[Pdbg] [PATCH v2 11/19] libpdbg: Add chiplet driver using sbefifo

Alistair Popple alistair at popple.id.au
Thu Apr 9 12:29:07 AEST 2020


On Tuesday, 7 April 2020 4:16:05 PM AEST Amitay Isaacs wrote:
> Signed-off-by: Amitay Isaacs <amitay at ozlabs.org>
> ---
>  libpdbg/sbefifo.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/libpdbg/sbefifo.c b/libpdbg/sbefifo.c
> index 112504f..72a58f9 100644
> --- a/libpdbg/sbefifo.c
> +++ b/libpdbg/sbefifo.c
> @@ -334,6 +334,15 @@ static struct pib sbefifo_pib = {
>  };
>  DECLARE_HW_UNIT(sbefifo_pib);
> 
> +static struct chiplet sbefifo_chiplet = {
> +        .target = {
> +                .name = "SBE FIFO Chip-op based Chiplet",
> +                .compatible = "ibm,sbefifo-chiplet",

I am wondering, shouldn't these have the same compatible property as the 
default chiplet implementation? We search the backend specific implementations 
first, so would just pick up the correct implementation based on backend first 
anyway right?

- Alistair

> +                .class = "chiplet",
> +        },
> +};
> +DECLARE_HW_UNIT(sbefifo_chiplet);
> +
>  static struct sbefifo kernel_sbefifo = {
>  	.target = {
>  		.name =	"Kernel based FSI SBE FIFO",
> @@ -352,6 +361,7 @@ static void register_sbefifo(void)
>  	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &kernel_sbefifo_hw_unit);
>  	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &sbefifo_chipop_hw_unit);
>  	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &sbefifo_pib_hw_unit);
> +	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &sbefifo_chiplet_hw_unit);
>  	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &sbefifo_mem_hw_unit);
>  	pdbg_hwunit_register(PDBG_DEFAULT_BACKEND, &sbefifo_pba_hw_unit);
>  }






More information about the Pdbg mailing list