* [PATCH] vmalloc/ioremap race fix.
@ 2002-09-13 22:08 David Woodhouse
0 siblings, 0 replies; only message in thread
From: David Woodhouse @ 2002-09-13 22:08 UTC (permalink / raw)
To: torvalds; +Cc: linux-kernel
CPU0 CPU1
---- ----
vfree() vmalloc()
remove vm_struct from list.
reallocate same virtual range
set up page tables
tear down page tables
So we leave the vm_struct in the list but marked as being deleted, while we
tear down the page tables.
===== fs/proc/kcore.c 1.4 vs edited =====
--- 1.4/fs/proc/kcore.c Fri Mar 15 11:18:21 2002
+++ edited/fs/proc/kcore.c Fri Sep 13 22:47:36 2002
@@ -119,12 +119,12 @@
*num_vma = 0;
size = ((size_t)high_memory - PAGE_OFFSET + PAGE_SIZE);
- if (!vmlist) {
+ if (list_empty(&vmlist)) {
*elf_buflen = PAGE_SIZE;
return (size);
}
- for (m=vmlist; m; m=m->next) {
+ list_for_each_entry(m, &vmlist, list) {
try = (size_t)m->addr + m->size;
if (try > size)
size = try;
@@ -244,7 +244,7 @@
phdr->p_align = PAGE_SIZE;
/* setup ELF PT_LOAD program header for every vmalloc'd area */
- for (m=vmlist; m; m=m->next) {
+ list_for_each_entry(m, &vmlist, list) {
if (m->flags & VM_IOREMAP) /* don't dump ioremap'd stuff! (TA) */
continue;
@@ -404,11 +404,15 @@
memset(elf_buf, 0, tsz);
read_lock(&vmlist_lock);
- for (m=vmlist; m && cursize; m=m->next) {
+ list_for_each_entry(m, &vmlist, list) {
unsigned long vmstart;
unsigned long vmsize;
- unsigned long msize = m->size - PAGE_SIZE;
+ unsigned long msize;
+ if (!cursize)
+ break;
+
+ msize = m->size - PAGE_SIZE;
if (((unsigned long)m->addr + msize) <
curstart)
continue;
===== mm/vmalloc.c 1.19 vs edited =====
--- 1.19/mm/vmalloc.c Mon Aug 19 01:37:40 2002
+++ edited/mm/vmalloc.c Fri Sep 13 22:28:09 2002
@@ -19,7 +19,7 @@
rwlock_t vmlist_lock = RW_LOCK_UNLOCKED;
-struct vm_struct *vmlist;
+LIST_HEAD(vmlist);
static inline void unmap_area_pte(pmd_t *pmd, unsigned long address,
unsigned long size)
@@ -190,7 +190,7 @@
*/
struct vm_struct *get_vm_area(unsigned long size, unsigned long flags)
{
- struct vm_struct **p, *tmp, *area;
+ struct vm_struct *tmp, *area;
unsigned long addr = VMALLOC_START;
area = kmalloc(sizeof(*area), GFP_KERNEL);
@@ -203,7 +203,7 @@
size += PAGE_SIZE;
write_lock(&vmlist_lock);
- for (p = &vmlist; (tmp = *p) ;p = &tmp->next) {
+ list_for_each_entry(tmp, &vmlist, list) {
if ((size + addr) < addr)
goto out;
if (size + addr <= (unsigned long)tmp->addr)
@@ -214,8 +214,7 @@
}
found:
- area->next = *p;
- *p = area;
+ list_add_tail(&area->list, &tmp->list);
area->flags = flags;
area->addr = (void *)addr;
@@ -244,18 +243,23 @@
*/
struct vm_struct *remove_vm_area(void *addr)
{
- struct vm_struct **p, *tmp;
+ struct vm_struct *tmp;
write_lock(&vmlist_lock);
- for (p = &vmlist ; (tmp = *p) ;p = &tmp->next) {
+ list_for_each_entry(tmp, &vmlist, list) {
if (tmp->addr == addr)
goto found;
}
write_unlock(&vmlist_lock);
return NULL;
-found:
- *p = tmp->next;
+ found:
+ if (tmp->flags & VM_DELETING) {
+ printk(KERN_ERR "Trying to vfree() region being freed (%p)\n", addr);
+ tmp = NULL;
+ }
+ else
+ tmp->flags |= VM_DELETING;
write_unlock(&vmlist_lock);
return tmp;
}
@@ -293,6 +297,10 @@
kfree(area->pages);
}
+ write_lock(&vmlist_lock);
+ list_del(&area->list);
+ write_unlock(&vmlist_lock);
+
kfree(area);
return;
}
@@ -302,7 +310,7 @@
*
* @addr: memory base address
*
- * Free the virtually continguos memory area starting at @addr, as
+ * Free the virtually contiguous memory area starting at @addr, as
* obtained from vmalloc(), vmalloc_32() or __vmalloc().
