Fixed
Created: Jul 24, 2016
Updated: Dec 3, 2018
Resolved Date: Aug 22, 2016
Found In Version: 8.0
Fix Version: 8.0.0.9
Severity: Standard
Applicable for: Wind River Linux 8
Component/s: Networking
Config line: /wrlinux-7/wrlinux/configure --enable-board=intel-x86-64 --enable-rootfs=glibc-small+initramfs --enable-kernel=standard --enable-parallel-pkgbuilds=8 --enable-jobs=8 --enable-reconfig --enable-ccache --enable-internet-download=no --with-template=feature/build_libc --enable-bootimage=iso --enable-rm-oldimgs=yes --enable-rm-work=yes --with-rcpl-version=0013
Igb Tx Unit Hangs on long scp transfers, usually between 5 and 15 minutes. Reproduction rate is quite low, it doesn't happen every time.
Example hang trace:
[ 810.868866] igb 0000:04:00.0: Detected Tx Unit Hang
Tx Queue <1>
TDH <82>
TDT <82>
next_to_use <84>
next_to_clean <82>
buffer_info[next_to_clean]
time_stamp <1000789f0>
next_to_watch <ffff880076839830>
jiffies <10007c70c>
desc.status <128200>
[ 810.872501] ------------[ cut here ]------------ [ 810.872541] WARNING: CPU: 0 PID: 198 at /build-big/wrlinux/build/bitbake_build/tmp/work-shared/intel-x86-64/kernel-source/net/sched/sch_generic.c:264 dev_watchdog+0x259/0x270() [ 810.872555] NETDEV WATCHDOG: exteth (igb): transmit queue 1 timed out [ 810.872564] Modules linked in: wl(PO) iTCO_wdt iTCO_vendor_support fjveincam(O) usbdpfp(O) intel_rapl intel_powerclamp coretemp crct10dif_pclmul lpc_ich i2c_i801 iosf_mbi thermal fan battery ac
[ 810.872646] CPU: 0 PID: 198 Comm: systemd-journal Tainted: P O 3.14.39ltsi-WR7.0.0.13_standard #1
[ 810.872657] Hardware name: N/A N/A/N/A, BIOS MVV1R112 X64 12/01/2014 [ 810.872668] 0000000000000009 ffff880079c03da0 ffffffff81b4555f ffff880079c03de8 [ 810.872690] ffff880079c03dd8 ffffffff81082ebd 0000000000000001 ffff8800359b8940 [ 810.872710] 0000000000000000 ffff88006ff7a000 0000000000000008 ffff880079c03e38 [ 810.872730] Call Trace:
[ 810.872741] <IRQ> [<ffffffff81b4555f>] dump_stack+0x4e/0x7a [ 810.872782] [<ffffffff81082ebd>] warn_slowpath_common+0x7d/0xa0 [ 810.872803] [<ffffffff81082f2c>] warn_slowpath_fmt+0x4c/0x50 [ 810.872825] [<ffffffff81984af9>] dev_watchdog+0x259/0x270 [ 810.872846] [<ffffffff819848a0>] ? dev_graft_qdisc+0x80/0x80 [ 810.872868] [<ffffffff8108fd3b>] call_timer_fn+0x3b/0x170 [ 810.872888] [<ffffffff819848a0>] ? dev_graft_qdisc+0x80/0x80 [ 810.872909] [<ffffffff810905d4>] run_timer_softirq+0x1c4/0x2d0 [ 810.872929] [<ffffffff81087e37>] __do_softirq+0xb7/0x2e0 [ 810.872949] [<ffffffff8108819e>] irq_exit+0x7e/0xa0 [ 810.872968] [<ffffffff81b57664>] smp_apic_timer_interrupt+0x44/0x50
[ 810.872990] [<ffffffff81b5658a>] apic_timer_interrupt+0x6a/0x70 [ 810.872999] <EOI> [<ffffffff811cd630>] ? page_follow_link_light+0x40/0x40 [ 810.873037] [<ffffffff811d256f>] ? user_path_at_empty+0x5f/0xb0 [ 810.873058] [<ffffffff814751fa>] ? __percpu_counter_add+0x6a/0xa0 [ 810.873078] [<ffffffff811889bb>] ? do_munmap+0x26b/0x410 [ 810.873097] [<ffffffff811c750c>] SyS_readlinkat+0x4c/0x110 [ 810.873119] [<ffffffff81b558f6>] system_call_fastpath+0x1a/0x1f [ 810.873133] ---[ end trace 63f4234e1acecc96 ]--- [ 810.873254] igb 0000:04:00.0 exteth: Reset adapter
[ 810.885875] systemd-networkd[254]: exteth : lost carrier
[ 810.885920] systemd-networkd[254]: exteth : DHCP lease lost
[ 813.160111] igb: exteth NIC Link is Up 10 Mbps Half Duplex, Flow Control: None
[ 813.162193] systemd-networkd[254]: exteth : gained carrier
[ 815.530361] systemd-networkd[254]: exteth : DHCPv4 address 192.168.8.20/24 via 192.168.8.1
Network adapter used on customer's side:
"Our system has two Ethernet ports. The port exhibiting the problem is the I210 and uses the igb driver. The 82547L uses the e1000e driver.
root@localhost:~# lspci | grep -i 'ethernet'
02:00.0 Ethernet controller: Intel Corporation 82574L Gigabit Network Connection
03:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03)"
Occurrence of hang lessened after RCPL 0010 when a fix for the hangs gas been released, but it still happens from time to time
