From: acrn-users@... <acrn-users@...> On Behalf Of Alfonso Sanchez-Beato
Sent: Wednesday, January 8, 2020 7:39 PM
To: acrn-users@...
Subject: Re: [acrn-users] Cannot launch hart RT UOS when running Ubuntu as SOS

 

From: acrn-users@... <acrn-users@...> On Behalf Of Steve Osselton
Sent: Wednesday, January 8, 2020 8:32 PM
To: acrn-users@...
Subject: [acrn-users] Problem with Industry Scenario on a NUC7i7DNH

 

Hi,

Am essentially following the deployment instructions for the ACRN industry scenario
using a NUC7i7DNH with a serial port connected.

https://projectacrn.github.io/latest/getting-started/rt_industry.html

Flashing clear-31670-live-server.iso worked fine. Then followed the "Use the pre-installed industry ACRN hypervisor"
section and installed serial port support. Again this all worked fine and looked good.

I then rebooted the NUC. At HDMI console get "Intel NUC" banner with the following at top LHS of screen:

detected 8 processes, 8 enabled
current on process 0

And on serial port:

E0S0 S0

Not sure what has happened here. Am I missing something obvious ?

Thanks Steve

From: acrn-users@... <acrn-users@...> On Behalf Of alfonso.sanchez-beato@...
Sent: Wednesday, January 8, 2020 5:44 PM
To: acrn-users@...
Subject: [acrn-users] Cannot launch hart RT UOS when running Ubuntu as SOS

 

Hi,

I have set-up Ubuntu as service OS by following the instructions in https://projectacrn.github.io/latest/tutorials/using_ubuntu_as_sos.html. ACRN is recognized as hypervisor and all looks good in principle. However, after that I followed the instructions on how to start an RT UOS and things fail:

-----Original Message-----
From: acrn-users@... <acrn-users@...> On Behalf Of Geoffroy Van Cutsem
Sent: Thursday, November 28, 2019 3:23 PM
To: acrn-users@...
Subject: Re: [acrn-users] How to get the 2nd vCPU running in the Industry scenario?

On Thu, 2019-11-28 at 02:04 +0000, Liu, Fuzhong wrote:

Hi Tonny
You should copy
https://github.com/projectacrn/acrn-hypervisor/blob/master/devicemodel
/samples/nuc/launch_hard_rt_vm.sh for RTVM; Not maxcpus parameter in
it.
Not
https://github.com/projectacrn/acrn-hypervisor/blob/master/devicemodel
/samples/nuc/launch_uos.sh

Is there any value at all in having this maxcpus parameter *anywhere*?
My thinking is that it's ACRN that determines how many vCPUs are visible to any User VM, and we'd want our User VM to always use as many as are available, regardless of whether this is an RTVM or a standard User VM, wouldn't we?

Thanks,
Geoffroy

From: acrn-users@... <acrn-users@...> On Behalf Of Tzeng, Tonny
Sent: Thursday, November 28, 2019 9:21 AM
To: acrn-users@...
Subject: Re: [acrn-users] How to get the 2nd vCPU running in the Industry scenario?

 

Hi Fuzhong,

 

You are right! I just copied this comment line to my kernel command line without noticing it limits the vCPU to 1. Both vCPU are running now, though you may consider to remove this misleading parameter in the sample launch script. 😊

 

Regards,

Tonny

 

 

Hi Tonny

About the secondary vCPU is in “Init” mode after I launch  user VM. The HV console indicates:

 

ACRN:\>vcpu_list

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

I think you add maxcpus=1 in your kernel comline; Could you please share the user VM kernel comline?

 

ACRN:\>vcpu_list(without maxcpus=1)

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Running

 

ACRN:\>vcpu_list(with maxcpus=1)

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

 

BR.

Fuzhong

 

Hi Tonny,

This is a known issue. In current implementation, SOS will offline all vcpus except BSP before launch any user VM(i.e. offline pCPU1/2/3 in your case), and your launched user VM2 only init pCPU2/3 to run, this cause pCPU1 always be offlined until you launch user VM1.

In our new implementation, acrn-config tool will do a modification in launch script, that only pCPUs belong to the launched VM could be offlined, then it will resolve your issue.

or you can modify the launch script by manual:

  for i in `ls -d /sys/devices/system/cpu/cpu[1-99]`; do

+ # do not offline the CPU which not belong to the UOS

+ if [ "i" = "/sys/devices/system/cpu/cpu1" ]; then continue; fi

        online=`cat $i/online`

Hi,

 

It looks to me two cores are now assigned to the VM2 in the Industry scenario, but only the primary vCPU is shown in “Running” state, the secondary vCPU is in “Init” mode after I launch my user VM. The HV console indicates:

 

ACRN:\>vcpu_list

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

Is it a bug or how can I move the secondary vCPU to “Running” mode? Thanks.

 

Regards,

Tonny

From: acrn-users@... <acrn-users@...> On Behalf Of Tzeng, Tonny
Sent: Thursday, November 28, 2019 9:21 AM
To: acrn-users@...
Subject: Re: [acrn-users] How to get the 2nd vCPU running in the Industry scenario?

 

Hi Fuzhong,

 

You are right! I just copied this comment line to my kernel command line without noticing it limits the vCPU to 1. Both vCPU are running now, though you may consider to remove this misleading parameter in the sample launch script. 😊

 

Regards,

Tonny

 

 

Hi Tonny

About the secondary vCPU is in “Init” mode after I launch  user VM. The HV console indicates:

 

ACRN:\>vcpu_list

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

I think you add maxcpus=1 in your kernel comline; Could you please share the user VM kernel comline?

 

ACRN:\>vcpu_list(without maxcpus=1)

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Running

 

ACRN:\>vcpu_list(with maxcpus=1)

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

 

BR.

Fuzhong

 

Hi Tonny,

This is a known issue. In current implementation, SOS will offline all vcpus except BSP before launch any user VM(i.e. offline pCPU1/2/3 in your case), and your launched user VM2 only init pCPU2/3 to run, this cause pCPU1 always be offlined until you launch user VM1.

In our new implementation, acrn-config tool will do a modification in launch script, that only pCPUs belong to the launched VM could be offlined, then it will resolve your issue.

or you can modify the launch script by manual:

  for i in `ls -d /sys/devices/system/cpu/cpu[1-99]`; do

+ # do not offline the CPU which not belong to the UOS

+ if [ "i" = "/sys/devices/system/cpu/cpu1" ]; then continue; fi

        online=`cat $i/online`

Hi,

 

It looks to me two cores are now assigned to the VM2 in the Industry scenario, but only the primary vCPU is shown in “Running” state, the secondary vCPU is in “Init” mode after I launch my user VM. The HV console indicates:

 

ACRN:\>vcpu_list

 

VM ID    PCPU ID    VCPU ID    VCPU ROLE    VCPU STATE

=====    =======    =======    =========    ==========

  0         0          0       PRIMARY      Running

  2         2          0       PRIMARY      Running

  2         3          1       SECONDARY    Init

 

Is it a bug or how can I move the secondary vCPU to “Running” mode? Thanks.

 

Regards,

Tonny