Upgrading 2013-2015 Macbook Pro SSD to M.2 NVMe

[tetet]

Has anybody tried patching the NVMe driver?

It might be the only way to get those drives with a lot of KPs to work?

https://github.com/RehabMan/patch-nvme

I believe @wawon tried to install kernel extensions before, but it did not work in MacBook. It was somewhere about page 15-20 in this thread.

I also think this is a promising direction. Any more insight, @wawon?

 

[daniele_zing]

Thank you Gilles for your fantastic work and for the analysis of connector pinout.
I'd like to add my two cents:

1) the devslp signal on the M.2 connector as far as I know (I found some interesting article) is used to
activate a DEVICE SLEEP mode on the disks ... they should turn off all what they can and reduce their consumption to a few milliwatts ... This means that on the M.2 connector this is an INPUT SIGNAL.
I don't know if the signal on the Apple connector is an OUT signal, and with the correct polarity, and if it was intended for the same scope.
If somebody could find more info on how the DEVSLP signal works on both the Apple and M.2 side we could better understand how to connect them. Probably the sleep/reboot problems are related to an incorrect design in this part.

2) I believe that we should pay more attention to POWER consumption of drives: the original Apple drives as far as I know consume about 6 Watt ... at 3.3V that's less than 2 amps. In the table that you created about your tests I see that some replacement drive goes to much higher power values, such as 8 or 9 ... In this case, although I believe that the Apple power supply is well designed and has some margin, it's not probable that it can deliver 50% more current .... and this could explain the kernel panics or other unstable behavior.
So we should prefer, for the sake of stability, the drives with amperage similar to the Apple ones, even if this means slightly lower speeds ...

 

[gilles_polysoft]

Hi again,
Has anybody tried patching the NVMe driver?

It might be the only way to get those drives with a lot of KPs to work?

https://github.com/RehabMan/patch-nvme

/// Flemming

Hi Flemming,

NVMe patches were only aimed at getting tiers NVMe SSD recognized on 10.11 and 10.12.
They are not intended to make the NVMe kext work better...

Once your are in 10.13 the patchs don't do anything and aren't needed... On 10.12 they are also unnecessary with 4K formatted drives.

Patches may allow to get 512B formatted SSD like the 960 Evo / Pro work under 10.12 : I've tried and didn't succeed but one may make it work !

Does the compatibility matrix for 512GB drives of gilles_polysoft also apply to the 1tb versions of those disks?

Yes, sure... only difference is that 1TB may have better write performances over 512 GB...

 

[zalg]

Hi,

I'm looking for 2TB nvme. Are there any models out there that have been tested?
I'm based in Japan but often travel to China
There's also the OWC Aura but not really convinced by their customer support when there were issues with the previous version and high sierra and I don't live in the US.

Thanks

 

[Marno]

->zalg: what macbook do you have ?

For MBpro 13" 2015, I now have all the pieces :
a KC1000 960GB, Kapton tape and... a black small NGFF adapter from Sintech.
I was on the edge to update the bootROM but noticed that mine was already : MBP121.0171.B00 (I bought this MBpro 2015 in summer 2017, maybe BootROM were already updated in production?).
Anyway I saved myself the hassle to update the bootROM to a potentially newer version, because I'm staying on Sierra for a while.
I will start the 'surgery' later and will come back to keep you posted (hopefully I won't have issues for sleep/reboot/strong IO).

 

[Audit13]

Should I install High Sierra onto an early 2014 13" MBP to ensure that the MBP has the latest bootrom even though my friend intends to go back to a clean install of Sierra? By doing this, it makes it easier to can jump to High Sierra with a 3rd party nvme drive in the future?

 

[zalg]

->zalg: what macbook do you have ?

I have a mbpro 15" late 2013

 

[gilles_polysoft]

Thank you Gilles for your fantastic work and for the analysis of connector pinout.
I'd like to add my two cents:

1) the devslp signal on the M.2 connector as far as I know (I found some interesting article) is used to
activate a DEVICE SLEEP mode on the disks ... they should turn off all what they can and reduce their consumption to a few milliwatts ... This means that on the M.2 connector this is an INPUT SIGNAL.
I don't know if the signal on the Apple connector is an OUT signal, and with the correct polarity, and if it was intended for the same scope.

