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

Groups > linux.kernel > #1502246 > unrolled thread

[PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness()

Back to article view | Back to linux.kernel

Contents

  [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Andy Lutomirski <luto@kernel.org> - 2016-10-17 19:10 +0200
    Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Stephan Mueller <smueller@chronox.de> - 2016-10-17 19:20 +0200
      Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Andy Lutomirski <luto@amacapital.net> - 2016-10-17 19:40 +0200
        Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Stephan Mueller <smueller@chronox.de> - 2016-10-17 20:40 +0200
          Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Andy Lutomirski <luto@amacapital.net> - 2016-10-17 23:10 +0200
            Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness() Stephan Mueller <smueller@chronox.de> - 2016-10-17 23:20 +0200
    Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in  add_early_randomness() Herbert Xu <herbert@gondor.apana.org.au> - 2016-10-19 06:00 +0200

#1502246 — [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness()

hw_random carefully avoids using a stack buffer except in
add_early_randomness().  This causes a crash in virtio_rng if
CONFIG_VMAP_STACK=y.

Reported-by: Matt Mullins <mmullins@mmlx.us>
Tested-by: Matt Mullins <mmullins@mmlx.us>
Fixes: d3cc7996473a ("hwrng: fetch randomness only after device init")
Signed-off-by: Andy Lutomirski <luto@kernel.org>
---

This fixes a crash in 4.9-rc1.

resending because I typoed the git send-email command.  I stealthily added
Matt's Tested-by, too.

 drivers/char/hw_random/core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/char/hw_random/core.c b/drivers/char/hw_random/core.c
index 9203f2d130c0..340f96e44642 100644
--- a/drivers/char/hw_random/core.c
+++ b/drivers/char/hw_random/core.c
@@ -84,14 +84,14 @@ static size_t rng_buffer_size(void)
 
 static void add_early_randomness(struct hwrng *rng)
 {
-	unsigned char bytes[16];
 	int bytes_read;
+	size_t size = min_t(size_t, 16, rng_buffer_size());
 
 	mutex_lock(&reading_mutex);
-	bytes_read = rng_get_data(rng, bytes, sizeof(bytes), 1);
+	bytes_read = rng_get_data(rng, rng_buffer, size, 1);
 	mutex_unlock(&reading_mutex);
 	if (bytes_read > 0)
-		add_device_randomness(bytes, bytes_read);
+		add_device_randomness(rng_buffer, bytes_read);
 }
 
 static inline void cleanup_rng(struct kref *kref)
-- 
2.7.4

[toc] | [next] | [standalone]

#1502282

Am Montag, 17. Oktober 2016, 10:06:27 CEST schrieb Andy Lutomirski:

Hi Andy,

> diff --git a/drivers/char/hw_random/core.c b/drivers/char/hw_random/core.c
> index 9203f2d130c0..340f96e44642 100644
> --- a/drivers/char/hw_random/core.c
> +++ b/drivers/char/hw_random/core.c
> @@ -84,14 +84,14 @@ static size_t rng_buffer_size(void)
> 
>  static void add_early_randomness(struct hwrng *rng)
>  {
> -	unsigned char bytes[16];
>  	int bytes_read;
> +	size_t size = min_t(size_t, 16, rng_buffer_size());
> 
>  	mutex_lock(&reading_mutex);
> -	bytes_read = rng_get_data(rng, bytes, sizeof(bytes), 1);
> +	bytes_read = rng_get_data(rng, rng_buffer, size, 1);
>  	mutex_unlock(&reading_mutex);
>  	if (bytes_read > 0)
> -		add_device_randomness(bytes, bytes_read);
> +		add_device_randomness(rng_buffer, bytes_read);

Shouldn't there be a memset(0) of the rng_buffer at this point to avoid having 
such data lingering in memory?


