From owner-freebsd-hackers  Wed Nov 29 13:11:21 1995
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id NAA20725
          for hackers-outgoing; Wed, 29 Nov 1995 13:11:21 -0800
Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id NAA20715
          for <hackers@freebsd.org>; Wed, 29 Nov 1995 13:11:13 -0800
Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id OAA28617; Wed, 29 Nov 1995 14:05:44 -0700
From: Terry Lambert <terry@lambert.org>
Message-Id: <199511292105.OAA28617@phaeton.artisoft.com>
Subject: Re: uio question...
To: msmith@atrad.adelaide.edu.au (Michael Smith)
Date: Wed, 29 Nov 1995 14:05:44 -0700 (MST)
Cc: hackers@freebsd.org
In-Reply-To: <199511290134.BAA09808@genesis.atrad.adelaide.edu.au> from "Michael Smith" at Nov 29, 95 01:34:51 am
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1203      
Sender: owner-hackers@freebsd.org
Precedence: bulk

> Okay, my entry for the lame-device-driver-question of the month 8)
> 
> Given a device read function passed a uio struct, what's the correct way
> to determine the total size of the read request?  
> 
> I have two possibles here : walk the array of iovec structures, or
> contemplate the uio_resid field, which appears to be used by a number of 
> other drivers...
> 
> Experience?

The "resid" is used to determine residual read remaining.  Typically,
a uio struct will have an iovec count of one, or the alternate iovec's
will be non-contiguous.

In either case , for a non-one iovec count in the uio, you *must* walk
the iovec list to return any information.

The question is whether your driver does scatter/gather itself, or whether
it will be handling I/O one iovec at a time.

You should look at the tty, generic (controller independent) scsi code,
one of the SCSI controller dependendt code pieces (the 1540/1740?  I
recommend against the NCR/AHA2xxx code because of the additional complexity)
and the readv/writev system call code in /usr/src/sys/kern.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.