linux-kernel - Re: [PATCH 2/2] ASoC: add driver for Rockchip RK3xxx I2S controller

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20140702102310.GC410@sirena.org.uk>
Date:	Wed, 2 Jul 2014 11:23:10 +0100
From:	Mark Brown <broonie@...nel.org>
To:	Huang Tao <huangtao@...k-chips.com>
Cc:	jianqun <xjq@...k-chips.com>, heiko@...ech.de, lgirdwood@...il.com,
	perex@...ex.cz, tiwai@...e.de, grant.likely@...aro.org,
	robh+dt@...nel.org, linux-kernel@...r.kernel.org,
	alsa-devel@...a-project.org, devicetree@...r.kernel.org,
	zhangqing@...k-chips.com, hj@...k-chips.com,
	kever.yang@...k-chips.com, zyw@...k-chips.com, yzq@...k-chips.com,
	zhenfu.fang@...k-chips.com, cf@...k-chips.com, kfx@...k-chips.com
Subject: Re: [PATCH 2/2] ASoC: add driver for Rockchip RK3xxx I2S controller

On Wed, Jul 02, 2014 at 09:01:49AM +0800, Huang Tao wrote:
> 于 2014年07月02日 01:07, Mark Brown 写道:
> >> +static inline void i2s_writel(struct rk_i2s_dev *i2s, u32 value,
> >> > +			      unsigned int offset)
> >> > +{
> >> > +	writel_relaxed(value, i2s->regs + offset);
> >> > +}
> >> > +
> >> > +static inline u32 i2s_readl(struct rk_i2s_dev *i2s, unsigned int offset)
> >> > +{
> >> > +	return readl_relaxed(i2s->regs + offset);
> >> > +}
> > Perhaps use regmap?  The main advantage would be the debug
> > infrastructure, though you could also use _update_bits() to reduce the
> > amount of time spent locked.

> Are you sure? This is a I2S driver, we can write the register directly,
> do not through I2C or SPI bus.
> Write a register is only a few instructions on ARM, but write through
> regmap, it may take a long path.
> I think it will just consume CPU power and make the whole thing more
> complex.
> Could you tell me what benefits we can get if use regmap? Or something I
> just missing?

The main thing is the diagnostic infrastructure, plus the cache if you
find you need that for suspend and resume (which you didn't implement
yet) - that's the main reason things use regmap for memory mapped
registers.  There is some overhead but it's not *that* big.

Download attachment "signature.asc" of type "application/pgp-signature" (820 bytes)