linux-kernel - Re: [PATCH 2/2] nbd: add support for nbd as root device

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 13 Jun 2019 12:54:46 -0400
From:   Josef Bacik <josef@...icpanda.com>
To:     Roman Stratiienko <roman.stratiienko@...ballogic.com>
Cc:     Josef Bacik <josef@...icpanda.com>, linux-kernel@...r.kernel.org,
        nbd@...er.debian.org,
        Aleksandr Bulyshchenko <A.Bulyshchenko@...ballogic.com>,
        linux-block@...r.kernel.org, axboe@...nel.dkn.org
Subject: Re: [PATCH 2/2] nbd: add support for nbd as root device

On Thu, Jun 13, 2019 at 07:21:43PM +0300, Roman Stratiienko wrote:
> > I don't doubt you have a good reason to want it, I'm just not clear on why an
> > initramfs isn't an option?  You have this special kernel with your special
> > option, and you manage to get these things to boot your special kernel right?
> > So why is a initramfs with a tiny nbd-client binary in there not an option?
> >
> > Also I mean that there are a bunch of different nbd servers out there.  We have
> > our own here at Facebook, qemu has one, IIRC there's a ceph one.  They all have
> > their own connection protocols.  The beauty of NBD is that it doesn't have to
> > know about that part, it just does the block device part, and I'd really rather
> > leave it that way.  Thanks,
> >
> > Josef
> 
> 
> The only reason I prefer embed client into the kernel is to save
> valuable engineering time creating and supporting custom initramfs,
> that is not so easy especially on Android-based systems.
> 

I'm unconvinced that creating an initramfs is any harder than building your own
kernel with this patch and providing the configuration information to the
device, especially for android where we inside of Facebook provision android
devices with a custom initramfs all the time, and have done so for many, many
years.

> Taking into account that if using NBD and creating custom initramfs
> required only for automated testing, many companies would choose
> manual deployment instead, that is bad for the human progress.
> 
> I believe that all users of non-standard NBD handshake protocols could
> continue to use custom nbd-clients.

There is no "standard NBD handshake protocol" is my point.  That part exists
outside of the NBD spec that the kernel is concerned with.  The client is happy
to do whatever it pleases.  Now there's no doubt that the standard nbd client
and server that is shipped is the reference spec, but my point is that it is
still outside the kernel and so able to change as it sees fit.

> 
> Either you accept this patch or not I would like to pass review from
> maintainers and other persons that was involved in NBD development,
> thus making a step closer to get this mainlined in some future.

...I am the maintainer, but feel free to try to get Jens to take a patch that I
think is unnecessary.  Thanks,

Josef