Lockzi Posted January 17, 2017 Share Posted January 17, 2017 @Lockzi My setup is running without HWPEnabler.kext in S/L/E with hibernate mode set to 0. I remember that goodwin_c mentioned at the previous thread that we do not need the kext since we can enable it at clover configurator and it is supported by Clover EFI (38XX ver and higher supports HWP if I am not wrong, have to check it at changelogs). So far I haven't encounter data corruption. You do need the kext if you want to be able to control/change the power setting. And exactly, if you do not use HWPEnabler.kext, but have HWPEnabled = true in config.plist that would mean that you enable HWPEnable before kernel loads which if my theory is correct - would not result in corruption. So if data corruption occurs. The system becomes useless? Or are you guys checking for data corruption? Sent from my Pixel XL using Tapatalk You will not be able to boot anymore if this occurs to your machine. It's a "you can't miss it" situation Link to comment Share on other sites More sharing options...
Deniryer Posted January 18, 2017 Share Posted January 18, 2017 According to wmchris post here either the clover option, or the kext. no need for both. There is no need to have both. Also I didn't found the kext to be attached at the repo of this tutorial, so I skipped it and so far my machine work with intel speed shift enabled via clover without corrupting my file system or data.Maybe the corruptions only occurs when you use both at the same time? Don't really know. Link to comment Share on other sites More sharing options...
tns1111 Posted January 18, 2017 Share Posted January 18, 2017 Thanks everybody for reporting about (possible) NVMe data corruption. i list my "setup" here if it helps. i have no problems until now: As i closely followed the instructions of this thread i think - i don't have HWPEnabler.kext installed (i cant find it in system EFI) - My config.plist says HWPEnabled = true - i did pmset -a hibernatemode 0 sleep and wake up is fine do you guys keep a clone (clonezilla?) of your system for the case of data corruption? i also have a windows10 partition. will it also be affected? thanks! EDIT: My SSD: THNSN5512GPUK NVMe TOSHIBA 512GB Link to comment Share on other sites More sharing options...
albydnc Posted January 18, 2017 Share Posted January 18, 2017 Hi, is anyone working on the SD card Reader? I'm a photographer, it would be really useful for me. Link to comment Share on other sites More sharing options...
Lockzi Posted January 18, 2017 Share Posted January 18, 2017 Hi, is anyone working on the SD card Reader? I'm a photographer, it would be really useful for me. At This point (and probably ever) there is no known way to make the SD card reader working. According to wmchris post here There is no need to have both. Also I didn't found the kext to be attached at the repo of this tutorial, so I skipped it and so far my machine work with intel speed shift enabled via clover without corrupting my file system or data. Maybe the corruptions only occurs when you use both at the same time? Don't really know. Thanks for giving out your setup. To clarify, the reason for using both is that you would be able to alternate your HWP configuration (Power Saver/Maximum Performance). From my information in addition with: Thanks everybody for reporting about (possible) NVMe data corruption. i list my "setup" here if it helps. i have no problems until now: As i closely followed the instructions of this thread i think - i don't have HWPEnabler.kext installed (i cant find it in system EFI) - My config.plist says HWPEnabled = true - i did pmset -a hibernatemode 0 sleep and wake up is fine do you guys keep a clone (clonezilla?) of your system for the case of data corruption? i also have a windows10 partition. will it also be affected? thanks! In mind, my theory (as long as my system is not corrupted) is correct: 1. pmset hibernatemode 0 2. HWPEnabler (by only kext) causes corruption. (Would like more experience except only mine to confirm this.) 3. If you still want to have HWP, then you need to enable it through Clover (bootloader(before kernel loads)). 4. And if you enable it through Clover, you can as well enable the KEXT (as long as my system doesn't corrupt anytime soon, then I would have to eat my words). 1 Link to comment Share on other sites More sharing options...
