)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"42187f47_33690aba","updated":"2023-03-15 13:55:27.000000000","message":"i forget to submit these a few days ago sorry","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"bea635664ecad2f8315a76f814684c59c0050ada","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"57afa58c_41c98445","updated":"2023-04-21 10:53:28.000000000","message":"I\u0027m not against this just wondering about the naming.","commit_id":"d382551d4bdf91e618453fe0b7716b03f655fbd6"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"43e992aa8c2a91fa13dc51261564294deddb147f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":8,"id":"e2558446_ba98d7c2","updated":"2023-05-03 12:18:13.000000000","message":"This looks good to me. (The pep8 result is valid as we have L54 longer than 79 chars)","commit_id":"4787148e4d68282b5bfbed277cc46087a3d8d220"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"f987db1ccf239553f98b329dd8441ef167ff3c9f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":9,"id":"0fbea575_6c867933","updated":"2023-05-03 12:42:56.000000000","message":"Let\u0027s do this.","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"}],"specs/2023.1/approved/virtio_packedring_configuration_support.rst":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"85d9c48d794a7b2d23cbf424b9a6a41f8c48690e","unresolved":true,"context_lines":[{"line_number":55,"context_line":"option to `true`. By enabling this option, operators will allow VMs to choose"},{"line_number":56,"context_line":"the preferred virtq layout. By disabling this option, the operator will retain"},{"line_number":57,"context_line":"OpenStack working as-is."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Alternatives"},{"line_number":60,"context_line":"------------"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"1804fa98_90138ffa","line":58,"updated":"2023-01-02 05:35:50.000000000","message":"this is very anti cloud.\n\nif the ring layout is something that the guest driver can configure when packed ring fromat is advertiesed then we should not have any config option.\n\nwe should enable the capablity advertisement at both the qemu/libvirt level and the placment/scheduler level.\n\n\nwhat we should do is add a new standard trait COMPUTE_NET_VIRTIO_PACKED\non hypervisors that meet the qemu and libvirt requirements the compute agent will\nreport that new trait to placement.\n\nwhere the local hypervisor supports packed format the feature bit should be always enabled so that the guest can negociate teh ring layout.\n\nas an end user that requires packed format for my VNF or application i can upload an image with a required trait request via the image properites.\n\nas an operator i can simiarly use required or forbidin traits in the flavor to \nprovide affinity and anti affinity.\n\n\nthe largest chalange i see with this or the config driven approch is live migrations.\n\nthe virtio ring format and feature bit in general are not allowed to change during a llive migration. more specificially the source and destination host must both support the negociated feature set. the soucce node gets that for free via the inital feature negociation in the guest when the vm is booted. the destintaiton requirement must be ensure by the schduler/placement and it will be validated by qemu/libvirt when we invoke the lower level live migration.\n\n\nthat means that if we have a vm that has packed format negocation enabled we need to be able to ensure we select host where that can also be enabled.\n\nas far as i am aware there is no way to detect from libvirt what virtio fetures were negocated in the guest and we cant ask libvirt what virtio features are supported by the host vhost backend.\n\n\nwe have encountered issues in this regrad in the past where adding the mtu to the libvirt xml changes the virtio feature bitmask form a 32bit int to a 64bit int.\n\nwhen we first started adding the mtu it broke live migation because we did not detect if the running vm had the mtu element and uncontionally added it when we live migrated.\n\nthat worked fine for vms booted on the current release but broke live migrtion for\nvm booted on the previous release and there for boke updates.\n\nadding a config option is one way to work around that but its a bad one as it add operational complexity and still has upgrade impacts if you ever want to change the default or explictly set it to another value.\n\nso to make live migration and upgrades work we need to be carful to ensure we know if a vm has been booted with support for packed format.\n\nwe normally do that via and image property.\nfor example we enable virtio-net multiqueue support via an image property or flavor extra spec.