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

Groups > linux.kernel > #1322382 > unrolled thread

[patch] staging: rtl8712: memory corruption in wpa_set_encryption()

Back to article view | Back to linux.kernel

Contents

  [patch] staging: rtl8712: memory corruption in wpa_set_encryption() Dan Carpenter <dan.carpenter@oracle.com> - 2016-01-30 15:50 +0100
    Re: [patch] staging: rtl8712: memory corruption in wpa_set_encryption() Joshua Clayton <stillcompiling@gmail.com> - 2016-01-31 01:40 +0100

#1322382 — [patch] staging: rtl8712: memory corruption in wpa_set_encryption()

->KeyMaterial is declared as a 16 byte array, but we only ever allocate
either 5 or 13 bytes of it.  The problem is that we memset() all 16
bytes to zero so we're memsetting past the end of the allocated memory.

I fixed this in slightly lazy way, by just allocating 16 bytes.  This
works but there is a lot more cleanup you could do to this code if you
wanted.  Which is why this code is in staging.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>

diff --git a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
index edfc680..db2e31bc 100644
--- a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
+++ b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
@@ -398,12 +398,9 @@ static int wpa_set_encryption(struct net_device *dev, struct ieee_param *param,
 			wep_key_idx = 0;
 		if (wep_key_len > 0) {
 			wep_key_len = wep_key_len <= 5 ? 5 : 13;
-			pwep = kmalloc((u32)(wep_key_len +
-				FIELD_OFFSET(struct NDIS_802_11_WEP,
-				KeyMaterial)), GFP_ATOMIC);
+			pwep = kzalloc(sizeof(*pwep), GFP_ATOMIC);
 			if (pwep == NULL)
 				return -ENOMEM;
-			memset(pwep, 0, sizeof(struct NDIS_802_11_WEP));
 			pwep->KeyLength = wep_key_len;
 			pwep->Length = wep_key_len +
 				 FIELD_OFFSET(struct NDIS_802_11_WEP,

[toc] | [next] | [standalone]

#1322531

On Saturday, January 30, 2016 05:41:10 PM Dan Carpenter wrote:
> ->KeyMaterial is declared as a 16 byte array, but we only ever allocate
> either 5 or 13 bytes of it.  The problem is that we memset() all 16
> bytes to zero so we're memsetting past the end of the allocated memory.
> 
> I fixed this in slightly lazy way, by just allocating 16 bytes.  This
> works but there is a lot more cleanup you could do to this code if you
> wanted.  Which is why this code is in staging.
Better in every way than a crazy variable alloc if you ask me.
> 
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> 
> diff --git a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> index edfc680..db2e31bc 100644
> --- a/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> +++ b/drivers/staging/rtl8712/rtl871x_ioctl_linux.c
> @@ -398,12 +398,9 @@ static int wpa_set_encryption(struct net_device *dev, struct ieee_param *param,
>  			wep_key_idx = 0;
>  		if (wep_key_len > 0) {
>  			wep_key_len = wep_key_len <= 5 ? 5 : 13;
> -			pwep = kmalloc((u32)(wep_key_len +
> -				FIELD_OFFSET(struct NDIS_802_11_WEP,
> -				KeyMaterial)), GFP_ATOMIC);
> +			pwep = kzalloc(sizeof(*pwep), GFP_ATOMIC);
>  			if (pwep == NULL)
>  				return -ENOMEM;
> -			memset(pwep, 0, sizeof(struct NDIS_802_11_WEP));

Should there be a newline after the "if" statement?
>  			pwep->KeyLength = wep_key_len;
>  			pwep->Length = wep_key_len +
>  				 FIELD_OFFSET(struct NDIS_802_11_WEP,

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web