Dell XPS 1340 mostly working with OSX 10.5.6,10.5.7

[pratik.np]

hey bcc9, i know this might be too much to ask but can you try to patch the attached VoodooHDA to work with our sound cards... i mean currently it does work but with a constant disturbance... and i remember i read somewhere in this thread that it was the same problem with AppleHDA when you started to work on it at first... so may be making the same modifications should work.... VoodooHDA will be better cuz it as of now doesnt break the shutdown issues and this same kext works under snow leopard, and in snow leopard with video wifi and ps2 devices working already audio is the only thing left... i mean audio without disturbance that is...

 

i used this kext with and without the custom efi strings that we have and in any case the result is the same... any help will be greatly appreciated... thanx in advance

voodoohda.rar

[pmcnano]

Afaik, VoodooHDA has nothing to do with AppleHDA. 2nd, lol I guess bcc will answer "if you want it..do it yourself".....at least research it. 3rd..i dont have third.....

[pratik.np]

hmmm.... i dont mind doing it on my own... if only i knew wat to do....

 

i know it would take a lot of work if its a different case then the one with applehda.. i only asked cuz may be making the exact same changes to voodoohda as you ppl made to applehda might work... i dont know anything abt this...

 

and i did research abt it also.. tried googling it a lot... but 99% of ppl have voodoohda working without any modifications....and nobody wrote abt ne problems so i thought i'd ask....

 

neways if he doesnt answer then no problem.... life goes on right...lol

[ElekTron]

By the way will repair my disk permissions will messup the OS? it is safe to do? as when verifying there are like a million discrepancies... what do u guys think?

 

I did is sometimes, no problems, think it doesnt get trouble when u use it. It also shows me some discrepancies, think there are the same everytime i start it. But the system runs n runs n runs... :-)

[n/a/n/a/n/a]

I did is sometimes, no problems, think it doesnt get trouble when u use it. It also shows me some discrepancies, think there are the same everytime i start it. But the system runs n runs n runs... :-)

 

 

Thanks a lot mate, i did it an no problems. Some of them disappear a few still there, but who cares as long as the computer works am happy!

 

Take care

[pmcnano]

I was just trying some new kext for using Vanilla Speedstep, and maybe I did something wrong..but...I just noticed that disabling CPupowermanagement.kext (disabler.kext or anything similar) IM PRETTY SURE reduces battery life...Im getting almost 3.30 - 3.45 hours with the kexts loaded when using disabler I got 2.45-3 hours....

[ElekTron]

I was just trying some new kext for using Vanilla Speedstep, and maybe I did something wrong..but...I just noticed that disabling CPupowermanagement.kext (disabler.kext or anything similar) IM PRETTY SURE reduces battery life...Im getting almost 3.30 - 3.45 hours with the kexts loaded when using disabler I got 2.45-3 hours....

 

You get 3.30 - 3.45?!?! With normal 6 cell battery in the 1340 Dell? Which components you have inside? I only have around 2-2,5 hours with the 6 cell... When i open system status it shows me that my battery has a Full loading capacity (in mAh): 5042, mostly around 5000. :-(

[pmcnano]

Lol no, I posted a while ago that i Bought my 9cell battery! But then again, loading intelcpupowermanagement do helps with the battery life ! ^^

[bcc9]

hey bcc9, i know this might be too much to ask but can you try to patch the attached VoodooHDA to work with our sound cards... i mean currently it does work but with a constant disturbance... and i remember i read somewhere in this thread that it was the same problem with AppleHDA when you started to work on it at first... so may be making the same modifications should work.... VoodooHDA will be better cuz it as of now doesnt break the shutdown issues and this same kext works under snow leopard, and in snow leopard with video wifi and ps2 devices working already audio is the only thing left... i mean audio without disturbance that is...

 

i used this kext with and without the custom efi strings that we have and in any case the result is the same... any help will be greatly appreciated... thanx in advance

I don't remember reports of AppleHDA working with distorted audio, just that it would sometimes not work. Can you point me to such a post? It could help in figuring out what is going wrong with voodoohda if there was a way to make applehda misbehave the same way.

