Potential Fix for S3 sleep and headphone jack #3
Comments
|
@zachs78 Also, I forgot to mention-karabiner elements will let you map your function keys, including brightness. I recommend checking it out. You might have to use the top-case apple logo setting to change brightness if the default brightness option doesn't work. You'll understand if you check it out. Another quick question: given the amazingly large trackpad, do you have any palm rejection issues? Good luck! |
|
rEFInd isn't going to help. This is an OpenCore EFI? AFAIK rEFI is an EFI to re-enable S3 for Windows 2004. I am not on Windows 2004 yet and CsEnable works well there. S3 is just as broken though (Dell BIOS bug) and S2 sleep is terrible! I want to map the brightness key properly via SSDT. Any idea how to do that? I haven't had any problems with palm rejection to be honest, but I use an external keyboard and mouse most of the time. The keyboard layout with the home button sharing the same physical button as F11 doesn't quite work for a programmer. Given that I have an external dock and bluetooth headset, I haven't found the need to use a headphone jack either. The only real annoyance is I can't put the laptop to sleep, but it's no different to Windows. |
|
I could remap brightness keys I had also own a Dell XPS, I would share my effort here |
|
@zachs78 I'll contact dell support, and also found your dell threads about broken sleep, dell don't want to test it internally and want to replace the motherboard of my device, what do you think? |
|
There might be another way for enabling sleep support. Apparently S3 sleep mode can be successfully hacktivated on either the 13” 9300, as well as the 17” 9700 model. That said it is know that the same setting do not yield in a useable S3 mode on the 9500 model due to the aforementioned firmware bug. However what might be a doable approach, is to flash a 9700 bios to the 9500 model as those are fairly similar. This could also be done incrementally by comparing both bios and replacing only certain modules (UEFI tool + bin-Walk). |
|
AcpiPatcher.efi can enable s3 standby but when computer wake it will be black screen and have to be hard reset
|
|
How did you enabled s3? |
Yeah, and cannot wake :))) |
Use rEFind to enable s3 with driver: acpipatcher.efi .By this way linux and win wake with black screen |
|
yep you can sleep but when wake up laptop screen will be black just, backlight just turn on |
|
But external Screen Turn on? theres a new bios update 1.7 or something like this |
My bios is newest. Maybe flash bios 1.4 will slove the problem |
|
@zachs78 bios 1.8 slove the S3 Sleep problem? I tried the DeepinOS, this OS can support Sleep |
|
For me 1.8 also not enable s3 sleep |
|
one of many threads is here : https://www.dell.com/community/Precision-Mobile-Workstations/precision-5550-S3-sleep-not-working/td-p/7759319 ( 5550 precision is pretty much xps 9500 with different wifi and nvidia chip). Dell does a terrible job here and should really fix this in the BIOS. |
It would make me curious to believe then if its just platformid/ACPI patching for the display part just to get working again otherwise, thats what I'd first suspect since its coming out of sleep |
|
For the record, Dell did replace my motherboard, did a terrible job at it, with a screw missing, screws loose, protection pads loose, etc. The support was the worst experience ever. On the plus side, I got them to admit that the problem is in the BIOS ( firmware ) and also that they will fix it, but that they could not give a timeline - so don't be surprised if it is in fact never. I also tried to patch the DSDT by having S3 do S0 but I am not sure if it did anything, I got the same issue with the wake to logo hang issue. I did try to compare an XPS 9700 BIOS with my Precision 5550 but there were to many differences. Could anyone dump the full DSDT and SSDT files for the latest BIOS of that? I hope someone else with more knowledge on this may be of help, I think looking with more understanding to the diff might point to the cause of the problem, or at least makes it work with the sign-of-life options off, like on the XPS 9700. I would love to see S3 fixed. Let's keep going :) |
|
Precision 5550 BIOS got an update to 1.22 without option to go back to 1.21 or earlier. |
|
update to 1.23.1 , no change. yes Dell, you suck. |
|
I am thinking about selling the Precision 5550 because of the sleep issue. Does anyone have any suggestions on last attempts tp try? I bought an HP and intent to never look back if I get a good price for this machine. cheers |
|
Got my hands on an XPS 17 and thought sleep did not work, but it turned out it only works when power is plugged in at the time the sleep is triggered, see also : https://www.reddit.com/r/Dell/comments/hla8yk/how_to_enable_s3_deep_sleep_on_xps_17_9700_in/ . This does not work on the XPS 15. I am not sure if I have enough time and skill to find the XPS 17 issue - this seems fixable. If anyone is up for the task or can help, please let me know. Tips on tooling, testing, debugging, methods etc. |
I have asked the DELL service, the recover is XPS9500 The hardware design is flawed, so I give up trying |
|
