Description of problem: [machine type 9.2]Failed to migrate VM from RHEL 9.3 to RHEL 9.2 Version-Release number of selected component (if applicable): SRC: RHEL 9.3 qemu-kvm-8.0.0-7.el9.x86_64 kernel-5.14.0-338.el9.x86_64 seabios-bin-1.16.1-1.el9.noarch DST: RHEL 9.2 kernel-5.14.0-284.18.1.el9_2.x86_64 qemu-kvm-7.2.0-14.el9_2.3.x86_64 seabios-bin-1.16.1-1.el9.noarch How reproducible: 5/5 Steps to Reproduce: 1.In src : # /usr/libexec/qemu-kvm -name guest=rhel9 -machine pc-q35-rhel9.2.0 -monitor stdio -vnc :1 -cpu Broadwell-noTSX-IBRS,enforce QEMU 8.0.0 monitor - type 'help' for more information (qemu) 2.In dst : # /usr/libexec/qemu-kvm -name guest=rhel9 -machine pc-q35-rhel9.2.0 -monitor stdio -vnc :1 -cpu Broadwell-noTSX-IBRS,enforce -incoming defer QEMU 7.2.0 monitor - type 'help' for more information (qemu) migrate_incoming tcp::4444 (qemu) 3.Migrate VM from RHEL9.3.0 to RHEL9.2.0 (qemu) migrate -d tcp:10.73.212.86:4444 Actual results: (qemu) qemu-kvm: get_pci_config_device: Bad config data: i=0x10a read: 40 device: 0 cmask: ff wmask: 0 w1cmask:0 qemu-kvm: Failed to load PCIDevice:config qemu-kvm: Failed to load e1000e:parent_obj qemu-kvm: error while loading state for instance 0x0 of device '0000:00:02.0/e1000e' qemu-kvm: load of migration failed: Invalid argument Expected results: Migrate vm successfully. Additional info:
Seems to be another PCI breakage besides bz2215819. Needinfo Michael / Igor for this.
maybe yes, maybe no. could you test with the fix for bz2215819 please?
(In reply to Michael S. Tsirkin from comment #2) > maybe yes, maybe no. could you test with the fix for bz2215819 please? Michael, please see https://bugzilla.redhat.com/show_bug.cgi?id=2215819#c35 - note that the config index is different. Min, would you please double check with what Michael said (by checking whether Leo's fix there can fix this problem)? Thanks!
yes i saw that, it's a different device so could still be same capability at a different index.
It blocks test from rhel 9.3 and rhel 9.2 and it should be a blocker from QE's perspective. Also cc fjin Thank you.
I found the bug: patch 5ed3dabe57d was applied to 9.3, and hw_compat_7_2 got { TYPE_PCI_DEVICE, "x-pcie-err-unc-mask", "off" }, but I did not add this line to hw_compat_rhel_9_2, so rhel machine type 9.2 still set the undesired bit I will send a one-liner downstream to fix this
QE bot(pre verify): Set 'Verified:Tested,SanityOnly' as gating/tier1 test pass.
Verified the bug on following builds SRC:RHEL 9.3 kernel-5.14.0-348.el9.x86_64 qemu-kvm-8.0.0-10.el9.x86_64 /usr/libexec/qemu-kvm -name guest=rhel9 -machine pc-q35-rhel9.2.0 -monitor stdio -vnc :1 -cpu Broadwell-noTSX-IBRS,enforce QEMU 8.0.0 monitor - type 'help' for more information (qemu) migrate -d tcp:10.73.212.86:4444 DST:RHEL 9.2 5.14.0-284.26.1.el9_2.x86_64 qemu-kvm-7.2.0-14.el9_2.3.x86_64 /usr/libexec/qemu-kvm -name guest=rhel9 -machine pc-q35-rhel9.2.0 -monitor stdio -vnc :1 -cpu Broadwell-noTSX-IBRS,enforce -incoming defer QEMU 7.2.0 monitor - type 'help' for more information (qemu) migrate_incoming tcp::4444 Actual results Migration passed Expected results Migration pass
Base on comment 18 and comment 19, move this bug to be verified. Thank YOU !