Re: newly compiled kernel no .img file

Matt Bernstein (matt@theBachChoir.org.uk)
Wed, 5 Dec 2001 16:56:43 +0000 (GMT)


At 11:40 -0500 Brian Gerst wrote:

>> If you notice the first declaration of image the >
>> "initrd=/boot/initrd-2.4.7-10.img" is not present . Of course I removed
>> it so that there would be no kernel panic and I am able to boot into the
>> new kernel (xunil). > What I want to know is what is this .img file why
>> is it required in the original kernel compilation and not in the newer .
>
>Your distribution put that there so that it can use modules for drivers
>that are required to mount the root filesystem (ie. SCSI, fs driver,
>etc.). If you build your own kernel, those drivers should be built
>non-modular, therefore you won't need an initrd.

Yes for most users.

I'd like to build a quite generic 2.4 tree with everything as a module
where possible. I am forced to compile in binfmt_elf, initrd and romfs.

This allows me to use one tree for several different machines (some might
have an ext3 / on an IDE HDD; others maybe reiserfs on a gdth controller
etc..) without a huge amount of dead code in the kernel. So.. when a
subset of them are buggy we can see what modules they have in common..

My question: is this mega-module setup likely to be less stable than a
monolith? I'm not fussed about a % or two performance loss.

Matt

-
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/