vmware-workstation 17.5.1 issue with kernel 6.8.8-200.fc39 #243
Comments
|
|
Any ideas? |
|
How you installed the vmware-modules? |
|
I can confirm everything works perfectly with Guess I'll skip the kernel update for now. |
Yes. As I said, and as confirmed by @priiduneemre it works perfectly on 6.8.7, while I'm getting the above kernel oops on 6.8.8. Haven't seen any ill effects on 6.8.8-200.fc39 though. |
|
vmx86.c:
Ok, index 0 in an array is invalid? Why? Because at vmmon-only/include/x86msr.h, line 95 we have: In typical C this means an expandable array. But UBSAN did not like it. BTW: This is probably just a warning, I don't think it is crashing anything. |
|
I'll test this patch at home: Suggested by looking at https://gist.github.com/joanbm/9cd5fda1dcfab9a67b42cc6195b7b269 |
|
Hi, running on bare linux Fedora as of now as following Fedora Linux 40 (Workstation Edition) 6.8.8-300.fc40.x86_64 when i update to the 6.8.8 i also encountered some issue, instead of build and install it i had to replace the original tarballs and then build and install it. |
|
Hello, I confirm the same error on the following configuration: Fedora Linux 40 6.8.8-300.fc40.x86_64, VM seems to run fine regardless of the error in dmesg, but after shutting a VM down CPU load goes to almost 100%, which can be stopped with "modprobe -r vmmon". kernel-6.8.7-300.fc40.x86_64 was fine with modules from this repo. |
|
I confirm Fedora Linux 40, kernels 6.8.8 and 6.8.9 are affected. Worked fine prior to 6.8.8. https://discussion.fedoraproject.org/t/fedora-40-amd-gpu-kernel-6-8-8-issue-anyone-else/116646 |
|
I add the following output for this kernel: Linux clr-3a9b24c3451f4ded968a21702b4bc5de 6.8.9-1432.native #1 SMP Mon May 6 12:48:36 PDT 2024 x86_64 GNU/Linux |
|
|
I have similar messages in any kernel greater than 6.5.12 and it does cause problems. |
|
I was able to fix the warnings by deleting the zero in these two locations ./vmmon-only/include/x86cpuid.h typedef struct CPUIDQuery { ./vmmon-only/include/x86msr.h typedef struct MSRQuery { However if I boot to kernel 6.8.9 and bring up a VM in workstation 17.5.2 the host has lockup problems with these messages. [ 103.805075] /dev/vmmon[4226]: PTSC: initialized at 3499998000 Hz using TSC, TSCs are synchronized. |
|
My kernel (Fedora 40) just updated to 6.9.1, and I'm still getting the dev_base_lock error - won't compile. Any suggestions? Attempting to install 17.5.1 (uname -r = 6.9.1-200.fc40.x86_64). Further, 17.5.2 won't let me check it out via git checkout
|
|
Fixed my issue by disabling anything not needed in firmware setup and using these kernel modules https://github.com/mkubecek/vmware-host-modules/ I get UBSAN: array-index-out-of-bounds messages but no lockups and successfully upgraded to Fedora 40 on the host. |
|
It's works fine in Linux cr-lan 6.8.10-1434.native #1 SMP Fri May 17 04:09:08 PDT 2024 x86_64 GNU/Linux The virtual network editor works. The network in the VM works. But It does not compile in 6.9.1 |
|
Interesting, I can successfully make vmmon-only, but not vmnet-only...has everything to do with *_lock
|
|
The above two fixes from @dioni21 work fine for me on kernels > 6.8.7. I am running VMware Workstation 17.5.2 with 6.8.10-200.fc39 without any issues. `$ diff -autrB vmware-host-modules-workstation-17.5.1 vmware-host-modules-workstation-17.5.1-torel
Only in vmware-host-modules-workstation-17.5.1-torel/vmmon-only/include: x86msr.h.ORIG |
|
Patch works fine with the latest kernel, 6.8.11-200.fc39.x86_64. |
|
Still no luck on my end. Seems like all of the above patches are fixing vmmon, and I can complete make on that module. However, vmnet is NOT completing without the dev_base_lock error. Has anyone found a fix for that? |
|
dev_base_lock is defined in this file for me. /usr/src/kernels/$(uname -r)/include/linux/netdevice.h |
|
dev_base_lock has been removed as of kernel 6.9 |
Yes, you are on the way. 
|
|
The 17.5.1 module still tries to compile against it, so how do I solve for it? |
At this moment try to use 6.8.n kernel version. |
Sadly not an option right now - I'm running the Asus ROG 6.9 kernel for a thousand other fixes. :) |
So, you only have to wait to fix it then... |
And wait I will - this is a great community, thanks for the answers. |
|
FYI these modules worked on 6.9.1 |
|
Works for me with today's kernel upgrade to Fedora W40. Thanks |
Great. Thank you! |
|
I have a kernel crash that looks like your errors. The difference is i use VMware 17.6.0 and kernel 6.8.0-41-generic, i'm on Mint 22 based on Ubuntu 24.04 according to the vmware release notes, 24.04 should be supported. `[28720.724660] ------------[ cut here ]------------ [29247.035083] task:P2P_DISCOVER state:D stack:0 pid:3948 tgid:3098 ppid:1 flags:0x00000002 |
Same here. My system hangs after it's been idle for a while (cannot get past the lock screen, network drops, etc.). Not sure if it's related, but it started happening when I upgraded to 17.6. |
|
Same crash with the last kernel update 6.8.0-44 |
|
Same crash with 6.8.0-45. What the similarity with Version Mismatch error? I haven´t this error in the stack trace. |
|
I am on 22.04 and 24.04 with a 6.8.0 kernel and after this update, I have to disable vmware service on startup. It will freeze the machine and crash and cannot even reboot remotely so I guess that's pretty bad. |
|
I was on ubuntu 24.04 6.8 as well and having the same issues (machine freeze, error in the dmesg, cannot reboot etc.). Solved installing kernel 6.1 with https://github.com/bkw777/mainline. |
Similar to mkubecek#214, add akmod configuration/support to vmware-host-modules. This will ensure Fedora users are able to get a fresh build of vmware-host-modules each time the kernel updates. Spec configuration and installation based on https://github.com/Fred78290/nct6687d. Installation for RPM: ``` make akmod ``` Test/Dev flow: ``` make clean make akmod/clean && make akmod && sudo akmods sudo systemctl start vmware.service ``` Applied fix for newer kernel: mkubecek#239 Applied another fix for newer kernel: mkubecek#243 (comment)
Works perfectly with 6.8.7-200.fc39. VMware Wkst 17.5.1 still works.
The text was updated successfully, but these errors were encountered: