Re: Bugs in parted's pte handling?

[Jim Meyering]

Robert Herndon wrote:
> [Apologies for any protocol violations here.  I've just
> joined the group, and don't see this addressed in the
> recently archived mailing list issues.]
>
> There seem to be bugs in the handling of partition table
> entries in parted, relating to cases where
>
> gpt->NumberOfPartitionEntries * gpt->SizeOfPartitionEntry
>
> is not a multiple of the sector size.  This is true of all
> parted versions I've looked at (1.8.1, 1.8.8, and 2.3),
> although the particulars vary somewhat between versions.

Thank you for a fine bug report.
How do the offending GPT partition tables arise?  Do you know of a tool
that produces them?  By default?  I would imagine that such tables are
rather rare.

Of course, that's no reason to read uninitialized memory and fail for
a technically-valid GPT partition table.

I'll fix it one way or another, but will need a test case.  Do you
have code handy to generate an offending table?  Ideally it would use
only parted and say, dd, awk and similarly portable tools.  A C program
would work, too, but I'd probably have to get a copyright assignment or
disclaimer from you in order to be able to use it.

At worst, I will cook up something starting with a parted-created image,
and poking into it a few new values and manually-computed checksums
to match.