\n\nif we were to make this configurable we would need to do that via a flaovr/image property not via a host level config option.\n\nthis however is quite a low level option to expsoe to an enduser/cloud operator and \nreally is not something that likely should be exposed.\n\nif we can instead detect this cabplity and automatically enable it where possible that would be better.\n\nso the main question i have is if we were to resolve the migration/upgrade problem\nis there a reason why we should not always enable this feature transparently and make it uncondtional from a user/operator level . (with some caveats for upgrades)","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"},{"author":{"_account_id":35239,"name":"Justas Poderys","email":"jpo@napatech.com","username":"justas_napa"},"change_message_id":"3fb6c0d0c5031b1c32bfeed7a046de5d83ce8890","unresolved":true,"context_lines":[{"line_number":55,"context_line":"option to `true`. By enabling this option, operators will allow VMs to choose"},{"line_number":56,"context_line":"the preferred virtq layout. By disabling this option, the operator will retain"},{"line_number":57,"context_line":"OpenStack working as-is."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Alternatives"},{"line_number":60,"context_line":"------------"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"c181cbf3_e1183d8a","line":58,"in_reply_to":"1804fa98_90138ffa","updated":"2023-01-02 08:32:41.000000000","message":"Few notes:\n- I don\u0027t see a single reason not to have this enabled, and leave it to VM to decide the preferred layout.\n- There is no reliable way to know the negotiated option. Furthermore, nothing stops the VM from unbinding the driver at some point and then re-binding with different layout selected.\n- I agree that going the \"trait\" way is best for now. Should we take a jab and make a change PR, or should be wait for more community input?","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"b994f21e4208d2f84fdad52cbd21066a6978aafe","unresolved":false,"context_lines":[{"line_number":55,"context_line":"option to `true`. By enabling this option, operators will allow VMs to choose"},{"line_number":56,"context_line":"the preferred virtq layout. By disabling this option, the operator will retain"},{"line_number":57,"context_line":"OpenStack working as-is."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Alternatives"},{"line_number":60,"context_line":"------------"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"8da97a8a_e677bc2a","line":58,"in_reply_to":"6b0d4684_141bb15f","updated":"2023-03-16 09:31:30.000000000","message":"Done","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"ab7060fb8c0e85d3c0758f25602f4794f7de3f7e","unresolved":true,"context_lines":[{"line_number":55,"context_line":"option to `true`. By enabling this option, operators will allow VMs to choose"},{"line_number":56,"context_line":"the preferred virtq layout. By disabling this option, the operator will retain"},{"line_number":57,"context_line":"OpenStack working as-is."},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Alternatives"},{"line_number":60,"context_line":"------------"},{"line_number":61,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"6b0d4684_141bb15f","line":58,"in_reply_to":"c181cbf3_e1183d8a","updated":"2023-01-06 11:59:51.000000000","message":"we have extened the spec freeze until next thursday so technially the spec coudl be improved and reviewd in that time but it proably wont happen until next cycle\n\nif you want to start working on an implemenation that is ok but note that this really is proably not going to get merged until the B/2023.2 cycle\n\nthe main thing you need to ensure is that if a vm is booted on an non upgraded node it can live migrate to an upgraded node.\n\nif its booted on an upgraded nodes its ok if it cannot migrate to a non upgraded node. with that said its preferable if that still works too.\n\nfor upgraded nodes the migration code that updates the xml should ensure that if the feature is not enabeld in the orginal xml its not enabled in the updated xml.\n\nhowever on cold migrate/normal boot its fine to enabel the feature automatically when runnign through the upgraded code path.","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"ab7060fb8c0e85d3c0758f25602f4794f7de3f7e","unresolved":true,"context_lines":[{"line_number":107,"context_line":"Upgrade impact"},{"line_number":108,"context_line":"--------------"},{"line_number":109,"context_line":""},{"line_number":110,"context_line":"None"},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"Implementation"},{"line_number":113,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"d7fb4155_399e0433","line":110,"updated":"2023-01-06 11:59:51.000000000","message":"this need to detail how upgrade will be supported.\n\nspecifical as i noted above how live migration between an non upgraded host and an upgraded host will be supported. including back and fort migration.\n\nback and forth migration is where a vm is booted on a non upgraded host and then migrated to an upgraded host and then back to the orginal host without reboooting the vm.\n\nwe test this in the grenade job to ensure that durign a roling upgrade operators can move vms around to enable the upgrade without workload downtime.\n\nif the vm is not rebooted we do not expect the new feature ot be enable even when movign to/form the upgrade host.\n\nyou will need to provide functional tests to test thsi too.\n\nthe way this shoudl be enabled is for new spawns on upgraded hosts.\nsuch as boot, cold migrate, unshleve, resize, evacuate or hard_reboot.","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"},{"author":{"_account_id":35239,"name":"Justas Poderys","email":"jpo@napatech.com","username":"justas_napa"},"change_message_id":"4f55e16913bd3fcc15650423c5e418ab907411bc","unresolved":false,"context_lines":[{"line_number":107,"context_line":"Upgrade impact"},{"line_number":108,"context_line":"--------------"},{"line_number":109,"context_line":""},{"line_number":110,"context_line":"None"},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"Implementation"},{"line_number":113,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":1,"id":"dcd7828d_736e6c24","line":110,"in_reply_to":"d7fb4155_399e0433","updated":"2023-03-03 08:17:44.000000000","message":"This was now addressed in the spec.","commit_id":"46a9922419b4ebb5886f5847fa9039d888194bad"}],"specs/2023.2/approved/virtio_packedring_configuration_support.rst":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":37,"context_line":"as a mask, allowing the backed to advertise the support when set. The driver in"},{"line_number":38,"context_line":"the VM is still responsible for choosing the layout of virtqs."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"This blueprint proposes to add a Nova configuration option, that sets the"},{"line_number":41,"context_line":"`packed` option to `true` on the node. This way all VMs running on the node"},{"line_number":42,"context_line":"are allowed to choose the virtq layout based on what is offered by the backed,"},{"line_number":43,"context_line":"rather than being froced to use split-ring."}],"source_content_type":"text/x-rst","patch_set":6,"id":"62ca19e1_d15154d8","line":40,"range":{"start_line":40,"start_character":33,"end_line":40,"end_character":58},"updated":"2023-03-15 13:55:27.000000000","message":"nit: a Nova flavor extra_spec and glance image property\n\nit not actully adding any config options.","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"b4694f5d226e2a1611b960f0c90bdc09a433bc95","unresolved":false,"context_lines":[{"line_number":37,"context_line":"as a mask, allowing the backed to advertise the support when set. The driver in"},{"line_number":38,"context_line":"the VM is still responsible for choosing the layout of virtqs."},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"This blueprint proposes to add a Nova configuration option, that sets the"},{"line_number":41,"context_line":"`packed` option to `true` on the node. This way all VMs running on the node"},{"line_number":42,"context_line":"are allowed to choose the virtq layout based on what is offered by the backed,"},{"line_number":43,"context_line":"rather than being froced to use split-ring."}],"source_content_type":"text/x-rst","patch_set":6,"id":"fc8efd26_25a37360","line":40,"range":{"start_line":40,"start_character":33,"end_line":40,"end_character":58},"in_reply_to":"62ca19e1_d15154d8","updated":"2023-03-16 09:31:12.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":59,"context_line":"  hw:packed_ring\u003dtrue|false  (default false)"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"* Provide new compute ``COMPUTE_NET_VIRTIO_PACKED`` capablity trait."},{"line_number":62,"context_line":"  This trait can be required/forbidden by user. Scheduler will automatically"},{"line_number":63,"context_line":"  set this trait to the resource provider summary if libvirt version is higher"},{"line_number":64,"context_line":"  than 6.3"},{"line_number":65,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"f62b29c8_d438209c","line":62,"range":{"start_line":62,"start_character":48,"end_line":62,"end_character":57},"updated":"2023-03-15 13:55:27.000000000","message":"nova-compute agent","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"b4694f5d226e2a1611b960f0c90bdc09a433bc95","unresolved":false,"context_lines":[{"line_number":59,"context_line":"  hw:packed_ring\u003dtrue|false  (default false)"},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"* Provide new compute ``COMPUTE_NET_VIRTIO_PACKED`` capablity trait."