lmb cleanups and additions
Subsystems (6)
This is my current WIP series. It's compile tested on ppc and sparc64 and quickly boot tested on ppc64. The "advanced" features such as the new array resize are completely untested in this version.
My aim is still to replace the bottom part of Yinghai's patch series rather than build on top of it, and from there, add whatever he needs to successfully port x86 over and turn NO_BOOTMEM into something half decent without adding a ton of unneeded crap to the core lmb.
This is not finished tho. Here's a peek at my TODO list:
- Move to mm/lmb.c
- Various random return types with non-useful return codes (lmb_add, lmb_remove, ...) needs cleaning & documenting
- Add docbook for all API functions
- lmb_add or lmb_reserve of overlapping regions are going to wreck things in very interesting ways. We could easily error out but that's sub-optimal, instead we should break them up to only add/reserve the bits that aren't yet
- Add some pr_debug in various places in there
- Improve the NUMA interaction
In the meantime, comments are welcome :-)
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 1/25 lmb: Rename lmb_region to lmb_type and lmb_property to lmb_region
--- arch/microblaze/mm/init.c | 4 +- arch/powerpc/mm/hash_utils_64.c | 2 +- arch/powerpc/mm/mem.c | 26 +++--- arch/powerpc/platforms/embedded6xx/wii.c | 2 +- arch/sparc/mm/init_64.c | 6 +- include/linux/lmb.h | 24 ++-- lib/lmb.c | 168 +++++++++++++++--------------- 7 files changed, 115 insertions(+), 117 deletions(-)
arch/microblaze/mm/init.c +2 -2
@@ -76,8 +76,8 @@ void __init setup_memory(void)/* Find main memory where is the kernel */ for (i =0 ; i < lmb.memory.cnt; i++) { - memory_start = (u32 ) lmb.memory.region[i].base; - memory_end = (u32 ) lmb.memory.region[i].base + memory_start = (u32 ) lmb.memory.regions[i].base; + memory_end = (u32 ) lmb.memory.regions[i].base + (u32 ) lmb.memory.region[i].size;if ((memory_start <= (u32 )_text) && ((u32 )_text <= memory_end)) {
arch/powerpc/mm/hash_utils_64.c +1 -1
@@ -660,7 +660,7 @@ static void __init htab_initialize(void)/* create bolted the linear mapping in the hash table */ for (i=0 ; i < lmb.memory.cnt; i++) { - base = (unsigned long )__va (lmb.memory.region[i].base); + base = (unsigned long )__va (lmb.memory.regions[i].base); size = lmb.memory.region[i].size; DBG("creating mapping for region: %lx..%lx (prot: %lx)\n",
arch/powerpc/mm/mem.c +13 -13
@@ -86,10 +86,10 @@ int page_is_ram(unsigned long pfn)for (i=0 ; i < lmb.memory.cnt; i++) {unsigned long base; - base = lmb.memory.region[i].base; + base = lmb.memory.regions[i].base;if ((paddr >= base) && - (paddr < (base + lmb.memory.region[i].size))) { + (paddr < (base + lmb.memory.regions[i].size))) {return 1 ; } }
@@ -149,7 +149,7 @@ int walk_system_ram_range(unsigned long start_pfn, unsigned long nr_pages, void *arg, int (*func)(unsigned long, unsigned long, void *)) { - struct lmb_property res; + struct lmb_region res;unsigned long pfn, len;u64 end;int ret =-1 ;
@@ -206,7 +206,7 @@ void __init do_init_bootmem(void)/* Add active regions with valid PFNs */ for (i =0 ; i < lmb.memory.cnt; i++) {unsigned long start_pfn, end_pfn; - start_pfn = lmb.memory.region[i].base >>PAGE_SHIFT ; + start_pfn = lmb.memory.regions[i].base >>PAGE_SHIFT ; end_pfn = start_pfn +lmb_size_pages (&lmb.memory, i);add_active_range (0 , start_pfn, end_pfn); }
@@ -219,16 +219,16 @@ void __init do_init_bootmem(void)/* reserve the sections we're already using */ for (i =0 ; i < lmb.reserved.cnt; i++) { -unsigned long addr = lmb.reserved.region[i].base + +unsigned long addr = lmb.reserved.regions[i].base +lmb_size_bytes (&lmb.reserved, i) -1 ;if (addr < lowmem_end_addr) -reserve_bootmem (lmb.reserved.region[i].base, +reserve_bootmem (lmb.reserved.regions[i].base,lmb_size_bytes (&lmb.reserved, i),BOOTMEM_DEFAULT ); -else if (lmb.reserved.region[i].base < lowmem_end_addr) { +else if (lmb.reserved.regions[i].base < lowmem_end_addr) {unsigned long adjusted_size = lowmem_end_addr - - lmb.reserved.region[i].base; -reserve_bootmem (lmb.reserved.region[i].base, + lmb.reserved.regions[i].base; +reserve_bootmem (lmb.reserved.regions[i].base, adjusted_size,BOOTMEM_DEFAULT ); } }
@@ -237,7 +237,7 @@ void __init do_init_bootmem(void)/* reserve the sections we're already using */ for (i =0 ; i < lmb.reserved.cnt; i++) -reserve_bootmem (lmb.reserved.region[i].base, +reserve_bootmem (lmb.reserved.regions[i].base,lmb_size_bytes (&lmb.reserved, i),BOOTMEM_DEFAULT );
@@ -257,10 +257,10 @@ static int __init mark_nonram_nosave(void)for (i =0 ; i < lmb.memory.cnt -1 ; i++) { lmb_region_max_pfn = - (lmb.memory.region[i].base >>PAGE_SHIFT ) + - (lmb.memory.region[i].size >>PAGE_SHIFT ); + (lmb.memory.regions[i].base >>PAGE_SHIFT ) + + (lmb.memory.regions[i].size >>PAGE_SHIFT ); lmb_next_region_start_pfn = - lmb.memory.region[i+1 ].base >>PAGE_SHIFT ; + lmb.memory.regions[i+1 ].base >>PAGE_SHIFT ;if (lmb_region_max_pfn < lmb_next_region_start_pfn) register_nosave_region(lmb_region_max_pfn,
arch/powerpc/platforms/embedded6xx/wii.c +1 -1
@@ -65,7 +65,7 @@ static int __init page_aligned(unsigned long x)void __init wii_memory_fixups (void ) { -struct lmb_property *p = lmb.memory.region; +struct lmb_region *p = lmb.memory.region;/* * This is part of a workaround to allow the use of two
arch/sparc/mm/init_64.c +3 -3
@@ -978,7 +978,7 @@ static void __init add_node_ranges(void)unsigned long size =lmb_size_bytes (&lmb.memory, i);unsigned long start, end; - start = lmb.memory.region[i].base; + start = lmb.memory.regions[i].base; end = start + size;while (start < end) {unsigned long this_end;
@@ -1299,7 +1299,7 @@ static void __init bootmem_init_nonnuma(void)if (!size)continue ; - start_pfn = lmb.memory.region[i].base >>PAGE_SHIFT ; + start_pfn = lmb.memory.regions[i].base >>PAGE_SHIFT ; end_pfn = start_pfn +lmb_size_pages (&lmb.memory, i);add_active_range (0 , start_pfn, end_pfn); }
@@ -1339,7 +1339,7 @@ static void __init trim_reserved_in_node(int nid)numadbg (" trim_reserved_in_node(%d)\n" , nid);for (i =0 ; i < lmb.reserved.cnt; i++) { -unsigned long start = lmb.reserved.region[i].base; +unsigned long start = lmb.reserved.regions[i].base;unsigned long size =lmb_size_bytes (&lmb.reserved, i);unsigned long end = start + size;
include/linux/lmb.h +12 -12
@@ -18,22 +18,22 @@#define MAX_LMB_REGIONS 128 -struct lmb_property { +struct lmb_region {u64 base;u64 size; }; -struct lmb_region { +struct lmb_type {unsigned long cnt;u64 size; -struct lmb_property region[MAX_LMB_REGIONS +1 ]; +struct lmb_region regions[MAX_LMB_REGIONS +1 ]; };struct lmb {unsigned long debug;u64 rmo_size; -struct lmb_region memory; -struct lmb_region reserved; +struct lmb_type memory; +struct lmb_type reserved; };extern struct lmb lmb;
@@ -56,27 +56,27 @@ extern u64 lmb_end_of_DRAM(void);extern void __init lmb_enforce_memory_limit (u64 memory_limit);extern int __init lmb_is_reserved (u64 addr);extern int lmb_is_region_reserved (u64 base,u64 size); -extern int lmb_find (struct lmb_property *res); +extern int lmb_find (struct lmb_region *res);extern void lmb_dump_all (void );static inline u64 -lmb_size_bytes (struct lmb_region *type,unsigned long region_nr) +lmb_size_bytes (struct lmb_type *type,unsigned long region_nr) { -return type->region[region_nr].size; +return type->regions[region_nr].size; }static inline u64 -lmb_size_pages (struct lmb_region *type,unsigned long region_nr) +lmb_size_pages (struct lmb_type *type,unsigned long region_nr) {return lmb_size_bytes (type, region_nr) >>PAGE_SHIFT ; }static inline u64 -lmb_start_pfn (struct lmb_region *type,unsigned long region_nr) +lmb_start_pfn (struct lmb_type *type,unsigned long region_nr) { -return type->region[region_nr].base >>PAGE_SHIFT ; +return type->regions[region_nr].base >>PAGE_SHIFT ; }static inline u64 -lmb_end_pfn (struct lmb_region *type,unsigned long region_nr) +lmb_end_pfn (struct lmb_type *type,unsigned long region_nr) {return lmb_start_pfn (type, region_nr) +lmb_size_pages (type, region_nr);
lib/lmb.c +83 -85
402 lines not on this page · raw message
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 2/25 lmb: No reason to include asm/lmb.h late
--- include/linux/lmb.h | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)
include/linux/lmb.h +2 -2
@@ -16,6 +16,8 @@#include <linux/init.h> #include <linux/mm.h> +#include <asm/lmb.h> +#define MAX_LMB_REGIONS 128struct lmb_region {
@@ -82,8 +84,6 @@ lmb_end_pfn(struct lmb_type *type, unsigned long region_nr)lmb_size_pages (type, region_nr); } -#include <asm/lmb.h> -#endif /* __KERNEL__ */ #endif /* _LINUX_LMB_H */
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 3/25 lmb: Introduce for_each_lmb() and new accessors, and use it
Walk lmb's using for_each_lmb() and use lmb_region_base/end_pfn() for getting to PFNs. Update sparc, powerpc, microblaze and sh.
Note: This is -almost- a direct conversion. It doesn't fix some existing crap when/if lmb's aren't page aligned in the first place. This will be sorted out separately.
