Lenovo ThinkPad T420 with UEFI Only

TimeWalker75a · September 26, 2013

Guys, mind testing this please?

https://www.dropbox.com/s/mgz6jvcmx8eq6v5/T420-4236AT9.zip

Please remove your AppleACPIPlatform roll back (which comes from 10.6.2, are you kidding me?) and install the one enclosed in the archive, it's a genuine one from 10.8.5. I've patched your EC RAM registers to support new ACPI interpreter and thus you no longer require the rollback for battery to work properly. This will make the test in regards to shutdown more cleaner!

I basically split your registers into 8 bit ones down from 16, 32 and 128. Then rejoined them by appropriate methods so that the integrity would remain. Only those registers that are actually used in ACPI were changed:

                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBR1,   8, 
                        SBR2,   8, 
                        SBF1,   8, 
                        SBF2,   8, 
                        SBAE,   16, 
                        SBRS,   16, 
                        SBA1,   8, 
                        SBA2,   8, 
                        SBV1,   8, 
                        SBV2,   8, 
                        SBAF,   16, 
                        SBBS,   16
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBB1,   8, 
                        SBB2,   8, 
                        SBMD,   16, 
                        SBCC,   16
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBD1,   8, 
                        SBD2,   8, 
                        SBDA,   8, 
                        SBDB,   8, 
                        SBOM,   16, 
                        SBSI,   16, 
                        SBDT,   16, 
                        SBS1,   8, 
                        SBS2,   8
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBC1,   8, 
                        SBC2,   8, 
                        SBC3,   8, 
                        SBC4,   8
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBM1,   32, 
                        SBM2,   32, 
                        SBM3,   32, 
                        SBM4,   32
                    }
 
 

                    Method (B1B2, 2, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x08), Arg0, Local0)
                        Return (Local0)
                    }
 
                    Method (B1B4, 4, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x08), Arg0, Local0)
                        Or (ShiftLeft (Arg2, 0x10), Local0, Local0)
                        Or (ShiftLeft (Arg3, 0x18), Local0, Local0)
                        Return (Local0)
                    }
 
                    Method (L1L4, 4, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x20), Arg0, Local0)
                        Or (ShiftLeft (Arg2, 0x40), Local0, Local0)
                        Or (ShiftLeft (Arg3, 0x60), Local0, Local0)
                        Return (Local0)
                    }

tluck · September 26, 2013

@TimeWalker75a

well - my latest bundle posted had us using ErmaC's patch for AppleACPIPlatform (v1.7) (10.8.4 i believe?) + Rehabmans battery stuff.

i switched over this newer version ACPI when i found that rehabmans AppleSmartBatteryManager was found to work with DSDT unchanged.

previously yes was using a very old snow leopard 1.3.5 ACPIPlatform since VoodooBattery required it.

but sure, i will implement/test your package. hang on.

Tetonne · September 26, 2013

Guys, mind testing this please?

https://www.dropbox.com/s/mgz6jvcmx8eq6v5/T420-4236AT9.zip

Please remove your AppleACPIPlatform roll back (which comes from 10.6.2, are you kidding me?) and install the one enclosed in the archive, it's a genuine one from 10.8.5. I've patched your EC RAM registers to support new ACPI interpreter and thus you no longer require the rollback for battery to work properly. This will make the test in regards to shutdown more cleaner!

I basically split your registers into 8 bit ones down from 16, 32 and 128. Then rejoined them by appropriate methods so that the integrity would remain. Only those registers that are actually used in ACPI were changed:

                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBR1,   8, 
                        SBR2,   8, 
                        SBF1,   8, 
                        SBF2,   8, 
                        SBAE,   16, 
                        SBRS,   16, 
                        SBA1,   8, 
                        SBA2,   8, 
                        SBV1,   8, 
                        SBV2,   8, 
                        SBAF,   16, 
                        SBBS,   16
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBB1,   8, 
                        SBB2,   8, 
                        SBMD,   16, 
                        SBCC,   16
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBD1,   8, 
                        SBD2,   8, 
                        SBDA,   8, 
                        SBDB,   8, 
                        SBOM,   16, 
                        SBSI,   16, 
                        SBDT,   16, 
                        SBS1,   8, 
                        SBS2,   8
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBC1,   8, 
                        SBC2,   8, 
                        SBC3,   8, 
                        SBC4,   8
                    }
 
                    Field (ECOR, ByteAcc, NoLock, Preserve)
                    {
                                Offset (0xA0), 
                        SBM1,   32, 
                        SBM2,   32, 
                        SBM3,   32, 
                        SBM4,   32
                    }
 
