Re: PATCH: Initial support for Lenovo T430
Thank you for reply me.
It is acceptable to get the EC version directly from the EC register
space,
> as long as you do DMI matches beforehand. That might be better than a
> date-based quirk list.
>
> You say the string does exist. Where is it? WMI?
The string is visible if you use dmidecode (the string is Firmware
Version: 1.7 i suppose but i'm not sure that is the EC version because
on Lenovo documentation is 1.07)
>This is an incompatible change in the userspace ABI, which is
absolutely
> unacceptable except in the most extreme cases. It has to be fixed in
some
> other way.
>
> How does the temperature sensors behave in the newer Lenovo ECs?
Ok I expected it :) In T430 there is only one temperature sensor (as
reported in DSDT with name SB.PCI0.LPC.EC.TMP0) that correspond to 0x78
address in EC ram. The others (from 0x79 to 0x81) dont exist and report
0 so i thought that i removed theyr files from sysfs. Big error,
right? :)
> I discovered some interesting facts about mute LED, GPS, _OSI string,
> > and accelerometer. So I continue to work to implement these
features.Care to document them before you attempt to publish any code
changes?
> Depending on how you discovered them [e.g. you decompiled or
disassembled
> windows drivers], we will need a clean-room implementation aproach,
where
> you document how to do it, and someone else writes the code.
>
Oh no! I discovered them only with disassembled DSDT code, looking the
behavior under windows (i don't decompile drivers or else..), with some
informations found on the web.
I discovered that if i pass OSI string "Linux" at boot the machine it
deactivate the automatic control of mute speaker led when the button is
pressed, i discovered how to turn on it via software. It's present maybe
in DSDT a strange debug information about GPS device presents in the
machine (maybe its possible to interact with it through ACPI?) I read
documentation of HDAPS kernel module i tried to implement accelerometer
directly in thinkpad_acpi but i discovered that the EC commands valid
for old Thinkpad dont function anymore...
>It has some rough edges for sure.
No problem i must learn! :) I will follow your suggestions and i will
read the documentation.
Another question: how can i read directly in user mode the NVRAM? (i
simply read the /dev/nvram file? or /dev/port or /dev/mem?).
Sorry but I'm starting to get interested in this field seriously only
now
Thank you
Il giorno gio, 30/08/2012 alle 22.57 -0300, Henrique de Moraes Holschuh
ha scritto:
> On Fri, 31 Aug 2012, Andrea Oliveri wrote:
> > 1) Defines a new field in struct tpacpi_quirk: bios_date, an
unsigned
> > int field that contain the bios release date of the machine in this
> > format MMDDYYYY (e.s 04/23/1988 -> 04231988). I introduced it
because in
> > new bioses there isn't the EC version string (precisely this string
> > exists but it's not attainable from the kernel with the DMI syscall
> > family, no ACPI method exist in DSDT table to extract it). I set to
0
> > the bios release date of other thinkpad in quirks list and i used
for
> > the new generation machines (like T430) the quirk type TPV_QL0
>
> It is acceptable to get the EC version directly from the EC register
space,
> as long as you do DMI matches beforehand. That might be better than a
> date-based quirk list.
>
> You say the string does exist. Where is it? WMI?
>
> > 3) I change some parts of the thermal subdriver, so it's possibile
for
> > new generation thinkpad without EC version string to export in sysfs
the
> > temperature informations also i set the subdriver to export in sysfs
> > only the temperature sensors present in the machine (no an infinite
list
> > of tempX_input with 0 value).
>
> This is an incompatible change in the userspace ABI, which is
absolutely
> unacceptable except in the most extreme cases. It has to be fixed in
some
> other way.
>
> How does the temperature sensors behave in the newer Lenovo ECs?
>
> > I post here the patch file http://pastebin.com/ZDxZyw1P
> >
> > I discovered some interesting facts about mute LED, GPS, _OSI
string,
> > and accelerometer. So I continue to work to implement these
features.
>
> Care to document them before you attempt to publish any code changes?
> Depending on how you discovered them [e.g. you decompiled or
disassembled
> windows drivers], we will need a clean-room implementation aproach,
where
> you document how to do it, and someone else writes the code.
>
> > Be lenient, its my first patch at kernel driver that i make =)
>
> It has some rough edges for sure. If you want to improve on that
area, read
> Documentation/SubmittingPatches in the kernel source tree...
>
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
RSS Feed