All of lore.kernel.org
 help / color / mirror / Atom feed
From: Pablo Neira Ayuso <pablo@netfilter.org>
To: netfilter-devel@vger.kernel.org
Cc: davem@davemloft.net, netdev@vger.kernel.org, kuba@kernel.org,
	pabeni@redhat.com, edumazet@google.com, fw@strlen.de,
	horms@kernel.org
Subject: [PATCH net-next 01/26] selftests: netfilter: nft_concat_range.sh: add coverage for 4bit group representation
Date: Thu, 22 May 2025 18:52:13 +0200	[thread overview]
Message-ID: <20250522165238.378456-2-pablo@netfilter.org> (raw)
In-Reply-To: <20250522165238.378456-1-pablo@netfilter.org>

From: Florian Westphal <fw@strlen.de>

Pipapo supports a more compact '4 bit group' format that is chosen when
the memory needed for the default exceeds a threshold (2mb).

Add coverage for those code paths, the existing tests use small sets that
are handled by the default representation.

This comes with a test script run-time increase, but I think its ok:

 normal: 2m35s -> 3m9s
 debug:  3m24s -> 5m29s (with KSFT_MACHINE_SLOW=yes).

Cc: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
 .../net/netfilter/nft_concat_range.sh         | 165 +++++++++++++++++-
 1 file changed, 161 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/net/netfilter/nft_concat_range.sh b/tools/testing/selftests/net/netfilter/nft_concat_range.sh
index 1f5979c1510c..efea93cf23d4 100755
--- a/tools/testing/selftests/net/netfilter/nft_concat_range.sh
+++ b/tools/testing/selftests/net/netfilter/nft_concat_range.sh
@@ -15,10 +15,12 @@ source lib.sh
 # Available test groups:
 # - reported_issues: check for issues that were reported in the past
 # - correctness: check that packets match given entries, and only those
+# - correctness_large: same but with additional non-matching entries
 # - concurrency: attempt races between insertion, deletion and lookup
 # - timeout: check that packets match entries until they expire
 # - performance: estimate matching rate, compare with rbtree and hash baselines
-TESTS="reported_issues correctness concurrency timeout"
+TESTS="reported_issues correctness correctness_large concurrency timeout"
+
 [ -n "$NFT_CONCAT_RANGE_TESTS" ] && TESTS="${NFT_CONCAT_RANGE_TESTS}"
 
 # Set types, defined by TYPE_ variables below
@@ -1257,9 +1259,7 @@ send_nomatch() {
 # - add ranged element, check that packets match it
 # - check that packets outside range don't match it
 # - remove some elements, check that packets don't match anymore
-test_correctness() {
-	setup veth send_"${proto}" set || return ${ksft_skip}
-
+test_correctness_main() {
 	range_size=1
 	for i in $(seq "${start}" $((start + count))); do
 		end=$((start + range_size))
@@ -1293,6 +1293,163 @@ test_correctness() {
 	done
 }
 
