If X1C5 use alps, my patch will not be used. And this patch had been
verified on laptops with alps/elan sticks.

And on your laptop psmouse module already tried alps protocol, it failed
and fallback to PS/2.

My commit is to enable scroll mode with middle button and stick and
other trackpoint features.
Yes, like you said maybe you don't use these features, but other people
like to use.

I suggest you to use Ubuntu kernel that is built from mainline kernel
just for your convenience that you won't need to rebuild the kernel with
RMI4_SMB config enabled. *NOT* for you to always use.
Indeed I didn't know the evbug is enabled. Even for now I don't think
this kernel with evbug will hack or harm you system even I am not a
security guy.

Regards,
Aaron

On 12/30/2017 10:40 PM, Sebastian Schmidt wrote:

On Sat, Dec 30, 2017 at 10:17:55PM +0800, Aaron Ma wrote:

quoted

Sorry, I don't know anything about the firmware software.
So you can NOT expect me to fix the firmware issue.

You changed trackpoint.c in ec667683c532c93fb41e100e5d61a518971060e2 to,
according to the commit message, “support version 2 and 3”. Even though
I don’t understand the change, because according to the comment next to
TP_MAGIC_IDENT (and the code in trackpoint.c), the version is reported
in param[1], not in param[0]. Also it’s called “MAGIC_IDENT” and not
“SUPPORTED_FIRMWARE_VERSIONS”.

quoted

I have helped answer all the question I can.
I don't know why you blame me like this.

Because reverting your commit fixes the issue for me. In fact, I was
just starting to reverse engineer the differences between firmware
versions 3 and 4, when I noticed a file called
“Alps_Pointing-Device_Updater_amd64_1_4.exe”, and all the other binaries
also saying only “ALPS” and not “Synaptics”, in the
trackpoint_fw_updater_1.0.0.9.zip update package (for model 20HR). Are
you actually certain that Gen5 X1s do always contain a Synaptics
Trackpoint and not some models an ALPS one? Does changing the speed on
your FW version 0x03 or 0x02 work at all?

I’m suspecting that by changing the TP_MAGIC_IDENT to supposedly newer
firmware versions you just made that driver wrongly recognize an ALPS
trackpoint as whatever trackpoint.c is for.

And I am, in fact, upset, since you don’t appear to be supporting the
code you have written, even though it causes regressions. Then you ask
me to install a kernel that includes a keylogger without any warning
whatsoever and now “yeah, sysfs is barely used, just use GNOME”. Who
else should I blame, please?

Thanks,
Sebastian