Jump to content

[Guide] Dell XPS 15 L502X (Early 2011) Snow Leopard Install (possibly L702x too)

wen1

By wen1,
in Installation Guides

 Share
3,030 posts in this topic

Recommended Posts

DoiX
DoiX
Posted
Posted

Your speedstep is working, DawTaylor. Missing P-States are normal, they're rather random from install to install.

 

 

i7 Users: Does kozlek's latest revision detect your fan speed?

Link to comment
Share on other sites
TimeWalker75a
TimeWalker75a
Posted
Posted

Your speedstep is working, DawTaylor. Missing P-States are normal, they're rather random from install to install.

 

 

i7 Users: Does kozlek's latest revision detect your fan speed?

It shouldn't vary from install to install as the amount of p-states in your ssdt tables (or clover's generated tables based on your lowest and highest multipliers) remains constant. I haven't seen any stepper errors once. You have to troubleshoot this...

 

Regarding fan speed reading.. why should it be only related to i7 users? Fan monitoring is based on ACPI Sensors kext and is actually pulling data by executing methods from your DSDT table. 

https://github.com/Dolnor/Vostro-3450-Fan-Override

 

Remove the PLLD device and you leave just SCMD in place and compile an SSDT table from it.. will end up as this:

4BnXD.png

Link to comment
Share on other sites
DoiX
DoiX
Posted
Posted

Because it was working some revisions ago, i naturally assume something changed kozlek's code, since i didn't change a single thing in my SSDT/DSDT...

 

On my old install 10.8 -> 10.8.4 vanilla, i had like 12 P-States, now, on the straight 10.8.4 install, i have from 8x to 29x. Same DSDT/SSDT tables. So something does affect those power states. I also recall from 2 years ago when i messed around with SSDT tables that some i7 users had different P-States locked out(aka missing) than i had

Link to comment
Share on other sites
jkbuha
jkbuha
Posted
Posted

TW - For some reason I cannot compile (either with iasl or iasl64) your SSDT file - the compiler just ramps up to 100% CPU usage and freezes). Do you have a compiled aml I can use?

 

Also I have zipped up the modified/new kexts AFAIK - but if anyone knows of a more systematic method to detect new/changed kexts please enlighten me.

archive.zip

Link to comment
Share on other sites
TimeWalker75a
TimeWalker75a
Posted
Posted

Odd, I'm getting a perfect compilation:

4CkAz.png

 

Here's compiled table for you: http://puu.sh/4CkBE.zip

If your battery registers were split differently (ie. not SOT0 and SOT1) .. if ever split of course, then it wouldn't work for you, at least not with battery voltage sensor in place.

 

I see you are using AppleACPIPlatform from 10.6.2, which can't be a good thing! You'd better split your registers for battery to work and one less rollback would be required.

AppleBacklightExpert is not needed as far as I can tell, at least not on my vostro as backlight controls from the framebuffer work absolutely fine. ICHSmbus kext was intended for Andigilog SuperIO sensor which is not found on your 15z, so all it does is interferes with AppleSMBUS and prevents it from loading. I don't see why would you need IOGraphicsFamily rollback either. 


Also, I have fixed shutdown issue for UEFI. Actually, I was advised  by a person who goes by the name endless to look at it.

All along (almost a year now) it wasn't any of Clover's fault, it was ours. DSDT had to be fixed in order to get proper shutdown... 

For our machines just the soft-off case (0x05) is enough to fix the shutdown, even though intel recommends to write 0 before entering any of the sleep states. 

 

4Clbb.png

 

The most retarded thing is, Pheonix had even defined PM Base Address + 30, which is SLP_SMI_EN register as Name (PM30, 0x0430) but never used it for anything!

So you have to define the SLP Register yourself and change the _PTS method, adding the If (LEqual (Arg0, 0x05)) chunks of code..

    OperationRegion (SLPR, SystemIO, PM30, One)
    Field (SLPR, ByteAcc, NoLock, Preserve)
    {
            ,   4,
        SLPE,   1
    }

    Method (_PTS, 1, NotSerialized)
    {
        Store (Zero, P80D)
        P8XH (Zero, Arg0)
        Store (One, MY9F)
        If (LEqual (Arg0, 0x03))
        {
            If (LAnd (DTSE, LGreater (TCNT, One)))
            {
                TRAP (TRTD, 0x1E)
            }
        }

        If (LOr (LEqual (Arg0, 0x03), LEqual (Arg0, 0x04)))
        {
            If (LEqual (Arg0, 0x04))
            {
                Acquire (\_SB.PCI0.LPCB.EC0.MUT0, 0xFFFF)
                \_SB.PHSR (0xB2)
                Release (\_SB.PCI0.LPCB.EC0.MUT0)
            }

            Store (One, SLPO)
        }

        If (LEqual (Arg0, 0x05))
        {
            Store (Zero, SLPE)
            Sleep (0x10)
        }
    }
Link to comment
Share on other sites
jkbuha
jkbuha
Posted
Posted

Thanks TW. Still not sure why my iasl freezes when compiling the original table, but interestingly enough it decompiled & recompiled your aml fine.

 

Fair point re: AppleACPIPlatform - it's because I was using a patched AppleSmartBatteryManager.kext which only worked in combo. Do you have an unmodified 10.7.5 AACPIP kext and corresponding DSDT entry to split registers?

 

IOGraphicsFamily was to modify AGPM settings but I believe this can also be removed. Again - do you have the originals so that I can verify?

 

Oh, and it looks like shutdown on UEFI now works flawlessly :) Kudos to yourself + endless

Link to comment
Share on other sites
TimeWalker75a
TimeWalker75a
Posted
Posted

Dump me your original DSDT table via Clover (F4 in GUI, then get it from //EFI/ACPI/origin .. the folder has to be present or the files won't store).

Sadly, I don't have 10.7.5 anymore, you can download the combo update from Apple's Support page and extract it using free utility called unpkg  then just install the kexts you desire.

Link to comment
Share on other sites
TimeWalker75a
TimeWalker75a
Posted
Posted

http://puu.sh/4Cta0.zip

here's a patched one for you.

 

are you still using the FOWM method though? I've found it's broken with no way to fix it and as I have media keys enabled it's deeply unpleasant when you accidentally hit F2 and it freezes..

I ended up just making the method blank:

 

        Method (FOWM, 2, NotSerialized)
        {
        }
Link to comment
Share on other sites
jkbuha
jkbuha
Posted
Posted

Many thanks for this TW. I'm diffing the unpatched<->TW<->current versions as we speak just to see what has been changed and left in place.

 

My FOWM is still intact:

        Method (FOWM, 2, NotSerialized)
        {
            Acquire (PSMX, 0xFFFF)
            Store (0xE0, WCMD)
            Store (0x57, TPID)
            Store (Arg0, WCCL)
            Store (Arg1, WCSE)
            Store (Zero, PHXI)
            Release (PSMX)
        }

But as I don't normally boot UEFI I'm not entirely sure I'm affected by this. Also I don't really use F2 as I don't use the media keys.

 

On a related note; did the patches by pene allow you to boot (stable) into Win8 from Clover? Selecting exit from clover goes straight to \boot for me at the moment, which is just about fine cos I whack up chameleon and I can select whichever OS I want in legacy mode. Still means I cannot pass anything interesting to the target OS from Clover, but I'm guessing that's a dead duck anyway for Win/Linux...

Link to comment
Share on other sites
TimeWalker75a
TimeWalker75a
Posted
Posted

Yeah, if you change the FOWM nemethod booting with CloverEFI or Chameleon your Wi-Fi toggle shortkey will end up dead, so don't use that if you boot non-UEFI more frequently.

Actually, retreat from booting legacy with any of these patches. They were meant only for UEFI Clover and not CloverEFI.. same goes for Exit, it only properly works for UEFI as it gives the control back to BIOS and BIOS selects the second legacy-bootable entry from boot order (in your case it's Chameleon's boot file).

 

The patches Pene implemented ended up corrupting NVRAM after every legacy boot, deleting DIagnostic, Setup and Boot Manager variables from it so reflashing bios from Windows had to be used after every attempt. At one point it wiped HDD boot variable clean, so I was not even able to boot anything except CD.

Link to comment
Share on other sites
jkbuha
jkbuha
Posted
Posted

Hmm for some reason one of your dsdt changes disabled my broadcom wifi card. I'll need to manually turn on/off each one until I find out which, and why.

 

CloverEFI actually doesnt work on my machine - it freezes on displaying the drive screen. UEFI is the only way I can boot clover.

 

Wow. Pity about pene's method; looked like we were getting somewhere. So do you still boot Win 8 in legacy?

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...