[tip:kmemcheck] x86: introduce bootmem_state

From: Yinghai Lu
Date: Sun Mar 08 2009 - 14:57:08 EST

Next message: Ingo Molnar: "Re: [tip:kmemcheck] x86: introduce bootmem_state"
Previous message: Mhayk Whandson: "Re: doubt about path_walk() function"
Next in thread: Ingo Molnar: "Re: [tip:kmemcheck] x86: introduce bootmem_state"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Commit-ID: 4c76c04421dfe7be3e5a1d8ab1b2a3be0b02558e
Gitweb: http://git.kernel.org/tip/4c76c04421dfe7be3e5a1d8ab1b2a3be0b02558e
Author: "Yinghai Lu" <yinghai@xxxxxxxxxx>
AuthorDate: Fri, 6 Mar 2009 16:49:00 -0800
Commit: Ingo Molnar <mingo@xxxxxxx>
CommitDate: Sun, 8 Mar 2009 19:43:09 +0100

x86: introduce bootmem_state

Impact: cleanup

extend after_bootmem and after_init_bootmem to bootmem_state
and will have BEFORE_BOOTMEM, DURING_BOOTMEM, AFTER_BOOTMEM

v2: style changes according to ingo
v3: move bootmem_state declaring to arch/x86/include

Signed-off-by: Yinghai Lu <yinghai@xxxxxxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
LKML-Reference: <49B1C47C.5080401@xxxxxxxxxx>
Signed-off-by: Ingo Molnar <mingo@xxxxxxx>

---
arch/x86/include/asm/page_types.h | 8 ++++++++
arch/x86/kernel/setup.c | 1 +
arch/x86/mm/init.c | 13 +++++++------
arch/x86/mm/init_32.c | 28 ++++++++++++++++++++--------
arch/x86/mm/init_64.c | 33 +++++++++++++++++++--------------
5 files changed, 55 insertions(+), 28 deletions(-)

diff --git a/arch/x86/include/asm/page_types.h b/arch/x86/include/asm/page_types.h
index 826ad37..f8b9c88 100644
--- a/arch/x86/include/asm/page_types.h
+++ b/arch/x86/include/asm/page_types.h
@@ -40,6 +40,14 @@

#ifndef __ASSEMBLY__

+enum bootmem_state {
+ BEFORE_BOOTMEM,
+ DURING_BOOTMEM,
+ AFTER_BOOTMEM
+};
+
+extern enum bootmem_state bootmem_state;
+
extern int page_is_ram(unsigned long pagenr);
extern int devmem_is_allowed(unsigned long pagenr);

diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c
index f28c56e..ce9e888 100644
--- a/arch/x86/kernel/setup.c
+++ b/arch/x86/kernel/setup.c
@@ -885,6 +885,7 @@ void __init setup_arch(char **cmdline_p)
#endif

initmem_init(0, max_pfn);
+ bootmem_state = DURING_BOOTMEM;

#ifdef CONFIG_ACPI_SLEEP
/*
diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c
index a8d2933..f98b501 100644
--- a/arch/x86/mm/init.c
+++ b/arch/x86/mm/init.c
@@ -14,7 +14,7 @@ unsigned long __initdata e820_table_start;
unsigned long __meminitdata e820_table_end;
unsigned long __meminitdata e820_table_top;

-int after_bootmem;
+enum bootmem_state bootmem_state = BEFORE_BOOTMEM;

int direct_gbpages
#ifdef CONFIG_DIRECT_GBPAGES
@@ -143,7 +143,7 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,

printk(KERN_INFO "init_memory_mapping: %016lx-%016lx\n", start, end);

- if (!after_bootmem)
+ if (bootmem_state == BEFORE_BOOTMEM)
init_gbpages();

#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KMEMCHECK)
@@ -283,7 +283,7 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,
* memory mapped. Unfortunately this is done currently before the
* nodes are discovered.
*/
- if (!after_bootmem)
+ if (bootmem_state == BEFORE_BOOTMEM)
find_early_table_space(end, use_pse, use_gbpages);

#ifdef CONFIG_X86_32
@@ -304,16 +304,17 @@ unsigned long __init_refok init_memory_mapping(unsigned long start,
#endif

#ifdef CONFIG_X86_64
- if (!after_bootmem)
+ if (bootmem_state == BEFORE_BOOTMEM)
mmu_cr4_features = read_cr4();
#endif
__flush_tlb_all();

- if (!after_bootmem && e820_table_end > e820_table_start)
+ if (bootmem_state == BEFORE_BOOTMEM &&
+ e820_table_end > e820_table_start)
reserve_early(e820_table_start << PAGE_SHIFT,
e820_table_end << PAGE_SHIFT, "PGTABLE");

- if (!after_bootmem)
+ if (bootmem_state == BEFORE_BOOTMEM)
early_memtest(start, end);

return ret >> PAGE_SHIFT;
diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c
index ccadfb1..34ef5c7 100644
--- a/arch/x86/mm/init_32.c
+++ b/arch/x86/mm/init_32.c
@@ -82,14 +82,20 @@ static __init void *alloc_low_page(void)
static pmd_t * __init one_md_table_init(pgd_t *pgd)
{
pud_t *pud;
- pmd_t *pmd_table;
+ pmd_t *pmd_table = NULL;

#ifdef CONFIG_X86_PAE
if (!(pgd_val(*pgd) & _PAGE_PRESENT)) {
- if (after_bootmem)
+ switch (bootmem_state) {
+ case DURING_BOOTMEM:
pmd_table = (pmd_t *)alloc_bootmem_low_pages(PAGE_SIZE);
- else
+ break;
+ case BEFORE_BOOTMEM:
pmd_table = (pmd_t *)alloc_low_page();
+ break;
+ default:
+ panic("after bootmem call one_md_table_init\n");
+ }
paravirt_alloc_pmd(&init_mm, __pa(pmd_table) >> PAGE_SHIFT);
set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
pud = pud_offset(pgd, 0);
@@ -113,15 +119,21 @@ static pte_t * __init one_page_table_init(pmd_t *pmd)
if (!(pmd_val(*pmd) & _PAGE_PRESENT)) {
pte_t *page_table = NULL;

- if (after_bootmem) {
+ switch (bootmem_state) {
+ case DURING_BOOTMEM:
#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KMEMCHECK)
page_table = (pte_t *) alloc_bootmem_pages(PAGE_SIZE);
#endif
if (!page_table)
page_table =
(pte_t *)alloc_bootmem_low_pages(PAGE_SIZE);
- } else
+ break;
+ case BEFORE_BOOTMEM:
page_table = (pte_t *)alloc_low_page();
+ break;
+ default:
+ panic("after bootmem call one_page_table_init\n");
+ }

paravirt_alloc_pte(&init_mm, __pa(page_table) >> PAGE_SHIFT);
set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE));
@@ -169,7 +181,7 @@ static pte_t *__init page_table_kmap_check(pte_t *pte, pmd_t *pmd,
pte_t *newpte;
int i;

- BUG_ON(after_bootmem);
+ BUG_ON(bootmem_state != BEFORE_BOOTMEM);
newpte = alloc_low_page();
for (i = 0; i < PTRS_PER_PTE; i++)
set_pte(newpte + i, pte[i]);
@@ -855,8 +867,6 @@ void __init setup_bootmem_allocator(void)
bootmap = setup_node_bootmem(nodeid, start_pfn, end_pfn,
bootmap);
}
-
- after_bootmem = 1;
}

/*
@@ -923,6 +933,8 @@ void __init mem_init(void)
/* this will put all low memory onto the freelists */
totalram_pages += free_all_bootmem();

+ bootmem_state = AFTER_BOOTMEM;
+
reservedpages = 0;
for (tmp = 0; tmp < max_low_pfn; tmp++)
/*
diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c
index e127f2e..20ade92 100644
--- a/arch/x86/mm/init_64.c
+++ b/arch/x86/mm/init_64.c
@@ -140,20 +140,26 @@ __setup("noexec32=", nonx32_setup);

/*
* NOTE: This function is marked __ref because it calls __init function
- * (alloc_bootmem_pages). It's safe to do it ONLY when after_bootmem == 0.
+ * (alloc_bootmem_pages). It's safe to do it ONLY when DURING_BOOTMEM.
*/
static __ref void *spp_getpage(void)
{
- void *ptr;
+ void *ptr = NULL;

- if (after_bootmem)
+ switch (bootmem_state) {
+ case AFTER_BOOTMEM:
ptr = (void *) get_zeroed_page(GFP_ATOMIC | __GFP_NOTRACK);
- else
+ break;
+ case DURING_BOOTMEM:
ptr = alloc_bootmem_pages(PAGE_SIZE);
+ break;
+ default:
+ panic("calling spp_getpage before bootmem\n");
+ }

if (!ptr || ((unsigned long)ptr & ~PAGE_MASK)) {
panic("set_pte_phys: cannot allocate page data %s\n",
- after_bootmem ? "after bootmem" : "");
+ bootmem_state == AFTER_BOOTMEM ? "after bootmem" : "");
}

pr_debug("spp_getpage %p\n", ptr);
@@ -320,16 +326,17 @@ void __init cleanup_highmap(void)

static __ref void *alloc_low_page(unsigned long *phys)
{
- unsigned long pfn = e820_table_end++;
+ unsigned long pfn;
void *adr;

- if (after_bootmem) {
+ if (bootmem_state == AFTER_BOOTMEM) {
adr = (void *)get_zeroed_page(GFP_ATOMIC | __GFP_NOTRACK);
*phys = __pa(adr);

return adr;
}

+ pfn = e820_table_end++;
if (pfn >= e820_table_top)
panic("alloc_low_page: ran out of memory");

@@ -341,7 +348,7 @@ static __ref void *alloc_low_page(unsigned long *phys)

static __ref void unmap_low_page(void *adr)
{
- if (after_bootmem)
+ if (bootmem_state == AFTER_BOOTMEM)
return;

early_iounmap(adr, PAGE_SIZE);
@@ -360,7 +367,7 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
for(i = pte_index(addr); i < PTRS_PER_PTE; i++, addr += PAGE_SIZE, pte++) {

if (addr >= end) {
- if (!after_bootmem) {
+ if (bootmem_state != AFTER_BOOTMEM) {
for(; i < PTRS_PER_PTE; i++, pte++)
set_pte(pte, __pte(0));
}
@@ -416,7 +423,7 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
pgprot_t new_prot = prot;

if (address >= end) {
- if (!after_bootmem) {
+ if (bootmem_state != AFTER_BOOTMEM) {
for (; i < PTRS_PER_PMD; i++, pmd++)
set_pmd(pmd, __pmd(0));
}
@@ -502,7 +509,7 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
if (addr >= end)
break;

- if (!after_bootmem &&
+ if (bootmem_state != AFTER_BOOTMEM &&
!e820_any_mapped(addr, addr+PUD_SIZE, 0)) {
set_pud(pud, __pud(0));
continue;
@@ -693,8 +700,6 @@ void __init mem_init(void)

/* clear_bss() already clear the empty_zero_page */

- reservedpages = 0;
-
/* this will put all low memory onto the freelists */
#ifdef CONFIG_NUMA
totalram_pages = numa_free_all_bootmem();
@@ -702,9 +707,9 @@ void __init mem_init(void)
totalram_pages = free_all_bootmem();
#endif

+ bootmem_state = AFTER_BOOTMEM;
absent_pages = absent_pages_in_range(0, max_pfn);
reservedpages = max_pfn - totalram_pages - absent_pages;
- after_bootmem = 1;

codesize = (unsigned long) &_etext - (unsigned long) &_text;
datasize = (unsigned long) &_edata - (unsigned long) &_etext;
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Ingo Molnar: "Re: [tip:kmemcheck] x86: introduce bootmem_state"
Previous message: Mhayk Whandson: "Re: doubt about path_walk() function"
Next in thread: Ingo Molnar: "Re: [tip:kmemcheck] x86: introduce bootmem_state"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]