gmane.linux.drivers.hdaps.devel

Re: Leonovo Thinpad W510 - Axes totally messed up

Bjoern Olausson — 2011-04-14T11:04:27

Yes

Initial mode latch is 0x05

Thanks for taking a look.

Cheers,
Bjoern

------------------------------------------------------------------------------
Benefiting from Server Virtualization: Beyond Initial Workload 
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve 
application availability and disaster protection. Learn more about boosting 
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev

Re: Leonovo Thinpad W510 - Axes totally messed up

Henrique de Moraes Holschuh — 2011-04-10T11:06:32

No, you have to rotate them.

What is the "latch value" reported by hdaps?

Leonovo Thinpad W510 - Axes totally messed up

Bjoern Olausson — 2011-04-09T23:09:59

Hi,

I just wanted to ask if it s possible to support Lenovo ThinkPad W510 properly.

I added the following to hdaps.c and it turned out to work - in some way:

--- hdaps.c.org 2011-04-09 23:31:57.029871784 +0200
+++ hdaps.c     2011-04-09 23:57:44.059274533 +0200
< at >< at > -526,6 +526,7 < at >< at > static struct dmi_system_id __initdata h
        HDAPS_DMI_MATCH_INVERT("LENOVO", "ThinkPad T60", HDAPS_BOTH_AXES),
        HDAPS_DMI_MATCH_INVERT("LENOVO", "ThinkPad T61p", HDAPS_BOTH_AXES),
        HDAPS_DMI_MATCH_INVERT("LENOVO", "ThinkPad T61", HDAPS_BOTH_AXES),
+       HDAPS_DMI_MATCH_NORMAL("LENOVO", "ThinkPad W510"),
        HDAPS_DMI_MATCH_NORMAL("IBM", "ThinkPad X40"),
        HDAPS_DMI_MATCH_INVERT("IBM", "ThinkPad X41", HDAPS_Y_AXIS),
        HDAPS_DMI_MATCH_INVERT("LENOVO", "ThinkPad X60", HDAPS_BOTH_AXES),

The problem is that the directions are completely of. I tested it with
hdaps_gl and got the following:

Tilt to the left: hdaps_gl --> backward.
Tilt to the right: hdaps_gl --> front.
Tilt to the back: hdaps_gl --> right.
Tilt to the front: hdaps_gl --> left.

Since I can't code in C I failed to fix it on my own :-( and inverting
axis does not resolve this problem.
I guess the directions are not essential for HDAPS since it works
flawless, except this little cosmetic problem mentioned above.

Kind regards,
Bjoern

------------------------------------------------------------------------------
Xperia(TM) PLAY
It's a major breakthrough. An authentic gaming
smartphone on the nation's most reliable network.
And it wants your games.
http://p.sf.net/sfu/verizon-sfdev

CPU usage

Tomasz Torcz — 2010-12-22T13:49:39

Hello list,

  Let me introduce myself:  I'm packaging hdapsd in Fedora.
Recently I've received bugreport about hdapsd constantly using
about 2% CPU on older T60 laptop.  Powertop shows about 50 wakeups
per second which looks normal.  Is this CPU consumption per-design
or should I ask user for ltrace/strace dump of hdapsd activity?

  (In Fedora we do not ship latest tp_smapi, because it's not
in mainstream kernel.  But I think T60 should be supported by
upstream Linux).

Re: Lenovo S10-3t hdaps research

Henrique de Moraes Holschuh — 2010-11-28T00:22:03

Check if it is not a time series, where you have a window that shows the
last two measurements.

Check if the values going to the "command port" are the same ones used by
the out-of-tree HDAPS driver...

Check if some of the data you see in that window is not really SBS battery
pack data (go look at tp-smapi if you want to know why).

Check if the DSDT knows about these ports (probably not).

Find which EC is used on the S10, and get the datasheets.  This likely will
tell you the protocol to use to talk to the EC.


The in-kernel driver is extremely crippled.

Good luck.  I advise you to attempt to do a proper clean-room
implementation, if you want your code to go to mainline some day.

