Re: Stalls when starting a VSOCK listening socket: soft lockups, RCU stalls, timeout

[Jiri Slaby <jirislaby@kernel.org>]

On 06. 03. 26, 10:57, Thomas Gleixner wrote:
> On Fri, Mar 06 2026 at 06:48, Jiri Slaby wrote:
>> On 05. 03. 26, 20:25, Thomas Gleixner wrote:
>>> Is there simple way to reproduce?
>>
>> Unfortunately not at all. To date, I even cannot reproduce locally, it
>> reproduces exclusively in opensuse build service (and github CI as per
>> Matthieu's report). I have a project in there with packages which fail
>> more often than others:
>>     https://build.opensuse.org/project/monitor/home:jirislaby:softlockup
>> But it's all green ATM.
>>
>> Builds of Go 1.24 and tests of rust 1.90 fail the most. The former even
>> takes only ~ 8 minutes, so it's not that intensive build at all. So the
>> reasons are unknown to me. At least, Go apparently uses threads for
>> building (unlike gcc/clang with forks/processes). Dunno about rust.
> 
> I tried with tons of test cases which stress test mmcid with threads and
> failed.

Me too, with many artificial pthread or fork or combined loads/bombs 
with loops and yields.

I was once successful to see the failure in a local build of Go: using 
"osc build --vm-type=kvm" which is what the build service (see below) calls.

It's extremely hard to hit it locally. So there is likely some rather 
small race window or whatnot.

> Can you provide me your .config

Sure, it's the standard openSUSE kernel, i.e.:
https://github.com/openSUSE/kernel-source/blob/9c1596772e0/config/x86_64/default

source version,

It happens with 6.19+, the current failures are with the commit above 
which is 6.19.5.

I added 7.0-rc2 as well now to:
   https://build.opensuse.org/project/monitor/home:jirislaby:softlockup

Well, it already failed for Go:
  
https://build.opensuse.org/package/live_build_log/home:jirislaby:softlockup/go1.24/617/x86_64

So at least it is consistent, and not stable tree related ;).


If that helps, I would be likely able to "bisect" the 4 your mm_cid 
patches if they can be reverted on the top of 6.19 easily. (By letting 
the kernel run in the build service.)

> VM setup (Number of CPUs, memory etc.)?

For example, the currently failing build:
https://build.opensuse.org/package/live_build_log/home:jirislaby:softlockup/rust1.90:test/openSUSE_Factory/x86_64

says:
[   10s] /usr/bin/qemu-kvm -nodefaults -no-reboot -nographic -vga none 
-cpu host -M pc,accel=kvm,usb=off,dump-guest-core=off,vmport=off 
-sandbox on -bios /usr/share/qemu/qboot.rom -object 
rng-random,filename=/dev/random,id=rng0 -device virtio-rng-pci,rng=rng0 
-object iothread,id=io0 -run-with user=qemu -net none -kernel 
/var/cache/obs/worker/root_4/.mount/boot/kernel -initrd 
/var/cache/obs/worker/root_4/.mount/boot/initrd -append 
root=/dev/disk/by-id/virtio-0 rootfstype=ext4 rootflags=noatime 
elevator=noop nmi_watchdog=0 rw ia32_emulation=1 oops=panic panic=1 
quiet console=hvc0 init=/.build/build -m 40960 -drive 
file=/var/cache/obs/worker/root_4/root,format=raw,if=none,id=disk,cache=unsafe,aio=io_uring 
-device virtio-blk-pci,iothread=io0,drive=disk,serial=0 -drive 
file=/var/cache/obs/worker/root_4/swap,format=raw,if=none,id=swap,cache=unsafe,aio=io_uring 
-device virtio-blk-pci,iothread=io0,drive=swap,serial=1 -device 
virtio-serial,max_ports=2 -device virtconsole,chardev=virtiocon0 
-chardev stdio,mux=on,id=virtiocon0 -mon chardev=virtiocon0 -chardev 
socket,id=monitor,server=on,wait=off,path=/var/cache/obs/worker/root_4/root.qemu/monitor 
-mon chardev=monitor,mode=readline -smp 12



The with-7.0-rc2 Go fail above runs:

[    4s] /usr/bin/qemu-kvm -nodefaults -no-reboot -nographic -vga none 
-cpu host -M pc,accel=kvm,usb=off,dump-guest-core=off,vmport=off 
-sandbox on -bios /usr/share/qemu/qboot.rom -object 
rng-random,filename=/dev/random,id=rng0 -device virtio-rng-pci,rng=rng0 
-object iothread,id=io0 -run-with user=qemu -net none -kernel 
/var/cache/obs/worker/root_12/.mount/boot/kernel -initrd 
/var/cache/obs/worker/root_12/.mount/boot/initrd -append 
root=/dev/disk/by-id/virtio-0 rootfstype=ext4 rootflags=noatime 
elevator=noop nmi_watchdog=0 rw ia32_emulation=1 oops=panic panic=1 
quiet console=hvc0 init=/.build/build -m 16384 -drive 
file=/var/cache/obs/worker/root_12/root,format=raw,if=none,id=disk,cache=unsafe,aio=io_uring 
-device virtio-blk-pci,iothread=io0,drive=disk,serial=0 -drive 
file=/var/cache/obs/worker/root_12/swap,format=raw,if=none,id=swap,cache=unsafe,aio=io_uring 
-device virtio-blk-pci,iothread=io0,drive=swap,serial=1 -device 
virtio-serial,max_ports=2 -device virtconsole,chardev=virtiocon0 
-chardev stdio,mux=on,id=virtiocon0 -mon chardev=virtiocon0 -chardev 
socket,id=monitor,server=on,wait=off,path=/var/cache/obs/worker/root_12/root.qemu/monitor 
-mon chardev=monitor,mode=readline -smp 4



> I tried to find it on that github page Matthiue mentioned but I'm
> probably too stupid to navigate this clicky interface.

I haven't looked into the details of the github failure yet...

thanks,
-- 
js
suse labs