mm: bug in set_page_dirty_buffers

This was triggered, but not the fault of, the dirty page accounting
patches. Suitable for -stable as well, after it goes upstream.

Unable to handle kernel NULL pointer dereference at virtual address 0000004c
EIP is at _spin_lock+0x12/0x66
Call Trace:
 [<401766e7>] __set_page_dirty_buffers+0x15/0xc0
 [<401401e7>] set_page_dirty+0x2c/0x51
 [<40140db2>] set_page_dirty_balance+0xb/0x3b
 [<40145d29>] __do_fault+0x1d8/0x279
 [<40147059>] __handle_mm_fault+0x125/0x951
 [<401133f1>] do_page_fault+0x440/0x59f
 [<4034d0c1>] error_code+0x39/0x40
 [<08048a33>] 0x8048a33
 =======================

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

diff --git a/fs/buffer.c b/fs/buffer.c
index 71649ef9b6586696c695e2d80c000ceedaa51282..5b329f0e4f09bc5ce186e904463ba05b873447b8 100644 (file)
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -838,7 +838,10 @@ EXPORT_SYMBOL(mark_buffer_dirty_inode);
  */
 int __set_page_dirty_buffers(struct page *page)
 {
-       struct address_space * const mapping = page->mapping;
+       struct address_space * const mapping = page_mapping(page);
+
+       if (unlikely(!mapping))
+               return !TestSetPageDirty(page);
 
        spin_lock(&mapping->private_lock);
        if (page_has_buffers(page)) {