Lenovo S10-3t hdaps research

Javier — 2010-11-27T18:31:41

So the other day I received a S10-3t as a present. Kept the stock Windows 
7 and decided to fire up WinDbg and see why the Linux HDAPS driver wasn't 
working when the device clearly has an accelerometer.

Don't read the rest of the post if your local legislation forbids you 
from doing so.

Seemingly, HDAPS on the IdeaPad is an entirely different beast. 

There seems to be a memory accessible by the IO ports 0x702 (address) and 
0x703 (data), and some device at ports 0x6C (seems like a command port) 
and 0x68 (seems like a port where error codes are returned, also reading 
it resets "something"). The memory is around 64 bytes long (rest reads as 
zeros). Sending a 0x11 command causes values on the memory to change.

- < at >0x10 the driver writes here commands it also sends to IO port 0x6C.
- < at >0x12-0x13 this seems to be coordinate X of the accelerometer, as a 
s16. Ranges are from 400..740.
- < at >0x14-0x15 coordinate Y. Range 420..740
- < at >0x17-0x18 mirror of coordinate X?
- < at >0x19-0x21 mirror of coordinate Y?

No idea about the rest of data, some of it also changes when sending 
command 0x11. Might be temperature.

I've made a preliminary driver based of the current inkernel one and the 
above that seems to work on my S10-3t. It does the minimal amount of 
chatter that will make the device work (the sleeps are required, 
otherwise error condition "3" happens) even after a cold boot (so, no 
Windows required). Input device will give the raw values.

http://gitorious.org/iaps/iaps/blobs/master/iaps.c

Remember that this all comes from a WinDbg session so I have no idea if 
this will cripple the hardware on the longterm because of a missing step 
or whatever. Use at your own risk.


------------------------------------------------------------------------------
Increase Visibility of Your 3D Game App & Earn a Chance To Win $500!
Tap into the largest installed PC base & get more eyes on your game by
optimizing for Intel(R) Graphics Technology. Get started today with the
Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs.
http://p.sf.net/sfu/intelisp-dev2dev

Re: hdapsd on intel macbook

Stefan /*St0fF*/ Hübner — 2010-10-01T07:17:23

Hi list,

well, from what you say - and I second that it works well enough for my
X61T - it looks like our macbook-guys need to inquire Apple tech support
for what kind of readings they get.  As soon as they know, it'll be not
the hardest task to implement.  The only problem will be that I fear
Apple will give out as much information as IBM/Lenovo: none.

Another testing-scenario would probably be helpful: turn on you macbook
in a weird position: like roll it 45 degrees to the side and another 45
degrees backwards, then power it on and have a look if the readings from
turning it 90° to the side are still the same.  If that's the case, my
guess would be that the firmware already integrates the real readings
from the sensors into positions - which makes it weird stuff, as you'll
have to differentiate out the real accelerations. (Actually doing the
backward calculations from what the firmware does.)

All the best,
Stefan

Am 30.09.2010 16:48, schrieb Elias Oltmanns:


------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Elias Oltmanns — 2010-09-30T14:48:15

[...]

As it happens, I had a discussion about this with Shem a long time ago.
In my opinion, the algorithm in hdapsd is based very heavily on trial
and error and only superficially on the intuitions taught by physics. In
fact, the assumptions underlying the algorithm as suggested by the
comments in the source code are plain wrong. Since things seemed to work
well enough for me and I did not indulge in lengthy experiments to get
more clues about what the sensor readings actually represent, I did not
pursue the matter.

As far as I know, Shem's main argument against my objections to the
algorithm still stands: We really don't know the exact meaning of the
sensor readings (at least as far as IBM/Lenovo is concerned). For one
thing we have no idea about the orientation of the 2d sensor which, in
my estimation, has to be unable to detect acceleration in one particular
direction. Secondly, the firmware might do all sorts of things to the
data before passing them on to the interface. How are we to compensate
for either of these unknowns?


We have to act way before free fall. Besides, acceleration is dangerous,
whatever the cause (falling, accelerationg in a car or on the train, etc).

Regards,

Elias

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Stefan /*St0fF*/ Hübner — 2010-09-29T20:28:27

Am 29.09.2010 10:34, schrieb Evgeni Golov:

I just don't know which way your readings are, but making 2d to 3d is
pretty easy.  What do we want to detect: acceleration!

