public inbox for pve-devel@lists.proxmox.com
 help / color / mirror / Atom feed
From: Hannes Laimer <h.laimer@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [PATCH pve-network v4 06/20] sdn: ipam: key the MAC cache by lower-case MAC
Date: Wed, 21 Oct 2026 10:44:31 +0200	[thread overview]
Message-ID: <20261021084445.462195-7-h.laimer@proxmox.com> (raw)
In-Reply-To: <20261021084445.462195-1-h.laimer@proxmox.com>

Generated guest MACs are upper case, but the API accepts either
spelling. The cache was keyed by the MAC as its writer spelled it. A
deletion in the other spelling missed the entry, so the cache kept
answering with the deleted address. The cache is now read with every
MAC in lower case and written under the lower-case MAC. An entry
cached before this change under another spelling is read under the
lower-case key and leaves the file with the next write. Two spellings
that disagree on an address are both dropped on the read, since one
of them is stale and the cache cannot tell which. The next lookup
then asks the IPAM. The dnsmasq reservation file is written and read
with its MACs in lower case as well, since its lines from before keep
their writer's spelling.

Signed-off-by: Hannes Laimer <h.laimer@proxmox.com>
---

Notes:
    v4:
     * fold the spelling handling into the cache read
     * split the miss and lock fixes into their own commits

 src/PVE/Network/SDN/Dhcp/Dnsmasq.pm |  3 ++
 src/PVE/Network/SDN/Ipams.pm        | 39 ++++++++++----
 src/test/run_test_vnets_blackbox.pl | 82 +++++++++++++++++++++++++++++
 3 files changed, 113 insertions(+), 11 deletions(-)

diff --git a/src/PVE/Network/SDN/Dhcp/Dnsmasq.pm b/src/PVE/Network/SDN/Dhcp/Dnsmasq.pm
index 477b700f..d3b78fe9 100644
--- a/src/PVE/Network/SDN/Dhcp/Dnsmasq.pm
+++ b/src/PVE/Network/SDN/Dhcp/Dnsmasq.pm
@@ -53,6 +53,7 @@ sub update_lease {
 
 sub add_ip_mapping {
     my ($class, $dhcpid, $macdb, $mac, $ip4, $ip6) = @_;
+    $mac = lc($mac);
 
     my $ethers_file = ethers_file($dhcpid);
     my $ethers_tmp_file = "$ethers_file.tmp";
@@ -76,6 +77,8 @@ sub add_ip_mapping {
                 warn "failed to parse MAC from $dhcpid ethers file on line $line_no: '$line'\n";
                 next;
             }
+            # the cache is keyed in lower case, a line from before keeps its writer's spelling
+            $parsed_mac = lc($parsed_mac);
 
             my ($parsed_ip4, $parsed_ip6);
             if ($parsed_ip2) {
diff --git a/src/PVE/Network/SDN/Ipams.pm b/src/PVE/Network/SDN/Ipams.pm
index 3d849d79..0e795a93 100644
--- a/src/PVE/Network/SDN/Ipams.pm
+++ b/src/PVE/Network/SDN/Ipams.pm
@@ -23,17 +23,31 @@ PVE::Network::SDN::Ipams::Plugin->init();
 my $macdb_filename = "sdn/mac-cache.json";
 cfs_register_file($macdb_filename, \&json_reader, \&json_writer);
 
-# an entry without an address was left behind by a lookup that missed, and
-# says nothing about its MAC
+# the file may hold a MAC under more than one spelling, and an entry without
+# an address says nothing about its MAC. Two spellings that disagree on an
+# address are both dropped, since one of them is stale and nothing here says
+# which. The next lookup then asks the IPAM
+my sub fold_macdb {
+    my ($db) = @_;
+
+    my $macs = {};
+    my %clash;
+    for my $key (sort keys %{ $db->{macs} // {} }) {
+        my $entry = $db->{macs}->{$key};
+        for my $family (grep { defined($entry->{$_}) } qw(ip4 ip6)) {
+            my $seen = $macs->{ lc($key) }->{$family};
+            $clash{ lc($key) } = 1 if defined($seen) && lc($seen) ne lc($entry->{$family});
+            $macs->{ lc($key) }->{$family} //= $entry->{$family};
+        }
+    }
+    delete $macs->{$_} for keys %clash;
+    return { %$db, macs => $macs };
+}
+
 sub json_reader {
     my ($filename, $data) = @_;
 