Deniryer Posted January 18, 2017 Share Posted January 18, 2017 Hi, is anyone working on the SD card Reader? I'm a photographer, it would be really useful for me. You can only use sd card reader via usb 3.1 Type C hub like this. At This point (and probably ever) there is no known way to make the SD card reader working. Thanks for giving out your setup. To clarify, the reason for using both is that you would be able to alternate your HWP configuration (Power Saver/Maximum Performance). From my information in addition with: In mind, my theory (as long as my system is not corrupted) is correct: 1. pmset hibernatemode 0 2. HWPEnabler (by only kext) causes corruption. (Would like more experience except only mine to confirm this.) 3. If you still want to have HWP, then you need to enable it through Clover (bootloader(before kernel loads)). 4. And if you enable it through Clover, you can as well enable the KEXT (as long as my system doesn't corrupt anytime soon, then I would have to eat my words). Is possible to be related to type/manufacturer of the NVMe drive? In your signature I see that you have Toshiba SSD, I have Samsung PM961 (Polaris) 512 NVMe Drive and so far nothing happen to my system. I will backup my system today, install the kext and then if something happen I will share it here. Link to comment Share on other sites More sharing options...
mikolatero Posted January 19, 2017 Share Posted January 19, 2017 I have THNSN5512GPUK NVMe TOSHIBA 512GB too... ¿Maybe it's a problem with this nvme? Another question... ¿it's possible to install into a external drive via usb 3 instead internal nvme? Link to comment Share on other sites More sharing options...
Krim404 Posted January 19, 2017 Author Share Posted January 19, 2017 I have THNSN5512GPUK NVMe TOSHIBA 512GB too... ¿Maybe it's a problem with this nvme? Another question... ¿it's possible to install into a external drive via usb 3 instead internal nvme? ssd works and yes you can install on external drive. Link to comment Share on other sites More sharing options...
mdub Posted January 19, 2017 Share Posted January 19, 2017 Hi guys, I thought I'd chime in here as I'm trying to get my dell e5470 latitude working with sierra and have run into the same issues with HWP, SSDTprgen and the latest MacBook Pro smbios. I have an intel 6820hq cpu and MacBook Pro 13,2 is the only smbios that matches my CPU so naturally I'd like to get that working. I also get a KP with macbook 9,1 and many other smbios definitions. Like others I get a KP when booting with an ssdt generated by SSDTprgen using MacBook 13,2 smbios. I switched to HWPEnabler kext which seemed to work well at first but I got data corruption that reset my user account after a couple boots. This with was a fully patched DSDT/SSDT and hibernate mode off. I have a Samsung 850 pro ssd so the problem is not NVMe related. Questions: - is HWP retained after sleep if only using HWPEnable=true in clover (no kext). I could live without power profiles if the setting is definitely retained. How would I check this in AppleIntelInfo.kext? - what is the exact cause of the KP when using a generated SSDT? Is it an smbios serial number problem? Link to comment Share on other sites More sharing options...
mikolatero Posted January 20, 2017 Share Posted January 20, 2017 ssd works and yes you can install on external drive. I know that the ssd works, but maybe one nvme has more chances of corruption than others... We need to limit the problem from the base. Link to comment Share on other sites More sharing options...
Krim404 Posted January 20, 2017 Author Share Posted January 20, 2017 I know that the ssd works, but maybe one nvme has more chances of corruption than others... We need to limit the problem from the base think i said before i'm using the toshiba and never had one corruption so far. So i doubt it has something to do with the SSD itself. Hi guys, I thought I'd chime in here as I'm trying to get my dell e5470 latitude working with sierra and have run into the same issues with HWP, SSDTprgen and the latest MacBook Pro smbios. I have an intel 6820hq cpu and MacBook Pro 13,2 is the only smbios that matches my CPU so naturally I'd like to get that working. I also get a KP with macbook 9,1 and many other smbios definitions. Like others I get a KP when booting with an ssdt generated by SSDTprgen using MacBook 13,2 smbios. I switched to HWPEnabler kext which seemed to work well at first but I got data corruption that reset my user account after a couple boots. This with was a fully patched DSDT/SSDT and hibernate mode off. I have a Samsung 850 pro ssd so the problem is not NVMe related. Questions: - is HWP retained after sleep if only using HWPEnable=true in clover (no kext). I could live without power profiles if the setting is definitely retained. How would I check this in AppleIntelInfo.kext? - what is the exact cause of the KP when using a generated SSDT? Is it an smbios serial number problem? i used the HWP kext for months in the past with ElCap and had no corruption at all. Additionally i dont think that this simple command in the kext is able to make the SSD corrupt. If you want to use the kext see gymnaes 10.11 tutorial for the 9550. He has published a script for reloading HWP after sleep. I personally suggest using the clover flag -> much easier and from the same developer. they're are two different KP. 1. if you use a different SMBIOS you should always be sure to set it correctly and 2. if you use a skylake smbios (which provides cpu informations) you should not provide additional CPU informations as ssdt Link to comment Share on other sites More sharing options...
