diff options
Diffstat (limited to 'arch/mips/mm')
-rw-r--r-- | arch/mips/mm/c-octeon.c | 20 | ||||
-rw-r--r-- | arch/mips/mm/c-r4k.c | 102 | ||||
-rw-r--r-- | arch/mips/mm/dma-default.c | 33 | ||||
-rw-r--r-- | arch/mips/mm/fault.c | 8 | ||||
-rw-r--r-- | arch/mips/mm/gup.c | 3 | ||||
-rw-r--r-- | arch/mips/mm/hugetlbpage.c | 5 | ||||
-rw-r--r-- | arch/mips/mm/init.c | 6 | ||||
-rw-r--r-- | arch/mips/mm/page.c | 1 | ||||
-rw-r--r-- | arch/mips/mm/sc-mips.c | 3 | ||||
-rw-r--r-- | arch/mips/mm/tlb-r4k.c | 38 | ||||
-rw-r--r-- | arch/mips/mm/tlbex.c | 310 |
11 files changed, 289 insertions, 240 deletions
diff --git a/arch/mips/mm/c-octeon.c b/arch/mips/mm/c-octeon.c index a0bcdbb81d41..c8efdb5b6ee0 100644 --- a/arch/mips/mm/c-octeon.c +++ b/arch/mips/mm/c-octeon.c @@ -19,6 +19,7 @@ #include <asm/bootinfo.h> #include <asm/cacheops.h> #include <asm/cpu-features.h> +#include <asm/cpu-type.h> #include <asm/page.h> #include <asm/pgtable.h> #include <asm/r4kcache.h> @@ -186,9 +187,10 @@ static void probe_octeon(void) unsigned long dcache_size; unsigned int config1; struct cpuinfo_mips *c = ¤t_cpu_data; + int cputype = current_cpu_type(); config1 = read_c0_config1(); - switch (c->cputype) { + switch (cputype) { case CPU_CAVIUM_OCTEON: case CPU_CAVIUM_OCTEON_PLUS: c->icache.linesz = 2 << ((config1 >> 19) & 7); @@ -199,7 +201,7 @@ static void probe_octeon(void) c->icache.sets * c->icache.ways * c->icache.linesz; c->icache.waybit = ffs(icache_size / c->icache.ways) - 1; c->dcache.linesz = 128; - if (c->cputype == CPU_CAVIUM_OCTEON_PLUS) + if (cputype == CPU_CAVIUM_OCTEON_PLUS) c->dcache.sets = 2; /* CN5XXX has two Dcache sets */ else c->dcache.sets = 1; /* CN3XXX has one Dcache set */ @@ -224,6 +226,20 @@ static void probe_octeon(void) c->options |= MIPS_CPU_PREFETCH; break; + case CPU_CAVIUM_OCTEON3: + c->icache.linesz = 128; + c->icache.sets = 16; + c->icache.ways = 39; + c->icache.flags |= MIPS_CACHE_VTAG; + icache_size = c->icache.sets * c->icache.ways * c->icache.linesz; + + c->dcache.linesz = 128; + c->dcache.ways = 32; + c->dcache.sets = 8; + dcache_size = c->dcache.sets * c->dcache.ways * c->dcache.linesz; + c->options |= MIPS_CPU_PREFETCH; + break; + default: panic("Unsupported Cavium Networks CPU type"); break; diff --git a/arch/mips/mm/c-r4k.c b/arch/mips/mm/c-r4k.c index f749f687ee87..62ffd20ea869 100644 --- a/arch/mips/mm/c-r4k.c +++ b/arch/mips/mm/c-r4k.c @@ -12,6 +12,7 @@ #include <linux/highmem.h> #include <linux/kernel.h> #include <linux/linkage.h> +#include <linux/preempt.h> #include <linux/sched.h> #include <linux/smp.h> #include <linux/mm.h> @@ -24,6 +25,7 @@ #include <asm/cacheops.h> #include <asm/cpu.h> #include <asm/cpu-features.h> +#include <asm/cpu-type.h> #include <asm/io.h> #include <asm/page.h> #include <asm/pgtable.h> @@ -344,14 +346,8 @@ static void r4k_blast_scache_setup(void) static inline void local_r4k___flush_cache_all(void * args) { -#if defined(CONFIG_CPU_LOONGSON2) - r4k_blast_scache(); - return; -#endif - r4k_blast_dcache(); - r4k_blast_icache(); - switch (current_cpu_type()) { + case CPU_LOONGSON2: case CPU_R4000SC: case CPU_R4000MC: case CPU_R4400SC: @@ -359,7 +355,18 @@ static inline void local_r4k___flush_cache_all(void * args) case CPU_R10000: case CPU_R12000: case CPU_R14000: + /* + * These caches are inclusive caches, that is, if something + * is not cached in the S-cache, we know it also won't be + * in one of the primary caches. + */ r4k_blast_scache(); + break; + + default: + r4k_blast_dcache(); + r4k_blast_icache(); + break; } } @@ -570,8 +577,17 @@ static inline void local_r4k_flush_icache_range(unsigned long start, unsigned lo if (end - start > icache_size) r4k_blast_icache(); - else - protected_blast_icache_range(start, end); + else { + switch (boot_cpu_type()) { + case CPU_LOONGSON2: + protected_blast_icache_range(start, end); + break; + + default: + protected_loongson23_blast_icache_range(start, end); + break; + } + } } static inline void local_r4k_flush_icache_range_ipi(void *args) @@ -601,11 +617,13 @@ static void r4k_dma_cache_wback_inv(unsigned long addr, unsigned