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

[Vladimir 'φ-coder/phcoder' Serbinenko]

On 22.04.2012 20:24, Samuel Thibault wrote:
> Hello,
>
> GNU/Hurd is moving to userland partition support, i.e. /dev/hd0s* slices
> are actually served by userland translators which use libparted.  The
> grub probe code needs to be fixed for that, here is a proposed patch.
>
> The principle is that it gets the store translator behind the file being
> probed for, and compares the storage area there with what grub detects
> itself.  That is why it makes use of the map array, which I preferred
> to rename to hostdisk_map (thus big but trivial changes in hostdisk.c):
> even if it was already an external symbol, it's preferrable to prefix
> it.
map was never intended to be globally visible. At most perhaps
"semi-local" due to difficulties of split between getroot.c and hostdisk.c
Moreover the behaviour of find_root_device must be independent of what
is already in map since we don't rely anymore at map being populated at
all and at entries only on as-needed basis.
Additional problems include:

+      if (disk->dev->id != GRUB_DISK_DEVICE_BIOSDISK_ID && disk->dev->id != 
GRUB_DISK_DEVICE_HOSTDISK_ID)
+       {
No need to check for BIOSDISK. It's not used for hostdisk.
+      store_offset_t start = partition->start * GRUB_DISK_SECTOR_SIZE,
+                    length = partition->len * GRUB_DISK_SECTOR_SIZE,
This is both stylistically incorrect and wrong.
Stylistically at declaration we repeat type on every line.
Additionally partition->start is only relative to container, not whole disk. 
You have to use grub_disk_partition_get_start for this.


>
> Samuel
>
>
> _______________________________________________
> Grub-devel mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/grub-devel


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

signature.asc
Description: OpenPGP digital signature