Re: [PATCH,HURD] Fix grub-probe with userland partition support

[Vladimir 'φ-coder/phcoder' Serbinenko]

On 24.04.2012 11:19, Samuel Thibault wrote:
> Because path_start is the start of the file, not of the partition
> containing the file.
>
>> > Some partitions may contain more partitions, like logical container
>> > contains logical partitions.
> Yes, but we do not have defined /dev namings for these yet, so we don't
> know how to look for them grom GRUB yet.
That 2 comments together open the can of worms. Second one would imply
we should take the smallest partition containing the given file. E.g.
One could have
hd0s1 containing hd0s5 and hd0s6. In this case we want hd0s5 and not
hd0s1. On the other hand a file may reside in partition but be part of a
filesystem spanning through the whole disk. This is common for all kinds
of hybrid CDROMs, including the ones created by grub-mkrescue. They have
iso9660 spanning through the whole disk but all of the disk other than
the first sector is in some kind of partition table to avoid it being
accidentally overwritten. So even though the file itself is inside a
partition, we want the whole disk.
Also if file is 0-sized, sparse, compressed or encrypted it won't have
blocklist and this approach would fail, possibly even with out-of-range
array access.
I'm surprised that Hurd doesn't offer a way to just ask "What does this
filesystem translator consume?"

-- 
Regards
Vladimir 'φ-coder/phcoder' Serbinenko

signature.asc
Description: OpenPGP digital signature