cpu_topology_test: Add test case to test cpu topology #2337

vivianQizhu · 2020-07-29T08:06:44Z

id: 1856239
Signed-off-by: Qianqian Zhu qizhu@redhat.com

qemu/tests/cfg/cpu_topology_test.cfg

vivianQizhu · 2020-08-03T06:43:57Z

@nanliu-r Updated, please help review. @Kuhn-Chen @PaulYuuu Please help review as well if you are available, thanks.

qemu/tests/cpu_topology_test.py

nanliu-r · 2020-08-04T03:59:54Z

Hi, guest refuses to start for this error when the host has 4 CPUs, please help to check this and do we need to boot guest with 240 smp on every host?

/usr/libexec/qemu-kvm -smp 240,maxcpus=240,cores=5,threads=1,dies=1,sockets=48 -M q35
VNC server running on ::1:5900
qemu-kvm: current -smp configuration requires Extended Interrupt Mode enabled. You can add an IOMMU using: -device intel-iommu,intremap=on,eim=on

vivianQizhu · 2020-08-04T04:57:15Z

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

nanliu-r · 2020-08-04T06:08:12Z

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

Hi, seems we can't always get the right list, sometimes guest refuses to start for this:
-smp 70,maxcpus=70,cores=7,threads=1,dies=1,sockets=2
[qemu output] qemu-kvm: cpu topology: sockets (2) * dies (1) * cores (7) * threads (1) < smp_cpus (70)

please help to check this, thanks.

vivianQizhu · 2020-08-04T06:30:04Z

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

Hi, seems we can't always get the right list, sometimes guest refuses to start for this:
-smp 70,maxcpus=70,cores=7,threads=1,dies=1,sockets=2
[qemu output] qemu-kvm: cpu topology: sockets (2) * dies (1) * cores (7) * threads (1) < smp_cpus (70)

please help to check this, thanks.

Hi @nanliu-r Is this a Windows guest?

Kuhn-Chen · 2020-08-04T06:38:34Z

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

Hi, seems we can't always get the right list, sometimes guest refuses to start for this:
-smp 70,maxcpus=70,cores=7,threads=1,dies=1,sockets=2
[qemu output] qemu-kvm: cpu topology: sockets (2) * dies (1) * cores (7) * threads (1) < smp_cpus (70)

please help to check this, thanks.

According to this statement, the smp should not be greater than the (vcpu_cores * vcpu_threads * vcpu_sockets).
“ params['smp'] = params['vcpu_maxcpus'] = (vcpu_cores *
vcpu_threads * vcpu_sockets)“

nanliu-r · 2020-08-04T07:24:51Z

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

Hi, seems we can't always get the right list, sometimes guest refuses to start for this:
-smp 70,maxcpus=70,cores=7,threads=1,dies=1,sockets=2
[qemu output] qemu-kvm: cpu topology: sockets (2) * dies (1) * cores (7) * threads (1) < smp_cpus (70)
please help to check this, thanks.

Hi @nanliu-r Is this a Windows guest?

Yes

Thanks @nanliu-r Updated to vcpu_sockets = min(240 // (vcpu_cores * vcpu_threads), 10), what do you think?

Hi, seems we can't always get the right list, sometimes guest refuses to start for this:
-smp 70,maxcpus=70,cores=7,threads=1,dies=1,sockets=2
[qemu output] qemu-kvm: cpu topology: sockets (2) * dies (1) * cores (7) * threads (1) < smp_cpus (70)
please help to check this, thanks.

Hi @nanliu-r Is this a Windows guest?

Yes, this issue can reproduce with windows guest, and for rhel guest, I can met the issue output: 'qemu-kvm: current -smp configuration requires Extended Interrupt Mode enabled. You can add an IOMMU using: -device intel-iommu,intremap=on,eim=on' sometimes when boot guest with -smp 200,maxcpus=200,cores=10,threads=2,dies=1,sockets=10.

seems we suggest to overcommit host CPU number five times in one VM .@PaulYuuu does your case need a special machine as your comments "On ppc,the most extreme case is smp = 10 * 8 * 10, which is larger than vCPU limitation(240/384), please limit vcpu_cores and vcpu_sockets."

vivianQizhu · 2020-08-04T07:34:57Z

seems we suggest to overcommit host CPU number five times in one VM .@PaulYuuu does your case need a special machine as your comments "On ppc,the most extreme case is smp = 10 * 8 * 10, which is larger than vCPU limitation(240/384), please limit vcpu_cores and vcpu_sockets."

