Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 19 Apr 2016 16:26:29 +0100
From:      Ruslan Bukin <ruslan.bukin@cl.cam.ac.uk>
To:        Ian Lepore <ian@freebsd.org>
Cc:        Ruslan Bukin <br@FreeBSD.org>, src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org
Subject:   Re: svn commit: r298268 - head/sys/dev/spibus
Message-ID:  <20160419152629.GA28888@bsdpad.com>
In-Reply-To: <1461076899.1232.23.camel@freebsd.org>
References:  <201604191418.u3JEICrY023254@repo.freebsd.org> <1461076899.1232.23.camel@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Apr 19, 2016 at 08:41:39AM -0600, Ian Lepore wrote:
> On Tue, 2016-04-19 at 14:18 +0000, Ruslan Bukin wrote:
> > Author: br
> > Date: Tue Apr 19 14:18:12 2016
> > New Revision: 298268
> > URL: https://svnweb.freebsd.org/changeset/base/298268
> > 
> > Log:
> >   Add optional chip_select/deselect methods. This is required
> >   when we want to keep CS asserted for multiple transfers.
> > 
> 
> This is missing a now-required mechanism to arbitrate ownership of the
> bus.  If a leaf driver can take over the bus for longer than a single
> transfer, then it needs the same kind of logic as the iicbus uses
> (well, conceptually the same, it doesn't need the same horrible
> implementation).
> 
> I also don't see any default implementation of the new methods, or a
> commit that adds the new functionality to all existing SPI controller
> drivers.
> 

Thanks, Ian.
I just realized we don't require to hold CS for multiple transfers
in our hardware. So I reverted this.

Ruslan




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160419152629.GA28888>