Hi guys, spotted this just now when I was looking at S3 sleep for the 9520. This is going to sound a bit controversial, but under one of the latest mainline linux kernels (mine's kali 6.1.0) it is possible to successfully trigger S3 sleep/wake under Linux. Logs hereunder: `──(jkbuha㉿kali)-[~/Downloads/XPS-9520-AlderLake-OpenCore-main/EFI] ┌──(jkbuha㉿kali)-[~/Downloads/XPS-9520-AlderLake-OpenCore-main/EFI] ┌──(jkbuha㉿kali)-[~/Downloads/XPS-9520-AlderLake-OpenCore-main/EFI] ┌──(jkbuha㉿kali)-[~/Downloads/XPS-9520-AlderLake-OpenCore-main/EFI] To validate this you'll need to (a) set the ACPI UEFI IFR 'S3 Sleep' variable to enabled (default is off for 9510/20/30 but should be set for 9500) and also boot linux from an existing partition where mem_sleep is already set to [deep], as it's usually not enabled by default on a live usb. I suspect the kernel devs have activated a different wake path via _GPRW but will spend some more time investigating this to see if it can be backported onto MacOS and injected via SSDT. |
|
hi, thanks for sharing, that may be interesting. Just for clarity, when running Linux with mem_sleep set to deep ( kernel param ) on both 9500 and 9700 models, they hang, but the 9700 works when on power. |
|
I can leave here the dmesg log for a successful 9520 deep (S3) mem_sleep
In the kernel dev chats I can see some unorthodox return to wake through a different EC mapping. I'll see where it leads. In the meantime if you can see if can spot any differences in the 9700 sleep logs (or at least where the 9500 fails). |
|
This is my dmesg output for a lid closed sleep on Fedora 38, kernel 6.4.14-fc200 on the XPS 9700 ( latest BIOS ). [ 153.230924] wlp0s20f3: deauthenticating from ba:15:a2:3b:f0:cd by local choice (Reason: 3=DEAUTH_LEAVING) might be an issue with NVram because now the touchscreen does not work on wake and there is a strange effect where the logo shows, then gets garbeled and another sleep wake action seems to be needed. I will try to post again when that is fixed, but thought it might be useful to share anyway. This sing of life ( logo ) has something to do with the issue too I believe. Would be fantastic to get to the bottom of this :)! |
|
here is the output of dmesg after an NVram reset. Sign of life logo was on in Bios, after disabling that the behaviour disappeared and I get a 'normal S3 sleep' on this XPS 9700: [ 15.421439] evdi: [I] (card3) Closed by Task 3099 (gnome-shell) of process 3099 (gnome-shell) The touchscreen stops working after the wake, which is something I read about happening for other models. That does not happen on MacOS last time I checked. |
|
Can you attach here the current (linux) DSDT.aml for the 9700? sudo cat /sys/firmware/acpi/tables/DSDT > DSDT.aml |
|
sure, there you go. |
|
hi, any updates ? does anyone know how I could find what changes on the ACPI level when the power is plugged/unplugged ? With that I could look at the methods of the BIOS file to see where such a change can affect the sleep. Thank you. |
|
similar issue : EETagent/T480-OpenCore-Hackintosh#20 |
|
The BIOS errors are because you reduced the RTC window. You need to keep it
to A0-A7 at least as that range covers the BIOS error checksums (if you
expose it to macOS hibernate it will write some incorrect values to the
BIOS causing those errors.
If it is indeed WiFi, then worth trying to change/disable driver and see if
there’s any difference in power consumption.
…On Fri, 29 Sep 2023 at 20:38, OpenMinded ***@***.***> wrote:
similar issue : EETagent/T480-OpenCore-Hackintosh#20
<EETagent/T480-OpenCore-Hackintosh#20>
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMC6SBG3SE6DMGRH5E6TX44PSHANCNFSM4R3ZR2FQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Great, thanks, put it back. I disabled the wifi in the BIOS and it looked at first there was some difference, but a bit after not. Depends if this Coconutbattery is accurate, it takes a while to change. Perhaps it has to do with that linked issue, so the iGPU. I added igfxonln=1 back and igfxfw=2 for testing , can't notice difference but wake is fast. Another finding : 4 TB crucial only works with hibernate in SSD slot 2, which I think is enough to conclude it may not work with any other drive in SSD slot 1. I can image that if an EFI on a disk in slot 2 may work on a disk in slot 1, I will probably not spend more time to test unless anybody thinks that is really useful. |
|
There's definitely an increased power draw post S4 hibernation, I estimate
it's about 1.5-2x idle (just as if eg: mds_stores was running). You can
measure it with non-CPU/GPU apps such as coconutbattery and wattagio. Given
that there's no os impact (geekbench has similar scores pre- and post-
sleep) it's most likely (a) an ACPI/PCIe device that has powered up or not
respecting low-power limits or (b) some ACPI code that is misbehaving.
To find the culprit we need to:
(a) switch off/isolate some/most devices on boot time (eg: camera, nvme,
Bluetooth etc) and see which ones are triggering power drain (and then
write an additional piece of ACPI code to properly shut down/low-power them)
(b) debug ACPI routines (using kernel+ACPI debugging) pre-/post-
hibernation to see which specific procedures/methods are buggy and throwing
errors post-S4. This is something I need to teach myself in the coming days
when I've got some spare time.