I think he means the extreme case from my previous code.

qemu/tests/cpu_topology_test.py

PaulYuuu · 2020-08-05T10:23:53Z

qemu/tests/cpu_topology_test.py

+        vcpu_sockets = min(max(host_cpu * 5 // (vcpu_cores * vcpu_threads), 1),
+                           random.randint(1, 10))


Some hosts have a lot of CPU, so there is a risk:

min(max(192 * 5 // (8 * 10), 1), random.randint(1, 10)) 10

And then, smp = vcpu_maxcpus = 8 * 10 * 10
I think using fixed sockets (2) is okay for this test case.

IIUC, AMD CPUs do not support multiple threads except EYPC. @nanliu-r please help to confirm.

Some hosts have a lot of CPU, so there is a risk:

min(max(192 * 5 // (8 * 10), 1), random.randint(1, 10)) 10

And then, smp = vcpu_maxcpus = 8 * 10 * 10
I think using fixed sockets (2) is okay for this test case.

@nanliu-r What do you think about this, to have a fixed sockets(2)?

nanliu-r · 2020-08-10T02:40:58Z

Hi, I still can meet the issue "[qemu output] qemu-kvm: cpu topology: sockets (8) * cores (2) * threads (1) < smp_cpus (32)" on EPYC (Milan) machine, both windows and rhel guest, can you help to check again?

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              256
On-line CPU(s) list: 0-255
Thread(s) per core:  2
Core(s) per socket:  64
Socket(s):           2
NUMA node(s):        2
Vendor ID:           AuthenticAMD
CPU family:          25
Model:               0
Model name:          AMD Eng Sample: 100-000000114-07_22/15_N
Stepping:            0

PaulYuuu · 2020-08-10T02:51:30Z

Hi, I still can meet the issue "[qemu output] qemu-kvm: cpu topology: sockets (8) * cores (2) * threads (1) < smp_cpus (32)" on EPYC (Milan) machine, both windows and rhel guest, can you help to check again?
Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              256
On-line CPU(s) list: 0-255
Thread(s) per core:  2
Core(s) per socket:  64
Socket(s):           2
NUMA node(s):        2
Vendor ID:           AuthenticAMD
CPU family:          25
Model:               0
Model name:          AMD Eng Sample: 100-000000114-07_22/15_N
Stepping:            0

This is the known issue of all AMD platforms. https://github.com/avocado-framework/avocado-vt/blob/666b5e5bb353272835c8d6bf618e11e8890fff3e/virttest/qemu_vm.py#L1532-L1540

nanliu-r · 2020-08-10T03:22:08Z

Hi, I still can meet the issue "[qemu output] qemu-kvm: cpu topology: sockets (8) * cores (2) * threads (1) < smp_cpus (32)" on EPYC (Milan) machine, both windows and rhel guest, can you help to check again?
Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              256
On-line CPU(s) list: 0-255
Thread(s) per core:  2
Core(s) per socket:  64
Socket(s):           2
NUMA node(s):        2
Vendor ID:           AuthenticAMD
CPU family:          25
Model:               0
Model name:          AMD Eng Sample: 100-000000114-07_22/15_N
Stepping:            0
This is the known issue of all AMD platforms. https://github.com/avocado-framework/avocado-vt/blob/666b5e5bb353272835c8d6bf618e11e8890fff3e/virttest/qemu_vm.py#L1532-L1540

@vivianQizhu Hi, since EPYC support multi-threads now can we do some changes in vt?

