Re: [PATCH v2 20/28] ARCv2: barriers

From: Will Deacon
Date: Tue Jun 23 2015 - 05:26:40 EST


On Tue, Jun 23, 2015 at 10:03:25AM +0100, Vineet Gupta wrote:
> On Tuesday 23 June 2015 02:19 PM, Will Deacon wrote:
> >> +/*
> >> > + * MMIO can also get buffered/optimized in micro-arch, so barriers needed
> >> > + * Based on ARM model for the typical use case
> >> > + *
> >> > + * <writel_relaxed DMA buffer>
> >> > + * <writel MMIO "go" reg>
> >> > + * or:
> >> > + * <readl MMIO "status" reg>
> >> > + * <readl_relaxed DMA buffer>
> > The writel_relaxed/readl_relaxed parts here would actually just be
> > bog-standard loads and stores to an in-memory buffer. I was trying too hard
> > to show the barrier semantics and accidentally turned the DMA buffers into
> > __iomem regions.
>
> Not sure if I follow you completely :-)

D'oh, sorry.

> IMHO, It doesn't matter if we are dealing with a typical DMA buffer (cached) or a
> buffer descriptor (typically uncached unless there's hardware IO-coh or some
> such). Both the cases assume a vanilla ld/st to buffer (using relaxed API) with a
> surrounding MMIO access.

It's more that you should only pass __iomem pointers (i.e. stuff you got
back from something like ioremap) to readl_relaxed/writel_relaxed and that's
not typically how you would allocate your DMA buffer.

> > If you fix the comment:
>
> Does this look better ?
>
> - * <writel_relaxed DMA buffer>
> + * <writel_relaxed DMA buffer (cached or uncached)>

I'd just replace 'writel_relaxed' with whatever your store instruction is
(ST)?

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