linux-kernel - Re: [PATCH V2 7/7] dmaengine/dw_dmac: implement pause and resume in dwc

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: <20110419085514.GG2820@pulham.picochip.com>
Date:	Tue, 19 Apr 2011 09:55:15 +0100
From:	Jamie Iles <jamie@...ieiles.com>
To:	Viresh Kumar <viresh.kumar@...com>
Cc:	linux-kernel@...r.kernel.org, vinod.koul@...el.com,
	dan.j.williams@...el.com, linux-arm-kernel@...ts.infradead.org,
	armando.visconti@...com, shiraz.hashim@...com, amit.goel@...com,
	viresh.linux@...il.com, linus.walleij@...aro.org
Subject: Re: [PATCH V2 7/7] dmaengine/dw_dmac: implement pause and resume in
 dwc_control

On Tue, Apr 19, 2011 at 02:02:12PM +0530, Viresh Kumar wrote:
> From: Linus Walleij <linus.walleij@...aro.org>
> 
> Some peripherals like amba-pl011 needs pause to be implemented in DMA controller
> drivers. This also returns correct status from dwc_tx_status() in case chan is
> paused.
> 
> Signed-off-by: Linus Walleij <linus.walleij@...aro.org>
> Signed-off-by: Viresh Kumar <viresh.kumar@...com>
> ---
>  drivers/dma/dw_dmac.c      |   25 +++++++++++++++++++++++--
>  drivers/dma/dw_dmac_regs.h |    1 +
>  2 files changed, 24 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
> index c30f0ba..48d2d7e 100644
> --- a/drivers/dma/dw_dmac.c
> +++ b/drivers/dma/dw_dmac.c
> @@ -832,8 +832,7 @@ static int dwc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
>  	struct dw_desc		*desc, *_desc;
>  	LIST_HEAD(list);
>  
> -	/* Only supports DMA_TERMINATE_ALL */
> -	if (cmd != DMA_TERMINATE_ALL)
> +	if (cmd != DMA_TERMINATE_ALL && cmd != DMA_PAUSE && cmd != DMA_RESUME)
>  		return -ENXIO;
>  
>  	/*
> @@ -844,11 +843,30 @@ static int dwc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
>  	 */
>  	spin_lock_irqsave(&dwc->lock, dwc->lflags);
>  
> +	if (cmd == DMA_RESUME) {
> +		if (dwc->paused) {
> +			channel_set_bit(dw, CH_EN, dwc->mask);
> +			while (!(dma_readl(dw, CH_EN) & dwc->mask))
> +				cpu_relax();
> +		}
> +		spin_unlock_irqrestore(&dwc->lock, dwc->lflags);
> +		return 0;
> +	}
> +
>  	channel_clear_bit(dw, CH_EN, dwc->mask);
>  
>  	while (dma_readl(dw, CH_EN) & dwc->mask)
>  		cpu_relax();
>  
> +	if (cmd == DMA_PAUSE) {
> +		dwc->paused = true;
> +		spin_unlock_irqrestore(&dwc->lock, dwc->lflags);
> +		return 0;
> +	}

>From the Synopsys datasheet it looks to me like pausing by clearing the 
channel enable bit could cause data to be lost.  I think you need to set 
the CH_SUSP bit and wait for the FIFO_EMPTY flag to go high then disable 
the channel.

Jamie
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/