 

The patches done for applehda do not directly translate to voodoohda as there should be no need for EFI strings or info.plist hackery to make voodoohda work. The voodoohda code, being based upon the freebsd code, is able to pick up the configuration from the hardware codec directly (more completely than the linux code I might add). Manually adding pinconfig data that was already in the hardware shouldn't be necessary with a working driver, btw. Quirks for hardware workarounds could be added to voodoohda like the linux code has all over the place, but it looks to me the speakers should be working OK without any quirks. I tried manually adding some quirks of my own anyways, (to disable unused DACs, disable the mixer), but no improvement. I've tried many other tricks as well, such as assigning the HDA hardware its own private IRQ via DSDT, no improvement with the private interrupt vector.

 

The only thing that cleans up the audio so far is running a process that generates a lot of disk or display interrupts. (on the order of 25 interrupts per second).

 

and i did research abt it also.. tried googling it a lot... but 99% of ppl have voodoohda working without any modifications....and nobody wrote abt ne problems so i thought i'd ask....

There are quite a lot of posts about voodoohda distorting the audio in exactly the same way as we're seeing here. In fact there are several bug reports on file for this already over at http://code.google.com/p/voodoohda/ and the voodooprojects forum http://forum.voodooprojects.org/index.php; no evidence that anyone else is looking into any of the problem reports.

 

Lol no, I posted a while ago that i Bought my 9cell battery! But then again, loading intelcpupowermanagement do helps with the battery life ! ^^

To figure out the difference, maybe you could check with cpu-x to see if the cpu is being throttled to a lower multiplier in one case vs the other.

 

Why when i install the hid2hci (to fix bluethooth) program from bcc9, everything comes out successful but still dosent work, is there any little step that am missing? [ i just downloaded the file and installed ]. I have even tried to install it manually from terminal on sudo -s priviledges. Any Ideas why?? am using 10.5.7.

 

Once again, Thanks for your help guys.

Works for me (root access not necessary). Not sure why you'd be having problems. I assume the bluetooth device is showing up OK under system-profiler's usb device list. Maybe you have bluetooth partially disabled via the bios/windows config setting that configures the devices affected by the wireless control key.

[pmcnano]

bcc, no theres no other multiplier used, I think its using lower voltages??? But I cant check it since cpu-x doesnt report voltage here. Dont know why.

 

On the other side, VoodooHDA is actually distorted, I tried using it upon its release but it never worked and since it seems a dead project...well It really didnt matter.

 

=)!

[GAOO]

To bcc9:

 

Hi, I've been reading through a lot of your progress on the XPS 1340, and I'm very impressed by your dedication I was wondering if you would be willing to answer a few questions of mine?

 

I have a Dell Studio 14z (1440). The specs of the 14z are very similar to the XPS 1340, and the GPU is identical. I was hoping that I could just use your provided IOPCIFamily.kext and the EFI plist file to make the magic happen for my GPU, but sadly nothing gets displayed. I have previously done some EFI string stuff for my other laptop, where its 8400GS GPU worked 100% perfectly after. However, the community doesn't seem to have much info on the GeForce 9 series and I'm at my wits end on how to proceed.

 

In essence, I am curious what changes you have made to the IOPCIFamily.kext (is it specific only to the XPS 1340, or to the 9400M GPU in general?), and how you came up with the NVCAP value in your EFI plist? To my understanding on what I read so far, getting the correct EFI string would essentially make the 9400 M GPU work, but the modified IOPCIFamily.kext is to help get the sound working as well.

 

 

Any help would be greatly appreciated, thank you.

[bcc9]

To bcc9:

 

Hi, I've been reading through a lot of your progress on the XPS 1340, and I'm very impressed by your dedication smile.gif I was wondering if you would be willing to answer a few questions of mine?

Thanks, I've spent too much time on this haven't I?

I have a Dell Studio 14z (1440). The specs of the 14z are very similar to the XPS 1340, and the GPU is identical. I was hoping that I could just use your provided IOPCIFamily.kext and the EFI plist file to make the magic happen for my GPU, but sadly nothing gets displayed. I have previously done some EFI string stuff for my other laptop, where its 8400GS GPU worked 100% perfectly after. However, the community doesn't seem to have much info on the GeForce 9 series and I'm at my wits end on how to proceed.

Someone just PM'd me the very same question yesterday. (Looks like I'm going to have to stop responding to PMs it seems as I hate repeating myself and I don't have time for private tech support anyways. Please could people just post Qs publicly. Anyways...)

 

The IOPCIFamily change I made isn't necessary if you instead use natit. I'd recommend the natit approach at this point for 9400m users as it avoids having to modify apple's stock kexts. So try the approach here: http://www.insanelymac.com/forum/index.php...t&p=1204995

The version of natit I posted there has been modified with an NVCAP that supports the 3 external video ports found on the xps 1340. I've also added the pci device ids for the 9400m and 9200m so that they are identified by name.

I've been running my xps13 with 9400m with this for weeks now. Works under snow leopard too.

 

Also if you install osx86tools and select 'View PCI Device/Vendor ID', then it should install lspci. Then if you run lspci -nnv and lspci -vt and post the results, then we can compare exactly what hardware should work across both systems. I have not seen any lspci details for the xps14 on the net as of yet.

 

In essence, I am curious what changes you have made to the IOPCIFamily.kext (is it specific only to the XPS 1340, or to the 9400M GPU in general?), and how you came up with the NVCAP value in your EFI plist? To my understanding on what I read so far, getting the correct EFI string would essentially make the 9400 M GPU work, but the modified IOPCIFamily.kext is to help get the sound working as well.

I came up with the nvcap myself by inferring things from the partially explained reverse engineering effort here:

http://nvinject.free.fr/forums/viewtopic.php?t=214

and also by using the ioreg output from genuine macbooks.

 

The IOPCIFamily change was just the first way I was able to get the video EFI strings to be picked up by the stock nvidia kext. Not specific to the 9400m. No, the IOPCIFamily change has nothing to do with audio.

 

bcc, no theres no other multiplier used, I think its using lower voltages??? But I cant check it since cpu-x doesnt report voltage here. Dont know why.

Well superhai's code doesn't work with super-lfm modes, like I've mentioned, so that might be the distinction. You could check your cpu voltages to see if you're managing a super-lfm mode in the case where your battery is lasting longer.

On the other side, VoodooHDA is actually distorted,

Exactly (unless you busy the system with a lot of extra i/o)

I tried using it upon its release but it never worked and since it seems a dead project...well It really didnt matter.

But it will matter as you can't keep using the 10.5.3 based applehda solution with 10.6.

[raider73]

bcc9,

 

I don't see it mentioned on the first post so I'll ask. It's not on your list of hardware working/not working. Does your card reader work in OS X? Over the weekend, I tried using the card reader and it doesn't recognized the SD card that I put in. I wouldn't be surprised that it doesn't work since the Macbooks didn't come a card reader. The new Macbook Pros do have a card reader though. Just asking? I'm already happy as is since this laptop runs like {censored} on Vista or Windows 7--freezes and lockups all the time. It works slightly better in Linux and it has been running over a week on OS X 10.5.7 w/o any crashes/lockups yet(fingers crossed).

[pmcnano]

As I said before, theres a Voodoo kext in their forums, feel free to try!

[bcc9]

bcc9,

 

I don't see it mentioned on the first post so I'll ask. It's not on your list of hardware working/not working. Does your card reader work in OS X? Over the weekend, I tried using the card reader and it doesn't recognized the SD card that I put in. I wouldn't be surprised that it doesn't work since the Macbooks didn't come a card reader. The new Macbook Pros do have a card reader though. Just asking?

Under OSX, I haven't tried to get working (and nobody else seems to have either): SD card, expresscard, or infrared.

As pmcnano has mentioned a few times now, there's the new voodoo project for SD support, so I'd recommend you try that and report back. Guess I'll add these 3 components to post #1. [update: done]

I'm already happy as is since this laptop runs like {censored} on Vista or Windows 7--freezes and lockups all the time. It works slightly better in Linux and it has been running over a week on OS X 10.5.7 w/o any crashes/lockups yet(fingers crossed).

Ha ha. W7 with the latest&greatest nvidia drivers for the chipset&video seems to work pretty well. Everyone else seems to report problems with windows vista until the dell bloat-ware is uninstalled. I only have hanging problems with esata at 3gbps; esata issues seem to be a fundamental nvidia chipset problem....

[raider73]

I only have hanging problems with esata at 3gbps; esata issues seem to be a fundamental nvidia chipset problem....

 

Maybe that is my problem since I have an external eSATA drive connected at all times when the laptop is home.

[GAOO]

Thanks a lot for your help bcc9, I will be trying the natit method now.

 

I have also dumped the lspci outputs to text files, included in the attachments. Looking at the list now, I'm not so sure if the XPS 1340 specs is so similar to the Studio 1440

 

I've also done the "reverse" engineering work to get the NVCAP for my previous laptop's 8400 Unfortunately I just can't seem to get it for the Studio 1440. So thanks a lot for all your help! I was also thinking about PMing you, but I figured your inbox should be filled with questions by now

 

I'll let you know how everything goes for the Studio 14z. Once I get the video, sound, and ethernet working, I'll be making a separate thread on how to make things work in the 14z. Unfortunately, I don't think I can be as dedicated as you though

 

 

Thanks, I've spent too much time on this haven't I?

 

Someone just PM'd me the very same question yesterday. (Looks like I'm going to have to stop responding to PMs it seems as I hate repeating myself and I don't have time for private tech support anyways. Please could people just post Qs publicly. Anyways...)

 

The IOPCIFamily change I made isn't necessary if you instead use natit. I'd recommend the natit approach at this point for 9400m users as it avoids having to modify apple's stock kexts. So try the approach here: http://www.insanelymac.com/forum/index.php...t&p=1204995

The version of natit I posted there has been modified with an NVCAP that supports the 3 external video ports found on the xps 1340. I've also added the pci device ids for the 9400m and 9200m so that they are identified by name.

I've been running my xps13 with 9400m with this for weeks now. Works under snow leopard too.

 

Also if you install osx86tools and select 'View PCI Device/Vendor ID', then it should install lspci. Then if you run lspci -nnv and lspci -vt and post the results, then we can compare exactly what hardware should work across both systems. I have not seen any lspci details for the xps14 on the net as of yet.

 

I came up with the nvcap myself by inferring things from the partially explained reverse engineering effort here:

http://nvinject.free.fr/forums/viewtopic.php?t=214

and also by using the ioreg output from genuine macbooks.

 

The IOPCIFamily change was just the first way I was able to get the video EFI strings to be picked up by the stock nvidia kext. Not specific to the 9400m. No, the IOPCIFamily change has nothing to do with audio.

 

Well superhai's code doesn't work with super-lfm modes, like I've mentioned, so that might be the distinction. You could check your cpu voltages to see if you're managing a super-lfm mode in the case where your battery is lasting longer.

Exactly (unless you busy the system with a lot of extra i/o) But it will matter as you can't keep using the 10.5.3 based applehda solution with 10.6.

lspci_nnv.txt

lspci_vt.txt

[GAOO]

Just tried the natit file you uploaded and unfortunately, black screen only. I don't quite get the difference between using an EFI string and an injector, anyhow...

 

I'm certain that the GPU device path is correct (identical to the XPS1340), but nothing other than that. I still feel like the NVCAP value doesn't seem to work for the Studio 14z, so I'll be playing around with that in the Natit's Info.plist file.

