[SOLVED] not-present page / kernel panic

sibidharan · 10-14-2023, 08:25 PM

I did enable slub_debug=F and made the computer work for sometime, but I was unable to boot live USB. A power failure trigged the behaviour said in the forum and adding maxcpus=1 in kernel parameter via grub booted the server, but only one core is useable. As said here I am able to enable other cores but that is causing some type of softlockups and hard lockups at random times, since thjs is a server with 32 cores and 128GB RAM running on intel i9 13th GEN.

Without maxcpus=1, the kernel panic happens during the initrd phase something to do with udev loading some modules, but its the same in any live USB, I tried Alma, Void and older kernels, all panics without maxcpus=1

I am unable to boildown to a kernelmodule that is causing this, but one thing is common,

The kmem_cache_alloc fails at 0x360/0x360 in both my server and in live USB on same server with same kernel. I removed SSDs, GPUs all possible things and ran live USB and same panic happens. With maxcpus=1 I am able to boot.

https://ibb.co/MZR0H1c

sibidharan · 10-14-2023, 08:27 PM

Quote:

Originally Posted by hazel

Here's a sample of the error taken from one of the photos. There are a lot of repetitions of this sort of thing.

Code:

[2.911260] kernel tried to execute NX-protected page - exploit attempt?
(uid: 0)
[2.911260] BUG: unable to handle page fault for address:
fffffe00000453a8
[2.911261] #PF: supervisor instruction fetch in kernel mode
[2.911261] #PF: error_code(0x0011) - permissions violation
[2.911262] PGD 87efc6067 P4D 87efc6067 PUD 87efc4067 PMD 87efc3067 PTE
000000085fc4d163
[2.911264] Thread overran stack, or stack corrupted
[2.911264] Oops: 0011:0xfffffc000000453a8

If I enable all performance and efficiency cores on i9-13990K after successfully booting with one core, this error happens.

sibidharan · 10-15-2023, 01:51 AM

I have an interesting update, since my CPU has performance and efficiency cores, if I just turn on all efficiency cores and have only one performance core, the system boots!

On a 24C system, I have only 17 cores as of now.
I believe maxcpus=1 and limiting CPU via BIOS both has same effect, but enabling all efficiency cores didn’t cause any crash. Weird. No Random crashes, once the computer boots, it works very well. But now the next restart is what I am worried. Since its a sevrer, it’s serving now.

Is there any chance the CPU has gone faulty or the motherboard since memtest passed successfully, so can rule out RAM, and this is only a boot time issue. I did removed all SSD and HDD and tried booting liveUSB and experienced similar issues, all are non tainted.

business_kid · 10-15-2023, 10:08 AM

Have you guys with the cpu problems gone searching the kernel bugs or reported one? Especially if you're running servers you want the answer to this. This now sounds like a bus timing issue. Linux isn't windows and this should not be happening.

sibidharan · 10-16-2023, 06:22 PM

https://answers.launchpad.net/ubuntu/+question/708200

Just asked here!! Lets see if something comes up!