[libvirt] virsh nodecpustats --cpu <invalid cpu no> is not failing error

[bugproxy]

== Comment: #0 - Satheesh Rajendran <email address hidden> - 2020-02-12 01:42:11 ==
---Problem Description---
virsh nodecpustats --cpu <invalid cpu no> is not failing error

Contact Information = <email address hidden>

---uname output---
5.4.0-13-generic

Machine Type = power9 ppc64le DD2.3

---Debugger---
A debugger is not configured

---Steps to Reproduce---
 1. Execute virsh nodecpustats with a invalid cpu number, ideally it should fail with error msg rather it returns nothing and commands succeeds.

# lscpu
Architecture: ppc64le
Byte Order: Little Endian
CPU(s): 160
On-line CPU(s) list: 0-159
Thread(s) per core: 4
Core(s) per socket: 20
Socket(s): 2
NUMA node(s): 2
Model: 2.3 (pvr 004e 1203)
Model name: POWER9, altivec supported
Frequency boost: enabled
CPU max MHz: 3800.0000
CPU min MHz: 2300.0000
L1d cache: 1.3 MiB
L1i cache: 1.3 MiB
L2 cache: 10 MiB
L3 cache: 200 MiB
NUMA node0 CPU(s): 0-79
NUMA node8 CPU(s): 80-159
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Mitigation; RFI Flush, L1D private per thread
Vulnerability Mds: Not affected
Vulnerability Meltdown: Mitigation; RFI Flush, L1D private per thread
Vulnerability Spec store bypass: Mitigation; Kernel entry/exit barrier (eieio)
Vulnerability Spectre v1: Mitigation; __user pointer sanitization, ori31 speculation barrier enabled
Vulnerability Spectre v2: Mitigation; Software count cache flush (hardware accelerated), Software link stack flush
Vulnerability Tsx async abort: Not affected

the system has total 160 cpus, any cpu number beyond should return error.

Actual:
# virsh nodecpustats --cpu 2000

# echo $?
0

Expected:
# virsh nodecpustats --cpu 2000
error: Unable to get node cpu stats
error: Invalid cpuNum in virHostCPUGetStatsLinux

Userspace tool common name: ii libvirt0:ppc64el 6.0.0-0ubuntu2 ppc64el library for interfacing with different virtualization systems

The userspace tool has the following bit modes: both

Userspace rpm: ii libvirt0:ppc64el 6.0.0-0ubuntu2 ppc64el library for interfacing with different virtualization systems

Userspace tool obtained from project website: na

*Additional Instructions for <email address hidden>:
-Post a private note with access information to the machine that the bug is occuring on.
-Attach ltrace and strace of userspace application.

== Comment: #3 - Satheesh Rajendran <email address hidden> - 2020-02-12 04:07:48 ==

== Comment: #8 - Mauro Rodrigues <email address hidden> - 2020-02-14 06:54:20 ==

== Comment: #9 - Mauro Rodrigues <email address hidden> - 2020-02-14 06:55:38 ==
It does happen upstream. Seeteena would you please test the attached patch?

Thing is that, even though virHostCPUGetStatsLinux was logging the invalid cpuNum error, it wasn't returning a failure (see the message below), so the upper layers correctly couldn't report as such.

2020-02-14 12:44:03.524+0000: 46611: error : virHostCPUGetStatsLinux:833 : Invalid cpuNum in virHostCPUGetStatsLinux

Another thing that bothers me, but considere I'm new to libvirt/qemu/kvm so it might just be my own incorrect understanding of it, is the fact that when we use "LIBVIRT_DEBUG=1 virsh cmd" we are just logging the command, not libvirt (ie libvirtd) execution.. The above message could be seen in libvirtd logs, but not from the command execution with LIBVIRT_DEBUG=1. So for me it's a bit misleading, but again it might just be my own mistake understanding this.

I recommend to always attach libvirtd logs as well in future issues.

== Comment: #13 - Mauro Rodrigues <email address hidden> - 2020-02-25 06:50:24 ==
FYI The patch, including a unit test, was sent to the libvirt mail list last friday. No feedback so far, unfortunately (which I think is fairly usual at libvirt ml).

https://www.redhat.com/archives/libvir-list/2020-February/msg00757.html

... which reads:
"util: virhostcpu: Fail when fetching CPU Stats for invalid cpu"

[bugproxy (bugproxy)]

Default Comment by Bridge

[bugproxy (bugproxy)]

Default Comment by Bridge

[Christian Ehrhardt  (paelzer)]

Fail:
root@f:~# virsh nodecpustats --cpu 47

root@f:~# echo $?
0

While older releases worked:
$ virsh nodecpustats --cpu 47
error: Unable to get node cpu stats
error: Invalid cpuNum in virHostCPUGetStatsLinux

Confirmed the issue, will create a test build with the upstream submission and try it ...

[Christian Ehrhardt  (paelzer)]

PPA: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3986
MP: https://code.launchpad.net/~paelzer/ubuntu/+source/libvirt/+git/libvirt/+merge/380942

[bugproxy (bugproxy)]

------- Comment From <email address hidden> 2020-03-23 09:34 EDT-------
(In reply to comment #19)
> PPA: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3986
> MP:
> https://code.launchpad.net/~paelzer/ubuntu/+source/libvirt/+git/libvirt/
> +merge/380942

Tested above PPA and found the reported issue is fixed.

~# uname -a
Linux xyz 5.4.0-18-generic #22-Ubuntu SMP Sat Mar 7 18:06:34 UTC 2020 ppc64le ppc64le ppc64le GNU/Linux

# dpkg -l|grep libvirt
ii libvirt-clients 6.0.0-0ubuntu6~ppa2 ppc64el Programs for the libvirt library
ii libvirt-daemon 6.0.0-0ubuntu6~ppa2 ppc64el Virtualization daemon
ii libvirt-daemon-driver-qemu 6.0.0-0ubuntu6~ppa2 ppc64el Virtualization daemon QEMU connection driver
ii libvirt-daemon-driver-storage-rbd 6.0.0-0ubuntu6~ppa2 ppc64el Virtualization daemon RBD storage driver
ii libvirt-daemon-system 6.0.0-0ubuntu6~ppa2 ppc64el Libvirt daemon configuration files
ii libvirt-daemon-system-systemd 6.0.0-0ubuntu6~ppa2 ppc64el Libvirt daemon configuration files (systemd)
ii libvirt-dev:ppc64el 6.0.0-0ubuntu6~ppa2 ppc64el development files for the libvirt library
ii libvirt-glib-1.0-0:ppc64el 3.0.0-1 ppc64el libvirt GLib and GObject mapping library
ii libvirt0:ppc64el 6.0.0-0ubuntu6~ppa2 ppc64el library for interfacing with different virtualization systems
ii python3-libvirt 6.0.0-0ubuntu3 ppc64el libvirt Python 3 bindings

# virsh nodecpustats --cpu 2000
error: Unable to get node cpu stats
error: Invalid cpuNum in virHostCPUGetStatsLinux

# echo $?
1

Regards,
-Satheesh

[Christian Ehrhardt  (paelzer)]

With the build from the PPA it works fine now.
I just need to make sure the other changes are good as well before an upload.

root@f-dcap-after-cap:~# virsh nodecpustats --cpu 47
error: Unable to get node cpu stats
error: Invalid cpuNum in virHostCPUGetStatsLinux

root@f-dcap-after-cap:~# echo $?
1

[Christian Ehrhardt  (paelzer)]

FYI I also replied on the upstream ML to add my test/review tags to get this going there.

[bugproxy (bugproxy)]

------- Comment From <email address hidden> 2020-03-24 09:49 EDT-------
Christian, thanks for your review and testing, it indeed helped to get things moving in the mail list. The patch is merged upstream:

https://libvirt.org/git/?p=libvirt.git;a=commit;h=75a4ec42f70b5324f95d7ffbbfbf7457620735e4

[Christian Ehrhardt  (paelzer)]

FYI Regression test is fine on this build, but waiting a bit to give it a chance to be upstream committed

[Launchpad Janitor (janitor)]

This bug was fixed in the package libvirt - 6.0.0-0ubuntu6

---------------
libvirt (6.0.0-0ubuntu6) focal; urgency=medium

  * d/p/ubuntu/lp-1867460-*: fix domcapabilities before capabilities
    and binary autodetection in general (LP: #1867460)
  * d/p/stable/lp-1868539-*: stabilize libvirt by backporting upstream
    fixes (LP: #1868539)
  * d/p/ubuntu/lp-1853200*: add cpu models without hle/rtm features to have
    modern types on kernels with recent security fixes (LP: #1853200)
  * d/p/ubuntu/lp-1868528-*: Fail when fetching CPU Status for invalid CPU
    (LP: #1868528)

 -- Christian Ehrhardt <email address hidden> Fri, 20 Mar 2020 10:34:19 +0100

[bugproxy (bugproxy)]

------- Comment From <email address hidden> 2020-05-11 15:09 EDT-------
Based on the previous comments, closing this one.