mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* Hugetlb page bug fix for i386 in PAE mode
       [not found] <B05667366EE6204181EABE9C1B1C0EB58022A3@scsmsx401.amr.corp.intel.com>
@ 2004-06-16 22:13 ` Chen, Kenneth W
  0 siblings, 0 replies; 2+ messages in thread
From: Chen, Kenneth W @ 2004-06-16 22:13 UTC (permalink / raw)
  To: linux-kernel

Hate my email client, just about to throw my *&^%$ out of the windows.
Resent with no line wrap, etc.

-----Original Message-----

Hit a bug check when unmap a hugetlb vma in PAE mode on i386 (and x86-64).

 Bad page state at free_hot_cold_page (in process 'a.out', page c165cc40)
 flags:0x20000000 mapping:f75e1d00 mapped:0 count:0
 Backtrace:
 Call Trace:
  [<c0133e0d>] bad_page+0x79/0x9e
  [<c0134550>] free_hot_cold_page+0x71/0xfa
  [<c0115d60>] unmap_hugepage_range+0xa3/0xbf
  [<c013d375>] unmap_vmas+0xac/0x252
  [<c0117691>] default_wake_function+0x0/0xc
  [<c0140bea>] unmap_region+0xd8/0x145
  [<c0140f2d>] do_munmap+0xfc/0x14d
  [<c01b8a56>] sys_shmdt+0xa5/0x126
  [<c010a2ad>] sys_ipc+0x23c/0x27f
  [<c014a85e>] sys_write+0x38/0x59
  [<c0103e1b>] syscall_call+0x7/0xb

It turns out there is a bug in hugetlb_prefault(): with 3 level page table,
huge_pte_alloc() might return a pmd that points to a PTE page.  It happens
if the virtual address for hugetlb mmap is recycled from previously used
normal page mmap.  free_pgtables() might not scrub the pmd entry on munmap
and hugetlb_prefault skips on any pmd presence regardless what type it is.
Patch to fix the bug.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>

diff -Nurp linux-2.6.7/arch/i386/mm/hugetlbpage.c linux-2.6.7.htlb/arch/i386/mm/hugetlbpage.c
--- linux-2.6.7/arch/i386/mm/hugetlbpage.c	2004-06-16 18:32:30.000000000 -0700
+++ linux-2.6.7.htlb/arch/i386/mm/hugetlbpage.c	2004-06-16 18:34:32.000000000 -0700
@@ -244,8 +244,15 @@ int hugetlb_prefault(struct address_spac
 			ret = -ENOMEM;
 			goto out;
 		}
-		if (!pte_none(*pte))
-			continue;
+
+		if (!pte_none(*pte)) {
+			pmd_t *pmd = (pmd_t *) pte;
+
+			page = pmd_page(*pmd);
+			pmd_clear(pmd);
+			dec_page_state(nr_page_table_pages);
+			page_cache_release(page);
+		}

 		idx = ((addr - vma->vm_start) >> HPAGE_SHIFT)
 			+ (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT));



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Hugetlb page bug fix for i386 in PAE mode
@ 2004-06-16 22:07 Chen, Kenneth W
  0 siblings, 0 replies; 2+ messages in thread
From: Chen, Kenneth W @ 2004-06-16 22:07 UTC (permalink / raw)
  To: linux-kernel

Hit a bug check when unmap a hugetlb vma in PAE mode on i386 (and
x86-64).

 Bad page state at free_hot_cold_page (in process 'a.out', page
c165cc40)
 flags:0x20000000 mapping:f75e1d00 mapped:0 count:0
 Backtrace:
 Call Trace:
  [<c0133e0d>] bad_page+0x79/0x9e
  [<c0134550>] free_hot_cold_page+0x71/0xfa
  [<c0115d60>] unmap_hugepage_range+0xa3/0xbf
  [<c013d375>] unmap_vmas+0xac/0x252
  [<c0117691>] default_wake_function+0x0/0xc
  [<c0140bea>] unmap_region+0xd8/0x145
  [<c0140f2d>] do_munmap+0xfc/0x14d
  [<c01b8a56>] sys_shmdt+0xa5/0x126
  [<c010a2ad>] sys_ipc+0x23c/0x27f
  [<c014a85e>] sys_write+0x38/0x59
  [<c0103e1b>] syscall_call+0x7/0xb

It turns out there is a bug in hugetlb_prefault(): with 3 level page
table,
huge_pte_alloc() might return a pmd that points to a PTE page.  It
happens
if the virtual address for hugetlb mmap is recycled from previously used
normal page mmap.  free_pgtables() might not scrub the pmd entry on
munmap
and hugetlb_prefault skips on any pmd presence regardless what type it
is.
Patch to fix the bug.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>

diff -Nurp linux-2.6.7/arch/i386/mm/hugetlbpage.c
linux-2.6.7.htlb/arch/i386/mm/hugetlbpage.c
--- linux-2.6.7/arch/i386/mm/hugetlbpage.c	2004-06-16
18:32:30.000000000 -0700
+++ linux-2.6.7.htlb/arch/i386/mm/hugetlbpage.c	2004-06-16
18:34:32.000000000 -0700
@@ -244,8 +244,15 @@ int hugetlb_prefault(struct address_spac
 			ret = -ENOMEM;
 			goto out;
 		}
-		if (!pte_none(*pte))
-			continue;
+
+		if (!pte_none(*pte)) {
+			pmd_t *pmd = (pmd_t *) pte;
+
+			page = pmd_page(*pmd);
+			pmd_clear(pmd);
+			dec_page_state(nr_page_table_pages);
+			page_cache_release(page);
+		}
 
 		idx = ((addr - vma->vm_start) >> HPAGE_SHIFT)
 			+ (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT));

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2004-06-16 22:14 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <B05667366EE6204181EABE9C1B1C0EB58022A3@scsmsx401.amr.corp.intel.com>
2004-06-16 22:13 ` Hugetlb page bug fix for i386 in PAE mode Chen, Kenneth W
2004-06-16 22:07 Chen, Kenneth W

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®