85% performance hit in iwlwifi from 12.04 to 14.04

[Pablo Piaggio]

Summary: Almost 10 times slower speeds using 14.04 compared with 12.04 (same machine). Using iperf, Precise: 172 Mbits/sec. Trusty: 27.4 Mbits/sec. Using rsync: Precise: 19.83MB/s. Trusty: 3.12MB/s.

Details:

Machine:
    Laptop Toshiba Satellite A665
    Triple boot Ubuntu 12.04, 13.04 and 14.04

Network (same network for both):
    d-link router
    5Ghz rage

Wireless card: Centrino

$ lspci -nnk | grep -iA2 network
07:00.0 Network controller [0280]: Intel Corporation Centrino Advanced-N + WiMAX 6250 [Kilmer Peak] [8086:0087] (rev 5f)
 Subsystem: Intel Corporation Centrino Advanced-N + WiMAX 6250 2x2 AGN [8086:1301]
 Kernel driver in use: iwlwifi

12.04 Precise
#######################################################################

$ lsb_release -a
Distributor ID: Ubuntu
Description: Ubuntu 12.04.4 LTS
Release: 12.04
Codename: precise

 $ uname -a
Linux vanhalen 3.2.0-29-generic #46-Ubuntu SMP Fri Jul 27 17:03:23 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

$ iperf -c reinhardt
------------------------------------------------------------
Client connecting to reinhardt, TCP port 5001
TCP window size: 23.5 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.105 port 43135 connected with 192.168.1.1 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 206 MBytes 172 Mbits/sec

$ rsync -vP LAS.s22e10.432p.mp4 reinhardt:./
user@reinhardt's password:
LAS.s22e10.432p.mp4
   235861112 100% 19.83MB/s 0:00:11 (xfer#1, to-check=0/1)

sent 235889984 bytes received 31 bytes 14296364.55 bytes/sec
total size is 235861112 speedup is 1.00

14.04 Trusty
#######################################################################

$ lsb_release -a
Distributor ID: Ubuntu
Description: Ubuntu 14.04 LTS
Release: 14.04
Codename: trusty

$ uname -a
Linux vanhalen 3.13.0-24-generic #47-Ubuntu SMP Fri May 2 23:30:00 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

$ iperf -c reinhardt
------------------------------------------------------------
Client connecting to reinhardt, TCP port 5001
TCP window size: 85.0 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.105 port 49807 connected with 192.168.1.1 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.1 sec 32.9 MBytes 27.4 Mbits/sec

$ rsync -e "ssh -p2312" -vP LAS.s22e10.432p.mp4 reinhardt:./
user@reinhardt's password:
LAS.s22e10.432p.mp4
    235,861,112 100% 3.12MB/s 0:01:12 (xfr#1, to-chk=0/1)

sent 235,918,776 bytes received 34 bytes 2,930,668.45 bytes/sec
total size is 235,861,112 speedup is 1.00

ProblemType: Bug
DistroRelease: Ubuntu 14.04
Package: linux-image-extra-3.13.0-24-generic 3.13.0-24.47
ProcVersionSignature: Ubuntu 3.13.0-24.47-generic 3.13.9
Uname: Linux 3.13.0-24-generic x86_64
NonfreeKernelModules: nvidia
ApportVersion: 2.14.1-0ubuntu3.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: pablo 1977 F.... pulseaudio
 /dev/snd/controlC0: pablo 1977 F.... pulseaudio
CurrentDesktop: Unity
Date: Wed May 14 22:38:47 2014
HibernationDevice: RESUME=UUID=fed461e0-ae63-4fb1-8d92-8140a1b06551
InstallationDate: Installed on 2014-05-13 (2 days ago)
InstallationMedia: Ubuntu 14.04 LTS "Trusty Tahr" - Release amd64 (20140417)
MachineType: TOSHIBA Satellite A665
ProcFB: 0 VESA VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.13.0-24-generic root=UUID=77c75c70-93b5-41fa-984c-4fa7f9473ffc ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-3.13.0-24-generic N/A
 linux-backports-modules-3.13.0-24-generic N/A
 linux-firmware 1.127.2
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 09/14/10
dmi.bios.vendor: TOSHIBA
dmi.bios.version: 1.80
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: NWQAA
dmi.board.vendor: TOSHIBA
dmi.board.version: 1.00
dmi.chassis.asset.tag: *
dmi.chassis.type: 9
dmi.chassis.vendor: TOSHIBA
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnTOSHIBA:bvr1.80:bd09/14/10:svnTOSHIBA:pnSatelliteA665:pvrPSAW3U-0DE039:rvnTOSHIBA:rnNWQAA:rvr1.00:cvnTOSHIBA:ct9:cvrN/A:
dmi.product.name: Satellite A665
dmi.product.version: PSAW3U-0DE039
dmi.sys.vendor: TOSHIBA

[Brad Figg (brad-figg)]

This change was made by a bot.

[Pablo Piaggio (papibe)]

The line:
12.04 Trusty
#######################################################################

should say:
12.04 Precise
#######################################################################

[Pablo Piaggio (papibe)]

Lines edited.

(comment #3 no longer applies),

[Joseph Salisbury (jsalisbury)]

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v3.15 kernel[0].

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

If you are unable to test the mainline kernel, for example it will not boot, please add the tag: 'kernel-unable-to-test-upstream'.
Once testing of the upstream kernel is complete, please mark this bug as "Confirmed".

Thanks in advance.

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.15-rc5-utopic/

[Pablo Piaggio (papibe)]

Hi Joseph. Thanks for your interest.

Working on it. I'll post my results ASAP.

P.S.: I'm a little confused. I know the iwlfifi driver is not in linux-image-generic*, but in linux-image-extra-*. Does this mean that the driver will remain the same? Is your assessment, then, that the problem lays on how the kernel itself works with the driver?

Regards.

[Pablo Piaggio (papibe)]

Done. Bad news. It does not fix the problem. The speeds remain slow as before.

Tagged as 'kernel-bug-exists-upstream' as requested.

Any other thought?

Details:
---------------------------------------------------------------------------

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04 LTS
Release: 14.04
Codename: trusty

$ uname -a
Linux vanhalen 3.15.0-031500rc5-generic #201405091635 SMP Fri May 9 20:36:31 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

$ iperf -c reinhardt
------------------------------------------------------------
Client connecting to reinhardt, TCP port 5001
TCP window size: 85.0 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.105 port 33675 connected with 192.168.1.1 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.1 sec 32.5 MBytes 27.1 Mbits/sec
pablo@vanhalen:~$ l
bookmarks-2014-05-12.json Documents/ Music/ report.txt Videos/
Desktop/ Downloads/ Pictures/ report.txt~
devproj/ examples.desktop Public/ Templates/

$ rsync -e "ssh -p2312" -vP LAS.s22e10.432p.mp4 reinhardt:./
user@reinhardt's password:
LAS.s22e10.432p.mp4
    235,861,112 100% 3.31MB/s 0:01:07 (xfr#1, to-chk=0/1)

sent 235,918,776 bytes received 34 bytes 3,209,779.73 bytes/sec
total size is 235,861,112 speedup is 1.00
------------------------------------------------------------------------------------------------------------

P.S.: I answered my own question regarding the driver: iwlwifi comes in the upstream kernel:

$dpkg -S $(modinfo iwlwifi | head -1 | awk '{print $2}')
linux-image-3.15.0-031500rc5-generic: /lib/modules/3.15.0-031500rc5-generic/kernel/drivers/net/wireless/iwlwifi/iwlwifi.ko

[Joseph Salisbury (jsalisbury)]

This issue appears to be an upstream bug, since you tested the latest upstream kernel. Would it be possible for you to open an upstream bug report[0]? That will allow the upstream Developers to examine the issue, and may provide a quicker resolution to the bug.

Please follow the instructions on the wiki page[0]. The first step is to email the appropriate mailing list. If no response is received, then a bug may be opened on bugzilla.kernel.org.

Once this bug is reported upstream, please add the tag: 'kernel-bug-reported-upstream'.

[0] https://wiki.ubuntu.com/Bugs/Upstream/kernel

[Pablo Piaggio (papibe)]

Yes. No problem.

Working on it.

I'll update when done.

[Pablo Piaggio (papibe)]

Got an answer right away. I'll test this tomorrow morning.

