Skip to contentSkip to conversation

ipvs: Add boundary check on ioctl arguments

PATCH Applied Simon Horman 3 participants 5 messages Dec 29, 2009 – Jan 4, 2010

Subsystems (4)
lore
Share & tools
lore

Simon Horman

author maintainer
#0 lore raw

PATCH 1/1 ipvs: Add boundary check on ioctl arguments

From: Arjan van de Ven <[email protected]>

ipvs: Add boundary check on ioctl arguments

The ipvs code has a nifty system for doing the size of ioctl command copies; it defines an array with values into which it indexes the cmd to find the right length.

Unfortunately, the ipvs code forgot to check if the cmd was in the range that the array provides, allowing for an index outside of the array, which then gives a "garbage" result into the length, which then gets used for copying into a stack buffer.

Fix this by adding sanity checks on these as well as the copy size.

---

 net/netfilter/ipvs/ip_vs_ctl.c |   14 +++++++++++++-
 1 files changed, 13 insertions(+), 1 deletions(-)

Hi Arjen,

this is the 4th response to your patch. I am guessing the previous ones didn't reach you for some reason. And I guess this one wont for the same reason.

I agree with Julian's assessment that your patch shouldn't be necessary, but on the other hand I think that the checks are reasonable. Your original patch made checks of the form of "cmd > IP_VS_SO_GET_MAX + 1". I have updated this to "cmd > IP_VS_SO_GET_MAX", as suggested by Julian, as the optmax elements of struct nf_sockopt_ops set a non-inclusive range.

http://lkml.indiana.edu/hypermail/linux/kernel/0910.0/00852.html

net/netfilter/ipvs/ip_vs_ctl.c +13 -1
Index: net-next-2.6/net/netfilter/ipvs/ip_vs_ctl.c
===================================================================
@@ -2077,6 +2077,10 @@ do_ip_vs_set_ctl(struct sock *sk, int cm
 	if (!capable(CAP_NET_ADMIN))
 		return -EPERM;
 
+	if (cmd < IP_VS_BASE_CTL || cmd > IP_VS_SO_SET_MAX)
+		return -EINVAL;
+	if (len < 0 || len >  MAX_ARG_LEN)
+		return -EINVAL;
 	if (len != set_arglen[SET_CMDID(cmd)]) {
 		pr_err("set_ctl: len %u != %u\n",
 		       len, set_arglen[SET_CMDID(cmd)]);
@@ -2352,17 +2356,25 @@ do_ip_vs_get_ctl(struct sock *sk, int cm
 {
 	unsigned char arg[128];
 	int ret = 0;
+	unsigned int copylen;
 
 	if (!capable(CAP_NET_ADMIN))
 		return -EPERM;
 
+	if (cmd < IP_VS_BASE_CTL || cmd > IP_VS_SO_GET_MAX)
+		return -EINVAL;
+
 	if (*len < get_arglen[GET_CMDID(cmd)]) {
 		pr_err("get_ctl: len %u < %u\n",
 		       *len, get_arglen[GET_CMDID(cmd)]);
 		return -EINVAL;
 	}
 
-	if (copy_from_user(arg, user, get_arglen[GET_CMDID(cmd)]) != 0)
+	copylen = get_arglen[GET_CMDID(cmd)];
+	if (copylen > 128)
+		return -EINVAL;
+
+	if (copy_from_user(arg, user, copylen) != 0)
 		return -EFAULT;
 
 	if (mutex_lock_interruptible(&__ip_vs_mutex))
#1 lore raw
Quoted 38 lines · Simon Horman

From: Arjan van de Ven <[email protected]>

ipvs: Add boundary check on ioctl arguments

The ipvs code has a nifty system for doing the size of ioctl command copies; it defines an array with values into which it indexes the cmd to find the right length.

Unfortunately, the ipvs code forgot to check if the cmd was in the range that the array provides, allowing for an index outside of the array, which then gives a "garbage" result into the length, which then gets used for copying into a stack buffer.

Fix this by adding sanity checks on these as well as the copy size.

