[PATCH] ARC: Incorrect mm reference used in vmalloc fault handler

[PATCH] ARC: Incorrect mm reference used in vmalloc fault handler

[Vineet Gupta]

A vmalloc fault needs to sync up PGD/PTE entry from init_mm to current task's
"active_mm". ARC vmalloc fault handler however was using mm.

A vmalloc fault for non user task context (actually pre-userland, from
init thread's open for /dev/console) caused the handler to deref NULL mm
(for mm->pgd)

The reasons it worked so far is amazing:

1. By default (!SMP), vmalloc fault handler uses a cached value of PGD.
   In SMP that MMU register is repurposed hence need for mm pointer deref.

2. In pre-3.12 SMP kernel, the problem triggering vmalloc didn't exist in
   pre-userland code path - it was introduced with commit 20bafb3d23d108bc
   "n_tty: Move buffers into n_tty_data"

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Cc: Gilad Ben-Yossef <gilad@benyossef.com>
Cc: Noam Camus <noamc@ezchip.com>
Cc: stable@vger.kernel.org    #3.10 and 3.11
Cc: Peter Hurley <peter@hurleysoftware.com>
Cc: linux-kernel@vger.kernel.org
---
 arch/arc/mm/fault.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/arch/arc/mm/fault.c b/arch/arc/mm/fault.c
index 81554ded1260..9c69552350c4 100644
--- a/arch/arc/mm/fault.c
+++ b/arch/arc/mm/fault.c
@@ -17,7 +17,7 @@
 #include <asm/pgalloc.h>
 #include <asm/mmu.h>
 
-static int handle_vmalloc_fault(struct mm_struct *mm, unsigned long address)
+static int handle_vmalloc_fault(unsigned long address)
 {
 	/*
 	 * Synchronize this task's top level page-table
@@ -27,7 +27,7 @@ static int handle_vmalloc_fault(struct mm_struct *mm, unsigned long address)
 	pud_t *pud, *pud_k;
 	pmd_t *pmd, *pmd_k;
 
-	pgd = pgd_offset_fast(mm, address);
+	pgd = pgd_offset_fast(current->active_mm, address);
 	pgd_k = pgd_offset_k(address);
 
 	if (!pgd_present(*pgd_k))
@@ -72,7 +72,7 @@ void do_page_fault(unsigned long address, struct pt_regs *regs)
 	 * nothing more.
 	 */
 	if (address >= VMALLOC_START && address <= VMALLOC_END) {
-		ret = handle_vmalloc_fault(mm, address);
+		ret = handle_vmalloc_fault(address);
 		if (unlikely(ret))
 			goto bad_area_nosemaphore;
 		else
-- 
1.8.1.2

Re: [PATCH] Critical ARC Fix for 3.12

[Geert Uytterhoeven]

Added CC stable, so Greg will see it.
Added commit 9c41f4eeb9d51f3ece20428d35a3ea32cf3b5622, so Greg knows
what to cherry-pick.

On Sat, Nov 2, 2013 at 1:17 PM, Vineet Gupta <Vineet.Gupta1@synopsys.com> wrote:
> Hi Linus,
>
> Sorry for this extemely late patch for 3.12 but it causes a SMP build of ARC
> to panic in boot.
>
> Greg, the patch as it is doesn't apply to 3.10 (but does to 3.11). So can you
> please queue it up for 3.11 and I'll send a fixup for 3.10.
>
> Thx,
> -Vineet
>
>
> Vineet Gupta (1):
>   ARC: Incorrect mm reference used in vmalloc fault handler
>
>  arch/arc/mm/fault.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH] Critical ARC Fix for 3.12

[Vineet Gupta]

On 11/03/2013 03:59 PM, Geert Uytterhoeven wrote:
> Added CC stable, so Greg will see it.
> Added commit 9c41f4eeb9d51f3ece20428d35a3ea32cf3b5622, so Greg knows
> what to cherry-pick.

Thanks Geert. While cover-letter was not, the patch itself was CC'ed to stable -
this cover letter, with PATCH prefix has caused enough confusion already :-(

Greg, I have verified that patch builds fine with 3.10 and 3.11. Please take it.

-Vineet

Re: [PATCH] Critical ARC Fix for 3.12

[Greg KH]

On Sun, Nov 03, 2013 at 11:29:54AM +0100, Geert Uytterhoeven wrote:
> Added CC stable, so Greg will see it.
> Added commit 9c41f4eeb9d51f3ece20428d35a3ea32cf3b5622, so Greg knows
> what to cherry-pick.

It was already tagged with a cc: stable marking, so all is good.

thanks,

greg k-h