Date: Fri, 4 Mar 2011 13:29:44 -0800
From: Greg KH <greg@kroah.com>
To: Andy Lutomirski <luto@MIT.EDU>
Cc: Matthew Wilcox <willy@linux.intel.com>, linux-kernel@vger.kernel.org
Subject: Re: [REVIEW] NVM Express driver
Message-ID: <20110304212944.GC28680@kroah.com>
References: <20110303204749.GY3663@linux.intel.com>
 <20110303211336.GA32645@kroah.com>
 <20110303214104.GZ3663@linux.intel.com>
 <20110303215155.GA30451@kroah.com>
 <20110303220735.GA3663@linux.intel.com>
 <20110303222226.GA30966@kroah.com>
 <4D704DAB.9080100@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4D704DAB.9080100@mit.edu>
User-Agent: Mutt/1.5.20 (2009-06-14)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 3130
Lines: 68

On Thu, Mar 03, 2011 at 09:25:47PM -0500, Andy Lutomirski wrote:
> On 03/03/2011 05:22 PM, Greg KH wrote:
> >On Thu, Mar 03, 2011 at 05:07:35PM -0500, Matthew Wilcox wrote:
> >>On Thu, Mar 03, 2011 at 01:51:55PM -0800, Greg KH wrote:
> >>>Heh, no, well, submit_io should just go through the block layer and not
> >>>be a separate ioctl, right?
> >>
> >>Just like with SG_IO, there are reasons to do I/Os without going through
> >>the block layer.
> >
> >Ok, that makes sense.
> >
> >>>>There's a bit of an impedence mismatch there.  Think of this as
> >>>>being drive firmware instead of controller firmware.  This isn't for
> >>>>request_firmware() kind of uses, it's for some admin tool to come along
> >>>>and tell the drive "Oh, here's some new firmware for you".
> >>>
> >>>That's fine, request_firmware will work wonderfully for that.
> >>
> >>How would the driver know that it should call request_firmware()?
> >>Do it every 60 seconds in case somebody's downloaded some new firmware?
> >
> >Ick, no, just use the function provided that lets you create a firmware
> >request and be notified when it is written to,
> >request_firmware_nowait().  That is what it is there for.
> >
> >>>>If you look at the spec [1], you'll see there are a number of firmware
> >>>>slots in the device, and it's up to the managability utility to decide
> >>>>which one to replace or activate.  I dno't think you want to pull all
> >>>>that gnarly decision making code into the kernel, do you?
> >>>>
> >>>>[1] http://download.intel.com/standards/nvmhci/NVM_Express_1_0_Gold.pdf
> >>>
> >>>No, just export multiple "slots" as firmware devices ready to be filled
> >>>in by userspace whenever it wants/needs to.  The management utility can
> >>>just dump the firmware to those sysfs files when it determines it needs
> >>>to update the firmware, no decision making in the kernel at all.
> >>
> >>OK ... glad we decided to limit the number of slots.  I still don't see
> >>(in Documentation/firmware_class/README) how this works for user-initiated
> >>firmware updates rather than kernel-initiated.
> >
> >I didn't even realize we had a firmware README file...
> >
> >Anyway, just use request_firmware_nowait(), you will be fine.
> >
> 
> Unless I'm misunderstanding the spec, this is for *nonvolatile*
> firmware updates.  So if I have two of these devices and I stick a
> firmware file into /lib/firmware, should it update both devices?
> 
> Shouldn't reflashing the device firmware be something that the users
> asks for by something a little more specific than just creating a
> file?  (In any case, creating /lib/firmware/nvmhci-3 to flash slot 3
> seems rather silly.)

Yes, a userspace tool should handle all of this, much like Dell does
this with BIOS updates through this same interface.  See their in-kernel
code, and userspace tools for details on how to do this sanely.

thanks,

greg k-h
--
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/