*
* May not be called in interrupt context.
@@ -317,7 +325,7 @@
*
* @addr: memory base address
*
- * Free the virtually continguos memory area starting at @addr,
+ * Free the virtually contiguous memory area starting at @addr,
* which was created from the page array passed to vmap().
*
* May not be called in interrupt context.
@@ -328,12 +336,12 @@
}
/**
- * vmap - map an array of pages into virtually continguos space
+ * vmap - map an array of pages into virtually contiguous space
*
* @pages: array of page pointers
* @count: number of pages to map
*
- * Maps @count pages from @pages into continguos kernel virtual
+ * Maps @count pages from @pages into contiguous kernel virtual
* space.
*/
void *vmap(struct page **pages, unsigned int count)
@@ -355,14 +363,14 @@
}
/**
- * __vmalloc - allocate virtually continguos memory
+ * __vmalloc - allocate virtually contiguous memory
*
* @size: allocation size
* @gfp_mask: flags for the page level allocator
* @prot: protection mask for the allocated pages
*
* Allocate enough pages to cover @size from the page level
- * allocator with @gfp_mask flags. Map them into continguos
+ * allocator with @gfp_mask flags. Map them into contiguous
* kernel virtual space, using a pagetable protection of @prot.
*/
void *__vmalloc(unsigned long size, int gfp_mask, pgprot_t prot)
@@ -410,12 +418,12 @@
}
/**
- * vmalloc - allocate virtually continguos memory
+ * vmalloc - allocate virtually contiguous memory
*
* @size: allocation size
*
* Allocate enough pages to cover @size from the page level
- * allocator and map them into continguos kernel virtual space.
+ * allocator and map them into contiguous kernel virtual space.
*
* For tight cotrol over page level allocator and protection flags
* use __vmalloc() instead.
@@ -426,12 +434,12 @@
}
/**
- * vmalloc_32 - allocate virtually continguos memory (32bit addressable)
+ * vmalloc_32 - allocate virtually contiguous memory (32bit addressable)
*
* @size: allocation size
*
* Allocate enough 32bit PA addressable pages to cover @size from the
- * page level allocator and map them into continguos kernel virtual space.
+ * page level allocator and map them into contiguous kernel virtual space.
*/
void *vmalloc_32(unsigned long size)
{
@@ -449,7 +457,7 @@
count = -(unsigned long) addr;
read_lock(&vmlist_lock);
- for (tmp = vmlist; tmp; tmp = tmp->next) {
+ list_for_each_entry(tmp, &vmlist, list) {
vaddr = (char *) tmp->addr;
if (addr >= vaddr + tmp->size - PAGE_SIZE)
continue;
@@ -487,7 +495,7 @@
count = -(unsigned long) addr;
read_lock(&vmlist_lock);
- for (tmp = vmlist; tmp; tmp = tmp->next) {
+ list_for_each_entry(tmp, &vmlist, list) {
vaddr = (char *) tmp->addr;
if (addr >= vaddr + tmp->size - PAGE_SIZE)
continue;
===== include/linux/vmalloc.h 1.7 vs edited =====
--- 1.7/include/linux/vmalloc.h Mon Aug 5 20:05:22 2002
+++ edited/include/linux/vmalloc.h Fri Sep 13 22:28:16 2002
@@ -2,11 +2,13 @@
#define _LINUX_VMALLOC_H
#include <linux/spinlock.h>
+#include <linux/list.h>
/* bits in vm_struct->flags */
#define VM_IOREMAP 0x00000001 /* ioremap() and friends */
#define VM_ALLOC 0x00000002 /* vmalloc() */
#define VM_MAP 0x00000004 /* vmap()ed pages */
+#define VM_DELETING 0x00000008 /* Being removed */
struct vm_struct {
void *addr;
@@ -15,7 +17,7 @@
struct page **pages;
unsigned int nr_pages;
unsigned long phys_addr;
- struct vm_struct *next;
+ struct list_head list;
};
/*
@@ -39,9 +41,9 @@
extern void unmap_vm_area(struct vm_struct *area);
/*
- * Internals. Dont't use..
+ * Internals. Don't use.
*/
extern rwlock_t vmlist_lock;
-extern struct vm_struct *vmlist;
+extern struct list_head vmlist;
#endif /* _LINUX_VMALLOC_H */
--
dwmw2
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2002-09-13 22:03 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-09-13 22:08 [PATCH] vmalloc/ioremap race fix David Woodhouse
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).