If something is falling, the acceleration is linear.  So you can use the
simple Newton-Equation:
w(t)=a/4 * t^2 + v_0 * t + w_0

w ... the way or length or even position
t ... time
a ... acceleration
v_0 ... starting speed

Now the needed changes to the equation depend very much upon if the
readings are positions, speeds or accelerations.  Now I know that
position detection needs a fixed point - which a laptop normally does
not have (except if it's gps-equipped) - the readings very likely are
not positions.  Also detecting a current speed is pretty hard.  Only
detecting acceleration works.  So if the readings are accelerations, the
whole story is as easy as calculating a square root.  The problem might
be, that there are different types of accelerations:
- movement along any axis
- rotation around any axis
-> that is what 6-way-accelerometers are for.

If you can find out the right readings for movement along x,y and z,
then calculate a = sqrt(a_x^2+a_y^2+a_z^2).  If a is around 9 or larger,
then most likely the device is falling.

I'm sorry I don't have the time to go thru the driver code - would you
just outline the algorithm used, so maybe I can help you better?


All the best,
Stefan

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Till Straumann — 2010-09-29T15:02:22

Yes . I don't see a difference between -y and otherwise (except for the
greatly increased CPU load, of course).
till< at >tillbook:~/lcls/evg$ cat /sys/devices/platform/applesmc.768/position
(9,-5,262)

till< at >tillbook:~/lcls/evg$ cat /sys/devices/platform/applesmc.768/position
(12,255,-9)
till< at >tillbook:~/lcls/evg$ cat /sys/devices/platform/applesmc.768/position
(264,2,2)

till< at >tillbook:~/lcls/evg$ cat /sys/devices/platform/applesmc.768/position
(-246,5,-4)

I'll have to delay that test for a few days - I need to get on with my 
work, sorry.


- T.


------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Evgeni Golov — 2010-09-29T08:34:27

Hmm, really looks same as mine.


Ok, 40 is exact the value I always use on my thinkpads.
Coul you try running hdapsd with -y to force the sysfs interface? Is 40
still enough?

And while at it, could you post the values of the position file at the
following situations?
- just standing on the desk
- "standing" on the display (flipped 90 degrees back)
- standing on the right side (flipped 90 degrees to the right)
- standing on the left side (flipped 90 degrees to the left)

And could you try my git from the new-interfaces branch? Is 40 fine
there too? If so, this sounds like a "oh, we are Apple, lets break our
own API in every hw-release" again :(


Hm, okay, didnt know the Z axxis isnt present there.


It does not, as the algorithm used for fall-detection is 2D and I dont
know how to rewrite it :)

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Till Straumann — 2010-09-28T17:36:55

Sure (attached).
The default value (15, IIRC) lets my type normally just fine but
energetic hacking and slight motion of the laptop results triggers
a 'park'.

I use 40 now which lets me hack and move and set down the box
carefully. If I move or set it down a bit abruptly then hdapsd triggers.

Having just repaired my filesystem with 1k bad sectors I'm getting
a bit cautious (this incident initiated my venturing into hdapsd land)
- I wouldn't want to test a value of 200.
Hmm - the problem is that the applesmc driver doesn't feed the
3rd axis into the input/event stream anyways.
z is available from sysfs, however.
The 'position' file has the format: "(%d,%d,%d)".
Hmm - I don't notice anything peculiar (neither using input layer, nor
sysfs); I have a MacBookPro 5,2, running ubuntu 10.04, 2.6.32-24-generic
(x86_64).

I figured it's still better to watch X and Y but it would of course be
better to use all three coordinates. (But it seems to me that hdapsd,
at least the 20090401 release doesn't use Z anyways).

