Re: [PATCH 15/26] asm-generic: ioremap_uc should behave the same with and without MMU

[Paul Walmsley <paul.walmsley@sifive.com>]

On Sat, 17 Aug 2019, Christoph Hellwig wrote:

> Whatever reason there is for the existence of ioremap_uc, and the fact 
> that it returns NULL by default on architectures with an MMU applies 
> equally to nommu architectures, so don't provide different defaults.
> 
> In practice the difference is meaningless as the only portable driver
> that uses ioremap_uc is atyfb which probably doesn't show up on nommu
> devices.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

[ ... ]

> diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
> index d02806513670..a98ed6325727 100644
> --- a/include/asm-generic/io.h
> +++ b/include/asm-generic/io.h

[ ... ]

> @@ -1004,6 +985,21 @@ static inline void __iomem *ioremap_wt(phys_addr_t offset, size_t size)
>  }
>  #endif
>  
> +/*
> + * ioremap_uc is special in that we do require an explicit architecture
> + * implementation.  In general you do now want to use this function in a
                                         ^^^ not

> + * driver and use plain ioremap, which is uncached by default.  Similarly
                ^ instead 

> + * architectures should not implement it unless they have a very good
> + * reason.
> + */

Looks like this mess is only needed on x86 with certain graphics drivers 
and conflicts between MTRR and page table-based MMU attributes.


Reviewed-by: Paul Walmsley <paul.walmsley@sifive.com>
Tested-by: Paul Walmsley <paul.walmsley@sifive.com> # rv32, rv64 boot


- Paul