long size) /* Catch bad driver code */ BUG_ON(size == 0); + preempt_disable(); if (cpu_has_inclusive_pcaches) { if (size >= scache_size) r4k_blast_scache(); else blast_scache_range(addr, addr + size); + preempt_enable(); __sync(); return; } @@ -621,6 +639,7 @@ static void r4k_dma_cache_wback_inv(unsigned long addr, unsigned long size) R4600_HIT_CACHEOP_WAR_IMPL; blast_dcache_range(addr, addr + size); } + preempt_enable(); bc_wback_inv(addr, size); __sync(); @@ -631,6 +650,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size) /* Catch bad driver code */ BUG_ON(size == 0); + preempt_disable(); if (cpu_has_inclusive_pcaches) { if (size >= scache_size) r4k_blast_scache(); @@ -645,6 +665,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size) */ blast_inv_scache_range(addr, addr + size); } + preempt_enable(); __sync(); return; } @@ -655,6 +676,7 @@ static void r4k_dma_cache_inv(unsigned long addr, unsigned long size) R4600_HIT_CACHEOP_WAR_IMPL; blast_inv_dcache_range(addr, addr + size); } + preempt_enable(); bc_inv(addr, size); __sync(); @@ -780,20 +802,30 @@ static inline void rm7k_erratum31(void) static inline void alias_74k_erratum(struct cpuinfo_mips *c) { + unsigned int imp = c->processor_id & PRID_IMP_MASK; + unsigned int rev = c->processor_id & PRID_REV_MASK; + /* * Early versions of the 74K do not update the cache tags on a * vtag miss/ptag hit which can occur in the case of KSEG0/KUSEG * aliases. In this case it is better to treat the cache as always * having aliases. */ - if ((c->processor_id & 0xff) <= PRID_REV_ENCODE_332(2, 4, 0)) - c->dcache.flags |= MIPS_CACHE_VTAG; - if ((c->processor_id & 0xff) == PRID_REV_ENCODE_332(2, 4, 0)) - write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND); - if (((c->processor_id & 0xff00) == PRID_IMP_1074K) && - ((c->processor_id & 0xff) <= PRID_REV_ENCODE_332(1, 1, 0))) { - c->dcache.flags |= MIPS_CACHE_VTAG; - write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND); + switch (imp) { + case PRID_IMP_74K: + if (rev <= PRID_REV_ENCODE_332(2, 4, 0)) + c->dcache.flags |= MIPS_CACHE_VTAG; + if (rev == PRID_REV_ENCODE_332(2, 4, 0)) + write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND); + break; + case PRID_IMP_1074K: + if (rev <= PRID_REV_ENCODE_332(1, 1, 0)) { + c->dcache.flags |= MIPS_CACHE_VTAG; + write_c0_config6(read_c0_config6() | MIPS_CONF6_SYND); + } + break; + default: + BUG(); } } @@ -809,7 +841,7 @@ static void probe_pcache(void) unsigned long config1; unsigned int lsize; - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_R4600: /* QED style two way caches? */ case CPU_R4700: case CPU_R5000: @@ -1025,7 +1057,8 @@ static void probe_pcache(void) * presumably no vendor is shipping his hardware in the "bad" * configuration. */ - if ((prid & 0xff00) == PRID_IMP_R4000 && (prid & 0xff) < 0x40 && + if ((prid & PRID_IMP_MASK) == PRID_IMP_R4000 && + (prid & PRID_REV_MASK) < PRID_REV_R4400 && !(config & CONF_SC) && c->icache.linesz != 16 && PAGE_SIZE <= 0x8000) panic("Improper R4000SC processor configuration detected"); @@ -1045,7 +1078,7 @@ static void probe_pcache(void) * normally they'd suffer from aliases but magic in the hardware deals * with that for us so we don't need to take care ourselves. */ - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_20KC: case CPU_25KF: case CPU_SB1: @@ -1065,7 +1098,7 @@ static void probe_pcache(void) case CPU_34K: case CPU_74K: case CPU_1004K: - if (c->cputype == CPU_74K) + if (current_cpu_type() == CPU_74K) alias_74k_erratum(c); if ((read_c0_config7() & (1 << 16))) { /* effectively physically indexed dcache, @@ -1078,7 +1111,7 @@ static void probe_pcache(void) c->dcache.flags |= MIPS_CACHE_ALIASES; } - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_20KC: /* * Some older 20Kc chips doesn't have the 'VI' bit in @@ -1090,15 +1123,14 @@ static void probe_pcache(void) case CPU_ALCHEMY: c->icache.flags |= MIPS_CACHE_IC_F_DC; break; - } -#ifdef CONFIG_CPU_LOONGSON2 - /* - * LOONGSON2 has 4 way icache, but when using indexed cache op, - * one op will act on all 4 ways - */ - c->icache.ways = 1; -#endif + case CPU_LOONGSON2: + /* + * LOONGSON2 has 4 way icache, but when using indexed cache op, + * one op will act on all 4 ways + */ + c->icache.ways = 1; + } printk("Primary instruction cache %ldkB, %s, %s, linesize %d bytes.