linux-kernel - Re: [PATCH v3 1/5] dma: imx-sdma: add clock ratio 1:1 check

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]

Date:   Wed, 23 Jan 2019 16:43:48 +0100
From:   Lucas Stach <l.stach@...gutronix.de>
To:     "Angus Ainslie (Purism)" <angus@...ea.ca>
Cc:     angus.ainslie@...i.sm, Vinod Koul <vkoul@...nel.org>,
        dmaengine@...r.kernel.org, NXP Linux Team <linux-imx@....com>,
        Pengutronix Kernel Team <kernel@...gutronix.de>,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        Daniel Baluta <daniel.baluta@...il.com>
Subject: Re: [PATCH v3 1/5] dma: imx-sdma: add clock ratio 1:1 check

Am Mittwoch, den 23.01.2019, 08:23 -0700 schrieb Angus Ainslie (Purism):
> On i.mx8 mscale B0 chip, AHB/SDMA clock ratio 2:1 can't be supportted,
> since SDMA clock ratio has to be increased to 250Mhz, AHB can't reach
> to 500Mhz, so use 1:1 instead.
> 
> > Based on NXP commit MLK-16841-1 by Robin Gong <yibin.gong@....com>
> 
> > Signed-off-by: Angus Ainslie (Purism) <angus@...ea.ca>
> ---
>  drivers/dma/imx-sdma.c | 23 +++++++++++++++++++----
>  1 file changed, 19 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> index 0b3a67ff8e82..531a9d8b032a 100644
> --- a/drivers/dma/imx-sdma.c
> +++ b/drivers/dma/imx-sdma.c
> @@ -440,6 +440,8 @@ struct sdma_engine {
> > >  	unsigned int			irq;
> > >  	dma_addr_t			bd0_phys;
> > >  	struct sdma_buffer_descriptor	*bd0;
> > +	/* clock ratio for AHB:SDMA core. 1:1 is 1, 2:1 is 0*/
> > > +	bool				clk_ratio;
>  };
>  
>  static int sdma_config_write(struct dma_chan *chan,
> @@ -662,8 +664,14 @@ static int sdma_run_channel0(struct sdma_engine *sdma)
> >  		dev_err(sdma->dev, "Timeout waiting for CH0 ready\n");
>  
> >  	/* Set bits of CONFIG register with dynamic context switching */
> > -	if (readl(sdma->regs + SDMA_H_CONFIG) == 0)
> > -		writel_relaxed(SDMA_H_CONFIG_CSM, sdma->regs + SDMA_H_CONFIG);
> +	if (readl(sdma->regs + SDMA_H_CONFIG) == 0) {

If the ACR bit gets set in sdma_init(), do we ever end up in this code
path? From a quick glance it seems we might wrongfully skip the CSM
enable here.

> +		if (sdma->clk_ratio)
> > +			reg = SDMA_H_CONFIG_CSM | SDMA_H_CONFIG_ACR;
> > +		else
> +			reg = SDMA_H_CONFIG_CSM;

That's a personal style preference, but I would write this as:

reg = SDMA_H_CONFIG_CSM;

if (sdma->clk_ratio);
	reg |= SDMA_H_CONFIG_ACR;

> +
> > +		writel_relaxed(reg, sdma->regs + SDMA_H_CONFIG);
> > +	}
>  
> >  	return ret;
>  }
> @@ -1840,6 +1848,11 @@ static int sdma_init(struct sdma_engine *sdma)
> >  	if (ret)
> >  		goto disable_clk_ipg;
>  
> > +	if (clk_get_rate(sdma->clk_ahb) == clk_get_rate(sdma->clk_ipg))
> > +		sdma->clk_ratio = 1;
> > +	else
> +		sdma->clk_ratio = 0;

sdma is zeroed at allocation, so the else path here isn't necessary.

> +
> >  	/* Be sure SDMA has not started yet */
> >  	writel_relaxed(0, sdma->regs + SDMA_H_C0PTR);
>  
> @@ -1880,8 +1893,10 @@ static int sdma_init(struct sdma_engine *sdma)
> >  	writel_relaxed(0x4050, sdma->regs + SDMA_CHN0ADDR);
>  
> >  	/* Set bits of CONFIG register but with static context switching */
> > -	/* FIXME: Check whether to set ACR bit depending on clock ratios */
> > -	writel_relaxed(0, sdma->regs + SDMA_H_CONFIG);
> > +	if (sdma->clk_ratio)
> > +		writel_relaxed(SDMA_H_CONFIG_ACR, sdma->regs + SDMA_H_CONFIG);
> > +	else
> > +		writel_relaxed(0, sdma->regs + SDMA_H_CONFIG);
>  
> >  	writel_relaxed(ccb_phys, sdma->regs + SDMA_H_C0PTR);
>