Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 MonteRand (probably the start of an ongoing saga)
- Thread starter Syncretic
- Start date
- Sort by reaction score
You need to change the
With
MaxKernel values from <string>21.5.0</string> (MacOS 12.4 Ceiling) to <string>21.6.0</string> (MacOS 12.5 Ceiling) or better still, <string>21.99.99</string> (Any Monterey Ceiling) or even <string></string> (No Ceiling).With
<string>21.5.0</string>, it is not applied beyond MacOS 12.4.
I know the MaxKernel value has to be changed to 21.6.0 in order for it to work...but the question is whether the actual patch is still valid. Anyone heard from @Syncretic about this? He stated a few months ago that with each release of MacOS we would need to basically check and verify that the patch is still valid...because the thing being patched may have changed, and then bump the MaxKernel value also..
I just updated to 12.5 and made to here without a panic....and I didn't bump the MaxKernel yet until I get confirmation that the actual patch is still valid. But so far no panic. What are the specific scenarios I can test in MacOS to see if it still panics compared to 12.4?
I just updated to 12.5 and made to here without a panic....and I didn't bump the MaxKernel yet until I get confirmation that the actual patch is still valid. But so far no panic. What are the specific scenarios I can test in MacOS to see if it still panics compared to 12.4?
While noting what @Syncretic may have said, there is no real need to be excessively cautious with such patches as it is invariably that it is either applied or if the code has changed, it does not get applied.
Having said that, I can understand any developer advising what he has advised and would do exactly the same in his shoes. It can also reveal whether the patch is actually still needed or not.
Having said that, I can understand any developer advising what he has advised and would do exactly the same in his shoes. It can also reveal whether the patch is actually still needed or not.
Correct. Those with multiple parts such as AVXpel might indeed need a bit more care as it could be that in a new release, one part matches and another does not due to changes. This might or might not be significant depending on the details of what is happening.
So, yeah, any developer would just give a blanket instruction on keeping things on known working setup.
So, yeah, any developer would just give a blanket instruction on keeping things on known working setup.
Last edited:
what i noticed:
with syncretic patch with parameters 21.5.0, I have a kernel panic
without a patch with noAVX kext with parameters 21.5.0 - there is no panic,
but a lot of cloned mds processes appeared on all cores
with a patch with parameters 21.6.0 everything works as it should be
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro w5700, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, oc 0.8.2, mac os 12.5
with syncretic patch with parameters 21.5.0, I have a kernel panic
without a patch with noAVX kext with parameters 21.5.0 - there is no panic,
but a lot of cloned mds processes appeared on all cores
with a patch with parameters 21.6.0 everything works as it should be
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro w5700, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, oc 0.8.2, mac os 12.5
I successfully updated my MacPro 4,1 > 5,1 from Monterey 12.4 to 12.5 under OCLP 0.4.9.
I changed MaxKernel values from <string>21.5.0</string> (MacOS 12.4 Ceiling) to <string>21.6.0</string> in my config.plst
No issues at all. The progress bar during the update took about 10 minutes with GPU fans at high speed (XFX Radeon RX580 8GB).
After all, it rebooted with a stop in the progress bar -right in the middle- for about 1 min. It is the same stop I had every time I rebooted in 12.4
I changed MaxKernel values from <string>21.5.0</string> (MacOS 12.4 Ceiling) to <string>21.6.0</string> in my config.plst
No issues at all. The progress bar during the update took about 10 minutes with GPU fans at high speed (XFX Radeon RX580 8GB).
After all, it rebooted with a stop in the progress bar -right in the middle- for about 1 min. It is the same stop I had every time I rebooted in 12.4
Hi @Gustav Holdoff ! How are you? I'm experiencing many kernel panics when my mac pro is idle, or at least, in previous setups with Monterey (12.5) and Big Sur (11.6.8). Examining my new config.plist (I returned to Monterey 12.5 after a couple weeks trying Big Sur - using OpenCore 0.8.0) I see only NoAVXFSCompressionTypeZlib kext, but not AVXpel. Reading your post, I understood you use both, editing MaxKernel string? Comparing Monterey 12.5 to Big Sur 11.6.8, both ran with same instability, with one only difference: in Console app, when I was running big sur, there are many crash reports over mds process, while in Monterey the villain was WindowServer... On Big Sur, windowserver appear in many Spin reports only.