},{"line_number":62,"context_line":"  This trait can be required/forbidden by user. Scheduler will automatically"},{"line_number":63,"context_line":"  set this trait to the resource provider summary if libvirt version is higher"},{"line_number":64,"context_line":"  than 6.3"},{"line_number":65,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"c7a9996a_df06d5f9","line":62,"range":{"start_line":62,"start_character":48,"end_line":62,"end_character":57},"in_reply_to":"f62b29c8_d438209c","updated":"2023-03-16 09:31:12.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":66,"context_line":"* This spec will update scheduling process. ALL_REQUEST_FILTERS will be"},{"line_number":67,"context_line":"  extended with new filter packed_virtqueue_filter. It will update RequestSpec"},{"line_number":68,"context_line":"  with new trait in case if image property or flavor extra_spec is enabled to"},{"line_number":69,"context_line":"  avoid migration to the node without packed virtqueue feature support."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":".. _\u0027libvirt format domain\u0027:"},{"line_number":72,"context_line":"  https://libvirt.org/formatdomain.html#virtio-related-options"}],"source_content_type":"text/x-rst","patch_set":6,"id":"b5c066c7_7b85ae12","line":69,"updated":"2023-03-15 13:55:27.000000000","message":"+1","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"a521d990b6c2c728715bb8cd4b6a4592ec4dbb7e","unresolved":false,"context_lines":[{"line_number":66,"context_line":"* This spec will update scheduling process. ALL_REQUEST_FILTERS will be"},{"line_number":67,"context_line":"  extended with new filter packed_virtqueue_filter. It will update RequestSpec"},{"line_number":68,"context_line":"  with new trait in case if image property or flavor extra_spec is enabled to"},{"line_number":69,"context_line":"  avoid migration to the node without packed virtqueue feature support."},{"line_number":70,"context_line":""},{"line_number":71,"context_line":".. _\u0027libvirt format domain\u0027:"},{"line_number":72,"context_line":"  https://libvirt.org/formatdomain.html#virtio-related-options"}],"source_content_type":"text/x-rst","patch_set":6,"id":"d0aa0868_1f6a1603","line":69,"in_reply_to":"b5c066c7_7b85ae12","updated":"2023-05-03 12:02:49.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":104,"context_line":"Performance Impact"},{"line_number":105,"context_line":"------------------"},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"VMs using virtio-net will see an increase in performance. The increase cane be"},{"line_number":108,"context_line":"anywhere between 10/20% (see DPDK Intel Vhost/virtio perf. reports) and 75%"},{"line_number":109,"context_line":"(using Napatech SmartNICs)."},{"line_number":110,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"1b83194b_15a13f22","line":107,"range":{"start_line":107,"start_character":71,"end_line":107,"end_character":75},"updated":"2023-03-15 13:55:27.000000000","message":"can","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"b4694f5d226e2a1611b960f0c90bdc09a433bc95","unresolved":false,"context_lines":[{"line_number":104,"context_line":"Performance Impact"},{"line_number":105,"context_line":"------------------"},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"VMs using virtio-net will see an increase in performance. The increase cane be"},{"line_number":108,"context_line":"anywhere between 10/20% (see DPDK Intel Vhost/virtio perf. reports) and 75%"},{"line_number":109,"context_line":"(using Napatech SmartNICs)."},{"line_number":110,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"be31519a_e8f6500f","line":107,"range":{"start_line":107,"start_character":71,"end_line":107,"end_character":75},"in_reply_to":"1b83194b_15a13f22","updated":"2023-03-16 09:31:12.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":106,"context_line":""},{"line_number":107,"context_line":"VMs using virtio-net will see an increase in performance. The increase cane be"},{"line_number":108,"context_line":"anywhere between 10/20% (see DPDK Intel Vhost/virtio perf. reports) and 75%"},{"line_number":109,"context_line":"(using Napatech SmartNICs)."},{"line_number":110,"context_line":""},{"line_number":111,"context_line":"Other deployer impact"},{"line_number":112,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"a205e383_6368dbd4","line":109,"updated":"2023-03-15 13:55:27.000000000","message":"ya so the split queue overhead for software implemation is much lower then the cost for hardware backends so this makes sense to me.","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"a521d990b6c2c728715bb8cd4b6a4592ec4dbb7e","unresolved":false,"context_lines":[{"line_number":106,"context_line":""},{"line_number":107,"context_line":"VMs using virtio-net will see an increase in performance. The increase cane be"},{"line_number":108,"context_line":"anywhere between 10/20% (see DPDK Intel Vhost/virtio perf. reports) and 75%"},{"line_number":109,"context_line":"(using Napatech SmartNICs)."