linux-kernel - Re: GPL only modules [was Re: [GIT PATCH] more Driver core patches for 2.6.19]

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]

Message-Id: <200612141231.17331.hjk@linutronix.de>
Date:	Thu, 14 Dec 2006 12:31:16 +0100
From:	Hans-Jürgen Koch <hjk@...utronix.de>
To:	Alan <alan@...rguk.ukuu.org.uk>
Cc:	"Hua Zhong" <hzhong@...il.com>,
	"'Martin J. Bligh'" <mbligh@...igh.org>,
	"'Linus Torvalds'" <torvalds@...l.org>,
	"'Greg KH'" <gregkh@...e.de>, "'Jonathan Corbet'" <corbet@....net>,
	"'Andrew Morton'" <akpm@...l.org>,
	"'Michael K. Edwards'" <medwards.linux@...il.com>,
	linux-kernel@...r.kernel.org
Subject: Re: GPL only modules [was Re: [GIT PATCH] more Driver core patches for 2.6.19]

Am Donnerstag, 14. Dezember 2006 12:14 schrieb Alan:
> On Wed, 13 Dec 2006 22:01:15 -0800
> "Hua Zhong" <hzhong@...il.com> wrote:
> 
> > > I think allowing binary hardware drivers in userspace hurts 
> > > our ability to leverage companies to release hardware specs. 
> > 
> > If filesystems can be in user space, why can't drivers be in user space? On what *technical* ground?
> 
> The FUSE file system interface provides a clean disciplined interface
> which allows an fs to live in user space. The uio layer (if its ever
> fixed and cleaned up) provides some basic hooks that allow a user space
> program to arbitarily control hardware and make a nasty undebuggable mess.

You think it's easier for a manufacturer of industrial IO cards to
debug a (large) kernel module?

> 
> uio also doesn't handle hotplug, pci and other "small" matters.

uio is supposed to be a very thin layer. Hotplug and PCI are already
handled by other subsystems. 

> 
> Now if you wanted to make uio useful at minimum you would need
> 

The majority of industrial IO cards have registers and/or dual port RAM
that can be mapped to user space (even today). We want to add a simple
way to handle interrupts for such cards. That's all.
The fact that there might be some sort of hardware/interrupts/situations
where this is not possible or not so simple isn't that important at the
moment. We can extend the UIO system if somebody actually requires these
extensions.

Hans

-
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/