Lockzi Posted January 20, 2017 Share Posted January 20, 2017 think i said before i'm using the toshiba and never had one corruption so far. So i doubt it has something to do with the SSD itself. i used the HWP kext for months in the past with ElCap and had no corruption at all. Additionally i dont think that this simple command in the kext is able to make the SSD corrupt. If you want to use the kext see gymnaes 10.11 tutorial for the 9550. He has published a script for reloading HWP after sleep. I personally suggest using the clover flag -> much easier and from the same developer. When you used the HWP kext for months in ElCap, did you also have the Clover HWPEnabler = true? I believe it was a release in november of Clover that got the HWPEnabled feature introduced. Link to comment Share on other sites More sharing options...
goodwin_c Posted January 20, 2017 Share Posted January 20, 2017 When you used the HWP kext for months in ElCap, did you also have the Clover HWPEnabler = true? I believe it was a release in november of Clover that got the HWPEnabled feature introduced. This option in Clover was added by me as well. And it makes same basic thing as my kext. So if you have one of them, or both - there is zero difference. But kext can additionally specify performance profile - if you want to set something different from default one. Link to comment Share on other sites More sharing options...
parseInt32 Posted January 20, 2017 Share Posted January 20, 2017 What is a good replacement for the Killer 1535 Wi-Fi adapter? Link to comment Share on other sites More sharing options...
Lockzi Posted January 20, 2017 Share Posted January 20, 2017 This option in Clover was added by me as well. And it makes same basic thing as my kext. So if you have one of them, or both - there is zero difference. But kext can additionally specify performance profile - if you want to set something different from default one. Hello Goodwin_c! Nice to see you here as well, and thanks for your work for the community! Do you have any input to my theory below, that it has to be enabled before kernel loads to avoid risk of corruption? Sorry to hear that @mikolatero. We really need to find a permanent fix for these NVMe corruptions... I have had the same experience, although this latest install has been running for 2 weeks without a problem. sudo pmset -a hibernatemode 0 is a must. Previous guides has been saying that it can be set to 3 (darkhandz), but I've had corruptions with that setting. I initially "blamed" goodwin_c's HWPEnabler.kext, which I am certain has something to do with it, but on this installation I have both HWPEnabled = true in EFI/CLOVER/config.plist as well as the HWPEnabler.kext in L/E. My thoughts are that we can't be enabling HWP as late as KEXT loads, it needs to be before the kernel loads, i.e. in clover bootloader. To sum up my experience: Corruption 1: No change to pmset hibernatemode HWPEnabler.kext in L/E Corruption 2: pmset hibernatemode 3 HWPEnabler.kext in L/E Corruption 3: pmset hibernatemode 0 HWPEnabler.kext in L/E Current (2 weeks running setup without corruption (yet?) ): pmset hibernatemode 0 HWPEnabler.kext in L/E HWPEnabled = true in config.plist What was your setup regarding HWPEnabler? The more information we have, the more likely we can find a solution. Link to comment Share on other sites More sharing options...
goodwin_c Posted January 20, 2017 Share Posted January 20, 2017 Hello Goodwin_c! Nice to see you here as well, and thanks for your work for the community! Do you have any input to my theory below, that it has to be enabled before kernel loads to avoid risk of corruption? No, it can't affect anything - after enabling HWP - normal kernel power management will stop woking. But stop working means - that if kernel will try to change frequency directly though Speed Step - then CPU will just ignore this request. No errors, no segfaults, no side effects. Just CPU will be doing own job and ignoring all Speed Step requests from OS. Link to comment Share on other sites More sharing options...
