Dr. Hurt

Here's a quick hack for z value. If the driver detects any motion it'll dispatch a fake z=60 (well, hopefully that is ) V7alpha8.zip

I'm gonna assume the pointer is not moving because the driver is reporting a negative z (pressure) value. V7 does not report pressure values in its packets so the driver is going crazy and reporting -ve values. OS X needs a z value above 30 to start translating coordinates into movements. I'll find a way to report fake pressure when movement is detected. BTW, does your touchpad even have buttons? ---------------------- In other news, Linux 4.2 has been released and has some nice fixes for bitmap processing (multi-finger). Will add those to the next R3RC.

R3RC1: This is for ALPS V1-V5 only Changes: - Zero compile warnings now after the clean up - Improved 2-finger detection - Slowed down trackstick speed by 2x (as suggested by you). - Improved single finger scroll acceleration - Separated Rushmore profile, ported updates from Linux 4.1 (hope it doesn't backfire). - Updated interrupt handling (more strict filtering should make things smoother). - Cleaned up logging Watch for pointer smoothness, 2-finger scroll/click, pointer jump with trackstick. R3RC1.zip

ALPS V7 is just way too different in the way it works compared to V1-V5. Developing a driver for it without the hardware is a bit challenging. If this build doesn't at least show some promise, I'll publish all the ported V7 code to my github for future devs with the hardware to takeover. I have done all the hard work. All it needs is just the finishing touches. alpha7: - integrated the packet validity check into the interrupt handler (as opposed to a separate function). V7alpha7.zip

V7alpha6: - Added V7 interrupt handler. - Saves multitouch data to modelData for future use. (probably needs improvement) - Updated/ported lots of stuff from Linux 4.1 Again, this is the debug version. Use only with V7 hw. ================= In other news: Release 3 is progressing, RC coming soon. - Zero compile warnings now after the clean up - Improved 2-finger detection - Slowed down trackstick speed by 2x - Improved single finger scroll acceleration - Separated Rushmore profile, ported updates from Linux 4.1 - Updated interrupt handling - Cleaned up logging V7alpha6.zip

Those coordinates are definitely incorrect. I'm starting to think that the V7 code in the Linux driver is incomplete. I have a couple of patches from people who worked on the driver which seem to complete the missing pieces. I'm currently having some real life issues so I probably won't be able to provide an update soon. Maybe in a week, hopefully sooner.

I'm too tired to think right now. V7 doesn't have a single touch mode so I'm dispatching x1, y1 of multitouch instead. This is the debug version this time. Check console for x, y coordinates, number of fingers. Anything other than 0 means its working internally. V7alpha5.zip

V7alpha4 -ported most of the missing code. V7alpha4.zip

The touchpad is now properly initialized and is in command mode which is why the regular mouse driver didn't load. That's a good thing. The ALPS driver is now taking over. Up next: work on processing packets and dispatching them to OS X Remember that this driver is still in alpha and is missing a lot of code. Don't expect it to work fully just yet.

V7alpha3: - Changed the way init is handled - Fix the log messages get mixed Hope this version can at least initialize the touchpad correctly. V7alpha3.zip

@Jake Lo Here's V7alpha2 The only change is that I have added logging to every step of the init code to see where its fails. Search the log for "ALPS". Some Qs: Does 2-finger horizontal scroll work? 2-finger tap for right click? 3/4 finger gestures? Drag? Does changing settings (tracking speed, tapping, draging,.. etc) via the trackpad settings actually work? Or have no effect? I'm asking because I wanna know whether the driver is partially working or all that just the PS2Mouse driver at work. V7alpha2.zip

I'll assume you're talking about the V7alpha. Does initialization also fail? (search for "initialization" in the log in the console app) Does 2-finger side scroll work? (in launchpad for example). 2-finger tap for right click?

Works as touchpad or as normal ps2 mouse? I'd recommend you delete the mouse kext from the plugins folder of kext and reinstall. Is 2-finger tap or scroll working? Touchpad settings in system preferences work?

V7 alpha is here: WARNING: This is extremely likely to crash. It's incomplete, contains very messy code, and ugly work arounds. Use only with V7 systems for now. The most I'm hoping for at this stage is that the system boots without a kernel panic :') Search the log (console.app) for: - ALPS touchpad driver loaded - Found V7 - initialization complete Good Luck V7alpha1.zip

First post updated: R2, changelog, pushed commits to github (messy, I know) -------- As for trackstick, it's currently being divided by 8. How much slower should I make it? For example, I could divide by 12 instead and make is 1.5x slower (compared to R2), or divide by 16 and make it 2x slower. Did 2-finger scroll work better for you with the old driver? My code is based on Linux 4.1. I could downgrade to Linux 3.9 (like bpedman's version). Overall, what do you think of this driver? How does it compare to the old one? What's gotten better, what's gotten worse? --------------------- Data decay is not coming out as nice as I hoped, and I'm having trouble understanding/tweaking it due to lack of documentation. In other news, I have ported a huge chunk of V7 code so I might be able to post an alpha version sometime soon. V7 doesn't share any code with older versions which is making the job a nightmare. I'm resorting to very dirty hacks (for now), hoping to at least get it to load. We'll see.

R2RC2 Changes: - Re-did some of the changes I reverted for test1. Hopefully that won't break it again!! - Fixed the plist defaults - Small cleanup I'll postpone the data decay to R3 since its a bit of work and needs some testing. If all goes well, final changes will be committed and first post will be updated. R2RC2.zip

Thats good news. We're making progress I'll post R2RC2 tonight: - revert the plist changes. - add data decay to make pointer smoother.

@Jake Lo: Try editing the plist of R2RC with fingerz=30, scrollresolution & resolution=400, divisorx, divisory=2 Those are pretty much the only changes I remember after test.zip (the first one). @wern: Keep in mind that inertia scrolling only works with 2-finger scrolling. And yes, the pointer movement is caused by the Linux driver (even on Linux). Things are better on OS X though due to rehabman's simpleaverage smoothing.

I have a V3 id E7=0x73 0x02 0x64, EC=0x88 0x07 0x9d and 2-finger scroll is working perfectly. Maybe wern (same id) can can give us feedback about that. I haven't done much changes since test.zip!!! I find it really strange that the two models with the same id are behaving differently. They are using the exact same code. - Do you have any other touchpad/mouse kexts installed? Or utilities that could possibly interfere such as smoothmouse? - Use this to install kexts: http://mac.softpedia.com/get/System-Utilities/iKext.shtml(my work so I trust that it removes old kexts and doesn't mess up caches like other utils) - If it doesn't work on first reboot, connect a USB mouse and reboot again. It did that for me a couple of times while testing. Before I do any changes: Could you please try "test1" again to confirm that everything still works, then "test2" to try to narrow down where things went wrong. Only test the Rushmore systems (EC=0x88 0x08) For time's sake, don't test the systems with V7 or V3 id E7=0x73 0x02 0x64, EC=0x88 0x07 0x9d (we already know how those work). PS. I'll add a trackstick divisor to the plist later so that you can have control over how fast you want it to be.

Release Two RC: - Added divisor to trackstick scroll speed. If this works well enough for Jake Lo, I'll upload it to the first post as R2 and post a changelog with all changes since R1. R2RC1.zip

Just to confirm: Touchpad and Trackstick are both working with normal speed. Trackstick scrolling is too fast. I thought Trackstick scrolling was broken so I didn't apply the divisor to it. That should be an easy fix. I'll wait for Jake Lo confirmation, then apply divisor to Trackstick scrolling, commit all changes to github and make a "Release Two" tomorrow.

I'M OPTIMISTIC @Jake Lo: Ignore test2, try this instead. @wern: The touchpad probably died because It was being divided by 64 (8x8). My modifications have absolutely nothing to do with the battery status. It could be due to the updated base of the driver (rehabman's work). My battery status is working perfectly using ACPIBatteryManager.kext test3 changes: - Trackstick speed division is now completely separate from touchpad. - side scroll works perfectly now. test3.zip

The double click and right click issue is either a pressure threshold issue, or a bitmap issue. If it is a pressure issue, It should be fixed in this build. Anyway, here's another build: - Lowered resolution to 100dpi, 4 counts/mm (the default value on Linux). - Applied a universal divisor of 8 for both the trackstick and touchpad. (hope that doesn't break things again). - Raised the pressure threshold back to Linux default of 64 (regression from last build!!!). - Enabled side scroll (not yet working, but give it a try anyway). If you have the source of Exx30 and Exx20 drivers, it'll really help. UPDATE: Got side scrolling working. Woohoo!!!!! It's outrageously fast but it works. Gotta go to work now. Will fix speed and post test3 later tonight. test2.zip

I'll do some more work on it tomorrow. Which of these issues was present in bpedman's version? Or are they all new (with my version only)?

It is possible, but I need to determine the correct right and bottom edge coordinates. It could be different from model to model. Will look into it later once we actually get the basics working