[PATCH] mm: init_mm without ptlock

First step in pushing down the page_table_lock. init_mm.page_table_lock has been used throughout the architectures (usually for ioremap): not to serialize kernel address space allocation (that's usually vmlist_lock), but because pud_alloc,pmd_alloc,pte_alloc_kernel expect caller holds it. Reverse that: don't lock or unlock init_mm.page_table_lock in any of the architectures; instead rely on pud_alloc,pmd_alloc,pte_alloc_kernel to take and drop it when allocating a new one, to check lest a racing task already did. Similarly no page_table_lock in vmalloc's map_vm_area. Some temporary ugliness in __pud_alloc and __pmd_alloc: since they also handle user mms, which are converted only by a later patch, for now they have to lock differently according to whether or not it's init_mm. If sources get muddled, there's a danger that an arch source taking init_mm.page_table_lock will be mixed with common source also taking it (or neither take it). So break the rules and make another change, which should break the build for such a mismatch: remove the redundant mm arg from pte_alloc_kernel (ppc64 scrapped its distinct ioremap_mm in 2.6.13). Exceptions: arm26 used pte_alloc_kernel on user mm, now pte_alloc_map; ia64 used pte_alloc_map on init_mm, now pte_alloc_kernel; parisc had bad args to pmd_alloc and pte_alloc_kernel in unused USE_HPPA_IOREMAP code; ppc64 map_io_page forgot to unlock on failure; ppc mmu_mapin_ram and ppc64 im_free took page_table_lock for no good reason. Signed-off-by: Hugh Dickins <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
chewitt · Oct 30, 2005 · 872fec1 · 872fec1
1 parent 46dea3d
commit 872fec1
Show file tree

Hide file tree

Showing 26 changed files with 54 additions and 114 deletions.
diff --git a/arch/alpha/mm/remap.c b/arch/alpha/mm/remap.c
@@ -2,7 +2,6 @@
 #include <asm/pgalloc.h>
 #include <asm/cacheflush.h>
 
-/* called with the page_table_lock held */
 static inline void 
 remap_area_pte(pte_t * pte, unsigned long address, unsigned long size, 
 	       unsigned long phys_addr, unsigned long flags)
@@ -31,7 +30,6 @@ remap_area_pte(pte_t * pte, unsigned long address, unsigned long size,
 	} while (address && (address < end));
 }
 
-/* called with the page_table_lock held */
 static inline int 
 remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size, 
 	       unsigned long phys_addr, unsigned long flags)
@@ -46,7 +44,7 @@ remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size,
 	if (address >= end)
 		BUG();
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, 
@@ -70,7 +68,6 @@ __alpha_remap_area_pages(unsigned long address, unsigned long phys_addr,
 	flush_cache_all();
 	if (address >= end)
 		BUG();
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pmd_t *pmd;
 		pmd = pmd_alloc(&init_mm, dir, address);
@@ -84,7 +81,6 @@ __alpha_remap_area_pages(unsigned long address, unsigned long phys_addr,
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (address && (address < end));
-	spin_unlock(&init_mm.page_table_lock);
 	return error;
 }
 
diff --git a/arch/arm/mm/consistent.c b/arch/arm/mm/consistent.c
@@ -397,8 +397,6 @@ static int __init consistent_init(void)
 	pte_t *pte;
 	int ret = 0;
 
-	spin_lock(&init_mm.page_table_lock);
-
 	do {
 		pgd = pgd_offset(&init_mm, CONSISTENT_BASE);
 		pmd = pmd_alloc(&init_mm, pgd, CONSISTENT_BASE);
@@ -409,7 +407,7 @@ static int __init consistent_init(void)
 		}
 		WARN_ON(!pmd_none(*pmd));
 
-		pte = pte_alloc_kernel(&init_mm, pmd, CONSISTENT_BASE);
+		pte = pte_alloc_kernel(pmd, CONSISTENT_BASE);
 		if (!pte) {
 			printk(KERN_ERR "%s: no pte tables\n", __func__);
 			ret = -ENOMEM;
@@ -419,8 +417,6 @@ static int __init consistent_init(void)
 		consistent_pte = pte;
 	} while (0);
 
-	spin_unlock(&init_mm.page_table_lock);
-
 	return ret;
 }
 

diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
@@ -75,7 +75,7 @@ remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size,
 
 	pgprot = __pgprot(L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_DIRTY | L_PTE_WRITE | flags);
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, address + phys_addr, pgprot);
@@ -97,7 +97,6 @@ remap_area_pages(unsigned long start, unsigned long phys_addr,
 	phys_addr -= address;
 	dir = pgd_offset(&init_mm, address);
 	BUG_ON(address >= end);
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pmd_t *pmd = pmd_alloc(&init_mm, dir, address);
 		if (!pmd) {
@@ -114,7 +113,6 @@ remap_area_pages(unsigned long start, unsigned long phys_addr,
 		dir++;
 	} while (address && (address < end));
 
-	spin_unlock(&init_mm.page_table_lock);
 	flush_cache_vmap(start, end);
 	return err;
 }

diff --git a/arch/arm26/mm/memc.c b/arch/arm26/mm/memc.c
@@ -92,7 +92,7 @@ pgd_t *get_pgd_slow(struct mm_struct *mm)
 	if (!new_pmd)
 		goto no_pmd;
 
-	new_pte = pte_alloc_kernel(mm, new_pmd, 0);
+	new_pte = pte_alloc_map(mm, new_pmd, 0);
 	if (!new_pte)
 		goto no_pte;
 
@@ -101,6 +101,7 @@ pgd_t *get_pgd_slow(struct mm_struct *mm)
 	init_pte = pte_offset(init_pmd, 0);
 
 	set_pte(new_pte, *init_pte);
+	pte_unmap(new_pte);
 
 	/*
 	 * the page table entries are zeroed

diff --git a/arch/cris/mm/ioremap.c b/arch/cris/mm/ioremap.c
@@ -52,7 +52,7 @@ static inline int remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned lo
 	if (address >= end)
 		BUG();
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, address + phys_addr, prot);
@@ -74,7 +74,6 @@ static int remap_area_pages(unsigned long address, unsigned long phys_addr,
 	flush_cache_all();
 	if (address >= end)
 		BUG();
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pud_t *pud;
 		pmd_t *pmd;
@@ -94,7 +93,6 @@ static int remap_area_pages(unsigned long address, unsigned long phys_addr,
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (address && (address < end));
-	spin_unlock(&init_mm.page_table_lock);
 	flush_tlb_all();
 	return error;
 }

diff --git a/arch/frv/mm/dma-alloc.c b/arch/frv/mm/dma-alloc.c
@@ -55,21 +55,18 @@ static int map_page(unsigned long va, unsigned long pa, pgprot_t prot)
 	pte_t *pte;
 	int err = -ENOMEM;
 
-	spin_lock(&init_mm.page_table_lock);
-
 	/* Use upper 10 bits of VA to index the first level map */
 	pge = pgd_offset_k(va);
 	pue = pud_offset(pge, va);
 	pme = pmd_offset(pue, va);
 
 	/* Use middle 10 bits of VA to index the second-level map */
-	pte = pte_alloc_kernel(&init_mm, pme, va);
+	pte = pte_alloc_kernel(pme, va);
 	if (pte != 0) {
 		err = 0;
 		set_pte(pte, mk_pte_phys(pa & PAGE_MASK, prot));
 	}
 
-	spin_unlock(&init_mm.page_table_lock);
 	return err;
 }
 

diff --git a/arch/i386/mm/ioremap.c b/arch/i386/mm/ioremap.c
@@ -28,7 +28,7 @@ static int ioremap_pte_range(pmd_t *pmd, unsigned long addr,
 	unsigned long pfn;
 
 	pfn = phys_addr >> PAGE_SHIFT;
-	pte = pte_alloc_kernel(&init_mm, pmd, addr);
+	pte = pte_alloc_kernel(pmd, addr);
 	if (!pte)
 		return -ENOMEM;
 	do {
@@ -87,14 +87,12 @@ static int ioremap_page_range(unsigned long addr,
 	flush_cache_all();
 	phys_addr -= addr;
 	pgd = pgd_offset_k(addr);
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		next = pgd_addr_end(addr, end);
 		err = ioremap_pud_range(pgd, addr, next, phys_addr+addr, flags);
 		if (err)
 			break;
 	} while (pgd++, addr = next, addr != end);
-	spin_unlock(&init_mm.page_table_lock);
 	flush_tlb_all();
 	return err;
 }

diff --git a/arch/ia64/mm/init.c b/arch/ia64/mm/init.c
@@ -275,26 +275,21 @@ put_kernel_page (struct page *page, unsigned long address, pgprot_t pgprot)
 
 	pgd = pgd_offset_k(address);		/* note: this is NOT pgd_offset()! */
 
-	spin_lock(&init_mm.page_table_lock);
 	{
 		pud = pud_alloc(&init_mm, pgd, address);
 		if (!pud)
 			goto out;
-
 		pmd = pmd_alloc(&init_mm, pud, address);
 		if (!pmd)
 			goto out;
-		pte = pte_alloc_map(&init_mm, pmd, address);
+		pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			goto out;
-		if (!pte_none(*pte)) {
-			pte_unmap(pte);
+		if (!pte_none(*pte))
 			goto out;
-		}
 		set_pte(pte, mk_pte(page, pgprot));
-		pte_unmap(pte);
 	}
-  out:	spin_unlock(&init_mm.page_table_lock);
+  out:
 	/* no need for flush_tlb */
 	return page;
 }

diff --git a/arch/m32r/mm/ioremap.c b/arch/m32r/mm/ioremap.c
@@ -67,7 +67,7 @@ remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size,
 	if (address >= end)
 		BUG();
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, address + phys_addr, flags);
@@ -90,7 +90,6 @@ remap_area_pages(unsigned long address, unsigned long phys_addr,
 	flush_cache_all();
 	if (address >= end)
 		BUG();
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pmd_t *pmd;
 		pmd = pmd_alloc(&init_mm, dir, address);
@@ -104,7 +103,6 @@ remap_area_pages(unsigned long address, unsigned long phys_addr,
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (address && (address < end));
-	spin_unlock(&init_mm.page_table_lock);
 	flush_tlb_all();
 	return error;
 }

diff --git a/arch/m68k/mm/kmap.c b/arch/m68k/mm/kmap.c
@@ -201,7 +201,7 @@ void *__ioremap(unsigned long physaddr, unsigned long size, int cacheflag)
 			virtaddr += PTRTREESIZE;
 			size -= PTRTREESIZE;
 		} else {
-			pte_dir = pte_alloc_kernel(&init_mm, pmd_dir, virtaddr);
+			pte_dir = pte_alloc_kernel(pmd_dir, virtaddr);
 			if (!pte_dir) {
 				printk("ioremap: no mem for pte_dir\n");
 				return NULL;

diff --git a/arch/m68k/sun3x/dvma.c b/arch/m68k/sun3x/dvma.c
@@ -116,7 +116,7 @@ inline int dvma_map_cpu(unsigned long kaddr,
 			pte_t *pte;
 			unsigned long end3;
 
-			if((pte = pte_alloc_kernel(&init_mm, pmd, vaddr)) == NULL) {
+			if((pte = pte_alloc_kernel(pmd, vaddr)) == NULL) {
 				ret = -ENOMEM;
 				goto out;
 			}

diff --git a/arch/mips/mm/ioremap.c b/arch/mips/mm/ioremap.c
@@ -55,7 +55,7 @@ static inline int remap_area_pmd(pmd_t * pmd, unsigned long address,
 	if (address >= end)
 		BUG();
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, address + phys_addr, flags);
@@ -77,7 +77,6 @@ static int remap_area_pages(unsigned long address, phys_t phys_addr,
 	flush_cache_all();
 	if (address >= end)
 		BUG();
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pud_t *pud;
 		pmd_t *pmd;
@@ -96,7 +95,6 @@ static int remap_area_pages(unsigned long address, phys_t phys_addr,
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (address && (address < end));
-	spin_unlock(&init_mm.page_table_lock);
 	flush_tlb_all();
 	return error;
 }

diff --git a/arch/parisc/kernel/pci-dma.c b/arch/parisc/kernel/pci-dma.c
@@ -114,7 +114,7 @@ static inline int map_pmd_uncached(pmd_t * pmd, unsigned long vaddr,
 	if (end > PGDIR_SIZE)
 		end = PGDIR_SIZE;
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, vaddr);
+		pte_t * pte = pte_alloc_kernel(pmd, vaddr);
 		if (!pte)
 			return -ENOMEM;
 		if (map_pte_uncached(pte, orig_vaddr, end - vaddr, paddr_ptr))

diff --git a/arch/parisc/mm/ioremap.c b/arch/parisc/mm/ioremap.c
@@ -52,7 +52,7 @@ static inline int remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned lo
 	if (address >= end)
 		BUG();
 	do {
-		pte_t * pte = pte_alloc_kernel(NULL, pmd, address);
+		pte_t * pte = pte_alloc_kernel(pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		remap_area_pte(pte, address, end - address, address + phys_addr, flags);
@@ -75,10 +75,9 @@ static int remap_area_pages(unsigned long address, unsigned long phys_addr,
 	flush_cache_all();
 	if (address >= end)
 		BUG();
-	spin_lock(&init_mm.page_table_lock);
 	do {
 		pmd_t *pmd;
-		pmd = pmd_alloc(dir, address);
+		pmd = pmd_alloc(&init_mm, dir, address);
 		error = -ENOMEM;
 		if (!pmd)
 			break;
@@ -89,7 +88,6 @@ static int remap_area_pages(unsigned long address, unsigned long phys_addr,
 		address = (address + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (address && (address < end));
-	spin_unlock(&init_mm.page_table_lock);
 	flush_tlb_all();
 	return error;
 }

diff --git a/arch/ppc/kernel/dma-mapping.c b/arch/ppc/kernel/dma-mapping.c
@@ -335,8 +335,6 @@ static int __init dma_alloc_init(void)
 	pte_t *pte;
 	int ret = 0;
 
-	spin_lock(&init_mm.page_table_lock);
-
 	do {
 		pgd = pgd_offset(&init_mm, CONSISTENT_BASE);
 		pmd = pmd_alloc(&init_mm, pgd, CONSISTENT_BASE);
@@ -347,7 +345,7 @@ static int __init dma_alloc_init(void)
 		}
 		WARN_ON(!pmd_none(*pmd));
 
-		pte = pte_alloc_kernel(&init_mm, pmd, CONSISTENT_BASE);
+		pte = pte_alloc_kernel(pmd, CONSISTENT_BASE);
 		if (!pte) {
 			printk(KERN_ERR "%s: no pte tables\n", __func__);
 			ret = -ENOMEM;
@@ -357,8 +355,6 @@ static int __init dma_alloc_init(void)
 		consistent_pte = pte;
 	} while (0);
 
-	spin_unlock(&init_mm.page_table_lock);
-
 	return ret;
 }
 

diff --git a/arch/ppc/mm/4xx_mmu.c b/arch/ppc/mm/4xx_mmu.c
@@ -110,13 +110,11 @@ unsigned long __init mmu_mapin_ram(void)
 		pmd_t *pmdp;
 		unsigned long val = p | _PMD_SIZE_16M | _PAGE_HWEXEC | _PAGE_HWWRITE;
 
-		spin_lock(&init_mm.page_table_lock);
 		pmdp = pmd_offset(pgd_offset_k(v), v);
 		pmd_val(*pmdp++) = val;
 		pmd_val(*pmdp++) = val;
 		pmd_val(*pmdp++) = val;
 		pmd_val(*pmdp++) = val;
-		spin_unlock(&init_mm.page_table_lock);
 
 		v += LARGE_PAGE_SIZE_16M;
 		p += LARGE_PAGE_SIZE_16M;
@@ -127,10 +125,8 @@ unsigned long __init mmu_mapin_ram(void)
 		pmd_t *pmdp;
 		unsigned long val = p | _PMD_SIZE_4M | _PAGE_HWEXEC | _PAGE_HWWRITE;
 
-		spin_lock(&init_mm.page_table_lock);
 		pmdp = pmd_offset(pgd_offset_k(v), v);
 		pmd_val(*pmdp) = val;
-		spin_unlock(&init_mm.page_table_lock);
 
 		v += LARGE_PAGE_SIZE_4M;
 		p += LARGE_PAGE_SIZE_4M;

diff --git a/arch/ppc/mm/pgtable.c b/arch/ppc/mm/pgtable.c
@@ -280,18 +280,16 @@ map_page(unsigned long va, phys_addr_t pa, int flags)
 	pte_t *pg;
 	int err = -ENOMEM;
 
-	spin_lock(&init_mm.page_table_lock);
 	/* Use upper 10 bits of VA to index the first level map */
 	pd = pmd_offset(pgd_offset_k(va), va);
 	/* Use middle 10 bits of VA to index the second-level map */
-	pg = pte_alloc_kernel(&init_mm, pd, va);
+	pg = pte_alloc_kernel(pd, va);
 	if (pg != 0) {
 		err = 0;
 		set_pte_at(&init_mm, va, pg, pfn_pte(pa >> PAGE_SHIFT, __pgprot(flags)));
 		if (mem_init_done)
 			flush_HPTE(0, va, pmd_val(*pd));
 	}
-	spin_unlock(&init_mm.page_table_lock);
 	return err;
 }