Bad disk performance with AMD Carrizo FCH controllers

[Evren Yurtesen]

I have had a problem where importing database was much slower in a dual-boot machine when booted into ubuntu compared to running ubuntu inside virtualbox in windows 10. It seems Ubuntu does not like Amd carrizo fch

I collected some information and posted here (and attached as text file to report):
https://answers.launchpad.net/ubuntu/+question/360373

To put it short, the disk performs up to 10 times better inside virtualbox. (yes I thought about caches but VM box also has half the memory and they must favor ubuntu running on real hardware with twice ram....)

I believe this is a kernel/driver related problem since it directly effects disk performance.

ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: linux-image-4.4.0-34-generic 4.4.0-34.53
ProcVersionSignature: Ubuntu 4.4.0-34.53-generic 4.4.15
Uname: Linux 4.4.0-34-generic x86_64
AlsaVersion: Advanced Linux Sound Architecture Driver Version k4.4.0-34-generic.
AplayDevices: Error: [Errno 2] No such file or directory: 'aplay'
ApportVersion: 2.20.1-0ubuntu2.1
Architecture: amd64
ArecordDevices: Error: [Errno 2] No such file or directory: 'arecord'
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: jobgo 3381 F.... pulseaudio
 /dev/snd/controlC0: jobgo 3381 F.... pulseaudio
Card0.Amixer.info: Error: [Errno 2] No such file or directory: 'amixer'
Card0.Amixer.values: Error: [Errno 2] No such file or directory: 'amixer'
Card1.Amixer.info: Error: [Errno 2] No such file or directory: 'amixer'
Card1.Amixer.values: Error: [Errno 2] No such file or directory: 'amixer'
Date: Thu Aug 25 12:36:05 2016
HibernationDevice: RESUME=UUID=ba38c925-a729-4e33-a08f-a7573abd2826
InstallationDate: Installed on 2016-08-23 (1 days ago)
InstallationMedia: Ubuntu-Server 16.04.1 LTS "Xenial Xerus" - Release amd64 (20160719)
IwConfig: Error: [Errno 2] No such file or directory: 'iwconfig'
MachineType: HP HP Pavilion Notebook
ProcEnviron:
 TERM=xterm
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcFB: 0 amdgpudrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.4.0-34-generic.efi.signed root=UUID=cd8e7352-609f-4ebd-83ba-24335f6eece9 ro
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon.
RelatedPackageVersions:
 linux-restricted-modules-4.4.0-34-generic N/A
 linux-backports-modules-4.4.0-34-generic N/A
 linux-firmware 1.157.3
RfKill: Error: [Errno 2] No such file or directory: 'rfkill'
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 03/09/2016
dmi.bios.vendor: American Megatrends Inc.
dmi.bios.version: F.17
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: 80B5
dmi.board.vendor: HP
dmi.board.version: 81.29
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnAmericanMegatrendsInc.:bvrF.17:bd03/09/2016:svnHP:pnHPPavilionNotebook:pvr:rvnHP:rn80B5:rvr81.29:cvnHP:ct10:cvrChassisVersion:
dmi.product.name: HP Pavilion Notebook
dmi.sys.vendor: HP

[Brad Figg (brad-figg)]

This change was made by a bot.

[Joseph Salisbury (jsalisbury)]

Did this issue start happening after an update/upgrade? Was there a prior kernel version where you were not having this particular problem?

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v4.8 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'.

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/v4.8-rc3

[Evren Yurtesen (eyurtese-g)]

Quoting Joseph Salisbury <email address hidden>:

> Did this issue start happening after an update/upgrade? Was there a
> prior kernel version where you were not having this particular problem?

It was a fresh install of xenial, I just got the laptop a short while
ago and installed ubuntu in virtualbox, but now I wanted to try to run
it natively.

I can try an older 12.04 release if you want that I test older kernels
or maybe install an older kernel directly if I find them around :),
because ubuntu is currently totally unusable anyway. So it is ok to
erase it.

> Would it be possible for you to test the latest upstream kernel? Refer
> to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest
> v4.8 kernel[0].

Yes, I will try to test it now. I will return back to you soon.

[Evren Yurtesen (eyurtese-g)]

The 4.8 kernels do not boot at all. I had very strange results. First it gave errors about soft lockup. But the hardware is perfectly fine, I am running windows and xenial (albeit slow) with no crashes or errors. Also ran memtest and everything was fine.

The generic 4.8 is somewhat trying to boot but there are a pile of errors ranging from

AMD-Vi: Completion loop timed out
invalid magic
invalid ELF header
IOTLB_INV_TIMEOUT

The low-latency well...

So, long story short, I can't tell if this is fixed in 4.8 kernels because they do not even boot correctly.

[penalvch (penalvch)]

Evren Yurtesen, could you please provide the full computer model as noted on the sticker of the computer itself (not from the Bug Description, or the result of a terminal command)?

[Evren Yurtesen (eyurtese-g)]

The sticker says:
Model 15-ab114no
ProdID P3L25EA#UUW

Meanwhile, I tried the amdgpu-pro drivers and they keep crashing one way or another. It feels like somebody made sure that linux won't function on this hardware.

[Evren Yurtesen (eyurtese-g)]

1)

# sudo dmidecode -s bios-version && sudo dmidecode -s bios-release-date
F.19
05/27/2016

2)

It makes no difference whatsoever the difference is within reasonable limits. Still much slower than windows host/ubuntu guest.

Test results with new bios:

Operations performed: 14832 Read, 9888 Write, 31616 Other = 56336 Total
Read 231.75Mb Written 154.5Mb Total transferred 386.25Mb (1.2875Mb/sec)
   82.40 Requests/sec executed

Test execution summary:
    total time: 300.0007s
    total number of events: 24720
    total time taken by event execution: 181.0891
    per-request statistics:
         min: 0.01ms
         avg: 7.33ms
         max: 148.26ms
         approx. 95 percentile: 18.11ms

Threads fairness:
    events (avg/stddev): 24720.0000/0.00
    execution time (avg/stddev): 181.0891/0.00

3)

DONE!

[penalvch (penalvch)]

Evren Yurtesen, to keep this relevant to upstream, one would want to test the latest mainline kernel (now 4.8-rc4) as it is released.

Could you please advise?

[Evren Yurtesen (eyurtese-g)]

Amazingly the RC4 kernels booted! I tried both low latency and generic and problem is still there.

It may be unrelated but I think it is worth mentioning never the less. I have problems with keyboard skipping keypresses (Eg. forgets that I keep my finger on CTRL key, does C instead of CTRL-C etc.), touchpad palm detection not functioning (yes tried to enable it), grub breaking windows hibernation/fastboot (had to put windows boot loader to do the booting), amdgpu-pro drivers crashing, wifi rtl8723be needing driver from github. Yeah, things are barely functioning here :(

4.8.0-040800rc4-generic:

Operations performed: 15360 Read, 10240 Write, 32675 Other = 58275 Total
Read 240Mb Written 160Mb Total transferred 400Mb (1.3333Mb/sec)
   85.33 Requests/sec executed

Test execution summary:
    total time: 300.0046s
    total number of events: 25600
    total time taken by event execution: 181.6251
    per-request statistics:
         min: 0.01ms
         avg: 7.09ms
         max: 73.11ms
         approx. 95 percentile: 17.50ms

Threads fairness:
    events (avg/stddev): 25600.0000/0.00
    execution time (avg/stddev): 181.6251/0.00

4.8.0-040800rc4-lowlatency:

Operations performed: 15360 Read, 10240 Write, 32641 Other = 58241 Total
Read 240Mb Written 160Mb Total transferred 400Mb (1.3332Mb/sec)
   85.32 Requests/sec executed

Test execution summary:
    total time: 300.0363s
    total number of events: 25600
    total time taken by event execution: 181.5607
    per-request statistics:
         min: 0.01ms
         avg: 7.09ms
         max: 54.21ms
         approx. 95 percentile: 17.46ms

Threads fairness:
    events (avg/stddev): 25600.0000/0.00
    execution time (avg/stddev): 181.5607/0.00

[Evren Yurtesen (eyurtese-g)]

In addition, I tried 14.04.4 with 4.2.0-27-generic, result is same.,.

Operations performed: 15420 Read, 10280 Write, 32818 Other = 58518 Total
Read 240.94Mb Written 160.62Mb Total transferred 401.56Mb (1.3385Mb/sec)
   85.67 Requests/sec executed

Test execution summary:
    total time: 300.0019s
    total number of events: 25700
    total time taken by event execution: 181.3283
    per-request statistics:
         min: 0.01ms
         avg: 7.06ms
         max: 117.32ms
         approx. 95 percentile: 17.33ms

Threads fairness:
    events (avg/stddev): 25700.0000/0.00
    execution time (avg/stddev): 181.3283/0.00

[penalvch (penalvch)]

Evren Yurtesen, the issue you are reporting is an upstream one. Could you please report this problem following the instructions verbatim at https://wiki.ubuntu.com/Bugs/Upstream/kernel to the appropriate mailing list (TO Hans de Goede, and Tejun Heo CC linux-ide)?

Please provide a direct URL to your post to the mailing list when it becomes available so that it may be tracked.

Thank you for your help.

[Evren Yurtesen (eyurtese-g)]

Here is the link to the mailing list archive:
http://www.spinics.net/lists/linux-ide/msg53207.html

[Srijan Kumar Sharma (raja-srijan)]

Hi,

My Model is 15-ab125ax

I Would like to report that I have same problem as OP (Disk speed, keyboard, AMDGPU-Pro) and problem with wifi. Ubuntu 16.10 does not boot ("AMD-Vi: Completion loop timed out" followed by kernel panic). I had a lot of phone and email conversations with HP support and it does seems like my Laptop is locked for WINDOWS 10 ONLY. Currently I'm using Ubuntu 16.04 with self compiled 4.7 kernel from AMD staging which allows me to install AMDGPU. I'll try 4.8-rc4.

Hope this is helpful to someone.

Regards

[penalvch (penalvch)]

Srijan Kumar Sharma, this report is only scoped to disk performance with AMD Carrizo FCH Controller (not keyboard, wifi, booting, or any other problem).

Hence, it will help immensely if you filed a new report with the Ubuntu repository kernel (not mainline/upstream) via a terminal:
ubuntu-bug linux

Please feel free to subscribe me to it.

For more on why this is helpful, please see https://wiki.ubuntu.com/ReportingBugs.

[Evren Yurtesen (eyurtese-g)]

Srijan Kumar Sharma, to be honest I thought about making separate bug reports but I think maybe Linux kernel is sponsored by Intel or something. Because all my bug reports related to machines with AMD processors got silently ignored after making me collect information after information. So I decided that I would wait and see if this specific bug report would get fixed or not before doing more bug reports. So far 3 months and nothing. Usually what happens is that after about a year the ticket will be updated and say that the hardware is obsolete, or this version of linux kernel is old so I need to re-collect all the information again..