Bug 15892 - Intel Compute Stick: RTWHALXT runs 100% CPU for a long time
Summary: Intel Compute Stick: RTWHALXT runs 100% CPU for a long time
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: x86_64 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Thomas Backlund
QA Contact:
URL:
Whiteboard: MGA5TOO
Keywords:
: 15789 (view as bug list)
Depends on: 16655
Blocks:
  Show dependency treegraph
 
Reported: 2015-05-09 04:19 CEST by Bjarne Thomsen
Modified: 2015-12-05 13:01 CET (History)
2 users (show)

See Also:
Source RPM: kernel
CVE:
Status comment:


Attachments
dmesg output after re-boot (52.38 KB, text/plain)
2015-05-09 14:11 CEST, Bjarne Thomsen
Details
journalctl -b output after re-boot (86.42 KB, text/plain)
2015-05-09 14:16 CEST, Bjarne Thomsen
Details
Error from drakroam during Wi-Fi configuration (881 bytes, text/plain)
2015-05-12 17:50 CEST, Bjarne Thomsen
Details
lspcidrake -v (1.50 KB, text/plain)
2015-05-12 17:52 CEST, Bjarne Thomsen
Details
journalctl for May 12 after crash and re-boot (36.06 KB, application/x-xz)
2015-05-12 17:57 CEST, Bjarne Thomsen
Details

Description Bjarne Thomsen 2015-05-09 04:19:04 CEST
Description of problem:
I was trying to update Mageia-5-RC-x86_64 to kernel-3.19.7, but
It takes forever to search the database for packages.
Top tells me that RTWHALXT 100% of CPU time. What does it do?
I used btrfs on all partitions. Could that be a problem.
Could it be swapping? But why?

Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.


Reproducible: 

Steps to Reproduce:
Comment 1 Bjarne Thomsen 2015-05-09 04:46:22 CEST
A re-boot did the trick (but still, why?). It is now updating 169 packages.
Comment 2 Bjarne Thomsen 2015-05-09 05:35:04 CEST
My Intel Compute Stick is now running kernel-3.19.7-desktop-1 with Wi-Fi working.
According to Tom's Hardware it should also have Bluetooth......
Comment 3 Bjarne Thomsen 2015-05-09 10:35:36 CEST
Unfortunately: this morning RTWHALXT is again using all CPU power.
Comment 4 Bjarne Thomsen 2015-05-09 11:29:22 CEST
The WiFi connection is hanging, but this is not all. Monotoring network tells me that there are 3 udp connections established to remote port 123 of these addresses
78.156.103.10
77.66.33.146
195.234.155.123
nslookup 195.234.155.123
123.155.234.195.in-addr.arpa    name = ntp.dvconsulting.dk
155.234.195.in-addr.arpa        nameserver = ns10.skaarupaps.dk.
155.234.195.in-addr.arpa        nameserver = ns11.skaarupaps.dk.
whois ntp.dvconsulting.dk
# Hello 87.104.6.180. Your session has been logged.
#
# Copyright (c) 2002 - 2015 by DK Hostmaster A/S
#
# Version: 2.0.2
No entries found for the selected source.
This looks as an attack.
I had the impression that chrony does not use ntp servers.
By the way. Are you aware that mageia.org is graded F by ssllabs.com?
Comment 5 Bjarne Thomsen 2015-05-09 12:19:23 CEST
Too fast!
whois dvconsulting.dk
Registrant
Handle:               DCA300-DK
Name:                 DV CONSULTING A/S
Address:              Strandvejen 100
Postalcode:           2900
City:                 Hellerup
Country:              DK
Phone:                +45 26707070

I am running top to look what happens.
Maybe the driveris unstable.
Comment 6 Bjarne Thomsen 2015-05-09 14:08:39 CEST
It does not look good. I was running ping for more than an hour without problems.
Then I installet traceroot with
urpmi traceroot
When it had finished, I started typing, an the screen froze.
I had to turn off power and re-boot.
After the reboot, the Wi-Fi did not come up again.
I ran dmesg and journalctl -b. The outputs of these are attached.
Comment 7 Bjarne Thomsen 2015-05-09 14:11:23 CEST
Created attachment 6483 [details]
dmesg output after re-boot

CC: (none) => bjarne.thomsen

Comment 8 Bjarne Thomsen 2015-05-09 14:16:25 CEST
Created attachment 6484 [details]
journalctl -b output after re-boot

Call Trace:
[<ffffffff816c19da>] dump_stack+0x45/0x57
etc. etc.
Comment 9 Thomas Backlund 2015-05-09 14:34:06 CEST
RTWHALXT is part of the new wifi driver yes, and its a little bit of a mess right now and needs more work...

I will look if there are more valid "safe" fixes to pull in...

CC: (none) => tmb

Comment 10 Bjarne Thomsen 2015-05-11 10:40:13 CEST
I have made some further tests of the Wi-Fi driver.
Normally it works between 1 and 4 hours. Then the system freezes, and I have to use the power button to re-boot.
But there has been sufficient time to update the system.
Comment 11 Thomas Backlund 2015-05-12 14:04:45 CEST
does the new kernel-3.19.8-1 work any better ?
Comment 12 Bjarne Thomsen 2015-05-12 14:52:13 CEST
I do not think so, but let be try again.
The system now completely freezes. In principle it could be something else.
Is there a way to use journalctl to trace what exactly happened before it
froze, that is after a reboot?
On the other hand, I have not seen it happen, if I remove the WiFi interface.
Comment 13 Bjarne Thomsen 2015-05-12 15:32:59 CEST
Actually, I had not updated to kernel-3.19.8-1 on the Compute Stick.
This is now done, and ping has been running since 15:15.
I did notice one thing that I had overlooked.
It only finds my Asus EA-N66 access point. I also have a Cisco AP.
Comment 14 Bjarne Thomsen 2015-05-12 16:03:08 CEST
After 45min of running ping the screen is now frozen.
Anything that I can do to figure out what happened?
Comment 15 Thomas Backlund 2015-05-12 16:33:16 CEST
So is it better or worse than before ?

Is the Cisco AP a 2.4 or 5G one

as for figuring out...

When you reboot / start a new ping...

open a second terminal window and as root run:

journalctl -f

to keep it tracking all logging

That could maybe catch the oops / crash that happends
Comment 16 Bjarne Thomsen 2015-05-12 17:50:53 CEST
Created attachment 6523 [details]
Error from drakroam during Wi-Fi configuration
Comment 17 Bjarne Thomsen 2015-05-12 17:52:58 CEST
Created attachment 6524 [details]
lspcidrake -v

lspcidrake -v after drakroam error.
Comment 18 Bjarne Thomsen 2015-05-12 17:57:03 CEST
Created attachment 6525 [details]
journalctl for May 12 after crash and re-boot

It only took a few minutes before it crashed.
I included the log for this afternoon.
I am not quite sure what to look for.
Comment 19 Bjarne Thomsen 2015-05-13 11:17:10 CEST
Are you really going th ship mga5 with an EOL kernel?
Comment 20 Rémi Verschelde 2015-05-13 13:08:04 CEST
(In reply to Bjarne Thomsen from comment #19)
> Are you really going th ship mga5 with an EOL kernel?

Well kernels get EOL'ed pretty fast nowadays, 3.19.0 was released in February and is already EOL... If we were to upgrade now to 4.0, it would most likely delay the release of Mageia 5 until we've sorted out all the new issues, and eventually 4.0.x would likely also be EOL in 2-3 months.

I guess the plan is to do a branch update sometimes after Mageia 5's release, and until then Thomas will likely backport fixes of the somewhat-longterm 3.18 branch.
Comment 21 Thomas Backlund 2015-05-13 13:31:20 CEST
This is OT for this bugreport...

but...

I/We will team up with Ubuntu kernel team as they too are using 3.19 ...


Later on we will do a version bump to a newer kernel/mesa/... in order to properly support the upcoming Intel Skylake...

By the looks of it, I think it will be ~4.2 that probably also will be the next -longterm kernel.
Comment 22 Bjarne Thomsen 2015-05-13 15:29:36 CEST
*** Bug 15789 has been marked as a duplicate of this bug. ***
Comment 23 Bjarne Thomsen 2015-05-13 15:57:59 CEST
This sounds good.
Samuel Verschelde 2015-06-06 15:30:10 CEST

Assignee: bugsquad => tmb
Source RPM: Mageia-5-RC-x86_64-DVD => kernel
Whiteboard: (none) => MGA5TOO

Comment 24 Bjarne Thomsen 2015-06-27 13:11:18 CEST
Linux 4.0, Linux 4.1 Brings Performance Boosts For Some Intel Low-Power Hardware:
For at least some Intel Bay Trail systems, the Linux 4.0 and Linux 4.1 kernels bring measurable performance improvements as shown by this latest round of Phoronix kernel benchmarking.
http://www.phoronix.com/scan.php?page=article&item=linux-41-byt&num=1
Comment 25 Thomas Backlund 2015-06-27 17:21:32 CEST
We already carry some of the performance fixes, but yes. 4.1+ will improve things...
Comment 26 Thomas Backlund 2015-09-18 15:26:13 CEST
kernel 4.1.7 is now in updates testing, please try it out

Depends on: (none) => 16655

Comment 27 Bjarne Thomsen 2015-09-19 15:24:30 CEST
Indeed. Wi-Fi has now been working for more than 2 hours.
I have surfed using firefox and downloaded IMAP mails in thunderbird.
Finally, I downloaded a Mageia DVD-iso without any problems.
Comment 28 Bjarne Thomsen 2015-12-05 13:01:10 CET
I am closing this report

Status: NEW => RESOLVED
Resolution: (none) => FIXED


Note You need to log in before you can comment on or make changes to this bug.