If somebody could find more info on how the DEVSLP signal works on both the Apple and M.2 side we could better understand how to connect them. Probably the sleep/reboot problems are related to an incorrect design in this part.

Hi daniele_zing

thanks a lot for your 2 cents !

You're right about the wiring that could be the source of the sleep/reboot problems : I think that it is the case at least for reboot problems on the black sintech adapter (in my tests, the green adapter never has reboot problems, the black one has)

Also, I think the lack of connection on DEVSLEEP may be the cause of battery drain on hibernation.
If I have time I may try to read values of the DEVSLEEP pin on the apple motherboard...
and then mod a green sintech adapter and connect the DEVSLP between M.2 and apple connectors...

If I could, I suggest we can now identify/resume as follow the 5 problems with NVMe drives :

-problem #1 - sleep/standby on 2013/2014 macs : those mac aren't able to resume from deepsleep/hibernation with NVMe drives.This problem also occurs on the same mac with genuine Apple NVMe SSDs : this shows that this problem is not related to sintech adapter. Resolution of the problem : problem is solved by disabling hibernation, which is done by typing "sudo pmset standby 0"

- problem #2 - battery drain during sleep on 2013-2014 macs : the battery is drained more quickly on sleep with NVMe drives than with AHCI drives. This is a consequence of problem #1 : we need to disable hibernation, this results as the RAM being always powered on and bigger battery drain. Resolution of the problem : no solution available.

- problem #3 - battery drain during sleep on 2013-2017 macs : it seems the battery drain during sleep with NVMe drives is still higher with NVMe drives than AHCI drives even on 2015-2017 MBPr - MBA. This problem should be related to the DEVSLP wire not being connected between the Apple logic board and the M.2 SSD through the sintech (and other) adapters. Resolution of the problem : wiring DEVSLP wire ?

- problem #4 : reboot problem with Samsung NVMe drives with cablecc or black sintech adapters : connexion to the NVMe drive is lost at reboot, resulting in reboot failing with a blinking folder icon.
Resolution of the problem : use another brand than Samsung, or a large green adapter.


- problem #5 : kernel panics with Samsung SSD : kernel panics occur only with Samsung SSDs with huge i/o - files copy. This only occurs under macOS 10.13 (not linux or Windows), and randomly, but mostly on big transfers.
Resolution of the problem : wait for an updated IONVMe.kext - change Samsung SSD for other brand...

2) I believe that we should pay more attention to POWER consumption of drives: the original Apple drives as far as I know consume about 6 Watt ... at 3.3V that's less than 2 amps. In the table that you created about your tests I see that some replacement drive goes to much higher power values, such as 8 or 9 ... In this case, although I believe that the Apple power supply is well designed and has some margin, it's not probable that it can deliver 50% more current .... and this could explain the kernel panics or other unstable behavior.
So we should prefer, for the sake of stability, the drives with amperage similar to the Apple ones, even if this means slightly lower speeds ...

You are totally right to think about the power consumption, unfortunately, this doesn't seems to be the explanation of any problems here : the 960 Evo max power is rated at 6,04W which is exactly the same amount as Apple SMxxxL NVMe drives. Also, the same 960 Evo/Pro work perfectly under Linux/Windows on the same computers (rMBP).

To my knowledge, the kernel panic problem with samsung drives may be a software/driver problem... I don't think Apple did test any time its IONVMe.kext with OEM Samsung drives.. As of Apple SMxxxL SSDs, they are Samsung Polaris SSD but the firmware is customized to Apple needs, they have a 4K bloc size...

 

[gilles_polysoft]

Hi,

I'm looking for 2TB nvme. Are there any models out there that have been tested?
I'm based in Japan but often travel to China
Thanks

Hello zalg

I've tested so far two 2TB drives : the "genuine" Apple SM2048L, Being an apple genuine part it doesn't need any adapter and works natively... it is based samsung polaris SSD with custom firmware and formated in 4K) natively from 10.10.2 but is all but not cheap ($1300)

I've also tested the Samsung 960 Pro but as with Samsun 960 Evo I encountered with it some kernel panics on big i/o..

There is here a user (monofase) which reports not encountering any kernel panic with 960Evo (and large green Sintech adapter)..
I have also being reported by a german DJ twitter user that the 2TB Samsung pro works perfectly on his rMBP : https://twitter.com/me_ringo/status/954662455292170240

 

[daniele_zing]

[QUOTE="

-problem #1 - sleep/standby on 2013/2014 macs : those mac aren't able to resume from deepsleep/hibernation with NVMe drives......
[/QUOTE]

I have a retina MacBook Pro late 2013. I’ve been living with standby for 4 years ... its reliability is one of the best features of my mac. I never shut it down and reboot.
Now, that’s with the stock Apple sm256 drive. It’s clear that:
- the 12 +16 connector has a way to ask the drive to go low power
- the drive is well behaved, and always goes in and out of its low power state
- NVMe drives have also the possibility to be put in a low power state.
- I don’t see any reason to think that the pin for this is NOT the DevSlp pin on the M.2 connector
- if trying to connect this pin to the relative one in the Apple 12+16 connector we see that there are still problems, this would make me think about driver timing problems.... Maybe the NVMe drives react in a different way, or maybe they are too slow or too fast (with respect to the driver parameters which were modulated on the behaviour of the stock drives).

As far as I know one of the best features of the new interface is that it has lower latency: maybe the NVMe drives are too quick and their response is lost?

 

[gilles_polysoft]

hi,
i sadly started encountering a panics (which lead to immediate reboots) over heavy disk i/o with 1T 960 pro on mbp13 early 2015. reproducible. using sintech adapter.
not reproducible with linux or windows.

i removed it from 2014 mac mini. if i set it up as main drive - same panics. no panics if i use linux or windows. if i set it up as a secondary drive for my 2014 mac mini - it works perfectly fine on osx and no issues so far. linear speed is lower because interface has just two pci-e lines/lanes...

sadly had to revert back to original apple ssd on macbook pro;(

Hi damex,

please can you tell me if your boot drive was HFS+ or APFS formatted ?

I just gave my EVO 960 1TB a chance yesterday. Restored MacOs 10.13.3 from Time Machine backup and Restored Winclone Windows 10 backup. (750G / 250G)

Did the proper kaptan taping as gilles suggests (the part where the adapter mates the MBP connector)

So far only seen a KP during boot of MacOs.

Hi Flemming,

same question here, in your tests were you using an HFS+ or APFS formatted boot drive ?

 

[flemmingdjensen]

Hi damex,

please can you tell me if your boot drive was HFS+ or APFS formatted ?



Hi Flemming,

same question here, in your tests were you using an HFS+ or APFS formatted boot drive ?

Using APFS. Has been using this machine all evening without any KP!!! I think the kapton tape on top of the connector same way as the Sintech new adapter might have helped me.

Or is it the 10.13.3 update? This update rebooted several times where it might have updated EFI and things?

So right now I am pretty happy - only seen KPs when I completely reboot the computer

/// Flemming

 

[Tiberiu S]

Hi,

I saw that is a new adapter that is working for sleeping? Mine, on every sleep, it stopes. Do you have a link for the new adapter that is working well? Thanks.

 

[ravic]

Mine is APFS and I do have sleep issues with a Samsung P961 with a cablecc adapter.

 

[gilles_polysoft]

Hi everyone,

Little update with the test of a Toshiba / OCZ RD400. It's not a new drive, but it just works perfectly.
A few thoughts about this drive :
- it is the exact same drive as the Toshiba XG3
- MLC cells
- consumes only 6W.
- 4K bloc size capable so works from 10.12 and up
- quite expensive...

 

[gilles_polysoft]

I have done long tests with Samsung drives this evening, hoping I could figure it out..

I tested again the 960 Evo and PM981 (which is newer than the 960)

Tests were made with both macOS 10.13.4 (beta 0) and Linux Xubuntu 17.10 on a 13" rMBP late 2013, and with both black v.3 and large green v.2 sintech adapters.

