I have an ancient UP i586 machine which successfully runs 6.4.16 but which crashes without logging an oops shortly after booting either 6.5.13 or 6.6.8. This bug *might* be network-related, but is not fixed by: ``` --- linux-6.5/include/net/neighbour.h.orig 2023-12-10 22:11:54.079741645 +0000 +++ linux-6.5/include/net/neighbour.h 2023-12-10 22:12:24.920364781 +0000 @@ -162,7 +162,7 @@ struct rcu_head rcu; struct net_device *dev; netdevice_tracker dev_tracker; - u8 primary_key[0]; + u8 primary_key[]; } __randomize_layout; struct neigh_ops { ``` The dmesg log seems to get this far before stopping: ``` NET: Registered PF_INET6 protocol family Segment Routing with IPv6 In-situ OAM (IOAM) with IPv6 bridge: filtering via arp/ip/ip6tables is no longer available by default. Update your scripts to load br_netfilter if you need this. br0: port 1(eth1) entered blocking state br0: port 1(eth1) entered disabled state e100 0000:01:04.0 eth1: entered allmulticast mode e100 0000:01:04.0 eth1: entered promiscuous mode e100 0000:01:04.0 eth1: NIC Link is Up 100 Mbps Full Duplex br0: port 2(eth2) entered blocking state br0: port 2(eth2) entered disabled state e100 0000:01:05.0 eth2: entered allmulticast mode e100 0000:01:05.0 eth2: entered promiscuous mode e100 0000:01:05.0 eth2: NIC Link is Up 100 Mbps Full Duplex br0: port 2(eth2) entered blocking state br0: port 2(eth2) entered forwarding state br0: port 1(eth1) entered blocking state br0: port 1(eth1) entered forwarding state RPC: Registered named UNIX socket transport module. RPC: Registered udp transport module. RPC: Registered tcp transport module. RPC: Registered tcp-with-tls transport module. RPC: Registered tcp NFSv4.1 backchannel transport module. e100 0000:00:0f.0 eth0: NIC Link is Up 100 Mbps Full Duplex ```
Created attachment 305637 [details] 6.5.13 config
Created attachment 305638 [details] 6.6.8 config
Created attachment 305642 [details] dmesg log for Linux 6.5.0 Linux 6.5.0 also fails on this UP machine.
Chris, does a 6.4 kernel run successfully? Please post a successful boot log and kernel .config file.
Created attachment 305643 [details] 6.4.16 config
Created attachment 305644 [details] dmesg log for 6.4.16 Dmesg log for successful boot with 6.4.16.
Created attachment 305645 [details] Crash dump - 1 6.6.8 again, except recompiled with an updated toolchain. Once the kernel had locked up, I managed to trigger an oops via SysRq-"kill all tasks". I was obviously only able to capture what would fit on my screen at the time.
Created attachment 305646 [details] Crash dump - 2 Same crash, screenshot 2.
Created attachment 305647 [details] Crash dump - 3 Same crash, screenshot 3.
Created attachment 305648 [details] Crash dump - 4 Same crash, screenshot 4. I gave up after this and rebooted the box back to 6.4.16.
One obvious theory is that the e100 driver could be broken as of 6.5.0.