ᐧ
…On Fri, 29 Sept 2023 at 23:21, OpenMinded ***@***.***> wrote:
Great, thanks, put it back. I disabled the wifi in the BIOS and it looked
at first there was some difference, but a bit after not. Depends if this
Coconutbattery is accurate, it takes a while to change. Perhaps it has to
do with that linked issue, so the iGPU. I added igfxonln=1 back and
igfxfw=2 for testing , can't notice difference but wake is fast.
Then replaced igfxfw=2 with igfxnorpsc=1 .
I wonder if the issue is just the coconut measuring, like an average that
suddenly adds ?
I don't see any obvious issues when running Intel Power Gadget, it's a bit
of a mystery. Do you have concrete measurements that show the coco is not
nuts ?
The most Intel Power Gadget shows is that the CPU temperature seems to
have gone up and the PKG line seems to go up a bit but it does not seem to
add up to what Coconut shows.
I also noticed the voltage is reset to default when using voltageshift (
from command line ).
Another finding : 4 TB crucial only works with hibernate in SSD slot 2,
which I think is enough to conclude it may not work with any other drive in
SSD slot 1. I can image that if an EFI on a disk in slot 2 may work on a
disk in slot 1, I will probably not spend more time to test unless anybody
thinks that is really useful.
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMCZBFR3NKPK3JFFI35TX45CWZANCNFSM4R3ZR2FQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Finally fixed issue with power drain post hibernation. As suspected it was due to the 9500's ACPI logic powering the dGPU up again which has been fixed now thanks to a change in powering down logic and a new introductory transition to sleep state which powers up the GPU then shuts it down properly. All existing XPS 9500 hackintoshes would have had this problem post-sleep/wake cycle, but it wasn't obvious because, well, sleep or hibernation wasn't working until now. I've also changed the RTC device properties which means you don't need to add any rtx_exclude parameters anymore. You can get the latest changes from my github repo. Enjoy! |
|
hi, that's great! I tested it but I cannot get it working. I have Sonoma which worked with the previous, exchanged EFI and updated the Airportdriver version but I only see a quick Apple logo after verbose boot, then a black screen and after some time a reboot happens. Any idea what may cause this? Did you upgrade to latest BIOS given undervolting works with : |
|
Hmm that might be a Sonoma problem. Any chance you can re-image and try
Ventura again?
Am still on bios 1.15.1, works fine for all the things I need.
…On Thu, 5 Oct 2023 at 09:11, OpenMinded ***@***.***> wrote:
hi, that's great! I tested it but I cannot get it working. I have Sonoma
which worked with the previous, exchanged EFI and updated the Airportdriver
version but I only see a quick Apple logo after verbose boot, then a black
screen and after some time a reboot happens. Any idea what may cause this?
Did you upgrade to latest BIOS given undervolting works with :
setup_var_cv CpuSetup 0xda 0x01 0x00
?
Thanks for all your work, I hope mine will work too soon :).
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMCZKOM3RSKLB3NGOQ6TX5ZTUZAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZUHAZTIMJXGU3A>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
hi, I just reinstalled Ventura on a clean disk with the original working EFI - no problems. |
|
Could you change boot-args in EFI with the following pls: -v keepsyms=1 kext-dev-mode=1 -igfxblr hbfx-ahbm=3 igfxnorpsc=1 |
|
hi, thank you for the continued help, it's really appreciated! I have some stack traces, perhaps they help. I'm beginning to suspect BIOS version difference. I hope you can be persuaded to upgrade to see if it keeps working :). The traces : The last one I think was the wireless driver that was mismatched, that was fixed after updating it. |
|
Ok let's try and troubleshoot exactly where it happened. It's clearly not a bios version problem (otherwise the first commit wouldn't have worked for you), and we also need to make the S4 fix universal across all the XPS and Precision versions, irrespective of the configuration. A lot of further power and performance optimisation fixes have been committed since the first version so you'll get a lot more benefits by being on the same branch. Let's narrow down exactly which commit broke restore functionality. Could you download my second major version here (Download ZIP) and see if it works for you? |
|
oh, I see my post wasn't posted. Feedback: this boots but the screen shows a divison in 4 blocks : 2 big ones created by a line across the middle of the screen horizontally then another vertical line a few cm from the left. Interestingly, when the machine is put to sleep and then wake, sometimes the screen recovers. I do seem to have more power use than before sleep ( about 15 vs 30 ). |
|
I suspect I know what this quad bug is but just want to double check.
If/when it does appear, could you shut the lid immediately, count to 10 and
then open the lid again and see if the issue goes?
…On Tue, 10 Oct 2023 at 09:56, OpenMinded ***@***.***> wrote:
oh, I see my post wasn't posted. Feedback: this boots but the screen shows
a divison in 4 blocks : 2 big ones created by a line across the middle of
the screen horizontally then another vertical line a few cm from the left.
Interestingly, when the machine is put to sleep and then wake, sometimes
the screen recovers. I do seem to have more power use than before sleep (
about 15 vs 30 ).
By the way, small typo in sudo pmset hibernatemode 25 on your page.
Thanks again for all the help, I hope we can nail it down :)!
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMC6PE3UVYOJPWHULZTDX6UEVFAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ3TKNRRHEYQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
hi,
I tried that when the apple logo is visible and at login, but 10 seconds
does not seem to do the trick.
After login I get some garbles on top ( like it looks black and white and
vertical stripes ).
With the lid closed longer the garbles went and sometimes ( actually first
time I tried last time ) the screen gets normal.
Strangly at the moment it just keeps waking up although the settings from
pmset -g look good.
When I select sleep however, the screen went well after waking from
hibernation.
And again, use straight after wake was like 39 Watts, it went to 28 shortly
after, which is still more that around 11-14 initially.
I hope this helps, please let me know if I can test more.
Thank you again!
…On Tue, Oct 10, 2023 at 11:48 AM jkbuha ***@***.***> wrote:
I suspect I know what this quad bug is but just want to double check.
If/when it does appear, could you shut the lid immediately, count to 10
and
then open the lid again and see if the issue goes?
On Tue, 10 Oct 2023 at 09:56, OpenMinded ***@***.***> wrote:
> oh, I see my post wasn't posted. Feedback: this boots but the screen
shows
> a divison in 4 blocks : 2 big ones created by a line across the middle
of
> the screen horizontally then another vertical line a few cm from the
left.
> Interestingly, when the machine is put to sleep and then wake, sometimes
> the screen recovers. I do seem to have more power use than before sleep
(
> about 15 vs 30 ).
> By the way, small typo in sudo pmset hibernatemode 25 on your page.
> Thanks again for all the help, I hope we can nail it down :)!
>
> —
> Reply to this email directly, view it on GitHub
> <
#3 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAJMMC6PE3UVYOJPWHULZTDX6UEVFAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ3TKNRRHEYQ>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAGB3COFDBOV7JYLQWFA3TLX6UKW5AVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ4DINBRGQZQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
Could you try swapping out the deviceproperty settings for the iGPU with
the first version of my EFI? This might just be a simple device-id and/or
ig-platform-id fix.
Specifically use these properties:
<key>PciRoot(0x0)/Pci(0x2,0x0)</key>
<dict>
<key>AAPL,ig-platform-id</key>
<data>
CQClPg==
</data>
<key>device-id</key>
<data>
mz4AAA==
</data>
<key>enable-dpcd-max-link-rate-fix</key>
<data>
AQAAAA==
</data>
<key>enable-hdmi-dividers-fix</key>
<data>
AQAAAA==
</data>
<key>enable-hdmi20</key>
<data>
AQAAAA==
</data>
<key>enable-lspcon-support</key>
<data>
AQAAAA==
</data>
<key>enable-max-pixel-clock-override</key>
<data>
AQAAAA==
</data>
<key>framebuffer-con1-alldata</key>
<data>
AQUJAAAEAACHAQAA
</data>
<key>framebuffer-con1-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-con2-alldata</key>
<data>
AgYJAAAEAACHAQAA
</data>
<key>framebuffer-con2-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-con3-alldata</key>
<data>
AwQKAAAIAACHAQAA
</data>
<key>framebuffer-con3-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-con3-has-lspcon</key>
<data>
AQAAAA==
</data>
<key>framebuffer-patch-enable</key>
<data>
AQAAAA==
</data>
<key>framebuffer-portcount</key>
<data>
BAAAAA==
</data>
<key>framebuffer-unifiedmem</key>
<data>
AAAAgA==
</data>
</dict>
…On Tue, 10 Oct 2023 at 11:07, OpenMinded ***@***.***> wrote:
hi,
I tried that when the apple logo is visible and at login, but 10 seconds
does not seem to do the trick.
After login I get some garbles on top ( like it looks black and white and
vertical stripes ).
With the lid closed longer the garbles went and sometimes ( actually first
time I tried last time ) the screen gets normal.
Strangly at the moment it just keeps waking up although the settings from
pmset -g look good.
When I select sleep however, the screen went well after waking from
hibernation.
And again, use straight after wake was like 39 Watts, it went to 28
shortly
after, which is still more that around 11-14 initially.
I hope this helps, please let me know if I can test more.
Thank you again!
On Tue, Oct 10, 2023 at 11:48 AM jkbuha ***@***.***> wrote:
> I suspect I know what this quad bug is but just want to double check.
> If/when it does appear, could you shut the lid immediately, count to 10
> and
> then open the lid again and see if the issue goes?
>
> On Tue, 10 Oct 2023 at 09:56, OpenMinded ***@***.***> wrote:
>
> > oh, I see my post wasn't posted. Feedback: this boots but the screen
> shows
> > a divison in 4 blocks : 2 big ones created by a line across the middle
> of
> > the screen horizontally then another vertical line a few cm from the
> left.
> > Interestingly, when the machine is put to sleep and then wake,
sometimes
> > the screen recovers. I do seem to have more power use than before
sleep
> (
> > about 15 vs 30 ).
> > By the way, small typo in sudo pmset hibernatemode 25 on your page.
> > Thanks again for all the help, I hope we can nail it down :)!
> >
> > —
> > Reply to this email directly, view it on GitHub
> > <
>
#3 (comment)>,
>
> > or unsubscribe
> > <
>
https://github.com/notifications/unsubscribe-auth/AAJMMC6PE3UVYOJPWHULZTDX6UEVFAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ3TKNRRHEYQ>
>
> > .
> > You are receiving this because you were mentioned.Message ID:
> > ***@***.***>
> >
>
> —
> Reply to this email directly, view it on GitHub
> <
#3 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAGB3COFDBOV7JYLQWFA3TLX6UKW5AVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ4DINBRGQZQ>
> .
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMC4ITCIXKP2MLSYL22LX6UM63AVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGQ4DONBTHEZA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
I have tried to make this the same as the one that worked earlier : PciRoot(0x0)/Pci(0x2,0x0) - I took various steps. |
|
Yes now just use my latest full EFI update from github (comment out
dpcd-max-link-rate) and then see if the DGPU is correctly powered down post
hibernate.
If it's still consuming high power then the dgpu is on a different ACPI
path. Just post the actual device path ( you can get it from windows or
hackintool) and I'll create a new SSDT for you.
…On Wed, 11 Oct 2023 at 08:08, OpenMinded ***@***.***> wrote:
I have tried to make this the same as the one that worked earlier :
PciRoot(0x0)/Pci(0x2,0x0) - I took various steps.
When I eventually commented : dpcd-max-link-rate like #dpcd-max-link-rate
it worked properly.
Regarding power use I pretty much see the same as before : at boot about
31, going down to about 17. Then after hibernate wake currently around 45 ,
but not going down a lot, to about 30-35 Watts.
I will try to enable the other options again, but it didn't seem to affect
the power use.
If you have suggestions on what to investigate, please let me know.
And once again thank you very much for your continued support!
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMCYRNFRX5MCRXFXHT3LX6ZAVVAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZVGY4TQNRUGQZQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
pulled latest of your EFI, changed the option, upgraded itlwm and airport equivalent. |
|
By mistake I left my DSDT enabled (to debug ACPI). Just double check that
it’s disabled in your config.plist.
If the acpi path doesn’t show in hackintool you’ll need to boot into
windows and extract from device manager.
…On Wed, 11 Oct 2023 at 12:46, OpenMinded ***@***.***> wrote:
pulled latest of your EFI, changed the option, upgraded itlwm and airport
equivalent.
It hangs at boot just after iGPU - PWRD something I see AppleALC too I
think then a hang with a reference to disk0s4 I believe.
I didn't see the pci id for the Nvidia card in hackintool, where can I
find it and which line in the config aims to block it?
thanks!
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMC7GWJLIYEO6FRWKXDLX62BKJANCNFSM4R3ZR2FQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Hi, I use yours, the DSDT.aml is disabled.
…On Wed, Oct 11, 2023, 13:53 jkbuha ***@***.***> wrote:
By mistake I left my DSDT enabled (to debug ACPI). Just double check that
it’s disabled in your config.plist.
If the acpi path doesn’t show in hackintool you’ll need to boot into
windows and extract from device manager.
On Wed, 11 Oct 2023 at 12:46, OpenMinded ***@***.***> wrote:
> pulled latest of your EFI, changed the option, upgraded itlwm and
airport
> equivalent.
> It hangs at boot just after iGPU - PWRD something I see AppleALC too I
> think then a hang with a reference to disk0s4 I believe.
> I didn't see the pci id for the Nvidia card in hackintool, where can I
> find it and which line in the config aims to block it?
> thanks!
>
> —
> Reply to this email directly, view it on GitHub
> <
#3 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAJMMC7GWJLIYEO6FRWKXDLX62BKJANCNFSM4R3ZR2FQ>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAGB3CIMKI3CT46TIKDFMZDX62CDTANCNFSM4R3ZR2FQ>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
my Nvidia pci path is pci bus 1, device 0, function 0 , so I suppose 01:00.0 . can't get the latest EFI to work |
|
At this point I'm unsure what's not working with your config, which is a shame as I've got the latest updates working well and have even upgraded to Ventura and hibernate/resume and low-power after wake is all working fine, avg power on idle = 3W. I still need to do some further aspm optimisation but will do so once you've managed to use my updated EFI as well. What I'd suggest is to upload your most up to date working EFI together with your DSDT.aml and ioreg file (or point me to your github repo if you have one) and I'll try and see if I can send a few patches your way. |
|
hi, thank you for hanging in there and all the help, greatly appreciated! That can be found here: This is probably the stacktrace from your latest commit, with as only change again the dpcd-max-link-rate commented: If there's anything more I can do, I am happy to. This was all running on Ventura 13.6. Glad you upgraded too, BIOS not yet? good night! |
|
Ok so we're finally getting somewhere :)
This is what you should do next:
1. I've just updated my latest EFI config on my github. This should make
hibernate work for you (but not the power drain on resume for hibernation -
yet). Just test it to see if hibernate works correctly for now.
2. I've looked at your ioreg and can't find the dGPU which means it's being
switched off at boot-up. However you need to provide me with the exact
path, and I suspect it's slightly different for you, so you need to boot to
Windows, open device manager, select the Nvidia/Quadro card, right click
into properties and under Details tab, scroll down to the Location Paths
field and copy down the strings you find there. (More info here:
https://dortania.github.io/Getting-Started-With-ACPI/Desktops/desktop-disable.html
)
Then just reply with the copy/paste of the location paths fields here :)
ᐧ
ᐧ
ᐧ
…On Thu, 12 Oct 2023 at 23:58, OpenMinded ***@***.***> wrote:
hi, thank you for hanging in there and all the help, greatly appreciated!
That sounds like an amazing result, can't wait to get there too :)!
I will try my best to facilitate. I forked your repo and create an archive
folder with working EFI folders.
The last working one is this commit :
https://github.com/jkbuha/XPS-9500-IceLake-OpenCore/tree/9f12bc3e757e8f8119e73b5dad95b20a54abd5c8
when I just disable dpcd-max-link-rate.
That can be found here:
https://github.com/VGerris/XPS-9500-IceLake-OpenCore/tree/main/EFI_archives/EFI.095-last-working-9f12bc3e757e8f8119e73b5dad95b20a54abd5c8
I also added the ioreg there.
This is probably the stacktrace from your latest commit, with as only
change again the dpcd-max-link-rate commented:
panic(cpu 6 caller 0xffffff80019b60f3): Kernel trap at 0xffffff7f967db65b, type 0=divide error, registers:
CR0: 0x000000008001003b, CR2: 0x0000000109f6d000, CR3: 0x0000001045a87002, CR4: 0x00000000003626e0
RAX: 0x01a7c9fd80000000, RBX: 0x01a7c9fd80000000, RCX: 0x0000000000006000, RDX: 0x0000000000000000
RSP: 0xffffffc13bea3280, RBP: 0xffffffc13bea32b0, RSI: 0x00000000000000a1, RDI: 0xffffff8002516a38
R8: 0x0000000000000000, R9: 0xffffff9089bf6000, R10: 0x0000000000000001, R11: 0x000000000000000a
R12: 0x000000002350d520, R13: 0x0000000000000000, R14: 0xffffffc13bea32c8, R15: 0x0000000009a7ec80
RFL: 0x0000000000010246, RIP: 0xffffff7f967db65b, CS: 0x0000000000000008, SS: 0x0000000000000010
Fault CR2: 0x0000000109f6d000, Error code: 0x0000000000000000, Fault CPU: 0x6, PL: 0, VF: 0
Panicked task 0xffffff8bbe484698: 10 threads: pid 146: WindowServer
Backtrace (CPU 6), panicked thread: 0xffffff8bbecdf0c8, Frame : Return Address
0xffffffc13bea2c60 : 0xffffff800187205d mach_kernel : _handle_debugger_trap + 0x4ad
0xffffffc13bea2cb0 : 0xffffff80019c6144 mach_kernel : _kdp_i386_trap + 0x114
0xffffffc13bea2cf0 : 0xffffff80019b5c57 mach_kernel : _kernel_trap + 0x3b7
0xffffffc13bea2d40 : 0xffffff8001812951 mach_kernel : _return_from_trap + 0xc1
0xffffffc13bea2d60 : 0xffffff800187233d mach_kernel : _DebuggerTrapWithState + 0x5d
0xffffffc13bea2e50 : 0xffffff80018719e7 mach_kernel : _panic_trap_to_debugger + 0x1a7
0xffffffc13bea2eb0 : 0xffffff8001fdb32b mach_kernel : _panic + 0x84
0xffffffc13bea2fa0 : 0xffffff80019b60f3 mach_kernel : _sync_iss_to_iks + 0x2c3
0xffffffc13bea3120 : 0xffffff80019b5ddd mach_kernel : _kernel_trap + 0x53d
0xffffffc13bea3170 : 0xffffff8001812951 mach_kernel : _return_from_trap + 0xc1
0xffffffc13bea3190 : 0xffffff7f967db65b com.apple.driver.AppleIntelCFLGraphicsFramebuffer : __ZN31AppleIntelFramebufferController14SetupDPTimingsEP21AppleIntelFramebufferP21AppleIntelDisplayPathPNS_10CRTCParamsE + 0xf5
0xffffffc13bea32b0 : 0xffffff7f967d8afc com.apple.driver.AppleIntelCFLGraphicsFramebuffer : __ZN31AppleIntelFramebufferController9hwSetModeEP21AppleIntelFramebufferP21AppleIntelDisplayPathPK29IODetailedTimingInformationV2 + 0x272