This removes lmb_find() as well, which isn't used anymore
--- arch/microblaze/mm/init.c | 18 +++---- arch/powerpc/mm/hash_utils_64.c | 8 ++-- arch/powerpc/mm/mem.c | 92 ++++++++++++++------------------------- arch/powerpc/mm/numa.c | 17 ++++--- arch/sh/kernel/setup.c | 14 +++--- arch/sparc/mm/init_64.c | 30 +++++-------- include/linux/lmb.h | 56 ++++++++++++++++++------ lib/lmb.c | 32 ------------- 8 files changed, 114 insertions(+), 153 deletions(-)
arch/microblaze/mm/init.c +8 -10
@@ -69,16 +69,16 @@ static void __init paging_init(void)void __init setup_memory (void ) { -int i;unsigned long map_size; +struct lmb_region *reg; +#ifndef CONFIG_MMU u32 kernel_align_start, kernel_align_size;/* Find main memory where is the kernel */ -for (i =0 ; i < lmb.memory.cnt; i++) { - memory_start = (u32 ) lmb.memory.regions[i].base; - memory_end = (u32 ) lmb.memory.regions[i].base - + (u32 ) lmb.memory.region[i].size; +for_each_lmb (memory, reg) { + memory_start = (u32 )reg->base; + memory_end = (u32 ) reg->base + reg->size;if ((memory_start <= (u32 )_text) && ((u32 )_text <= memory_end)) { memory_size = memory_end - memory_start;
@@ -146,12 +146,10 @@ void __init setup_memory(void)free_bootmem (memory_start, memory_size);/* reserve allocate blocks */ -for (i =0 ; i < lmb.reserved.cnt; i++) { +for_each_lmb (reserved, reg) {pr_debug ("reserved %d - 0x%08x-0x%08x\n" , i, - (u32 ) lmb.reserved.region[i].base, - (u32 )lmb_size_bytes (&lmb.reserved, i)); -reserve_bootmem (lmb.reserved.region[i].base, -lmb_size_bytes (&lmb.reserved, i) -1 ,BOOTMEM_DEFAULT ); + (u32 ) reg->base, (u32 ) reg->size); +reserve_bootmem (reg->base, reg->size,BOOTMEM_DEFAULT ); }#ifdef CONFIG_MMU init_bootmem_done =1 ;
arch/powerpc/mm/hash_utils_64.c +4 -4
@@ -588,7 +588,7 @@ static void __init htab_initialize(void)unsigned long pteg_count;unsigned long prot;unsigned long base =0 , size =0 , limit; -int i; +struct lmb_region *reg;DBG (" -> htab_initialize()\n" );
@@ -659,9 +659,9 @@ static void __init htab_initialize(void)*/ /* create bolted the linear mapping in the hash table */ -for (i=0 ; i < lmb.memory.cnt; i++) { - base = (unsigned long )__va (lmb.memory.regions[i].base); - size = lmb.memory.region[i].size; +for_each_lmb (memory, reg) { + base = (unsigned long )__va (reg->base); + size = reg->size;DBG ("creating mapping for region: %lx..%lx (prot: %lx)\n" , base, size, prot);
arch/powerpc/mm/mem.c +33 -59
144 lines not on this page · raw message
arch/powerpc/mm/numa.c +9 -8
@@ -742,16 +742,17 @@ static void __init setup_nonnuma(void)unsigned long top_of_ram =lmb_end_of_DRAM ();unsigned long total_ram =lmb_phys_mem_size ();unsigned long start_pfn, end_pfn; -unsigned int i, nid =0 ; +unsigned int nid =0 ; +struct lmb_region *reg;printk (KERN_DEBUG "Top of RAM: 0x%lx, Total RAM: 0x%lx\n" , top_of_ram, total_ram);printk (KERN_DEBUG "Memory hole size: %ldMB\n" , (top_of_ram - total_ram) >>20 ); -for (i =0 ; i < lmb.memory.cnt; ++i) { - start_pfn = lmb.memory.region[i].base >>PAGE_SHIFT ; - end_pfn = start_pfn +lmb_size_pages (&lmb.memory, i); +for_each_lmb (memory, reg) { + start_pfn =lmb_region_base_pfn (reg); + end_pfn =lmb_region_end_pfn (reg);fake_numa_create_new_node (end_pfn, &nid);add_active_range (nid, start_pfn, end_pfn);
@@ -887,11 +888,11 @@ static struct notifier_block __cpuinitdata ppc64_numa_nb = {static void mark_reserved_regions_for_nid (int nid) {struct pglist_data *node =NODE_DATA (nid); -int i; +struct lmb_region *reg; -for (i =0 ; i < lmb.reserved.cnt; i++) { -unsigned long physbase = lmb.reserved.region[i].base; -unsigned long size = lmb.reserved.region[i].size; +for_each_lmb (reserved, reg) { +unsigned long physbase = reg->base; +unsigned long size = reg->size;unsigned long start_pfn = physbase >>PAGE_SHIFT ;unsigned long end_pfn =PFN_UP (physbase + size);struct node_active_region node_ar;
arch/sh/kernel/setup.c +6 -8
@@ -237,7 +237,7 @@ void __init setup_bootmem_allocator(unsigned long free_pfn)unsigned long bootmap_size;unsigned long bootmap_pages, bootmem_paddr;u64 total_pages = (lmb_end_of_DRAM () - __MEMORY_START) >>PAGE_SHIFT ; -int i; +struct lmb_region *reg; bootmap_pages =bootmem_bootmap_pages (total_pages);
@@ -253,10 +253,10 @@ void __init setup_bootmem_allocator(unsigned long free_pfn) min_low_pfn, max_low_pfn);/* Add active regions with valid PFNs. */ -for (i =0 ; i < lmb.memory.cnt; i++) { +for_each_lmb (memory, reg) {unsigned long start_pfn, end_pfn; - start_pfn = lmb.memory.region[i].base >>PAGE_SHIFT ; - end_pfn = start_pfn +lmb_size_pages (&lmb.memory, i); + start_pfn =lmb_region_base_pfn (reg); + end_pfn =lmb_region_end_pfn (reg);__add_active_range (0 , start_pfn, end_pfn); }
@@ -267,10 +267,8 @@ void __init setup_bootmem_allocator(unsigned long free_pfn)register_bootmem_low_pages ();/* Reserve the sections we're already using. */ -for (i =0 ; i < lmb.reserved.cnt; i++) -reserve_bootmem (lmb.reserved.region[i].base, -lmb_size_bytes (&lmb.reserved, i), -BOOTMEM_DEFAULT ); +for_each_lmb (reserved, reg) +reserve_bootmem (reg->base, reg->size,BOOTMEM_DEFAULT );node_set_online (0 );
arch/sparc/mm/init_64.c +12 -18
@@ -972,13 +972,13 @@ int of_node_to_nid(struct device_node *dp)static void __init add_node_ranges (void ) { -int i; +struct lmb_region *reg; -for (i =0 ; i < lmb.memory.cnt; i++) { -unsigned long size =lmb_size_bytes (&lmb.memory, i); +for_each_lmb (memory, reg) { +unsigned long size = reg->size;unsigned long start, end; - start = lmb.memory.regions[i].base; + start = reg->base; end = start + size;while (start < end) {unsigned long this_end;
@@ -1281,7 +1281,7 @@ static void __init bootmem_init_nonnuma(void) {unsigned long top_of_ram =lmb_end_of_DRAM ();unsigned long total_ram =lmb_phys_mem_size (); -unsigned int i; +struct lmb_region *reg;numadbg ("bootmem_init_nonnuma()\n" );
@@ -1292,15 +1292,14 @@ static void __init bootmem_init_nonnuma(void)init_node_masks_nonnuma (); -for (i =0 ; i < lmb.memory.cnt; i++) { -unsigned long size =lmb_size_bytes (&lmb.memory, i); +for_each_lmb (memory, reg) {unsigned long start_pfn, end_pfn; -if (!size) +if (!reg->size)continue ; - start_pfn = lmb.memory.regions[i].base >>PAGE_SHIFT ; - end_pfn = start_pfn +lmb_size_pages (&lmb.memory, i); + start_pfn =lmb_region_base_pfn (reg); + end_pfn =lmb_region_end_pfn (reg);add_active_range (0 , start_pfn, end_pfn); }
@@ -1334,17 +1333,12 @@ static void __init reserve_range_in_node(int nid, unsigned long start,static void __init trim_reserved_in_node (int nid) { -int i; +struct lmb_region *reg;numadbg (" trim_reserved_in_node(%d)\n" , nid); -for (i =0 ; i < lmb.reserved.cnt; i++) { -unsigned long start = lmb.reserved.regions[i].base; -unsigned long size =lmb_size_bytes (&lmb.reserved, i); -unsigned long end = start + size; - -reserve_range_in_node (nid, start, end); - } +for_each_lmb (reserved, reg) +reserve_range_in_node (nid, reg->base, reg->base + reg->size); }static void __init bootmem_init_one_node (int nid)
include/linux/lmb.h +42 -14
74 lines not on this page · raw message
lib/lmb.c +0 -32
@@ -505,35 +505,3 @@ int lmb_is_region_reserved(u64 base, u64 size)return lmb_overlaps_region (&lmb.reserved, base, size); } -/* -* Given a <base, len>, find which memory regions belong to this range. -* Adjust the request and return a contiguous chunk. -*/ -int lmb_find (struct lmb_region *res) -{ -int i; -u64 rstart, rend; - - rstart = res->base; - rend = rstart + res->size -1 ; - -for (i =0 ; i < lmb.memory.cnt; i++) { -u64 start = lmb.memory.regions[i].base; -u64 end = start + lmb.memory.regions[i].size -1 ; - -if (start > rend) -return -1 ; - -if ((end >= rstart) && (start < rend)) { -/* adjust the request */ -if (rstart < start) - rstart = start; -if (rend > end) - rend = end; - res->base = rstart; - res->size = rend - rstart +1 ; -return 0 ; - } - } -return -1 ; -}
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 4/25 lmb: Remove nid_range argument, arch provides lmb_nid_range() instead
--- arch/sparc/mm/init_64.c | 16 ++++++---------- include/linux/lmb.h | 7 +++++-- lib/lmb.c | 13 ++++++++----- 3 files changed, 19 insertions(+), 17 deletions(-)
arch/sparc/mm/init_64.c +6 -10
@@ -785,8 +785,7 @@ static int find_node(unsigned long addr)return -1 ; } -static unsigned long long nid_range (unsigned long long start, -unsigned long long end,int *nid) +u64 lmb_nid_range (u64 start,u64 end,int *nid) { *nid =find_node (start); start +=PAGE_SIZE ;
@@ -804,8 +803,7 @@ static unsigned long long nid_range(unsigned long long start, return start; }#else -static unsigned long long nid_range (unsigned long long start, -unsigned long long end,int *nid) +u64 lmb_nid_range (u64 start,u64 end,int *nid) { *nid =0 ;return end;
@@ -822,8 +820,7 @@ static void __init allocate_node_data(int nid)struct pglist_data *p;#ifdef CONFIG_NEED_MULTIPLE_NODES - paddr =lmb_alloc_nid (sizeof (struct pglist_data ), -SMP_CACHE_BYTES , nid, nid_range); + paddr =lmb_alloc_nid (sizeof (struct pglist_data ),SMP_CACHE_BYTES , nid);if (!paddr) {prom_printf ("Cannot allocate pglist_data for nid[%d]\n" , nid);prom_halt ();
@@ -843,8 +840,7 @@ static void __init allocate_node_data(int nid)if (p->node_spanned_pages) { num_pages =bootmem_bootmap_pages (p->node_spanned_pages); - paddr =lmb_alloc_nid (num_pages <<PAGE_SHIFT ,PAGE_SIZE , nid, - nid_range); + paddr =lmb_alloc_nid (num_pages <<PAGE_SHIFT ,PAGE_SIZE , nid);if (!paddr) {prom_printf ("Cannot allocate bootmap for nid[%d]\n" , nid);
@@ -984,7 +980,7 @@ static void __init add_node_ranges(void)unsigned long this_end;int nid; - this_end =nid_range (start, end, &nid); + this_end =lmb_nid_range (start, end, &nid); numadbg("Adding active range nid[%d] " "start[%lx] end[%lx]\n",
@@ -1317,7 +1313,7 @@ static void __init reserve_range_in_node(int nid, unsigned long start,unsigned long this_end;int n; - this_end =nid_range (start, end, &n); + this_end =lmb_nid_range (start, end, &n);if (n == nid) {numadbg (" MATCH reserving range [%lx:%lx]\n" , start, this_end);
include/linux/lmb.h +5 -2
@@ -46,8 +46,7 @@ extern long lmb_add(u64 base, u64 size);extern long lmb_remove (u64 base,u64 size);extern long __init lmb_free (u64 base,u64 size);extern long __init lmb_reserve (u64 base,u64 size); -extern u64 __init lmb_alloc_nid (u64 size,u64 align,int nid, -u64 (*nid_range)(u64 ,u64 ,int *)); +extern u64 __init lmb_alloc_nid (u64 size,u64 align,int nid);extern u64 __init lmb_alloc (u64 size,u64 align);extern u64 __init lmb_alloc_base (u64 size,u64 ,u64 max_addr);
@@ -61,6 +60,10 @@ extern int lmb_is_region_reserved(u64 base, u64 size);extern void lmb_dump_all (void ); +/* Provided by the architecture */ +extern u64 lmb_nid_range (u64 start,u64 end,int *nid); + +/* * pfn conversion functions *
lib/lmb.c +8 -5
@@ -319,7 +319,6 @@ static u64 __init lmb_alloc_nid_unreserved(u64 start, u64 end, }static u64 __init lmb_alloc_nid_region (struct lmb_region *mp, -u64 (*nid_range)(u64 ,u64 ,int *),u64 size,u64 align,int nid) {u64 start, end;
@@ -332,7 +331,7 @@ static u64 __init lmb_alloc_nid_region(struct lmb_region *mp,u64 this_end;int this_nid; - this_end =nid_range (start, end, &this_nid); + this_end =lmb_nid_range (start, end, &this_nid);if (this_nid == nid) {u64 ret =lmb_alloc_nid_unreserved (start, this_end, size, align);
@@ -345,8 +344,7 @@ static u64 __init lmb_alloc_nid_region(struct lmb_region *mp,return ~(u64 )0 ; } -u64 __init lmb_alloc_nid (u64 size,u64 align,int nid, -u64 (*nid_range)(u64 start,u64 end,int *nid)) +u64 __init lmb_alloc_nid (u64 size,u64 align,int nid) {struct lmb_type *mem = &lmb.memory;int i;
@@ -357,7 +355,6 @@ u64 __init lmb_alloc_nid(u64 size, u64 align, int nid,for (i =0 ; i < mem->cnt; i++) {u64 ret =lmb_alloc_nid_region (&mem->regions[i], - nid_range, size, align, nid);if (ret != ~(u64 )0 )return ret;
@@ -505,3 +502,9 @@ int lmb_is_region_reserved(u64 base, u64 size)return lmb_overlaps_region (&lmb.reserved, base, size); } +u64 __weak lmb_nid_range (u64 start,u64 end,int *nid) +{ + *nid =0 ; + +return end; +}
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 5/25 lmb: Factor the lowest level alloc function
--- lib/lmb.c | 59 +++++++++++++++++++++++++++-------------------------------- 1 files changed, 27 insertions(+), 32 deletions(-)
lib/lmb.c +27 -32
@@ -294,8 +294,8 @@ static u64 lmb_align_up(u64 addr, u64 size)return (addr + (size -1 )) & ~(size -1 ); } -static u64 __init lmb_alloc_nid_unreserved (u64 start,u64 end, -u64 size,u64 align) +static u64 __init lmb_alloc_region (u64 start,u64 end, +u64 size,u64 align) {u64 base, res_base;long j;
@@ -318,6 +318,13 @@ static u64 __init lmb_alloc_nid_unreserved(u64 start, u64 end,return ~(u64 )0 ; } +u64 __weak __init lmb_nid_range (u64 start,u64 end,int *nid) +{ + *nid =0 ; + +return end; +} +static u64 __init lmb_alloc_nid_region (struct lmb_region *mp,u64 size,u64 align,int nid) {
@@ -333,8 +340,7 @@ static u64 __init lmb_alloc_nid_region(struct lmb_region *mp, this_end =lmb_nid_range (start, end, &this_nid);if (this_nid == nid) { -u64 ret =lmb_alloc_nid_unreserved (start, this_end, - size, align); +u64 ret =lmb_alloc_region (start, this_end, size, align);if (ret != ~(u64 )0 )return ret; }
@@ -351,6 +357,10 @@ u64 __init lmb_alloc_nid(u64 size, u64 align, int nid)BUG_ON (0 == size); +/* We do a bottom-up search for a region with the right +* nid since that's easier considering how lmb_nid_range() +* works +*/ size =lmb_align_up (size, align);for (i =0 ; i < mem->cnt; i++) {
@@ -383,7 +393,7 @@ u64 __init lmb_alloc_base(u64 size, u64 align, u64 max_addr)u64 __init __lmb_alloc_base (u64 size,u64 align,u64 max_addr) { -long i, j; +long i;u64 base =0 ;u64 res_base;
@@ -396,33 +406,24 @@ u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr)if (max_addr ==LMB_ALLOC_ANYWHERE ) max_addr =LMB_REAL_LIMIT ; +/* Pump up max_addr */ +if (max_addr ==LMB_ALLOC_ANYWHERE ) + max_addr = ~(u64 )0 ; + +/* We do a top-down search, this tends to limit memory +* fragmentation by keeping early boot allocs near the +* top of memory +*/ for (i = lmb.memory.cnt -1 ; i >=0 ; i--) {u64 lmbbase = lmb.memory.regions[i].base;u64 lmbsize = lmb.memory.regions[i].size;if (lmbsize < size)continue ; -if (max_addr ==LMB_ALLOC_ANYWHERE ) - base =lmb_align_down (lmbbase + lmbsize - size, align); -else if (lmbbase < max_addr) { - base =min (lmbbase + lmbsize, max_addr); - base =lmb_align_down (base - size, align); - }else -continue ; - -while (base && lmbbase <= base) { - j =lmb_overlaps_region (&lmb.reserved, base, size); -if (j <0 ) { -/* this area isn't reserved, take it */ -if (lmb_add_region (&lmb.reserved, base, size) <0 ) -return 0 ; -return base; - } - res_base = lmb.reserved.regions[j].base; -if (res_base < size) -break ; - base =lmb_align_down (res_base - size, align); - } + base =min (lmbbase + lmbsize, max_addr); + res_base =lmb_alloc_region (lmbbase, base, size, align); +if (res_base != ~(u64 )0 ) +return res_base; }return 0 ; }
@@ -502,9 +503,3 @@ int lmb_is_region_reserved(u64 base, u64 size)return lmb_overlaps_region (&lmb.reserved, base, size); } -u64 __weak lmb_nid_range (u64 start,u64 end,int *nid) -{ - *nid =0 ; - -return end; -}
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 6/25 lmb: Expose LMB_ALLOC_ANYWHERE
--- arch/powerpc/mm/hash_utils_64.c | 2 +- include/linux/lmb.h | 1 + lib/lmb.c | 2 -- 3 files changed, 2 insertions(+), 3 deletions(-)
arch/powerpc/mm/hash_utils_64.c +1 -1
@@ -625,7 +625,7 @@ static void __init htab_initialize(void)if (machine_is (cell)) limit =0x80000000 ;else - limit =0 ; + limit =LMB_ALLOC_ANYWHERE ; table =lmb_alloc_base (htab_size_bytes, htab_size_bytes, limit);
include/linux/lmb.h +1 -0
@@ -50,6 +50,7 @@ extern u64 __init lmb_alloc_nid(u64 size, u64 align, int nid);extern u64 __init lmb_alloc (u64 size,u64 align);extern u64 __init lmb_alloc_base (u64 size,u64 ,u64 max_addr); +#define LMB_ALLOC_ANYWHERE 0extern u64 __init __lmb_alloc_base (u64 size,u64 align,u64 max_addr);extern u64 __init lmb_phys_mem_size (void );
lib/lmb.c +0 -2
@@ -15,8 +15,6 @@#include <linux/bitops.h> #include <linux/lmb.h> -#define LMB_ALLOC_ANYWHERE 0 -struct lmb lmb;static int lmb_debug;
Signature · 1 line
1.6.3.3
PATCH 7/25 lmb: Introduce default allocation limit and use it to replace explicit ones
This introduce lmb.current_limit which is used to limit allocations from lmb_alloc() or lmb_alloc_base(..., LMB_ALLOC_ACCESSIBLE). The old LMB_ALLOC_ANYWHERE changes value from 0 to ~(u64)0 and can still be used to alloc really anywhere. It is -no-longer- cropped to LMB_REAL_LIMIT which disappears.
--- arch/microblaze/include/asm/lmb.h | 3 --- arch/powerpc/include/asm/lmb.h | 7 ------- arch/powerpc/kernel/prom.c | 20 +++++++++++++++++++- arch/powerpc/kernel/setup_32.c | 2 +- arch/powerpc/mm/40x_mmu.c | 5 +++-- arch/powerpc/mm/fsl_booke_mmu.c | 3 ++- arch/powerpc/mm/hash_utils_64.c | 3 ++- arch/powerpc/mm/init_32.c | 29 +++++++---------------------- arch/powerpc/mm/ppc_mmu_32.c | 3 +-- arch/powerpc/mm/tlb_nohash.c | 2 ++ arch/sh/include/asm/lmb.h | 2 -- arch/sparc/include/asm/lmb.h | 2 -- include/linux/lmb.h | 16 +++++++++++++++- lib/lmb.c | 19 +++++++++++-------- 14 files changed, 63 insertions(+), 53 deletions(-)
arch/microblaze/include/asm/lmb.h +0 -3
@@ -9,9 +9,6 @@#ifndef _ASM_MICROBLAZE_LMB_H#define _ASM_MICROBLAZE_LMB_H -/* LMB limit is OFF */ -#define LMB_REAL_LIMIT 0xFFFFFFFF -#endif /* _ASM_MICROBLAZE_LMB_H */
arch/powerpc/include/asm/lmb.h +0 -7
@@ -5,11 +5,4 @@ #define LMB_DBG(fmt...) udbg_printf(fmt) -#ifdef CONFIG_PPC32 -extern phys_addr_t lowmem_end_addr; -#define LMB_REAL_LIMIT lowmem_end_addr -#else -#define LMB_REAL_LIMIT 0 -#endif -#endif /* _ASM_POWERPC_LMB_H */
arch/powerpc/kernel/prom.c +19 -1
@@ -98,7 +98,7 @@ static void __init move_device_tree(void)if ((memory_limit && (start + size) > memory_limit) ||overlaps_crashkernel (start, size)) { - p =__va (lmb_alloc_base (size,PAGE_SIZE , lmb.rmo_size)); + p =__va (lmb_alloc (size,PAGE_SIZE ));memcpy (p, initial_boot_params, size); initial_boot_params = (struct boot_param_header *)p;DBG ("Moved device tree to 0x%p\n" , p);
@@ -655,6 +655,21 @@ static void __init phyp_dump_reserve_mem(void)static inline void __init phyp_dump_reserve_mem (void ) {}#endif /* CONFIG_PHYP_DUMP && CONFIG_PPC_RTAS */ +static void set_boot_memory_limit (void ) +{ +#ifdef CONFIG_PPC32 +/* 601 can only access 16MB at the moment */ +if (PVR_VER (mfspr (SPRN_PVR )) ==1 ) +lmb_set_current_limit (0x01000000 ); +/* 8xx can only access 8MB at the moment */ +else if (PVR_VER (mfspr (SPRN_PVR )) ==0x50 ) +lmb_set_current_limit (0x00800000 ); +else +lmb_set_current_limit (0x10000000 ); +#else +lmb_set_current_limit (lmb.rmo_size); +#endif +}void __init early_init_devtree (void *params) {
@@ -683,6 +698,7 @@ void __init early_init_devtree(void *params)/* Scan memory nodes and rebuild LMBs */ lmb_init (); +of_scan_flat_dt (early_init_dt_scan_root,NULL );of_scan_flat_dt (early_init_dt_scan_memory_ppc,NULL );
@@ -718,6 +734,8 @@ void __init early_init_devtree(void *params)DBG ("Phys. mem: %llx\n" ,lmb_phys_mem_size ()); +set_boot_memory_limit (); +/* We may need to relocate the flat tree, do it now. * FIXME .. and the initrd too? */ move_device_tree ();
arch/powerpc/kernel/setup_32.c +1 -1
@@ -247,7 +247,7 @@ static void __init irqstack_early_init(void)unsigned int i;/* interrupt stacks must be in lowmem, we get that for free on ppc32 -* as the lmb is limited to lowmem by LMB_REAL_LIMIT */ +* as the lmb is limited to lowmem by default */ for_each_possible_cpu (i) { softirq_ctx[i] = (struct thread_info *)__va (lmb_alloc (THREAD_SIZE ,THREAD_SIZE ));
arch/powerpc/mm/40x_mmu.c +3 -2
@@ -35,6 +35,7 @@#include <linux/init.h> #include <linux/delay.h> #include <linux/highmem.h> +#include <linux/lmb.h> #include <asm/pgalloc.h> #include <asm/prom.h>
@@ -47,6 +48,7 @@#include <asm/bootx.h> #include <asm/machdep.h> #include <asm/setup.h> +#include "mmu_decl.h" extern int __map_without_ltlbs;
@@ -139,8 +141,7 @@ unsigned long __init mmu_mapin_ram(unsigned long top)* coverage with normal-sized pages (or other reasons) do not * attempt to allocate outside the allowed range. */ - - __initial_memory_limit_addr = memstart_addr + mapped; +lmb_set_current_limit (memstart_addr + mapped);return mapped; }
arch/powerpc/mm/fsl_booke_mmu.c +2 -1
@@ -40,6 +40,7 @@#include <linux/init.h> #include <linux/delay.h> #include <linux/highmem.h> +#include <linux/lmb.h> #include <asm/pgalloc.h> #include <asm/prom.h>
@@ -231,5 +232,5 @@ void __init adjust_total_lowmem(void)pr_cont ("%lu Mb, residual: %dMb\n" ,tlbcam_sz (tlbcam_index -1 ) >>20 , (unsigned int )((total_lowmem - __max_low_memory) >>20 )); - __initial_memory_limit_addr = memstart_addr + __max_low_memory; +lmb_set_current_limit (memstart_addr + __max_low_memory); }
arch/powerpc/mm/hash_utils_64.c +2 -1
@@ -696,7 +696,8 @@ static void __init htab_initialize(void)#endif /* CONFIG_U3_DART */ BUG_ON (htab_bolt_mapping (base, base + size,__pa (base), prot, mmu_linear_psize, mmu_kernel_ssize)); - } + } +lmb_set_current_limit (LMB_ALLOC_ANYWHERE );/* * If we have a memory_limit and we've allocated TCEs then we need to
arch/powerpc/mm/init_32.c +7 -22
@@ -92,12 +92,6 @@ int __allow_ioremap_reserved;unsigned long __max_low_memory =MAX_LOW_MEM ;/* -* address of the limit of what is accessible with initial MMU setup - -* 256MB usually, but only 16MB on 601. -*/ -phys_addr_t __initial_memory_limit_addr = (phys_addr_t )0x10000000 ; - -/* * Check for command-line options that affect what MMU_init will do. */ void MMU_setup (void )
@@ -126,13 +120,6 @@ void __init MMU_init(void)if (ppc_md.progress) ppc_md.progress ("MMU:enter" ,0x111 ); -/* 601 can only access 16MB at the moment */ -if (PVR_VER (mfspr (SPRN_PVR )) ==1 ) - __initial_memory_limit_addr =0x01000000 ; -/* 8xx can only access 8MB at the moment */ -if (PVR_VER (mfspr (SPRN_PVR )) ==0x50 ) - __initial_memory_limit_addr =0x00800000 ; -/* parse args from command line */ MMU_setup ();
@@ -190,20 +177,18 @@ void __init MMU_init(void)#ifdef CONFIG_BOOTX_TEXT btext_unmap ();#endif + +/* Shortly after that, the entire linear mapping will be available */ +lmb_set_current_limit (lowmem_end_addr); }/* This is only called until mem_init is done. */ void __init *early_get_page (void ) { -void *p; - -if (init_bootmem_done) { - p =alloc_bootmem_pages (PAGE_SIZE ); - }else { - p =__va (lmb_alloc_base (PAGE_SIZE ,PAGE_SIZE , - __initial_memory_limit_addr)); - } -return p; +if (init_bootmem_done) +return alloc_bootmem_pages (PAGE_SIZE ); +else +return __va (lmb_alloc (PAGE_SIZE ,PAGE_SIZE )); }/* Free up now-unused memory */
arch/powerpc/mm/ppc_mmu_32.c +1 -2
@@ -223,8 +223,7 @@ void __init MMU_init_hw(void)* Find some memory for the hash table. */ if ( ppc_md.progress ) ppc_md.progress ("hash:find piece" ,0x322 ); - Hash =__va (lmb_alloc_base (Hash_size, Hash_size, - __initial_memory_limit_addr)); + Hash =__va (lmb_alloc (Hash_size, Hash_size));cacheable_memzero (Hash, Hash_size); _SDR1 =__pa (Hash) |SDR1_LOW_BITS ;
arch/powerpc/mm/tlb_nohash.c +2 -0
@@ -432,6 +432,8 @@ static void __early_init_mmu(int boot_cpu)* the MMU configuration */ mb (); + +lmb_set_current_limit (linear_map_top); }void __init early_init_mmu (void )
arch/sh/include/asm/lmb.h +0 -2
@@ -1,6 +1,4 @@#ifndef __ASM_SH_LMB_H#define __ASM_SH_LMB_H -#define LMB_REAL_LIMIT 0 -#endif /* __ASM_SH_LMB_H */
arch/sparc/include/asm/lmb.h +0 -2
@@ -5,6 +5,4 @@ #define LMB_DBG(fmt...) prom_printf(fmt) -#define LMB_REAL_LIMIT 0 -#endif /* !(_SPARC64_LMB_H) */
include/linux/lmb.h +15 -1
@@ -34,6 +34,7 @@ struct lmb_type {struct lmb {unsigned long debug;u64 rmo_size; +u64 current_limit;struct lmb_type memory;struct lmb_type reserved; };
@@ -46,11 +47,16 @@ extern long lmb_add(u64 base, u64 size);extern long lmb_remove (u64 base,u64 size);extern long __init lmb_free (u64 base,u64 size);extern long __init lmb_reserve (u64 base,u64 size); +extern u64 __init lmb_alloc_nid (u64 size,u64 align,int nid);extern u64 __init lmb_alloc (u64 size,u64 align); + +/* Flags for lmb_alloc_base() amd __lmb_alloc_base() */ +#define LMB_ALLOC_ANYWHERE (~(u64)0) +#define LMB_ALLOC_ACCESSIBLE 0 +extern u64 __init lmb_alloc_base (u64 size,u64 ,u64 max_addr); -#define LMB_ALLOC_ANYWHERE 0extern u64 __init __lmb_alloc_base (u64 size,u64 align,u64 max_addr);extern u64 __init lmb_phys_mem_size (void );
@@ -64,6 +70,14 @@ extern void lmb_dump_all(void);/* Provided by the architecture */ extern u64 lmb_nid_range (u64 start,u64 end,int *nid); +/** +* lmb_set_current_limit - Set the current allocation limit to allow +* limiting allocations to what is currently +* accessible during boot +* @limit: New limit value (physical address) +*/ +extern void lmb_set_current_limit (u64 limit); +/* * pfn conversion functions
lib/lmb.c +11 -8
@@ -115,6 +115,8 @@ void __init lmb_init(void) lmb.reserved.regions[0 ].base =0 ; lmb.reserved.regions[0 ].size =0 ; lmb.reserved.cnt =1 ; + + lmb.current_limit =LMB_ALLOC_ANYWHERE ; }void __init lmb_analyze (void )
@@ -373,7 +375,7 @@ u64 __init lmb_alloc_nid(u64 size, u64 align, int nid)u64 __init lmb_alloc (u64 size,u64 align) { -return lmb_alloc_base (size, align,LMB_ALLOC_ANYWHERE ); +return lmb_alloc_base (size, align,LMB_ALLOC_ACCESSIBLE ); }u64 __init lmb_alloc_base (u64 size,u64 align,u64 max_addr)
@@ -399,14 +401,9 @@ u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr) size =lmb_align_up (size, align); -/* On some platforms, make sure we allocate lowmem */ -/* Note that LMB_REAL_LIMIT may be LMB_ALLOC_ANYWHERE */ -if (max_addr ==LMB_ALLOC_ANYWHERE ) - max_addr =LMB_REAL_LIMIT ; -/* Pump up max_addr */ -if (max_addr ==LMB_ALLOC_ANYWHERE ) - max_addr = ~(u64 )0 ; +if (max_addr ==LMB_ALLOC_ACCESSIBLE ) + max_addr = lmb.current_limit;/* We do a top-down search, this tends to limit memory * fragmentation by keeping early boot allocs near the
@@ -501,3 +498,9 @@ int lmb_is_region_reserved(u64 base, u64 size)return lmb_overlaps_region (&lmb.reserved, base, size); } + +void __init lmb_set_current_limit (u64 limit) +{ + lmb.current_limit = limit; +} +
Signature · 1 line
1.6.3.3
PATCH 8/25 lmb: Remove rmo_size, burry it in arch/powerpc where it belongs
The RMA (RMO is a misnomer) is a concept specific to ppc64 (in fact server ppc64 though I hijack it on embedded ppc64 for similar purposes) and represents the area of memory that can be accessed in real mode (aka with MMU off), or on embedded, from the exception vectors (which is bolted in the TLB) which pretty much boils down to the same thing.
We take that out of the generic LMB data structure and move it into arch/powerpc where it belongs, renaming it to "RMA" while at it.
--- arch/powerpc/include/asm/mmu.h | 12 ++++++++++++ arch/powerpc/kernel/head_40x.S | 6 +----- arch/powerpc/kernel/paca.c | 2 +- arch/powerpc/kernel/prom.c | 29 ++++++++--------------------- arch/powerpc/kernel/rtas.c | 2 +- arch/powerpc/kernel/setup_64.c | 2 +- arch/powerpc/mm/40x_mmu.c | 14 +++++++++++++- arch/powerpc/mm/44x_mmu.c | 14 ++++++++++++++ arch/powerpc/mm/fsl_booke_mmu.c | 9 +++++++++ arch/powerpc/mm/hash_utils_64.c | 22 +++++++++++++++++++++- arch/powerpc/mm/init_32.c | 14 ++++++++++++++ arch/powerpc/mm/init_64.c | 1 + arch/powerpc/mm/ppc_mmu_32.c | 15 +++++++++++++++ arch/powerpc/mm/tlb_nohash.c | 14 ++++++++++++++ include/linux/lmb.h | 1 - lib/lmb.c | 8 -------- 16 files changed, 125 insertions(+), 40 deletions(-)
arch/powerpc/include/asm/mmu.h +12 -0
@@ -2,6 +2,8 @@#define _ASM_POWERPC_MMU_H_#ifdef __KERNEL__ +#include <linux/types.h> +#include <asm/asm-compat.h> #include <asm/feature-fixups.h>
@@ -82,6 +84,16 @@ extern unsigned int __start___mmu_ftr_fixup, __stop___mmu_ftr_fixup;extern void early_init_mmu (void );extern void early_init_mmu_secondary (void ); +extern void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size); + +#ifdef CONFIG_PPC64 +/* This is our real memory area size on ppc64 server, on embedded, we +* make it match the size our of bolted TLB area +*/ +extern u64 ppc64_rma_size; +#endif /* CONFIG_PPC64 */ +#endif /* !__ASSEMBLY__ */ /* The kernel use the constants below to index in the page sizes array.
arch/powerpc/kernel/head_40x.S +1 -5
@@ -923,11 +923,7 @@ initial_mmu: mtspr SPRN_PID,r0 sync - /* Configure and load two entries into TLB slots 62 and 63. - * In case we are pinning TLBs, these are reserved in by the - * other TLB functions. If not reserving, then it doesn't - * matter where they are loaded. - */ + /* Configure and load one entry into TLB slots 63 */ clrrwi r4,r4,10 /* Mask off the real page number */ ori r4,r4,(TLB_WR | TLB_EX) /* Set the write and execute bits */
arch/powerpc/kernel/paca.c +1 -1
@@ -115,7 +115,7 @@ void __init allocate_pacas(void)* the first segment. On iSeries they must be within the area mapped * by the HV, which is HvPagesToMap * HVPAGESIZE bytes. */ - limit =min (0x10000000ULL , lmb.rmo_size); + limit =min (0x10000000ULL , ppc64_rma_size);if (firmware_has_feature (FW_FEATURE_ISERIES )) limit =min (limit, HvPagesToMap *HVPAGESIZE );
arch/powerpc/kernel/prom.c +8 -21
@@ -66,6 +66,7 @@int __initdata iommu_is_off;int __initdata iommu_force_on;unsigned long tce_alloc_start, tce_alloc_end; +u64 ppc64_rma_size;#endif static int __init early_parse_mem (char *p)
@@ -492,7 +493,7 @@ static int __init early_init_dt_scan_memory_ppc(unsigned long node,void __init early_init_dt_add_memory_arch (u64 base,u64 size) { -#if defined(CONFIG_PPC64 ) +#ifdef CONFIG_PPC64 if (iommu_is_off) {if (base >=0x80000000ul )return ;
@@ -501,9 +502,13 @@ void __init early_init_dt_add_memory_arch(u64 base, u64 size) }#endif -lmb_add (base, size); - +/* First LMB added, do some special initializations */ +if (memstart_addr == ~(phys_addr_t )0 ) +setup_initial_memory_limit (base, size); memstart_addr =min ((u64 )memstart_addr, base); + +/* Add the chunk to the LMB list */ +lmb_add (base, size); }u64 __init early_init_dt_alloc_memory_arch (u64 size,u64 align)
@@ -655,22 +660,6 @@ static void __init phyp_dump_reserve_mem(void)static inline void __init phyp_dump_reserve_mem (void ) {}#endif /* CONFIG_PHYP_DUMP && CONFIG_PPC_RTAS */ -static void set_boot_memory_limit (void ) -{ -#ifdef CONFIG_PPC32 -/* 601 can only access 16MB at the moment */ -if (PVR_VER (mfspr (SPRN_PVR )) ==1 ) -lmb_set_current_limit (0x01000000 ); -/* 8xx can only access 8MB at the moment */ -else if (PVR_VER (mfspr (SPRN_PVR )) ==0x50 ) -lmb_set_current_limit (0x00800000 ); -else -lmb_set_current_limit (0x10000000 ); -#else -lmb_set_current_limit (lmb.rmo_size); -#endif -} -void __init early_init_devtree (void *params) {phys_addr_t limit;
@@ -734,8 +723,6 @@ void __init early_init_devtree(void *params)DBG ("Phys. mem: %llx\n" ,lmb_phys_mem_size ()); -set_boot_memory_limit (); -/* We may need to relocate the flat tree, do it now. * FIXME .. and the initrd too? */ move_device_tree ();
arch/powerpc/kernel/rtas.c +1 -1
@@ -934,7 +934,7 @@ void __init rtas_initialize(void)*/ #ifdef CONFIG_PPC64 if (machine_is (pseries) &&firmware_has_feature (FW_FEATURE_LPAR )) { - rtas_region =min (lmb.rmo_size,RTAS_INSTANTIATE_MAX ); + rtas_region =min (ppc64_rma_size,RTAS_INSTANTIATE_MAX ); ibm_suspend_me_token =rtas_token ("ibm,suspend-me" ); }#endif
arch/powerpc/kernel/setup_64.c +1 -1
@@ -482,7 +482,7 @@ static void __init emergency_stack_init(void)* bringup, we need to get at them in real mode. This means they * must also be within the RMO region. */ - limit =min (0x10000000ULL , lmb.rmo_size); + limit =min (0x10000000ULL , ppc64_rma_size);for_each_possible_cpu (i) {unsigned long sp;
arch/powerpc/mm/40x_mmu.c +13 -1
@@ -141,7 +141,19 @@ unsigned long __init mmu_mapin_ram(unsigned long top)* coverage with normal-sized pages (or other reasons) do not * attempt to allocate outside the allowed range. */ -lmb_set_current_limit (memstart_addr + mapped); +lmb_set_current_limit (mapped);return mapped; } + +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* We don't currently support the first LMB not mapping 0 +* physical on those processors +*/ +BUG_ON (first_lmb_base !=0 ); + +/* 40x can only access 16MB at the moment (see head_40x.S) */ +lmb_set_current_limit (min_t (u64, first_lmb_size,0x00800000 )); +}
arch/powerpc/mm/44x_mmu.c +14 -0
@@ -24,6 +24,8 @@*/ #include <linux/init.h> +#include <linux/lmb.h> +#include <asm/mmu.h> #include <asm/system.h> #include <asm/page.h>
@@ -213,6 +215,18 @@ unsigned long __init mmu_mapin_ram(unsigned long top)return total_lowmem; } +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* We don't currently support the first LMB not mapping 0 +* physical on those processors +*/ +BUG_ON (first_lmb_base !=0 ); + +/* 44x has a 256M TLB entry pinned at boot */ +lmb_set_current_limit (min_t (u64, first_lmb_size,PPC_PIN_SIZE )); +} +#ifdef CONFIG_SMP void __cpuinit mmu_init_secondary (int cpu) {
arch/powerpc/mm/fsl_booke_mmu.c +9 -0
@@ -234,3 +234,12 @@ void __init adjust_total_lowmem(void)lmb_set_current_limit (memstart_addr + __max_low_memory); } + +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +phys_addr_t limit = first_lmb_base + first_lmb_size; + +/* 64M mapped initially according to head_fsl_booke.S */ +lmb_set_current_limit (min_t (u64, limit,0x04000000 )); +}
arch/powerpc/mm/hash_utils_64.c +21 -1
@@ -649,7 +649,7 @@ static void __init htab_initialize(void)#ifdef CONFIG_DEBUG_PAGEALLOC linear_map_hash_count =lmb_end_of_DRAM () >>PAGE_SHIFT ; linear_map_hash_slots =__va (lmb_alloc_base (linear_map_hash_count, -1 , lmb.rmo_size)); +1 , ppc64_rma_size));memset (linear_map_hash_slots,0 , linear_map_hash_count);#endif /* CONFIG_DEBUG_PAGEALLOC */
@@ -1221,3 +1221,23 @@ void kernel_map_pages(struct page *page, int numpages, int enable) local_irq_restore(flags); }#endif /* CONFIG_DEBUG_PAGEALLOC */ + +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* We don't currently support the first LMB not mapping 0 +* physical on those processors +*/ +BUG_ON (first_lmb_base !=0 ); + +/* On LPAR systems, the first entry is our RMA region, +* non-LPAR 64-bit hash MMU systems don't have a limitation +* on real mode access, but using the first entry works well +* enough. We also clamp it to 1G to avoid some funky things +* such as RTAS bugs etc... +*/ + ppc64_rma_size =min_t (u64, first_lmb_size,0x40000000 ); + +/* Finally limit subsequent allocations */ +lmb_set_current_limit (ppc64_rma_size); +}
arch/powerpc/mm/init_32.c +14 -0
@@ -237,3 +237,17 @@ void free_initrd_mem(unsigned long start, unsigned long end) }#endif + +#ifdef CONFIG_8xx/* No 8xx specific .c file to put that in ... */ +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* We don't currently support the first LMB not mapping 0 +* physical on those processors +*/ +BUG_ON (first_lmb_base !=0 ); + +/* 8xx can only access 8MB at the moment */ +lmb_set_current_limit (min_t (u64, first_lmb_size,0x00800000 )); +} +#endif /* CONFIG_8xx */
arch/powerpc/mm/init_64.c +1 -0
@@ -328,3 +328,4 @@ int __meminit vmemmap_populate(struct page *start_page, return 0; }#endif /* CONFIG_SPARSEMEM_VMEMMAP */ +
arch/powerpc/mm/ppc_mmu_32.c +15 -0
@@ -271,3 +271,18 @@ void __init MMU_init_hw(void)if ( ppc_md.progress ) ppc_md.progress ("hash:done" ,0x205 ); } + +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* We don't currently support the first LMB not mapping 0 +* physical on those processors +*/ +BUG_ON (first_lmb_base !=0 ); + +/* 601 can only access 16MB at the moment */ +if (PVR_VER (mfspr (SPRN_PVR )) ==1 ) +lmb_set_current_limit (min_t (u64, first_lmb_size,0x01000000 )); +else /* Anything else has 256M mapped */ +lmb_set_current_limit (min_t (u64, first_lmb_size,0x10000000 )); +}
arch/powerpc/mm/tlb_nohash.c +14 -0
@@ -446,4 +446,18 @@ void __cpuinit early_init_mmu_secondary(void) __early_init_mmu(0); } +void setup_initial_memory_limit (phys_addr_t first_lmb_base, +phys_addr_t first_lmb_size) +{ +/* On Embedded 64-bit, we adjust the RMA size to match +* the bolted TLB entry. We know for now that only 1G +* entries are supported though that may eventually +* change. We crop it to the size of the first LMB to +* avoid going over total available memory just in case... +*/ + ppc64_rma_size =min_t (u64, first_lmb_size,0x40000000 ); + +/* Finally limit subsequent allocations */ +lmb_set_current_limit (ppc64_lmb_base + ppc64_rma_size); +}#endif /* CONFIG_PPC64 */
include/linux/lmb.h +0 -1
@@ -33,7 +33,6 @@ struct lmb_type {struct lmb {unsigned long debug; -u64 rmo_size;u64 current_limit;struct lmb_type memory;struct lmb_type reserved;
lib/lmb.c +0 -8
@@ -49,7 +49,6 @@ void lmb_dump_all(void)return ;pr_info ("LMB configuration:\n" ); -pr_info (" rmo_size = 0x%llx\n" , (unsigned long long )lmb.rmo_size);pr_info (" memory.size = 0x%llx\n" , (unsigned long long )lmb.memory.size);lmb_dump (&lmb.memory,"memory" );
@@ -195,10 +194,6 @@ static long lmb_add_region(struct lmb_type *type, u64 base, u64 size)long lmb_add (u64 base,u64 size) { -/* On pSeries LPAR systems, the first LMB is our RMO region. */ -if (base ==0 ) - lmb.rmo_size = size; -return lmb_add_region (&lmb.memory, base, size); }
@@ -459,9 +454,6 @@ void __init lmb_enforce_memory_limit(u64 memory_limit)break ; } -if (lmb.memory.regions[0 ].size < lmb.rmo_size) - lmb.rmo_size = lmb.memory.regions[0 ].size; - memory_limit =lmb_end_of_DRAM ();/* And truncate any reserves above the limit also. */
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 9/25 lmb: Change u64 to phys_addr_t
Let's not waste space and cycles on archs that don't support >32-bit physical address space.
--- include/linux/lmb.h | 44 ++++++++++---------- lib/lmb.c | 114 ++++++++++++++++++++++++++------------------------- 2 files changed, 80 insertions(+), 78 deletions(-)
include/linux/lmb.h +22 -22
@@ -21,19 +21,19 @@#define MAX_LMB_REGIONS 128struct lmb_region { -u64 base; -u64 size; +phys_addr_t base; +phys_addr_t size; };struct lmb_type {unsigned long cnt; -u64 size; +phys_addr_t size;struct lmb_region regions[MAX_LMB_REGIONS +1 ]; };struct lmb {unsigned long debug; -u64 current_limit; +phys_addr_t current_limit;struct lmb_type memory;struct lmb_type reserved; };
@@ -42,32 +42,32 @@ extern struct lmb lmb;extern void __init lmb_init (void );extern void __init lmb_analyze (void ); -extern long lmb_add (u64 base,u64 size); -extern long lmb_remove (u64 base,u64 size); -extern long __init lmb_free (u64 base,u64 size); -extern long __init lmb_reserve (u64 base,u64 size); +extern long lmb_add (phys_addr_t base,phys_addr_t size); +extern long lmb_remove (phys_addr_t base,phys_addr_t size); +extern long __init lmb_free (phys_addr_t base,phys_addr_t size); +extern long __init lmb_reserve (phys_addr_t base,phys_addr_t size); -extern u64 __init lmb_alloc_nid (u64 size,u64 align,int nid); -extern u64 __init lmb_alloc (u64 size,u64 align); +extern phys_addr_t __init lmb_alloc_nid (phys_addr_t size,phys_addr_t align,int nid); +extern phys_addr_t __init lmb_alloc (phys_addr_t size,phys_addr_t align);/* Flags for lmb_alloc_base() amd __lmb_alloc_base() */ -#define LMB_ALLOC_ANYWHERE (~(u64)0) +#define LMB_ALLOC_ANYWHERE (~(phys_addr_t)0)#define LMB_ALLOC_ACCESSIBLE 0 -extern u64 __init lmb_alloc_base (u64 size, -u64 ,u64 max_addr); -extern u64 __init __lmb_alloc_base (u64 size, -u64 align,u64 max_addr); -extern u64 __init lmb_phys_mem_size (void ); -extern u64 lmb_end_of_DRAM (void ); -extern void __init lmb_enforce_memory_limit (u64 memory_limit); -extern int __init lmb_is_reserved (u64 addr); -extern int lmb_is_region_reserved (u64 base,u64 size); +extern phys_addr_t __init lmb_alloc_base (phys_addr_t size, +phys_addr_t ,phys_addr_t max_addr); +extern phys_addr_t __init __lmb_alloc_base (phys_addr_t size, +phys_addr_t align,phys_addr_t max_addr); +extern phys_addr_t __init lmb_phys_mem_size (void ); +extern phys_addr_t lmb_end_of_DRAM (void ); +extern void __init lmb_enforce_memory_limit (phys_addr_t memory_limit); +extern int __init lmb_is_reserved (phys_addr_t addr); +extern int lmb_is_region_reserved (phys_addr_t base,phys_addr_t size);extern void lmb_dump_all (void );/* Provided by the architecture */ -extern u64 lmb_nid_range (u64 start,u64 end,int *nid); +extern phys_addr_t lmb_nid_range (phys_addr_t start,phys_addr_t end,int *nid);/** * lmb_set_current_limit - Set the current allocation limit to allow
@@ -75,7 +75,7 @@ extern u64 lmb_nid_range(u64 start, u64 end, int *nid);* accessible during boot * @limit: New limit value (physical address) */ -extern void lmb_set_current_limit (u64 limit); +extern void lmb_set_current_limit (phys_addr_t limit);/*
lib/lmb.c +58 -56
308 lines not on this page · raw message
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 10/25 lmb: Remove unused lmb.debug struct member
--- include/linux/lmb.h | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-)
include/linux/lmb.h +4 -3
@@ -32,7 +32,6 @@ struct lmb_type { };struct lmb { -unsigned long debug;phys_addr_t current_limit;struct lmb_type memory;struct lmb_type reserved;
@@ -55,9 +54,11 @@ extern phys_addr_t __init lmb_alloc(phys_addr_t size, phys_addr_t align);#define LMB_ALLOC_ACCESSIBLE 0extern phys_addr_t __init lmb_alloc_base (phys_addr_t size, -phys_addr_t ,phys_addr_t max_addr); +phys_addr_t align, +phys_addr_t max_addr);extern phys_addr_t __init __lmb_alloc_base (phys_addr_t size, -phys_addr_t align,phys_addr_t max_addr); +phys_addr_t align, +phys_addr_t max_addr);extern phys_addr_t __init lmb_phys_mem_size (void );extern phys_addr_t lmb_end_of_DRAM (void );extern void __init lmb_enforce_memory_limit (phys_addr_t memory_limit);
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 11/25 lmb: Remove lmb_type.size and add lmb.memory_size instead
Right now, both the "memory" and "reserved" lmb_type structures have a "size" member. It's unused in the later case, and represent the calculated memory size in the later case.
This moves it out to the main lmb structure instead
--- arch/powerpc/mm/mem.c | 2 +- include/linux/lmb.h | 2 +- lib/lmb.c | 8 ++++---- 3 files changed, 6 insertions(+), 6 deletions(-)
arch/powerpc/mm/mem.c +1 -1
@@ -301,7 +301,7 @@ void __init mem_init(void)swiotlb_init (1 );#endif - num_physpages = lmb.memory.size >>PAGE_SHIFT ; + num_physpages =lmb_phys_mem_size () >>PAGE_SHIFT ; high_memory = (void *)__va (max_low_pfn *PAGE_SIZE );#ifdef CONFIG_NEED_MULTIPLE_NODES
include/linux/lmb.h +1 -1
@@ -27,12 +27,12 @@ struct lmb_region {struct lmb_type {unsigned long cnt; -phys_addr_t size;struct lmb_region regions[MAX_LMB_REGIONS +1 ]; };struct lmb {phys_addr_t current_limit; +phys_addr_t memory_size;/* Updated by lmb_analyze() */ struct lmb_type memory;struct lmb_type reserved; };
lib/lmb.c +4 -4
@@ -49,7 +49,7 @@ void lmb_dump_all(void)return ;pr_info ("LMB configuration:\n" ); -pr_info (" memory.size = 0x%llx\n" , (unsigned long long )lmb.memory.size); +pr_info (" memory size = 0x%llx\n" , (unsigned long long )lmb.memory_size);lmb_dump (&lmb.memory,"memory" );lmb_dump (&lmb.reserved,"reserved" );
@@ -123,10 +123,10 @@ void __init lmb_analyze(void) {int i; - lmb.memory.size =0 ; + lmb.memory_size =0 ;for (i =0 ; i < lmb.memory.cnt; i++) - lmb.memory.size += lmb.memory.regions[i].size; + lmb.memory_size += lmb.memory.regions[i].size; }static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size)
@@ -423,7 +423,7 @@ phys_addr_t __init __lmb_alloc_base(phys_addr_t size, phys_addr_t align, phys_ad/* You must call lmb_analyze() before this. */ phys_addr_t __init lmb_phys_mem_size (void ) { -return lmb.memory.size; +return lmb.memory_size; }phys_addr_t lmb_end_of_DRAM (void )
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 12/25 lmb: Move lmb arrays to static storage in lmb.c and make their size a variable
This is in preparation for having resizable arrays.
Note that we still allocate one more than needed, this is unchanged from the previous implementation.
--- include/linux/lmb.h | 7 ++++--- lib/lmb.c | 10 +++++++++- 2 files changed, 13 insertions(+), 4 deletions(-)
include/linux/lmb.h +4 -3
@@ -18,7 +18,7 @@#include <asm/lmb.h> -#define MAX_LMB_REGIONS 128 +#define INIT_LMB_REGIONS 128struct lmb_region {phys_addr_t base;
@@ -26,8 +26,9 @@ struct lmb_region { };struct lmb_type { -unsigned long cnt; -struct lmb_region regions[MAX_LMB_REGIONS +1 ]; +unsigned long cnt;/* number of regions */ +unsigned long max;/* size of the allocated array */ +struct lmb_region *regions; };struct lmb {
lib/lmb.c +9 -1
@@ -18,6 +18,8 @@struct lmb lmb;static int lmb_debug; +static struct lmb_region lmb_memory_init_regions[INIT_LMB_REGIONS +1 ]; +static struct lmb_region lmb_reserved_init_regions[INIT_LMB_REGIONS +1 ];static int __init early_lmb (char *p) {
@@ -104,6 +106,12 @@ static void lmb_coalesce_regions(struct lmb_type *type,void __init lmb_init (void ) { +/* Hookup the initial arrays */ + lmb.memory.regions = lmb_memory_init_regions; + lmb.memory.max =INIT_LMB_REGIONS ; + lmb.reserved.regions = lmb_reserved_init_regions; + lmb.reserved.max =INIT_LMB_REGIONS ; +/* Create a dummy zero size LMB which will get coalesced away later. * This simplifies the lmb_add() code below... */
@@ -169,7 +177,7 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_tif (coalesced)return coalesced; -if (type->cnt >=MAX_LMB_REGIONS ) +if (type->cnt >= type->max)return -1 ;/* Couldn't coalesce the LMB, so add it to the sorted table. */
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 13/25 lmb: Add debug markers at the end of the array
Since we allocate one more than needed, why not do a bit of sanity checking here to ensure we don't walk past the end of the array ?
--- lib/lmb.c | 11 +++++++++++ 1 files changed, 11 insertions(+), 0 deletions(-)
lib/lmb.c +11 -0
@@ -13,6 +13,7 @@#include <linux/kernel.h> #include <linux/init.h> #include <linux/bitops.h> +#include <linux/poison.h> #include <linux/lmb.h> struct lmb lmb;
@@ -112,6 +113,10 @@ void __init lmb_init(void) lmb.reserved.regions = lmb_reserved_init_regions; lmb.reserved.max =INIT_LMB_REGIONS ; +/* Write a marker in the unused last array entry */ + lmb.memory.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; + lmb.reserved.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; +/* Create a dummy zero size LMB which will get coalesced away later. * This simplifies the lmb_add() code below... */
@@ -131,6 +136,12 @@ void __init lmb_analyze(void) {int i; +/* Check marker in the unused last array entry */ +WARN_ON (lmb_memory_init_regions[INIT_LMB_REGIONS ].base + != (phys_addr_t )RED_INACTIVE ); +WARN_ON (lmb_reserved_init_regions[INIT_LMB_REGIONS ].base + != (phys_addr_t )RED_INACTIVE ); + lmb.memory_size =0 ;for (i =0 ; i < lmb.memory.cnt; i++)
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 14/25 lmb: Make lmb_find_region() out of lmb_alloc_region()
This function will be used to locate a free area to put the new lmb arrays when attempting to resize them. lmb_alloc_region() is gone, the two callsites now call lmb_add_region().
--- lib/lmb.c | 20 +++++++++----------- 1 files changed, 9 insertions(+), 11 deletions(-)
lib/lmb.c +9 -11
@@ -309,8 +309,8 @@ static phys_addr_t lmb_align_up(phys_addr_t addr, phys_addr_t size)return (addr + (size -1 )) & ~(size -1 ); } -static phys_addr_t __init lmb_alloc_region (phys_addr_t start,phys_addr_t end, -phys_addr_t size,phys_addr_t align) +static phys_addr_t __init lmb_find_region (phys_addr_t start,phys_addr_t end, +phys_addr_t size,phys_addr_t align) {phys_addr_t base, res_base;long j;
@@ -318,12 +318,8 @@ static phys_addr_t __init lmb_alloc_region(phys_addr_t start, phys_addr_t end, base =lmb_align_down ((end - size), align);while (start <= base) { j =lmb_overlaps_region (&lmb.reserved, base, size); -if (j <0 ) { -/* this area isn't reserved, take it */ -if (lmb_add_region (&lmb.reserved, base, size) <0 ) - base = ~(phys_addr_t )0 ; +if (j <0 )return base; - } res_base = lmb.reserved.regions[j].base;if (res_base < size)break ;
@@ -356,8 +352,9 @@ static phys_addr_t __init lmb_alloc_nid_region(struct lmb_region *mp, this_end =lmb_nid_range (start, end, &this_nid);if (this_nid == nid) { -phys_addr_t ret =lmb_alloc_region (start, this_end, size, align); -if (ret != ~(phys_addr_t )0 ) +phys_addr_t ret =lmb_find_region (start, this_end, size, align); +if (ret != ~(phys_addr_t )0 && +lmb_add_region (&lmb.reserved, start, size) >=0 )return ret; } start = this_end;
@@ -432,8 +429,9 @@ phys_addr_t __init __lmb_alloc_base(phys_addr_t size, phys_addr_t align, phys_adif (lmbsize < size)continue ; base =min (lmbbase + lmbsize, max_addr); - res_base =lmb_alloc_region (lmbbase, base, size, align); -if (res_base != ~(phys_addr_t )0 ) + res_base =lmb_find_region (lmbbase, base, size, align); +if (res_base != ~(phys_addr_t )0 && +lmb_add_region (&lmb.reserved, res_base, size) >=0 )return res_base; }return 0 ;
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 15/25 lmb: Define LMB_ERROR internally instead of using ~(phys_addr_t)0
--- lib/lmb.c | 12 +++++++----- 1 files changed, 7 insertions(+), 5 deletions(-)
lib/lmb.c +7 -5
@@ -22,6 +22,8 @@ static int lmb_debug;static struct lmb_region lmb_memory_init_regions[INIT_LMB_REGIONS +1 ];static struct lmb_region lmb_reserved_init_regions[INIT_LMB_REGIONS +1 ]; +#define LMB_ERROR (~(phys_addr_t)0) +static int __init early_lmb (char *p) {if (p &&strstr (p,"debug" ))
@@ -326,7 +328,7 @@ static phys_addr_t __init lmb_find_region(phys_addr_t start, phys_addr_t end, base =lmb_align_down (res_base - size, align); } -return ~(phys_addr_t )0 ; +return LMB_ERROR ; }phys_addr_t __weak __init lmb_nid_range (phys_addr_t start,phys_addr_t end,int *nid)
@@ -353,14 +355,14 @@ static phys_addr_t __init lmb_alloc_nid_region(struct lmb_region *mp, this_end =lmb_nid_range (start, end, &this_nid);if (this_nid == nid) {phys_addr_t ret =lmb_find_region (start, this_end, size, align); -if (ret != ~(phys_addr_t )0 && +if (ret !=LMB_ERROR &&lmb_add_region (&lmb.reserved, start, size) >=0 )return ret; } start = this_end; } -return ~(phys_addr_t )0 ; +return LMB_ERROR ; }phys_addr_t __init lmb_alloc_nid (phys_addr_t size,phys_addr_t align,int nid)
@@ -379,7 +381,7 @@ phys_addr_t __init lmb_alloc_nid(phys_addr_t size, phys_addr_t align, int nid)for (i =0 ; i < mem->cnt; i++) {phys_addr_t ret =lmb_alloc_nid_region (&mem->regions[i], size, align, nid); -if (ret != ~(phys_addr_t )0 ) +if (ret !=LMB_ERROR )return ret; }
@@ -430,7 +432,7 @@ phys_addr_t __init __lmb_alloc_base(phys_addr_t size, phys_addr_t align, phys_adcontinue ; base =min (lmbbase + lmbsize, max_addr); res_base =lmb_find_region (lmbbase, base, size, align); -if (res_base != ~(phys_addr_t )0 && +if (res_base !=LMB_ERROR &&lmb_add_region (&lmb.reserved, res_base, size) >=0 )return res_base; }
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 16/25 lmb: Move lmb_init() to the bottom of the file
It's a real PITA to have to search for it in the middle
--- lib/lmb.c | 54 +++++++++++++++++++++++++++--------------------------- 1 files changed, 27 insertions(+), 27 deletions(-)
lib/lmb.c +27 -27
@@ -107,33 +107,6 @@ static void lmb_coalesce_regions(struct lmb_type *type,lmb_remove_region (type, r2); } -void __init lmb_init (void ) -{ -/* Hookup the initial arrays */ - lmb.memory.regions = lmb_memory_init_regions; - lmb.memory.max =INIT_LMB_REGIONS ; - lmb.reserved.regions = lmb_reserved_init_regions; - lmb.reserved.max =INIT_LMB_REGIONS ; - -/* Write a marker in the unused last array entry */ - lmb.memory.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; - lmb.reserved.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; - -/* Create a dummy zero size LMB which will get coalesced away later. -* This simplifies the lmb_add() code below... -*/ - lmb.memory.regions[0 ].base =0 ; - lmb.memory.regions[0 ].size =0 ; - lmb.memory.cnt =1 ; - -/* Ditto. */ - lmb.reserved.regions[0 ].base =0 ; - lmb.reserved.regions[0 ].size =0 ; - lmb.reserved.cnt =1 ; - - lmb.current_limit =LMB_ALLOC_ANYWHERE ; -} -void __init lmb_analyze (void ) {int i;
@@ -517,3 +490,30 @@ void __init lmb_set_current_limit(phys_addr_t limit) lmb.current_limit = limit; } +void __init lmb_init (void ) +{ +/* Hookup the initial arrays */ + lmb.memory.regions = lmb_memory_init_regions; + lmb.memory.max =INIT_LMB_REGIONS ; + lmb.reserved.regions = lmb_reserved_init_regions; + lmb.reserved.max =INIT_LMB_REGIONS ; + +/* Write a marker in the unused last array entry */ + lmb.memory.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; + lmb.reserved.regions[INIT_LMB_REGIONS ].base = (phys_addr_t )RED_INACTIVE ; + +/* Create a dummy zero size LMB which will get coalesced away later. +* This simplifies the lmb_add() code below... +*/ + lmb.memory.regions[0 ].base =0 ; + lmb.memory.regions[0 ].size =0 ; + lmb.memory.cnt =1 ; + +/* Ditto. */ + lmb.reserved.regions[0 ].base =0 ; + lmb.reserved.regions[0 ].size =0 ; + lmb.reserved.cnt =1 ; + + lmb.current_limit =LMB_ALLOC_ANYWHERE ; +} +
Signature · 1 line
1.6.3.3
PATCH 17/25 lmb: split lmb_find_base() out of __lmb_alloc_base()
This will be used by the array resize code and might prove useful to some arch code as well at which point it can be made non-static.
--- lib/lmb.c | 43 +++++++++++++++++++++++++++---------------- 1 files changed, 27 insertions(+), 16 deletions(-)
lib/lmb.c +27 -16
@@ -366,20 +366,7 @@ phys_addr_t __init lmb_alloc(phys_addr_t size, phys_addr_t align)return lmb_alloc_base (size, align,LMB_ALLOC_ACCESSIBLE ); } -phys_addr_t __init lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) -{ -phys_addr_t alloc; - - alloc =__lmb_alloc_base (size, align, max_addr); - -if (alloc ==0 ) -panic ("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n" , - (unsigned long long ) size, (unsigned long long ) max_addr); - -return alloc; -} - -phys_addr_t __init __lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) {long i;phys_addr_t base =0 ;
@@ -405,13 +392,37 @@ phys_addr_t __init __lmb_alloc_base(phys_addr_t size, phys_addr_t align, phys_adcontinue ; base =min (lmbbase + lmbsize, max_addr); res_base =lmb_find_region (lmbbase, base, size, align); -if (res_base !=LMB_ERROR && -lmb_add_region (&lmb.reserved, res_base, size) >=0 ) +if (res_base !=LMB_ERROR )return res_base; }return 0 ; } +phys_addr_t __init __lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +{ +phys_addr_t found =lmb_find_base (size, align, max_addr); + +if (found !=LMB_ERROR && +lmb_add_region (&lmb.reserved, found, size) >=0 ) +return found; + +return 0 ; +} + +phys_addr_t __init lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +{ +phys_addr_t alloc; + + alloc =__lmb_alloc_base (size, align, max_addr); + +if (alloc ==0 ) +panic ("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n" , + (unsigned long long ) size, (unsigned long long ) max_addr); + +return alloc; +} + +/* You must call lmb_analyze() before this. */ phys_addr_t __init lmb_phys_mem_size (void ) {
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 18/25 lmb: Move functions around into a more sensible order
Some shuffling is needed for doing array resize so we may as well put some sense into the ordering of the functions in the whole lmb.c file. No code change. Added some comments.
--- lib/lmb.c | 295 ++++++++++++++++++++++++++++++++---------------------------- 1 files changed, 157 insertions(+), 138 deletions(-)
lib/lmb.c +157 -138
@@ -24,40 +24,18 @@ static struct lmb_region lmb_reserved_init_regions[INIT_LMB_REGIONS + 1];#define LMB_ERROR (~(phys_addr_t)0) -static int __init early_lmb (char *p) -{ -if (p &&strstr (p,"debug" )) - lmb_debug =1 ; -return 0 ; -} -early_param ("lmb" , early_lmb); +/* +* Address comparison utilities +*/ -static void lmb_dump (struct lmb_type *region,char *name) +static phys_addr_t lmb_align_down (phys_addr_t addr,phys_addr_t size) { -unsigned long long base, size; -int i; - -pr_info (" %s.cnt = 0x%lx\n" , name, region->cnt); - -for (i =0 ; i < region->cnt; i++) { - base = region->regions[i].base; - size = region->regions[i].size; - -pr_info (" %s[0x%x]\t0x%016llx - 0x%016llx, 0x%llx bytes\n" , - name, i, base, base + size -1 , size); - } +return addr & ~(size -1 ); } -void lmb_dump_all (void ) +static phys_addr_t lmb_align_up (phys_addr_t addr,phys_addr_t size) { -if (!lmb_debug) -return ; - -pr_info ("LMB configuration:\n" ); -pr_info (" memory size = 0x%llx\n" , (unsigned long long )lmb.memory_size); - -lmb_dump (&lmb.memory,"memory" ); -lmb_dump (&lmb.reserved,"reserved" ); +return (addr + (size -1 )) & ~(size -1 ); } static unsigned long lmb_addrs_overlap(phys_addr_t base1, phys_addr_t size1,
@@ -88,6 +66,77 @@ static long lmb_regions_adjacent(struct lmb_type *type,return lmb_addrs_adjacent (base1, size1, base2, size2); } +long lmb_overlaps_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size) +{ +unsigned long i; + +for (i =0 ; i < type->cnt; i++) { +phys_addr_t rgnbase = type->regions[i].base; +phys_addr_t rgnsize = type->regions[i].size; +if (lmb_addrs_overlap (base, size, rgnbase, rgnsize)) +break ; + } + +return (i < type->cnt) ? i :-1 ; +} + +/* +* Find, allocate, deallocate or reserve unreserved regions. All allocations +* are top-down. +*/ + +static phys_addr_t __init lmb_find_region (phys_addr_t start,phys_addr_t end, +phys_addr_t size,phys_addr_t align) +{ +phys_addr_t base, res_base; +long j; + + base =lmb_align_down ((end - size), align); +while (start <= base) { + j =lmb_overlaps_region (&lmb.reserved, base, size); +if (j <0 ) +return base; + res_base = lmb.reserved.regions[j].base; +if (res_base < size) +break ; + base =lmb_align_down (res_base - size, align); + } + +return LMB_ERROR ; +} + +static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +{ +long i; +phys_addr_t base =0 ; +phys_addr_t res_base; + +BUG_ON (0 == size); + + size =lmb_align_up (size, align); + +/* Pump up max_addr */ +if (max_addr ==LMB_ALLOC_ACCESSIBLE ) + max_addr = lmb.current_limit; + +/* We do a top-down search, this tends to limit memory +* fragmentation by keeping early boot allocs near the +* top of memory +*/ +for (i = lmb.memory.cnt -1 ; i >=0 ; i--) { +phys_addr_t lmbbase = lmb.memory.regions[i].base; +phys_addr_t lmbsize = lmb.memory.regions[i].size; + +if (lmbsize < size) +continue ; + base =min (lmbbase + lmbsize, max_addr); + res_base =lmb_find_region (lmbbase, base, size, align); +if (res_base !=LMB_ERROR ) +return res_base; + } +return 0 ; +} +static void lmb_remove_region (struct lmb_type *type,unsigned long r) {unsigned long i;
@@ -107,22 +156,6 @@ static void lmb_coalesce_regions(struct lmb_type *type,lmb_remove_region (type, r2); } -void __init lmb_analyze (void ) -{ -int i; - -/* Check marker in the unused last array entry */ -WARN_ON (lmb_memory_init_regions[INIT_LMB_REGIONS ].base - != (phys_addr_t )RED_INACTIVE ); -WARN_ON (lmb_reserved_init_regions[INIT_LMB_REGIONS ].base - != (phys_addr_t )RED_INACTIVE ); - - lmb.memory_size =0 ; - -for (i =0 ; i < lmb.memory.cnt; i++) - lmb.memory_size += lmb.memory.regions[i].size; -} -static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size) {unsigned long coalesced =0 ;
@@ -260,50 +293,42 @@ long __init lmb_reserve(phys_addr_t base, phys_addr_t size)return lmb_add_region (_rgn, base, size); } -long lmb_overlaps_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size) +phys_addr_t __init __lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) { -unsigned long i; +phys_addr_t found =lmb_find_base (size, align, max_addr); -for (i =0 ; i < type->cnt; i++) { -phys_addr_t rgnbase = type->regions[i].base; -phys_addr_t rgnsize = type->regions[i].size; -if (lmb_addrs_overlap (base, size, rgnbase, rgnsize)) -break ; - } +if (found !=LMB_ERROR && +lmb_add_region (&lmb.reserved, found, size) >=0 ) +return found; -return (i < type->cnt) ? i :-1 ; +return 0 ; } -static phys_addr_t lmb_align_down (phys_addr_t addr,phys_addr_t size) +phys_addr_t __init lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) { -return addr & ~(size -1 ); -} +phys_addr_t alloc; -static phys_addr_t lmb_align_up (phys_addr_t addr,phys_addr_t size) -{ -return (addr + (size -1 )) & ~(size -1 ); -} + alloc =__lmb_alloc_base (size, align, max_addr); -static phys_addr_t __init lmb_find_region (phys_addr_t start,phys_addr_t end, -phys_addr_t size,phys_addr_t align) -{ -phys_addr_t base, res_base; -long j; +if (alloc ==0 ) +panic ("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n" , + (unsigned long long ) size, (unsigned long long ) max_addr); - base =lmb_align_down ((end - size), align); -while (start <= base) { - j =lmb_overlaps_region (&lmb.reserved, base, size); -if (j <0 ) -return base; - res_base = lmb.reserved.regions[j].base; -if (res_base < size) -break ; - base =lmb_align_down (res_base - size, align); - } +return alloc; +} -return LMB_ERROR ; +phys_addr_t __init lmb_alloc (phys_addr_t size,phys_addr_t align) +{ +return lmb_alloc_base (size, align,LMB_ALLOC_ACCESSIBLE ); } + +/* +* Additional node-local allocators. Search for node memory is bottom up +* and walks lmb regions within that node bottom-up as well, but allocation +* within an lmb region is top-down. +*/ +phys_addr_t __weak __init lmb_nid_range (phys_addr_t start,phys_addr_t end,int *nid) { *nid =0 ;
@@ -361,67 +386,9 @@ phys_addr_t __init lmb_alloc_nid(phys_addr_t size, phys_addr_t align, int nid)return lmb_alloc (size, align); } -phys_addr_t __init lmb_alloc (phys_addr_t size,phys_addr_t align) -{ -return lmb_alloc_base (size, align,LMB_ALLOC_ACCESSIBLE ); -} - -static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) -{ -long i; -phys_addr_t base =0 ; -phys_addr_t res_base; - -BUG_ON (0 == size); - - size =lmb_align_up (size, align); - -/* Pump up max_addr */ -if (max_addr ==LMB_ALLOC_ACCESSIBLE ) - max_addr = lmb.current_limit; - -/* We do a top-down search, this tends to limit memory -* fragmentation by keeping early boot allocs near the -* top of memory -*/ -for (i = lmb.memory.cnt -1 ; i >=0 ; i--) { -phys_addr_t lmbbase = lmb.memory.regions[i].base; -phys_addr_t lmbsize = lmb.memory.regions[i].size; - -if (lmbsize < size) -continue ; - base =min (lmbbase + lmbsize, max_addr); - res_base =lmb_find_region (lmbbase, base, size, align); -if (res_base !=LMB_ERROR ) -return res_base; - } -return 0 ; -} - -phys_addr_t __init __lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) -{ -phys_addr_t found =lmb_find_base (size, align, max_addr); - -if (found !=LMB_ERROR && -lmb_add_region (&lmb.reserved, found, size) >=0 ) -return found; - -return 0 ; -} - -phys_addr_t __init lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) -{ -phys_addr_t alloc; - - alloc =__lmb_alloc_base (size, align, max_addr); - -if (alloc ==0 ) -panic ("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n" , - (unsigned long long ) size, (unsigned long long ) max_addr); - -return alloc; -} - +/* +* Remaining API functions +*/ /* You must call lmb_analyze() before this. */ phys_addr_t __init lmb_phys_mem_size (void )
@@ -501,6 +468,50 @@ void __init lmb_set_current_limit(phys_addr_t limit) lmb.current_limit = limit; } +static void lmb_dump (struct lmb_type *region,char *name) +{ +unsigned long long base, size; +int i; + +pr_info (" %s.cnt = 0x%lx\n" , name, region->cnt); + +for (i =0 ; i < region->cnt; i++) { + base = region->regions[i].base; + size = region->regions[i].size; + +pr_info (" %s[0x%x]\t0x%016llx - 0x%016llx, 0x%llx bytes\n" , + name, i, base, base + size -1 , size); + } +} + +void lmb_dump_all (void ) +{ +if (!lmb_debug) +return ; + +pr_info ("LMB configuration:\n" ); +pr_info (" memory size = 0x%llx\n" , (unsigned long long )lmb.memory_size); + +lmb_dump (&lmb.memory,"memory" ); +lmb_dump (&lmb.reserved,"reserved" ); +} + +void __init lmb_analyze (void ) +{ +int i; + +/* Check marker in the unused last array entry */ +WARN_ON (lmb_memory_init_regions[INIT_LMB_REGIONS ].base + != (phys_addr_t )RED_INACTIVE ); +WARN_ON (lmb_reserved_init_regions[INIT_LMB_REGIONS ].base + != (phys_addr_t )RED_INACTIVE ); + + lmb.memory_size =0 ; + +for (i =0 ; i < lmb.memory.cnt; i++) + lmb.memory_size += lmb.memory.regions[i].size; +} +void __init lmb_init (void ) {/* Hookup the initial arrays */
@@ -528,3 +539,11 @@ void __init lmb_init(void) lmb.current_limit =LMB_ALLOC_ANYWHERE ; } +static int __init early_lmb (char *p) +{ +if (p &&strstr (p,"debug" )) + lmb_debug =1 ; +return 0 ; +} +early_param ("lmb" , early_lmb); +
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 19/25 lmb: Add array resizing support
When one of the array gets full, we resize it. After much thinking and a few iterations of that code, I went back to on-demand resizing using the (new) internal lmb_find_base() function, which is pretty much what Yinghai initially proposed, though there some differences in the details.
To work this relies on the default alloc limit being set sensibly by the architecture.
--- lib/lmb.c | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 92 insertions(+), 1 deletions(-)
lib/lmb.c +92 -1
@@ -11,6 +11,7 @@*/ #include <linux/kernel.h> +#include <linux/slab.h> #include <linux/init.h> #include <linux/bitops.h> #include <linux/poison.h>
@@ -24,6 +25,17 @@ static struct lmb_region lmb_reserved_init_regions[INIT_LMB_REGIONS + 1];#define LMB_ERROR (~(phys_addr_t)0) +/* inline so we don't get a warning when pr_debug is compiled out */ +static inline const char *lmb_type_name (struct lmb_type *type) +{ +if (type == &lmb.memory) +return "memory" ; +else if (type == &lmb.reserved) +return "reserved" ; +else +return "unknown" ; +} +/* * Address comparison utilities */
@@ -156,6 +168,73 @@ static void lmb_coalesce_regions(struct lmb_type *type,lmb_remove_region (type, r2); } +/* Defined below but needed now */ +static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size); + +static int lmb_double_array (struct lmb_type *type) +{ +struct lmb_region *new_array, *old_array; +phys_addr_t old_size, new_size, addr; +int use_slab =slab_is_available (); + +pr_debug ("lmb: %s array full, doubling..." ,lmb_type_name (type)); + +/* Calculate new doubled size */ + old_size = type->max *sizeof (struct lmb_region ); + new_size = old_size <<1 ; + +/* Try to find some space for it. +* +* WARNING: We assume that either slab_is_available() and we use it or +* we use LMB for allocations. That means that this is unsafe to use +* when bootmem is currently active (unless bootmem itself is implemented +* on top of LMB which isn't the case yet) +* +* This should however not be an issue for now, as we currently only +* call into LMB while it's still active, or much later when slab is +* active for memory hotplug operations +*/ +if (use_slab) { + new_array =kmalloc (new_size,GFP_KERNEL ); + addr = new_array ==NULL ?LMB_ERROR :__pa (new_array); + }else + addr =lmb_find_base (new_size,sizeof (phys_addr_t),LMB_ALLOC_ACCESSIBLE ); +if (addr ==LMB_ERROR ) { +pr_err ("lmb: Failed to double %s array from %ld to %ld entries !\n" , +lmb_type_name (type), type->max, type->max *2 ); +return -1 ; + } + new_array =__va (addr); + +/* Found space, we now need to move the array over before +* we add the reserved region since it may be our reserved +* array itself that is full. +*/ +memcpy (new_array, type->regions, old_size); +memset (new_array + type->max,0 , old_size); + old_array = type->regions; + type->regions = new_array; + type->max <<=1 ; + +/* If we use SLAB that's it, we are done */ +if (use_slab) +return 0 ; + +/* Add the new reserved region now. Should not fail ! */ +BUG_ON (lmb_add_region (&lmb.reserved, addr, new_size) <0 ); + +/* If the array wasn't our static init one, then free it. We only do +* that before SLAB is available as later on, we don't know whether +* to use kfree or free_bootmem_pages(). Shouldn't be a big deal +* anyways +*/ +if (old_array != lmb_memory_init_regions && + old_array != lmb_reserved_init_regions) +lmb_free (__pa (old_array), old_size); + +return 0 ; +} +static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size) {unsigned long coalesced =0 ;
@@ -196,7 +275,11 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_tif (coalesced)return coalesced; -if (type->cnt >= type->max) + +/* If we are out of space, we fail. It's too late to resize the array +* but then this shouldn't have happened in the first place. +*/ +if (WARN_ON (type->cnt >= type->max))return -1 ;/* Couldn't coalesce the LMB, so add it to the sorted table. */
@@ -217,6 +300,14 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t } type->cnt++; +/* The array is full ? Try to resize it. If that fails, we undo +* our allocation and return an error +*/ +if (type->cnt == type->max &&lmb_double_array (type)) { + type->cnt--; +return -1 ; + } +return 0 ; }
Signature · 1 line
1.6.3.3
PATCH 20/25 lmb: Add arch function to control coalescing of lmb memory regions
Some archs such as ARM want to avoid coalescing accross things such as the lowmem/highmem boundary or similar. This provides the option to control it via an arch callback for which a weak default is provided which always allows coalescing.
--- include/linux/lmb.h | 2 ++ lib/lmb.c | 19 ++++++++++++++++++- 2 files changed, 20 insertions(+), 1 deletions(-)
include/linux/lmb.h +2 -0
@@ -70,6 +70,8 @@ extern void lmb_dump_all(void);/* Provided by the architecture */ extern phys_addr_t lmb_nid_range (phys_addr_t start,phys_addr_t end,int *nid); +extern int lmb_memory_can_coalesce (phys_addr_t addr1,phys_addr_t size1, +phys_addr_t addr2,phys_addr_t size2);/** * lmb_set_current_limit - Set the current allocation limit to allow
lib/lmb.c +18 -1
@@ -235,6 +235,12 @@ static int lmb_double_array(struct lmb_type *type)return 0 ; } +extern int __weak lmb_memory_can_coalesce (phys_addr_t addr1,phys_addr_t size1, +phys_addr_t addr2,phys_addr_t size2) +{ +return 1 ; +} +static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size) {unsigned long coalesced =0 ;
@@ -256,6 +262,10 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_treturn 0 ; adjacent =lmb_addrs_adjacent (base, size, rgnbase, rgnsize); +/* Check if arch allows coalescing */ +if (adjacent !=0 && type == &lmb.memory && + !lmb_memory_can_coalesce (base, size, rgnbase, rgnsize)) +break ;if (adjacent >0 ) { type->regions[i].base -= size; type->regions[i].size += size;
@@ -268,7 +278,14 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t } } -if ((i < type->cnt -1 ) &&lmb_regions_adjacent (type, i, i+1 )) { +/* If we plugged a hole, we may want to also coalesce with the +* next region +*/ +if ((i < type->cnt -1 ) &&lmb_regions_adjacent (type, i, i+1 ) && + ((type != &lmb.memory ||lmb_memory_can_coalesce (type->regions[i].base, + type->regions[i].size, + type->regions[i+1 ].base, + type->regions[i+1 ].size)))) {lmb_coalesce_regions (type, i, i+1 ); coalesced++; }
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 21/25 lmb: Add "start" argument to lmb_find_base()
To constraint the search of a region between two boundaries, which will be used by the new NUMA aware allocator among others.
--- lib/lmb.c | 27 ++++++++++++++++----------- 1 files changed, 16 insertions(+), 11 deletions(-)
lib/lmb.c +16 -11
@@ -117,19 +117,18 @@ static phys_addr_t __init lmb_find_region(phys_addr_t start, phys_addr_t end,return LMB_ERROR ; } -static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align, +phys_addr_t start,phys_addr_t end) {long i; -phys_addr_t base =0 ; -phys_addr_t res_base;BUG_ON (0 == size); size =lmb_align_up (size, align);/* Pump up max_addr */ -if (max_addr ==LMB_ALLOC_ACCESSIBLE ) - max_addr = lmb.current_limit; +if (end ==LMB_ALLOC_ACCESSIBLE ) + end = lmb.current_limit;/* We do a top-down search, this tends to limit memory * fragmentation by keeping early boot allocs near the
@@ -138,13 +137,19 @@ static phys_addr_t __init lmb_find_base(phys_addr_t size, phys_addr_t align, phyfor (i = lmb.memory.cnt -1 ; i >=0 ; i--) {phys_addr_t lmbbase = lmb.memory.regions[i].base;phys_addr_t lmbsize = lmb.memory.regions[i].size; +phys_addr_t bottom, top, found;if (lmbsize < size)continue ; - base =min (lmbbase + lmbsize, max_addr); - res_base =lmb_find_region (lmbbase, base, size, align); -if (res_base !=LMB_ERROR ) -return res_base; +if ((lmbbase + lmbsize) <= start) +break ; + bottom =max (lmbbase, start); + top =min (lmbbase + lmbsize, end); +if (bottom >= top) +continue ; + found =lmb_find_region (lmbbase, top, size, align); +if (found !=LMB_ERROR ) +return found; }return 0 ; }
@@ -198,7 +203,7 @@ static int lmb_double_array(struct lmb_type *type) new_array =kmalloc (new_size,GFP_KERNEL ); addr = new_array ==NULL ?LMB_ERROR :__pa (new_array); }else - addr =lmb_find_base (new_size,sizeof (phys_addr_t),LMB_ALLOC_ACCESSIBLE ); + addr =lmb_find_base (new_size,sizeof (phys_addr_t),0 ,LMB_ALLOC_ACCESSIBLE );if (addr ==LMB_ERROR ) {pr_err ("lmb: Failed to double %s array from %ld to %ld entries !\n" ,lmb_type_name (type), type->max, type->max *2 );
@@ -403,7 +408,7 @@ long __init lmb_reserve(phys_addr_t base, phys_addr_t size)phys_addr_t __init __lmb_alloc_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) { -phys_addr_t found =lmb_find_base (size, align, max_addr); +phys_addr_t found =lmb_find_base (size, align,0 , max_addr);if (found !=LMB_ERROR &&lmb_add_region (&lmb.reserved, found, size) >=0 )
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 22/25 lmb: NUMA allocate can now use early_pfn_map
We now provide a default (weak) implementation of lmb_nid_range() which uses the early_pfn_map[] if CONFIG_ARCH_POPULATES_NODE_MAP is set. Sparc still needs to use its own method due to the way the pages can be scattered between nodes.
This implementation is inefficient due to our main algorithm and callback construct wanting to work on an ascending addresses bases while early_pfn_map[] would rather work with nid's (it's unsorted at that stage). But it should work and we can look into improving it subsequently, possibly using arch compile options to chose a different algorithm alltogether.
--- include/linux/lmb.h | 3 +++ lib/lmb.c | 28 +++++++++++++++++++++++++++- 2 files changed, 30 insertions(+), 1 deletions(-)
include/linux/lmb.h +3 -0
@@ -47,6 +47,9 @@ extern long lmb_remove(phys_addr_t base, phys_addr_t size);extern long __init lmb_free (phys_addr_t base,phys_addr_t size);extern long __init lmb_reserve (phys_addr_t base,phys_addr_t size); +/* The numa aware allocator is only available if +* CONFIG_ARCH_POPULATES_NODE_MAP is set +*/ extern phys_addr_t __init lmb_alloc_nid (phys_addr_t size,phys_addr_t align,int nid);extern phys_addr_t __init lmb_alloc (phys_addr_t size,phys_addr_t align);
lib/lmb.c +27 -1
@@ -15,6 +15,7 @@#include <linux/init.h> #include <linux/bitops.h> #include <linux/poison.h> +#include <linux/pfn.h> #include <linux/lmb.h> struct lmb lmb;
@@ -439,11 +440,36 @@ phys_addr_t __init lmb_alloc(phys_addr_t size, phys_addr_t align)/* * Additional node-local allocators. Search for node memory is bottom up * and walks lmb regions within that node bottom-up as well, but allocation -* within an lmb region is top-down. +* within an lmb region is top-down. XXX I plan to fix that at some stage +* +* WARNING: Only available after early_node_map[] has been populated, +* on some architectures, that is after all the calls to add_active_range() +* have been done to populate it. */ phys_addr_t __weak __init lmb_nid_range (phys_addr_t start,phys_addr_t end,int *nid) { +#ifdef CONFIG_ARCH_POPULATES_NODE_MAP +/* +* This code originates from sparc which really wants use to walk by addresses +* and returns the nid. This is not very convenient for early_pfn_map[] users +* as the map isn't sorted yet, and it really wants to be walked by nid. +* +* For now, I implement the inefficient method below which walks the early +* map multiple times. Eventually we may want to use an ARCH config option +* to implement a completely different method for both case. +*/ +unsigned long start_pfn, end_pfn; +int i; + +for (i =0 ; i <MAX_NUMNODES ; i++) { +get_pfn_range_for_nid (i, &start_pfn, &end_pfn); +if (start <PFN_PHYS (start_pfn) || start >=PFN_PHYS (end_pfn)) +continue ; + *nid = i; +return min (end,PFN_PHYS (end_pfn)); + } +#endif *nid =0 ;return end;
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
PATCH 23/25 lmb: Separate lmb_alloc_nid() and lmb_alloc_try_nid()
The former is now strict, it will fail if it cannot honor the allocation within the node, while the later implements the previous semantic which falls back to allocating anywhere.
--- arch/sparc/mm/init_64.c | 4 ++-- include/linux/lmb.h | 6 +++++- lib/lmb.c | 10 ++++++++++ 3 files changed, 17 insertions(+), 3 deletions(-)
arch/sparc/mm/init_64.c +2 -2
@@ -820,7 +820,7 @@ static void __init allocate_node_data(int nid)struct pglist_data *p;#ifdef CONFIG_NEED_MULTIPLE_NODES - paddr =lmb_alloc_nid (sizeof (struct pglist_data ),SMP_CACHE_BYTES , nid); + paddr =lmb_alloc_try_nid (sizeof (struct pglist_data ),SMP_CACHE_BYTES , nid);if (!paddr) {prom_printf ("Cannot allocate pglist_data for nid[%d]\n" , nid);prom_halt ();
@@ -840,7 +840,7 @@ static void __init allocate_node_data(int nid)if (p->node_spanned_pages) { num_pages =bootmem_bootmap_pages (p->node_spanned_pages); - paddr =lmb_alloc_nid (num_pages <<PAGE_SHIFT ,PAGE_SIZE , nid); + paddr =lmb_alloc_try_nid (num_pages <<PAGE_SHIFT ,PAGE_SIZE , nid);if (!paddr) {prom_printf ("Cannot allocate bootmap for nid[%d]\n" , nid);
include/linux/lmb.h +5 -1
@@ -50,7 +50,11 @@ extern long __init lmb_reserve(phys_addr_t base, phys_addr_t size);/* The numa aware allocator is only available if * CONFIG_ARCH_POPULATES_NODE_MAP is set */ -extern phys_addr_t __init lmb_alloc_nid (phys_addr_t size,phys_addr_t align,int nid); +extern phys_addr_t __init lmb_alloc_nid (phys_addr_t size,phys_addr_t align, +int nid); +extern phys_addr_t __init lmb_alloc_try_nid (phys_addr_t size,phys_addr_t align, +int nid); +extern phys_addr_t __init lmb_alloc (phys_addr_t size,phys_addr_t align);/* Flags for lmb_alloc_base() amd __lmb_alloc_base() */
lib/lmb.c +10 -0
@@ -522,9 +522,19 @@ phys_addr_t __init lmb_alloc_nid(phys_addr_t size, phys_addr_t align, int nid)return ret; } +return 0 ; +} + +phys_addr_t __init lmb_alloc_try_nid (phys_addr_t size,phys_addr_t align,int nid) +{ +phys_addr_t res =lmb_alloc_nid (size, align, nid); + +if (res) +return res;return lmb_alloc (size, align); } +/* * Remaining API functions */
Signature · 7 lines
1.6.3.3 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 21/25] lmb: Add "start" argument to lmb_find_base()
In reply to Benjamin HerrenschmidtQuoted 45 lines · Benjamin Herrenschmidt
To constraint the search of a region between two boundaries, which will be used by the new NUMA aware allocator among others.
Signed-off-by Benjamin Herrenschmidt <[email protected]>---
lib/lmb.c | 27 ++++++++++++++++-----------1 files changed, 16 insertions(+), 11 deletions(-)lib/lmb.c +5
-6@@ -117,19 +117,18 @@ static phys_addr_t __init lmb_find_region(phys_addr_t start, phys_addr_t end, return LMB_ERROR; } -static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align,phys_addr_t max_addr) +static phys_addr_t __init lmb_find_base (phys_addr_t size,phys_addr_t align, + phys_addr_t start, phys_addr_t end) { long i; - phys_addr_t base = 0; - phys_addr_t res_base; BUG_ON(0 == size); size = lmb_align_up(size, align); /* Pump up max_addr */ - if (max_addr == LMB_ALLOC_ACCESSIBLE) - max_addr = lmb.current_limit; + if (end == LMB_ALLOC_ACCESSIBLE) + end = lmb.current_limit; /* We do a top-down search, this tends to limit memory* fragmentation by keeping early boot allocs near the @@ -138,13 +137,19 @@ static phys_addr_t __init lmb_find_base(phys_addr_t size, phys_addr_t align, phy for (i = lmb.memory.cnt - 1; i >= 0; i--) { phys_addr_t lmbbase = lmb.memory.regions[i].base; phys_addr_t lmbsize = lmb.memory.regions[i].size; + phys_addr_t bottom, top, found; if (lmbsize < size) continue; - base = min(lmbbase + lmbsize, max_addr);
lib/lmb.c static phys_addr_t __init lmb_find_base(phys_addr_t size, phys_addr_t align, phy · line 150
Quoted 10 lines
lib/lmb.c +7
-3- res_base = lmb_find_region(lmbbase, base, size, align); - if (res_base != LMB_ERROR) - return res_base; + if ((lmbbase + lmbsize) <= start) + break; + bottom = max(lmbbase, start); + top = min(lmbbase + lmbsize, end); + if (bottom >= top) + continue; + found = lmb_find_region(lmbbase, top, size, align);
^^^^^^^^^
should use bottom here
YH
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 21/25] lmb: Add "start" argument to lmb_find_base()
Quoted 13 lines · Yinghai Lu
lib/lmb.c static phys_addr_t __init lmb_find_base(phys_addr_t size, phys_addr_t align, phy · line 150
Quoted 13 lines
if (lmbsize < size)
continue;lib/lmb.c +7
-4- base =min (lmbbase + lmbsize, max_addr); - res_base =lmb_find_region (lmbbase, base, size, align); -if (res_base !=LMB_ERROR ) -return res_base; +if ((lmbbase + lmbsize) <= start) +break ; + bottom =max (lmbbase, start); + top =min (lmbbase + lmbsize, end); +if (bottom >= top) +continue ; + found =lmb_find_region (lmbbase, top, size, align);
Quoted 2 lines
^^^^^^^^^should use bottom here
Correct, I missed that when converting.
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 19/25] lmb: Add array resizing support
In reply to Benjamin HerrenschmidtQuoted 131 lines · Benjamin Herrenschmidt
When one of the array gets full, we resize it. After much thinking and a few iterations of that code, I went back to on-demand resizing using the (new) internal
lmb_find_base()function, which is pretty much what Yinghai initially proposed, though there some differences in the details.To work this relies on the default alloc limit being set sensibly by the architecture.
Signed-off-by Benjamin Herrenschmidt <[email protected]>---
lib/lmb.c | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-1 files changed, 92 insertions(+), 1 deletions(-)lib/lmb.c +84
-1@@ -11,6 +11,7 @@*/ #include <linux/kernel.h> +#include <linux/slab.h> #include <linux/init.h> #include <linux/bitops.h> #include <linux/poison.h>@@ -24,6 +25,17 @@ static struct lmb_region lmb_reserved_init_regions[INIT_LMB_REGIONS + 1]; #define LMB_ERROR (~(phys_addr_t)0) +/* inline so we don't get a warning when pr_debug is compiled out */ +static inline const char *lmb_type_name (struct lmb_type *type) +{ + if (type == &lmb.memory) + return "memory"; + else if (type == &lmb.reserved) + return "reserved"; + else + return "unknown"; +} + /** Address comparison utilities */ @@ -156,6 +168,73 @@ static void lmb_coalesce_regions(struct lmb_type *type, lmb_remove_region(type, r2); } +/* Defined below but needed now */ +static long lmb_add_region (struct lmb_type *type,phys_addr_t base,phys_addr_t size); + +static int lmb_double_array (struct lmb_type *type) +{ + struct lmb_region *new_array, *old_array; + phys_addr_t old_size, new_size, addr; + int use_slab = slab_is_available(); + + pr_debug("lmb: %s array full, doubling...", lmb_type_name(type)); + + /* Calculate new doubled size */ + old_size = type->max * sizeof(struct lmb_region); + new_size = old_size << 1; + + /* Try to find some space for it. + * + * WARNING: We assume that either slab_is_available() and we use it or + * we use LMB for allocations. That means that this is unsafe to use + * when bootmem is currently active (unless bootmem itself is implemented + * on top of LMB which isn't the case yet) + * + * This should however not be an issue for now, as we currently only + * call into LMB while it's still active, or much later when slab is + * active for memory hotplug operations + */ + if (use_slab) { + new_array = kmalloc(new_size, GFP_KERNEL); + addr = new_array == NULL ? LMB_ERROR : __pa(new_array); + } else + addr = lmb_find_base(new_size, sizeof(phys_addr_t), LMB_ALLOC_ACCESSIBLE); + if (addr == LMB_ERROR) { + pr_err("lmb: Failed to double %s array from %ld to %ld entries !\n", + lmb_type_name(type), type->max, type->max * 2); + return -1; + } + new_array = __va(addr); + + /* Found space, we now need to move the array over before + * we add the reserved region since it may be our reserved + * array itself that is full. + */ + memcpy(new_array, type->regions, old_size); + memset(new_array + type->max, 0, old_size); + old_array = type->regions; + type->regions = new_array; + type->max <<= 1; + + /* If we use SLAB that's it, we are done */ + if (use_slab) + return 0; + + /* Add the new reserved region now. Should not fail ! */ + BUG_ON(lmb_add_region(&lmb.reserved, addr, new_size) < 0); + + /* If the array wasn't our static init one, then free it. We only do + * that before SLAB is available as later on, we don't know whether + * to use kfree or free_bootmem_pages(). Shouldn't be a big deal + * anyways + */ + if (old_array != lmb_memory_init_regions && + old_array != lmb_reserved_init_regions) + lmb_free(__pa(old_array), old_size); + + return 0; +} + static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t size) { unsigned long coalesced = 0;@@ -196,7 +275,11 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t if (coalesced) return coalesced; - if (type->cnt >= type->max) + + /* If we are out of space, we fail. It's too late to resize the array + * but then this shouldn't have happened in the first place. + */ + if (WARN_ON(type->cnt >= type->max)) return -1; /* Couldn't coalesce the LMB, so add it to the sorted table. */
lib/lmb.c static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t · line 306
Quoted 7 lines
lib/lmb.c +4
-0@@ -217,6 +300,14 @@ static long lmb_add_region(struct lmb_type *type, phys_addr_t base, phys_addr_t } type->cnt++; + /* The array is full ? Try to resize it. If that fails, we undo + * our allocation and return an error + */ + if (type->cnt == type->max && lmb_double_array(type)) {
you need to pass base, base+size with lmb_double_array()
otherwise when you are using lmb_reserve(base, size), double_array() array could have chance to get
new buffer that is overlapped with [base, base + size).
to keep it simple, should check_double_array() after lmb_reserve, lmb_add, lmb_free (yes, that need it too).
that was suggested by Michael Ellerman.
YH
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 19/25] lmb: Add array resizing support
Quoted 10 lines · Yinghai Lu
you need to pass base, base+size with
lmb_double_array()otherwise when you are using
lmb_reserve(base, size),double_array()array could have chance to get
new buffer that is overlapped with [base, base + size).to keep it simple, should
check_double_array()afterlmb_reserve,lmb_add,lmb_free(yes, that need it too).
that was suggested by Michael Ellerman.
No. You may notice that I addressed this problem by moving the call to lmb_double_array() to -after- we record the entry in the array, so it shouldn't be able to pickup the same one.
I dislike the idea of sprinkling the check for resize everywhere at the top level.
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 19/25] lmb: Add array resizing support
Quoted 7 lines · Benjamin Herrenschmidt
Quoted 6 lines · Yinghai Lu
you need to pass base, base+size with
lmb_double_array()otherwise when you are using
lmb_reserve(base, size),double_array()array could have chance to get
new buffer that is overlapped with [base, base + size).
Quoted 8 lines
Quoted 4 lines
to keep it simple, should
check_double_array()afterlmb_reserve,lmb_add,lmb_free(yes, that need it too).
that was suggested by Michael Ellerman.No. You may notice that I addressed this problem by moving the call to
lmb_double_array()to -after- we record the entry in the array, so it shouldn't be able to pickup the same one.
oh, you are right.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 12/25] lmb: Move lmb arrays to static storage in lmb.c and make their size a variable
In reply to Benjamin Herrenschmidtinclude/linux/lmb.h line 24
Quoted 13 lines · Benjamin Herrenschmidt
include/linux/lmb.h +1
-1@@ -18,7 +18,7 @@#include <asm/lmb.h> -#define MAX_LMB_REGIONS 128 +#define INIT_LMB_REGIONS 128struct lmb_region {phys_addr_t base;
Perhaps it would be better to weight this against MAX_ACTIVE_REGIONS for the ARCH_POPULATES_NODE_MAP case? The early node map is already using that size, at least.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 12/25] lmb: Move lmb arrays to static storage in lmb.c and make their size a variable
Quoted 3 lines · Paul Mundt
Perhaps it would be better to weight this against
MAX_ACTIVE_REGIONSfor theARCH_POPULATES_NODE_MAPcase? The early node map is already using that size, at least.
Well, one of the next patches implement dynamic resize of the LMB array so I was actually thinking about shrinking the static one to a minimum...
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 05/25] lmb: Factor the lowest level alloc function
In reply to Benjamin Herrenschmidtlib/lmb.c u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr) · line 412
Quoted 9 lines · Benjamin Herrenschmidt
lib/lmb.c +4
-0@@ -396,33 +406,24 @@ u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr)if (max_addr ==LMB_ALLOC_ANYWHERE ) max_addr =LMB_REAL_LIMIT ; +/* Pump up max_addr */ +if (max_addr ==LMB_ALLOC_ANYWHERE ) + max_addr = ~(u64 )0 ; +
That if is pretty useless as you set max_addr to LMB_REAL_LIMIT right above.
Thanks,
tglx
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 05/25] lmb: Factor the lowest level alloc function
Quoted 5 lines · Thomas Gleixner
Review u64 __init __lmb_alloc_base(u64 size, u64
Quoted 1 line
(quoted diff) +0
-0@@ -396,33 +406,24 @@ u64 __init __lmb_alloc_base(u64 size, u64align, u64
max_addr)Quoted 3 lines
lib/lmb.c +0
-0if (max_addr ==LMB_ALLOC_ANYWHERE ) max_addr =LMB_REAL_LIMIT ;
Quoted 7 lines
lib/lmb.c u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr) · line 412
Quoted 4 lines
lib/lmb.c +4
-0+/* Pump up max_addr */ +if (max_addr ==LMB_ALLOC_ANYWHERE ) + max_addr = ~(u64 )0 ; +That if is pretty useless as you set
max_addrtoLMB_REAL_LIMITright above.
This is a mismerge/mis-rebase of one of my patches actually. I'll dbl check what's up but I think the first hunk should go along with LMB_REAL_LIMIT in favor of the new default limit thing. But we first need to make sure the default is set sensibly and I haven't fixed all archs yet.
I'll dbl check what's up there.
Cheers,
Benm.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 05/25] lmb: Factor the lowest level alloc function
In reply to Thomas GleixnerQuoted 6 lines · Thomas Gleixner
Quoted 5 lines · Benjamin Herrenschmidt
lib/lmb.c +0
-0@@ -396,33 +406,24 @@ u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr)if (max_addr ==LMB_ALLOC_ANYWHERE ) max_addr =LMB_REAL_LIMIT ;
Quoted 7 lines
lib/lmb.c u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr) · line 412
Quoted 4 lines
lib/lmb.c +4
-0+/* Pump up max_addr */ +if (max_addr ==LMB_ALLOC_ANYWHERE ) + max_addr = ~(u64 )0 ; +That if is pretty useless as you set
max_addrtoLMB_REAL_LIMITright above.
Well, actually no :-) LMB_REAL_LIMIT can actually be == to LMB_ALLOC_ANYWHERE which at this stage in the series is still 0, so we must not miss the second bit.
Subsequent patches remove this anyways as LMB_ALLOC_ANYWHERE gets turned into ~0 which makes more sense.
So it's a bit weird looking, but it's correct and transitory code only.
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 03/25] lmb: Introduce for_each_lmb() and new accessors, and use it
In reply to Benjamin Herrenschmidtinclude/linux/lmb.h extern u64 lmb_end_of_DRAM(void); · line 113
Quoted 6 lines · Benjamin Herrenschmidt
include/linux/lmb.h +5
-0+#define for_each_lmb (lmb_type, region) \ + for (reg = lmb.lmb_type.regions; \ + region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ + region++) +
Can you please make this:
#define for_each_lmb(lmb_type, region, __iter) \ for (__iter = lmb.lmb_type.regions; \ region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ region++)
so we do not have the iterator name hardcoded inside the macro body.
Thanks,
tglx
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 03/25] lmb: Introduce for_each_lmb() and new accessors, and use it
Quoted 8 lines · Thomas Gleixner
Can you please make this:
#define for_each_lmb(lmb_type, region, __iter) \ for (__iter = lmb.lmb_type.regions; \ region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ region++)so we do not have the iterator name hardcoded inside the macro body.
Oops, you are right, that's a thinko. I'll fix that.
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Re: [PATCH 03/25] lmb: Introduce for_each_lmb() and new accessors, and use it
In reply to Thomas GleixnerQuoted 8 lines · Thomas Gleixner
include/linux/lmb.h extern u64 lmb_end_of_DRAM(void); · line 113
Quoted 5 lines
include/linux/lmb.h +5
-0+#define for_each_lmb (lmb_type, region) \ + for (reg = lmb.lmb_type.regions; \ + region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ + region++) +Can you please make this:
Quoted 6 lines
#define for_each_lmb(lmb_type, region, __iter) \ for (__iter = lmb.lmb_type.regions; \ region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ region++)so we do not have the iterator name hardcoded inside the macro body.
As a matter of fact, we didn't, but I made a typo on the first one:
- for (reg = lmb.lmb_type.regions; \
+ for (region = lmb.lmb_type.regions; \ region < (lmb.lmb_type.regions + lmb.lmb_type.cnt); \ region++)
I'll fix that in my series.
Cheers,
Ben.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>