\n", icache_size >> 10, @@ -1174,7 +1206,6 @@ static int probe_scache(void) return 1; } -#if defined(CONFIG_CPU_LOONGSON2) static void __init loongson2_sc_init(void) { struct cpuinfo_mips *c = ¤t_cpu_data; @@ -1190,7 +1221,6 @@ static void __init loongson2_sc_init(void) c->options |= MIPS_CPU_INCLUSIVE_CACHES; } -#endif extern int r5k_sc_init(void); extern int rm7k_sc_init(void); @@ -1207,7 +1237,7 @@ static void setup_scache(void) * processors don't have a S-cache that would be relevant to the * Linux memory management. */ - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_R4000SC: case CPU_R4000MC: case CPU_R4400SC: @@ -1240,11 +1270,10 @@ static void setup_scache(void) #endif return; -#if defined(CONFIG_CPU_LOONGSON2) case CPU_LOONGSON2: loongson2_sc_init(); return; -#endif + case CPU_XLP: /* don't need to worry about L2, fully coherent */ return; @@ -1384,9 +1413,8 @@ static void r4k_cache_error_setup(void) { extern char __weak except_vec2_generic; extern char __weak except_vec2_sb1; - struct cpuinfo_mips *c = ¤t_cpu_data; - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_SB1: case CPU_SB1A: set_uncached_handler(0x100, &except_vec2_sb1, 0x80); diff --git a/arch/mips/mm/dma-default.c b/arch/mips/mm/dma-default.c index aaccf1c10699..2e9418562258 100644 --- a/arch/mips/mm/dma-default.c +++ b/arch/mips/mm/dma-default.c @@ -18,6 +18,7 @@ #include <linux/highmem.h> #include <asm/cache.h> +#include <asm/cpu-type.h> #include <asm/io.h> #include <dma-coherence.h> @@ -50,16 +51,20 @@ static inline struct page *dma_addr_to_page(struct device *dev, } /* + * The affected CPUs below in 'cpu_needs_post_dma_flush()' can + * speculatively fill random cachelines with stale data at any time, + * requiring an extra flush post-DMA. + * * Warning on the terminology - Linux calls an uncached area coherent; * MIPS terminology calls memory areas with hardware maintained coherency * coherent. */ - -static inline int cpu_is_noncoherent_r10000(struct device *dev) +static inline int cpu_needs_post_dma_flush(struct device *dev) { return !plat_device_is_coherent(dev) && - (current_cpu_type() == CPU_R10000 || - current_cpu_type() == CPU_R12000); + (boot_cpu_type() == CPU_R10000 || + boot_cpu_type() == CPU_R12000 || + boot_cpu_type() == CPU_BMIPS5000); } static gfp_t massage_gfp_flags(const struct device *dev, gfp_t gfp) @@ -230,7 +235,7 @@ static inline void __dma_sync(struct page *page, static void mips_dma_unmap_page(struct device *dev, dma_addr_t dma_addr, size_t size, enum dma_data_direction direction, struct dma_attrs *attrs) { - if (cpu_is_noncoherent_r10000(dev)) + if (cpu_needs_post_dma_flush(dev)) __dma_sync(dma_addr_to_page(dev, dma_addr), dma_addr & ~PAGE_MASK, size, direction); @@ -284,7 +289,7 @@ static void mips_dma_unmap_sg(struct device *dev, struct scatterlist *sg, static void mips_dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, size_t size, enum dma_data_direction direction) { - if (cpu_is_noncoherent_r10000(dev)) + if (cpu_needs_post_dma_flush(dev)) __dma_sync(dma_addr_to_page(dev, dma_handle), dma_handle & ~PAGE_MASK, size, direction); } @@ -292,7 +297,6 @@ static void mips_dma_sync_single_for_cpu(struct device *dev, static void mips_dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, size_t size, enum dma_data_direction direction) { - plat_extra_sync_for_device(dev); if (!plat_device_is_coherent(dev)) __dma_sync(dma_addr_to_page(dev, dma_handle), dma_handle & ~PAGE_MASK, size, direction); @@ -303,12 +307,10 @@ static void mips_dma_sync_sg_for_cpu(struct device *dev, { int i; - /* Make sure that gcc doesn't leave the empty loop body. */ - for (i = 0; i < nelems; i++, sg++) { - if (cpu_is_noncoherent_r10000(dev)) + if (cpu_needs_post_dma_flush(dev)) + for (i = 0; i < nelems; i++, sg++) __dma_sync(sg_page(sg), sg->offset, sg->length, direction); - } } static void mips_dma_sync_sg_for_device(struct device *dev, @@ -316,17 +318,15 @@ static void mips_dma_sync_sg_for_device(struct device *dev, { int i; - /* Make sure that gcc doesn't leave the empty loop body. */ - for (i = 0; i < nelems; i++, sg++) { - if (!plat_device_is_coherent(dev)) + if (!plat_device_is_coherent(dev)) + for (i = 0; i < nelems; i++, sg++) __dma_sync(sg_page(sg), sg->offset, sg->length, direction); - } } int mips_dma_mapping_error(struct device *dev, dma_addr_t dma_addr) { - return plat_dma_mapping_error(dev, dma_addr); + return 0; } int mips_dma_supported(struct device *dev, u64 mask) @@ -339,7 +339,6 @@ void dma_cache_sync(struct device *dev, void *vaddr, size_t size, { BUG_ON(direction == DMA_NONE); - plat_extra_sync_for_device(dev); if (!plat_device_is_coherent(dev)) __dma_sync_virtual(vaddr, size, direction); } diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c index 85df1cd8d446..becc42bb1849 100644 --- a/arch/mips/mm/fault.c +++ b/arch/mips/mm/fault.c @@ -42,8 +42,7 @@ static void __kprobes __do_page_fault(struct pt_regs *regs, unsigned long write, const int field = sizeof(unsigned long) * 2; siginfo_t info; int fault; - unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE | - (write ? FAULT_FLAG_WRITE : 0); + unsigned int flags = FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE; #if 0 printk("Cpu%d[%s:%d:%0*lx:%ld:%0*lx]\n", raw_smp_processor_id(), @@ -93,6 +92,8 @@ static void __kprobes __do_page_fault(struct pt_regs *regs, unsigned long write, if (in_atomic() || !mm) goto bad_area_nosemaphore; + if (user_mode(regs)) + flags |= FAULT_FLAG_USER; retry: down_read(&mm->mmap_sem); vma = find_vma(mm, address); @@ -114,6 +115,7 @@ good_area: if (write) { if (!(vma->vm_flags & VM_WRITE)) goto bad_area; + flags |= FAULT_FLAG_WRITE; } else { if (cpu_has_rixi) { if (address == regs->cp0_epc && !(vma->vm_flags & VM_EXEC)) { @@ -241,6 +243,8 @@ out_of_memory: * (which will retry the fault, or kill us if we got oom-killed). */ up_read(&mm->mmap_sem); + if (!user_mode(regs)) + goto no_context; pagefault_out_of_memory(); return; diff --git a/arch/mips/mm/gup.c b/arch/mips/mm/gup.c index d4ea5c9c4a93..06ce17c2a905 100644 --- a/arch/mips/mm/gup.c +++ b/arch/mips/mm/gup.c @@ -12,6 +12,7 @@ #include <linux/swap.h> #include <linux/hugetlb.h> +#include <asm/cpu-features.h> #include <asm/pgtable.h> static inline pte_t gup_get_pte(pte_t *ptep) @@ -273,7 +274,7 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write, len = (unsigned long) nr_pages << PAGE_SHIFT; end = start + len; - if (end < start) + if (end < start || cpu_has_dc_aliases) goto slow_irqon; /* XXX: batch / limit 'nr' */ diff --git a/arch/mips/mm/hugetlbpage.c b/arch/mips/mm/hugetlbpage.c index a7fee0dfb7a9..01fda4419ed0 100644 --- a/arch/mips/mm/hugetlbpage.c +++ b/arch/mips/mm/hugetlbpage.c @@ -85,6 +85,11 @@ int pud_huge(pud_t pud) return (pud_val(pud) & _PAGE_HUGE) != 0; } +int pmd_huge_support(void) +{ + return 1; +} + struct page * follow_huge_pmd(struct mm_struct *mm, unsigned long address, pmd_t *pmd, int write) diff --git a/arch/mips/mm/init.c b/arch/mips/mm/init.c index 4e73f10a7519..12156176c7ca 100644 --- a/arch/mips/mm/init.c +++ b/arch/mips/mm/init.c @@ -124,7 +124,7 @@ void *kmap_coherent(struct page *page, unsigned long addr) BUG_ON(Page_dcache_dirty(page)); - inc_preempt_count(); + pagefault_disable(); idx = (addr >> PAGE_SHIFT) & (FIX_N_COLOURS - 1); #ifdef CONFIG_MIPS_MT_SMTC idx += FIX_N_COLOURS * smp_processor_id() + @@ -193,8 +193,7 @@ void kunmap_coherent(void) write_c0_entryhi(old_ctx); EXIT_CRITICAL(flags); #endif - dec_preempt_count(); - preempt_check_resched(); + pagefault_enable(); } void copy_user_highpage(struct page *to, struct page *from, @@ -254,6 +253,7 @@ void copy_from_user_page(struct vm_area_struct *vma, SetPageDcacheDirty(page); } } +EXPORT_SYMBOL_GPL(copy_from_user_page); void __init fixrange_init(unsigned long start, unsigned long end, pgd_t *pgd_base) diff --git a/arch/mips/mm/page.c b/arch/mips/mm/page.c index 218c2109a55d..cbd81d17793a 100644 --- a/arch/mips/mm/page.c +++ b/arch/mips/mm/page.c @@ -18,6 +18,7 @@ #include <asm/bugs.h> #include <asm/cacheops.h> +#include <asm/cpu-type.h> #include <asm/inst.h> #include <asm/io.h> #include <asm/page.h> diff --git a/arch/mips/mm/sc-mips.c b/arch/mips/mm/sc-mips.c index 5d01392e3518..08d05aee8788 100644 --- a/arch/mips/mm/sc-mips.c +++ b/arch/mips/mm/sc-mips.c @@ -6,6 +6,7 @@ #include <linux/sched.h> #include <linux/mm.h> +#include <asm/cpu-type.h> #include <asm/mipsregs.h> #include <asm/bcache.h> #include <asm/cacheops.h> @@ -71,7 +72,7 @@ static inline int mips_sc_is_activated(struct cpuinfo_mips *c) unsigned int tmp; /* Check the bypass bit (L2B) */ - switch (c->cputype) { + switch (current_cpu_type()) { case CPU_34K: case CPU_74K: case CPU_1004K: diff --git a/arch/mips/mm/tlb-r4k.c b/arch/mips/mm/tlb-r4k.c index 00b26a67a06d..da3b0b9c9eae 100644 --- a/arch/mips/mm/tlb-r4k.c +++ b/arch/mips/mm/tlb-r4k.c @@ -16,6 +16,7 @@ #include <linux/module.h> #include <asm/cpu.h> +#include <asm/cpu-type.h> #include <asm/bootinfo.h> #include <asm/mmu_context.h> #include <asm/pgtable.h> @@ -51,21 +52,26 @@ extern void build_tlb_refill_handler(void); #endif /* CONFIG_MIPS_MT_SMTC */ -#if defined(CONFIG_CPU_LOONGSON2) /* * LOONGSON2 has a 4 entry itlb which is a subset of dtlb, * unfortrunately, itlb is not totally transparent to software. */ -#define FLUSH_ITLB write_c0_diag(4); - -#define FLUSH_ITLB_VM(vma) { if ((vma)->vm_flags & VM_EXEC) write_c0_diag(4); } - -#else - -#define FLUSH_ITLB -#define FLUSH_ITLB_VM(vma) +static inline void flush_itlb(void) +{ + switch (current_cpu_type()) { + case CPU_LOONGSON2: + write_c0_diag(4); + break; + default: + break; + } +} -#endif +static inline void flush_itlb_vm(struct vm_area_struct *vma) +{ + if (vma->vm_flags & VM_EXEC) + flush_itlb(); +} void local_flush_tlb_all(void) { @@ -92,7 +98,7 @@ void local_flush_tlb_all(void) } tlbw_use_hazard(); write_c0_entryhi(old_ctx); - FLUSH_ITLB; + flush_itlb(); EXIT_CRITICAL(flags); } EXPORT_SYMBOL(local_flush_tlb_all); @@ -154,7 +160,7 @@ void local_flush_tlb_range(struct vm_area_struct *vma, unsigned long start, } else { drop_mmu_context(mm, cpu); } - FLUSH_ITLB; + flush_itlb(); EXIT_CRITICAL(flags); } } @@ -196,7 +202,7 @@ void local_flush_tlb_kernel_range(unsigned long start, unsigned long end) } else { local_flush_tlb_all(); } - FLUSH_ITLB; + flush_itlb(); EXIT_CRITICAL(flags); } @@ -229,7 +235,7 @@ void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long page) finish: write_c0_entryhi(oldpid); - FLUSH_ITLB_VM(vma); + flush_itlb_vm(vma); EXIT_CRITICAL(flags); } } @@ -261,7 +267,7 @@ void local_flush_tlb_one(unsigned long page) tlbw_use_hazard(); } write_c0_entryhi(oldpid); - FLUSH_ITLB; + flush_itlb(); EXIT_CRITICAL(flags); } @@ -334,7 +340,7 @@ void __update_tlb(struct vm_area_struct * vma, unsigned long address, pte_t pte) tlb_write_indexed(); } tlbw_use_hazard(); - FLUSH_ITLB_VM(vma); + flush_itlb_vm(vma); EXIT_CRITICAL(flags); } diff --git a/arch/mips/mm/tlbex.c b/arch/mips/mm/tlbex.c index 556cb4815770..183f2b583e4d 100644 --- a/arch/mips/mm/tlbex.c +++ b/arch/mips/mm/tlbex.c @@ -30,6 +30,7 @@ #include <linux/cache.h> #include <asm/cacheflush.h> +#include <asm/cpu-type.h> #include <asm/pgtable.h> #include <asm/war.h> #include <asm/uasm.h> @@ -85,6 +86,7 @@ static int use_bbit_insns(void) case CPU_CAVIUM_OCTEON: case CPU_CAVIUM_OCTEON_PLUS: case CPU_CAVIUM_OCTEON2: + case CPU_CAVIUM_OCTEON3: return 1; default: return 0; @@ -95,6 +97,7 @@ static int use_lwx_insns(void) { switch (current_cpu_type()) { case CPU_CAVIUM_OCTEON2: + case CPU_CAVIUM_OCTEON3: return 1; default: return 0; @@ -337,10 +340,6 @@ static struct work_registers build_get_work_registers(u32 **p) { struct work_registers r; - int smp_processor_id_reg; - int smp_processor_id_sel; - int smp_processor_id_shift; - if (scratch_reg >= 0) { /* Save in CPU local C0_KScratch? */ UASM_i_MTC0(p, 1, c0_kscratch(), scratch_reg); @@ -351,25 +350,9 @@ static struct work_registers build_get_work_registers(u32 **p) } if (num_possible_cpus() > 1) { -#ifdef CONFIG_MIPS_PGD_C0_CONTEXT - smp_processor_id_shift = 