[coreboot] [commit] r6633 - trunk/src/mainboard/asrock/e350m1

[Stefan Reinauer]

* repository service <svn at coreboot.org> [110604 17:47]:
> Author: stuge
> Date: Sat Jun  4 17:47:05 2011
> New Revision: 6633
> URL: https://tracker.coreboot.org/trac/coreboot/changeset/6633
> 
> Log:
> Port persimmon r6591 to e350m1: ROM cache early
> 
> Enable rom cache early to reduce boot time.
> 
> Signed-off-by: Marshall Buschman <mbuschman at lucidmachines.com>
> Acked-by: Peter Stuge <peter at stuge.se>
> 
> Modified:
>    trunk/src/mainboard/asrock/e350m1/agesawrapper.c
>    trunk/src/mainboard/asrock/e350m1/romstage.c
> 
> Modified: trunk/src/mainboard/asrock/e350m1/agesawrapper.c
> ==============================================================================
> --- trunk/src/mainboard/asrock/e350m1/agesawrapper.c	Sat Jun  4 17:46:50 2011	(r6632)
> +++ trunk/src/mainboard/asrock/e350m1/agesawrapper.c	Sat Jun  4 17:47:05 2011	(r6633)
> @@ -157,13 +157,6 @@
>    PciData = (AMD_APU_SSID<<0x10)|AMD_APU_SVID;
>    LibAmdPciWrite(AccessWidth32, PciAddress, &PciData, &StdHeader); 
>    
> -
> -  /* Set ROM cache onto WP to decrease post time */
> -  MsrReg = (0x0100000000 - CONFIG_ROM_SIZE) | 5;
> -  LibAmdMsrWrite (0x20C, &MsrReg, &StdHeader);
> -  MsrReg = (0x1000000000 - CONFIG_ROM_SIZE) | 0x800;
> -  LibAmdMsrWrite (0x20D, &MsrReg, &StdHeader);
> -
>    Status = AGESA_SUCCESS;
>    return (UINT32)Status;
>  }
> 
> Modified: trunk/src/mainboard/asrock/e350m1/romstage.c
> ==============================================================================
> --- trunk/src/mainboard/asrock/e350m1/romstage.c	Sat Jun  4 17:46:50 2011	(r6632)
> +++ trunk/src/mainboard/asrock/e350m1/romstage.c	Sat Jun  4 17:47:05 2011	(r6633)
> @@ -47,6 +47,11 @@
>    u32 val;
>    u8 reg8;
>  
> +  // all cores: allow caching of flash chip code and data
> +  // (there are no cache-as-ram reliability concerns with family 14h)
> +  __writemsr (0x20c, (0x0100000000ull - CONFIG_ROM_SIZE) | 5);
> +  __writemsr (0x20d, (0x1000000000ull - CONFIG_ROM_SIZE) | 0x800);
> +


again, why __writemsr instead of writemsr?

I think this should go into cache_as_ram.S like on (all the?) other CPUs