diff options
Diffstat (limited to 'arch/powerpc/mm/book3s64/hash_utils.c')
| -rw-r--r-- | arch/powerpc/mm/book3s64/hash_utils.c | 887 |
1 files changed, 684 insertions, 203 deletions
diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c index 468169e33c86..9dc5889d6ecb 100644 --- a/arch/powerpc/mm/book3s64/hash_utils.c +++ b/arch/powerpc/mm/book3s64/hash_utils.c @@ -7,7 +7,7 @@ * * SMP scalability work: * Copyright (C) 2001 Anton Blanchard <anton@au.ibm.com>, IBM - * + * * Module name: htab.c * * Description: @@ -36,25 +36,30 @@ #include <linux/hugetlb.h> #include <linux/cpu.h> #include <linux/pgtable.h> +#include <linux/debugfs.h> +#include <linux/random.h> +#include <linux/elf-randomize.h> +#include <linux/of_fdt.h> +#include <linux/kfence.h> -#include <asm/debugfs.h> +#include <asm/interrupt.h> #include <asm/processor.h> #include <asm/mmu.h> #include <asm/mmu_context.h> #include <asm/page.h> +#include <asm/pgalloc.h> #include <asm/types.h> #include <linux/uaccess.h> #include <asm/machdep.h> -#include <asm/prom.h> #include <asm/io.h> #include <asm/eeh.h> #include <asm/tlb.h> #include <asm/cacheflush.h> #include <asm/cputable.h> #include <asm/sections.h> -#include <asm/copro.h> +#include <asm/spu.h> #include <asm/udbg.h> -#include <asm/code-patching.h> +#include <asm/text-patching.h> #include <asm/fadump.h> #include <asm/firmware.h> #include <asm/tm.h> @@ -63,6 +68,7 @@ #include <asm/pte-walk.h> #include <asm/asm-prototypes.h> #include <asm/ultravisor.h> +#include <asm/kfence.h> #include <mm/mmu_decl.h> @@ -98,8 +104,6 @@ */ static unsigned long _SDR1; -struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT]; -EXPORT_SYMBOL_GPL(mmu_psize_defs); u8 hpte_page_sizes[1 << LP_BITS]; EXPORT_SYMBOL_GPL(hpte_page_sizes); @@ -112,9 +116,7 @@ int mmu_linear_psize = MMU_PAGE_4K; EXPORT_SYMBOL_GPL(mmu_linear_psize); int mmu_virtual_psize = MMU_PAGE_4K; int mmu_vmalloc_psize = MMU_PAGE_4K; -#ifdef CONFIG_SPARSEMEM_VMEMMAP -int mmu_vmemmap_psize = MMU_PAGE_4K; -#endif +EXPORT_SYMBOL_GPL(mmu_vmalloc_psize); int mmu_io_psize = MMU_PAGE_4K; int mmu_kernel_ssize = MMU_SEGSIZE_256M; EXPORT_SYMBOL_GPL(mmu_kernel_ssize); @@ -124,12 +126,7 @@ EXPORT_SYMBOL_GPL(mmu_slb_size); #ifdef CONFIG_PPC_64K_PAGES int mmu_ci_restrictions; #endif -#ifdef CONFIG_DEBUG_PAGEALLOC -static u8 *linear_map_hash_slots; -static unsigned long linear_map_hash_count; -static DEFINE_SPINLOCK(linear_map_hash_lock); -#endif /* CONFIG_DEBUG_PAGEALLOC */ -struct mmu_hash_ops mmu_hash_ops; +struct mmu_hash_ops mmu_hash_ops __ro_after_init; EXPORT_SYMBOL(mmu_hash_ops); /* @@ -173,6 +170,376 @@ static struct mmu_psize_def mmu_psize_defaults_gp[] = { }, }; +static inline void tlbiel_hash_set_isa206(unsigned int set, unsigned int is) +{ + unsigned long rb; + + rb = (set << PPC_BITLSHIFT(51)) | (is << PPC_BITLSHIFT(53)); + + asm volatile("tlbiel %0" : : "r" (rb)); +} + +/* + * tlbiel instruction for hash, set invalidation + * i.e., r=1 and is=01 or is=10 or is=11 + */ +static __always_inline void tlbiel_hash_set_isa300(unsigned int set, unsigned int is, + unsigned int pid, + unsigned int ric, unsigned int prs) +{ + unsigned long rb; + unsigned long rs; + unsigned int r = 0; /* hash format */ + + rb = (set << PPC_BITLSHIFT(51)) | (is << PPC_BITLSHIFT(53)); + rs = ((unsigned long)pid << PPC_BITLSHIFT(31)); + + asm volatile(PPC_TLBIEL(%0, %1, %2, %3, %4) + : : "r"(rb), "r"(rs), "i"(ric), "i"(prs), "i"(r) + : "memory"); +} + + +static void tlbiel_all_isa206(unsigned int num_sets, unsigned int is) +{ + unsigned int set; + + asm volatile("ptesync": : :"memory"); + + for (set = 0; set < num_sets; set++) + tlbiel_hash_set_isa206(set, is); + + ppc_after_tlbiel_barrier(); +} + +static void tlbiel_all_isa300(unsigned int num_sets, unsigned int is) +{ + unsigned int set; + + asm volatile("ptesync": : :"memory"); + + /* + * Flush the partition table cache if this is HV mode. + */ + if (early_cpu_has_feature(CPU_FTR_HVMODE)) + tlbiel_hash_set_isa300(0, is, 0, 2, 0); + + /* + * Now invalidate the process table cache. UPRT=0 HPT modes (what + * current hardware implements) do not use the process table, but + * add the flushes anyway. + * + * From ISA v3.0B p. 1078: + * The following forms are invalid. + * * PRS=1, R=0, and RIC!=2 (The only process-scoped + * HPT caching is of the Process Table.) + */ + tlbiel_hash_set_isa300(0, is, 0, 2, 1); + + /* + * Then flush the sets of the TLB proper. Hash mode uses + * partition scoped TLB translations, which may be flushed + * in !HV mode. + */ + for (set = 0; set < num_sets; set++) + tlbiel_hash_set_isa300(set, is, 0, 0, 0); + + ppc_after_tlbiel_barrier(); + + asm volatile(PPC_ISA_3_0_INVALIDATE_ERAT "; isync" : : :"memory"); +} + +void hash__tlbiel_all(unsigned int action) +{ + unsigned int is; + + switch (action) { + case TLB_INVAL_SCOPE_GLOBAL: + is = 3; + break; + case TLB_INVAL_SCOPE_LPID: + is = 2; + break; + default: + BUG(); + } + + if (early_cpu_has_feature(CPU_FTR_ARCH_300)) + tlbiel_all_isa300(POWER9_TLB_SETS_HASH, is); + else if (early_cpu_has_feature(CPU_FTR_ARCH_207S)) + tlbiel_all_isa206(POWER8_TLB_SETS, is); + else if (early_cpu_has_feature(CPU_FTR_ARCH_206)) + tlbiel_all_isa206(POWER7_TLB_SETS, is); + else + WARN(1, "%s called on pre-POWER7 CPU\n", __func__); +} + +#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KFENCE) +static void kernel_map_linear_page(unsigned long vaddr, unsigned long idx, + u8 *slots, raw_spinlock_t *lock) +{ + unsigned long hash; + unsigned long vsid = get_kernel_vsid(vaddr, mmu_kernel_ssize); + unsigned long vpn = hpt_vpn(vaddr, vsid, mmu_kernel_ssize); + unsigned long mode = htab_convert_pte_flags(pgprot_val(PAGE_KERNEL), HPTE_USE_KERNEL_KEY); + long ret; + + hash = hpt_hash(vpn, PAGE_SHIFT, mmu_kernel_ssize); + + /* Don't create HPTE entries for bad address */ + if (!vsid) + return; + + if (slots[idx] & 0x80) + return; + + ret = hpte_insert_repeating(hash, vpn, __pa(vaddr), mode, + HPTE_V_BOLTED, + mmu_linear_psize, mmu_kernel_ssize); + + BUG_ON (ret < 0); + raw_spin_lock(lock); + BUG_ON(slots[idx] & 0x80); + slots[idx] = ret | 0x80; + raw_spin_unlock(lock); +} + +static void kernel_unmap_linear_page(unsigned long vaddr, unsigned long idx, + u8 *slots, raw_spinlock_t *lock) +{ + unsigned long hash, hslot, slot; + unsigned long vsid = get_kernel_vsid(vaddr, mmu_kernel_ssize); + unsigned long vpn = hpt_vpn(vaddr, vsid, mmu_kernel_ssize); + + hash = hpt_hash(vpn, PAGE_SHIFT, mmu_kernel_ssize); + raw_spin_lock(lock); + if (!(slots[idx] & 0x80)) { + raw_spin_unlock(lock); + return; + } + hslot = slots[idx] & 0x7f; + slots[idx] = 0; + raw_spin_unlock(lock); + if (hslot & _PTEIDX_SECONDARY) + hash = ~hash; + slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; + slot += hslot & _PTEIDX_GROUP_IX; + mmu_hash_ops.hpte_invalidate(slot, vpn, mmu_linear_psize, + mmu_linear_psize, + mmu_kernel_ssize, 0); +} +#endif + +static inline bool hash_supports_debug_pagealloc(void) +{ + unsigned long max_hash_count = ppc64_rma_size / 4; + unsigned long linear_map_count = memblock_end_of_DRAM() >> PAGE_SHIFT; + + if (!debug_pagealloc_enabled() || linear_map_count > max_hash_count) + return false; + return true; +} + +#ifdef CONFIG_DEBUG_PAGEALLOC +static u8 *linear_map_hash_slots; +static unsigned long linear_map_hash_count; +static DEFINE_RAW_SPINLOCK(linear_map_hash_lock); +static __init void hash_debug_pagealloc_alloc_slots(void) +{ + if (!hash_supports_debug_pagealloc()) + return; + + linear_map_hash_count = memblock_end_of_DRAM() >> PAGE_SHIFT; + linear_map_hash_slots = memblock_alloc_try_nid( + linear_map_hash_count, 1, MEMBLOCK_LOW_LIMIT, + ppc64_rma_size, NUMA_NO_NODE); + if (!linear_map_hash_slots) + panic("%s: Failed to allocate %lu bytes max_addr=%pa\n", + __func__, linear_map_hash_count, &ppc64_rma_size); +} + +static inline void hash_debug_pagealloc_add_slot(phys_addr_t paddr, + int slot) +{ + if (!debug_pagealloc_enabled() || !linear_map_hash_count) + return; + if ((paddr >> PAGE_SHIFT) < linear_map_hash_count) + linear_map_hash_slots[paddr >> PAGE_SHIFT] = slot | 0x80; +} + +static int hash_debug_pagealloc_map_pages(struct page *page, int numpages, + int enable) +{ + unsigned long flags, vaddr, lmi; + int i; + + if (!debug_pagealloc_enabled() || !linear_map_hash_count) + return 0; + + local_irq_save(flags); + for (i = 0; i < numpages; i++, page++) { + vaddr = (unsigned long)page_address(page); + lmi = __pa(vaddr) >> PAGE_SHIFT; + if (lmi >= linear_map_hash_count) + continue; + if (enable) + kernel_map_linear_page(vaddr, lmi, + linear_map_hash_slots, &linear_map_hash_lock); + else + kernel_unmap_linear_page(vaddr, lmi, + linear_map_hash_slots, &linear_map_hash_lock); + } + local_irq_restore(flags); + return 0; +} + +#else /* CONFIG_DEBUG_PAGEALLOC */ +static inline void hash_debug_pagealloc_alloc_slots(void) {} +static inline void hash_debug_pagealloc_add_slot(phys_addr_t paddr, int slot) {} +static int __maybe_unused +hash_debug_pagealloc_map_pages(struct page *page, int numpages, int enable) +{ + return 0; +} +#endif /* CONFIG_DEBUG_PAGEALLOC */ + +#ifdef CONFIG_KFENCE +static u8 *linear_map_kf_hash_slots; +static unsigned long linear_map_kf_hash_count; +static DEFINE_RAW_SPINLOCK(linear_map_kf_hash_lock); + +static phys_addr_t kfence_pool; + +static __init void hash_kfence_alloc_pool(void) +{ + if (!kfence_early_init_enabled()) + goto err; + + /* allocate linear map for kfence within RMA region */ + linear_map_kf_hash_count = KFENCE_POOL_SIZE >> PAGE_SHIFT; + linear_map_kf_hash_slots = memblock_alloc_try_nid( + linear_map_kf_hash_count, 1, + MEMBLOCK_LOW_LIMIT, ppc64_rma_size, + NUMA_NO_NODE); + if (!linear_map_kf_hash_slots) { + pr_err("%s: memblock for linear map (%lu) failed\n", __func__, + linear_map_kf_hash_count); + goto err; + } + + /* allocate kfence pool early */ + kfence_pool = memblock_phys_alloc_range(KFENCE_POOL_SIZE, PAGE_SIZE, + MEMBLOCK_LOW_LIMIT, MEMBLOCK_ALLOC_ANYWHERE); + if (!kfence_pool) { + pr_err("%s: memblock for kfence pool (%lu) failed\n", __func__, + KFENCE_POOL_SIZE); + memblock_free(linear_map_kf_hash_slots, + linear_map_kf_hash_count); + linear_map_kf_hash_count = 0; + goto err; + } + memblock_mark_nomap(kfence_pool, KFENCE_POOL_SIZE); + + return; +err: + pr_info("Disabling kfence\n"); + disable_kfence(); +} + +static __init void hash_kfence_map_pool(void) +{ + unsigned long kfence_pool_start, kfence_pool_end; + unsigned long prot = pgprot_val(PAGE_KERNEL); + unsigned int pshift = mmu_psize_defs[mmu_linear_psize].shift; + + if (!kfence_pool) + return; + + kfence_pool_start = (unsigned long) __va(kfence_pool); + kfence_pool_end = kfence_pool_start + KFENCE_POOL_SIZE; + __kfence_pool = (char *) kfence_pool_start; + BUG_ON(htab_bolt_mapping(kfence_pool_start, kfence_pool_end, + kfence_pool, prot, mmu_linear_psize, + mmu_kernel_ssize)); + update_page_count(mmu_linear_psize, KFENCE_POOL_SIZE >> pshift); + memblock_clear_nomap(kfence_pool, KFENCE_POOL_SIZE); +} + +static inline void hash_kfence_add_slot(phys_addr_t paddr, int slot) +{ + unsigned long vaddr = (unsigned long) __va(paddr); + unsigned long lmi = (vaddr - (unsigned long)__kfence_pool) + >> PAGE_SHIFT; + + if (!kfence_pool) + return; + BUG_ON(!is_kfence_address((void *)vaddr)); + BUG_ON(lmi >= linear_map_kf_hash_count); + linear_map_kf_hash_slots[lmi] = slot | 0x80; +} + +static int hash_kfence_map_pages(struct page *page, int numpages, int enable) +{ + unsigned long flags, vaddr, lmi; + int i; + + WARN_ON_ONCE(!linear_map_kf_hash_count); + local_irq_save(flags); + for (i = 0; i < numpages; i++, page++) { + vaddr = (unsigned long)page_address(page); + lmi = (vaddr - (unsigned long)__kfence_pool) >> PAGE_SHIFT; + + /* Ideally this should never happen */ + if (lmi >= linear_map_kf_hash_count) { + WARN_ON_ONCE(1); + continue; + } + + if (enable) + kernel_map_linear_page(vaddr, lmi, + linear_map_kf_hash_slots, + &linear_map_kf_hash_lock); + else + kernel_unmap_linear_page(vaddr, lmi, + linear_map_kf_hash_slots, + &linear_map_kf_hash_lock); + } + local_irq_restore(flags); + return 0; +} +#else +static inline void hash_kfence_alloc_pool(void) {} +static inline void hash_kfence_map_pool(void) {} +static inline void hash_kfence_add_slot(phys_addr_t paddr, int slot) {} +static int __maybe_unused +hash_kfence_map_pages(struct page *page, int numpages, int enable) +{ + return 0; +} +#endif + +#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KFENCE) +int hash__kernel_map_pages(struct page *page, int numpages, int enable) +{ + void *vaddr = page_address(page); + + if (is_kfence_address(vaddr)) + return hash_kfence_map_pages(page, numpages, enable); + else + return hash_debug_pagealloc_map_pages(page, numpages, enable); +} + +static void hash_linear_map_add_slot(phys_addr_t paddr, int slot) +{ + if (is_kfence_address(__va(paddr))) + hash_kfence_add_slot(paddr, slot); + else + hash_debug_pagealloc_add_slot(paddr, slot); +} +#else +static void hash_linear_map_add_slot(phys_addr_t paddr, int slot) {} +#endif + /* * 'R' and 'C' update notes: * - Under pHyp or KVM, the updatepp path will not set C, thus it *will* @@ -186,7 +553,7 @@ static struct mmu_psize_def mmu_psize_defaults_gp[] = { * - We make sure R is always set and never lost * - C is _PAGE_DIRTY, and *should* always be set for a writeable mapping */ -unsigned long htab_convert_pte_flags(unsigned long pteflags) +unsigned long htab_convert_pte_flags(unsigned long pteflags, unsigned long flags) { unsigned long rflags = 0; @@ -210,9 +577,16 @@ unsigned long htab_convert_pte_flags(unsigned long pteflags) else rflags |= 0x3; } + VM_WARN_ONCE(!(pteflags & _PAGE_RWX), "no-access mapping request"); } else { if (pteflags & _PAGE_RWX) rflags |= 0x2; + /* + * We should never hit this in normal fault handling because + * a permission check (check_pte_access()) will bubble this + * to higher level linux handler even for PAGE_NONE. + */ + VM_WARN_ONCE(!(pteflags & _PAGE_RWX), "no-access mapping request"); if (!((pteflags & _PAGE_WRITE) && (pteflags & _PAGE_DIRTY))) rflags |= 0x1; } @@ -240,7 +614,7 @@ unsigned long htab_convert_pte_flags(unsigned long pteflags) */ rflags |= HPTE_R_M; - rflags |= pte_to_hpte_pkey_bits(pteflags); + rflags |= pte_to_hpte_pkey_bits(pteflags, flags); return rflags; } @@ -255,13 +629,17 @@ int htab_bolt_mapping(unsigned long vstart, unsigned long vend, shift = mmu_psize_defs[psize].shift; step = 1 << shift; - prot = htab_convert_pte_flags(prot); + prot = htab_convert_pte_flags(prot, HPTE_USE_KERNEL_KEY); DBG("htab_bolt_mapping(%lx..%lx -> %lx (%lx,%d,%d)\n", vstart, vend, pstart, prot, psize, ssize); - for (vaddr = vstart, paddr = pstart; vaddr < vend; - vaddr += step, paddr += step) { + /* Carefully map only the possible range */ + vaddr = ALIGN(vstart, step); + paddr = ALIGN(pstart, step); + vend = ALIGN_DOWN(vend, step); + + for (; vaddr < vend; vaddr += step, paddr += step) { unsigned long hash, hpteg; unsigned long vsid = get_kernel_vsid(vaddr, ssize); unsigned long vpn = hpt_vpn(vaddr, vsid, ssize); @@ -301,7 +679,7 @@ repeat: ssize); if (ret == -1) { /* - * Try to to keep bolted entries in primary. + * Try to keep bolted entries in primary. * Remove non bolted entries and try insert again */ ret = mmu_hash_ops.hpte_remove(hpteg); @@ -320,11 +698,8 @@ repeat: break; cond_resched(); -#ifdef CONFIG_DEBUG_PAGEALLOC - if (debug_pagealloc_enabled() && - (paddr >> PAGE_SHIFT) < linear_map_hash_count) - linear_map_hash_slots[paddr >> PAGE_SHIFT] = ret | 0x80; -#endif /* CONFIG_DEBUG_PAGEALLOC */ + /* add slot info in debug_pagealloc / kfence linear map */ + hash_linear_map_add_slot(paddr, ret); } return ret < 0 ? ret : 0; } @@ -332,7 +707,7 @@ repeat: int htab_remove_mapping(unsigned long vstart, unsigned long vend, int psize, int ssize) { - unsigned long vaddr; + unsigned long vaddr, time_limit; unsigned int step, shift; int rc; int ret = 0; @@ -343,8 +718,21 @@ int htab_remove_mapping(unsigned long vstart, unsigned long vend, if (!mmu_hash_ops.hpte_removebolted) return -ENODEV; - for (vaddr = vstart; vaddr < vend; vaddr += step) { + /* Unmap the full range specificied */ + vaddr = ALIGN_DOWN(vstart, step); + time_limit = jiffies + HZ; + + for (;vaddr < vend; vaddr += step) { rc = mmu_hash_ops.hpte_removebolted(vaddr, psize, ssize); + + /* + * For large number of mappings introduce a cond_resched() + * to prevent softlockup warnings. + */ + if (time_after(jiffies, time_limit)) { + cond_resched(); + time_limit = jiffies + HZ; + } if (rc == -ENOENT) { ret = -ENOENT; continue; @@ -356,7 +744,7 @@ int htab_remove_mapping(unsigned long vstart, unsigned long vend, return ret; } -static bool disable_1tb_segments = false; +static bool disable_1tb_segments __ro_after_init; static int __init parse_disable_1tb_segments(char *p) { @@ -365,6 +753,40 @@ static int __init parse_disable_1tb_segments(char *p) } early_param("disable_1tb_segments", parse_disable_1tb_segments); +bool stress_hpt_enabled __initdata; + +static int __init parse_stress_hpt(char *p) +{ + stress_hpt_enabled = true; + return 0; +} +early_param("stress_hpt", parse_stress_hpt); + +__ro_after_init DEFINE_STATIC_KEY_FALSE(stress_hpt_key); + +/* + * per-CPU array allocated if we enable stress_hpt. + */ +#define STRESS_MAX_GROUPS 16 +struct stress_hpt_struct { + unsigned long last_group[STRESS_MAX_GROUPS]; +}; + +static inline int stress_nr_groups(void) +{ + /* + * LPAR H_REMOVE flushes TLB, so need some number > 1 of entries + * to allow practical forward progress. Bare metal returns 1, which + * seems to help uncover more bugs. + */ + if (firmware_has_feature(FW_FEATURE_LPAR)) + return STRESS_MAX_GROUPS; + else + return 1; +} + +static struct stress_hpt_struct *stress_hpt_struct; + static int __init htab_dt_scan_seg_sizes(unsigned long node, const char *uname, int depth, void *data) @@ -533,7 +955,7 @@ static int __init htab_dt_scan_hugepage_blocks(unsigned long node, block_size = be64_to_cpu(addr_prop[1]); if (block_size != (16 * GB)) return 0; - printk(KERN_INFO "Huge page(16GB) memory: " + pr_info("Huge page(16GB) memory: " "addr = 0x%lX size = 0x%lX pages = %d\n", phys_addr, block_size, expected_pages); if (phys_addr + block_size * expected_pages <= memblock_end_of_DRAM()) { @@ -544,7 +966,7 @@ static int __init htab_dt_scan_hugepage_blocks(unsigned long node, } #endif /* CONFIG_HUGETLB_PAGE */ -static void mmu_psize_set_default_penc(void) +static void __init mmu_psize_set_default_penc(void) { int bpsize, apsize; for (bpsize = 0; bpsize < MMU_PAGE_COUNT; bpsize++) @@ -554,7 +976,7 @@ static void mmu_psize_set_default_penc(void) #ifdef CONFIG_PPC_64K_PAGES -static bool might_have_hea(void) +static bool __init might_have_hea(void) { /* * The HEA ethernet adapter requires awareness of the @@ -596,7 +1018,7 @@ static void __init htab_scan_page_sizes(void) } #ifdef CONFIG_HUGETLB_PAGE - if (!hugetlb_disabled) { + if (!hugetlb_disabled && !early_radix_enabled() ) { /* Reserve 16G huge page memory sections for huge pages */ of_scan_flat_dt(htab_dt_scan_hugepage_blocks, NULL); } @@ -625,7 +1047,7 @@ static void __init htab_scan_page_sizes(void) * low-order N bits as the encoding for the 2^(12+N) byte page size * (if it exists). */ -static void init_hpte_page_sizes(void) +static void __init init_hpte_page_sizes(void) { long int ap, bp; long int shift, penc; @@ -658,16 +1080,15 @@ static void __init htab_init_page_sizes(void) bool aligned = true; init_hpte_page_sizes(); - if (!debug_pagealloc_enabled()) { + if (!hash_supports_debug_pagealloc() && !kfence_early_init_enabled()) { /* * Pick a size for the linear mapping. Currently, we only * support 16M, 1M and 4K which is the default */ - if (IS_ENABLED(STRICT_KERNEL_RWX) && + if (IS_ENABLED(CONFIG_STRICT_KERNEL_RWX) && (unsigned long)_stext % 0x1000000) { if (mmu_psize_defs[MMU_PAGE_16M].shift) - pr_warn("Kernel not 16M aligned, " - "disabling 16M linear map alignment"); + pr_warn("Kernel not 16M aligned, disabling 16M linear map alignment\n"); aligned = false; } @@ -717,7 +1138,7 @@ static void __init htab_init_page_sizes(void) mmu_vmemmap_psize = mmu_virtual_psize; #endif /* CONFIG_SPARSEMEM_VMEMMAP */ - printk(KERN_DEBUG "Page orders: linear mapping = %d, " + pr_info("Page orders: linear mapping = %d, " "virtual = %d, io = %d" #ifdef CONFIG_SPARSEMEM_VMEMMAP ", vmemmap = %d" @@ -788,7 +1209,7 @@ static unsigned long __init htab_get_table_size(void) } #ifdef CONFIG_MEMORY_HOTPLUG -int resize_hpt_for_hotplug(unsigned long new_mem_size) +static int resize_hpt_for_hotplug(unsigned long new_mem_size) { unsigned target_hpt_shift; @@ -816,12 +1237,15 @@ int