-------- Original Message --------
Date: Tue, 20 May 2014 07:58:10 +0300
From: Emmanuel Grumbach <email address hidden>

On Tue, May 20, 2014 at 7:19 AM, Pablo Piaggio <email address hidden> wrote:
> [1.] One line summary of the problem:
> 85% decrease performance in iwlwifi from 3.2.0-29-generic to
> 3.15.0-031500rc5-generic

add 11n_disable=8 as a module parameter.
We disabled TX AMPDU by default due to issues users were seeing. If
you are after performance, add this module parameter and you'll get
your performance back.

[Seth Forshee (sforshee)]

Pablo: Did you test Emmanuel's suggestion, and did it help?

If so, you can make the workaround permanent by creating a file in the /etc/modprobe.d directory (e.g. iwlwifi-11n.conf) which contains the line "options iwlwifi 11n_disable=8". Tx AMPDU has caused problems for enough people that I don't think there's much chance of changing the default back to on.

[Pablo Piaggio (papibe)]

Hi again Joseph. Thanks for your guidance through out this process :)

Yes, this bring the speed back close as it were in 12.04. I tested on both the upstream kernel, and the current Trusty 14.04 kernel. Both are working at fairly high speeds again (see below).

So I guess this one was not exactly a bug, but a conscious decision to compromise speed in favor of stability.

Kind Regards.

Upstream kernel
######################################################################

$ uname -a
Linux vanhalen 3.15.0-031500rc5-generic #201405091635 SMP Fri May 9 20:36:31 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

$ iperf -c reinhardt
------------------------------------------------------------
Client connecting to reinhardt, TCP port 5001
TCP window size: 85.0 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.105 port 49994 connected with 192.168.1.1 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 104 MBytes 87.1 Mbits/sec

$ rsync -e "ssh -p2312" -vP LAS.s22e10.432p.mp4 reinhardt:
user@reinhardt's password:
LAS.s22e10.432p.mp4
    235,861,112 100% 13.49MB/s 0:00:16 (xfr#1, to-chk=0/1)

sent 235,918,776 bytes received 34 bytes 10,039,098.30 bytes/sec
total size is 235,861,112 speedup is 1.00

Current Trusty kernel
######################################################################

$ uname -a
Linux vanhalen 3.13.0-24-generic #47-Ubuntu SMP Fri May 2 23:30:00 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

$ iperf -c reinhardt
------------------------------------------------------------
Client connecting to reinhardt, TCP port 5001
TCP window size: 85.0 KByte (default)
------------------------------------------------------------
[ 3] local 192.168.1.105 port 54472 connected with 192.168.1.1 port 5001
[ ID] Interval Transfer Bandwidth
[ 3] 0.0-10.0 sec 132 MBytes 110 Mbits/sec

$ rsync -vP LAS.s22e10.432p.mp4 reinhardt:
user@reinhardt's password:
LAS.s22e10.432p.mp4
    235,861,112 100% 12.12MB/s 0:00:18 (xfr#1, to-chk=0/1)

sent 235,918,776 bytes received 34 bytes 8,902,596.60 bytes/sec
total size is 235,861,112 speedup is 1.00

[Seth Forshee (sforshee)]

Thanks, Pablo. In that case I'm going to close the bug.

[madbiologist (me-again)]

This might be fixed in the Ubuntu 15.04 "Vivid Vervet" kernel 3.19.0-17.17 by this commit:

iwlwifi: fix max_ht_ampdu_exponent for older devices

The diff file mentions your Intel Corporation Centrino Advanced-N + WiMAX 6250 2x2 AGN.

[Thulasiram (thulasiram-valleru)]

My Dell XPS 14Z has Intel Centrino Advanced 6230 WLAN adapter and with Linux driver, the connection often drops and even the bandwidth is reduced to 5 Mpbs from 25 Mpbs which I get in Windows. I am aware that the bug was present even in Windows but a firmware update is released.

I tried to follow Ubuntu forums and wireless Kernel page to keep track of this bug and I didn't find any change to this Linux driver or even a work around. I followed the suggestions from the forums and the result is unfruitful and remains the same with each iteration of Kernel version.

I tried 15.04, other variants of Linux distribution although knowing that every distribution use the same firmware from kernel.org. I lost my hope already and looking for an alternative PCI-e Wifi adapter. For almost three years, Linux community couldn't give a solution and I think they even hate the idea of releasing a patch.