bug-grub
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: a patch for working with HDD images and for newer Intel cards


From: Yoshinori K. Okuji
Subject: Re: a patch for working with HDD images and for newer Intel cards
Date: Wed, 14 May 2003 17:54:27 +0900
User-agent: Wanderlust/2.8.1 (Something) SEMI/1.14.3 (Ushinoya) FLIM/1.14.3 (Unebigoryƍmae) APEL/10.3 Emacs/21.2 (i386-debian-linux-gnu) MULE/5.0 (SAKAKI)

At Tue, 06 May 2003 11:08:45 +0100,
address@hidden wrote:
> I don't agree with the patch as it treats files (disk images)
> and disks differently which is confusing at best.
> The exact same logic holds true for disks as files.
> I.E. if you do try to open /dev/hda1 or /dev/loop1 for e.g.
> in write_to_partition and it's already mounted then
> disk corruption may occur.

However, a disk image is very different from a real disk indeed,
because Linux maintains the information about partitions for the
latter, but not for the former. Thus, for an image file, Linux forgets
everything about it once you have Linux to unmount it. For real disks,
you cannot do that.

> Note also in principle I don't agree with the last documentation
> point above. I.E. in what situations would you need a filesystem
> mounted rw and also be sure the cache is coherent with the disk
> and not being changed while grub is running. IMHO grub should
> never open a mounted (rw) partition (on linux at least).
> It would be nice if the linux kernel gave an error when you tried
> to open(/dev/hda1,O_RDWR), but grub should probably work around
> this by getting a list of mounted (rw) partitions that it can't open.

If we do that, the user couldn't install GRUB into a running system
any longer.

Okuji




reply via email to

[Prev in Thread] Current Thread [Next in Thread]