---

 net/netfilter/ipvs/ip_vs_ctl.c |   14 +++++++++++++-
 1 files changed, 13 insertions(+), 1 deletions(-)

Hi Arjen,

this is the 4th response to your patch. I am guessing the previous ones didn't reach you for some reason. And I guess this one wont for the same reason.

I agree with Julian's assessment that your patch shouldn't be necessary, but on the other hand I think that the checks are reasonable. Your original patch made checks of the form of "cmd > IP_VS_SO_GET_MAX + 1". I have updated this to "cmd > IP_VS_SO_GET_MAX", as suggested by Julian, as the optmax elements of struct nf_sockopt_ops set a non-inclusive range.

Quoted 3 lines

http://lkml.indiana.edu/hypermail/linux/kernel/0910.0/00852.html

Index: net-next-2.6/net/netfilter/ipvs/ip_vs_ctl.c

As a bugfix, this seems more appropriate for net-2.6.git. Please let me know which tree you want me to apply this to.

#2 lore raw

On 1/4/2010 5:59, Patrick McHardy wrote:

[sorry for the late response, just got back from a good holiday, which means no work email access ;-) ]

Quoted 9 lines
Quoted 7 lines · Simon Horman

I agree with Julian's assessment that your patch shouldn't be necessary, but on the other hand I think that the checks are reasonable. Your original patch made checks of the form of "cmd> IP_VS_SO_GET_MAX + 1". I have updated this to "cmd> IP_VS_SO_GET_MAX", as suggested by Julian, as the optmax elements of struct nf_sockopt_ops set a non-inclusive range.

Quoted 6 lines
Quoted 3 lines

http://lkml.indiana.edu/hypermail/linux/kernel/0910.0/00852.html

Index: net-next-2.6/net/netfilter/ipvs/ip_vs_ctl.c

As a bugfix, this seems more appropriate for net-2.6.git. Please let me know which tree you want me to apply this to.

this really ought to go into 2.6.33.....

#3 lore raw
Quoted 14 lines · Arjan van de Ven

On 1/4/2010 5:59, Patrick McHardy wrote:

[sorry for the late response, just got back from a good holiday, which means no work email access ;-) ]

Quoted 9 lines
Quoted 7 lines · Simon Horman

I agree with Julian's assessment that your patch shouldn't be necessary, but on the other hand I think that the checks are reasonable. Your original patch made checks of the form of "cmd> IP_VS_SO_GET_MAX + 1". I have updated this to "cmd> IP_VS_SO_GET_MAX", as suggested by Julian, as the optmax elements of struct nf_sockopt_ops set a non-inclusive range.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

http://lkml.indiana.edu/hypermail/linux/kernel/0910.0/00852.html

Index: net-next-2.6/net/netfilter/ipvs/ip_vs_ctl.c

As a bugfix, this seems more appropriate for net-2.6.git. Please let me know which tree you want me to apply this to.

this really ought to go into 2.6.33.....

Thanks, applied and will send it upstream soon.

author maintainer
#4 lore raw
Quoted 17 lines · Patrick McHardy
Quoted 16 lines · Arjan van de Ven

On 1/4/2010 5:59, Patrick McHardy wrote:

[sorry for the late response, just got back from a good holiday, which means no work email access ;-) ]

Quoted 11 lines
Quoted 9 lines · Simon Horman

I agree with Julian's assessment that your patch shouldn't be necessary, but on the other hand I think that the checks are reasonable. Your original patch made checks of the form of "cmd> IP_VS_SO_GET_MAX + 1". I have updated this to "cmd> IP_VS_SO_GET_MAX", as suggested by Julian, as the optmax elements of struct nf_sockopt_ops set a non-inclusive range.

http://lkml.indiana.edu/hypermail/linux/kernel/0910.0/00852.html

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines
Quoted 1 line

Index: net-next-2.6/net/netfilter/ipvs/ip_vs_ctl.c

As a bugfix, this seems more appropriate for net-2.6.git. Please let me know which tree you want me to apply this to.

this really ought to go into 2.6.33.....

Thanks, applied and will send it upstream soon.

Thanks