Re: Minimum size of GPT partition table header

[Jim Meyering]

rahul dev wrote:
> Guys,
>
>    There seems to be a bug in function _header_is_valid() in file gpt.c.
>
> static int
> _header_is_valid (PedDisk const *disk, GuidPartitionTableHeader_t *gpt,
>                   PedSector my_lba)
> {
>   uint32_t crc, origcrc;
>   PedDevice const *dev = disk->dev;
>
>   if (PED_LE64_TO_CPU (gpt->Signature) != GPT_HEADER_SIGNATURE)
>     return 0;
>   /*
>    * "While the GUID Partition Table Header's size may increase
>    * in the future it cannot span more than one block on the
>    * device."  EFI Specification, version 1.10, 11.2.2.1
>    */
>   if (PED_LE32_TO_CPU (gpt->HeaderSize) < pth_get_size_static (dev)
>       || PED_LE32_TO_CPU (gpt->HeaderSize) > dev->sector_size)
>     return 0;
>
>
> The check gpt->HeaderSize < pth_get_size_static seems to be incorrect.
> I think minimum size of gpt header is 92 bytes. So, correct check should be
> gtp->HeaderSize < 92.
>
> As mentioned in the EFI spec that the header size may increase in future, 
> pth_get_size_static() will also return the increased size. In that case, if 
> we are running parted on a disk partitioned by old versions, the above check 
> would fail. For older versions, gpt->HeaderSize is 92, whereas 
> pth_get_size_static() will be > 92 for newer versions (where gpt hdr size has 
> been increased).

Thanks for reviewing the code.
However, currently, pth_get_size_static() always returns 92, and I do
not expect it (or the official header size) to change any time soon.

Has this code caused an actual problem for you?