netfilter-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Thomas Haller <thaller@redhat.com>
To: NetFilter <netfilter-devel@vger.kernel.org>
Cc: Pablo Neira Ayuso <pablo@netfilter.org>,
	Thomas Haller <thaller@redhat.com>
Subject: [nft v2 PATCH 3/3] py: add input_{set,get}_flags() API to helpers
Date: Fri, 14 Jul 2023 10:48:53 +0200	[thread overview]
Message-ID: <20230714084943.1080757-3-thaller@redhat.com> (raw)
In-Reply-To: <20230714084943.1080757-1-thaller@redhat.com>

Note that the corresponding API for output flags does not expose the
plain numeric flags. Instead, it exposes the underlying, flag-based C
API more directly.

Reasons:

- a flags property has the benefits that adding new flags is very light
  weight. Otherwise, every addition of a flag requires new API. That new
  API increases the documentation and what the user needs to understand.
  With a flag API, we just need new documentation what the new flag is.
  It's already clear how to use it.

- opinionated, also the usage of "many getter/setter API" is not have
  better usability. Its convenient when we can do similar things (setting
  a boolean flag) depending on an argument of a function, instead of
  having different functions.

  Compare

     ctx.set_reversedns_output(True)
     ctx.set_handle_output(True)

  with

     ctx.ouput_set_flags(NFT_CTX_OUTPUT_REVERSEDNS | NFT_CTX_OUTPUT_HANDLE)

  Note that the vast majority of users of this API will just create one
  nft_ctx instance and set the flags once. Each user application
  probably has only one place where they call the setter once. So
  while I think flags have better usability, it doesn't matter much
  either way.

- if individual properties are preferable over flags, then the C API
  should also do that. In other words, the Python API should be similar
  to the underlying C API.

- I don't understand how to do this best. Is Nftables.output_flags
  public API? It appears to be, as it has no underscore. Why does this
  additional mapping from function (get_reversedns_output()) to name
  ("reversedns") to number (1<<0) exist?

Downside is the inconsistency with the existing output flags API.

Signed-off-by: Thomas Haller <thaller@redhat.com>
---

This is probably a controversial approach :)

 py/nftables.py | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/py/nftables.py b/py/nftables.py
index b9fa63bb8789..795700db45ef 100644
--- a/py/nftables.py
+++ b/py/nftables.py
@@ -21,6 +21,16 @@ import os
 
 NFTABLES_VERSION = "0.1"
 
+"""Prevent blocking DNS lookups for IP addresses.
+
+By default, nftables will try to resolve IP addresses with blocking getaddrinfo() API.
+By setting this flag, only literal IP adddresses are supported in input.
+
+This numeric flag can be passed to Nftables.input_get_flags() and Nftables.input_set_flags().
+"""
+NFT_CTX_INPUT_NO_DNS = 1
+
+
 class SchemaValidator:
     """Libnftables JSON validator using jsonschema"""
 
@@ -159,6 +169,27 @@ class Nftables:
     def __del__(self):
         self.nft_ctx_free(self.__ctx)
 
+    def input_get_flags(self):
+        """Query input flags for the nft context.
+
+        See input_get_flags() for supported flags.
+
+        Returns the currently set input flags as number.
+        """
+        return self.nft_ctx_input_get_flags(self.__ctx)
+
+    def input_set_flags(self, flags):
+        """Set input flags for the nft context as number.
+
+        By default, a new context objects has no flags set.
+
+        Supported flags are NFT_CTX_INPUT_NO_DNS (0x1) to disable blocking address
+        lookup via getaddrinfo.
+
+        Returns nothing.
+        """
+        self.nft_ctx_input_set_flags(self.__ctx, flags)
+
     def __get_output_flag(self, name):
         flag = self.output_flags[name]
         return self.nft_ctx_output_get_flags(self.__ctx) & flag
-- 
2.41.0


  parent reply	other threads:[~2023-07-14  8:50 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-10 17:46 [nft PATCH] nftables: add flag for nft context to avoid blocking getaddrinfo() Thomas Haller
2023-07-10 17:58 ` Pablo Neira Ayuso
2023-07-14  8:48   ` [nft v2 PATCH 1/3] nftables: add input flags for nft_ctx Thomas Haller
2023-07-14  8:48     ` [nft v2 PATCH 2/3] nftables: add input flag NFT_CTX_INPUT_NO_DNS to avoid blocking getaddrinfo() Thomas Haller
2023-07-14 10:07       ` Phil Sutter
2023-07-18  9:12         ` Thomas Haller
2023-07-14  8:48     ` Thomas Haller [this message]
2023-07-14  9:59       ` [nft v2 PATCH 3/3] py: add input_{set,get}_flags() API to helpers Phil Sutter
2023-07-18 10:07         ` Thomas Haller
2023-07-14 10:16     ` [nft v2 PATCH 1/3] nftables: add input flags for nft_ctx Phil Sutter
2023-07-18  9:05       ` Thomas Haller
2023-07-18  9:33         ` Phil Sutter
2023-07-18 10:31           ` Thomas Haller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230714084943.1080757-3-thaller@redhat.com \
    --to=thaller@redhat.com \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pablo@netfilter.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).