Besides, I still can reproduce this error on intel machine with windows guest with probability （6/10）

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              4
On-line CPU(s) list: 0-3
Thread(s) per core:  1
Core(s) per socket:  4
Socket(s):           1
NUMA node(s):        1
Vendor ID:           GenuineIntel
CPU family:          6
Model:               94
Model name:          Intel(R) Xeon(R) CPU E3-1225 v5 @ 3.30GHz
Stepping:            3

qemu/tests/cpu_topology_test.py

vivianQizhu · 2020-08-31T02:07:28Z

@nanliu-r Any more comments?

qemu/tests/cpu_topology_test.py

nanliu-r · 2020-09-01T09:26:00Z

qemu/tests/cpu_topology_test.py

+    if params['machine_type'] == 'pseries':
+        vcpu_threads_list = [1, 2, 4, 8]
+    params['vcpu_cores'] = vcpu_cores = random.randint(1, 10)
+    host_cpu = cpu.online_count()


Can we get the vcpu_cores according to the host_cpu to avoid CPU overcommit?
I mean if the host_cpu is less than 10 we can decrease the range?

@nanliu-r Sure, what do you expect? Same as host cores?

@nanliu-r Sure, what do you expect? Same as host cores?

Sounds good for me.

How about half of host_cpu considering of multiple threads?

@nanliu-r What if host cpu core is 2? Will that be too small? Let's try with =host cpu and see if it works well first, what do you think?

@nanliu-r Update to vcpu_cores = random.randint(1, max(6, host_cpu//2)). Please help review, thanks.

Shouldn't it be vcpu_cores = random.randint(1, min(6, host_cpu//2)) ?

@nanliu-r No, you said, when host_cpu<12(hostcpu//2 < 6), it should be 6, when host_cpu >=12(host_cpu//2 > 6), it should be host_cpu//2, so it should be max(6, host_cpu//2).

@vivianQizhu vcpu_cores = random.randint(1, min(6, host_cpu//2)) This is what I want to say at first.
If we use max(6, host_cpu//2) we maybe get a large maxcpus and we will hit the issue intel-iommu.

@nanliu-r Okay, updated.

Signed-off-by: Qianqian Zhu <qizhu@redhat.com>

nanliu-r · 2020-09-02T08:18:38Z

Test pass on AMD and Intel. Thanks.
(5/6) repeat3.Host_RHEL.m8.u3.product_rhel.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.8.3.0.x86_64.io-github-autotest-qemu.cpu_topology_test.q35: PASS (73.69 s)
(6/6) repeat3.Host_RHEL.m8.u3.product_rhel.qcow2.virtio_scsi.up.virtio_net.Guest.Win10.x86_64.io-github-autotest-qemu.cpu_topology_test.q35: PASS (201.28 s)
RESULTS : PASS 6 | ERROR 0 | FAIL 0 | SKIP 0 | WARN 0 | INTERRUPT 0 | CANCEL 0

Ack.

vivianQizhu · 2020-09-02T08:49:52Z

Thanks @nanliu-r @PaulYuuu @Kuhn-Chen

vivianQizhu force-pushed the cpu_topology branch 4 times, most recently from 252e902 to 6d9bf2e Compare July 30, 2020 07:02

Kuhn-Chen reviewed Jul 31, 2020

View reviewed changes

qemu/tests/cfg/cpu_topology_test.cfg Outdated Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch 2 times, most recently from 911e248 to 685688b Compare August 3, 2020 06:42

vivianQizhu force-pushed the cpu_topology branch from 685688b to 34eea44 Compare August 3, 2020 06:46

PaulYuuu reviewed Aug 3, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Outdated Show resolved Hide resolved

qemu/tests/cpu_topology_test.py Outdated Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch 2 times, most recently from 63760f2 to 5ea3681 Compare August 3, 2020 10:41

PaulYuuu approved these changes Aug 4, 2020

View reviewed changes

vivianQizhu force-pushed the cpu_topology branch from 5ea3681 to d5c7172 Compare August 4, 2020 04:56

vivianQizhu force-pushed the cpu_topology branch from d5c7172 to 5bd3a11 Compare August 4, 2020 06:31

vivianQizhu force-pushed the cpu_topology branch from 5bd3a11 to 2561706 Compare August 4, 2020 06:47