 

                    Method (B1B2, 2, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x08), Arg0, Local0)
                        Return (Local0)
                    }
 
                    Method (B1B4, 4, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x08), Arg0, Local0)
                        Or (ShiftLeft (Arg2, 0x10), Local0, Local0)
                        Or (ShiftLeft (Arg3, 0x18), Local0, Local0)
                        Return (Local0)
                    }
 
                    Method (L1L4, 4, NotSerialized)
                    {
                        Or (ShiftLeft (Arg1, 0x20), Arg0, Local0)
                        Or (ShiftLeft (Arg2, 0x40), Local0, Local0)
                        Or (ShiftLeft (Arg3, 0x60), Local0, Local0)
                        Return (Local0)
                    }

Thanks à lot

Report : Boot error smcreadkeyaction errorr

Fall back error to old smc :-(

TimeWalker75a · September 26, 2013

Thanks à lot

Report : Boot error smcreadkeyaction errorr

Fall back error to old smc :-(

Has nothing to do with what you were testing. If you want this error gone patch AppleSMC binary

4885c075 10488d3d 7d50 to 4885c0eb 6e488d3d 7d50, this will make it fall back to old mode right away without trying the new one. This error is even present on genuine Apple hardware.

TimeWalker75a · September 26, 2013

Original battery kext was never supposed to work on hackintos. Disable extended bif and extra bif flags inside info.plistt of RehabMan's kext and it should work, I'm using it myself.

TimeWalker75a · September 27, 2013

I might have messed up the registers, will take a look.

TimeWalker75a · September 27, 2013

I actually fail to see what may have gone wrong. Could you post IOReg dump from Registry Explorer with AC cord connected?

TimeWalker75a · September 27, 2013

It seems _BIF is not properly parsed. The easiest way to debug would be to compile the debug version of the kext.

Also I still see you haven't disabled the "UseExtraBatteryInformation" key .. your laptop doesn't have a register to store cycle count as far as I can see, so approximation has to happen..

Please install this: http://puu.sh/4BttQ.zip

And see Console for errors..

TimeWalker75a · September 27, 2013

Yeah, pretty much _BIF method fails to be evaluated.. gonna look into that.

 AppleSmartBatteryManager::getBatteryBIF called
 AppleSmartBatteryManager::getBatteryBIF: validateObject return 0x0
 AppleSmartBatteryManager::getBatteryBIF: evaluateObject error 0xe00002bc

TimeWalker75a · September 27, 2013

Ok, i've simplified the thing for testing a little bit.. so lets see if it works: http://puu.sh/4BvBq.zip

                    Method (GBIF, 3, NotSerialized)
                    {
                        Acquire (BATM, 0xFFFF)
                        If (Arg2)
                        {
                            Store (B1B2 (SBB1, SBB2), Local7)
                            ShiftRight (Local7, 0x0F, Local7)
                            XOr (Local7, One, Index (Arg1, Zero))                 // Power Units 
                            Store (B1B2 (SBD1, SBD2), Index (Arg1, One))          // Design Capacity
                            Store (B1B2 (SBF1, SBF2), Index (Arg1, 0x02))         // Last Full Charge Capacity
                            Store (B1B2 (SBDA, SBDB), Index (Arg1, 0x04))         // Design Voltage
                            Divide (B1B2 (SBF1, SBF2), 0x14, Local2, Index (Arg1, 0x05)) // Design Capacity of Warning
                            Divide (0x00030D40, B1B2 (SBDA, SBDB), Local2, Index (Arg1, 0x06)) // Design Capacity of Low
                            Store (B1B2 (SBD1, SBD2), Index (Arg1, 0x09))         // Model Number
                            Store (B1B2 (SBS1, SBS2), Local0)
                            Name (SERN, Buffer (0x06)
                            {
                                "     "
                            })
                            Store (0x04, Local2)
                            While (Local0)
                            {
                                Divide (Local0, 0x0A, Local1, Local0)
                                Add (Local1, 0x30, Index (SERN, Local2))
                                Decrement (Local2)
                            }
 
                            Store (SERN, Index (Arg1, 0x0A))                      // Serial Number
                            Store (B1B4 (SBC1, SBC2, SBC3, SBC4), Index (Arg1, 0x0B)) // Battery Type
                            Store (L1L4 (SBM1, SBM2, SBM3, SBM4), Index (Arg1, 0x0C)) // OEM Information
                        }
                        Else
                        {
                            Store (Ones, Index (Arg1, One))
                            Store (Zero, Index (Arg1, 0x05))
                            Store (Zero, Index (Arg1, 0x06))
                            Store (Ones, Index (Arg1, 0x02))
                        }
 
                        Release (BATM)
                        Return (Arg1)
                    }

Edit:

ok I see the issue ...

