[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20141105094146.GB27686@lukather>
Date: Wed, 5 Nov 2014 10:41:46 +0100
From: Maxime Ripard <maxime.ripard@...e-electrons.com>
To: Russell King - ARM Linux <linux@....linux.org.uk>
Cc: Chen-Yu Tsai <wens@...e.org>,
Kishon Vijay Abraham I <kishon@...com>,
Mike Turquette <mturquette@...aro.org>,
Grant Likely <grant.likely@...aro.org>,
Rob Herring <robh+dt@...nel.org>,
Hans de Goede <hdegoede@...hat.com>,
linux-sunxi@...glegroups.com, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH 1/6] clk: sunxi: Add support for sun9i a80 usb clocks and
resets
Hi Russell,
On Tue, Nov 04, 2014 at 06:12:19PM +0000, Russell King - ARM Linux wrote:
> On Tue, Nov 04, 2014 at 12:07:14PM +0800, Chen-Yu Tsai wrote:
> > + spin_lock_irqsave(data->lock, flags);
> > +
> > + reg = readl(data->reg);
> > + writel(reg & ~BIT(id), data->reg);
> > +
> > + spin_unlock_irqrestore(data->lock, flags);
>
> Don't we have generic support for atomic modification of register
> values? Hmm, we have it for ARM only - atomic_io_modify() and
> atomic_io_modify_relaxed().
>
> I guess we should push for those to become cross-arch if we end up
> with generic drivers shared between other architectures.
IIRC, the atomic MMIO accessors are doing exactly that, but with a
global lock for all MMIO accesses, while here we protect a single
register.
I'm not really sure that sharing this spinlock across the whole system
is worth it and scales that well.
Maxime
--
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)
Powered by blists - more mailing lists