From d0ec2c6f2c2f0478b34ae78b3e65f60a561ac807 Mon Sep 17 00:00:00 2001 From: "Huang, Ying" Date: Mon, 2 Jun 2008 14:26:18 +0800 Subject: x86: reserve highmem pages via reserve_early This patch makes early reserved highmem pages become reserved pages. This can be used for highmem pages allocated by bootloader such as EFI memory map, linked list of setup_data, etc. Signed-off-by: Huang Ying Cc: andi@firstfloor.org Cc: mingo@redhat.com Signed-off-by: Thomas Gleixner --- arch/x86/mm/init_32.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86/mm/init_32.c') diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index ec30d10154b6..0e7bb5e81670 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -289,7 +289,8 @@ static void __init permanent_kmaps_init(pgd_t *pgd_base) void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro) { - if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn))) { + if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn)) && + !page_is_reserved_early(pfn)) { ClearPageReserved(page); init_page_count(page); __free_page(page); -- cgit v1.2.3 From d2dbf343329dc777d77488743465f7be4245971d Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 13 Jun 2008 02:00:56 -0700 Subject: x86: clean up reserve_bootmem_generic() and port it to 32-bit 1. add reserve_bootmem_generic for 32bit 2. change len to unsigned long 3. make early_res_to_bootmem to use it Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/e820.c | 2 +- arch/x86/kernel/mpparse.c | 18 ++++++------------ arch/x86/mm/init_32.c | 6 ++++++ arch/x86/mm/init_64.c | 3 ++- include/asm-x86/proto.h | 2 -- include/linux/bootmem.h | 2 ++ 6 files changed, 17 insertions(+), 16 deletions(-) (limited to 'arch/x86/mm/init_32.c') diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index 4f2cd5d179e2..774063f11be0 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -635,7 +635,7 @@ void __init early_res_to_bootmem(u64 start, u64 end) continue; printk(KERN_INFO " early res: %d [%llx-%llx] %s\n", i, final_start, final_end - 1, r->name); - reserve_bootmem(final_start, final_end - final_start, + reserve_bootmem_generic(final_start, final_end - final_start, BOOTMEM_DEFAULT); } } diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c index 7ac1b689b70a..b62ac6ba1410 100644 --- a/arch/x86/kernel/mpparse.c +++ b/arch/x86/kernel/mpparse.c @@ -859,10 +859,11 @@ static int __init smp_scan_config(unsigned long base, unsigned long length, if (!reserve) return 1; -#ifdef CONFIG_X86_32 - reserve_bootmem(virt_to_phys(mpf), PAGE_SIZE, + reserve_bootmem_generic(virt_to_phys(mpf), PAGE_SIZE, BOOTMEM_DEFAULT); if (mpf->mpf_physptr) { + unsigned long size = PAGE_SIZE; +#ifdef CONFIG_X86_32 /* * We cannot access to MPC table to compute * table size yet, as only few megabytes from @@ -872,22 +873,15 @@ static int __init smp_scan_config(unsigned long base, unsigned long length, * PAGE_SIZE from mpg->mpf_physptr yields BUG() * in reserve_bootmem. */ - unsigned long size = PAGE_SIZE; unsigned long end = max_low_pfn * PAGE_SIZE; if (mpf->mpf_physptr + size > end) size = end - mpf->mpf_physptr; - reserve_bootmem(mpf->mpf_physptr, size, +#endif + reserve_bootmem_generic(mpf->mpf_physptr, size, BOOTMEM_DEFAULT); } -#else - reserve_bootmem_generic(virt_to_phys(mpf), PAGE_SIZE, - BOOTMEM_DEFAULT); - if (mpf->mpf_physptr) - reserve_bootmem_generic(mpf->mpf_physptr, - PAGE_SIZE, BOOTMEM_DEFAULT); -#endif - return 1; + return 1; } bp += 4; length -= 16; diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 0e7bb5e81670..abadb1da70df 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -785,3 +785,9 @@ void free_initrd_mem(unsigned long start, unsigned long end) free_init_pages("initrd memory", start, end); } #endif + +int __init reserve_bootmem_generic(unsigned long phys, unsigned long len, + int flags) +{ + return reserve_bootmem(phys, len, flags); +} diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index bf7bf1de6c25..b8c2c1ef7ad5 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -799,7 +799,8 @@ void free_initrd_mem(unsigned long start, unsigned long end) } #endif -int __init reserve_bootmem_generic(unsigned long phys, unsigned len, int flags) +int __init reserve_bootmem_generic(unsigned long phys, unsigned long len, + int flags) { #ifdef CONFIG_NUMA int nid, next_nid; diff --git a/include/asm-x86/proto.h b/include/asm-x86/proto.h index a9f51472521e..3dd458c385c0 100644 --- a/include/asm-x86/proto.h +++ b/include/asm-x86/proto.h @@ -14,8 +14,6 @@ extern void ia32_syscall(void); extern void ia32_cstar_target(void); extern void ia32_sysenter_target(void); -extern int reserve_bootmem_generic(unsigned long phys, unsigned len, int flags); - extern void syscall32_cpu_init(void); extern void check_efer(void); diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h index 686895bacd9d..a1d9b79078ea 100644 --- a/include/linux/bootmem.h +++ b/include/linux/bootmem.h @@ -84,6 +84,8 @@ extern int reserve_bootmem(unsigned long addr, unsigned long size, int flags); __alloc_bootmem_low(x, PAGE_SIZE, 0) #endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */ +extern int reserve_bootmem_generic(unsigned long addr, unsigned long size, + int flags); extern unsigned long free_all_bootmem(void); extern unsigned long free_all_bootmem_node(pg_data_t *pgdat); extern void *__alloc_bootmem_node(pg_data_t *pgdat, -- cgit v1.2.3 From b5bc6c0e55000dab86b73f838f5ad02908b23755 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sat, 14 Jun 2008 18:32:52 -0700 Subject: x86, mm: use add_highpages_with_active_regions() for high pages init v2 use early_node_map to init high pages, so we can remove page_is_ram() and page_is_reserved_early() in the big loop with add_one_highpage also remove page_is_reserved_early(), it is not needed anymore. v2: fix the build of other platforms Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/e820.c | 11 -------- arch/x86/mm/discontig_32.c | 19 ++++++-------- arch/x86/mm/init_32.c | 62 ++++++++++++++++++++++++++++++++++++++-------- include/asm-x86/e820.h | 1 - include/asm-x86/highmem.h | 3 +++ include/linux/mm.h | 2 ++ mm/page_alloc.c | 8 ++++++ 7 files changed, 71 insertions(+), 35 deletions(-) (limited to 'arch/x86/mm/init_32.c') diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index 5051ce744b4e..ed46b7a6bc13 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -612,17 +612,6 @@ void __init free_early(u64 start, u64 end) early_res[j - 1].end = 0; } -int __init page_is_reserved_early(unsigned long pagenr) -{ - u64 start = (u64)pagenr << PAGE_SHIFT; - int i; - struct early_res *r; - - i = find_overlapped_early(start, start + PAGE_SIZE); - r = &early_res[i]; - return (i < MAX_EARLY_RES && r->end); -} - void __init early_res_to_bootmem(u64 start, u64 end) { int i; diff --git a/arch/x86/mm/discontig_32.c b/arch/x86/mm/discontig_32.c index c3f119e99e0d..7c4d0255f8d8 100644 --- a/arch/x86/mm/discontig_32.c +++ b/arch/x86/mm/discontig_32.c @@ -100,7 +100,6 @@ unsigned long node_memmap_size_bytes(int nid, unsigned long start_pfn, #endif extern unsigned long find_max_low_pfn(void); -extern void add_one_highpage_init(struct page *, int, int); extern unsigned long highend_pfn, highstart_pfn; #define LARGE_PAGE_BYTES (PTRS_PER_PTE * PAGE_SIZE) @@ -432,10 +431,10 @@ void __init set_highmem_pages_init(int bad_ppro) { #ifdef CONFIG_HIGHMEM struct zone *zone; - struct page *page; + int nid; for_each_zone(zone) { - unsigned long node_pfn, zone_start_pfn, zone_end_pfn; + unsigned long zone_start_pfn, zone_end_pfn; if (!is_highmem(zone)) continue; @@ -443,16 +442,12 @@ void __init set_highmem_pages_init(int bad_ppro) zone_start_pfn = zone->zone_start_pfn; zone_end_pfn = zone_start_pfn + zone->spanned_pages; + nid = zone_to_nid(zone); printk("Initializing %s for node %d (%08lx:%08lx)\n", - zone->name, zone_to_nid(zone), - zone_start_pfn, zone_end_pfn); - - for (node_pfn = zone_start_pfn; node_pfn < zone_end_pfn; node_pfn++) { - if (!pfn_valid(node_pfn)) - continue; - page = pfn_to_page(node_pfn); - add_one_highpage_init(page, node_pfn, bad_ppro); - } + zone->name, nid, zone_start_pfn, zone_end_pfn); + + add_highpages_with_active_regions(nid, zone_start_pfn, + zone_end_pfn, bad_ppro); } totalram_pages += totalhigh_pages; #endif diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index abadb1da70df..ba07a489230e 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -287,10 +287,10 @@ static void __init permanent_kmaps_init(pgd_t *pgd_base) pkmap_page_table = pte; } -void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro) +static void __init +add_one_highpage_init(struct page *page, int pfn, int bad_ppro) { - if (page_is_ram(pfn) && !(bad_ppro && page_kills_ppro(pfn)) && - !page_is_reserved_early(pfn)) { + if (!(bad_ppro && page_kills_ppro(pfn))) { ClearPageReserved(page); init_page_count(page); __free_page(page); @@ -299,18 +299,58 @@ void __init add_one_highpage_init(struct page *page, int pfn, int bad_ppro) SetPageReserved(page); } +struct add_highpages_data { + unsigned long start_pfn; + unsigned long end_pfn; + int bad_ppro; +}; + +static void __init add_highpages_work_fn(unsigned long start_pfn, + unsigned long end_pfn, void *datax) +{ + int node_pfn; + struct page *page; + unsigned long final_start_pfn, final_end_pfn; + struct add_highpages_data *data; + int bad_ppro; + + data = (struct add_highpages_data *)datax; + bad_ppro = data->bad_ppro; + + final_start_pfn = max(start_pfn, data->start_pfn); + final_end_pfn = min(end_pfn, data->end_pfn); + if (final_start_pfn >= final_end_pfn) + return; + + for (node_pfn = final_start_pfn; node_pfn < final_end_pfn; + node_pfn++) { + if (!pfn_valid(node_pfn)) + continue; + page = pfn_to_page(node_pfn); + add_one_highpage_init(page, node_pfn, bad_ppro); + } + +} + +void __init add_highpages_with_active_regions(int nid, unsigned long start_pfn, + unsigned long end_pfn, + int bad_ppro) +{ + struct add_highpages_data data; + + data.start_pfn = start_pfn; + data.end_pfn = end_pfn; + data.bad_ppro = bad_ppro; + + work_with_active_regions(nid, add_highpages_work_fn, &data); +} + #ifndef CONFIG_NUMA static void __init set_highmem_pages_init(int bad_ppro) { - int pfn; + add_highpages_with_active_regions(0, highstart_pfn, highend_pfn, + bad_ppro); - for (pfn = highstart_pfn; pfn < highend_pfn; pfn++) { - /* - * Holes under sparsemem might not have no mem_map[]: - */ - if (pfn_valid(pfn)) - add_one_highpage_init(pfn_to_page(pfn), pfn, bad_ppro); - } totalram_pages += totalhigh_pages; } #endif /* !CONFIG_NUMA */ diff --git a/include/asm-x86/e820.h b/include/asm-x86/e820.h index 6b0ce745a60c..55d310596907 100644 --- a/include/asm-x86/e820.h +++ b/include/asm-x86/e820.h @@ -86,7 +86,6 @@ extern u64 find_e820_area_size(u64 start, u64 *sizep, u64 align); extern void reserve_early(u64 start, u64 end, char *name); extern void free_early(u64 start, u64 end); extern void early_res_to_bootmem(u64 start, u64 end); -extern int page_is_reserved_early(unsigned long pagenr); extern u64 early_reserve_e820(u64 startt, u64 sizet, u64 align); extern unsigned long e820_end_of_ram(void); diff --git a/include/asm-x86/highmem.h b/include/asm-x86/highmem.h index e153f3b44774..85c4fea41ff6 100644 --- a/include/asm-x86/highmem.h +++ b/include/asm-x86/highmem.h @@ -74,6 +74,9 @@ struct page *kmap_atomic_to_page(void *ptr); #define flush_cache_kmaps() do { } while (0) +extern void add_highpages_with_active_regions(int nid, unsigned long start_pfn, + unsigned long end_pfn, int bad_ppro); + #endif /* __KERNEL__ */ #endif /* _ASM_HIGHMEM_H */ diff --git a/include/linux/mm.h b/include/linux/mm.h index 034a3156d2f0..e4de460907c1 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1011,6 +1011,8 @@ extern unsigned long find_min_pfn_with_active_regions(void); extern unsigned long find_max_pfn_with_active_regions(void); extern void free_bootmem_with_active_regions(int nid, unsigned long max_low_pfn); +typedef void (*work_fn_t)(unsigned long, unsigned long, void *); +extern void work_with_active_regions(int nid, work_fn_t work_fn, void *data); extern void sparse_memory_present_with_active_regions(int nid); #ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID extern int early_pfn_to_nid(unsigned long pfn); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index d80e1868e570..41c6e3aa059f 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2929,6 +2929,14 @@ void __init free_bootmem_with_active_regions(int nid, } } +void __init work_with_active_regions(int nid, work_fn_t work_fn, void *data) +{ + int i; + + for_each_active_range_index_in_nid(i, nid) + work_fn(early_node_map[i].start_pfn, early_node_map[i].end_pfn, + data); +} /** * sparse_memory_present_with_active_regions - Call memory_present for each active range * @nid: The node to call memory_present for. If MAX_NUMNODES, all nodes will be used. -- cgit v1.2.3 From cc9f7a0ccf000d4db5fbdc7b0ae48eefea102f69 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 16 Jun 2008 16:11:08 -0700 Subject: x86: kill bad_ppro so don't punish all other cpus without that problem when init highmem Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/kernel/setup_32.c | 9 +++++++++ arch/x86/mm/discontig_32.c | 4 ++-- arch/x86/mm/init_32.c | 43 ++++++++++++------------------------------- include/asm-x86/highmem.h | 2 +- include/asm-x86/numa_32.h | 2 +- 5 files changed, 25 insertions(+), 35 deletions(-) (limited to 'arch/x86/mm/init_32.c') diff --git a/arch/x86/kernel/setup_32.c b/arch/x86/kernel/setup_32.c index f3ddba5ed9a7..9692aeb8ecae 100644 --- a/arch/x86/kernel/setup_32.c +++ b/arch/x86/kernel/setup_32.c @@ -68,6 +68,7 @@ #include #include #include +#include /* This value is set up by the early boot code to point to the value immediately after the boot time page tables. It contains a *physical* @@ -764,6 +765,14 @@ void __init setup_arch(char **cmdline_p) if (efi_enabled) efi_init(); + if (ppro_with_ram_bug()) { + e820_update_range(0x70000000ULL, 0x40000ULL, E820_RAM, + E820_RESERVED); + sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map); + printk(KERN_INFO "fixed physical RAM map:\n"); + e820_print_map("bad_ppro"); + } + e820_register_active_regions(0, 0, -1UL); /* * partially used pages are not usable - thus diff --git a/arch/x86/mm/discontig_32.c b/arch/x86/mm/discontig_32.c index 7c4d0255f8d8..6216e43b6e95 100644 --- a/arch/x86/mm/discontig_32.c +++ b/arch/x86/mm/discontig_32.c @@ -427,7 +427,7 @@ void __init zone_sizes_init(void) return; } -void __init set_highmem_pages_init(int bad_ppro) +void __init set_highmem_pages_init(void) { #ifdef CONFIG_HIGHMEM struct zone *zone; @@ -447,7 +447,7 @@ void __init set_highmem_pages_init(int bad_ppro) zone->name, nid, zone_start_pfn, zone_end_pfn); add_highpages_with_active_regions(nid, zone_start_pfn, - zone_end_pfn, bad_ppro); + zone_end_pfn); } totalram_pages += totalhigh_pages; #endif diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index ba07a489230e..fb5694d788bf 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -220,13 +220,6 @@ static void __init kernel_physical_mapping_init(pgd_t *pgd_base) } } -static inline int page_kills_ppro(unsigned long pagenr) -{ - if (pagenr >= 0x70000 && pagenr <= 0x7003F) - return 1; - return 0; -} - /* * devmem_is_allowed() checks to see if /dev/mem access to a certain address * is valid. The argument is a physical page number. @@ -287,22 +280,17 @@ static void __init permanent_kmaps_init(pgd_t *pgd_base) pkmap_page_table = pte; } -static void __init -add_one_highpage_init(struct page *page, int pfn, int bad_ppro) +static void __init add_one_highpage_init(struct page *page, int pfn) { - if (!