2017-12-26 20:52:39

by Toralf Förster

[permalink] [raw]
Subject: 4.14.9 boot error (regression), 4.14.8 is fine

at a headless server: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz, stable hardened Gentoo Linux
The hang does occur before any log message is written.
A ping command shows, that the server does reboot and dies after 20 pings are received.
The .config is attached

FWIW I'm surprised a little bit about having a new kernel option in a LTS release.

--
Toralf
PGP C4EACDDE 0076E94E


Attachments:
.config (72.63 kB)

2017-12-26 21:24:48

by Ozgur

[permalink] [raw]
Subject: Re: 4.14.9 boot error (regression), 4.14.8 is fine

26.12.2017, 23:52, "Toralf Förster" <[email protected]>:
> at a headless server: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz, stable hardened Gentoo Linux
> The hang does occur before any log message is written.
> A ping command shows, that the server does reboot and dies after 20 pings are received.
> The .config is attached

Hello,
please could you little more detail of the have received errors?
You compile the kernel right? So, system is boot but not the network respond?

Regards

> FWIW I'm surprised a little bit about having a new kernel option in a LTS release.
>
> --
> Toralf
> PGP C4EACDDE 0076E94E

Ozgur

2017-12-26 21:35:41

by Toralf Förster

[permalink] [raw]
Subject: Re: 4.14.9 boot error (regression), 4.14.8 is fine

On 12/26/2017 10:16 PM, Ozgur wrote:
> You compile the kernel right? So, system is boot but not the network respond?
>
> Regards
Yes, I compile was fine the kernel (tried both new "kernel unwinder" kernel options", made a distclean before).

All, what I can tell is in moment, it looks like that the network is up for a short while before the hang does offur.

--
Toralf
PGP C4EACDDE 0076E94E

2017-12-26 21:52:20

by Randy Dunlap

[permalink] [raw]
Subject: Re: 4.14.9 boot error (regression), 4.14.8 is fine

On 12/26/2017 12:52 PM, Toralf Förster wrote:
> at a headless server: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz, stable hardened Gentoo Linux
> The hang does occur before any log message is written.
> A ping command shows, that the server does reboot and dies after 20 pings are received.
> The .config is attached
>
> FWIW I'm surprised a little bit about having a new kernel option in a LTS release.

It's just a renamed kernel config option, right?
and a different default setting (UNWINDER_ORC became the new default).


--
~Randy

2017-12-26 23:54:34

by Ozgur

[permalink] [raw]
Subject: Re: 4.14.9 boot error (regression), 4.14.8 is fine


27.12.2017, 00:52, "Randy Dunlap" <[email protected]>:
> On 12/26/2017 12:52 PM, Toralf Förster wrote:
>>  at a headless server: Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz, stable hardened Gentoo Linux
>>  The hang does occur before any log message is written.
>>  A ping command shows, that the server does reboot and dies after 20 pings are received.
>>  The .config is attached

Hello,

Your .config file requires "yes" for most parameters.
Do you just run the "make" command by default config working copy /boot directory?

$ copy /boot/config* .config
$ make

Have you try this default .config? Everything is good?

>>  FWIW I'm surprised a little bit about having a new kernel option in a LTS release.
>
> It's just a renamed kernel config option, right?
> and a different default setting (UNWINDER_ORC became the new default).
>
> --
> ~Randy

Ozgur

2017-12-27 10:23:37

by Toralf Förster

[permalink] [raw]
Subject: Re: 4.14.9 boot error (regression), 4.14.8 is fine

On 12/26/2017 11:34 PM, Toralf Förster wrote:
> For now I couldn't get a higher resolution working to catch the whole stack trace.

I attached the last lines of the stack trace.
The kernel parameter vga=ask doesn't work any longer so can't get more than 80x25 in moment.

--
Toralf
PGP C4EACDDE 0076E94E


Attachments:
IMG_20171227_111425.jpg.jpg (271.23 kB)