Add device address parameter for s390x by maxujun · Pull Request #4010 · avocado-framework/avocado-vt

maxujun · 2024-10-08T07:34:04Z

ID:1360

maxujun · 2024-10-08T07:35:30Z

(1/1) Host_RHEL.m8.u10.product_rhel.nographic.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.8.10.0.s390x.io-github-autotest-qemu.boot.s390-virtio: STARTED
(1/1) Host_RHEL.m8.u10.product_rhel.nographic.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.8.10.0.s390x.io-github-autotest-qemu.boot.s390-virtio: PASS (26.29 s)

maxujun · 2024-12-25T12:12:45Z

@PaulYuuu @fbq815 Could you help review?

YvanY0

@maxujun code almost good to me, but could you avoid handle blank lines on qdevices.py, as it's not a part of this PR.

And I leave few comments for review, also I saw 1 CI failed you need to take a look.

YvanY0 · 2025-01-17T03:07:53Z

+    def __init__(self, busid, bus_type, aobject):
+        """bus&ssid&devno, 4 ssid and 65536 device numbers"""
+        super(QCSSBus, self).__init__(
+            "bus", [["ssid", "devno"], [4, 65536]], busid, bus_type, aobject


Well, here only cover ssid and devno, as cssidset fe by default. However, should we cover unusual test scenarios? from the qemu doc

* a virtio-mouse device in a non-standard channel subsystem image:: -device virtio-mouse-ccw,devno=2.0.2222 This would not show up in a standard Linux guest. The properties for the device would be ``dev_id = "2.0.2222"`` and ``subch_id = "2.0.0000"``. * a virtio-keyboard device in another non-standard channel subsystem image:: -device virtio-keyboard-ccw,devno=0.0.1234 This would not show up in a standard Linux guest, either, as ``0`` is not the standard channel subsystem image id. The properties for the device would be ``dev_id = "0.0.1234"`` and ``subch_id = "0.0.0000"``.

we can set cssid value rather than default fe, so how about includes the ssid but set fe by default.

YvanY0 · 2025-01-17T03:24:03Z

+        if device.get_param("addr"):
+            del device.params["addr"]
+        device.set_param("devno", "fe.%s.%s" % (
+            hex(addr[0])[2:], f"{addr[1]:04x}"))


as we can mark sure the ssid only with length 0-3, so hex(addr[0])[2:] can easy transfer to addr[0]

And for f"{addr[1]:04x}"), of course your can "fe.%s.%4x" and addr[1] to do the same thing. Or just f"fe.{addr[0]}.{addr[1]:4x}"

maxujun · 2025-03-04T04:46:41Z

Updated.

Secure exectuion: remove hardcode of device number for secure execution related cases on s390x, cause we are adding device number by default, reference to avocado-framework/avocado-vt#4010 Signed-off-by: bfu <bfu@redhat.com>

YvanY0 · 2025-04-03T01:48:07Z

+                qdevices.QCSSBus(
                    "virtual-css",
                    "virtual-css",
                    "virtual-css",


I think busid, bus_type, aobject should not be the same.

@xuma as the info qtree, please change the bus_type to "virtual-css-bus",
reference:
dev: virtual-css-bridge, id ""
css_dev_path = true
bus: virtual-css
type virtual-css-bus
dev: virtio-keyboard-ccw, id "input_keyboard1"
ioeventfd = true
max_revision = 2 (0x2)
devno = ""
dev_id = "fe.0.0004"
subch_id = "fe.0.0004"

YvanY0 · 2025-04-03T01:54:03Z

+        """bus&cssid&ssid&devno, 255 cssid,4 ssid and 65536 device numbers"""
+        super(QCSSBus, self).__init__(
+            "bus",
+            [["cssid", "ssid", "devno"], [255, 4, 65536]],


Set some default values in arguments; this is more flexible. If we have other requirements, then we can init QCSSBus with diff length.

And their range should be 0-255, 0-3, 0-65535.
I would suggest you refer https://gitlab.com/qemu-project/qemu/-/blob/master/docs/system/s390x/css.rst?ref_type=heads

@maxujun should be [["cssid", "ssid", "devno"], [256, 4, 65536]]

YvanY0 · 2025-04-03T02:06:02Z

+        if device.get_param("addr"):
+            del device.params["addr"]
+        device.set_param("devno", f"{addr[0]:x}.{addr[1]}.{addr[2]:04x}")


Delete it and reset? For s390x, we should ensure addr is not set.

@maxujun please use devno instead of addr for s390x

maxujun · 2025-04-27T14:26:13Z

@PaulYuuu @fbq815 Updated，Please help review.thanks.

maxujun · 2025-04-29T08:14:24Z

(1/1) Host_RHEL.m9.u6.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.boot.s390-virtio: STARTED
(1/1) Host_RHEL.m9.u6.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.boot.s390-virtio: PASS (25.37 s)

fbq815

LGTM

fbq815 · 2025-04-29T11:16:29Z

@maxujun the code LGTM but seems you should handle the CI check issue

maxujun · 2025-05-08T07:19:38Z

@maxujun the code LGTM but seems you should handle the CI check issue

Updated and passed.

YvanY0 · 2025-05-12T06:52:17Z

+        """Read the values in base of 16 (hex)"""
+        addr = device.get_param("devno")
+        if isinstance(addr, int):
+            return [254, 0, addr]


If use hard code, then cssid and ssid in the __init__ are meaningless.

I mean, if we definitely use fe.0 as the prefix, we can handle devno only, to reduce the get free slot logic.

YvanY0 · 2025-05-12T06:56:33Z

+        )
+
+    @staticmethod
+    def _addr2stor(addr):


Seems this is exactly the same as QSparseBus._addr2stor, so why do we need it?

YvanY0 · 2025-05-12T07:09:23Z

+        elif not addr:  # not defined
+            return [254, 0, None]
+        elif isinstance(addr, six.string_types):
+            addr = [int(_, 16) for _ in addr.split(".", 2)]


The devno can range from 0~65535, it isn't hex, so int(_, 16) is incorrect.

Can you refine the _dev2addr function?

YvanY0

@maxujun can you provide the result, should contains command line with generated devno

YvanY0 · 2025-06-03T06:30:23Z

+    """
+
+    def __init__(
+        self, busid, bus_type, aobject, cssid_len=256, ssid_len=4, devno_len=65536


So with the latest code, it seems cssid_len and ssid_len are not necessary, as they are fixed values.

maxujun · 2025-06-10T07:06:06Z

@maxujun can you provide the result, should contains command line with generated devno

(1/1) Host_RHEL.m9.u6.nographic.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.boot.s390-virtio: STARTED
(1/1) Host_RHEL.m9.u6.nographic.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.boot.s390-virtio: PASS (21.70 s)
[stdlog] MALLOC_PERTURB_=1 /usr/libexec/qemu-kvm
[stdlog] -S
[stdlog] -name 'avocado-vt-vm1'
[stdlog] -sandbox on,elevateprivileges=deny,obsolete=deny,resourcecontrol=deny,spawn=deny
[stdlog] -machine s390-ccw-virtio,memory-backend=mem-machine_mem
[stdlog] -nodefaults
[stdlog] -vga none
[stdlog] -m 3072
[stdlog] -object '{"size": 3221225472, "id": "mem-machine_mem", "qom-type": "memory-backend-ram"}'
[stdlog] -smp 1,maxcpus=1,cores=1,threads=1,drawers=1,books=1,sockets=1
[stdlog] -cpu 'host'
[stdlog] -chardev socket,wait=off,id=qmp_id_qmpmonitor1,path=/var/tmp/avocado_zknrsajp/monitor-qmpmonitor1-20250610-030435-4qevaZBS,server=on
[stdlog] -mon chardev=qmp_id_qmpmonitor1,mode=control
[stdlog] -chardev socket,wait=off,id=qmp_id_catch_monitor,path=/var/tmp/avocado_zknrsajp/monitor-catch_monitor-20250610-030435-4qevaZBS,server=on
[stdlog] -mon chardev=qmp_id_catch_monitor,mode=control
[stdlog] -chardev socket,wait=off,id=chardev_serial0,path=/var/tmp/avocado_zknrsajp/serial-serial0-20250610-030435-4qevaZBS,server=on
[stdlog] -device '{"id": "serial0", "driver": "sclpconsole", "chardev": "chardev_serial0"}'
[stdlog] -device '{"id": "virtio_scsi_ccw0", "driver": "virtio-scsi-ccw", "devno": "fe.0.0000"}'
[stdlog] -blockdev '{"node-name": "file_image1", "driver": "file", "auto-read-only": true, "discard": "unmap", "aio": "threads", "filename": "/home/kar/vt_test_images/rhel960-s390x-virtio-scsi.qcow2", "cache": {"direct": true, "no-flush": false}}'
[stdlog] -blockdev '{"node-name": "drive_image1", "driver": "qcow2", "read-only": false, "cache": {"direct": true, "no-flush": false}, "file": "file_image1"}'
[stdlog] -device '{"driver": "scsi-hd", "id": "image1", "drive": "drive_image1", "write-cache": "on"}'
[stdlog] -device '{"driver": "virtio-net-ccw", "mac": "9a:87:a3:0c:af:4f", "id": "iddod4qr", "netdev": "idrXydF2", "devno": "fe.0.0001"}'
[stdlog] -netdev '{"id": "idrXydF2", "type": "tap", "vhost": true, "vhostfd": "16", "fd": "13"}'
[stdlog] -nographic
[stdlog] -rtc base=utc,clock=host
[stdlog] -boot strict=on
[stdlog] -enable-kvm
[stdlog] -device '{"driver": "virtio-mouse-ccw", "id": "input_mouse1", "devno": "fe.0.0002"}'
[stdlog] -device '{"driver": "virtio-keyboard-ccw", "id": "input_keyboard1", "devno": "fe.0.0003"}'
[stdlog] 2025-06-10 03:04:38,232 avocado.virttest.qemu_vm INFO | Created qemu process with parent PID 88114

fbq815

others LGTM, but when I tried with the SE guest, I found there's no devno, @maxujun could you please have a look? Details in jira tickets.
@PaulYuuu sorry for the wrong approval, I'll take a try again.

maxujun · 2025-06-11T07:54:33Z

(1/1) Host_RHEL.m9.u7.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.nic_hotplug.one_pci.nic_virtio.s390-virtio: STARTED
(1/1) Host_RHEL.m9.u7.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.9.6.0.s390x.io-github-autotest-qemu.nic_hotplug.one_pci.nic_virtio.s390-virtio: PASS (326.57 s)

[stdlog] 2025-06-11 03:03:59,316 avocado.test nic_hotplug L0217 DEBUG| Hotplug 1th 'virtio-net-ccw' nic named 'hotplug_nic1'
[stdlog] 2025-06-11 03:03:59,317 avocado.virttest.virt_vm virt_vm L0965 DEBUG| Generating random mac address for nic
[stdlog] 2025-06-11 03:03:59,332 avocado.virttest.qemu_vm error_context L0084 DEBUG| Context: Start test iteration 1 --> (hotplug_nic) --> Opening tap device node for t1-Ufxuby
[stdlog] 2025-06-11 03:03:59,335 avocado.virttest.qemu_vm error_context L0084 INFO | Context: Start test iteration 1 --> (hotplug_nic) --> Assigning tap idojSaoc to qemu by fd
[stdlog] 2025-06-11 03:03:59,335 avocado.virttest.qemu_monitor qemu_monitor L0448 DEBUG| (monitor avocado-vt-vm1.qmpmonitor1) Sending command 'getfd'
[stdlog] 2025-06-11 03:03:59,335 avocado.virttest.qemu_monitor qemu_monitor L2074 DEBUG| Send command: {"execute": "getfd", "arguments": {"fdname": "idojSaoc"}, "id": "VDZJRCt3"}
[stdlog] 2025-06-11 03:03:59,336 avocado.virttest.qemu_vm error_context L0084 DEBUG| Context: Start test iteration 1 --> (hotplug_nic) --> Raising interface for nic hotplug_nic1 on vm avocado-vt-vm1
[stdlog] 2025-06-11 03:03:59,338 avocado.virttest.qemu_vm error_context L0084 DEBUG| Context: Start test iteration 1 --> (hotplug_nic) --> Raising bridge for nic hotplug_nic1 on vm avocado-vt-vm1
[stdlog] 2025-06-11 03:03:59,342 avocado.virttest.qemu_vm error_context L0084 DEBUG| Context: Start test iteration 1 --> (hotplug_nic) --> Hotplugging nic hotplug_nic1 on vm avocado-vt-vm1
[stdlog] 2025-06-11 03:03:59,342 avocado.virttest.qemu_monitor qemu_monitor L0448 DEBUG| (monitor avocado-vt-vm1.qmpmonitor1) Sending command 'netdev_add'
[stdlog] 2025-06-11 03:03:59,343 avocado.virttest.qemu_monitor qemu_monitor L2074 DEBUG| Send command: {"execute": "netdev_add", "arguments": {"id": "id7d3b9P", "fd": "38", "type": "tap"}, "id": "k5r7xZeK"}
[stdlog] 2025-06-11 03:03:59,346 avocado.virttest.qemu_monitor qemu_monitor L0448 DEBUG| (monitor avocado-vt-vm1.qmpmonitor1) Sending command 'device_add'
[stdlog] 2025-06-11 03:03:59,346 avocado.virttest.qemu_monitor qemu_monitor L2074 DEBUG| Send command: {"execute": "device_add", "arguments": {"id": "idULY8ro", "driver": "virtio-net-ccw", "netdev": "id7d3b9P", "mac": "9a:71:84:a6:a9:65", "bus": "virtual-css", "devno": "fe.0.0005"}, "id": "cO7Lsllo"}
[stdlog] 2025-06-11 03:03:59,348 avocado.test nic_hotplug L0221 INFO | Check if new interface gets ip address