 

BTW, I always have NVDAResman.kext loaded first followed by 3 of the same message "IGPU: Family Specific Matching failed". I'm assuming this is normal since this is before Natit.kext is injecting values.

 

Thanks a lot for your help bcc9, I will be trying the natit method now.

 

I have also dumped the lspci outputs to text files, included in the attachments. Looking at the list now, I'm not so sure if the XPS 1340 specs is so similar to the Studio 1440

 

I've also done the "reverse" engineering work to get the NVCAP for my previous laptop's 8400 Unfortunately I just can't seem to get it for the Studio 1440. So thanks a lot for all your help! I was also thinking about PMing you, but I figured your inbox should be filled with questions by now

 

I'll let you know how everything goes for the Studio 14z. Once I get the video, sound, and ethernet working, I'll be making a separate thread on how to make things work in the 14z. Unfortunately, I don't think I can be as dedicated as you though

[HannibalX]

First off, I'd like to thank bcc9 for this excellent thread! :-)

I have a Dell Studio 14z, and I've used his guide to get things going. I used a retail OSX 10.5.6 install DVD; I copied it to a USB thumb drive and used Chameleon 2.0 RC2 to make the device bootable. The laptop is dual boot with Windows 7 and OSX, with Chameleon 2.0 RC2 as the bootloader, installed using 'Munky's method' in the EFI partition. I don't have everything going yet, as I'm all about keeping the system as vanilla as possible. But I can confirm that the natit.kext that bcc9 uploaded is working just fine. I'm even loading it from my EFI partition with no issues at all.

 

GAOO -

Did you remove any EFI strings you may have added, or any other type of custom video kexts?

 

 

Regards,

Hannibal

[GAOO]

Hey HannibalX, thanks for the reply. I don't want to hijack this thread, so do you mind if you can PM me with your e-mail address or some method of communicating with you?

 

I don't have any other EFI strings. But I'll do a clean install just to make sure there's nothing else interfering. [Edited out stuff here] Anyhow, we should make a completely separate thread for the 14z, but I'll probably do that once I have mine working and gather more info.

 

EDIT: sorry just noticed what steps you've taken. My bad for rushing and not reading it carefully.

 

First off, I'd like to thank bcc9 for this excellent thread! :-)

I have a Dell Studio 14z, and I've used his guide to get things going. I used a retail OSX 10.5.6 install DVD; I copied it to a USB thumb drive and used Chameleon 2.0 RC2 to make the device bootable. The laptop is dual boot with Windows 7 and OSX, with Chameleon 2.0 RC2 as the bootloader, installed using 'Munky's method' in the EFI partition. I don't have everything going yet, as I'm all about keeping the system as vanilla as possible. But I can confirm that the natit.kext that bcc9 uploaded is working just fine. I'm even loading it from my EFI partition with no issues at all.

 

GAOO -

Did you remove any EFI strings you may have added, or any other type of custom video kexts?

 

 

Regards,

Hannibal

[bcc9]

Maybe that is my problem since I have an external eSATA drive connected at all times when the laptop is home.

Recommend you make sure any esata drive is jumpered to 1.5gbps else I think you'll have I/O problems (hangs) even after the disk mounts up fine. Apple got a firmware fix for their nvidia SATA speed issues with the mcp79 but nothing from dell so far.

 

I have also dumped the lspci outputs to text files, included in the attachments. Looking at the list now, I'm not so sure if the XPS 1340 specs is so similar to the Studio 1440

That shows that the chipset is identical, same revision even, and so your graphics, sata, dsdt, gpe handler (lid/battery) issues are likely the same. The ethernet & 802.11 are different so you guys should make a separate thread for those at least.

 

I don't quite get the difference between using an EFI string and an injector, anyhow...

 

Well I don't fully understand why the EFI string solution seems to require IOPCIFamily to be hacked for the 9400m whereas the injector solution does not. I think it stems from how the EFI solution adds the attributes to the IOACPIPlane whereas the injector solution adds the attributes to the IODeviceTree. What I know is the former breaks the initialization of the nvidia kexts if IOPCIFamily is allowed to match on the video device and adds it to the IODeviceTree plane.