hash__create_section_mapping(unsigned long start, unsigned long end, int nid, pgprot_t prot) { int rc; + unsigned int pshift = mmu_psize_defs[mmu_linear_psize].shift; if (end >= H_VMALLOC_START) { pr_warn("Outside the supported range\n"); return -1; } + resize_hpt_for_hotplug(memblock_phys_mem_size()); + rc = htab_bolt_mapping(start, end, __pa(start), pgprot_val(prot), mmu_linear_psize, mmu_kernel_ssize); @@ -831,14 +1255,22 @@ int hash__create_section_mapping(unsigned long start, unsigned long end, mmu_kernel_ssize); BUG_ON(rc2 && (rc2 != -ENOENT)); } + update_page_count(mmu_linear_psize, (end - start) >> pshift); return rc; } int hash__remove_section_mapping(unsigned long start, unsigned long end) { + unsigned int pshift = mmu_psize_defs[mmu_linear_psize].shift; + int rc = htab_remove_mapping(start, end, mmu_linear_psize, mmu_kernel_ssize); - WARN_ON(rc < 0); + + if (resize_hpt_for_hotplug(memblock_phys_mem_size()) == -ENOSPC) + pr_warn("Hash collision while resizing HPT\n"); + + if (!rc) + update_page_count(mmu_linear_psize, -((end - start) >> pshift)); return rc; } #endif /* CONFIG_MEMORY_HOTPLUG */ @@ -857,29 +1289,68 @@ static void __init hash_init_partition_table(phys_addr_t hash_table, pr_info("Partition table %p\n", partition_tb); } +void hpt_clear_stress(void); +static struct timer_list stress_hpt_timer; +static void stress_hpt_timer_fn(struct timer_list *timer) +{ + int next_cpu; + + hpt_clear_stress(); + if (!firmware_has_feature(FW_FEATURE_LPAR)) + tlbiel_all(); + + next_cpu = cpumask_next(raw_smp_processor_id(), cpu_online_mask); + if (next_cpu >= nr_cpu_ids) + next_cpu = cpumask_first(cpu_online_mask); + stress_hpt_timer.expires = jiffies + msecs_to_jiffies(10); + add_timer_on(&stress_hpt_timer, next_cpu); +} + static void __init htab_initialize(void) { unsigned long table; unsigned long pteg_count; unsigned long prot; - unsigned long base = 0, size = 0; - struct memblock_region *reg; + phys_addr_t base = 0, size = 0, end, limit = MEMBLOCK_ALLOC_ANYWHERE; + u64 i; + unsigned int pshift = mmu_psize_defs[mmu_linear_psize].shift; DBG(" -> htab_initialize()\n"); + if (firmware_has_feature(FW_FEATURE_LPAR)) + limit = ppc64_rma_size; + if (mmu_has_feature(MMU_FTR_1T_SEGMENT)) { mmu_kernel_ssize = MMU_SEGSIZE_1T; mmu_highuser_ssize = MMU_SEGSIZE_1T; - printk(KERN_INFO "Using 1TB segments\n"); + pr_info("Using 1TB segments\n"); } if (stress_slb_enabled) static_branch_enable(&stress_slb_key); + if (no_slb_preload) + static_branch_enable(&no_slb_preload_key); + + if (stress_hpt_enabled) { + unsigned long tmp; + static_branch_enable(&stress_hpt_key); + // Too early to use nr_cpu_ids, so use NR_CPUS + tmp = memblock_phys_alloc_range(sizeof(struct stress_hpt_struct) * NR_CPUS, + __alignof__(struct stress_hpt_struct), + MEMBLOCK_LOW_LIMIT, limit); + memset((void *)tmp, 0xff, sizeof(struct stress_hpt_struct) * NR_CPUS); + stress_hpt_struct = __va(tmp); + + timer_setup(&stress_hpt_timer, stress_hpt_timer_fn, 0); + stress_hpt_timer.expires = jiffies + msecs_to_jiffies(10); + add_timer(&stress_hpt_timer); + } + /* * Calculate the required size of the htab. We want the number of * PTEGs to equal one half the number of real pages. - */ + */ htab_size_bytes = htab_get_table_size(); pteg_count = htab_size_bytes >> 7; @@ -889,7 +1360,7 @@ static void __init htab_initialize(void) firmware_has_feature(FW_FEATURE_PS3_LV1)) { /* Using a hypervisor which owns the htab */ htab_address = NULL; - _SDR1 = 0; + _SDR1 = 0; #ifdef CONFIG_FA_DUMP /* * If firmware assisted dump is active firmware preserves @@ -901,23 +1372,10 @@ static void __init htab_initialize(void) mmu_hash_ops.hpte_clear_all(); #endif } else { - unsigned long limit = MEMBLOCK_ALLOC_ANYWHERE; - -#ifdef CONFIG_PPC_CELL - /* - * Cell may require the hash table down low when using the - * Axon IOMMU in order to fit the dynamic region over it, see - * comments in cell/iommu.c - */ - if (fdt_subnode_offset(initial_boot_params, 0, "axon") > 0) { - limit = 0x80000000; - pr_info("Hash table forced below 2G for Axon IOMMU\n"); - } -#endif /* CONFIG_PPC_CELL */ table = memblock_phys_alloc_range(htab_size_bytes, htab_size_bytes, - 0, limit); + MEMBLOCK_LOW_LIMIT, limit); if (!table) panic("ERROR: Failed to allocate %pa bytes below %pa\n", &htab_size_bytes, &limit); @@ -942,25 +1400,15 @@ static void __init htab_initialize(void) prot = pgprot_val(PAGE_KERNEL); -#ifdef CONFIG_DEBUG_PAGEALLOC - if (debug_pagealloc_enabled()) { - linear_map_hash_count = memblock_end_of_DRAM() >> PAGE_SHIFT; - linear_map_hash_slots = memblock_alloc_try_nid( - linear_map_hash_count, 1, MEMBLOCK_LOW_LIMIT, - ppc64_rma_size, NUMA_NO_NODE); - if (!linear_map_hash_slots) - panic("%s: Failed to allocate %lu bytes max_addr=%pa\n", - __func__, linear_map_hash_count, &ppc64_rma_size); - } -#endif /* CONFIG_DEBUG_PAGEALLOC */ - + hash_debug_pagealloc_alloc_slots(); + hash_kfence_alloc_pool(); /* create bolted the linear mapping in the hash table */ - for_each_memblock(memory, reg) { - base = (unsigned long)__va(reg->base); - size = reg->size; + for_each_mem_range(i, &base, &end) { + size = end - base; + base = (unsigned long)__va(base); - DBG("creating mapping for region: %lx..