lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date:   Tue, 24 Dec 2019 11:24:09 -0800
From:   PGNet Dev <pgnet.dev@...il.com>
To:     netdev@...r.kernel.org
Subject: with kernel 5.4.6, two Eth interfaces -- one 'reliably named', the
 other not. used to work , what's changed?

 I recently upgraded a linux/64 box to

	uname -rm
		5.4.6-24.ge5f8301-default x86_64

For 'ages' prior, I've had two functional Eth interfaces on it

	inxi -n
(1)		Network:   Card-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller driver: r8169
		           IF: eth0 state: down mac: 18:d6:c7:01:15:11
(2)		           Card-2: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller driver: r8169
		           IF: enp3s0 state: up speed: 1000 Mbps duplex: full mac: 00:52:35:50:44:04

where (2)'s the Mobo ETH, and (1)'s an ETH PCI-e card

Both expect/use the same driver,

	lspci -tv | grep -i eth
		+-04.0-[02]----00.0  Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
		+-06.0-[03]----00.0  Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller


The driver's up

	lsmod | grep 8169
		r8169                  94208  0
		libphy                 98304  2 r8169,realtek

provided by

	rpm -q --whatprovides /lib/modules/5.4.6-24.ge5f8301-default/kernel/drivers/net/ethernet/realtek/r8169.ko
		kernel-default-5.4.6-24.1.ge5f8301.x86_64

the cards are available,

I've had reliable naming enabled

	cat /proc/cmdline
		... net.ifnames=1 biosdevname=0 

and the two interfaces, Mobo & PCI, _had_ always appeared as enp2s0 & enp3s0

with current kernel,

	uname -rm
		5.4.6-24.ge5f8301-default x86_64

& firmware packages,

	rpm -qa | grep -i kernel-firmware | sort
		kernel-firmware-all-20191118-36.13.noarch
		kernel-firmware-amdgpu-20191118-36.13.noarch
		kernel-firmware-ath10k-20191118-36.13.noarch
		kernel-firmware-atheros-20191118-36.13.noarch
		kernel-firmware-bluetooth-20191118-36.13.noarch
		kernel-firmware-bnx2-20191118-36.13.noarch
		kernel-firmware-brcm-20191118-36.13.noarch
		kernel-firmware-chelsio-20191118-36.13.noarch
		kernel-firmware-dpaa2-20191118-36.13.noarch
		kernel-firmware-i915-20191118-36.13.noarch
		kernel-firmware-intel-20191118-36.13.noarch
		kernel-firmware-iwlwifi-20191118-36.13.noarch
		kernel-firmware-liquidio-20191118-36.13.noarch
		kernel-firmware-marvell-20191118-36.13.noarch
		kernel-firmware-media-20191118-36.13.noarch
		kernel-firmware-mediatek-20191118-36.13.noarch
		kernel-firmware-mellanox-20191118-36.13.noarch
		kernel-firmware-mwifiex-20191118-36.13.noarch
		kernel-firmware-network-20191118-36.13.noarch
		kernel-firmware-nfp-20191118-36.13.noarch
		kernel-firmware-nvidia-20191118-36.13.noarch
		kernel-firmware-platform-20191118-36.13.noarch
		kernel-firmware-qlogic-20191118-36.13.noarch
		kernel-firmware-radeon-20191118-36.13.noarch
		kernel-firmware-realtek-20191118-36.13.noarch
		kernel-firmware-serial-20191118-36.13.noarch
		kernel-firmware-sound-20191118-36.13.noarch
		kernel-firmware-ti-20191118-36.13.noarch
		kernel-firmware-ueagle-20191118-36.13.noarch
		kernel-firmware-usb-network-20191118-36.13.noarch

The TPLINK PCI card no longer comes up as an 'en*'-named card, per

	https://www.freedesktop.org/software/systemd/man/systemd.net-naming-scheme.html

but rather, incorrectly (?), as 'eth0'

	hwinfo --netcard | egrep -i "Ethernet controller|Driver|addr|Model:|Device:|Device file"
		07: PCI 300.0: 0200 Ethernet controller
		  Model: "Realtek RTL8111/8168 PCI Express Gigabit Ethernet controller"
		  Device: pci 0x8168 "RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller"
		  SubDevice: pci 0x8168 "RTL8111/8168 PCI Express Gigabit Ethernet controller"
		  Driver: "r8169"
		  Driver Modules: "r8169"
		  Device File: enp3s0
		  HW Address: 00:52:35:50:44:04
		  Permanent HW Address: 00:52:35:50:44:04
		  Driver Info #0:
		    Driver Status: r8169 is active
		    Driver Activation Cmd: "modprobe r8169"
		11: PCI 200.0: 0200 Ethernet controller
		  Model: "TP-LINK TG-3468 Gigabit PCI Express Network Adapter"
		  Device: pci 0x8168 "RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller"
		  SubDevice: pci 0x3468 "TG-3468 Gigabit PCI Express Network Adapter"
		  Driver: "r8169"
		  Driver Modules: "r8169"
??		  Device File: eth0
		  HW Address: 18:d6:c7:01:15:11
		  Permanent HW Address: 18:d6:c7:01:15:11
		  Driver Info #0:
		    Driver Status: r8169 is active
		    Driver Activation Cmd: "modprobe r8169"

noting,

	ls -1 /sys/class/net/
		enp3s0@
		eth0@
		lo@ 

in `dmesg`

	dmesg | egrep -i "eth|enp"
		[    4.564854] r8169 0000:02:00.0 eth0: RTL8168e/8111e, 18:d6:c7:01:15:11, XID 2c2, IRQ 27
		[    4.564856] r8169 0000:02:00.0 eth0: jumbo features [frames: 9200 bytes, tx checksumming: ko]
		[    4.568641] r8169 0000:03:00.0 eth1: RTL8168c/8111c, 00:52:35:50:44:04, XID 3c4, IRQ 18
		[    4.568643] r8169 0000:03:00.0 eth1: jumbo features [frames: 6128 bytes, tx checksumming: ko]
		[    4.614030] r8169 0000:03:00.0 enp3s0: renamed from eth1
		[   28.179613] RTL8211B Gigabit Ethernet r8169-300:00: attached PHY driver [RTL8211B Gigabit Ethernet] (mii_bus:phy_addr=r8169-300:00, irq=IGNORE)
		[   28.283488] r8169 0000:03:00.0 enp3s0: Link is Down
		[   30.498955] r8169 0000:03:00.0 enp3s0: Link is Up - 1Gbps/Full - flow control rx/tx
		[   30.498976] IPv6: ADDRCONF(NETDEV_CHANGE): enp3s0: link becomes ready

Something's changed -- as both interfaces used to be properly named per reliable-naming standard.

I _can_ bring up the network on the Mobo's renamed enp3s0 interface ... but no longer on the PCI card.

I'm not clear on why one interface IS using the reliable-naming scheme, and the other is NOT.

Any hints/clues as to the problem &/or a fix?

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