linux-kernel - Re: [PATCH] scsi: allow persistent reservations without CAP_SYS

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:	Tue, 12 Jun 2012 19:08:53 +0200
From:	Paolo Bonzini <pbonzini@...hat.com>
To:	Alan Cox <alan@...rguk.ukuu.org.uk>
CC:	linux-kernel@...r.kernel.org, axboe@...nel.dk,
	linux-scsi@...r.kernel.org, jbottomley@...allels.com
Subject: Re: [PATCH] scsi: allow persistent reservations without CAP_SYS_RAWIO

Il 12/06/2012 18:55, Alan Cox ha scritto:
> On Tue, 12 Jun 2012 18:08:32 +0200
> Paolo Bonzini <pbonzini@...hat.com> wrote:
> 
>> Persistent reservations commands cannot be issued right now without
>> giving CAP_SYS_RAWIO to the process who wishes to send them.  This
>> is a bit heavy-handed, allow these two commands.
>>
>> Signed-off-by: Paolo Bonzini <pbonzini@...hat.com>
>> ---
>>         Ok for 3.5 as well?
> 
> NAK.
> 
> Persistent reservations are exactly the kind of command that should have
> a security model attached to them.

There is.  It's called "chmod"; you don't give write access to LUNs to
random users.  and SCM_RIGHTS is what lets you override it securely.

> Red Hat seems to be an ever growing source of "mummy its hard, lets
> disable all the security" type fixes. Please stop it.

Last time you were complaining that I was turning things *off* (SG_IO to
partitions for root).  Now you complain that I'm turning things *on*.
It's difficult to say they are the same thing.  Though perhaps you were
talking about someone else.

> There is a sensible debate to be had about whether a lesser privilege
> ought to be allowed. The real fix to this as with half of the other
> crazy attempts to break all the security models that seem to keep spewing
> forth is for someone who cares about it (that seems to me Red Hat) add
> support for pushing a BPF filter onto a block device command queue. 

Sure; however, doing so requires access to some member of "struct file"
from SG_IO.  Thus, ioctl would need to take a "struct file" rather than
just an fmode_t.

The switch to fmode_t was done in 2007 by Al Viro.  I would like to
understand the reasons for the switch; it seems to me that it was part
of the big kernel lock removal.  If it's acceptable to undo it, I would
very much would like to add generic BPF filtering to SG_IO.

Paolo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/