+test_correctness() {
+	setup veth send_"${proto}" set || return ${ksft_skip}
+
+	test_correctness_main
+}
+
+# Repeat the correctness tests, but add extra non-matching entries.
+# This exercises the more compact '4 bit group' representation that
+# gets picked when the default 8-bit representation exceed
+# NFT_PIPAPO_LT_SIZE_HIGH bytes of memory.
+# See usage of NFT_PIPAPO_LT_SIZE_HIGH in pipapo_lt_bits_adjust().
+#
+# The format() helper is way too slow when generating lots of
+# entries so its not used here.
+test_correctness_large() {
+	setup veth send_"${proto}" set || return ${ksft_skip}
+	# number of dummy (filler) entries to add.
+	local dcount=16385
+
+	(
+	echo -n "add element inet filter test { "
+
+	case "$type_spec" in
+	"ether_addr . ipv4_addr")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_mac $((1000000 + i))
+			printf ". 172.%i.%i.%i " $((RANDOM%256)) $((RANDOM%256)) $((i%256))
+		done
+		;;
+	"inet_proto . ipv6_addr")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "%i . " $((RANDOM%256))
+			format_addr6 $((1000000 + i))
+		done
+		;;
+	"inet_service . inet_proto")
+		# smaller key sizes, need more entries to hit the
+		# 4-bit threshold.
+		dcount=65536
+		for i in $(seq 1 $dcount); do
+			local proto=$((RANDOM%256))
+
+			# Test uses UDP to match, as it also fails when matching
+			# an entry that doesn't exist, so skip 'udp' entries
+			# to not trigger a wrong failure.
+			[ $proto -eq 17 ] && proto=18
+			[ $i -gt 1 ] && echo ", "
+			printf "%i . %i " $(((i%65534) + 1)) $((proto))
+		done
+		;;
+	"inet_service . ipv4_addr")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "%i . 172.%i.%i.%i " $(((RANDOM%65534) + 1)) $((RANDOM%256)) $((RANDOM%256)) $((i%256))
+		done
+		;;
+	"ipv4_addr . ether_addr")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . " $((RANDOM%256)) $((RANDOM%256)) $((i%256))
+			format_mac $((1000000 + i))
+		done
+		;;
+	"ipv4_addr . inet_service")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . %i" $((RANDOM%256)) $((RANDOM%256)) $((i%256)) $(((RANDOM%65534) + 1))
+		done
+		;;
+	"ipv4_addr . inet_service . ether_addr . inet_proto . ipv4_addr")
+		dcount=65536
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . %i . " $((RANDOM%256)) $((RANDOM%256)) $((i%256)) $(((RANDOM%65534) + 1))
+			format_mac $((1000000 + i))
+			printf ". %i . 192.168.%i.%i" $((RANDOM%256)) $((RANDOM%256)) $((i%256))
+		done
+		;;
+	"ipv4_addr . inet_service . inet_proto")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . %i . %i " $((RANDOM%256)) $((RANDOM%256)) $((i%256)) $(((RANDOM%65534) + 1)) $((RANDOM%256))
+		done
+		;;
+	"ipv4_addr . inet_service . inet_proto . ipv4_addr")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . %i . %i . 192.168.%i.%i " $((RANDOM%256)) $((RANDOM%256)) $((i%256)) $(((RANDOM%65534) + 1)) $((RANDOM%256)) $((RANDOM%256)) $((RANDOM%256))
+		done
+		;;
+	"ipv4_addr . inet_service . ipv4_addr")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			printf "172.%i.%i.%i . %i . 192.168.%i.%i " $((RANDOM%256)) $((RANDOM%256)) $((i%256)) $(((RANDOM%65534) + 1)) $((RANDOM%256)) $((RANDOM%256))
+		done
+		;;
+	"ipv6_addr . ether_addr")
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_addr6 $((i + 1000000))
+			echo -n " . "
+			format_mac $((1000000 + i))
+		done
+		;;
+	"ipv6_addr . inet_service")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_addr6 $((i + 1000000))
+			echo -n " .  $(((RANDOM%65534) + 1))"
+		done
+		;;
+	"ipv6_addr . inet_service . ether_addr")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_addr6 $((i + 1000000))
+			echo -n " .  $(((RANDOM%65534) + 1)) . "
+			format_mac $((i + 1000000))
+		done
+		;;
+	"ipv6_addr . inet_service . ether_addr . inet_proto")
+		dcount=65536
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_addr6 $((i + 1000000))
+			echo -n " .  $(((RANDOM%65534) + 1)) . "
+			format_mac $((i + 1000000))
+			echo -n " .  $((RANDOM%256))"
+		done
+		;;
+	"ipv6_addr . inet_service . ipv6_addr . inet_service")
+		dcount=32768
+		for i in $(seq 1 $dcount); do
+			[ $i -gt 1 ] && echo ", "
+			format_addr6 $((i + 1000000))
+			echo -n " .  $(((RANDOM%65534) + 1)) . "
+			format_addr6 $((i + 2123456))
+			echo -n " .  $((RANDOM%256))"
+		done
+		;;
+	*)
+		"Unhandled $type_spec"
+		return 1
+	esac
+	echo -n "}"
+
+	) | nft -f - || return 1
+
+	test_correctness_main
+}
+
 # Concurrency test template:
 # - add all the elements
 # - start a thread for each physical thread that:
-- 
2.30.2