Results with Linux and the 960 Evo/Pro :
- never had any hang or panic... Sleep works and resume properly. Reboot works and never present a blinking folder even on 2013-2014 mac, with both green and black adapters !
Copy of 400GB of datas can be made without trouble up to a fast external Samsung T3 , at a medium speed higher than 300 MB/sec


Results with macOS and the 960 Evo/Pro
- reboot fails with the black sintech adapter with a blinking icon
- reboot works 100% times without issue with the large green adapter
- huge data copy-I/O transfers work without issue from a NVMe APFS boot drive running 10.3.3
- huge data copy-i/o transfers randomly causes kernel panics while booted from a NVMe APFS drive under 10.3.2. Same thing while booted from an HFS+ drive (any version of High Sierra)


Results with Linux and the PM981 :
- no problem at all. Disk is fast, reboot works 100% of the time, sleep seems to work perfectly; no single issue

Results with macOS and the PM981 :
- disk impossible to use. Read works fast (1500MB/s) but write almost always fails ! (with both models of adapters).
Write on a APFS or HFS disk can write up to 100MB a 5 MB/sec only... after a while, I got spinning ball cursor and no transfer at all, then all this followed by a KP...


Conclusions so far :

- as for kernel panics issues : it sems macOS doesn't handle well the combinaison of Samsung NVMe drives formates in HFS+... (or Samsung NVMe drives formatted in APFS under 10.13.2 or lower)

- also macOS doesn't like the combinaison of both Samsung NVMe and Sintech black adapter... it results of a blinking icon folder



I have noticed no problem at all with any combinaison of linux / apple hardware / samsung NVMe drive / sintech adapter

Also all 2015-2017 macs are free from reboot problems so they can use a black sintech adapter with no issue !

And last, I feel quite optimistic, at least the fact that it works perfectly with linux/ windows makes me hope that the remaining issue (very noticeable with the PM981) is only with the macOS kext and the way macOS handles samsung NVMe drives..

 

[zalg]

Hello zalg

I've tested so far two 2TB drives : the "genuine" Apple SM2048L, Being an apple genuine part it doesn't need any adapter and works natively... it is based samsung polaris SSD with custom firmware and formated in 4K) natively from 10.10.2 but is all but not cheap ($1300)

I've also tested the Samsung 960 Pro but as with Samsun 960 Evo I encountered with it some kernel panics on big i/o..

There is here a user (monofase) which reports not encountering any kernel panic with 960Evo (and large green Sintech adapter)..
I have also being reported by a german DJ twitter user that the 2TB Samsung pro works perfectly on his rMBP : https://twitter.com/me_ringo/status/954662455292170240

Thanks Gilles!

Where were you able to get the genuine part SM2048L at that price? And was it new or extracted from a second hand mbp? It's not much more expensive than the Samsung pro or the Aura Pro X so would seem like a better solution.

I'm already complaining about kernel panics happening every once in a while since upgrading to Sierra (less now though since 10.12.5), don't really want to have more of them

 

[daniele_zing]

More about power....
I just stumbled upon some pictures of Samsung 960 Pro on Amazon ...
Take a look at the labels:
- 512gb and 1 tb are rated at 3.3v, 2.7 A
- 2 tb is rated at 3.3v, 1A !!!

If I remember well my original sm256 is about 1.9A.
So, if the label on the 2tb is correct, it looks like there’s a huge difference between different sizes.
Moreover, in general, we are well above the original amperage.
Now these numbers should be “peak” current, as the drives consume different amount of power for read and write. I have seen very different numbers in the specs, sometimes you see that somebody declares let’s say 6w for write, somebody else (like Toshiba) much lower numbers.

The problem is that if we have PEAK currents much higher than the original drives it’s like that the subsystem “sits down” in certain operations, like resuming from sleep, where most of the subsystem must be energized again.

I think that we should check if the drives that behave better have also the lower requirements of peak current.

 

[gilles_polysoft]

Thanks Gilles!

Where were you able to get the genuine part SM2048L at that price? And was it new or extracted from a second hand mbp? It's not much more expensive than the Samsung pro or the Aura Pro X so would seem like a better solution.

Hello zalg,
as an aasp I was able to order it directly from Apple. So I received it new, as a spare stock part for an iMac 5K 27".
You just need to be aware that Apple only gives a 3 month warranty on spare parts ordered as stock...

If you are interested you can find a lot of those drives on ebay for 1500$ : just search for "Apple 2TB polaris" or MZ-KKW2T00-0A6, of course it is grey market with no warranty at all.. but it works !

Sometimes there are bargains on ebay : I found a brand new 2TB 960 Pro for 700€ (800$)..
[doublepost=1517047381][/doublepost]

More about power....
I just stumbled upon some pictures of Samsung 960 Pro on Amazon ...
Take a look at the labels:
- 512gb and 1 tb are rated at 3.3v, 2.7 A
- 2 tb is rated at 3.3v, 1A !!!

[...]
I think that we should check if the drives that behave better have also the lower requirements of peak current.

I could have a look at each label, but as for the max power I have quoted for each SSD, it is the value I have read with the smartctl command under linux.

Also I don't know what to think of that : the Samsung PM981 I tested thouroughly yesterday do exhibits the same symptoms as other Samsung drives, only worse ! so this drive is to me the perfect test : if this drive works, all other Samsung will !
What is surprising is that all those problems only occurs with macOS and never with windows nor Linux... I do think that if it were a too much power requirement problem I would encounter the same exact problems under linux or windows, no ?

Thinking about that when I installed windows I did it without the help of Bootcamp assistant so I used Windows native nvme driver...
I may be curious to see if with bootcamp Apple provides its own NVMe drivers or let Windows use its own ?

Truly in my test this PM981 which is rated at more than 7W did perfectly behave with linux and windows : I litteraly made tons of duplications of 200 GB folders from the NVMe SSD to itself or to an external T3 ssd, did multiple reboots with the black adapter on a late 2013 mac... I tried everything to make it bug and it didn't failed once with linux nor windows...

But it did with High sierra...

 

[zalg]

Thanks again Gilles!

So, to get the Apple SM2048L, I'd have to ask an aasp to order it for me which I'm not sure they'd be willing to do.

To summarize my other options and make sure I understood correctly:
If I go the samsung 960 Pro route, since I have a late 2013 mbp, I'd have to get the green adapter from sintech (only v2 has been tested to work right?) and I'd have to format the disk in APFS.
The cheapest I found in Japan is 1350 usd, so I guess I'll have to wait for a nice offer on ebay if I go down this road .

In term of performance, they are both basically the same based on your benchmark (and should be since apple is using a customized version of the OEM version of the 960 pro).

 

[daniele_zing]

What is surprising is that all those problems only occurs with macOS and never with windows nor Linux... I do think that if it were a too much power requirement problem I would encounter the same exact problems under linux or windows, no ?

Yes, perfect. This leads me to think that:
- only the transient current is a problem (probably higher requests make the supply sit down and then rise to the nominal 3.3v
- the real problem is in timings: in windows/linux probably there are some delays that allow the situation to resume to normal (3.3v), while under OS X some timeout is hit.
If this is the case we can only hope in a revised driver/setup for NVMe drives from Apple ... or some hack from somebody who knows where to put the hands ....

 

[ravic]

I have done long tests with Samsung drives this evening, hoping I could figure it out..

I tested again the 960 Evo and PM981 (which is newer than the 960)

Tests were made with both macOS 10.13.4 (beta 0) and Linux Xubuntu 17.10 on a 13" rMBP late 2013, and with both black v.3 and large green v.2 sintech adapters.

Results with Linux and the 960 Evo/Pro :
- never had any hang or panic... Sleep works and resume properly. Reboot works and never present a blinking folder even on 2013-2014 mac, with both green and black adapters !
Copy of 400GB of datas can be made without trouble up to a fast external Samsung T3 , at a medium speed higher than 300 MB/sec


Results with macOS and the 960 Evo/Pro
- reboot fails with the black sintech adapter with a blinking icon
- reboot works 100% times without issue with the large green adapter
- huge data copy-I/O transfers work without issue from a NVMe APFS boot drive running 10.3.3
- huge data copy-i/o transfers randomly causes kernel panics while booted from a NVMe APFS drive under 10.3.2. Same thing while booted from an HFS+ drive (any version of High Sierra)


Results with Linux and the PM981 :
- no problem at all. Disk is fast, reboot works 100% of the time, sleep seems to work perfectly; no single issue

Results with macOS and the PM981 :
- disk impossible to use. Read works fast (1500MB/s) but write almost always fails ! (with both models of adapters).
Write on a APFS or HFS disk can write up to 100MB a 5 MB/sec only... after a while, I got spinning ball cursor and no transfer at all, then all this followed by a KP...


