On Thu, Mar 07, 2002 at 08:47:24AM -0600, Steve Lord wrote:
> On Thu, 2002-03-07 at 06:06, Etienne Lorrain wrote:
> > > With "allocate on flush", (aka delayed allocation), file data is
> > > assigned a disk mapping when the data is being written out, rather than
> > > at write(2) time. This has the following advantages:
> >
> > I do agree that this is a better solution than current one,
> > but (even if I did not had time to test the patch), I have
> > a question: How about bootloaders?
> >
> > IHMO all current bootloaders need to write to disk a "chain" of sector
> > to load for their own initialisation, i.e. loading the remainning
> > part of code stored on a file in one filesystem from the 512 bytes
> > bootcode. This "chain" of sector can only be known once the file
> > has been allocated to disk - and it has to be written on the same file,
> > at its allocated space.
> >
> > So can you upgrade LILO or GRUB with your patch installed?
> > It is not a so big problem (the solution being to install the
> > bootloader on an unmounted filesystem with tools like e2fsprogs),
> > but it seems incompatible with the current executables.
>
> The interface used by lilo to read the kernel location needs to flush
> data out to disk before returning results. It's not too hard to do.
>
Also, GRUB shouldn't have this problem since it reads the filesystems
directly on boot.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 07 2002 - 21:01:05 EST