Messages in this thread | | | Date | Tue, 11 Jul 2000 00:42:14 +0200 (CEST) | From | Andrea Arcangeli <> | Subject | Re: linux-2.4.0 breaks grub install into partition |
| |
On Mon, 10 Jul 2000, Daniel Kobras wrote:
>Do we run into the same aliasing issues if the fs doesn't reside on >/dev/hda1 itself but on a block dev on top of it - like /dev/md0 or a
Right, you run in exactly the same aliasing issue.
Also on top or LVM or MD you can't trust bmap because it will return the number of the logical block, but of course at boot you need the number of the physical block and the name of the physical device.
>logical volume -, or does the current raid/LVM code already take care of >those cases?
You don't have to deal with LVM/MD but with the fs. The fs doesn't know the underlying blockdevice is logical (LVM/MD are a trasparent) so it won't take care of those cases.
If you want to write to the bootsector of a partition that is part of a VG and that is allocated in a logical volume with a live filesystem on top of on it, you have to know the first block of the logical volume corresponds to the first block of the physical partiton and then you can write to the first 512 bytes of the logical volume knowing it will be written to the first 512 bytes of the physical partition and you won't run into aliasing issues for the same reason of the non-LVM case. You can get this information using pvdata -a /dev/hda1 for LVM. (for MD you probably need to look at /proc/mdstat)
Andrea
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |