[PATCH 2.5.41-mm2+] Shared page table bugfix for mprotect

[PATCH 2.5.41-mm2+] Shared page table bugfix for mprotect

[Dave McCracken]

[-- Attachment #1: Type: text/plain, Size: 382 bytes --]


I discovered I hadn't done the right things for mprotect, so here's a patch
that fixes it.  It goes on top of the #ifdef cleanup I sent earlier today.

Dave McCracken

======================================================================
Dave McCracken          IBM Linux Base Kernel Team      1-512-838-3059
dmccr@us.ibm.com                                        T/L   678-3059

[-- Attachment #2: shpte-2.5.41-mm2-2.diff --]
[-- Type: text/plain, Size: 4278 bytes --]

--- 2.5.41-mm2-shsent/./include/linux/mm.h	2002-10-10 10:34:58.000000000 -0500
+++ 2.5.41-mm2-shpte/./include/linux/mm.h	2002-10-10 13:23:30.000000000 -0500
@@ -360,6 +360,7 @@
 
 extern void zap_page_range(struct vm_area_struct *vma, unsigned long address, unsigned long size);
 #ifdef CONFIG_SHAREPTE
+extern pte_t *pte_unshare(struct mm_struct *mm, pmd_t *pmd, unsigned long address);
 extern int share_page_range(struct mm_struct *dst, struct mm_struct *src, struct vm_area_struct *vma, pmd_t **prev_pmd);
 #else
 extern int copy_page_range(struct mm_struct *dst, struct mm_struct *src, struct vm_area_struct *vma);
--- 2.5.41-mm2-shsent/./include/asm-i386/pgtable.h	2002-10-10 10:17:55.000000000 -0500
+++ 2.5.41-mm2-shpte/./include/asm-i386/pgtable.h	2002-10-10 13:14:48.000000000 -0500
@@ -212,6 +212,7 @@
 static inline pte_t pte_mkwrite(pte_t pte)	{ (pte).pte_low |= _PAGE_RW; return pte; }
 static inline int pmd_write(pmd_t pmd)		{ return (pmd).pmd & _PAGE_RW; }
 static inline pmd_t pmd_wrprotect(pmd_t pmd)	{ (pmd).pmd &= ~_PAGE_RW; return pmd; }
+static inline pmd_t pmd_mkwrite(pmd_t pmd)	{ (pmd).pmd |= _PAGE_RW; return pmd; }
 
 static inline  int ptep_test_and_clear_dirty(pte_t *ptep)	{ return test_and_clear_bit(_PAGE_BIT_DIRTY, &ptep->pte_low); }
 static inline  int ptep_test_and_clear_young(pte_t *ptep)	{ return test_and_clear_bit(_PAGE_BIT_ACCESSED, &ptep->pte_low); }
--- 2.5.41-mm2-shsent/./mm/mprotect.c	2002-10-10 10:17:57.000000000 -0500
+++ 2.5.41-mm2-shpte/./mm/mprotect.c	2002-10-10 13:31:14.000000000 -0500
@@ -16,6 +16,7 @@
 #include <linux/fs.h>
 #include <linux/highmem.h>
 #include <linux/security.h>
+#include <linux/rmap-locking.h>
 
 #include <asm/uaccess.h>
 #include <asm/pgalloc.h>
@@ -24,10 +25,11 @@
 #include <asm/tlbflush.h>
 
 static inline void
-change_pte_range(pmd_t *pmd, unsigned long address,
+change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, unsigned long address,
 		unsigned long size, pgprot_t newprot)
 {
 	pte_t * pte;
+	struct page *ptepage;
 	unsigned long end;
 
 	if (pmd_none(*pmd))
@@ -37,11 +39,32 @@
 		pmd_clear(pmd);
 		return;
 	}
-	pte = pte_offset_map(pmd, address);
+	ptepage = pmd_page(*pmd);
+	pte_page_lock(ptepage);
 	address &= ~PMD_MASK;
 	end = address + size;
 	if (end > PMD_SIZE)
 		end = PMD_SIZE;
+
+#ifdef CONFIG_SHAREPTE
+	if (page_count(ptepage) > 1) {
+		if ((address == 0) && (end == PMD_SIZE)) {
+			pmd_t	pmdval = *pmd;
+
+			if (vma->vm_flags & VM_MAYWRITE)
+				pmdval = pmd_mkwrite(pmdval);
+			else
+				pmdval = pmd_wrprotect(pmdval);
+			set_pmd(pmd, pmdval);
+			pte_page_unlock(ptepage);
+			return;
+		}
+		pte = pte_unshare(vma->vm_mm, pmd, address);
+		ptepage = pmd_page(*pmd);
+	} else
+#endif
+		pte = pte_offset_map(pmd, address);
+
 	do {
 		if (pte_present(*pte)) {
 			pte_t entry;
@@ -56,11 +79,12 @@
 		address += PAGE_SIZE;
 		pte++;
 	} while (address && (address < end));
+	pte_page_unlock(ptepage);
 	pte_unmap(pte - 1);
 }
 
 static inline void
-change_pmd_range(pgd_t *pgd, unsigned long address,
+change_pmd_range(struct vm_area_struct *vma, pgd_t *pgd, unsigned long address,
 		unsigned long size, pgprot_t newprot)
 {
 	pmd_t * pmd;
@@ -79,7 +103,7 @@
 	if (end > PGDIR_SIZE)
 		end = PGDIR_SIZE;
 	do {
-		change_pte_range(pmd, address, end - address, newprot);
+		change_pte_range(vma, pmd, address, end - address, newprot);
 		address = (address + PMD_SIZE) & PMD_MASK;
 		pmd++;
 	} while (address && (address < end));
@@ -98,7 +122,7 @@
 		BUG();
 	spin_lock(&current->mm->page_table_lock);
 	do {
-		change_pmd_range(dir, start, end - start, newprot);
+		change_pmd_range(vma, dir, start, end - start, newprot);
 		start = (start + PGDIR_SIZE) & PGDIR_MASK;
 		dir++;
 	} while (start && (start < end));
--- 2.5.41-mm2-shsent/./mm/memory.c	2002-10-10 10:35:34.000000000 -0500
+++ 2.5.41-mm2-shpte/./mm/memory.c	2002-10-10 13:22:29.000000000 -0500
@@ -218,7 +218,7 @@
  * held.
  */
 
-static pte_t *pte_unshare(struct mm_struct *mm, pmd_t *pmd, unsigned long address)
+pte_t *pte_unshare(struct mm_struct *mm, pmd_t *pmd, unsigned long address)
 {
 	pte_t	*src_ptb, *dst_ptb;
 	struct page *oldpage, *newpage, *tmppage;

Fork timing numbers for shared page tables

[Dave McCracken]

--On Thursday, October 10, 2002 12:44:19 -0700 Andrew Morton
<akpm@digeo.com> wrote:

> Be nice to get some compelling benchmark figures onto the
> mailing lists to help push these.  They're pretty late...

I've done some basic timing tests for shared page tables using a simple
fork test I wrote.  It has three modes:

The first mode forks as fast as it can, then calculates how long each fork
took.  This measures the time the fork() system call took.

The second mode adds a wait() for the child after the fork.  The child just
calls exit(0).  This measures how long the child ran.

The third mode adds an exec() in the child of a very small executable,
which just exits.  This adds the exec() time to the mix.

The program also optionally allocates a shared memory object and touches
all the pages in it before the start of the test.  This adds extra pages to
be dealt with by fork/exec/exit.  None of the pages are touched after the
test starts.

I ran this test in three cases, 2.5.41, 2.5.41-mm2 without share, and
2.5.41-mm2 with share.

Now for the results (all times are in ms):

		2.5.41	mm2-unshared	mm2-shared
		------	------------	----------
fork
----

400K		 1.7	 1.6		 0.5
4M		 5.0	 5.0		 3.4
40M		28.4	29.5		 3.4

fork/exit
---------

400K		 1.7	 1.6		 1.6
4M		 4.9	 5.3		 4.1
40M		44.2	45.1		 4.1

fork/exec/exit
--------------

400K		 6.5	 7.5		 7.7
4M		10.3	11.9		10.7
40M		49.3	51.4		10.7


I don't know why exec introduces a small penalty for small tasks. I'm
working on some optimizations that might help.

Dave McCracken

======================================================================
Dave McCracken          IBM Linux Base Kernel Team      1-512-838-3059
dmccr@us.ibm.com                                        T/L   678-3059

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/

Re: Fork timing numbers for shared page tables

[William Lee Irwin III]

On Thu, Oct 10, 2002 at 02:59:07PM -0500, Dave McCracken wrote:
> I ran this test in three cases, 2.5.41, 2.5.41-mm2 without share, and
> 2.5.41-mm2 with share.
> Now for the results (all times are in ms):

Hrm, it'd be nice to see how nicely this does things for things like
500GB-sized processes on 64-bit boxen...


Any chance you could pass this test along for randomized benchmark
type stuff?



Thanks,
Bill
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/

Re: Fork timing numbers for shared page tables

[Benjamin LaHaise]

On Thu, Oct 10, 2002 at 02:59:07PM -0500, Dave McCracken wrote:
> I don't know why exec introduces a small penalty for small tasks. I'm
> working on some optimizations that might help.

Compare against vfork() to see what kind of best case is possible, and 
how much of the overhead in small tasks is just in non-vm overhead.

		-ben
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/