* [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
@ 2012-02-29 6:35 Dan Carpenter
2012-02-29 8:21 ` walter harms
2012-03-01 10:19 ` [patch 1/4] " Jussi Kivilinna
0 siblings, 2 replies; 7+ messages in thread
From: Dan Carpenter @ 2012-02-29 6:35 UTC (permalink / raw)
To: Jussi Kivilinna; +Cc: John W. Linville, linux-wireless, kernel-janitors
If "offset" is negative then we can get past this check:
if (offset > CONTROL_BUFFER_SIZE)
Or if we pick a very high "req_ie_len" then we can get around the check:
if (offset + req_ie_len > CONTROL_BUFFER_SIZE)
I made "resp_ie_len" and "req_ie_len" unsigned. I don't know if it was
intentional that they were signed in the original.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
diff --git a/drivers/net/wireless/rndis_wlan.c b/drivers/net/wireless/rndis_wlan.c
index a330c69..6d8a986 100644
--- a/drivers/net/wireless/rndis_wlan.c
+++ b/drivers/net/wireless/rndis_wlan.c
@@ -2755,9 +2755,10 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
struct rndis_wlan_private *priv = get_rndis_wlan_priv(usbdev);
struct ndis_80211_assoc_info *info = NULL;
u8 bssid[ETH_ALEN];
- int resp_ie_len, req_ie_len;
+ unsigned int resp_ie_len, req_ie_len;
+ unsigned int offset;
u8 *req_ie, *resp_ie;
- int ret, offset;
+ int ret;
bool roamed = false;
bool match_bss;
@@ -2785,6 +2786,8 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
ret = get_association_info(usbdev, info, CONTROL_BUFFER_SIZE);
if (!ret) {
req_ie_len = le32_to_cpu(info->req_ie_length);
+ if (req_ie_len > CONTROL_BUFFER_SIZE)
+ req_ie_len = CONTROL_BUFFER_SIZE;
if (req_ie_len > 0) {
offset = le32_to_cpu(info->offset_req_ies);
@@ -2799,6 +2802,8 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
}
resp_ie_len = le32_to_cpu(info->resp_ie_length);
+ if (resp_ie_len > CONTROL_BUFFER_SIZE)
+ resp_ie_len = CONTROL_BUFFER_SIZE;
if (resp_ie_len > 0) {
offset = le32_to_cpu(info->offset_resp_ies);
^ permalink raw reply related [flat|nested] 7+ messages in thread* Re: [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-02-29 6:35 [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work() Dan Carpenter
@ 2012-02-29 8:21 ` walter harms
2012-03-01 6:58 ` Dan Carpenter
2012-03-01 7:02 ` [patch 1/4 v2] " Dan Carpenter
2012-03-01 10:19 ` [patch 1/4] " Jussi Kivilinna
1 sibling, 2 replies; 7+ messages in thread
From: walter harms @ 2012-02-29 8:21 UTC (permalink / raw)
To: Dan Carpenter
Cc: Jussi Kivilinna, John W. Linville, linux-wireless,
kernel-janitors
Am 29.02.2012 07:35, schrieb Dan Carpenter:
> If "offset" is negative then we can get past this check:
> if (offset > CONTROL_BUFFER_SIZE)
> Or if we pick a very high "req_ie_len" then we can get around the check:
> if (offset + req_ie_len > CONTROL_BUFFER_SIZE)
>
> I made "resp_ie_len" and "req_ie_len" unsigned. I don't know if it was
> intentional that they were signed in the original.
>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
>
> diff --git a/drivers/net/wireless/rndis_wlan.c b/drivers/net/wireless/rndis_wlan.c
> index a330c69..6d8a986 100644
> --- a/drivers/net/wireless/rndis_wlan.c
> +++ b/drivers/net/wireless/rndis_wlan.c
> @@ -2755,9 +2755,10 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
> struct rndis_wlan_private *priv = get_rndis_wlan_priv(usbdev);
> struct ndis_80211_assoc_info *info = NULL;
> u8 bssid[ETH_ALEN];
> - int resp_ie_len, req_ie_len;
> + unsigned int resp_ie_len, req_ie_len;
> + unsigned int offset;
> u8 *req_ie, *resp_ie;
> - int ret, offset;
> + int ret;
> bool roamed = false;
> bool match_bss;
>
> @@ -2785,6 +2786,8 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
> ret = get_association_info(usbdev, info, CONTROL_BUFFER_SIZE);
> if (!ret) {
> req_ie_len = le32_to_cpu(info->req_ie_length);
> + if (req_ie_len > CONTROL_BUFFER_SIZE)
> + req_ie_len = CONTROL_BUFFER_SIZE;
> if (req_ie_len > 0) {
> offset = le32_to_cpu(info->offset_req_ies);
>
> @@ -2799,6 +2802,8 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
> }
>
> resp_ie_len = le32_to_cpu(info->resp_ie_length);
> + if (resp_ie_len > CONTROL_BUFFER_SIZE)
> + resp_ie_len = CONTROL_BUFFER_SIZE;
> if (resp_ie_len > 0) {
> offset = le32_to_cpu(info->offset_resp_ies);
>
hi dan,
the check below "if (resp_ie_len > 0)" looks strange for an unsigned.
re,
wh
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-02-29 8:21 ` walter harms
@ 2012-03-01 6:58 ` Dan Carpenter
2012-03-01 7:02 ` [patch 1/4 v2] " Dan Carpenter
1 sibling, 0 replies; 7+ messages in thread
From: Dan Carpenter @ 2012-03-01 6:58 UTC (permalink / raw)
To: walter harms
Cc: Jussi Kivilinna, John W. Linville, linux-wireless,
kernel-janitors
[-- Attachment #1: Type: text/plain, Size: 462 bytes --]
On Wed, Feb 29, 2012 at 09:21:29AM +0100, walter harms wrote:
> > resp_ie_len = le32_to_cpu(info->resp_ie_length);
> > + if (resp_ie_len > CONTROL_BUFFER_SIZE)
> > + resp_ie_len = CONTROL_BUFFER_SIZE;
> > if (resp_ie_len > 0) {
> > offset = le32_to_cpu(info->offset_resp_ies);
> >
>
>
> hi dan,
> the check below "if (resp_ie_len > 0)" looks strange for an unsigned.
>
Good point. I'll resend.
regards,
dan carpenter
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread* [patch 1/4 v2] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-02-29 8:21 ` walter harms
2012-03-01 6:58 ` Dan Carpenter
@ 2012-03-01 7:02 ` Dan Carpenter
2012-03-01 9:51 ` bojan prtvar
1 sibling, 1 reply; 7+ messages in thread
From: Dan Carpenter @ 2012-03-01 7:02 UTC (permalink / raw)
To: walter harms
Cc: Jussi Kivilinna, John W. Linville, linux-wireless,
kernel-janitors
[-- Attachment #1: Type: text/plain, Size: 1929 bytes --]
If "offset" is negative then we can get past this check:
if (offset > CONTROL_BUFFER_SIZE)
Or if we pick a very high "req_ie_len" then we can get around the check:
if (offset + req_ie_len > CONTROL_BUFFER_SIZE)
I made "resp_ie_len" and "req_ie_len" unsigned. I don't know if it was
intentional that they were signed in the original.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
v2: Fixed a style issue for Walter Harms. Changed > 0 to != 0.
diff --git a/drivers/net/wireless/rndis_wlan.c b/drivers/net/wireless/rndis_wlan.c
index a330c69..dde45ef 100644
--- a/drivers/net/wireless/rndis_wlan.c
+++ b/drivers/net/wireless/rndis_wlan.c
@@ -2755,9 +2755,10 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
struct rndis_wlan_private *priv = get_rndis_wlan_priv(usbdev);
struct ndis_80211_assoc_info *info = NULL;
u8 bssid[ETH_ALEN];
- int resp_ie_len, req_ie_len;
+ unsigned int resp_ie_len, req_ie_len;
+ unsigned int offset;
u8 *req_ie, *resp_ie;
- int ret, offset;
+ int ret;
bool roamed = false;
bool match_bss;
@@ -2785,7 +2786,9 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
ret = get_association_info(usbdev, info, CONTROL_BUFFER_SIZE);
if (!ret) {
req_ie_len = le32_to_cpu(info->req_ie_length);
- if (req_ie_len > 0) {
+ if (req_ie_len > CONTROL_BUFFER_SIZE)
+ req_ie_len = CONTROL_BUFFER_SIZE;
+ if (req_ie_len != 0) {
offset = le32_to_cpu(info->offset_req_ies);
if (offset > CONTROL_BUFFER_SIZE)
@@ -2799,7 +2802,9 @@ static void rndis_wlan_do_link_up_work(struct usbnet *usbdev)
}
resp_ie_len = le32_to_cpu(info->resp_ie_length);
- if (resp_ie_len > 0) {
+ if (resp_ie_len > CONTROL_BUFFER_SIZE)
+ resp_ie_len = CONTROL_BUFFER_SIZE;
+ if (resp_ie_len != 0) {
offset = le32_to_cpu(info->offset_resp_ies);
if (offset > CONTROL_BUFFER_SIZE)
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply related [flat|nested] 7+ messages in thread* Re: [patch 1/4 v2] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-03-01 7:02 ` [patch 1/4 v2] " Dan Carpenter
@ 2012-03-01 9:51 ` bojan prtvar
2012-03-01 12:57 ` Dan Carpenter
0 siblings, 1 reply; 7+ messages in thread
From: bojan prtvar @ 2012-03-01 9:51 UTC (permalink / raw)
To: Dan Carpenter
Cc: walter harms, Jussi Kivilinna, John W. Linville, linux-wireless,
kernel-janitors
Hi,
On Thu, Mar 1, 2012 at 8:02 AM, Dan Carpenter <dan.carpenter@oracle.com> wrote:
>
>
>
>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> v2: Fixed a style issue for Walter Harms. Changed > 0 to != 0.
