[PATCH net-next v2 0/4] tools: ynl: add --dbg-small-recv for easier kernel testing

[PATCH net-next v2 0/4] tools: ynl: add --dbg-small-recv for easier kernel testing

[Jakub Kicinski]

When testing netlink dumps I usually hack some user space up
to constrain its user space buffer size (iproute2, ethtool or ynl).
Netlink will try to fill the messages up, so since these apps use
large buffers by default, the dumps are rarely fragmented.

I was hoping to figure out a way to create a selftest for dump
testing, but so far I have no idea how to do that in a useful
and generic way.

Until someone does that, make manual dump testing easier with YNL.
Create a special option for limiting the buffer size, so I don't
have to make the same edits each time, and maybe others will benefit,
too :)

Example:

  $ ./cli.py [...] --dbg-small-recv >/dev/null
  Recv: read 3712 bytes, 29 messages
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
    [...]
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
  Recv: read 3968 bytes, 31 messages
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
    [...]
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
  Recv: read 532 bytes, 5 messages
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
    [...]
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
     nl_len = 20 (4) nl_flags = 0x2 nl_type = 3

Now let's make the DONE not fit in the last message:

  $ ./cli.py [...] --dbg-small-recv 4499 >/dev/null
  Recv: read 3712 bytes, 29 messages
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
    [...]
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
  Recv: read 4480 bytes, 35 messages
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
    [...]
     nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
  Recv: read 20 bytes, 1 messages
     nl_len = 20 (4) nl_flags = 0x2 nl_type = 3


A real test would also have to check the messages are complete
and not duplicated. That part has to be done manually right now.

Note that the first message is always conservatively sized by the kernel.
Still, I think this is good enough to be useful.

v2:
 - patch 2:
   - move the recv_size setting up
   - change the default to 0 so that cli.py doesn't have to worry
     what the "unset" value is
v1: https://lore.kernel.org/all/20240301230542.116823-1-kuba@kernel.org/

Jakub Kicinski (4):
  tools: ynl: move the new line in NlMsg __repr__
  tools: ynl: allow setting recv() size
  tools: ynl: support debug printing messages
  tools: ynl: add --dbg-small-recv for easier kernel testing

 tools/net/ynl/cli.py     |  7 ++++++-
 tools/net/ynl/lib/ynl.py | 42 ++++++++++++++++++++++++++++++++++------
 2 files changed, 42 insertions(+), 7 deletions(-)

-- 
2.44.0

[PATCH net-next v2 1/4] tools: ynl: move the new line in NlMsg __repr__

[Jakub Kicinski]

We add the new line even if message has no error or extack,
which leads to print(nl_msg) ending with two new lines.

Reviewed-by: Donald Hunter <donald.hunter@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 tools/net/ynl/lib/ynl.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/tools/net/ynl/lib/ynl.py b/tools/net/ynl/lib/ynl.py
index ac55aa5a3083..92ade9105f31 100644
--- a/tools/net/ynl/lib/ynl.py
+++ b/tools/net/ynl/lib/ynl.py
@@ -213,11 +213,11 @@ from .nlspec import SpecFamily
         return self.nl_type
 
     def __repr__(self):
-        msg = f"nl_len = {self.nl_len} ({len(self.raw)}) nl_flags = 0x{self.nl_flags:x} nl_type = {self.nl_type}\n"
+        msg = f"nl_len = {self.nl_len} ({len(self.raw)}) nl_flags = 0x{self.nl_flags:x} nl_type = {self.nl_type}"
         if self.error:
-            msg += '\terror: ' + str(self.error)
+            msg += '\n\terror: ' + str(self.error)
         if self.extack:
-            msg += '\textack: ' + repr(self.extack)
+            msg += '\n\textack: ' + repr(self.extack)
         return msg
 
 
-- 
2.44.0

[PATCH net-next v2 2/4] tools: ynl: allow setting recv() size

[Jakub Kicinski]

Make the size of the buffer we use for recv() configurable.
The details of the buffer sizing in netlink are somewhat
arcane, we could spend a lot of time polishing this API.
Let's just leave some hopefully helpful comments for now.
This is a for-developers-only feature, anyway.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
v2:
 - move the handling somewhere a bit more sensible
 - use 0 as default, and if to set it
---
 tools/net/ynl/lib/ynl.py | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)

diff --git a/tools/net/ynl/lib/ynl.py b/tools/net/ynl/lib/ynl.py
index 92ade9105f31..c3ff5be33e4e 100644
--- a/tools/net/ynl/lib/ynl.py
+++ b/tools/net/ynl/lib/ynl.py
@@ -84,6 +84,10 @@ from .nlspec import SpecFamily
     return f"Netlink error: {os.strerror(-self.nl_msg.error)}\n{self.nl_msg}"
 
