diff options
author | Lennert Buytenhek <buytenh@wantstofly.org> | 2006-03-28 21:00:40 +0100 |
---|---|---|
committer | Russell King <rmk+kernel@arm.linux.org.uk> | 2006-03-28 21:00:40 +0100 |
commit | 23bdf86aa06ebe71bcbf6b7d25de9958c6ab33fa (patch) | |
tree | 56636558e8cdeee0739e7d8c82d66ffe625340b3 /include | |
parent | de4533a04eb4f66dbef71f59a9c118256b886823 (diff) | |
download | linux-23bdf86aa06ebe71bcbf6b7d25de9958c6ab33fa.tar.gz linux-23bdf86aa06ebe71bcbf6b7d25de9958c6ab33fa.tar.bz2 linux-23bdf86aa06ebe71bcbf6b7d25de9958c6ab33fa.zip |
[ARM] 3377/2: add support for intel xsc3 core
Patch from Lennert Buytenhek
This patch adds support for the new XScale v3 core. This is an
ARMv5 ISA core with the following additions:
- L2 cache
- I/O coherency support (on select chipsets)
- Low-Locality Reference cache attributes (replaces mini-cache)
- Supersections (v6 compatible)
- 36-bit addressing (v6 compatible)
- Single instruction cache line clean/invalidate
- LRU cache replacement (vs round-robin)
I attempted to merge the XSC3 support into proc-xscale.S, but XSC3
cores have separate errata and have to handle things like L2, so it
is simpler to keep it separate.
L2 cache support is currently a build option because the L2 enable
bit must be set before we enable the MMU and there is no easy way to
capture command line parameters at this point.
There are still optimizations that can be done such as using LLR for
copypage (in theory using the exisiting mini-cache code) but those
can be addressed down the road.
Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'include')
-rw-r--r-- | include/asm-arm/cacheflush.h | 8 | ||||
-rw-r--r-- | include/asm-arm/domain.h | 18 | ||||
-rw-r--r-- | include/asm-arm/page.h | 9 | ||||
-rw-r--r-- | include/asm-arm/proc-fns.h | 8 | ||||
-rw-r--r-- | include/asm-arm/system.h | 19 |
5 files changed, 62 insertions, 0 deletions
diff --git a/include/asm-arm/cacheflush.h b/include/asm-arm/cacheflush.h index 09e19a783a51..746be56b1b70 100644 --- a/include/asm-arm/cacheflush.h +++ b/include/asm-arm/cacheflush.h @@ -71,6 +71,14 @@ # endif #endif +#if defined(CONFIG_CPU_XSC3) +# ifdef _CACHE +# define MULTI_CACHE 1 +# else +# define _CACHE xsc3 +# endif +#endif + #if defined(CONFIG_CPU_V6) //# ifdef _CACHE # define MULTI_CACHE 1 diff --git a/include/asm-arm/domain.h b/include/asm-arm/domain.h index da1d960387d9..f8ea2de4848e 100644 --- a/include/asm-arm/domain.h +++ b/include/asm-arm/domain.h @@ -16,11 +16,29 @@ * DOMAIN_IO - domain 2 includes all IO only * DOMAIN_USER - domain 1 includes all user memory only * DOMAIN_KERNEL - domain 0 includes all kernel memory only + * + * The domain numbering depends on whether we support 36 physical + * address for I/O or not. Addresses above the 32 bit boundary can + * only be mapped using supersections and supersections can only + * be set for domain 0. We could just default to DOMAIN_IO as zero, + * but there may be systems with supersection support and no 36-bit + * addressing. In such cases, we want to map system memory with + * supersections to reduce TLB misses and footprint. + * + * 36-bit addressing and supersections are only available on + * CPUs based on ARMv6+ or the Intel XSC3 core. */ +#ifndef CONFIG_IO_36 #define DOMAIN_KERNEL 0 #define DOMAIN_TABLE 0 #define DOMAIN_USER 1 #define DOMAIN_IO 2 +#else +#define DOMAIN_KERNEL 2 +#define DOMAIN_TABLE 2 +#define DOMAIN_USER 1 +#define DOMAIN_IO 0 +#endif /* * Domain types diff --git a/include/asm-arm/page.h b/include/asm-arm/page.h index 416320d95419..a404d2bf0c68 100644 --- a/include/asm-arm/page.h +++ b/include/asm-arm/page.h @@ -40,6 +40,7 @@ * v4wb - ARMv4 with writeback cache, without minicache * v4_mc - ARMv4 with minicache * xscale - Xscale + * xsc3 - XScalev3 */ #undef _USER #undef MULTI_USER @@ -84,6 +85,14 @@ # endif #endif +#ifdef CONFIG_CPU_XSC3 +# ifdef _USER +# define MULTI_USER 1 +# else +# define _USER xsc3_mc +# endif +#endif + #ifdef CONFIG_CPU_COPY_V6 # define MULTI_USER 1 #endif diff --git a/include/asm-arm/proc-fns.h b/include/asm-arm/proc-fns.h index 7bef2bf6be51..106045edb862 100644 --- a/include/asm-arm/proc-fns.h +++ b/include/asm-arm/proc-fns.h @@ -138,6 +138,14 @@ # define CPU_NAME cpu_xscale # endif # endif +# ifdef CONFIG_CPU_XSC3 +# ifdef CPU_NAME +# undef MULTI_CPU +# define MULTI_CPU +# else +# define CPU_NAME cpu_xsc3 +# endif +# endif # ifdef CONFIG_CPU_V6 # ifdef CPU_NAME # undef MULTI_CPU diff --git a/include/asm-arm/system.h b/include/asm-arm/system.h index ec91d1ff032a..95b3abf4851b 100644 --- a/include/asm-arm/system.h +++ b/include/asm-arm/system.h @@ -108,6 +108,25 @@ extern void __show_regs(struct pt_regs *); extern int cpu_architecture(void); extern void cpu_init(void); +/* + * Intel's XScale3 core supports some v6 features (supersections, L2) + * but advertises itself as v5 as it does not support the v6 ISA. For + * this reason, we need a way to explicitly test for this type of CPU. + */ +#ifndef CONFIG_CPU_XSC3 +#define cpu_is_xsc3() 0 +#else +static inline int cpu_is_xsc3(void) +{ + extern unsigned int processor_id; + + if ((processor_id & 0xffffe000) == 0x69056000) + return 1; + + return 0; +} +#endif + #define set_cr(x) \ __asm__ __volatile__( \ "mcr p15, 0, %0, c1, c0, 0 @ set CR" \ |