},{"line_number":110,"context_line":""},{"line_number":111,"context_line":"Other deployer impact"},{"line_number":112,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"10bdc41b_0c2701f5","line":109,"in_reply_to":"a205e383_6368dbd4","updated":"2023-05-03 12:02:49.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"56de7d0e5f4ebe04ab89b3cefc231de2eb059f81","unresolved":true,"context_lines":[{"line_number":142,"context_line":"Feature Liaison"},{"line_number":143,"context_line":"---------------"},{"line_number":144,"context_line":""},{"line_number":145,"context_line":"None"},{"line_number":146,"context_line":""},{"line_number":147,"context_line":"Work Items"},{"line_number":148,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"5c60fe21_ad211f7c","line":145,"updated":"2023-03-15 13:55:27.000000000","message":"you can add me \"sean-k-mooney\" if you like","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"b4694f5d226e2a1611b960f0c90bdc09a433bc95","unresolved":false,"context_lines":[{"line_number":142,"context_line":"Feature Liaison"},{"line_number":143,"context_line":"---------------"},{"line_number":144,"context_line":""},{"line_number":145,"context_line":"None"},{"line_number":146,"context_line":""},{"line_number":147,"context_line":"Work Items"},{"line_number":148,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"95d9f80a_0d2246e7","line":145,"in_reply_to":"5c60fe21_ad211f7c","updated":"2023-03-16 09:31:12.000000000","message":"Done","commit_id":"ea6635da3b93e80a40162ed277b8fc530595ab70"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"bea635664ecad2f8315a76f814684c59c0050ada","unresolved":true,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_packed_ring`` for image property and ``hw:packed_ring`` for"},{"line_number":55,"context_line":"  flavor extra specs. Users will control the packed virtqueue feature,"},{"line_number":56,"context_line":"  and be able to disable it if desired."},{"line_number":57,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"3a490318_9d1fbf08","line":54,"updated":"2023-04-21 10:53:28.000000000","message":"Would it make sense to use a bit more specific name like `hw_virtio_packed_ring` and `hw:virtio_packed_ring`.","commit_id":"d382551d4bdf91e618453fe0b7716b03f655fbd6"},{"author":{"_account_id":35489,"name":"Danylo Vodopianov","email":"dvo-plv@napatech.com","username":"dvo-plv"},"change_message_id":"a521d990b6c2c728715bb8cd4b6a4592ec4dbb7e","unresolved":false,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_packed_ring`` for image property and ``hw:packed_ring`` for"},{"line_number":55,"context_line":"  flavor extra specs. Users will control the packed virtqueue feature,"},{"line_number":56,"context_line":"  and be able to disable it if desired."},{"line_number":57,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"bb3b25c1_045ca177","line":54,"in_reply_to":"0ca89914_02a1a68a","updated":"2023-05-03 12:02:49.000000000","message":"Done","commit_id":"d382551d4bdf91e618453fe0b7716b03f655fbd6"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"a325c6d24714022a16c32b0606f7a11899358ec7","unresolved":true,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_packed_ring`` for image property and ``hw:packed_ring`` for"},{"line_number":55,"context_line":"  flavor extra specs. Users will control the packed virtqueue feature,"},{"line_number":56,"context_line":"  and be able to disable it if desired."},{"line_number":57,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"0ca89914_02a1a68a","line":54,"in_reply_to":"3a490318_9d1fbf08","updated":"2023-04-21 11:44:51.000000000","message":"ya that is proably clearer so im +1 on the rename","commit_id":"d382551d4bdf91e618453fe0b7716b03f655fbd6"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"d74dc766abfe70f17562cfc76fe9c7f7a4404f17","unresolved":true,"context_lines":[{"line_number":18,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"VM using a Virtio-net paravirtual network device uses Virtual queues (virtqs)"},{"line_number":21,"context_line":"to send and recveive data between the virtio-net driver and the virtual or"},{"line_number":22,"context_line":"physical backed. The VirtIO standard originally defined a single type of virtq"},{"line_number":23,"context_line":"called split-ring queue. The latest edition of the standard (v1.1) adds a"},{"line_number":24,"context_line":"different type of the virtq, called packed-ring queue. A different layout of"}],"source_content_type":"text/x-rst","patch_set":9,"id":"31af6a67_a7b2438c","line":21,"range":{"start_line":21,"start_character":12,"end_line":21,"end_character":20},"updated":"2023-05-30 12:38:02.000000000","message":"receive","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"d74dc766abfe70f17562cfc76fe9c7f7a4404f17","unresolved":true,"context_lines":[{"line_number":31,"context_line":"virtq layout based on what it supports. As both options are identical features"},{"line_number":32,"context_line":"wise, and the packed-ring is more efficient, the latter is typically chosen."