[net,1/2] netlink: add NLA_U8_BUGGY attribute type

Commit Message

David Ahern Dec. 5, 2017, 4:41 p.m. UTC

On 12/5/17 9:34 AM, Johannes Berg wrote:
> On Tue, 2017-12-05 at 11:31 -0500, David Miller wrote:
>>
>>> We could try to fix up the big endian problem here, but we
>>> don't know *how* userspace misbehaved - if using nla_put_u32
>>> then we could, but we also found a debug tool (which we'll
>>> ignore for the purposes of this regression) that was putting
>>> the padding into the length.
> 
>> We're stuck with this thing forever... I'd like to consider other
>> options.
>>
>> I've seen this problem at least one time before, therefore I
>> suggest when we see a U8 attribute with a U32's length:
>>
>> 1) We access it as a u32, this takes care of all endianness
>>    issues.
> 
> Possible, but as I said above, I've seen at least one tool (a debug
> only script) now that will actually emit a U8 followed by 3 bytes of
> padding to make it netlink-aligned, but set the length to 4. That would
> be broken by making this change.
> 
> I'm not saying this is bad - but there are different levels of
> compatibility and I'd probably go for "bug compatibility" here rather
> than "fix-it-up compatibility".
> 
> Your call, ultimately - I've already fixed the tool I had found :-)
> 
>> 2) We emit a warning so that the app gets fixes.
> 

The attached is my proposal. Basically, allow it the length mismatch but
print a warning. This restores previous behavior and tells users of bad
commands.
From 29a504c8de553c17d4aafd5a2cb9384a28672fe4 Mon Sep 17 00:00:00 2001
From: David Ahern <[e-mail]>
Date: Sun, 3 Dec 2017 07:22:20 -0800
Subject: [PATCH] netlink: Relax attr validation for fixed length types

Commit 28033ae4e0f5 ("net: netlink: Update attr validation to require
exact length for some types") requires attributes using types NLA_U* and
NLA_S* to have an exact length. This change is exposing bugs in various
userspace commands that are sending attributes with an invalid length
(e.g., attribute has type NLA_U8 and userspace sends NLA_U32). While
the commands are clearly broken and need to be fixed, users are arguing
that the sudden change in enforcement is breaking older commands on
newer kernels for use cases that otherwise "worked".

Relax the validation to print a warning mesage similar to what is done
for messages containing extra bytes after parsing.

Fixes: 28033ae4e0f5 ("net: netlink: Update attr validation to require exact length for some types")
Signed-off-by: David Ahern <[e-mail]>
---
 lib/nlattr.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

Comments

David Miller Dec. 5, 2017, 4:51 p.m. UTC | #1

From: David Ahern <[e-mail]>
Date: Tue, 5 Dec 2017 09:41:21 -0700

> On 12/5/17 9:34 AM, Johannes Berg wrote:
>> On Tue, 2017-12-05 at 11:31 -0500, David Miller wrote:
>>>
>>>> We could try to fix up the big endian problem here, but we
>>>> don't know *how* userspace misbehaved - if using nla_put_u32
>>>> then we could, but we also found a debug tool (which we'll
>>>> ignore for the purposes of this regression) that was putting
>>>> the padding into the length.
>> 
>>> We're stuck with this thing forever... I'd like to consider other
>>> options.
>>>
>>> I've seen this problem at least one time before, therefore I
>>> suggest when we see a U8 attribute with a U32's length:
>>>
>>> 1) We access it as a u32, this takes care of all endianness
>>>    issues.
>> 
>> Possible, but as I said above, I've seen at least one tool (a debug
>> only script) now that will actually emit a U8 followed by 3 bytes of
>> padding to make it netlink-aligned, but set the length to 4. That would
>> be broken by making this change.
>> 
>> I'm not saying this is bad - but there are different levels of
>> compatibility and I'd probably go for "bug compatibility" here rather
>> than "fix-it-up compatibility".
>> 
>> Your call, ultimately - I've already fixed the tool I had found :-)
>> 
>>> 2) We emit a warning so that the app gets fixes.
>> 
> 
> The attached is my proposal. Basically, allow it the length mismatch but
> print a warning. This restores previous behavior and tells users of bad
> commands.

Where is the "access the U8 attribute as a U32 if length is 4" part
of my #1 above?

That's essential to handle this properly on all endianness.

Patch

diff --git a/lib/nlattr.c b/lib/nlattr.c
index 8bf78b4b78f0..6122662906c8 100644
--- a/lib/nlattr.c
+++ b/lib/nlattr.c
@@ -28,8 +28,16 @@  static const u8 nla_attr_len[NLA_TYPE_MAX+1] = {
 };
 
 static const u8 nla_attr_minlen[NLA_TYPE_MAX+1] = {
+	[NLA_U8]	= sizeof(u8),
+	[NLA_U16]	= sizeof(u16),
+	[NLA_U32]	= sizeof(u32),
+	[NLA_U64]	= sizeof(u64),
 	[NLA_MSECS]	= sizeof(u64),
 	[NLA_NESTED]	= NLA_HDRLEN,
+	[NLA_S8]	= sizeof(s8),
+	[NLA_S16]	= sizeof(s16),
+	[NLA_S32]	= sizeof(s32),
+	[NLA_S64]	= sizeof(s64),
 };
 
 static int validate_nla_bitfield32(const struct nlattr *nla,
@@ -70,10 +78,9 @@  static int validate_nla(const struct nlattr *nla, int maxtype,
 	BUG_ON(pt->type > NLA_TYPE_MAX);
 
 	/* for data types NLA_U* and NLA_S* require exact length */
-	if (nla_attr_len[pt->type]) {
-		if (attrlen != nla_attr_len[pt->type])
-			return -ERANGE;
-		return 0;
+	if (nla_attr_len[pt->type] && attrlen != nla_attr_len[pt->type]) {
+		pr_warn_ratelimited("netlink: '%s': attribute type %d has an invalid length.\n",
+				    current->comm, type);
 	}
 
 	switch (pt->type) {