Thanks in advance!
As far as I know, Monterey 12.4 and then 12.5 implemented a different approach to working with the processor, with memory addresses and with apfs.
if you returned to bigsur and have a panic, then you probably have patches with wrong parameters or you didn’t set the kexts necessary for working with bigsur in the config
for 12.4 and 12.5 you have to use either a syncretic patch (with parameters that match the OS version) or noAVXFSC... kext - now for me these two different approaches work equally correctly - the main thing is to use the correct parameters
Experts on the forum know better, but I can describe my experience:
1. starting from 12.4, some programs that work with indexing - for example, I had an AVAST antivirus - can cause a panic
2. if instead of trim you did a shrink on your ssd, then macOS does not correctly understand what to do with the wrong disk volumes and programs that access the disk - for example: SIRI and other mds processes - can cause panic - I removed all the wrong partitions to avoid panic
3. to work with 12.5, I also tested 0.8.0 made by Martin Lo and OCLP (4.9 and 4.10) - it is also based on 0.8.0 - with the right configs it works without panic - BUT I REMOVED AVAST antivirus just in any case (I had a correspondence with AVAST developers - I let them know - hopefully they will release their patches soon and I will install it again). I myself use 0.8.3 (according to the cdf instructions - everything is described in detail and very clearly - and if you carefully follow the instructions, everything should work)
OpenCore on the Mac Pro
This guide explains how to use the excellent OpenCore boot loader to install, run, and update the latest version of macOS on the MacPro5,1, resulting in a clean, unaltered operating system just like on a supported Mac.
forums.macrumors.com
I think for many users, the reason for the problems is that they try to leave several systems on their macpro, but Mac does not provide for this in the original - each system now has a different principle of operation and you are forced to make a config that will satisfy all systems, which can probably lead to mistakes
I decided for myself that I don’t need to have several systems i use 12.5 due to capabilities of new GPU (rescue mojave is enough, and opencore is not needed for mojave)
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro w5700, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, oc 0.8.2, mac os 12.5
Last edited:
I'm getting KP's since updating to 12.5.1 (using avxpel). I don't know what is causing them because they always happen when I am away from my computer...possibly the screensaver or something? I have Mac configured never to sleep automatically.
I'm going to disable avxpel for a while now to see if the panics go away.
I'm going to disable avxpel for a while now to see if the panics go away.
This is just an educated guess, but I believe the panics will not occur when AVXpel is disabled.
The panic logs show that the panics are similar to those that were happening before @Syncretic came up with AVXpel.
What was happening previously was that they happened when the ZlibComp kext was used to process AVX. With AVXpel, the ZlibComp kext was patched and all was good.
Now, the panic happens when the AppleImageSecurity kext is used. This is used just before the ZlibComp kext is run (it was used this way previously as well).
My hypothesis is that following changes in the Mac OS 12.5.1 update, the AppleImageSecurity kext now panics on the ZlibComp kext failing a security check because of AVXpel patching.
If correct, the upshot is:
- You have AVXpel
- You get the SECURITY panic when some condition, most probably not AVX related, requires ZlibComp kext.
- This happens whether you actually need AVX patching or not
- If your use pattern does not trigger such a hit, all works well and you report no issues with AVXpel
- You do not have AVXpel
- You do not need AVX patching
- You DO NOT get a panic when some condition requires ZlibComp kext as it will pass the check.
- You need AVX patching
- Some condition requires ZlibComp kext but not AVX
- You DO NOT get a panic because ZlibComp passes the check and AVX is not needed
- Some condition requires ZlibComp kext and AVX
- You get the AVX panic as while ZlibComp passes the check, AVX is needed
- Some condition requires ZlibComp kext but not AVX
- You do not need AVX patching
If the hypothesis is correct, it could bypass the issues (assuming it doesn't also fail the security check).
