Re: [PATCH 1/3] mtd: spi-nor: add optional DMA-safe bounce buffer for data transfer

From: Boris Brezillon
Date: Sun Jan 07 2018 - 15:07:48 EST


On Tue, 26 Dec 2017 14:59:00 +0100
Cyrille Pitchen <cyrille.pitchen@xxxxxxxxxx> wrote:

> Hi Vignesh
>
> Le 26/12/2017 Ã 14:42, Vignesh R a ÃcritÂ:
> > Hi Cyrille,
> >
> > Thanks for doing this series! One comment below.
> >
> > On 24-Dec-17 10:06 AM, Cyrille Pitchen wrote:
> > [...]
> >> diff --git a/drivers/mtd/spi-nor/spi-nor.c b/drivers/mtd/spi-nor/spi-nor.c
> >> index 8bafd462f0ae..59f9fbd45234 100644
> >> --- a/drivers/mtd/spi-nor/spi-nor.c
> >> +++ b/drivers/mtd/spi-nor/spi-nor.c
> >> @@ -14,8 +14,10 @@
> >> #include <linux/errno.h>
> >> #include <linux/module.h>
> >> #include <linux/device.h>
> >> +#include <linux/highmem.h>
> >> #include <linux/mutex.h>
> >> #include <linux/math64.h>
> >> +#include <linux/mm.h>
> >> #include <linux/sizes.h>
> >> #include <linux/slab.h>
> >>
> >> @@ -1232,6 +1234,56 @@ static const struct flash_info spi_nor_ids[] = {
> >> { },
> >> };
> >>
> >> +static bool spi_nor_is_dma_safe(const void *buf)
> >> +{
> >> + if (is_vmalloc_addr(buf))
> >> + return false;
> >> +
> >> +#ifdef CONFIG_HIGHMEM
> >> + if ((unsigned long)buf >= PKMAP_BASE &&
> >> + (unsigned long)buf < (PKMAP_BASE + (LAST_PKMAP * PAGE_SIZE)))
> >> + return false;
> >> +#endif
> >> +
> >> + return true;
> >> +}
> >> +
> >
> > Better way would be to use virt_addr_valid():
> > static bool spi_nor_is_dma_safe(const void *buf)
> > {
> > return virt_addr_valid(buf);
> > }
> >
> > Regards
> > Vignesh
> >
>
> Thanks for the advice :)
>
> https://patchwork.kernel.org/patch/9768341/
> Maybe I could check both virt_addr_valid() and object_is_on_stack() too ?

Yep, see the explanation given here [1].

[1]http://elixir.free-electrons.com/linux/v4.15-rc6/source/Documentation/DMA-API-HOWTO.txt#L132