%lx (prot: %lx)\n", - base, size, prot); + pr_debug("creating mapping for region: 0x%pa..0x%pa (prot: %lx)\n", + &base, &size, prot); if ((base + size) >= H_VMALLOC_START) { pr_warn("Outside the supported range\n"); @@ -969,7 +1417,10 @@ static void __init htab_initialize(void) BUG_ON(htab_bolt_mapping(base, base + size, __pa(base), prot, mmu_linear_psize, mmu_kernel_ssize)); + + update_page_count(mmu_linear_psize, size >> pshift); } + hash_kfence_map_pool(); memblock_set_current_limit(MEMBLOCK_ALLOC_ANYWHERE); /* @@ -989,6 +1440,8 @@ static void __init htab_initialize(void) BUG_ON(htab_bolt_mapping(tce_alloc_start, tce_alloc_end, __pa(tce_alloc_start), prot, mmu_linear_psize, mmu_kernel_ssize)); + update_page_count(mmu_linear_psize, + (tce_alloc_end - tce_alloc_start) >> pshift); } @@ -1043,10 +1496,6 @@ void __init hash__early_init_mmu(void) __pmd_table_size = H_PMD_TABLE_SIZE; __pud_table_size = H_PUD_TABLE_SIZE; __pgd_table_size = H_PGD_TABLE_SIZE; - /* - * 4k use hugepd format, so for hash set then to - * zero - */ __pmd_val_bits = HASH_PMD_VAL_BITS; __pud_val_bits = HASH_PUD_VAL_BITS; __pgd_val_bits = HASH_PGD_VAL_BITS; @@ -1068,7 +1517,7 @@ void __init hash__early_init_mmu(void) ps3_early_mm_init(); else if (firmware_has_feature(FW_FEATURE_LPAR)) hpte_init_pseries(); - else if (IS_ENABLED(CONFIG_PPC_NATIVE)) + else if (IS_ENABLED(CONFIG_PPC_HASH_MMU_NATIVE)) hpte_init_native(); if (!mmu_hash_ops.hpte_insert) @@ -1111,6 +1560,11 @@ void hash__early_init_mmu_secondary(void) if (cpu_has_feature(CPU_FTR_ARCH_206) && cpu_has_feature(CPU_FTR_HVMODE)) tlbiel_all(); + +#ifdef CONFIG_PPC_MEM_KEYS + if (mmu_has_feature(MMU_FTR_PKEY)) + mtspr(SPRN_UAMOR, default_uamor); +#endif } #endif /* CONFIG_SMP */ @@ -1119,25 +1573,25 @@ void hash__early_init_mmu_secondary(void) */ unsigned int hash_page_do_lazy_icache(unsigned int pp, pte_t pte, int trap) { - struct page *page; + struct folio *folio; if (!pfn_valid(pte_pfn(pte))) return pp; - page = pte_page(pte); + folio = page_folio(pte_page(pte)); /* page is dirty */ - if (!test_bit(PG_arch_1, &page->flags) && !PageReserved(page)) { - if (trap == 0x400) { - flush_dcache_icache_page(page); - set_bit(PG_arch_1, &page->flags); + if (!test_bit(PG_dcache_clean, &folio->flags.f) && + !folio_test_reserved(folio)) { + if (trap == INTERRUPT_INST_STORAGE) { + flush_dcache_icache_folio(folio); + set_bit(PG_dcache_clean, &folio->flags.f); } else pp |= HPTE_R_N; } return pp; } -#ifdef CONFIG_PPC_MM_SLICES static unsigned int get_paca_psize(unsigned long addr) { unsigned char *psizes; @@ -1154,12 +1608,6 @@ static unsigned int get_paca_psize(unsigned long addr) return (psizes[index >> 1] >> (mask_index * 4)) & 0xF; } -#else -unsigned int get_paca_psize(unsigned long addr) -{ - return get_paca()->mm_ctx_user_psize; -} -#endif /* * Demote a segment to using 4k pages. @@ -1171,7 +1619,9 @@ void demote_segment_4k(struct mm_struct *mm, unsigned long addr) if (get_slice_psize(mm, addr) == MMU_PAGE_4K) return; slice_set_range_psize(mm, addr, 1, MMU_PAGE_4K); - copro_flush_all_slbs(mm); +#ifdef CONFIG_SPU_BASE + spu_flush_all_slbs(mm); +#endif if ((get_paca_psize(addr) != MMU_PAGE_4K) && (current->mm == mm)) { copy_mm_to_paca(mm); @@ -1216,7 +1666,7 @@ static int subpage_protection(struct mm_struct *mm, unsigned long ea) spp >>= 30 - 2 * ((ea >> 12) & 0xf); /* - * 0 -> full premission + * 0 -> full permission * 1 -> Read only * 2 -> no access. * We return the flag that need to be cleared. @@ -1272,7 +1722,6 @@ int hash_page_mm(struct mm_struct *mm, unsigned long ea, unsigned long flags) { bool is_thp; - enum ctx_state prev_state = exception_enter(); pgd_t *pgdir; unsigned long vsid; pte_t *ptep; @@ -1301,12 +1750,14 @@ int hash_page_mm(struct mm_struct *mm, unsigned long ea, vsid = get_kernel_vsid(ea, mmu_kernel_ssize); psize = mmu_vmalloc_psize; ssize = mmu_kernel_ssize; + flags |= HPTE_USE_KERNEL_KEY; break; case IO_REGION_ID: vsid = get_kernel_vsid(ea, mmu_kernel_ssize); psize = mmu_io_psize; ssize = mmu_kernel_ssize; + flags |= HPTE_USE_KERNEL_KEY; break; default: /* @@ -1356,6 +1807,13 @@ int hash_page_mm(struct mm_struct *mm, unsigned long ea, goto bail; } + if (IS_ENABLED(CONFIG_PPC_4K_PAGES) && !radix_enabled()) { + if (hugeshift == PMD_SHIFT && psize == MMU_PAGE_16M) + hugeshift = mmu_psize_defs[MMU_PAGE_16M].shift; + if (hugeshift == PUD_SHIFT && psize == MMU_PAGE_16G) + hugeshift = mmu_psize_defs[MMU_PAGE_16G].shift; + } + /* * Add _PAGE_PRESENT to the required access perm. If there are parallel * updates to the pte that can possibly clear _PAGE_PTE, catch that too. @@ -1428,11 +1886,13 @@ int hash_page_mm(struct mm_struct *mm, unsigned long ea, * in vmalloc space, so switch vmalloc * to 4k pages */ - printk(KERN_ALERT "Reducing vmalloc segment " + pr_alert("Reducing vmalloc segment " "to 4kB pages because of " "non-cacheable mapping\n"); psize = mmu_vmalloc_psize = MMU_PAGE_4K; - copro_flush_all_slbs(mm); +#ifdef CONFIG_SPU_BASE + spu_flush_all_slbs(mm); +#endif } } @@ -1472,7 +1932,6 @@ int hash_page_mm(struct mm_struct *mm, unsigned long ea, DBG_LOW(" -> rc=%d\n", rc); bail: - exception_exit(prev_state); return rc; } EXPORT_SYMBOL_GPL(hash_page_mm); @@ -1494,16 +1953,26 @@ int hash_page(unsigned long ea, unsigned long access, unsigned long trap, } EXPORT_SYMBOL_GPL(hash_page); -int __hash_page(unsigned long trap, unsigned long ea, unsigned long dsisr, - unsigned long msr) +DEFINE_INTERRUPT_HANDLER(do_hash_fault) { + unsigned long ea = regs->dar; + unsigned long dsisr = regs->dsisr; unsigned long access = _PAGE_PRESENT | _PAGE_READ; unsigned long flags = 0; - struct mm_struct *mm = current->mm; - unsigned int region_id = get_region_id(ea); + struct mm_struct *mm; + unsigned int region_id; + long err; + if (unlikely(dsisr & (DSISR_BAD_FAULT_64S | DSISR_KEYFAULT))) { + hash__do_page_fault(regs); + return; + } + + region_id = get_region_id(ea); if ((region_id == VMALLOC_REGION_ID) || (region_id == IO_REGION_ID)) mm = &init_mm; + else + mm = current->mm; if (dsisr & DSISR_NOHPTE) flags |= HPTE_NOHPTE_UPDATE; @@ -1519,16 +1988,30 @@ int __hash_page(unsigned long trap, unsigned long ea, unsigned long dsisr, * 2) user space access kernel space. */ access |= _PAGE_PRIVILEGED; - if ((msr & MSR_PR) || (region_id == USER_REGION_ID)) + if (user_mode(regs) || (region_id == USER_REGION_ID)) access &= ~_PAGE_PRIVILEGED; - if (trap == 0x400) + if (TRAP(regs) == INTERRUPT_INST_STORAGE) access |= _PAGE_EXEC; - return hash_page_mm(mm, ea, access, trap, flags); + err = hash_page_mm(mm, ea, access, TRAP(regs), flags); + if (unlikely(err < 0)) { + // failed to insert a hash PTE due to an hypervisor error + if (user_mode(regs)) { + if (IS_ENABLED(CONFIG_PPC_SUBPAGE_PROT) && err == -2) + _exception(SIGSEGV, regs, SEGV_ACCERR, ea); + else + _exception(SIGBUS, regs, BUS_ADRERR, ea); + } else { + bad_page_fault(regs, SIGBUS); + } + err = 0; + + } else if (err) { + hash__do_page_fault(regs); + } } -#ifdef CONFIG_PPC_MM_SLICES static bool should_hash_preload(struct mm_struct *mm, unsigned long ea) { int psize = get_slice_psize(mm, ea); @@ -1545,12 +2028,6 @@ static bool should_hash_preload(struct mm_struct *mm, unsigned long ea) return true; } -#else -static bool should_hash_preload(struct mm_struct *mm, unsigned long ea) -{ - return true; -} -#endif static void hash_preload(struct mm_struct *mm, pte_t *ptep, unsigned long ea, bool is_exec, unsigned long trap) @@ -1559,6 +2036,7 @@ static void hash_preload(struct mm_struct *mm, pte_t *ptep, unsigned long ea, pgd_t *pgdir; int rc, ssize, update_flags = 0; unsigned long access = _PAGE_PRESENT | _PAGE_READ | (is_exec ? _PAGE_EXEC : 0); + unsigned long flags; BUG_ON(get_region_id(ea) != USER_REGION_ID); @@ -1592,6 +2070,20 @@ static void hash_preload(struct mm_struct *mm, pte_t *ptep, unsigned long ea, return; #endif /* CONFIG_PPC_64K_PAGES */ + /* + * __hash_page_* must run with interrupts off, including PMI interrupts + * off, as it sets the H_PAGE_BUSY bit. + * + * It's otherwise possible for perf interrupts to hit at any time and + * may take a hash fault reading the user stack, which could take a + * hash miss and deadlock on the same H_PAGE_BUSY bit. + * + * Interrupts must also be off for the duration of the + * mm_is_thread_local test and update, to prevent preempt running the + * mm on another CPU (XXX: this may be racy vs kthread_use_mm). + */ + powerpc_local_irq_pmu_save(flags); + /* Is that local to this CPU ? */ if (mm_is_thread_local(mm)) update_flags |= HPTE_LOCAL_UPDATE; @@ -1614,6 +2106,8 @@ static void hash_preload(struct mm_struct *mm, pte_t *ptep, unsigned long ea, mm_ctx_user_psize(&mm->context), mm_ctx_user_psize(&mm->context), pte_val(*ptep)); + + powerpc_local_irq_pmu_restore(flags); } /* @@ -1624,7 +2118,7 @@ static void hash_preload(struct mm_struct *mm, pte_t *ptep, unsigned long ea, * * This must always be called with the pte lock held. */ -void update_mmu_cache(struct vm_area_struct *vma, unsigned long address, +void __update_mmu_cache(struct vm_area_struct *vma, unsigned long address, pte_t *ptep) { /* @@ -1634,9 +2128,6 @@ void update_mmu_cache(struct vm_area_struct *vma, unsigned long address, unsigned long trap; bool is_exec; - if (radix_enabled()) - return; - /* We only want HPTEs for linux PTEs that have _PAGE_ACCESSED set */ if (!pte_young(*ptep) || address >= TASK_SIZE) return; @@ -1706,10 +2197,6 @@ unsigned long pte_get_hash_gslot(unsigned long vpn, unsigned long shift, return gslot; } -/* - * WARNING: This is called from hash_low_64.S, if you change this prototype, - * do not forget to update the assembly call site ! - */ void flush_hash_page(unsigned long vpn, real_pte_t pte, int psize, int ssize, unsigned long flags) { @@ -1804,27 +2291,6 @@ void flush_hash_range(unsigned long number, int local) } } -/* - * low_hash_fault is called when we the low level hash code failed - * to instert a PTE due to an hypervisor error - */ -void low_hash_fault(struct pt_regs *regs, unsigned long address, int rc) -{ - enum ctx_state prev_state = exception_enter(); - - if (user_mode(regs)) { -#ifdef CONFIG_PPC_SUBPAGE_PROT - if (rc == -2) - _exception(SIGSEGV, regs, SEGV_ACCERR, address); - else -#endif - _exception(SIGBUS, regs, BUS_ADRERR, address); - } else - bad_page_fault(regs, address, SIGBUS); - - exception_exit(prev_state); -} - long hpte_insert_repeating(unsigned long hash, unsigned long vpn, unsigned long pa, unsigned long rflags, unsigned long vflags, int psize, int ssize) @@ -1858,72 +2324,68 @@ repeat: return slot; } -#ifdef CONFIG_DEBUG_PAGEALLOC -static void kernel_map_linear_page(unsigned long vaddr, unsigned long lmi) +void hpt_clear_stress(void) { - unsigned long hash; - unsigned long vsid = get_kernel_vsid(vaddr, mmu_kernel_ssize); - unsigned long vpn = hpt_vpn(vaddr, vsid, mmu_kernel_ssize); - unsigned long mode = htab_convert_pte_flags(pgprot_val(PAGE_KERNEL)); - long ret; - - hash = hpt_hash(vpn, PAGE_SHIFT, mmu_kernel_ssize); - - /* Don't create HPTE entries for bad address */ - if (!vsid) - return; - - ret = hpte_insert_repeating(hash, vpn, __pa(vaddr), mode, - HPTE_V_BOLTED, - mmu_linear_psize, mmu_kernel_ssize); - - BUG_ON (ret < 0); - spin_lock(&linear_map_hash_lock); - BUG_ON(linear_map_hash_slots[lmi] & 0x80); - linear_map_hash_slots[lmi] = ret | 0x80; - spin_unlock(&linear_map_hash_lock); + int cpu = raw_smp_processor_id(); + int g; + + for (g = 0; g < stress_nr_groups(); g++) { + unsigned long last_group; + last_group = stress_hpt_struct[cpu].last_group[g]; + + if (last_group != -1UL) { + int i; + for (i = 0; i < HPTES_PER_GROUP; i++) { + if (mmu_hash_ops.hpte_remove(last_group) == -1) + break; + } + stress_hpt_struct[cpu].last_group[g] = -1; + } + } } -static void kernel_unmap_linear_page(unsigned long vaddr, unsigned long lmi) +void hpt_do_stress(unsigned long ea, unsigned long hpte_group) { - unsigned long hash, hidx, slot; - unsigned long vsid = get_kernel_vsid(vaddr, mmu_kernel_ssize); - unsigned long vpn = hpt_vpn(vaddr, vsid, mmu_kernel_ssize); + unsigned long last_group; + int cpu = raw_smp_processor_id(); - hash = hpt_hash(vpn, PAGE_SHIFT, mmu_kernel_ssize); - spin_lock(&linear_map_hash_lock); - BUG_ON(!(linear_map_hash_slots[lmi] & 0x80)); - hidx = linear_map_hash_slots[lmi] & 0x7f; - linear_map_hash_slots[lmi] = 0; - spin_unlock(&linear_map_hash_lock); - if (hidx & _PTEIDX_SECONDARY) - hash = ~hash; - slot = (hash & htab_hash_mask) * HPTES_PER_GROUP; - slot += hidx & _PTEIDX_GROUP_IX; - mmu_hash_ops.hpte_invalidate(slot, vpn, mmu_linear_psize, - mmu_linear_psize, - mmu_kernel_ssize, 0); -} + last_group = stress_hpt_struct[cpu].last_group[stress_nr_groups() - 1]; + if (hpte_group == last_group) + return; -void __kernel_map_pages(struct page *page, int numpages, int enable) -{ - unsigned long flags, vaddr, lmi; - int i; + if (last_group != -1UL) { + int i; + /* + * Concurrent CPUs might be inserting into this group, so + * give up after a number of iterations, to prevent a live + * lock. + */ + for (i = 0; i < HPTES_PER_GROUP; i++) { + if (mmu_hash_ops.hpte_remove(last_group) == -1) + break; + } + stress_hpt_struct[cpu].last_group[stress_nr_groups() - 1] = -1; + } - local_irq_save(flags); - for (i = 0; i < numpages; i++, page++) { - vaddr = (unsigned long)page_address(page); - lmi = __pa(vaddr) >> PAGE_SHIFT; - if (lmi >= linear_map_hash_count) - continue; - if (enable) - kernel_map_linear_page(vaddr, lmi); - else - kernel_unmap_linear_page(vaddr, lmi); + if (ea >= PAGE_OFFSET) { + /* + * We would really like to prefetch to get the TLB loaded, then + * remove the PTE before returning from fault interrupt, to + * increase the hash fault rate. + * + * Unfortunately QEMU TCG does not model the TLB in a way that + * makes this possible, and systemsim (mambo) emulator does not + * bring in TLBs with prefetches (although loads/stores do + * work for non-CI PTEs). + * + * So remember this PTE and clear it on the next hash fault. + */ + memmove(&stress_hpt_struct[cpu].last_group[1], + &stress_hpt_struct[cpu].last_group[0], + (stress_nr_groups() - 1) * sizeof(unsigned long)); + stress_hpt_struct[cpu].last_group[0] = hpte_group; } - local_irq_restore(flags); } -#endif /* CONFIG_DEBUG_PAGEALLOC */ void hash__setup_initial_memory_limit(phys_addr_t first_memblock_base, phys_addr_t first_memblock_size) @@ -1989,7 +2451,9 @@ DEFINE_DEBUGFS_ATTRIBUTE(fops_hpt_order, hpt_order_get, hpt_order_set, "%llu\n") static int __init hash64_debugfs(void) { - debugfs_create_file("hpt_order", 0600, powerpc_debugfs_root, NULL, + if (radix_enabled()) + return 0; + debugfs_create_file("hpt_order", 0600, arch_debugfs_dir, NULL, &fops_hpt_order); return 0; } @@ -2003,3 +2467,20 @@ void __init print_system_hash_info(void) if (htab_hash_mask) pr_info("htab_hash_mask = 0x%lx\n", htab_hash_mask); } + +unsigned long arch_randomize_brk(struct mm_struct *mm) +{ + /* + * If we are using 1TB segments and we are allowed to randomise + * the heap, we can put it above 1TB so it is backed by a 1TB + * segment. Otherwise the heap will be in the bottom 1TB + * which always uses 256MB segments and this may result in a + * performance penalty. + */ + if (is_32bit_task()) + return randomize_page(mm->brk, SZ_32M); + else if (!radix_enabled() && mmu_highuser_ssize == MMU_SEGSIZE_1T) + return randomize_page(max_t(unsigned long, mm->brk, SZ_1T), SZ_1G); + else + return randomize_page(mm->brk, SZ_1G); +} |