 

Maybe someone can fill in the blanks above.

 

So I have two different solutions; both have their downsides. EFI strings are cryptic, have hard-codec PCI bus and slot numbers, are hard to edit, especially when combined with other devices such as audio. And for the 9400m, the EFI solution somehow requires IOPCIFamily to be modified as well.

Injectors require one to manage a kext which isn't quite set&forget (even with /Extra), may require tweaking across major OSX upgrades, and the binary-only injectors that are floating around add confusion and aren't well maintained.

 

So I'm assuming you guys now have working QE/CI with my modified natit kext.

[Cheesy]

Hi guys,

 

 

Just to tell you that I am not using voodoo power and voodoo battery, just using vanilla Appleintelcpupowermanagement.kext and recompiled vanilla AppleACPIBatteryManager.kext. The CPU throttles correctly and the battery meter works ok. Here are my AppleACPIBatteryManager kext and my dsdt (slightly modified from bcc9's as to enable Appleintelcpupowermanagement to work correctly). With this, it is easier to upgrade, you do not need any disablers and no script to remove appleintelcpupowermanagement during upgrades.

 

 

Regards to you all !

dsdt_fixed.txt

AppleACPIBatteryManager.kext.zip

[pmcnano]

afaik AppleACPIBatteryManager.kext isnt a vanilla kext, its just like VoodooBattery, its an aditional kext. The cpu throthling is interesting though..........

[GAOO]

I work during the day so I only have very little time at night to get OSX working. It's nice to know that the chipsets and what not are all the same. Wifi for our G-cards works, not for the N-cards. Someone else with the N-card is looking into it, and I'll be helping him once I get stuff set up on my end.

 

Anyhow, Natit doesn't work. And I am guessing this is because I am using iATKOS v7, where something in it is patched. Natit is unable to detect and show anything on the laptop LCD, but it is able to produce a correct resolution screen on an external HDMI monitor. However, there's a lot of incorrect graphics rendering due to memory allocation errors from what I see on the console outputs.

 

HannibalX said he got Natit working using Boot-132 and Retail OSX. Based on your first post bcc9, I am guessing that is also what you are doing, so I'll be trying that out tonight to get Graphics working with Natit.

 

Again, thanks a lot to bcc9 and everyone else who's involved as well.

 

 

Recommend you make sure any esata drive is jumpered to 1.5gbps else I think you'll have I/O problems (hangs) even after the disk mounts up fine. Apple got a firmware fix for their nvidia SATA speed issues with the mcp79 but nothing from dell so far.

 

That shows that the chipset is identical, same revision even, and so your graphics, sata, dsdt, gpe handler (lid/battery) issues are likely the same. The ethernet & 802.11 are different so you guys should make a separate thread for those at least.

 

 

 

Well I don't fully understand why the EFI string solution seems to require IOPCIFamily to be hacked for the 9400m whereas the injector solution does not. I think it stems from how the EFI solution adds the attributes to the IOACPIPlane whereas the injector solution adds the attributes to the IODeviceTree. What I know is the former breaks the initialization of the nvidia kexts if IOPCIFamily is allowed to match on the video device and adds it to the IODeviceTree plane.

 

Maybe someone can fill in the blanks above.

 

So I have two different solutions; both have their downsides. EFI strings are cryptic, have hard-codec PCI bus and slot numbers, are hard to edit, especially when combined with other devices such as audio. And for the 9400m, the EFI solution somehow requires IOPCIFamily to be modified as well.

Injectors require one to manage a kext which isn't quite set&forget (even with /Extra), may require tweaking across major OSX upgrades, and the binary-only injectors that are floating around add confusion and aren't well maintained.

 

So I'm assuming you guys now have working QE/CI with my modified natit kext.

[pmcnano]

Why dont you try a retail install??? REmember, using a 10.5.6 disk, not something previous......since MCP79 isnt supported before