Data collection necessary to understand and fix dynamic graphics (61xx series)

Welcome to Project BaconBits.

I figured I'd bring myself and the other Linux hybrid graphics gurus into the dynamic switching mess. Our goal is to help you all understand the driver situation on Windows, find out if manual switching is possible, and eventual Linux support.

To this end, here's a short guide. To follow it, you need to boot an Ubuntu Live CD/USB session. You may need enable the Community (Universe) repository.

DSDT collection can also be accomplished on Windows with RW-Everything by saving the ACPI tables to a .rw file. I do not recommend this method since the information is not well-organized or as complete compared to the Linux method.

Please post the DSDT file you generate on the bug report as an attachment or here as a mediafire link. Post the output of all bolded commands in this thread.

Quoted below is the Linux method.

Coming Soon
fglrx driver assessment (July?)

 

If they need someone to do this for the original manual switching, I can give them that info too(perhaps for comparisons sake).

 

That would be nice.

But manually switching works, doesn't it?

As long as there's a mux, brute-forcing the GPU on is easy and can be done with ACPI calls.

 

Yea works just fine, no issues with latest BIOS and drivers. Reason I'm offering is there maybe some info for the guys there that might help them work this problem out for the Dynamic Switching system, thats all.

 

Non-dynamic (60xx series) systems could serve as a nice control to measure against.

Feel free to contribute.

 

Is this what you're looking for? Generated from RW-Everything: AcpiTbls.rw

 

It doesn't look like there's a mux. Once I get my dv6, I will put either Ubuntu or (if RedHat supports the latest laptops which I very much doubt) Scientific Linux 6 on it and then I can help you. At the rate this order placement is going, it looks like that'll be another couple of weeks though so hopefully someone else will help first.

 

I see the PEG entries, so in a sense, yes.

The Linux approach is far more standardized and targetted. Plus, we get specific information about card models, vendor strings, and PCI bus location.

 

I wish all you guys would simply bombard HP with complaints over this issue rather than spending your time looking for a solution outside of HP. HP needs to be prodded into getting this resolved ASAP. This is not a "do it yourself" issue.

 

It is a DIY issue because we might be able to kill 2 bird with one stone: Manual driver hacking and Linux support

Plus, we'll be prepared in the event that HP is slow to act or does nothing.

 

The probability of HP providing a solution for Linux users is utterly negligible. I applaud your determination in pestering HP, but in the absolute best case, you will get them to fix the switching in Windows. Linux has always been "do it yourself" territory for new hardware and the people who work on it are pretty good at getting reasonable solutions. In the worst case (if HP does nothing), somebody who knows both UNIX and Windows will have to take the Linux solution and adapt it to work in Windows. I am hoping that this is not necessary and we'll get an "official" driver or BIOS patch, but it's best to keep one's options open.

 

good point!

 

Not sure if I got all of the instructions, but I'll give it a shot sometime tonight! Might as well since my return period is about to expire anyways.

 

Come on guys, I shouldn't be the only one with a 61XX model here. I know it's a slight bother to install a Linux OS for data collection, but think about what's at stake here. You either 1) return your laptop out of frustration with HP or 2) get a solution from the community because you contributed a day of your life to get the data collection done. I know that many of you probably got a great deal on your laptop, and it would suck if you had to give it up, right?

Besides, today's a Sunday! Spend your weekend on this project and you may be rewarded

 

You don't need to install a Linux OS, just boot off of a Live USB stick (or CD/DVD).

 

Really? Thanks for the info, Athernai! Just goes to show I have no knowledge whatsoever about what I'm about to do

 

Here is a website with a tools to make Live USBs. I've used it with Ubuntu before and it works. Maybe LLStarks can provide instructions for use with the linked tools by people not familiar with Linux? There aren't that many people who have the laptops in their hands.

 

Ok, get ready for a noobie question - when I enter the command:
sudo apt-get install acpidump iasl
it gives me an error:
E: Unable to locate package acpidump
and when I try the command:
sudo yum install pmtools iasl
it give me another error:
sudo: yum: command not found

What should I do? Did I do something wrong?
I'm running Ubuntu 11.04 x64 from a Live USB stick (thanks Althernai!), by the way.