(bad_ppro && page_kills_ppro(pfn))) { - ClearPageReserved(page); - init_page_count(page); - __free_page(page); - totalhigh_pages++; - } else - SetPageReserved(page); + ClearPageReserved(page); + init_page_count(page); + __free_page(page); + totalhigh_pages++; } struct add_highpages_data { unsigned long start_pfn; unsigned long end_pfn; - int bad_ppro; }; static void __init add_highpages_work_fn(unsigned long start_pfn, @@ -312,10 +300,8 @@ static void __init add_highpages_work_fn(unsigned long start_pfn, struct page *page; unsigned long final_start_pfn, final_end_pfn; struct add_highpages_data *data; - int bad_ppro; data = (struct add_highpages_data *)datax; - bad_ppro = data->bad_ppro; final_start_pfn = max(start_pfn, data->start_pfn); final_end_pfn = min(end_pfn, data->end_pfn); @@ -327,29 +313,26 @@ static void __init add_highpages_work_fn(unsigned long start_pfn, if (!pfn_valid(node_pfn)) continue; page = pfn_to_page(node_pfn); - add_one_highpage_init(page, node_pfn, bad_ppro); + add_one_highpage_init(page, node_pfn); } } void __init add_highpages_with_active_regions(int nid, unsigned long start_pfn, - unsigned long end_pfn, - int bad_ppro) + unsigned long end_pfn) { struct add_highpages_data data; data.start_pfn = start_pfn; data.end_pfn = end_pfn; - data.bad_ppro = bad_ppro; work_with_active_regions(nid, add_highpages_work_fn, &data); } #ifndef CONFIG_NUMA -static void __init set_highmem_pages_init(int bad_ppro) +static void __init set_highmem_pages_init(void) { - add_highpages_with_active_regions(0, highstart_pfn, highend_pfn, - bad_ppro); + add_highpages_with_active_regions(0, highstart_pfn, highend_pfn); totalram_pages += totalhigh_pages; } @@ -358,7 +341,7 @@ static void __init set_highmem_pages_init(int bad_ppro) #else # define kmap_init() do { } while (0) # define permanent_kmaps_init(pgd_base) do { } while (0) -# define set_highmem_pages_init(bad_ppro) do { } while (0) +# define set_highmem_pages_init() do { } while (0) #endif /* CONFIG_HIGHMEM */ pteval_t __PAGE_KERNEL = _PAGE_KERNEL; @@ -605,13 +588,11 @@ static struct kcore_list kcore_mem, kcore_vmalloc; void __init mem_init(void) { int codesize, reservedpages, datasize, initsize; - int tmp, bad_ppro; + int tmp; #ifdef CONFIG_FLATMEM BUG_ON(!mem_map); #endif - bad_ppro = ppro_with_ram_bug(); - #ifdef CONFIG_HIGHMEM /* check that fixmap and pkmap do not overlap */ if (PKMAP_BASE + LAST_PKMAP*PAGE_SIZE >= FIXADDR_START) { @@ -634,7 +615,7 @@ void __init mem_init(void) if (page_is_ram(tmp) && PageReserved(pfn_to_page(tmp))) reservedpages++; - set_highmem_pages_init(bad_ppro); + set_highmem_pages_init(); codesize = (unsigned long) &_etext - (unsigned long) &_text; datasize = (unsigned long) &_edata - (unsigned long) &_etext; diff --git a/include/asm-x86/highmem.h b/include/asm-x86/highmem.h index 85c4fea41ff6..4514b16cc723 100644 --- a/include/asm-x86/highmem.h +++ b/include/asm-x86/highmem.h @@ -75,7 +75,7 @@ struct page *kmap_atomic_to_page(void *ptr); #define flush_cache_kmaps() do { } while (0) extern void add_highpages_with_active_regions(int nid, unsigned long start_pfn, - unsigned long end_pfn, int bad_ppro); + unsigned long end_pfn); #endif /* __KERNEL__ */ diff --git a/include/asm-x86/numa_32.h b/include/asm-x86/numa_32.h index 03d0f7a9bf02..a02674f64869 100644 --- a/include/asm-x86/numa_32.h +++ b/include/asm-x86/numa_32.h @@ -5,7 +5,7 @@ extern int pxm_to_nid(int pxm); #ifdef CONFIG_NUMA extern void __init remap_numa_kva(void); -extern void set_highmem_pages_init(int); +extern void set_highmem_pages_init(void); #else static inline void remap_numa_kva(void) { -- cgit v1.2.3