Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1633401 > unrolled thread

[RFC] fix race in drivers/char/random.c:get_reg()

Back to article view | Back to linux.kernel

Contents

  [RFC] fix race in drivers/char/random.c:get_reg() Michael Schmitz <schmitzmic@gmail.com> - 2017-04-30 09:50 +0200
    Re: [RFC] fix race in drivers/char/random.c:get_reg() Geert Uytterhoeven <geert@linux-m68k.org> - 2017-05-12 09:20 +0200
    Re: [RFC] fix race in drivers/char/random.c:get_reg() Theodore Ts'o <tytso@mit.edu> - 2017-05-24 23:50 +0200

#1633401 — [RFC] fix race in drivers/char/random.c:get_reg()

get_reg() can be reentered on architectures with prioritized interrupts
(m68k in this case), causing f->reg_index to be incremented after the
range check. Out of bounds memory access past the pt_regs struct results.
This will go mostly undetected unless access is beyond end of memory.

Prevent the race by disabling interrupts in get_reg().

Tested on m68k (Atari Falcon, and ARAnyM emulator).

Kudos to Geert Uytterhoeven for helping to trace this race.

Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
---
 drivers/char/random.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/drivers/char/random.c b/drivers/char/random.c
index 1ef2640..7d1799b 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -1115,12 +1115,16 @@ static void add_interrupt_bench(cycles_t start)
 static __u32 get_reg(struct fast_pool *f, struct pt_regs *regs)
 {
 	__u32 *ptr = (__u32 *) regs;
+	unsigned long flags;
 
 	if (regs == NULL)
 		return 0;
+	local_irq_save(flags);
 	if (f->reg_idx >= sizeof(struct pt_regs) / sizeof(__u32))
 		f->reg_idx = 0;
-	return *(ptr + f->reg_idx++);
+	ptr += f->reg_idx++;
+	local_irq_restore(flags);
+	return *ptr;
 }
 
 void add_interrupt_randomness(int irq, int irq_flags)
-- 
1.7.0.4

[toc] | [next] | [standalone]

#1640249

On Sun, Apr 30, 2017 at 9:49 AM, Michael Schmitz <schmitzmic@gmail.com> wrote:
> get_reg() can be reentered on architectures with prioritized interrupts
> (m68k in this case), causing f->reg_index to be incremented after the
> range check. Out of bounds memory access past the pt_regs struct results.
> This will go mostly undetected unless access is beyond end of memory.
>
> Prevent the race by disabling interrupts in get_reg().
>
> Tested on m68k (Atari Falcon, and ARAnyM emulator).
>
> Kudos to Geert Uytterhoeven for helping to trace this race.
>
> Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>

No comments from anyone?

Perhaps this wasn't clear, but (a) an access beyond end of memory crashes
the system, and (b) this is reproducible on Atari systems.

Thanks!

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

[toc] | [prev] | [next] | [standalone]

#1649975

On Sun, Apr 30, 2017 at 07:49:21PM +1200, Michael Schmitz wrote:
> get_reg() can be reentered on architectures with prioritized interrupts
> (m68k in this case), causing f->reg_index to be incremented after the
> range check. Out of bounds memory access past the pt_regs struct results.
> This will go mostly undetected unless access is beyond end of memory.
> 
> Prevent the race by disabling interrupts in get_reg().
> 
> Tested on m68k (Atari Falcon, and ARAnyM emulator).
> 
> Kudos to Geert Uytterhoeven for helping to trace this race.
> 
> Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>

Thanks, applied.  It will go to Linus shortly.

		     	     	- Ted

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web