Kuhn-Chen reviewed Aug 4, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch from 2561706 to 52e3367 Compare August 4, 2020 08:18

PaulYuuu reviewed Aug 4, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Outdated Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch 4 times, most recently from ae06322 to b2fed6d Compare August 4, 2020 08:31

PaulYuuu reviewed Aug 5, 2020

View reviewed changes

vivianQizhu force-pushed the cpu_topology branch 4 times, most recently from b6a4148 to b48a505 Compare August 19, 2020 08:51

vivianQizhu commented Aug 19, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch from b48a505 to 8a48dee Compare August 20, 2020 08:01

nanliu-r reviewed Aug 20, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Show resolved Hide resolved

nanliu-r reviewed Aug 31, 2020

View reviewed changes

qemu/tests/cpu_topology_test.py Show resolved Hide resolved

vivianQizhu force-pushed the cpu_topology branch 2 times, most recently from e93b2e6 to f236198 Compare September 1, 2020 06:24

nanliu-r reviewed Sep 1, 2020

View reviewed changes

vivianQizhu force-pushed the cpu_topology branch 2 times, most recently from c5308b2 to 806d383 Compare September 2, 2020 05:35

cpu_topology_test: Add test case to test cpu topology

a1c1c56

Signed-off-by: Qianqian Zhu <qizhu@redhat.com>

vivianQizhu force-pushed the cpu_topology branch from 806d383 to a1c1c56 Compare September 2, 2020 06:46

vivianQizhu merged commit b269495 into autotest:master Sep 2, 2020

zhencliu added a commit to zhencliu/tp-qemu that referenced this pull request Sep 3, 2020

fixup! Merge pull request autotest#2337 from vivianQizhu/cpu_topology

00ec37e

vivianQizhu deleted the cpu_topology branch September 8, 2020 01:57

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

cpu_topology_test: Add test case to test cpu topology #2337

cpu_topology_test: Add test case to test cpu topology #2337

vivianQizhu commented Jul 29, 2020

vivianQizhu commented Aug 3, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

Kuhn-Chen commented Aug 4, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

PaulYuuu Aug 5, 2020

vivianQizhu Aug 19, 2020

nanliu-r commented Aug 10, 2020 •

edited

PaulYuuu commented Aug 10, 2020

nanliu-r commented Aug 10, 2020 •

edited

vivianQizhu commented Aug 31, 2020

nanliu-r Sep 1, 2020 •

edited

vivianQizhu Sep 1, 2020

nanliu-r Sep 1, 2020

nanliu-r Sep 1, 2020

vivianQizhu Sep 1, 2020

vivianQizhu Sep 2, 2020

nanliu-r Sep 2, 2020

vivianQizhu Sep 2, 2020

nanliu-r Sep 2, 2020 •

edited

vivianQizhu Sep 2, 2020

nanliu-r commented Sep 2, 2020

vivianQizhu commented Sep 2, 2020

		vcpu_sockets = min(max(host_cpu * 5 // (vcpu_cores * vcpu_threads), 1),
		random.randint(1, 10))

cpu_topology_test: Add test case to test cpu topology #2337

cpu_topology_test: Add test case to test cpu topology #2337

Conversation

vivianQizhu commented Jul 29, 2020

vivianQizhu commented Aug 3, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

Kuhn-Chen commented Aug 4, 2020

nanliu-r commented Aug 4, 2020

vivianQizhu commented Aug 4, 2020

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nanliu-r commented Aug 10, 2020 • edited

PaulYuuu commented Aug 10, 2020

nanliu-r commented Aug 10, 2020 • edited

vivianQizhu commented Aug 31, 2020

nanliu-r Sep 1, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nanliu-r Sep 2, 2020 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nanliu-r commented Sep 2, 2020

vivianQizhu commented Sep 2, 2020

nanliu-r commented Aug 10, 2020 •

edited

nanliu-r commented Aug 10, 2020 •

edited

nanliu-r Sep 1, 2020 •

edited

nanliu-r Sep 2, 2020 •

edited