0xffffffc13bea3720 : 0xffffff7f967ac5f5 com.apple.driver.AppleIntelCFLGraphicsFramebuffer : __ZN21AppleIntelFramebuffer14setDisplayModeEii + 0xf8d
0xffffffc13bea38e0 : 0xffffff7f975df7fc com.apple.iokit.IOGraphicsFamily : __ZN13IOFramebuffer16doSetDisplayModeEii + 0x128
0xffffffc13bea3970 : 0xffffff7f975df65f com.apple.iokit.IOGraphicsFamily : __ZN13IOFramebuffer17extSetDisplayModeEP8OSObjectPvP25IOExternalMethodArguments + 0x97
0xffffffc13bea3a00 : 0xffffff7f975e6034 com.apple.iokit.IOGraphicsFamily : __ZN23IOFramebufferUserClient14externalMethodEjP25IOExternalMethodArgumentsP24IOExternalMethodDispatchP8OSObjectPv + 0x82
0xffffffc13bea3a50 : 0xffffff8001f526de mach_kernel : _is_io_connect_method + 0x37e
0xffffffc13bea3bb0 : 0xffffff80019728e9 mach_kernel : _iokit_server_routine + 0x3609
0xffffffc13bea3cc0 : 0xffffff800184c51a mach_kernel : _ipc_kmsg_send + 0x41a
0xffffffc13bea3d60 : 0xffffff8001863282 mach_kernel : _mach_msg_overwrite_trap + 0x592
0xffffffc13bea3dd0 : 0xffffff8001863907 mach_kernel : _mach_msg2_trap + 0x3c7
0xffffffc13bea3ef0 : 0xffffff800199b513 mach_kernel : _mach_call_munger64 + 0x1f3
0xffffffc13bea3fa0 : 0xffffff8001812db6 mach_kernel : _hndl_mach_scall64 + 0x16
Kernel Extensions in backtrace:
***@***.***>0xffffff7f975effff
dependency: ***@***.***>0xffffff80042ddfff
***@***.***>0xffffff7f96853fff
dependency: ***@***.***>0xffffff7f96507fff
dependency: ***@***.***>0xffffff8003e3dfff
dependency: ***@***.***>0xffffff7f9748afff
dependency: ***@***.***>0xffffff7f975effff
dependency: ***@***.***>0xffffff80042ddfff
dependency: ***@***.***>0xffffff80042f0fff
Process name corresponding to current thread (0xffffff8bbecdf0c8): WindowServer
Boot args: -v keepsyms=1 kext-dev-mode=1 igfxonln=1 igfxfw=2 -igfxblr hbfx-ahbm=3 acpi_layer=0x8 acpi_level=0x2 debug=0x100
Mac OS version:
22G120
Kernel version:
Darwin Kernel Version 22.6.0: Fri Sep 15 13:39:52 PDT 2023; root:xnu-8796.141.3.700.8~1/RELEASE_X86_64
Kernel UUID: F75FC53F-FC1A-3AB6-8980-EF66A83DD51D
roots installed: 0
KernelCache slide: 0x0000000001400000
KernelCache base: 0xffffff8001600000
Kernel slide: 0x00000000014dc000
Kernel text base: 0xffffff80016dc000
__HIB text base: 0xffffff8001500000
System model name: MacBookPro16,4 (Mac-A61BADE1FDAD7B05)
System shutdown begun: NO
Panic diags file available: YES (0x0)
Hibernation exit count: 0
System uptime in nanoseconds: 35299179102
Last Sleep: absolute base_tsc base_nano
Uptime : 0x0000000837febd1e
Sleep : 0x0000000000000000 0x0000000000000000 0x0000000000000000
Wake : 0x0000000000000000 0x0000002c3b09b228 0x0000000000000000
Compressor Info: 0% of compressed pages limit (OK) and 0% of segments limit (OK) with 0 swapfiles and OK swap space
Zone info:
Zone map: 0xffffff8085e4d000 - 0xffffffa085e4d000
. PGZ : 0xffffff8085e4d000 - 0xffffff808de4e000
. VM : 0xffffff808de4e000 - 0xffffff85597e7000
. RO : 0xffffff85597e7000 - 0xffffff86f2b1a000
. GEN0 : 0xffffff86f2b1a000 - 0xffffff8bbe4b3000
. GEN1 : 0xffffff8bbe4b3000 - 0xffffff9089e4c000
. GEN2 : 0xffffff9089e4c000 - 0xffffff95557e5000
. GEN3 : 0xffffff95557e5000 - 0xffffff9a2117f000
. DATA : 0xffffff9a2117f000 - 0xffffffa085e4d000
Metadata: 0xffffffa085e5d000 - 0xffffffa0a5e5d000
Bitmaps : 0xffffffa0a5e5d000 - 0xffffffa0b1e5d000
Extra : 0 - 0
If there's anything more I can do, I am happy to. This was all running on
Ventura 13.6. Glad you upgraded too, BIOS not yet? good night!