 

What you're doing is telling a package manager to install things. "sudo" just gives you administrative access. "apt-get" and "yum" are the names of the installers. Ubuntu uses apt-get which is why yum is not found. The problem appears to be that Ubuntu doesn't know where to look for the packages.

LLStarks, can you say how to get it to see the correct repository? As far as I know, the list of repositories is located in /etc/apt/sources.list , but I don't have Ubuntu in front of me and I haven't used it in a while so I don't remember where it's supposed to look.

 

Ok, here's the DSDT table! I'm not sure if it's the correct output because the commands were long, but hopefully it can be of help!

Oh and a few tips of advice for noobs like me :
1. Make sure you set up the repository settings so that you can also use community packages. I think I found the setting somewhere in the Ubuntu Software Center; look for a checkbox that says something about community repositories.
2. Don't mix up the accent grave (`) key with a single quote ('). I know it's kinda obvious but just in case.

 

Umm LLStarks, can you confirm whether or not this is what you were looking for? I might end up returning this laptop so I want to know if I did my job right.

 

Exactly what I was looking for.

Can I have the output of the following as well?

sudo dmidecode -s system-product-name

sudo dmidecode -s system-version

lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA

 

ubuntu@ubuntu:~$ sudo dmidecode -s system-product-name
HP Pavilion dv6 Notebook PC

ubuntu@ubuntu:~$ sudo dmidecode -s system-version lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA
dmidecode: invalid option -- 'v'

I have a feeling that the second code didn't execute as it should've.

 

retry them, i screwed up by combining the second and third lines

here's my whole 3 lines

MXC061
Not Specified
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03) (prog-if 00 [VGA controller])

 

Ok, here it is
ubuntu@ubuntu:~$ sudo dmidecode -s system-product-name
HP Pavilion dv6 Notebook PC

ubuntu@ubuntu:~$ sudo dmidecode -s system-version
058C110000244720001620100

ubuntu@ubuntu:~$ lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA
00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09) (prog-if 00 [VGA controller])
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:6740] (prog-if 00 [VGA controller])

 

Cool.

Very useful.

Can you you try manually switching your GPU?

Ubuntu Control Center 0.6.1 Released ~ Web Upd8: Ubuntu / Linux blog

It should only require an X restart once the card is switched.

If that doesn't work, please try installing the latest Catalyst driver (fglrx) and see if there is a switching option.

Either way, if you manage to switch the cards, I'd like the output of glxinfo and glxgears before and after.

 

I tried installing Ubuntu Control Center, but it keeps giving me an error:


I tried to go both ways but both just went back to the start (i.e. Install icon appears again). I gave the 32 bit a try but it simply refused to install. Is there some other program, perhaps, that allows me to switch graphics devices?
Another question: how do I install the latest Catalyst drivers? I tried doing it through the Additional Drivers application, but it doesn't list any "proprietary drivers".

 

from the folder you downloaded to:
sudo dpkg -i ucc061_amd64.deb

as for catalyst (try if ucc's vga_switcheroo doesn't work):
sudo apt-get install fglrx

there's also byo-switcheroo, which might work, but is more advanced and is not recommended:
https://lists.launchpad.net/hybrid-graphics-linux/msg01219.html

 

I've never messed with linux before. Would it help if I installed it and posted my results as well? I'm a 6100 dv6 owner.

 

Couldn't hurt.

There might be general hardware revisions to consider or differences between the 1GB and 2GB 6770M.

 

I'm getting this error when I try to unpack the Ubuntu Control Center:
Can't open /usr/share/applications/ubiquity-gtkui.desktop: No such file or directory at -e line 1, <> line 2568.

Also, how do I go about installing vga_switcheroo?

Sorry with all the questions, I simply don't have enough experience with a Linux operating system.

 

I just installed ubuntu onto a usb drive, and i'm having problems using the catalyst control center. It says there's a problem initializing the CCC because it can't find any ATI graphics drivers. Are there any drivers I need for the graphics cards/other stuff?

 

vga_switcheroo is already installed, it comes in the kernel

also, i'm at a loss as to why even ucc isn't working properly. that shouldn't happen, at all.

try this:

sudo apt-get install gdebi
sudo gdebi ucc061_amd64.deb (or ucc061_i386.deb if you're running 32-bit ubuntu)

brian, judging from previous lspci outputs, the card is recognized by the system and is apparently receiving power, but can't be detected for xserver radeon driver use or catalyst driver installation.

you can both try this:

sudo echo DDIS > /sys/kernel/debug/vgaswitcheroo/switch

and then do an alt+sysreq(printscreen)+k (restarts x and logs you out)

you can also try the other vga_switcheroo commands: https://help.ubuntu.com/community/HybridGraphics

 

Ok, I just reinstalled Ubuntu and the installation went as smooth as butter with your commands However, I'm not sure how to open the Ucc. It doesn't seem to appear in the applications lists...
Also, I verified that switcheroo is working, but I have no idea how to test whether or not the 6770m is being used. Is there a command for figuring out the currently used graphics device or something?

 

for ucc: alt+f2 then ucc

for checking gpu:

sudo apt-get install mesa-utils

and then compare the output of these commands before and after switching with ucc (recommended) or manually

glxinfo | grep OpenGL
glxgears

 

Thanks for the info. It looks like I can't switch the graphics devices - it defaults to the high performance gpu for some reason. In any case, here's the output from the two commands:
ubuntu@ubuntu:~$ glxinfo | grep OpenGL
OpenGL vendor string: Tungsten Graphics, Inc
OpenGL renderer string: Mesa DRI Intel(R) Sandybridge Mobile GEM 20100330 DEVELOPMENT
OpenGL version string: 2.1 Mesa 7.10.2
OpenGL shading language version string: 1.20
OpenGL extensions:
ubuntu@ubuntu:~$ glxgears
Running synchronized to the vertical refresh. The framerate should be
approximately the same as the monitor refresh rate.
302 frames in 5.0 seconds = 60.213 FPS
300 frames in 5.0 seconds = 59.841 FPS
300 frames in 5.0 seconds = 59.841 FPS
300 frames in 5.0 seconds = 59.837 FPS
XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
after 2580 requests (2580 known processed) with 0 events remaining.

I also tried out the OpenGL Performance thing in UCC, and it worked so I guess OpenGL does work? However, looking at the performance of the glxgears program, I'm not exactly sure if the 6770m or the Intel 3000 is running. What do you think? It does state in the startup page of UCC that the Video card is by ATI Technologies Inc.

 

opengl, at least in your setup is on intel. i'm still not sure whether we can overcome that.

btw, the following code may be useful in disabling the radeon card to save power until we can find a way to switch to it.

sudo apt-get install git
git clone http://<wbr>github.<wbr>com/mkottman/<wbr>acpi_call.<wbr>git
cd acpi_call
make
sudo insmod acpi_call.ko
./test_off.sh

 

This is embarrassing but it will take a day or two to attempt to help with the linux data collection. :/ Sorry

 

Same result here with the UCC .deb.

I am also able to turn the Radeon on and off with the Switcheroo module but I have no way of testing to see If it is connected

 

I managed to overcome that issue by reinstalling Ubuntu 11.04 into my USB drive and following these steps:

Be warned, though. UCC is somewhat buggy on this machine, and it'll give you errors every now and then. Also, as I posted earlier, the switching mechanism doesn't seem to function properly.

 

Hmm. I guess there's work to be done then.

Our biggest problem at the moment is visibility.

The hybrid graphics group is awesome and all, but I'm looking bigger.

A nice thread on the xorg mailing list or a sensationalist Phoronix article should bring in devs. I also need to speak with David Airlie and Alex Williamson.

Short of that, we may need to wait a few days for the Llano Bacon laptops to become available.

 

Correct me if i am wrong, but switchable graphics work on linux but not windows?

 

No, manual switching is not possible on Linux or Windows, yet. However, if a manual switching method is to be found, it'll be found on Linux because access to the hardware is more open and direct.

 

I could still use the output of acpi_call from someone.

 

The last time I tried that my Ubuntu failed to start up after rebooting. I reinstalled Ubuntu and tried it out again, though, so here's the output:

ubuntu@ubuntu:~/acpi_call$ sudo insmod acpi_call.ko
ubuntu@ubuntu:~/acpi_call$ ./test_off.sh
Trying \_SB.PCI0.P0P1.VGA._OFF: failed
Trying \_SB.PCI0.P0P2.VGA._OFF: failed
Trying \_SB_.PCI0.OVGA.ATPX: failed
Trying \_SB_.PCI0.OVGA.XTPX: failed
Trying \_SB.PCI0.P0P3.PEGP._OFF: failed
Trying \_SB.PCI0.P0P2.PEGP._OFF: failed
Trying \_SB.PCI0.P0P1.PEGP._OFF: failed
Trying \_SB.PCI0.MXR0.MXM0._OFF: failed
Trying \_SB.PCI0.PEG1.GFX0._OFF: failed
Trying \_SB.PCI0.PEG0.GFX0.DOFF: failed
Trying \_SB.PCI0.PEG1.GFX0.DOFF: failed
Trying \_SB.PCI0.PEG0.PEGP._OFF: works!

Crossing my fingers that it won't fail to reboot again...

 

well, at least we know can definitely turn the card off.

can i have the output of the following before and after running test_off?

lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA

 

Here ya go :
ubuntu@ubuntu:~$ lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA
00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09) (prog-if 00 [VGA controller])
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:6740] (prog-if 00 [VGA controller])
ubuntu@ubuntu:~$ cd acpi_call
ubuntu@ubuntu:~/acpi_call$ make
make -C /lib/modules/2.6.38-8-generic/build M=/home/ubuntu/acpi_call modules
make[1]: Entering directory `/usr/src/linux-headers-2.6.38-8-generic'
Building modules, stage 2.
MODPOST 1 modules
make[1]: Leaving directory `/usr/src/linux-headers-2.6.38-8-generic'
ubuntu@ubuntu:~/acpi_call$ sudo insmod acpi_call.ko
ubuntu@ubuntu:~/acpi_call$ ./test_off.sh
Trying \_SB.PCI0.P0P1.VGA._OFF: failed
Trying \_SB.PCI0.P0P2.VGA._OFF: failed
Trying \_SB_.PCI0.OVGA.ATPX: failed
Trying \_SB_.PCI0.OVGA.XTPX: failed
Trying \_SB.PCI0.P0P3.PEGP._OFF: failed
Trying \_SB.PCI0.P0P2.PEGP._OFF: failed
Trying \_SB.PCI0.P0P1.PEGP._OFF: failed
Trying \_SB.PCI0.MXR0.MXM0._OFF: failed
Trying \_SB.PCI0.PEG1.GFX0._OFF: failed
Trying \_SB.PCI0.PEG0.GFX0.DOFF: failed
Trying \_SB.PCI0.PEG1.GFX0.DOFF: failed
Trying \_SB.PCI0.PEG0.PEGP._OFF: works!
ubuntu@ubuntu:~/acpi_call$ lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA
00:02.0 VGA compatible controller [0300]: Intel Corporation 2nd Generation Core Processor Family Integrated Graphics Controller [8086:0126] (rev 09) (prog-if 00 [VGA controller])
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:6740] (rev ff) (prog-if ff)

 

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Device [1002:6740] (prog-if ff [VGA controller])

That's some great data and it does indeed confirm that the card can be turned off completely to save power.

 

I'm all lost with what the data means, but good to hear that the off switch worked

Unfortunately, I'm going to have to request an RMA this Friday because that is when my return period will end. I will have five days after the RMA request to send it back, so I'll be able to help with the data collection until June 29. I'm sorry for the inconvenience, but since I plan to use whatever laptop I end up with for at least four years into the future, I need to be 100% sure that I won't encounter any fatal problems down the road.

If we still have some data to collect after the 29th, someone else (or elses?) will need to provide them. I know I said this before, but everyone, please participate in the data collection! LLStarks is spending a generous portion of his/her time to fix a laptop he/she doesn't even own, so it's only natural for us to assist LLStarks as best we can

 

Yup. I'm passionate about this because I was on the verge of buying a dv6-60xx and then the 61xx came out.

I'm also gonna need help for the Llano laptops.

Data collection principles should be the same, but the results may present unique challenges that differ from current data collected.