  reply	other threads:[~2025-05-22 16:52 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-22 16:52 [PATCH net-next 00/26] Netfilter updates for net-next Pablo Neira Ayuso
2025-05-22 16:52 ` Pablo Neira Ayuso [this message]
2025-05-22 16:52 ` [PATCH net-next 02/26] netfilter: xtables: support arpt_mark and ipv6 optstrip for iptables-nft only builds Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 03/26] selftests: netfilter: nft_fib.sh: add 'type' mode tests Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 04/26] selftests: netfilter: move fib vrf test to nft_fib.sh Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 05/26] netfilter: nf_tables: nft_fib_ipv6: fix VRF ipv4/ipv6 result discrepancy Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 06/26] netfilter: nf_tables: nft_fib: consistent l3mdev handling Pablo Neira Ayuso
2025-05-23  7:35   ` Simon Horman
2025-05-23  7:39     ` Pablo Neira Ayuso
2025-05-23 13:26       ` Simon Horman
2025-05-23 13:29         ` Pablo Neira Ayuso
2025-05-23 13:38           ` Simon Horman
2025-05-22 16:52 ` [PATCH net-next 07/26] selftests: netfilter: nft_fib.sh: add type and oif tests with and without VRFs Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 08/26] netfilter: nft_tunnel: fix geneve_opt dump Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 09/26] netfilter: nf_dup{4, 6}: Move duplication check to task_struct Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 10/26] netfilter: nft_inner: Use nested-BH locking for nft_pcpu_tun_ctx Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 11/26] netfilter: nf_dup_netdev: Move the recursion counter struct netdev_xmit Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 12/26] netfilter: conntrack: make nf_conntrack_id callable without a module dependency Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 13/26] netfilter: nf_tables: add packets conntrack state to debug trace info Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 14/26] netfilter: nf_tables: Introduce functions freeing nft_hook objects Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 15/26] netfilter: nf_tables: Introduce nft_hook_find_ops{,_rcu}() Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 16/26] netfilter: nf_tables: Introduce nft_register_flowtable_ops() Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 17/26] netfilter: nf_tables: Pass nf_hook_ops to nft_unregister_flowtable_hook() Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 18/26] netfilter: nf_tables: Have a list of nf_hook_ops in nft_hook Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 19/26] netfilter: nf_tables: Prepare for handling NETDEV_REGISTER events Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 20/26] netfilter: nf_tables: Respect " Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 21/26] netfilter: nf_tables: Wrap netdev notifiers Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 22/26] netfilter: nf_tables: Handle NETDEV_CHANGENAME events Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 23/26] netfilter: nf_tables: Sort labels in nft_netdev_hook_alloc() Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 24/26] netfilter: nf_tables: Support wildcard netdev hook specs Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 25/26] netfilter: nf_tables: Add notifications for hook changes Pablo Neira Ayuso
2025-05-22 16:52 ` [PATCH net-next 26/26] selftests: netfilter: Torture nftables netdev hooks Pablo Neira Ayuso
  -- strict thread matches above, loose matches on Subject: below --
2025-05-23 13:26 [PATCH net-next,v2 00/26] Netfilter updates for net-next Pablo Neira Ayuso
2025-05-23 13:26 ` [PATCH net-next 01/26] selftests: netfilter: nft_concat_range.sh: add coverage for 4bit group representation Pablo Neira Ayuso
2025-05-26 17:11   ` patchwork-bot+netdevbpf

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=20250522165238.378456-2-pablo@netfilter.org \
    --to=pablo@netfilter.org \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=fw@strlen.de \
    --cc=horms@kernel.org \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=netfilter-devel@vger.kernel.org \
    --cc=pabeni@redhat.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.