Ciao
Stephan

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

#1502303

On Mon, Oct 17, 2016 at 10:17 AM, Stephan Mueller <smueller@chronox.de> wrote:
> Am Montag, 17. Oktober 2016, 10:06:27 CEST schrieb Andy Lutomirski:
>
> Hi Andy,
>
>> diff --git a/drivers/char/hw_random/core.c b/drivers/char/hw_random/core.c
>> index 9203f2d130c0..340f96e44642 100644
>> --- a/drivers/char/hw_random/core.c
>> +++ b/drivers/char/hw_random/core.c
>> @@ -84,14 +84,14 @@ static size_t rng_buffer_size(void)
>>
>>  static void add_early_randomness(struct hwrng *rng)
>>  {
>> -     unsigned char bytes[16];
>>       int bytes_read;
>> +     size_t size = min_t(size_t, 16, rng_buffer_size());
>>
>>       mutex_lock(&reading_mutex);
>> -     bytes_read = rng_get_data(rng, bytes, sizeof(bytes), 1);
>> +     bytes_read = rng_get_data(rng, rng_buffer, size, 1);
>>       mutex_unlock(&reading_mutex);
>>       if (bytes_read > 0)
>> -             add_device_randomness(bytes, bytes_read);
>> +             add_device_randomness(rng_buffer, bytes_read);
>
> Shouldn't there be a memset(0) of the rng_buffer at this point to avoid having
> such data lingering in memory?

Sure, but shouldn't that be a separate patch covering the whole hw_crypto core?

--Andy


-- 
Andy Lutomirski
AMA Capital Management, LLC

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

#1502340

Am Montag, 17. Oktober 2016, 10:30:13 CEST schrieb Andy Lutomirski:

Hi Andy,
> 
> Sure, but shouldn't that be a separate patch covering the whole hw_crypto
> core?

I think that you are right -- there are many more cases where a memset(0) is 
warranted.

Do you want to make this change or should I send a patch?

Ciao
Stephan

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

#1502455

On Mon, Oct 17, 2016 at 11:36 AM, Stephan Mueller <smueller@chronox.de> wrote:
> Am Montag, 17. Oktober 2016, 10:30:13 CEST schrieb Andy Lutomirski:
>
> Hi Andy,
>>
>> Sure, but shouldn't that be a separate patch covering the whole hw_crypto
>> core?
>
> I think that you are right -- there are many more cases where a memset(0) is
> warranted.
>
> Do you want to make this change or should I send a patch?

Can you do it?  I have my work cut out for me making sure that all the
known regressions get stomped quickly...

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

#1502466

Am Montag, 17. Oktober 2016, 14:03:17 CEST schrieb Andy Lutomirski:

Hi Andy,

> On Mon, Oct 17, 2016 at 11:36 AM, Stephan Mueller <smueller@chronox.de> 
wrote:
> > Am Montag, 17. Oktober 2016, 10:30:13 CEST schrieb Andy Lutomirski:
> > 
> > Hi Andy,
> > 
> >> Sure, but shouldn't that be a separate patch covering the whole hw_crypto
> >> core?
> > 
> > I think that you are right -- there are many more cases where a memset(0)
> > is warranted.
> > 
> > Do you want to make this change or should I send a patch?
> 
> Can you do it?  I have my work cut out for me making sure that all the
> known regressions get stomped quickly...

Sure, will do.

Thanks.

Ciao
Stephan

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

#1503549 — Re: [PATCH resend 4.9] hw_random: Don't use a stack buffer in add_early_randomness()

On Mon, Oct 17, 2016 at 10:06:27AM -0700, Andy Lutomirski wrote:
> hw_random carefully avoids using a stack buffer except in
> add_early_randomness().  This causes a crash in virtio_rng if
> CONFIG_VMAP_STACK=y.
> 
> Reported-by: Matt Mullins <mmullins@mmlx.us>
> Tested-by: Matt Mullins <mmullins@mmlx.us>
> Fixes: d3cc7996473a ("hwrng: fetch randomness only after device init")
> Signed-off-by: Andy Lutomirski <luto@kernel.org>

Patch applied.  Thanks.
-- 
Email: Herbert Xu <herbert@gondor.apana.org.au>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web