-    my $db = defined($data) && length($data) > 0 ? decode_json($data) : {};
-    for my $mac (keys %{ $db->{macs} // {} }) {
-        my $entry = $db->{macs}->{$mac};
-        delete $db->{macs}->{$mac} if !defined($entry->{ip4}) && !defined($entry->{ip6});
-    }
-    return $db;
+    return fold_macdb(defined($data) && length($data) > 0 ? decode_json($data) : {});
 }
 
 sub json_writer {
@@ -54,8 +68,10 @@ sub write_macdb {
     cfs_write_file($macdb_filename, $data);
 }
 
+# the cache is keyed by the MAC in lower case, whatever spelling its writers use
 sub add_cache_mac_ip {
     my ($mac, $ip) = @_;
+    $mac = lc($mac);
 
     cfs_lock_file(
         $macdb_filename,
@@ -75,6 +91,7 @@ sub add_cache_mac_ip {
 
 sub del_cache_mac_ip {
     my ($mac, $ip) = @_;
+    $mac = lc($mac);
 
     cfs_lock_file(
         $macdb_filename,
@@ -142,7 +159,7 @@ sub complete_sdn_vnet {
 sub get_ips_from_mac {
     my ($mac, $zoneid, $zone) = @_;
 
-    my $cached = read_macdb()->{macs}->{$mac};
+    my $cached = read_macdb()->{macs}->{ lc($mac) };
     return ($cached->{ip4}, $cached->{ip6}) if $cached;
 
     my $plugin_config = get_plugin_config($zone);
@@ -157,8 +174,8 @@ sub get_ips_from_mac {
         undef,
         sub {
             my $db = read_macdb();
-            $db->{macs}->{$mac}->{ip4} = $ip4 if defined($ip4);
-            $db->{macs}->{$mac}->{ip6} = $ip6 if defined($ip6);
+            $db->{macs}->{ lc($mac) }->{ip4} = $ip4 if defined($ip4);
+            $db->{macs}->{ lc($mac) }->{ip6} = $ip6 if defined($ip6);
             write_macdb($db);
         },
     );
diff --git a/src/test/run_test_vnets_blackbox.pl b/src/test/run_test_vnets_blackbox.pl
index d8e35231..6b89778f 100755
--- a/src/test/run_test_vnets_blackbox.pl
+++ b/src/test/run_test_vnets_blackbox.pl
@@ -471,6 +471,88 @@ sub test_ipam_cache_misses {
 
 run_test(\&test_ipam_cache_misses);
 
+sub test_ipam_cache_case {
+    my $test_name = (split(/::/, (caller(0))[3]))[-1];
+    my $zoneid = "TESTZONE";
+    my $vnetid = "testvnet";
+    my $mac = "da:65:8f:18:9b:6f";
+
+    # an entry written under the config's spelling before the cache was keyed in
+    # lower case is read and released like any other
+    create_zone({
+        type => "simple",
+        dhcp => "dnsmasq",
+        ipam => "pve",
+        zone => $zoneid,
+    });
+    create_vnet({
+        type => "vnet",
+        zone => $zoneid,
+        vnet => $vnetid,
+    });
+    create_subnet({
+        type => "subnet",
+        vnet => $vnetid,
+        subnet => "10.0.0.0/24",
+        gateway => "10.0.0.1",
+        'dhcp-range' => ["start-address=10.0.0.100,end-address=10.0.0.200"],
+    });
+    $test_state->{macdb}->{macs}->{ uc($mac) } = { ip4 => '10.0.0.150' };
+    my ($ip4) = PVE::Network::SDN::Vnets::get_ips_from_mac($vnetid, $mac);
+    is($ip4, '10.0.0.150', "$test_name: an old entry is found however it is spelled");
+    PVE::Network::SDN::Ipams::del_cache_mac_ip($mac, '10.0.0.150');
+    ok(
+        !(grep { lc($_) eq $mac } keys $test_state->{macdb}->{macs}->%*),
+        "$test_name: and released",
+    );
+
+    # a write spelled either way moves such an entry to the lower-case key
+    # and keeps what it held
+    $test_state->{macdb}->{macs}->{ uc($mac) } = { ip4 => '10.0.0.150' };
+    PVE::Network::SDN::Ipams::add_cache_mac_ip(uc($mac), 'fd00::150');
+    is_deeply(
+        [grep { lc($_) eq $mac } keys $test_state->{macdb}->{macs}->%*],
+        [$mac],
+        "$test_name: a write leaves one entry for the MAC",
+    );
+    is_deeply(
+        $test_state->{macdb}->{macs}->{$mac},
+        { ip4 => '10.0.0.150', ip6 => 'fd00::150' },
+        "$test_name: holding both addresses",
+    );
+
+    # the lookup's write moves an address-less entry under another spelling as well
+    create_ip({
+        zone => $zoneid,
+        vnet => $vnetid,
+        mac => $mac,
+        ip => "10.0.0.160",
+    });
+    $test_state->{macdb}->{macs} = { uc($mac) => { ip4 => undef, ip6 => undef } };
+    my @ips = PVE::Network::SDN::Vnets::get_ips_from_mac($vnetid, $mac);
+    is($ips[0], '10.0.0.160', "$test_name: a lookup past an address-less entry asks the IPAM");
+    is_deeply(
+        [sort keys $test_state->{macdb}->{macs}->%*],
+        [$mac],
+        "$test_name: and its write leaves one entry for the MAC",
+    );
+
+    # two spellings that disagree are not trusted, the lookup asks the IPAM
+    $test_state->{macdb}->{macs} = {
+        uc($mac) => { ip4 => '10.0.0.150' },
+        $mac => { ip4 => '10.0.0.151' },
+    };
+    @ips = PVE::Network::SDN::Vnets::get_ips_from_mac($vnetid, $mac);
+    is($ips[0], '10.0.0.160', "$test_name: a clash of spellings counts as a miss");
+    is_deeply(
+        $test_state->{macdb}->{macs},
+        { $mac => { ip4 => '10.0.0.160' } },
+        "$test_name: and its write leaves the IPAM's answer alone",
+    );
+}
+
+run_test(\&test_ipam_cache_case);
+
 sub test_nic_join {
     my ($test_name, $subnets) = @_;
 