Conclusions so far :

- as for kernel panics issues : it sems macOS doesn't handle well the combinaison of Samsung NVMe drives formates in HFS+... (or Samsung NVMe drives formatted in APFS under 10.13.2 or lower)

- also macOS doesn't like the combinaison of both Samsung NVMe and Sintech black adapter... it results of a blinking icon folder



I have noticed no problem at all with any combinaison of linux / apple hardware / samsung NVMe drive / sintech adapter

Also all 2015-2017 macs are free from reboot problems so they can use a black sintech adapter with no issue !

And last, I feel quite optimistic, at least the fact that it works perfectly with linux/ windows makes me hope that the remaining issue (very noticeable with the PM981) is only with the macOS kext and the way macOS handles samsung NVMe drives..

Thanks a lot for all of this work. I am installing 10.13.3 right now to see how it improves my situation. I have 2015 MacBook Pro Retina 13" with a Samsung PM961 and a cablecc adapter.
I am still waiting for my sintech adapter. I am expecting to receive a black one.

Sleep issue is probably also related to big writes.

Edit: after installing 10.13.3 I still have issues with sleep (closing the lid). But it could be my cablecc(chenyang) adapter. Hopefully sintech adapter will help. I am using APFS.

 

[gilles_polysoft]

Thanks a lot for all of this work. I am installing 10.13.3 right now to see how it improves my situation. I have 2015 MacBook Pro Retina 13" with a Samsung PM961 and a cablecc adapter.
I am still waiting for my sintech adapter. I am expecting to receive a black one.

Sleep issue is probably also related to big writes.

Edit: after installing 10.13.3 I still have issues with sleep (closing the lid). But it could be my cablecc(chenyang) adapter. Hopefully sintech adapter will help. I am using APFS.

Hello ravic

the sleep problem with NVMe SSD concern every rMBP from late 2013 to late 2014 and every NVMe SSD, even Apple SSD NVMe ! All you have to do is to set standby to zero : type sudo pmset -a standby 0 in the terminal.

 

[ravic]

Hello ravic

the sleep problem with NVMe SSD concern every rMBP from late 2013 to late 2014 and every NVMe SSD, even Apple SSD NVMe ! All you have to do is to set standby to zero : type sudo pmset -a standby 0 in the terminal.

I set it that way, I still get kernel panics. I will try to post more details on the KP

Edit: Here are the KP details. Still researching what this is

Sat Jan 27 15:39:49 2018

*** Panic Report ***
panic(cpu 0 caller 0xffffff7f8eb52a3c): nvme: "Fatal error occurred. CSTS=0x3
. FW Revision=6L7QCXY7\n"@/BuildRoot/Library/Caches/com.apple.xbs/Sources/IONVMeFamily/IONVMeFamily-356.30.6/IONVMeController.cpp:5275
Backtrace (CPU 0), Frame : Return Address
0xffffff911e383a10 : 0xffffff800bc4f606
0xffffff911e383a60 : 0xffffff800bd7c654
0xffffff911e383aa0 : 0xffffff800bd6e149
0xffffff911e383b20 : 0xffffff800bc01120
0xffffff911e383b40 : 0xffffff800bc4f03c
0xffffff911e383c70 : 0xffffff800bc4edbc
0xffffff911e383cd0 : 0xffffff7f8eb52a3c
0xffffff911e383e30 : 0xffffff800c295ecc
0xffffff911e383ea0 : 0xffffff800c295df6
0xffffff911e383ed0 : 0xffffff800bc88ec4
0xffffff911e383f40 : 0xffffff800bc889b5
0xffffff911e383fa0 : 0xffffff800bc004f7
Kernel Extensions in backtrace:
com.apple.iokit.IONVMeFamily(2.1)[8C2350A2-A812-3B8C-8DE1-3A6911793417]@0xffffff7f8eb40000->0xffffff7f8eb7cfff
dependency: com.apple.driver.AppleMobileFileIntegrity(1.0.5)[498548AE-30A0-36C6-92F7-1714D01C0A98]@0xffffff7f8cbfb000
dependency: com.apple.iokit.IOPCIFamily(2.9)[8E6C654E-4A8F-3C6B-BBFE-BA8A68C9C146]@0xffffff7f8c494000
dependency: com.apple.driver.AppleEFINVRAM(2.1)[79C21A42-B04F-314F-BA88-F221F3D54094]@0xffffff7f8ca79000
dependency: com.apple.iokit.IOStorageFamily(2.1)[B0FCE898-1542-34C0-B845-46FAE81DD9E6]@0xffffff7f8c646000
dependency: com.apple.iokit.IOReportFamily(31)[D2F2FBDF-4EE4-38BA-99F5-B699F886F413]@0xffffff7f8d084000

BSD process name corresponding to current thread: kernel_task

Mac OS version:
17D47

Kernel version:
Darwin Kernel Version 17.4.0: Sun Dec 17 09:19:54 PST 2017; root:xnu-4570.41.2~1/RELEASE_X86_64
Kernel UUID: 18D901F1-4A03-3FF1-AE34-C26B2732F13C
Kernel slide: 0x000000000b800000
Kernel text base: 0xffffff800ba00000
__HIB text base: 0xffffff800b900000
System model name: MacBookPro12,1 (Mac-E43C1C25D4880AD6)

System uptime in nanoseconds: 42928190376
last loaded kext at 1932648810: com.apple.iokit.SCSITaskUserClient 404.30.2 (addr 0xffffff7f8e5ff000, size 45056)
loaded kexts:
com.apple.iokit.SCSITaskUserClient 404.30.2
com.apple.driver.AppleUSBStorageCoexistentDriver 439.30.4
com.apple.driver.AppleUSBCardReader 439.30.4
com.apple.iokit.IOBluetoothUSBDFU 6.0.2f2
com.apple.filesystems.apfs 748.41.3
com.apple.driver.AppleFileSystemDriver 3.0.1
com.apple.filesystems.hfs.kext 407.30.1
com.apple.AppleFSCompression.AppleFSCompressionTypeDataless 1.0.0d1
com.apple.BootCache 40
com.apple.AppleFSCompression.AppleFSCompressionTypeZlib 1.0.0
com.apple.driver.AppleTopCaseHIDEventDriver 127
com.apple.driver.AirPort.BrcmNIC 1220.28.1a3
com.apple.driver.AppleRTC 2.0
com.apple.driver.AppleSmartBatteryManager 161.0.0
com.apple.driver.AppleACPIButtons 6.1
com.apple.driver.AppleHPET 1.8
com.apple.driver.AppleSMBIOS 2.1
com.apple.driver.AppleACPIEC 6.1
com.apple.driver.AppleAPIC 1.7
com.apple.nke.applicationfirewall 183
com.apple.security.TMSafetyNet 8
com.apple.security.quarantine 3
com.apple.IOBufferCopyEngineTest 1
com.apple.iokit.IOUSBMassStorageClass 4.0.4
com.apple.driver.usb.IOUSBHostHIDDevice 1.2
com.apple.driver.usb.cdc 5.0.0
com.apple.driver.usb.networking 5.0.0
com.apple.driver.usb.AppleUSBHostCompositeDevice 1.2
com.apple.filesystems.hfs.encodings.kext 1
com.apple.driver.AppleUSBMergeNub 900.4.1
com.apple.driver.AppleThunderboltDPInAdapter 5.0.2
com.apple.driver.AppleThunderboltDPAdapterFamily 5.0.2
com.apple.driver.AppleThunderboltPCIDownAdapter 2.1.3
com.apple.driver.AppleActuatorDriver 1204.13
com.apple.driver.AppleHIDKeyboard 203
com.apple.driver.AppleHSBluetoothDriver 127
com.apple.driver.IOBluetoothHIDDriver 6.0.2f2
com.apple.iokit.IOBluetoothFamily 6.0.2f2
com.apple.driver.AppleMultitouchDriver 1204.13
com.apple.driver.AppleInputDeviceSupport 1201.8
com.apple.driver.AppleXsanScheme 3
com.apple.driver.AppleHSSPIHIDDriver 52
com.apple.iokit.IONVMeFamily 2.1.0
com.apple.driver.AppleThunderboltNHI 4.5.6
com.apple.iokit.IOThunderboltFamily 6.6.3
com.apple.iokit.IO80211Family 1200.12.2
com.apple.driver.mDNSOffloadUserClient 1.0.1b8
com.apple.driver.corecapture 1.0.4
com.apple.driver.AppleHSSPISupport 52
com.apple.driver.AppleIntelLpssSpiController 3.0.60
com.apple.driver.AppleIntelLpssGspi 3.0.60
com.apple.driver.AppleIntelLpssDmac 3.0.60
com.apple.driver.AppleIntelLpssI2C 3.0.60
com.apple.driver.usb.AppleUSBXHCIPCI 1.2
com.apple.driver.usb.AppleUSBXHCI 1.2
com.apple.driver.usb.AppleUSBHostPacketFilter 1.0
com.apple.iokit.IOUSBFamily 900.4.1
com.apple.driver.AppleUSBHostMergeProperties 1.2
com.apple.driver.AppleEFINVRAM 2.1
com.apple.driver.AppleEFIRuntime 2.1
com.apple.iokit.IOHIDFamily 2.0.0
com.apple.iokit.IOSMBusFamily 1.1
com.apple.security.sandbox 300.0
com.apple.kext.AppleMatch 1.0.0d1
com.apple.iokit.IOBufferCopyEngineFamily 1
com.apple.driver.DiskImages 480.30.2
com.apple.driver.AppleFDEKeyStore 28.30
com.apple.driver.AppleEffaceableStorage 1.0
com.apple.driver.AppleKeyStore 2
com.apple.driver.AppleUSBTDM 439.30.4
com.apple.driver.AppleMobileFileIntegrity 1.0.5
com.apple.iokit.IOUSBMassStorageDriver 140.30.1
com.apple.iokit.IOSCSIBlockCommandsDevice 404.30.2
com.apple.iokit.IOSCSIArchitectureModelFamily 404.30.2
com.apple.iokit.IOStorageFamily 2.1
com.apple.driver.AppleCredentialManager 1.0
com.apple.driver.KernelRelayHost 1
com.apple.iokit.IOUSBHostFamily 1.2
com.apple.driver.usb.AppleUSBCommon 1.0
com.apple.driver.AppleBusPowerController 1.0
com.apple.driver.AppleSEPManager 1.0.1
com.apple.driver.IOSlaveProcessor 1
com.apple.iokit.IOReportFamily 31
com.apple.iokit.IOTimeSyncFamily 650.5
com.apple.iokit.IONetworkingFamily 3.4
com.apple.driver.AppleACPIPlatform 6.1
com.apple.driver.AppleSMCRTC 1.0
com.apple.driver.AppleSMC 3.1.9
com.apple.iokit.IOPCIFamily 2.9
com.apple.iokit.IOACPIFamily 1.4
com.apple.kec.pthread 1
com.apple.kec.Libm 1
com.apple.kec.corecrypto 1.0

EOF

 

[flemmingdjensen]

Hi,

I set it that way, I still get kernel panics. I will try to post more details on the KP

Edit: Here are the KP details. Still researching what this is

Sat Jan 27 15:39:49 2018

*** Panic Report ***
panic(cpu 0 caller 0xffffff7f8eb52a3c): nvme: "Fatal error occurred. CSTS=0x3
. FW Revision=6L7QCXY7\n"@/BuildRoot/Library/Caches/com.apple.xbs/Sources/IONVMeFamily/IONVMeFamily-356.30.6/IONVMeController.cpp:5275

EOF

Yes the KP seems to always occur at this location: IONVMeController.cpp:5275

I have tried Disassembling the kext using Hopper Disassembler V4 and I find that there are 3 calls (or reasons) that this panic is called... See the attached image...

References to: 0x12936

• 0x1165d: FatalStateTimeoutEP18IOTimer
• 0x127e0: CommandTimeoutEP18IOTimer
• 0x12da0: HandleCompletionErrorsEP16AppleNVMeRequestj

/// Flemming