nfc: af_nfc: Fix Spectre v1 vulnerability
diff mbox series

Message ID 20181221214753.GA2799@embeddedor
State New
Delegated to: Samuel Ortiz
Headers show
Series
  • nfc: af_nfc: Fix Spectre v1 vulnerability
Related show

Commit Message

Gustavo A. R. Silva Dec. 21, 2018, 9:47 p.m. UTC
proto is indirectly controlled by user-space, hence leading to
a potential exploitation of the Spectre variant 1 vulnerability.

This issue was detected with the help of Smatch:

net/nfc/af_nfc.c:42 nfc_sock_create() warn: potential spectre issue 'proto_tab' [w] (local cap)

Fix this by sanitizing proto before using it to index proto_tab.

Notice that given that speculation windows are large, the policy is
to kill the speculation on the first load and not worry if it can be
completed with a dependent load/store [1].

[1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
---
 net/nfc/af_nfc.c | 2 ++
 1 file changed, 2 insertions(+)

Comments

David Miller Dec. 22, 2018, 11:09 p.m. UTC | #1
From: "Gustavo A. R. Silva" <gustavo@embeddedor.com>
Date: Fri, 21 Dec 2018 15:47:53 -0600

> proto is indirectly controlled by user-space, hence leading to
> a potential exploitation of the Spectre variant 1 vulnerability.
> 
> This issue was detected with the help of Smatch:
> 
> net/nfc/af_nfc.c:42 nfc_sock_create() warn: potential spectre issue 'proto_tab' [w] (local cap)
> 
> Fix this by sanitizing proto before using it to index proto_tab.
> 
> Notice that given that speculation windows are large, the policy is
> to kill the speculation on the first load and not worry if it can be
> completed with a dependent load/store [1].
> 
> [1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2
> 
> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>

I'll take this directly, and queued up for -stable.

Thanks.
Gustavo A. R. Silva Dec. 22, 2018, 11:37 p.m. UTC | #2
On 12/22/18 5:09 PM, David Miller wrote:
> From: "Gustavo A. R. Silva" <gustavo@embeddedor.com>
> Date: Fri, 21 Dec 2018 15:47:53 -0600
> 
>> proto is indirectly controlled by user-space, hence leading to
>> a potential exploitation of the Spectre variant 1 vulnerability.
>>
>> This issue was detected with the help of Smatch:
>>
>> net/nfc/af_nfc.c:42 nfc_sock_create() warn: potential spectre issue 'proto_tab' [w] (local cap)
>>
>> Fix this by sanitizing proto before using it to index proto_tab.
>>
>> Notice that given that speculation windows are large, the policy is
>> to kill the speculation on the first load and not worry if it can be
>> completed with a dependent load/store [1].
>>
>> [1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2
>>
>> Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
> 
> I'll take this directly, and queued up for -stable.
> 

Dave,

I wonder if you can take this one too:

https://lore.kernel.org/lkml/20181221212229.GA32635@embeddedor/

It's pretty similar to the af_nfc one.

Thanks
--
Gustavo
David Miller Dec. 23, 2018, 2:42 a.m. UTC | #3
From: "Gustavo A. R. Silva" <gustavo@embeddedor.com>
Date: Sat, 22 Dec 2018 17:37:35 -0600

> I wonder if you can take this one too:
> 
> https://lore.kernel.org/lkml/20181221212229.GA32635@embeddedor/
> 
> It's pretty similar to the af_nfc one.

Sure, done.
Gustavo A. R. Silva Dec. 23, 2018, 3:03 a.m. UTC | #4
On 12/22/18 8:42 PM, David Miller wrote:
> From: "Gustavo A. R. Silva" <gustavo@embeddedor.com>
> Date: Sat, 22 Dec 2018 17:37:35 -0600
> 
>> I wonder if you can take this one too:
>>
>> https://lore.kernel.org/lkml/20181221212229.GA32635@embeddedor/
>>
>> It's pretty similar to the af_nfc one.
> 
> Sure, done.
> 

Great. Thanks.
--
Gustavo

Patch
diff mbox series

diff --git a/net/nfc/af_nfc.c b/net/nfc/af_nfc.c
index d3e594eb36d0..256f3c57059e 100644
--- a/net/nfc/af_nfc.c
+++ b/net/nfc/af_nfc.c
@@ -21,6 +21,7 @@ 
 
 #include <linux/nfc.h>
 #include <linux/module.h>
+#include <linux/nospec.h>
 
 #include "nfc.h"
 
@@ -37,6 +38,7 @@  static int nfc_sock_create(struct net *net, struct socket *sock, int proto,
 
 	if (proto < 0 || proto >= NFC_SOCKPROTO_MAX)
 		return -EINVAL;
+	proto = array_index_nospec(proto, NFC_SOCKPROTO_MAX);
 
 	read_lock(&proto_tab_lock);
 	if (proto_tab[proto] &&	try_module_get(proto_tab[proto]->owner)) {