestly don't know if that's needed, but it can't hurt to try. Then
again, does it actually matter if fdisk is unhappy so long as things
work? The latest fdisk code explicitly sets minimum alignment to 1MB
on new partition creation to make sure that we don't trigger issues
with different sector sizes, but existing partitions on (read-mostly)
USB sticks should work fine as far as I can see.
Steve McIntyre, Cambridge, UK. firstname.lastname@example.org
"I've only once written 'SQL is my bitch' in a comment. But that code
is in use on a military site..." -- Simon Booth