Lockzi Posted January 20, 2017 Share Posted January 20, 2017 No, it can't affect anything - after enabling HWP - normal kernel power management will stop woking. But stop working means - that if kernel will try to change frequency directly though Speed Step - then CPU will just ignore this request. No errors, no segfaults, no side effects. Just CPU will be doing own job and ignoring all Speed Step requests from OS. Interesting! Do you have any input regarding NVMe corruption if we forget HWP for a moment? Link to comment Share on other sites More sharing options...
goodwin_c Posted January 20, 2017 Share Posted January 20, 2017 I'm using HWP starting from time when i created this kext. Dual boot, sometimes using Win10 (not very often), mostly MacOS. Always latest betas, a lot of reboots (development of system things needs a lot of them). A lot of freezes while experimenting with other things - this means a lot of unconditional reboots using 5-second-pressed power button. And still my data partition is ok, no corruption. And yes - i have one of highest configurations with 512Gb Toshiba SSD, 4K and i7. So far so good. So from my example you can see that there is no relation with HWP 2 Link to comment Share on other sites More sharing options...
mdub Posted January 20, 2017 Share Posted January 20, 2017 With HWP does X86platformplugin get loaded? Rehabman suggested it does not and therefore IGPU power management won't fully work. 1 Link to comment Share on other sites More sharing options...
tns1111 Posted January 21, 2017 Share Posted January 21, 2017 What is a good replacement for the Killer 1535 Wi-Fi adapter? hey, i had the same question here in this thread. see answer of wmchris. i'm also looking for a replacement. in the meantime i use this: "D-Link DWA-131 - Revision E" external USB-Wifi. D-Link provides official 10.12 drivers which work very well. Link to comment Share on other sites More sharing options...
joginek Posted January 21, 2017 Share Posted January 21, 2017 Hi wmchris, I just want to inform you that no devices work when plugged to TB3 port even after your latest config.plist, kexts and patched ACPI updates. Link to comment Share on other sites More sharing options...
AndersenTech Posted January 21, 2017 Share Posted January 21, 2017 Interesting note: when changing SMBIOS (both in Clover and kext) to MBP13,3, onboard HDMI and HDMI via the USB-C adapter stop working entirely. I can leave the kext alone (still MBP13,3) but change the Clover config back to MB9,1 and both ports start working again. Link to comment Share on other sites More sharing options...
Krim404 Posted January 22, 2017 Author Share Posted January 22, 2017 Hi wmchris, I just want to inform you that no devices work when plugged to TB3 port even after your latest config.plist, kexts and patched ACPI updates. these are not intended as final fixes. sys_cl told me TB3 would work by manually generating the SSDT/DSDT files with the correct refs.txt file in hand while deassembling. So maybe someone wants to give it a try. Tutorial is available in the ElCapitan Guide for the XPS15 from Gymnae. Interesting! Do you have any input regarding NVMe corruption if we forget HWP for a moment? it seems like the corruption occurs because of SSDT/DSDT differences on each device. Seems like on some setups its entering sleep state before writing is completed. Other setups are immune to this. possible solutions would be to modify the SSDT/DSDT to give it a lil bit more time or to force a flush on sleep/deep-sleep. so it has nothing to do with HWP. regarding the hynix devices: it seems like it's not really the device itself. i already gave a possible patch to one of my testers, i really hope i get a positive feedback :-) UPDATE: it works! HYNIX SSDs ARE NOW SUPPORTED! Tutorial is updated http://www.insanelymac.com/forum/topic/319929-pikeralphas-nvme-patch-with-hynix-m2-ssds/?do=findComment&comment=2354906 2 Link to comment Share on other sites More sharing options...
Krim404 Posted January 23, 2017 Author Share Posted January 23, 2017 Status update: OSX 10.12.3 works without any issue. No changes required. 1 Link to comment Share on other sites More sharing options...
atilla212 Posted January 24, 2017 Share Posted January 24, 2017 these are not intended as final fixes. sys_cl told me TB3 would work by manually generating the SSDT/DSDT files with the correct refs.txt file in hand while deassembling. So maybe someone wants to give it a try. Tutorial is available in the ElCapitan Guide for the XPS15 from Gymnae. The only tutorial I found online is at github by Gymnae and he mentions that tb3 is solved in 10.1 2, is there a link to his guide. I am willing to try. 1 Link to comment Share on other sites More sharing options...
Recommended Posts