51; - smp_processor_id_reg = 20; /* XContext */ - smp_processor_id_sel = 0; -#else -# ifdef CONFIG_32BIT - smp_processor_id_shift = 25; - smp_processor_id_reg = 4; /* Context */ - smp_processor_id_sel = 0; -# endif -# ifdef CONFIG_64BIT - smp_processor_id_shift = 26; - smp_processor_id_reg = 4; /* Context */ - smp_processor_id_sel = 0; -# endif -#endif /* Get smp_processor_id */ - UASM_i_MFC0(p, K0, smp_processor_id_reg, smp_processor_id_sel); - UASM_i_SRL_SAFE(p, K0, K0, smp_processor_id_shift); + UASM_i_CPUID_MFC0(p, K0, SMP_CPUID_REG); + UASM_i_SRL_SAFE(p, K0, K0, SMP_CPUID_REGSHIFT); /* handler_reg_save index in K0 */ UASM_i_SLL(p, K0, K0, ilog2(sizeof(struct tlb_reg_save))); @@ -816,11 +799,11 @@ build_get_pmde64(u32 **p, struct uasm_label **l, struct uasm_reloc **r, } /* No uasm_i_nop needed here, since the next insn doesn't touch TMP. */ -#ifdef CONFIG_MIPS_PGD_C0_CONTEXT if (pgd_reg != -1) { /* pgd is in pgd_reg */ UASM_i_MFC0(p, ptr, c0_kscratch(), pgd_reg); } else { +#if defined(CONFIG_MIPS_PGD_C0_CONTEXT) /* * &pgd << 11 stored in CONTEXT [23..63]. */ @@ -832,30 +815,18 @@ build_get_pmde64(u32 **p, struct uasm_label **l, struct uasm_reloc **r, /* 1 0 1 0 1 << 6 xkphys cached */ uasm_i_ori(p, ptr, ptr, 0x540); uasm_i_drotr(p, ptr, ptr, 11); - } #elif defined(CONFIG_SMP) -# ifdef CONFIG_MIPS_MT_SMTC - /* - * SMTC uses TCBind value as "CPU" index - */ - uasm_i_mfc0(p, ptr, C0_TCBIND); - uasm_i_dsrl_safe(p, ptr, ptr, 19); -# else - /* - * 64 bit SMP running in XKPHYS has smp_processor_id() << 3 - * stored in CONTEXT. - */ - uasm_i_dmfc0(p, ptr, C0_CONTEXT); - uasm_i_dsrl_safe(p, ptr, ptr, 23); -# endif - UASM_i_LA_mostly(p, tmp, pgdc); - uasm_i_daddu(p, ptr, ptr, tmp); - uasm_i_dmfc0(p, tmp, C0_BADVADDR); - uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr); + UASM_i_CPUID_MFC0(p, ptr, SMP_CPUID_REG); + uasm_i_dsrl_safe(p, ptr, ptr, SMP_CPUID_PTRSHIFT); + UASM_i_LA_mostly(p, tmp, pgdc); + uasm_i_daddu(p, ptr, ptr, tmp); + uasm_i_dmfc0(p, tmp, C0_BADVADDR); + uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr); #else - UASM_i_LA_mostly(p, ptr, pgdc); - uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr); + UASM_i_LA_mostly(p, ptr, pgdc); + uasm_i_ld(p, ptr, uasm_rel_lo(pgdc), ptr); #endif + } uasm_l_vmalloc_done(l, *p); @@ -950,31 +921,25 @@ build_get_pgd_vmalloc64(u32 **p, struct uasm_label **l, struct uasm_reloc **r, static void __maybe_unused build_get_pgde32(u32 **p, unsigned int tmp, unsigned int ptr) { - long pgdc = (long)pgd_current; + if (pgd_reg != -1) { + /* pgd is in pgd_reg */ + uasm_i_mfc0(p, ptr, c0_kscratch(), pgd_reg); + uasm_i_mfc0(p, tmp, C0_BADVADDR); /* get faulting address */ + } else { + long pgdc = (long)pgd_current; - /* 32 bit SMP has smp_processor_id() stored in CONTEXT. */ + /* 32 bit SMP has smp_processor_id() stored in CONTEXT. */ #ifdef CONFIG_SMP -#ifdef CONFIG_MIPS_MT_SMTC - /* - * SMTC uses TCBind value as "CPU" index - */ - uasm_i_mfc0(p, ptr, C0_TCBIND); - UASM_i_LA_mostly(p, tmp, pgdc); - uasm_i_srl(p, ptr, ptr, 19); -#else - /* - * smp_processor_id() << 2 is stored in CONTEXT. - */ - uasm_i_mfc0(p, ptr, C0_CONTEXT); - UASM_i_LA_mostly(p, tmp, pgdc); - uasm_i_srl(p, ptr, ptr, 23); -#endif - uasm_i_addu(p, ptr, tmp, ptr); + uasm_i_mfc0(p, ptr, SMP_CPUID_REG); + UASM_i_LA_mostly(p, tmp, pgdc); + uasm_i_srl(p, ptr, ptr, SMP_CPUID_PTRSHIFT); + uasm_i_addu(p, ptr, tmp, ptr); #else - UASM_i_LA_mostly(p, ptr, pgdc); + UASM_i_LA_mostly(p, ptr, pgdc); #endif - uasm_i_mfc0(p, tmp, C0_BADVADDR); /* get faulting address */ - uasm_i_lw(p, ptr, uasm_rel_lo(pgdc), ptr); + uasm_i_mfc0(p, tmp, C0_BADVADDR); /* get faulting address */ + uasm_i_lw(p, ptr, uasm_rel_lo(pgdc), ptr); + } uasm_i_srl(p, tmp, tmp, PGDIR_SHIFT); /* get pgd only bits */ uasm_i_sll(p, tmp, tmp, PGD_T_LOG2); uasm_i_addu(p, ptr, ptr, tmp); /* add in pgd offset */ @@ -1346,95 +1311,100 @@ static void build_r4000_tlb_refill_handler(void) * need three, with the second nop'ed and the third being * unused. */ - /* Loongson2 ebase is different than r4k, we have more space */ -#if defined(CONFIG_32BIT) || defined(CONFIG_CPU_LOONGSON2) - if ((p - tlb_handler) > 64) - panic("TLB refill handler space exceeded"); -#else - if (((p - tlb_handler) > (MIPS64_REFILL_INSNS * 2) - 1) - || (((p - tlb_handler) > (MIPS64_REFILL_INSNS * 2) - 3) - && uasm_insn_has_bdelay(relocs, - tlb_handler + MIPS64_REFILL_INSNS - 3))) - panic("TLB refill handler space exceeded"); -#endif - - /* - * Now fold the handler in the TLB refill handler space. - */ -#if defined(CONFIG_32BIT) || defined(CONFIG_CPU_LOONGSON2) - f = final_handler; - /* Simplest case, just copy the handler. */ - uasm_copy_handler(relocs, labels, tlb_handler, p, f); - final_len = p - tlb_handler; -#else /* CONFIG_64BIT */ - f = final_handler + MIPS64_REFILL_INSNS; - if ((p - tlb_handler) <= MIPS64_REFILL_INSNS) { - /* Just copy the handler. */ - uasm_copy_handler(relocs, labels, tlb_handler, p, f); - final_len = p - tlb_handler; - } else { -#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT - const enum label_id ls = label_tlb_huge_update; -#else - const enum label_id ls = label_vmalloc; -#endif - u32 *split; - int ov = 0; - int i; - - for (i = 0; i < ARRAY_SIZE(labels) && labels[i].lab != ls; i++) - ; - BUG_ON(i == ARRAY_SIZE(labels)); - split = labels[i].addr; - - /* - * See if we have overflown one way or the other. - */ - if (split > tlb_handler + MIPS64_REFILL_INSNS || - split < p - MIPS64_REFILL_INSNS) - ov = 1; - - if (ov) { + switch (boot_cpu_type()) { + default: + if (sizeof(long) == 4) { + case CPU_LOONGSON2: + /* Loongson2 ebase is different than r4k, we have more space */ + if ((p - tlb_handler) > 64) + panic("TLB refill handler space exceeded"); /* - * Split two instructions before the end. One - * for the branch and one for the instruction - * in the delay slot. + * Now fold the handler in the TLB refill handler space. */ - split = tlb_handler + MIPS64_REFILL_INSNS - 2; - + f = final_handler; + /* Simplest case, just copy the handler. */ + uasm_copy_handler(relocs, labels, tlb_handler, p, f); + final_len = p - tlb_handler; + break; + } else { + if (((p - tlb_handler) > (MIPS64_REFILL_INSNS * 2) - 1) + || (((p - tlb_handler) > (MIPS64_REFILL_INSNS * 2) - 3) + && uasm_insn_has_bdelay(relocs, + tlb_handler + MIPS64_REFILL_INSNS - 3))) + panic("TLB refill handler space exceeded"); /* - * If the branch would fall in a delay slot, - * we must back up an additional instruction - * so that it is no longer in a delay slot. + * Now fold the handler in the TLB refill handler space. */ - if (uasm_insn_has_bdelay(relocs, split - 1)) - split--; - } - /* Copy first part of the handler. */ - uasm_copy_handler(relocs, labels, tlb_handler, split, f); - f += split - tlb_handler; - - if (ov) { - /* Insert branch. */ - uasm_l_split(&l, final_handler); - uasm_il_b(&f, &r, label_split); - if (uasm_insn_has_bdelay(relocs, split)) - uasm_i_nop(&f); - else { - uasm_copy_handler(relocs, labels, - split, split + 1, f); - uasm_move_labels(labels, f, f + 1, -1); - f++; - split++; + f = final_handler + MIPS64_REFILL_INSNS; + if ((p - tlb_handler) <= MIPS64_REFILL_INSNS) { + /* Just copy the handler. */ + uasm_copy_handler(relocs, labels, tlb_handler, p, f); + final_len = p - tlb_handler; + } else { +#ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT + const enum label_id ls = label_tlb_huge_update; +#else + const enum label_id ls = label_vmalloc; +#endif + u32 *split; + int ov = 0; + int i; + + for (i = 0; i < ARRAY_SIZE(labels) && labels[i].lab != ls; i++) + ; + BUG_ON(i == ARRAY_SIZE(labels)); + split = labels[i].addr; + + /* + * See if we have overflown one way or the other. + */ + if (split > tlb_handler + MIPS64_REFILL_INSNS || + split < p - MIPS64_REFILL_INSNS) + ov = 1; + + if (ov) { + /* + * Split two instructions before the end. One + * for the branch and one for the instruction + * in the delay slot. + */ + split = tlb_handler + MIPS64_REFILL_INSNS - 2; + + /* + * If the branch would fall in a delay slot, + * we must back up an additional instruction + * so that it is no longer in a delay slot. + */ + if (uasm_insn_has_bdelay(relocs, split - 1)) + split--; + } + /* Copy