For anyone using MyBootMgr, while you have choice between AVXpel Patches and the NoZlibAVX Kext in ConfigFactory, you will have noticed NoZlibAVX is the default highlighted option that the guide suggests should be used when not certain about the available options.
Last edited:
Thanks for that analysis and taking time to write it out so well. I will look into that other approach instead of avxpel. I’m not sure I even need avx. I am just trying to be proactive to prevent any panics. But it sounds like apple is starting to introduce kext security measures which may constrain how much we can patch kext rolling forward.
Does anyone know what are the currently known avx traps to avoid in usage?
Does anyone know what are the currently known avx traps to avoid in usage?
Anyone know where I can find instructions for how to use the above old version of the kext properly? Or if there is some newer approach I am not aware, where can I find info?
Take a look at the "May 24, 2022" revision of the OpenCore guide (see the WikiPost History).
Unlikely to make a difference if happening without AVX patches in play but no harm I suppose.
Are you using FileVault by any chance?
No I don't use FileVault
Unless 12.5.1 introduced some new use-case where the AVX problem is encountered and perhaps by using the older Kext I will get through it. We shall see.
Its possible that tsialex is right and I have to try resetting my battery...but I want to try this first.
Unless 12.5.1 introduced some new use-case where the AVX problem is encountered and perhaps by using the older Kext I will get through it. We shall see.
Its possible that tsialex is right and I have to try resetting my battery...but I want to try this first.
I have been running the legacy NoAVX kext for 4-5 days now and no KP so far. So that appears to be the right solution, at least for me, rather then avxpel.
UPDATE - it finally did KP again after 5 days. So problem still not solved for me. Will be trying a new BR2032 battery next.
UPDATE - it finally did KP again after 5 days. So problem still not solved for me. Will be trying a new BR2032 battery next.
Last edited:
Damn.. Put in a new BR2032 battery (left it out for 15 minutes to fully reset). Lasted a few days but eventually BAM...Kernel panic again...same as before.... on reboot -v it hung on decompression step also...
This time I was able to watch it happen. what happened is that first the beach ball spun for a few minutes, still able to move it around with the mouse, but unable to click on anything and get any response. I can't tell if this was being caused by Mail or which app might have caused it. I was not actually doing anything other then surfing around the the web. Then eventually the mac became unresponsive....and finally after several minutes...the screen went black, but still didn't crash or restart yet.. had to turn off the power to complete the KP process.
Running out of ideas now. This started happening with 12.5.1 and continues with 12.6
The next thing I will try but not today because I have too much to get done, is fresh install of MacOS 12.6...
if that doesn't work I will roll back to 12.5 and that will likely be the last version of MacOS used on this 5,1
This time I was able to watch it happen. what happened is that first the beach ball spun for a few minutes, still able to move it around with the mouse, but unable to click on anything and get any response. I can't tell if this was being caused by Mail or which app might have caused it. I was not actually doing anything other then surfing around the the web. Then eventually the mac became unresponsive....and finally after several minutes...the screen went black, but still didn't crash or restart yet.. had to turn off the power to complete the KP process.
Running out of ideas now. This started happening with 12.5.1 and continues with 12.6
The next thing I will try but not today because I have too much to get done, is fresh install of MacOS 12.6...
if that doesn't work I will roll back to 12.5 and that will likely be the last version of MacOS used on this 5,1
Last edited:
Yeah, I'm sitting on 12.5 for now. Maybe the smoke clears a bit down the road and I'll update, but for now I'm more than pleased with Monterey on my 5,1 at the 12.5 version.
It really seems like you have some hardware problem. I've gotta say I have used Martin's package, OCLP, and even tried rolling my own opencore config for awhile, on three different mac pros (2 4.1 to 5.1, 1 actual 5.1) all with completely different hardware inside and have never had these sorts of problems. I've been doing it since Catalina was released. =)
Never had an issue upgrading to 12.5.1, or 12.6 on two machines now. If you haven't identified a specific application that is causing it at a specific moment, I would be troubleshooting hardware by swapping in/out components.
The problem with the hardware failing around at at the same time of the update may have made it seem like it was an issue with the update??