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"Qemu added support for the packed virtqs in v4.2 and LibVirt in v6.3. Qemu and"},{"line_number":35,"context_line":"LibVirt supports the packed-ring virtqs via the `packed` option. However, note"},{"line_number":36,"context_line":"that this option *does not* force the VM to use the packed-ring virtq. It acts"},{"line_number":37,"context_line":"as a mask, allowing the backed to advertise the support when set. The driver in"}],"source_content_type":"text/x-rst","patch_set":9,"id":"07f610b5_18ace638","line":34,"range":{"start_line":34,"start_character":0,"end_line":34,"end_character":4},"updated":"2023-05-30 12:38:02.000000000","message":"QEMU","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"d74dc766abfe70f17562cfc76fe9c7f7a4404f17","unresolved":true,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_virtio_packed_ring`` for image property and"},{"line_number":55,"context_line":"  ``hw:virtio_packed_ring`` for flavor extra specs."},{"line_number":56,"context_line":"  Users will control the packed virtqueue feature, and be able to disable"},{"line_number":57,"context_line":"  it if desired."}],"source_content_type":"text/x-rst","patch_set":9,"id":"15d96560_255aa2c4","line":54,"updated":"2023-05-30 12:38:02.000000000","message":"Usually we do prefer have the flavor limiting the usage of any feature.\n\nIn your purpose any flavor may have this feature enabled. Operators may want to give that opportunity of using this feature for flavors only.\n\n\nhw_virtu_packed_ring should be available only if hw:virtio_packet_ring\u003dimage (as an example)","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"b455dc9787f2e0f75f08286a578fb383d5ed31b4","unresolved":true,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_virtio_packed_ring`` for image property and"},{"line_number":55,"context_line":"  ``hw:virtio_packed_ring`` for flavor extra specs."},{"line_number":56,"context_line":"  Users will control the packed virtqueue feature, and be able to disable"},{"line_number":57,"context_line":"  it if desired."}],"source_content_type":"text/x-rst","patch_set":9,"id":"54f0c003_a9098c42","line":54,"in_reply_to":"0e79a09c_afaaee6a","updated":"2023-05-30 13:17:40.000000000","message":"I see my mistake, but should not we at least indicate that if operator specifically set extra_spec to false, the user will not be able to enable the feature by overriding it using image properties?","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b41d265f275e8e0ca70e8082357eab74db507460","unresolved":true,"context_lines":[{"line_number":51,"context_line":"Proposed change"},{"line_number":52,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* Add ``hw_virtio_packed_ring`` for image property and"},{"line_number":55,"context_line":"  ``hw:virtio_packed_ring`` for flavor extra specs."},{"line_number":56,"context_line":"  Users will control the packed virtqueue feature, and be able to disable"},{"line_number":57,"context_line":"  it if desired."}],"source_content_type":"text/x-rst","patch_set":9,"id":"0e79a09c_afaaee6a","line":54,"in_reply_to":"15d96560_255aa2c4","updated":"2023-05-30 13:03:02.000000000","message":"that is not correct.\n\nwe only use the flavor to restrict the usage of a feature if it change the resouce allocation.\n\nhistorocially things like multi queue or the nic model where exclivily configurable by image property.\n\nso any feature like hw_virtio_packed_ring would have previouly only been enabled via the image.\n\nafter some request form operator to support enabling multi queue form the flavor we have default to added most new option as both an image property and a flaovr extra spec at the same time if we think its useful.\n\n\nwe have also moved away form requiring flavor extra specs to be set to allow things to be enabeld via the image.\n\nwe have recived feedback in the past that operator find that annoying to manage\nas they have to update or duplicate there falvors.\n\nso i dont thnk we should required hw:virtio_packed_ring\u003dimage as that would not be consitent with the direction we have been moving in for the last 3-4 years.","commit_id":"3a80f8a42fa8b976600885ffe8b91db1f2a988c3"}]}