>
Why not just if (req_ie_len) and if (resp_ie_len) ?
Regards,
Bojan
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [patch 1/4 v2] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-03-01 9:51 ` bojan prtvar
@ 2012-03-01 12:57 ` Dan Carpenter
0 siblings, 0 replies; 7+ messages in thread
From: Dan Carpenter @ 2012-03-01 12:57 UTC (permalink / raw)
To: bojan prtvar
Cc: walter harms, Jussi Kivilinna, John W. Linville, linux-wireless,
kernel-janitors
[-- Attachment #1: Type: text/plain, Size: 825 bytes --]
On Thu, Mar 01, 2012 at 10:51:37AM +0100, bojan prtvar wrote:
> Hi,
>
>
> On Thu, Mar 1, 2012 at 8:02 AM, Dan Carpenter <dan.carpenter@oracle.com> wrote:
> >
> >
> >
> >
> > Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> > ---
> > v2: Fixed a style issue for Walter Harms. Changed > 0 to != 0.
> >
>
> Why not just if (req_ie_len) and if (resp_ie_len) ?
>
It could go either way. I wrote it that way first, then I decided
that zero was a special enough case to draw attention to it. In
this case it felt like zero was its own thing.
For allocation failures I would do:
foo = kmalloc();
if (!foo)
return -ENOMEM;
Allocation failures are not interesting and the NULL doesn't have a
special meaning and doesn't need explanation.
Hard to explain.
regards,
dan carpenter
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work()
2012-02-29 6:35 [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work() Dan Carpenter
2012-02-29 8:21 ` walter harms
@ 2012-03-01 10:19 ` Jussi Kivilinna
1 sibling, 0 replies; 7+ messages in thread
From: Jussi Kivilinna @ 2012-03-01 10:19 UTC (permalink / raw)
To: Dan Carpenter; +Cc: John W. Linville, linux-wireless, kernel-janitors
Quoting Dan Carpenter <dan.carpenter@oracle.com>:
> If "offset" is negative then we can get past this check:
> if (offset > CONTROL_BUFFER_SIZE)
> Or if we pick a very high "req_ie_len" then we can get around the check:
> if (offset + req_ie_len > CONTROL_BUFFER_SIZE)
>
> I made "resp_ie_len" and "req_ie_len" unsigned. I don't know if it was
> intentional that they were signed in the original.
>
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
>
> diff --git a/drivers/net/wireless/rndis_wlan.c
> b/drivers/net/wireless/rndis_wlan.c
> index a330c69..6d8a986 100644
> --- a/drivers/net/wireless/rndis_wlan.c
> +++ b/drivers/net/wireless/rndis_wlan.c
> @@ -2755,9 +2755,10 @@ static void rndis_wlan_do_link_up_work(struct
> usbnet *usbdev)
> struct rndis_wlan_private *priv = get_rndis_wlan_priv(usbdev);
> struct ndis_80211_assoc_info *info = NULL;
> u8 bssid[ETH_ALEN];
> - int resp_ie_len, req_ie_len;
> + unsigned int resp_ie_len, req_ie_len;
> + unsigned int offset;
> u8 *req_ie, *resp_ie;
> - int ret, offset;
> + int ret;
> bool roamed = false;
> bool match_bss;
>
> @@ -2785,6 +2786,8 @@ static void rndis_wlan_do_link_up_work(struct
> usbnet *usbdev)
> ret = get_association_info(usbdev, info, CONTROL_BUFFER_SIZE);
> if (!ret) {
> req_ie_len = le32_to_cpu(info->req_ie_length);
> + if (req_ie_len > CONTROL_BUFFER_SIZE)
> + req_ie_len = CONTROL_BUFFER_SIZE;
> if (req_ie_len > 0) {
> offset = le32_to_cpu(info->offset_req_ies);
>
> @@ -2799,6 +2802,8 @@ static void rndis_wlan_do_link_up_work(struct
> usbnet *usbdev)
> }
>
> resp_ie_len = le32_to_cpu(info->resp_ie_length);
> + if (resp_ie_len > CONTROL_BUFFER_SIZE)
> + resp_ie_len = CONTROL_BUFFER_SIZE;
> if (resp_ie_len > 0) {
(resp_ie_len > 0), (resp_ie_len != 0), (resp_ie_len) .. all fine by me,
Acked-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
> offset = le32_to_cpu(info->offset_resp_ies);
>
>
>
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2012-03-01 12:56 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-02-29 6:35 [patch 1/4] rndis_wlan: integer overflows in rndis_wlan_do_link_up_work() Dan Carpenter
2012-02-29 8:21 ` walter harms
2012-03-01 6:58 ` Dan Carpenter
2012-03-01 7:02 ` [patch 1/4 v2] " Dan Carpenter
2012-03-01 9:51 ` bojan prtvar
2012-03-01 12:57 ` Dan Carpenter
2012-03-01 10:19 ` [patch 1/4] " Jussi Kivilinna
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).