fbq815 · 2025-06-16T09:38:13Z

@maxujun I've seen that with SE guest, there's no devno, could you please have a check?

maxujun · 2025-06-24T11:11:07Z

@maxujun I've seen that with SE guest, there's no devno, could you please have a check?
No problem on se host, I noticed the default directory is “/root/.local/lib/python3.12/site-packages/virttest/”，Not “/usr/local/lib/python3.12/site-packages/”. So maybe you patched to a wrong directory.
(1/1) Host_RHEL.m10.u1.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.10.1.s390x.io-github-autotest-qemu.boot.s390-virtio: STARTED
(1/1) Host_RHEL.m10.u1.qcow2.virtio_scsi.up.virtio_net.Guest.RHEL.10.1.s390x.io-github-autotest-qemu.boot.s390-virtio: PASS (24.43 s)
[stdlog] MALLOC_PERTURB_=1 /usr/libexec/qemu-kvm
[stdlog] -S
[stdlog] -name 'avocado-vt-vm1'
[stdlog] -sandbox on,elevateprivileges=deny,obsolete=deny,resourcecontrol=deny,spawn=deny
[stdlog] -machine s390-ccw-virtio,memory-backend=mem-machine_mem
[stdlog] -nodefaults
[stdlog] -device '{"driver": "virtio-gpu-ccw", "devno": "fe.0.0000"}'
[stdlog] -m 14336
[stdlog] -object '{"size": 15032385536, "id": "mem-machine_mem", "qom-type": "memory-backend-ram"}'
[stdlog] -smp 8,maxcpus=8,threads=1,sockets=2
[stdlog] -cpu 'host'
[stdlog] -chardev socket,wait=off,id=qmp_id_qmpmonitor1,server=on,path=/var/tmp/avocado_yohjbf0d/monitor-qmpmonitor1-20250624-055049-LBqlihM7
[stdlog] -mon chardev=qmp_id_qmpmonitor1,mode=control
[stdlog] -chardev socket,wait=off,id=qmp_id_catch_monitor,server=on,path=/var/tmp/avocado_yohjbf0d/monitor-catch_monitor-20250624-055049-LBqlihM7
[stdlog] -mon chardev=qmp_id_catch_monitor,mode=control
[stdlog] -chardev socket,wait=off,id=chardev_serial0,server=on,path=/var/tmp/avocado_yohjbf0d/serial-serial0-20250624-055049-LBqlihM7
[stdlog] -device '{"id": "serial0", "driver": "sclpconsole", "chardev": "chardev_serial0"}'
[stdlog] -device '{"id": "virtio_scsi_ccw0", "driver": "virtio-scsi-ccw", "devno": "fe.0.0001"}'
[stdlog] -blockdev '{"node-name": "file_image1", "driver": "file", "auto-read-only": true, "discard": "unmap", "aio": "threads", "filename": "/home/kar/vt_test_images/rhel101-s390x-virtio-scsi.qcow2", "cache": {"direct": true, "no-flush": false}}'
[stdlog] -blockdev '{"node-name": "drive_image1", "driver": "qcow2", "read-only": false, "cache": {"direct": true, "no-flush": false}, "file": "file_image1"}'
[stdlog] -device '{"driver": "scsi-hd", "id": "image1", "drive": "drive_image1", "write-cache": "on"}'
[stdlog] -device '{"driver": "virtio-net-ccw", "mac": "9a:de:6d:99:3c:28", "id": "id7fUxfB", "netdev": "idg3GQpx", "devno": "fe.0.0002"}'
[stdlog] -netdev '{"id": "idg3GQpx", "type": "tap", "vhost": true, "vhostfd": "17", "fd": "14"}'
[stdlog] -vnc :0
[stdlog] -rtc base=utc,clock=host
[stdlog] -boot strict=on
[stdlog] -enable-kvm
[stdlog] -device '{"driver": "virtio-mouse-ccw", "id": "input_mouse1", "devno": "fe.0.0003"}'
[stdlog] -device '{"driver": "virtio-keyboard-ccw", "id": "input_keyboard1", "devno": "fe.0.0004"}'