Store (SBDN, Index (Arg1, 0x09))

I forgot to split the register and mistakenly used design capacity register there!! Gonna fix that now!

Edit2:

done.. please give this a test - > http://puu.sh/4BvRa.zip

TimeWalker75a · September 27, 2013

What a stubborn piece of code.. let's try the simplified one, with model number registers properly coded in:

http://puu.sh/4BxU3.zip

If it fails to evaluate _BIF again we are just going brutal for sakes of testing - not using any registers what so ever, just defining values based on the screenshots you had shown. It's only for the sakes of testing whether the issue is inside the algo in general or it's register fault (my fault, hehe).

http://puu.sh/4Byi2.zip

                        If (Arg2)
                        {
                            Store (0x00, Index (Arg1, 0x00))
                            Store (0xb9a0, Index (Arg1, 0x01))
                            Store (0xa474, Index (Arg1, 0x02))
                            Store (0x2a30, Index (Arg1, 0x04))
                            Store (0x839, Index (Arg1, 0x05))
                            Store (0xc8, Index (Arg1, 0x06))
                            Store ("45N1001", Index (Arg1, 0x09))
                            Store ("17832", Index (Arg1, 0x0A))
                            Store ("LION", Index (Arg1, 0x0B))
                            Store ("SANYO", Index (Arg1, 0x0C))
                        }

TimeWalker75a · September 27, 2013

just compare two binaries and you'll see

TimeWalker75a · September 27, 2013

Ok, so it's register operations after all. I will look more thoroughly!

While you are at it, can you check shutdown please?

Tetonne · September 27, 2013

If you want this error gone patch AppleSMC binary

replace 4885c075 10488d3d 7d50 to 4885c0eb 6e488d3d 7d50,

this will make it fall back to old mode right away without trying the new one.

This error is even present on genuine Apple hardware.

sorry could you please give more detail to do so

i thought about add this in clover config

KextsToPatch

Find

Replace

but not sure

tlucj surely understood but i did not and i would like to learn from what i read

TimeWalker75a · September 27, 2013

tluck,

give this a shot.. I have avoided using L1L4 and B1B4 for now, let's see if plain 8 to 16 bit registers would work.

http://puu.sh/4BCFp.zip

sorry could you please give more detail to do so

i thought about add this in clover config

KextsToPatch

Find

Replace

but not sure

tlucj surely understood but i did not and i would like to learn from what i read

Here you go:

	<key>KernelAndKextPatches</key>
	<dict>
		<key>Debug</key>
		<false/>
		<key>KextsToPatch</key>
		<array>
			<dict>
				<key>Comment</key>
				<string>Fall back to old SMC mode right away</string>
				<key>Name</key>
				<string>AppleSMC</string>
				<key>Find</key>
				<data>SIXAdRBIjT19UA==</data>
				<key>Replace</key>
				<data>SIXA625IjT19UA==</data>
			</dict>
		</array>
	</dict>

TimeWalker75a · September 27, 2013

It is hard coded for sakes of testing, as I've said already. I've purposely replaced every B1B4 and L1L4 conversions with hardcoded strings to see if it ends up working. But I fail to see how this is possible that it is reporting current state as capacity of warning.. unless it reads the values from SBFC and dividing it by 0x14 ends up lower as the actual current capacity reported from _BST..

This Lenovo unit is a mess :S

Can you jump to IRC for live chit chat ?

/join irc.nspwn.com

/query timewalker

tluck · September 27, 2013

ok. we have another break through of sorts. on top of the shutdown hang mixup, a further rewrite of the DSDT by TimeWalker75a has provided a way to eliminate the ACPIPlatform rollback - so we can have one less custom kext to deal with.

Note: Rehabman updated and renamed AppleSmartBatteryManager to ACPIBatteryManager.kext so it does not conflict with the native one. anyway ... Huge thanks to Timewalker75a on helping making the T420 practically a perfect Hack!

will repackage and post ASAP with new DSDT so hang on and wait if you have not upgraded to the bundle from yesterday.

FakeSMC + Plugins

VoodooPS2

ACPIBatteryManager

$ kextstat |grep -v apple
Index Refs Address            Size       Wired      Name (Version) <Linked Against>
   16    2 0xffffff7f8119f000 0x16000    0x16000    org.netkas.driver.FakeSMC (873) <9 7 5 4 3 1>
   17    0 0xffffff7f811b5000 0x4000     0x4000     org.hwsensors.driver.CPUSensors (873) <16 7 5 4 3>
   30    2 0xffffff7f80811000 0xe000     0xe000     org.rehabman.voodoo.driver.PS2Controller (1.8.8) <9 7 5 4 3 1>
   32    0 0xffffff7f811bc000 0x6000     0x6000     org.hwsensors.driver.ACPISensors (873) <16 9 7 5 4 3>
   35    0 0xffffff7f81d14000 0x6000     0x6000     org.rehabman.driver.AppleSmartBatteryManager (1.40) <9 4 3>
   38    0 0xffffff7f80e72000 0x19000    0x19000    com.insanelymac.driver.AppleIntelE1000e (2.4.14) <37 10 5 4 3 1>
   52    0 0xffffff7f80834000 0x5000     0x5000     org.rehabman.voodoo.driver.PS2Keyboard (1.8.8) <30 28 7 6 5 4 3 1>
   53    0 0xffffff7f8081f000 0xb000     0xb000     org.rehabman.voodoo.driver.PS2Trackpad (1.8.8) <30 28 7 6 5 4 3 1>

TimeWalker75a · September 28, 2013

Glad you have sorted it out. So, is the shutdown working?

Tetonne · September 28, 2013

my config with edid ID with DisplayVendorID-6af for tluck next pack

my kexts and dsdt are the one provided by tluck

huge thanks to TimeWalker75a last post s

and obviously to tluck who was there to take TimeWalker75a advices

DisplayVendorID-6af.zip

tluck · September 29, 2013

limited testing so far points toward goodness on dsdt shutdown fix. seems very consistent when i only have 1 disk.

but i have an esata and usb3 PCI express. And these extra disks seem to introduce same old random problem.

TimeWalker75a · September 29, 2013

limited testing so far points toward goodness on dsdt shutdown fix.

Good to know, hope it's fixed for good!

What is the custom ssdt that you were referring to in Clover's thread that gives you more P-States on your i5? I'm wondering because I'm not really happy about how things are on my i3.. I get 8 11 15 19 and 22 (whilst the highest one should be 23, really). Care to share what was the origin of that ssdt or just the table itself?

tluck · September 29, 2013

verleihnix was the one who created the ssdt.aml table. I recall he may have used revo girls ssdtPRgen script? the ssdt.aml is in the clover bundle i posted in #397 .

tluck · September 30, 2013

just posted Clover b2165 for T420 at post #397

Update 29-Sep-2013

Revised DSDT (again!)
- for Shutdown hang (thanks to TimeWalker75a)
- for native AppleACPIPlatform kext - no need to rollback or patch this kext - use unmodified v1.8 from 10.8.5
Updated Clover to b2165
- minor tweaks/refinements to config.plist (e.g. dont think darkwake is needed anymore)
- has 2 OEM configs - 1 for high-res and 1 for low-res
- Note: recommend use of ssdt.aml if it fits your cpu - seems to get a bit more out turbo mode (get 32x at times instead of 30x with just Clover Generate P/C states)
Updated FakeSMC
- includes modified ACPI/Plugin Info.plist to remove DISABLED features so FAN speed shows up in HWMonitor (with stock AppleACPIPlatform.kext )

verleihnix · September 30, 2013

verhleinix was the one who created the ssdt.aml table. I recall he may have used revo girls ssdtPRgen script? the ssdt.aml is in the clover bundle i posted in #397 .

Hi tluck,

the script is based on RevoGirl's script modified for i-core mobile processors found at tony's side.

Mobile_ssdtPRGen.sh.zip

tluck · September 30, 2013

Update - if you downloaded the zip bundle before now.... please redo it! I found a mistake in the DSDT.

(re-)posted Clover b2165 for T420 at post #397

Update 30-Sep-2013

Revised DSDT (again!)
- for Shutdown hang (thanks to TimeWalker75a)
- for native AppleACPIPlatform kext - no need to rollback or patch this kext - use unmodified v1.8 from 10.8.5
Updated Clover to b2165
- minor tweaks/refinements to config.plist (e.g. dont think darkwake is needed anymore)
- has 2 OEM configs - 1 for high-res and 1 for low-res
- Note: recommend use of ssdt.aml if it fits your cpu - seems to get a bit more out turbo mode (get 32x at times instead of 30x with just Clover Generate P/C states)
Updated FakeSMC
- includes modified ACPI/Plugin Info.plist to remove DISABLED features so FAN speed shows up in HWMonitor (with stock AppleACPIPlatform.kext )

Lenovo ThinkPad T420 with UEFI Only

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites