Re: [PATCH v5 0/4] drivers/soc/litex: support 32-bit subregisters, 64-bit CPUs

["Gabriel L. Somlo" <gsomlo@gmail.com>]

Ping :)

On Sun, Dec 27, 2020 at 11:13:16AM -0500, Gabriel Somlo wrote:
> This series expands on commit 22447a99c97e ("drivers/soc/litex: add LiteX
> SoC Controller driver"), adding support for handling both 8- and 32-bit
> LiteX CSR (MMIO) subregisters, on both 32- and 64-bit CPUs.
> 
> Notes v5:
> 	- added patch (4/4) taking 'litex_[set|get]_reg()' private
> 	- additional optimization of [_]litex_set_reg() in 3/4
> 
> Notes v4:
> 	- improved "eloquence" of some 3/3 commit blurb paragraphs
> 	- fixed instance of "disgusting" comment style :)
> 	- litex_[get|set]_reg() now using size_t for 'reg_size' argument
> 	- slightly tighter shift calculation in litex_set_reg()
> 
> Notes v3:
> 	- split into smaller, more self-explanatory patches
> 	- more detailed commit blurb for "main payload" (3/3) patch
> 	- eliminate compiler warning on 32-bit architectures
> 
> Notes v2:
> 	- fix typo (s/u32/u64/) in litex_read64().
> 
> Notes v1:
> 	- LITEX_SUBREG_SIZE now provided by Kconfig.
> 	- it's not LITEX_REG_SIZE, but rather LITEX_SUBREG_ALIGN!
> 	- move litex_[get|set]_reg() to include/linux/litex.h and mark
> 	  them as "static inline";
> 	- redo litex_[read|write][8|16|32|64]() using litex_[get|set]_reg()
> 	  (compiler should produce code as efficient as hardcoded shifts,
> 	   but also automatically matching LITEX_SUBREG_SIZE).
> 
> Gabriel Somlo (4):
>   drivers/soc/litex: move generic accessors to litex.h
>   drivers/soc/litex: separate MMIO from subregister offset calculation
>   drivers/soc/litex: support 32-bit subregisters, 64-bit CPUs
>   drivers/soc/litex: make 'litex_[set|get]_reg()' methods private
> 
>  drivers/soc/litex/Kconfig          |  14 ++-
>  drivers/soc/litex/litex_soc_ctrl.c |  76 +-------------
>  include/linux/litex.h              | 154 +++++++++++++++++++----------
>  3 files changed, 119 insertions(+), 125 deletions(-)
> 
> -- 
> 2.26.2
>