first part of the handler. */ + uasm_copy_handler(relocs, labels, tlb_handler, split, f); + f += split - tlb_handler; + + if (ov) { + /* Insert branch. */ + uasm_l_split(&l, final_handler); + uasm_il_b(&f, &r, label_split); + if (uasm_insn_has_bdelay(relocs, split)) + uasm_i_nop(&f); + else { + uasm_copy_handler(relocs, labels, + split, split + 1, f); + uasm_move_labels(labels, f, f + 1, -1); + f++; + split++; + } + } + + /* Copy the rest of the handler. */ + uasm_copy_handler(relocs, labels, split, p, final_handler); + final_len = (f - (final_handler + MIPS64_REFILL_INSNS)) + + (p - split); } } - - /* Copy the rest of the handler. */ - uasm_copy_handler(relocs, labels, split, p, final_handler); - final_len = (f - (final_handler + MIPS64_REFILL_INSNS)) + - (p - split); + break; } -#endif /* CONFIG_64BIT */ uasm_resolve_relocs(relocs, labels); pr_debug("Wrote TLB refill handler (%u instructions).\n", @@ -1448,28 +1418,30 @@ static void build_r4000_tlb_refill_handler(void) extern u32 handle_tlbl[], handle_tlbl_end[]; extern u32 handle_tlbs[], handle_tlbs_end[]; extern u32 handle_tlbm[], handle_tlbm_end[]; - -#ifdef CONFIG_MIPS_PGD_C0_CONTEXT extern u32 tlbmiss_handler_setup_pgd[], tlbmiss_handler_setup_pgd_end[]; -static void build_r4000_setup_pgd(void) +static void build_setup_pgd(void) { const int a0 = 4; - const int a1 = 5; + const int __maybe_unused a1 = 5; + const int __maybe_unused a2 = 6; u32 *p = tlbmiss_handler_setup_pgd; const int tlbmiss_handler_setup_pgd_size = tlbmiss_handler_setup_pgd_end - tlbmiss_handler_setup_pgd; - struct uasm_label *l = labels; - struct uasm_reloc *r = relocs; +#ifndef CONFIG_MIPS_PGD_C0_CONTEXT + long pgdc = (long)pgd_current; +#endif memset(tlbmiss_handler_setup_pgd, 0, tlbmiss_handler_setup_pgd_size * sizeof(tlbmiss_handler_setup_pgd[0])); memset(labels, 0, sizeof(labels)); memset(relocs, 0, sizeof(relocs)); - pgd_reg = allocate_kscratch(); - +#ifdef CONFIG_MIPS_PGD_C0_CONTEXT if (pgd_reg == -1) { + struct uasm_label *l = labels; + struct uasm_reloc *r = relocs; + /* PGD << 11 in c0_Context */ /* * If it is a ckseg0 address, convert to a physical @@ -1491,6 +1463,26 @@ static void build_r4000_setup_pgd(void) uasm_i_jr(&p, 31); UASM_i_MTC0(&p, a0, c0_kscratch(), pgd_reg); } +#else +#ifdef CONFIG_SMP + /* Save PGD to pgd_current[smp_processor_id()] */ + UASM_i_CPUID_MFC0(&p, a1, SMP_CPUID_REG); + UASM_i_SRL_SAFE(&p, a1, a1, SMP_CPUID_PTRSHIFT); + UASM_i_LA_mostly(&p, a2, pgdc); + UASM_i_ADDU(&p, a2, a2, a1); + UASM_i_SW(&p, a0, uasm_rel_lo(pgdc), a2); +#else + UASM_i_LA_mostly(&p, a2, pgdc); + UASM_i_SW(&p, a0, uasm_rel_lo(pgdc), a2); +#endif /* SMP */ + uasm_i_jr(&p, 31); + + /* if pgd_reg is allocated, save PGD also to scratch register */ + if (pgd_reg != -1) + UASM_i_MTC0(&p, a0, c0_kscratch(), pgd_reg); + else + uasm_i_nop(&p); +#endif if (p >= tlbmiss_handler_setup_pgd_end) panic("tlbmiss_handler_setup_pgd space exceeded"); @@ -1501,7 +1493,6 @@ static void build_r4000_setup_pgd(void) dump_handler("tlbmiss_handler", tlbmiss_handler_setup_pgd, tlbmiss_handler_setup_pgd_size); } -#endif static void iPTE_LW(u32 **p, unsigned int pte, unsigned int ptr) @@ -2194,10 +2185,8 @@ static void flush_tlb_handlers(void) (unsigned long)handle_tlbs_end); local_flush_icache_range((unsigned long)handle_tlbm, (unsigned long)handle_tlbm_end); -#ifdef CONFIG_MIPS_PGD_C0_CONTEXT local_flush_icache_range((unsigned long)tlbmiss_handler_setup_pgd, (unsigned long)tlbmiss_handler_setup_pgd_end); -#endif } void build_tlb_refill_handler(void) @@ -2229,6 +2218,7 @@ void build_tlb_refill_handler(void) if (!run_once) { if (!cpu_has_local_ebase) build_r3000_tlb_refill_handler(); + build_setup_pgd(); build_r3000_tlb_load_handler(); build_r3000_tlb_store_handler(); build_r3000_tlb_modify_handler(); @@ -2252,9 +2242,7 @@ void build_tlb_refill_handler(void) default: if (!run_once) { scratch_reg = allocate_kscratch(); -#ifdef CONFIG_MIPS_PGD_C0_CONTEXT - build_r4000_setup_pgd(); -#endif + build_setup_pgd(); build_r4000_tlb_load_handler(); build_r4000_tlb_store_handler(); build_r4000_tlb_modify_handler(); |