Two new stable kernels

[Posted September 29, 2015 by ris]

Greg Kroah-Hartman has released stable kernels 4.2.2 and 4.1.9. Both contain numerous fixes throughout the tree.

Two new stable kernels

Posted Sep 30, 2015 9:41 UTC (Wed) by Kioob (subscriber, #56482) [Link] (4 responses)

Hi,

Be careful : I have problems with the 4.1.9 kernel. I've just upgraded 6 servers from Linux 4.1.8 to 4.1.9, and all 6 servers hangs after several minutes.

It starts with :
[ 176.453241] NMI watchdog: BUG: soft lockup - CPU#4 stuck for 23s! [swapper/4:0]
[ 176.453300] Modules linked in: iptable_filter ip_tables x_tables x86_pkg_temp_thermal coretemp ghash_clmulni_intel aesni_intel aes_x86_64 lrw gf128mul glue_helper ablk_helper iTCO_wdt iTCO_vendor_support cryptd sb_edac microcode serio_raw edac_core evdev i2c_i801 lpc_ich mfd_core ioatdma ipmi_si ipmi_msghandler shpchp wmi button nf_conntrack autofs4 btrfs dm_mod raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor raid6_pq raid1 md_mod sg sd_mod isci crc32c_intel libsas mpt2sas igb ahci libahci raid_class i2c_algo_bit libata scsi_transport_sas dca ptp ehci_pci scsi_mod ehci_hcd pps_core i2c_core
[ 176.453326] CPU: 4 PID: 0 Comm: swapper/4 Not tainted 4.1-dae-intel #1
[ 176.453327] Hardware name: Supermicro X9SRE/X9SRE-3F/X9SRi/X9SRi-3F/X9SRE/X9SRE-3F/X9SRi/X9SRi-3F, BIOS 3.0a 01/03/2014
[ 176.453328] task: ffff88087b838920 ti: ffff88087b840000 task.ti: ffff88087b840000
[ 176.453329] RIP: 0010:[<ffffffff813a2e95>] [<ffffffff813a2e95>] _raw_spin_unlock_irqrestore+0x5/0x6
[ 176.453334] RSP: 0018:ffff88087fd03e40 EFLAGS: 00000286
[ 176.453335] RAX: 00000000ffffffff RBX: 000000000000001d RCX: 00000000000000ec
[ 176.453336] RDX: 00000000000000eb RSI: 0000000000000286 RDI: ffff88087fd0d600
[ 176.453336] RBP: ffff88085dc50800 R08: ffff88087fd0d600 R09: 0000000000000f50
[ 176.453337] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88087fd03db8
[ 176.453338] R13: ffffffff813a3e8b R14: ffff88085dc50800 R15: ffff88085cf70ee0
[ 176.453339] FS: 0000000000000000(0000) GS:ffff88087fd00000(0000) knlGS:0000000000000000
[ 176.453340] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 176.453341] CR2: 00007f8a42e21000 CR3: 000000000160b000 CR4: 00000000001406e0
[ 176.453341] Stack:
[ 176.453342] ffffffff81076121 ffffffff810760de ffffffff00000010 0000000000000286
[ 176.453343] 0000000000000286 ffff88085cf70ee0 ffffffff8107614b ffff88085cf70e48
[ 176.453345] ffffffff8131c6d7 ffff88085cf70e48 ffff88085c3b0000 0000000000000001
[ 176.453346] Call Trace:
[ 176.453347] <IRQ>
[ 176.453350] [<ffffffff81076121>] ? try_to_del_timer_sync+0x43/0x4d
[ 176.453352] [<ffffffff810760de>] ? del_timer+0x4d/0x4d
[ 176.453354] [<ffffffff8107614b>] ? del_timer_sync+0x20/0x3d
[ 176.453356] [<ffffffff8131c6d7>] ? inet_csk_reqsk_queue_drop+0x5f/0xa3
[ 176.453357] [<ffffffff8131c880>] ? reqsk_timer_handler+0x165/0x18e
[ 176.453358] [<ffffffff8131c71b>] ? inet_csk_reqsk_queue_drop+0xa3/0xa3
[ 176.453360] [<ffffffff81075c0a>] ? call_timer_fn.isra.30+0x12/0x67
[ 176.453361] [<ffffffff81075dfe>] ? run_timer_softirq+0x17f/0x1a1
[ 176.453365] [<ffffffff8103d6e9>] ? __do_softirq+0xc5/0x1b4
[ 176.453366] [<ffffffff8103d910>] ? irq_exit+0x34/0x78
[ 176.453369] [<ffffffff81024e7f>] ? smp_trace_apic_timer_interrupt+0x6f/0x7b
[ 176.453371] [<ffffffff813a3e8b>] ? apic_timer_interrupt+0x6b/0x70
[ 176.453372] <EOI>
[ 176.453375] [<ffffffff812bc896>] ? cpuidle_enter_state+0xa7/0xff
[ 176.453376] [<ffffffff812bc83b>] ? cpuidle_enter_state+0x4c/0xff
[ 176.453379] [<ffffffff8106082a>] ? cpu_startup_entry+0x1ac/0x258
[ 176.453380] Code: 78 88 ea ff 31 c0 eb 96 48 83 c4 48 89 d0 5b 5d 41 5c 41 5d c3 90 e8 14 ec cb ff 85 c0 75 01 c3 b8 01 00 00 00 c3 80 07 01 56 9d <c3> c6 07 00 56 9d c3 fa e9 3a ec cb ff e9 35 ec cb ff 8b 07 85

Then near 30s after that, servers hang (no ping at all). Downgrade to 4.1.8 kernel solved this issue. So I've forwarded that trace to Greg, maybe this can help.

Two new stable kernels

Posted Sep 30, 2015 15:43 UTC (Wed) by ppedroni (subscriber, #6592) [Link]

Same problem. I've been unable to keep a 4.1.9 kernel up for more than 5 minutes.

Two new stable kernels

Posted Sep 30, 2015 21:48 UTC (Wed) by yascor (guest, #104713) [Link] (1 responses)

Same problem... "stable" LTS. Thought linux kernel will stay away from this "rush rush rush" attitude.
Lesson taken to not trust any (fresh) updates.

Kernel 4.1.7 version was fine.

btw "According to Mr. Kroah-Hartman, all users of the Linux 4.1 LTS kernel series must update the kernel packages in their GNU/Linux operating systems to version 4.1.9 as soon as possible."

ASP? Never again.

Two new stable kernels

Posted Oct 4, 2015 12:54 UTC (Sun) by mattallmill (guest, #104755) [Link]

I'm appalled at the attitude that 'once burned, twice shy'; even kernel developers are not perfect. Nevertheless, Greg would be well-advised to tone down his message to account for the possibility of a regression.

Two new stable kernels

Posted Oct 3, 2015 5:09 UTC (Sat) by DG (subscriber, #16978) [Link]

Problem seen here too :-(

Two new stable kernels

Posted Oct 3, 2015 8:15 UTC (Sat) by Furinax (subscriber, #104748) [Link]

after a bit of digging in the kernel mailing list archives: it seems a fix for this known issue was accidentally missed when picking patches for the stable series...

...in the hope it isn't something else - try this one-liner fix:
https://kernel.googlesource.com/pub/scm/linux/kernel/git/...