Will do

Re: hdapsd on intel macbook

Evgeni Golov — 2010-09-28T16:48:51

Could I still see it? See below for the why :)


It can. As I do not have an Intel-MacBook, I can't test myself.
I got a reply in my blog that the guy needed a -s value of about 200 to
use hdapsd normally. What is yours?
And: how did you handle the fact the MacBook has 3 axxis in the sensor?
you move the macbook. (<< thats the why)


Currently, the only "Protection" is the setting of mlock, so the process
won't be swapped etc. If you have more, gimme :)

Regards
Evgeni

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

Re: hdapsd on intel macbook

Till Straumann — 2010-09-28T16:41:59

I see - great; that looks just like what I did.
It is a bit sensitive but can't that be tuned with -s ?
The only other contribution (since you support APPLESMC already)
I could suggest is giving the hdapsd process real-time scheduling
and a high priority - otherwise: how do you ensure hdapsd gets
to scheduled fast enough for the disk to be parked before it hits
the floor?

Cheers

Re: hdapsd on intel macbook

Evgeni Golov — 2010-09-28T16:23:58

Hi Till,

yes hdapsd is still maintained (by me) and we do have some basic macbook
support in git [1], which is reported to be too sensitive :(.

I'd love to see your patches and merge them into hdapsd.

Regards
Evgeni

[1] http://github.com/evgeni/hdapsd/tree/new-interfaces

------------------------------------------------------------------------------
Start uncovering the many advantages of virtual appliances
and start using them to simplify application deployment and
accelerate your shift to cloud computing.
http://p.sf.net/sfu/novell-sfdev2dev

hdapsd on intel macbook

Till Straumann — 2010-09-28T04:37:29

Hi.

Thanks for hdapsd. Is this software still maintained?
I found that it is straightforward to use it on newer
intel x86-based macbooks.

I can provide patches if anybody is interested.

Please cc me on any replies - I'm not subscribed to
this list.

WKR

Lenovo ThinkPad T510 Needs HDAPS Inversion

Tom Gelinas — 2010-03-17T08:35:40

Howdy,

The Lenovo ThinkPad T510 requires HDAPS to be inverted at "7" to
function correctly. Thanks so much for your efforts!

-Tom Gelinas



------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev

RFT: hdapsd for Intel MacBooks and Hewlett-Packardlaptops

Evgeni Golov — 2010-03-07T18:49:24

[ Yes, this is most probably the wrong list, but who else listens to 
  hdapsd stuff out there? Feel free to forward there ;)              ]

After quite some time (last release 01-04-2009), hdapsd got a bit of 
love.
Brice Arnould has contributed some code for Hewlett-Packard laptops 
(those supported by the hp_accel module, see drivers/hwmon/hp_accel.c[1]) 
and I finally crossed all lines and wrote support for Apple MacBooks 
(the Intel ones, via applesmc module, see drivers/hwmon/applesmc.c[2]).

The HP code is interesting, as it support a hardware-logic mode, where 
hdapsd only parks the heads when told so by the HP hardware. As I do not 
own any compatible hardware, this is only tested by Brice himself.
The Apple SMC code isn't tested at all, as I don't have the hardware 
either.

That's why I want YOU to test it further (both, on HP and Apple) and 
report me bugs (mail to evgeni-8fiUuRrzOP0dnm+yROfE0A< at >public.gmane.org preferred) if you find any :)

You can get the latest source either via git from github:

git clone -b new-interfaces git://github.com/evgeni/hdapsd.git

Or from githubs tarball generator: 
http://github.com/evgeni/hdapsd/tarball/eb711f30395ac9bc682b14c22d8445b7ddf0b4a0

After you got the source, a simple

./autogen.sh
make

should produce a src/hdapsd binary, that you can test.

I can provide Debian and Ubuntu .debs if needed too.

Regards
Evgeni

[1] 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/hwmon/hp_accel.c
[2] 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=drivers/hwmon/applesmc.c

Re: Could not find a suitable interface

alkos333 — 2009-12-15T20:54:10

Never mind guys, the problem is resolved - just had to make sure both
hdaps and tp_smapi modules were running.

Could not find a suitable interface

alkos333 — 2009-12-15T20:44:32

The title says it all - when I try to start hdapsd, I get that message above.

Re: Gemalto ExpressCard Smart Card Reader under linux?

Henrique de Moraes Holschuh — 2009-11-10T16:55:47

Wouldn't it make a lot more sense to ask this on the linux-thinkpad ML?

If it is USB, it should work perfectly (ExpressCard cards can be either PCIe
or USB 2.0).