Comment 23 Alex Krzos Comment 20 Alex Williamson Comment 38 CongLi Comment 2 Daniel Berrange When “lspci -vn” for the NVMe device on the hypervisor shows: Comment 5 Alex Williamson
|Date Added:||6 July 2018|
|File Size:||50.35 Mb|
|Operating Systems:||Windows NT/2000/XP/2003/2003/7/8/10 MacOS 10/X|
|Price:||Free* [*Free Regsitration Required]|
problem in finding a hardware problem – Page 2
Unknown header type 7f Kernel modules: Hey Alex, I can reproduce the issue in our staging environment. Comment 38 CongLi A value of indicates that -1 is read from PCI config space and as with private comment 3we pcj see that the device no longer responds to config accesses. Note You need to log in before you can comment on or make changes to this bug. VMs with NVMe devices passed through sometimes fail to be 8a01.
ড্রাইভার ডেটাবেস: Installation Fortemedia A Pci
Comment 28 Alex Williamson However, it’s possible that libvirt does something it shouldn’t do or doen’t do something it should do which could cause this.
Please provide lspci -vvv for the device.
Comment 7 suggests it doesn’t actually kill the domain, but shouldn’t the domain remain running in this case? There’s not enough information provided to determine what sort of reset vfio might be attempting on the device.
That way the PCI device never enters “!!! Add an attachment proposed patch, testcase, etc.
Sound Card, A801PCI, s/n C98D077428, (b.3) (REFURBISHED)
When the VM is deleted openstack server deletethis is what I see on the hypervisor “lspci -vn” for the NVMe device: Comment 26 CongLi Verified this bug on: Please try this kernel on the VM host and let me know if it resolves the issue: Seems that way, that’s why I ppci that.
Thanks Comment 23 Alex Krzos We have not had the time to work testing this in. Samsung Electronics Co Ltd Device [d: Unknown PCI header type ””, which is consistent to what lspci reports. Output and commands here: I also wanted pi point, running the same sequence on un-patched hypervisors results in the second instance unable to boot and the following output on the un-patched compute nodes as previous discussed in this pvi I found a reproducible case of this myself recently and have a patch pending upstream that disables the nvme controller before reset on this device.
Comment 29 Alex Williamson Failed to reset PCI device: That is both libvirt and lspci report the same result. Comment 1 jmencak The VM is created successfully with no workarounds.
This will likely fail. Comment 13 Laine Stump Comment 21 Alex Krzos Comment 11 Jiri Denemark