[syslinux] isohybrid and ISO images whose size is not a multiple of 2048 bytes vs. VirtualBox

[Ady]

> I just want to comment on this Oh nooo, virtualbox can boot because of
> image not an even multiple of 2048 sector, so we must do some crazy
> stuff.
> 
> 
> To Tails creators (and anyone else that might be interested):
> First I want to hear if anyone has failed to boot such non-even-
> multiple-of-2048-sectors on real hardware?
 
On real optical devices, it won't happen.
 
> 
> Because the virtualbox I tried 2013 or 2014 (don't remember exactly
> when) couldn't even boot an image of selfmade SLES 9.3 boot
> cd. Meanwhile that image burned to a real CD does boot on real
> hardware. I haven't bothered with virtualbox since.
> 
> 
> To Thomas (and anyone else that might be interested):
> That fdisk complains means nothing on current BIOSes and harddisks (and
> SSDs). Current hardware does use LBA. That cylinder misalignment is
> only for the obsolete CHS handling. Every partitioner on Linux seems
> to have their own weird ideas how it should be and noone is the same
> as any other.
> 
> However I don't know much about how CD booting works. So it's possible
> that that still uses that old stuff. Again, do anyone has some
> evidence that these non-even-multiple-of-2048-sectors doesn't work on
> real hardware? It seems to me from Thomas mails that Debian has
> (inadvertedly) tested this successfully.
> 
> 
> -- 
> MartinS
 
The CHS values still need to be adequate, and the resulting isohybrid 
image should still adhere to the 2048 bytes per sector. In other words, 
the isohybrid tool should only improve users' experience.

Today a problem is seen in virtualbox (and VBox is not the problem), 
tomorrow under some other circumstances.

The TAILS' team has reported this issue, but the problem is not related 
specifically to TAILS.

This is about the isohybrid tool.

The issue, IMO, should be solved. The suggested patch is simple-enough, 
and the "cost" is, IMO, not at all excessive.

An additional patch could be implemented so to introduce an alternative 
command-line option to relax the condition (reducing the zero-padding a 
little). The default behavior should be to respect / follow adequate 
formats / standards / requirements.
 
Regards,
Ady.
 
> _______________________________________________
> Syslinux mailing list
> Submissions to Syslinux at zytor.com
> Unsubscribe or set options at:
> http://www.zytor.com/mailman/listinfo/syslinux
>