From mboxrd@z Thu Jan  1 00:00:00 1970
From: scott.bauer@intel.com (Scott Bauer)
Date: Fri, 9 Dec 2016 11:50:51 -0700
Subject: [PATCH v2 2/4] block: Add Sed-opal library
In-Reply-To: <20161209183034.GA28669@infradead.org>
References: <1480456322-27339-1-git-send-email-scott.bauer@intel.com>
 <1480456322-27339-3-git-send-email-scott.bauer@intel.com>
 <20161201005006.GE21081@localhost.localdomain>
 <20161201100456.GA17592@infradead.org>
 <20161201175343.GA6554@sbauer-Z170X-UD5>
 <20161201182239.GH21081@localhost.localdomain>
 <20161209174529.GA7865@sbauer-Z170X-UD5>
 <20161209183034.GA28669@infradead.org>
Message-ID: <20161209185051.GB7865@sbauer-Z170X-UD5>

On Fri, Dec 09, 2016@10:30:34AM -0800, Christoph Hellwig wrote:
> On Fri, Dec 09, 2016@10:45:30AM -0700, Scott Bauer wrote:
> > Now, with this in mind, it sort of makes sense to move this from
> > block/ back into lib/ and interface with the character dev. Instead
> > of passing around block_devices, we can pass around struct file *'s.
> >
> 
> Even the character device is always backed by the queues, and I'd really
> not prefer to have a struct file here - that's not useful at all
> for in-kernel users.

Is your main concern that if an in-kernel user wants to use this functionality
they have to pass us a block device pointer, and may not be able to? Is that
why you wanted to pass a function pointer to the code because a driver/kernel-user
can probably always have that availiable to send to us?