Signed-off-by: Xujun Ma <xuma@redhat.com>

fbq815

LGTM:
-object s390-pv-guest,id=lsec0
[stdlog] -enable-kvm
[stdlog] -device '{"driver": "virtio-mouse-ccw", "id": "input_mouse1", "devno": "fe.0.0003"}'
[stdlog] -device '{"driver": "virtio-keyboard-ccw", "id": "input_keyboard1", "devno": "fe.0.0004"}'

Secure exectuion: remove hardcode of device number for secure execution related cases on s390x, cause we are adding device number by default, reference to avocado-framework/avocado-vt#4010 Signed-off-by: bfu <bfu@redhat.com>

maxujun force-pushed the devno branch from bfd86ab to 82b362a Compare October 8, 2024 07:40

maxujun marked this pull request as draft October 29, 2024 09:08

maxujun force-pushed the devno branch 3 times, most recently from bb0897e to 10065f5 Compare December 20, 2024 08:31

maxujun marked this pull request as ready for review December 25, 2024 12:11

YvanY0 reviewed Jan 17, 2025

View reviewed changes

maxujun force-pushed the devno branch 6 times, most recently from d1f1344 to 9b12a5c Compare March 4, 2025 03:27

fbq815 mentioned this pull request Mar 24, 2025

Remove device number hardcode for secure execution on s390x autotest/tp-qemu#4292

Merged

fbq815 reviewed Mar 24, 2025

View reviewed changes

Comment thread virttest/qemu_devices/qdevices.py Outdated

maxujun force-pushed the devno branch 2 times, most recently from 223171a to 8117f7b Compare March 29, 2025 06:27

YvanY0 reviewed Apr 3, 2025

View reviewed changes

maxujun force-pushed the devno branch 5 times, most recently from 3149364 to 33b0724 Compare April 27, 2025 14:21

fbq815 approved these changes Apr 29, 2025

View reviewed changes

maxujun force-pushed the devno branch from 33b0724 to 7872490 Compare May 8, 2025 06:28

YvanY0 reviewed May 12, 2025

View reviewed changes

maxujun force-pushed the devno branch 2 times, most recently from e807b91 to 5d0d25d Compare May 20, 2025 09:03

YvanY0 reviewed Jun 3, 2025

View reviewed changes

maxujun force-pushed the devno branch 3 times, most recently from 2e462ee to a987958 Compare June 10, 2025 06:52

fbq815 approved these changes Jun 11, 2025

View reviewed changes

maxujun force-pushed the devno branch from a987958 to 30a1f22 Compare June 11, 2025 07:15

Add device address parameter for s390x

8863286

Signed-off-by: Xujun Ma <xuma@redhat.com>

maxujun force-pushed the devno branch from 30a1f22 to 8863286 Compare June 24, 2025 11:13

fbq815 approved these changes Jun 30, 2025

View reviewed changes

YvanY0 approved these changes Jul 1, 2025

View reviewed changes

YvanY0 merged commit a74cb78 into avocado-framework:master Jul 1, 2025
46 checks passed

Conversation

maxujun commented Oct 8, 2024

Uh oh!

maxujun commented Oct 8, 2024

Uh oh!

maxujun commented Dec 25, 2024

Uh oh!

YvanY0 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maxujun commented Mar 4, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maxujun commented Apr 27, 2025

Uh oh!

maxujun commented Apr 29, 2025

Uh oh!

fbq815 left a comment

Choose a reason for hiding this comment

Uh oh!

fbq815 commented Apr 29, 2025

Uh oh!

maxujun commented May 8, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

YvanY0 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maxujun commented Jun 10, 2025

Uh oh!

fbq815 left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

maxujun commented Jun 11, 2025

Uh oh!

fbq815 commented Jun 16, 2025

Uh oh!

maxujun commented Jun 24, 2025

Uh oh!

fbq815 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

fbq815 left a comment •

edited

Loading