 
+class ConfigError(Exception):
+    pass
+
+
 class NlAttr:
     ScalarFormat = namedtuple('ScalarFormat', ['native', 'big', 'little'])
     type_formats = {
@@ -400,7 +404,8 @@ genl_family_name_to_id = None
 
 
 class YnlFamily(SpecFamily):
-    def __init__(self, def_path, schema=None, process_unknown=False):
+    def __init__(self, def_path, schema=None, process_unknown=False,
+                 recv_size=0):
         super().__init__(def_path, schema)
 
         self.include_raw = False
@@ -415,6 +420,16 @@ genl_family_name_to_id = None
         except KeyError:
             raise Exception(f"Family '{self.yaml['name']}' not supported by the kernel")
 
+        # Note that netlink will use conservative (min) message size for
+        # the first dump recv() on the socket, our setting will only matter
+        # from the second recv() on.
+        self._recv_size = recv_size if recv_size else 131072
+        # Netlink will always allocate at least PAGE_SIZE - sizeof(skb_shinfo)
+        # for a message, so smaller receive sizes will lead to truncation.
+        # Note that the min size for other families may be larger than 4k!
+        if self._recv_size < 4000:
+            raise ConfigError()
+
         self.sock = socket.socket(socket.AF_NETLINK, socket.SOCK_RAW, self.nlproto.proto_num)
         self.sock.setsockopt(Netlink.SOL_NETLINK, Netlink.NETLINK_CAP_ACK, 1)
         self.sock.setsockopt(Netlink.SOL_NETLINK, Netlink.NETLINK_EXT_ACK, 1)
@@ -799,7 +814,7 @@ genl_family_name_to_id = None
     def check_ntf(self):
         while True:
             try:
-                reply = self.sock.recv(128 * 1024, socket.MSG_DONTWAIT)
+                reply = self.sock.recv(self._recv_size, socket.MSG_DONTWAIT)
             except BlockingIOError:
                 return
 
@@ -854,7 +869,7 @@ genl_family_name_to_id = None
         done = False
         rsp = []
         while not done:
-            reply = self.sock.recv(128 * 1024)
+            reply = self.sock.recv(self._recv_size)
             nms = NlMsgs(reply, attr_space=op.attr_set)
             for nl_msg in nms:
                 if nl_msg.extack:
-- 
2.44.0

[PATCH net-next v2 3/4] tools: ynl: support debug printing messages

[Jakub Kicinski]

For manual debug, allow printing the netlink level messages
to stderr.

Reviewed-by: Donald Hunter <donald.hunter@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 tools/net/ynl/lib/ynl.py | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/tools/net/ynl/lib/ynl.py b/tools/net/ynl/lib/ynl.py
index c3ff5be33e4e..239e22b7a85f 100644
--- a/tools/net/ynl/lib/ynl.py
+++ b/tools/net/ynl/lib/ynl.py
@@ -7,6 +7,7 @@ import random
 import socket
 import struct
 from struct import Struct
+import sys
 import yaml
 import ipaddress
 import uuid
@@ -420,6 +421,7 @@ genl_family_name_to_id = None
         except KeyError:
             raise Exception(f"Family '{self.yaml['name']}' not supported by the kernel")
 
+        self._recv_dbg = False
         # Note that netlink will use conservative (min) message size for
         # the first dump recv() on the socket, our setting will only matter
         # from the second recv() on.
@@ -453,6 +455,17 @@ genl_family_name_to_id = None
         self.sock.setsockopt(Netlink.SOL_NETLINK, Netlink.NETLINK_ADD_MEMBERSHIP,
                              mcast_id)
 
+    def set_recv_dbg(self, enabled):
+        self._recv_dbg = enabled
+
+    def _recv_dbg_print(self, reply, nl_msgs):
+        if not self._recv_dbg:
+            return
+        print("Recv: read", len(reply), "bytes,",
+              len(nl_msgs.msgs), "messages", file=sys.stderr)
+        for nl_msg in nl_msgs:
+            print("  ", nl_msg, file=sys.stderr)
+
     def _encode_enum(self, attr_spec, value):
         enum = self.consts[attr_spec['enum']]
         if enum.type == 'flags' or attr_spec.get('enum-as-flags', False):
@@ -819,6 +832,7 @@ genl_family_name_to_id = None
                 return
 
             nms = NlMsgs(reply)
+            self._recv_dbg_print(reply, nms)
             for nl_msg in nms:
                 if nl_msg.error:
                     print("Netlink error in ntf!?", os.strerror(-nl_msg.error))
@@ -871,6 +885,7 @@ genl_family_name_to_id = None
         while not done:
             reply = self.sock.recv(self._recv_size)
             nms = NlMsgs(reply, attr_space=op.attr_set)
+            self._recv_dbg_print(reply, nms)
             for nl_msg in nms:
                 if nl_msg.extack:
                     self._decode_extack(msg, op, nl_msg.extack)
-- 
2.44.0

[PATCH net-next v2 4/4] tools: ynl: add --dbg-small-recv for easier kernel testing

[Jakub Kicinski]

Most "production" netlink clients use large buffers to
make dump efficient, which means that handling of dump
continuation in the kernel is not very well tested.

Add an option for debugging / testing handling of dumps.
It enables printing of extra netlink-level debug and
lowers the recv() buffer size in one go. When used
without any argument (--dbg-small-recv) it picks
a very small default (4000), explicit size can be set,
too (--dbg-small-recv 5000).

Example:

$ ./cli.py [...] --dbg-small-recv
Recv: read 3712 bytes, 29 messages
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
 [...]
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
Recv: read 3968 bytes, 31 messages
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
 [...]
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
Recv: read 532 bytes, 5 messages
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
 [...]
   nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
   nl_len = 20 (4) nl_flags = 0x2 nl_type = 3

(the [...] are edits to shorten the commit message).

Note that the first message of the dump is sized conservatively
by the kernel.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 tools/net/ynl/cli.py | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/tools/net/ynl/cli.py b/tools/net/ynl/cli.py
index 0f8239979670..e8a65fbc3698 100755
--- a/tools/net/ynl/cli.py
+++ b/tools/net/ynl/cli.py
@@ -38,6 +38,8 @@ from lib import YnlFamily, Netlink
                         const=Netlink.NLM_F_APPEND)
     parser.add_argument('--process-unknown', action=argparse.BooleanOptionalAction)
     parser.add_argument('--output-json', action='store_true')
+    parser.add_argument('--dbg-small-recv', default=0, const=4000,
+                        action='store', nargs='?', type=int)
     args = parser.parse_args()
 
     def output(msg):
@@ -53,7 +55,10 @@ from lib import YnlFamily, Netlink
     if args.json_text:
         attrs = json.loads(args.json_text)
 
-    ynl = YnlFamily(args.spec, args.schema, args.process_unknown)
+    ynl = YnlFamily(args.spec, args.schema, args.process_unknown,
+                    recv_size=args.dbg_small_recv)
+    if args.dbg_small_recv:
+        ynl.set_recv_dbg(True)
 
     if args.ntf:
         ynl.ntf_subscribe(args.ntf)
-- 
2.44.0

Re: [PATCH net-next v2 2/4] tools: ynl: allow setting recv() size

[Donald Hunter]

Jakub Kicinski <kuba@kernel.org> writes:

> Make the size of the buffer we use for recv() configurable.
> The details of the buffer sizing in netlink are somewhat
> arcane, we could spend a lot of time polishing this API.
> Let's just leave some hopefully helpful comments for now.
> This is a for-developers-only feature, anyway.
>
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Reviewed-by: Donald Hunter <donald.hunter@gmail.com>

Re: [PATCH net-next v2 4/4] tools: ynl: add --dbg-small-recv for easier kernel testing

[Donald Hunter]

Jakub Kicinski <kuba@kernel.org> writes:

> Most "production" netlink clients use large buffers to
> make dump efficient, which means that handling of dump
> continuation in the kernel is not very well tested.
>
> Add an option for debugging / testing handling of dumps.
> It enables printing of extra netlink-level debug and
> lowers the recv() buffer size in one go. When used
> without any argument (--dbg-small-recv) it picks
> a very small default (4000), explicit size can be set,
> too (--dbg-small-recv 5000).
>
> Example:
>
> $ ./cli.py [...] --dbg-small-recv
> Recv: read 3712 bytes, 29 messages
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
>  [...]
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
> Recv: read 3968 bytes, 31 messages
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
>  [...]
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
> Recv: read 532 bytes, 5 messages
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
>  [...]
>    nl_len = 128 (112) nl_flags = 0x0 nl_type = 19
>    nl_len = 20 (4) nl_flags = 0x2 nl_type = 3
>
> (the [...] are edits to shorten the commit message).
>
> Note that the first message of the dump is sized conservatively
> by the kernel.
>
> Signed-off-by: Jakub Kicinski <kuba@kernel.org>

Reviewed-by: Donald Hunter <donald.hunter@gmail.com>

Re: [PATCH net-next v2 0/4] tools: ynl: add --dbg-small-recv for easier kernel testing

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by David S. Miller <davem@davemloft.net>:

On Mon,  4 Mar 2024 21:33:06 -0800 you wrote:
> When testing netlink dumps I usually hack some user space up
> to constrain its user space buffer size (iproute2, ethtool or ynl).
> Netlink will try to fill the messages up, so since these apps use
> large buffers by default, the dumps are rarely fragmented.
> 
> I was hoping to figure out a way to create a selftest for dump
> testing, but so far I have no idea how to do that in a useful
> and generic way.
> 
> [...]

Here is the summary with links:
  - [net-next,v2,1/4] tools: ynl: move the new line in NlMsg __repr__
    https://git.kernel.org/netdev/net-next/c/7df7231d6a6b
  - [net-next,v2,2/4] tools: ynl: allow setting recv() size
    https://git.kernel.org/netdev/net-next/c/7c93a88785da
  - [net-next,v2,3/4] tools: ynl: support debug printing messages
    https://git.kernel.org/netdev/net-next/c/a6a41521f95e
  - [net-next,v2,4/4] tools: ynl: add --dbg-small-recv for easier kernel testing
    https://git.kernel.org/netdev/net-next/c/c0111878d45e

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html