-- 
2.47.3





  parent reply	other threads:[~2026-10-21  8:46 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-21  8:44 [PATCH container/docs/manager/network/proxmox{-ebpf,-perl-rs,-ve-rs}/qemu-server v4 00/20] sdn: implement DHCP for all zones using eBPF Hannes Laimer
2026-10-21  8:44 ` [PATCH proxmox-ebpf v4 01/20] dhcp: add the per-tap DHCPv4 responder program Hannes Laimer
2026-10-21  8:44 ` [PATCH proxmox-ebpf v4 02/20] dhcp: add the subsystem that loads the responder and keeps its records Hannes Laimer
2026-10-21  8:44 ` [PATCH proxmox-perl-rs v4 03/20] pve-rs: sdn: add bindings for the eBPF DHCP responder Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 04/20] sdn: ipam: stop caching IPAM misses Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 05/20] sdn: ipam: write the lookup's cache entry under the cluster lock Hannes Laimer
2026-10-21  8:44 ` Hannes Laimer [this message]
2026-10-21  8:44 ` [PATCH pve-network v4 07/20] sdn: dhcp: replace a stale dnsmasq reservation instead of dropping it Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 08/20] sdn: dhcp: update reservations when ip mappings change Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 09/20] sdn: dhcp: make the lease time configurable per subnet Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 10/20] sdn: dhcp: require a backend only if a zone on this node uses it Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 11/20] sdn: dhcp: add an eBPF backend answering on each guest interface Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 12/20] sdn: zones: report interface plugs and unplugs to the DHCP backends Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 13/20] sdn: dhcp: run IP mapping edits on the guest's node Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-network v4 14/20] sdn: zones: allow the ebpf DHCP backend on every zone type Hannes Laimer
2026-10-21  8:44 ` [PATCH qemu-server v4 15/20] network: sdn: keep the DHCP backends current on guest NICs Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-container v4 16/20] net: report NIC plugs and removals to the SDN DHCP backends Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-manager v4 17/20] ui: sdn: add zone DHCP backend selector and subnet DHCP options Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-manager v4 18/20] sdn: bring up the DHCP backends at boot before the guests start Hannes Laimer
2026-10-21  8:44 ` [PATCH proxmox-ve-rs v4 19/20] ve-config: sdn: add the ebpf dhcp backend Hannes Laimer
2026-10-21  8:44 ` [PATCH pve-docs v4 20/20] sdn: dhcp: add eBPF backend section Hannes Laimer

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=20261021084445.462195-7-h.laimer@proxmox.com \
    --to=h.laimer@proxmox.com \
    --cc=pve-devel@lists.proxmox.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal