Web lists-archives.com

Re: DD bs=4M option on USB mem-stick creates false format




Hi,

kAt wrote:
> All I can say is that I feel honored that some useful code was produced
> with my problem statement as an inspiration.

Whether it's useful will still have to turn out. :)
Up to now it has one happy user.

I am pondering about the partition table mess since quite a while.
On the one hand it is what has been tested to halfways work since years.
(Although recently an 8 year old isohybrid bug was discovered which
 prevents very old BIOSes from booting off the stick. So the testing
 cannot have been all that intense and dilligent.)
On the other hand it just violates specs and software expectations.

A more standards compliant partition layout might exclude some exotic
or meanwhile antique machines.

For example it turned out in experiments with grub-mkrescue that some
EFI firmwares boot from an MBR-only layout only if the big ISO partition
has type 0x00 (= "Empty").
In the mjg layout of Debian ISOs, the big partition must have type 0x00,
because else most EFIs would object the nesting of partition 2 inside
partition 1. So the problem of some few EFIs was worked around by
accident.

It's a mine field.
So it might be best to keep the weird partition tables in the ISO
until it turned out that the particular machines, which shall use it,
are not in need of weirdness.
That would be the moment to consider use of make_isombr_part.


Have a nice day :)

Thomas