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMCZ65RSL4GFXNCJY25LX7BY3HANCNFSM4R3ZR2FQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
awesome! I hope it is what you say it is. If so, it should work on the XPS 9700 I suppose ? I haven't had time yet to test on that one, I need a working computer too :). |
|
We're going to need to troubleshoot this iteratively at an ACPI level. Ping
me on telegram (same handle as my github) and we'll get this resolved
quickly, and I'll then upload a fix for the Precision 5550 once it's
confirmed to work well on your machine.
ᐧ
…On Fri, 13 Oct 2023 at 01:09, OpenMinded ***@***.***> wrote:
awesome!
Pulled your EFI, same issue, hang at boot when almost there.
My paths are:
1 - same as in the linked guide :
https://dortania.github.io/Getting-Started-With-ACPI/assets/img/nvidia.672e8e16.png
( bios info )
2 - Location paths : PCIROOT(0)#PCI(0100)#PCI(0000)
ACPI(*SB*)#ACPI(PCI0)#ACPI(PEG0)#ACPI(PEGP)
I hope it is what you say it is. If so, it should work on the XPS 9700 I
suppose ? I haven't had time yet to test on that one, I need a working
computer too :).
cheers!
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJMMC2T5KV7YL73YFEU3GLX7CBCTANCNFSM4R3ZR2FQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
Hi, sounds awesome!
I installed telegram but it seems I need a phone number to connect.
You can reach me at my handle at ***@***.*** then we can use meet? Thank you!
…On Fri, Oct 13, 2023, 15:39 jkbuha ***@***.***> wrote:
We're going to need to troubleshoot this iteratively at an ACPI level.
Ping
me on telegram (same handle as my github) and we'll get this resolved
quickly, and I'll then upload a fix for the Precision 5550 once it's
confirmed to work well on your machine.
ᐧ
On Fri, 13 Oct 2023 at 01:09, OpenMinded ***@***.***> wrote:
> awesome!
> Pulled your EFI, same issue, hang at boot when almost there.
> My paths are:
> 1 - same as in the linked guide :
>
https://dortania.github.io/Getting-Started-With-ACPI/assets/img/nvidia.672e8e16.png
> ( bios info )
> 2 - Location paths : PCIROOT(0)#PCI(0100)#PCI(0000)
> ACPI(*SB*)#ACPI(PCI0)#ACPI(PEG0)#ACPI(PEGP)
>
> I hope it is what you say it is. If so, it should work on the XPS 9700 I
> suppose ? I haven't had time yet to test on that one, I need a working
> computer too :).
> cheers!
>
> —
> Reply to this email directly, view it on GitHub
> <
#3 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAJMMC2T5KV7YL73YFEU3GLX7CBCTANCNFSM4R3ZR2FQ>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAGB3COSEEW4CZJIFHJQUVDX7FAALAVCNFSM4R3ZR2F2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCNZWGE2TGNZWGM4Q>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
|
S4 Sleep (Hibernate+Resume) should now be fixed for all XPS 9500 and also Precision 5550 machines (thanks for confirming @VGerris!). Have pushed the latest changes to github... |
|
I can confirm the latest commit, which is : https://github.com/jkbuha/XPS-9500-IceLake-OpenCore/tree/878c183576497f4ab998601d0292f77fb383569b works like a charm! I understand more optimisations may come and who knows, perhaps S3. Big thanks and congratulations to @jkbuha for beating Dell to fixing their broken BIOS and giving us all a laptop that sucks less, or better yet is suddenly very useable. P.S I have a case open at Dell Pro support for the S3 issue and will keep chasing them, but it does not seem likely they will actually fix their crap, it has been 5 BIOS updates or more since without changes, even while they admitted the issue and promised a solution ( without ETA). I will probably spent some effort trying to get it to work if that does not work out, but for now I am very happy with the way it works with S4! |
|
Any tips on how to have S3 work on battery as well? I have Precision 5680 and it exhibits the same bug: S3 sleep works (after advertising is enabled with |
check out this to fix S3 sleep. rEFInd is a great tool, but AcpiPatcher.efi might also be able to be injected in the opencore/drivers folder if you don't want to use rEFInd. If you get a flickering screen or have the igpu lock to 350 mhz after sleep, add igfxfw=2 to your boot args to load apple's native GuC firmware. That also enables iGPU power management.
For the headphone jack, try adding verbstub.kext to your kexts folder. Jaromeyer has a folder with materials and instructions on his read.me on how to install combojack, which is a related utility. His solution worked for me.
Also, how efficient is S2 sleep? Please let me know, because I've had sleep related issues on my laptop and disabling S3 could be a solution.
Let me know if any of this works!
The text was updated successfully, but these errors were encountered: