<html><head><style>body{font-family:Helvetica,Arial;font-size:13px}</style></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><div id="bloop_customfont" style="font-family:Helvetica,Arial;font-size:13px; color: rgba(0,0,0,1.0); margin: 0px; line-height: auto;">On November 11, 2015 at 04:41:22, arch user (<a href="mailto:archuser@gmx.com">archuser@gmx.com</a>) wrote:</div> <div><blockquote type="cite" class="clean_bq" style="font-family: Helvetica, Arial; font-size: 13px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div><div></div><div>Hi, I've recently assembled a skylake system (i5 6400, AMD R9 380 [Tonga PRO], Gigabyte H170N), hoping to be able to use VFIO for PCI passthrough of the discrete graphics card to a Windows guest. I'm using qemu 2.4.1 on top of a bare Arch Linux instal; Windows 10 is installed as a KVM guest, using the UEFI images from [1]. After initial installation using the QXL video driver, I've swapped it for the AMD card. The guest boots without issues (aside from color corruption on HDMI output, which I assume has something to do with non-1080p output). However, roughly thirty seconds after initial bootup (guest is at the desktop) the guest appears to freeze (HDMI output drops, network connections disconnect). According to qemu metrics one guest core is pegged at 100% as it occurs. Interestingly, forcefully shutting down the guest and attempting to boot it again immediately puts it into the 'core pegged at 100%' state, without ever showing any sign of life. I've found a thread detailingr recent issues on Arch [2], but I don't experience slowdowns, just a full lockup. Similarly, the bug at [3] doesn't seem to be related either (temporarily assigned 2 GB to the guest just to be sure). Neither kernel 4.1.10, nor 4.2.5 works. Kernel and qemu logs show no interesting leads - qemu startup info and relevant dmesg entries + lspci/iommu is attached. Any ideas on how to debug this? [1]: https://www.kraxel.org/repos/jenkins/edk2/ [2]: https://bbs.archlinux.org/viewtopic.php?id=203240 [3]: https://bugzilla.kernel.org/show_bug.cgi?id=107561 _______________________________________________ vfio-users mailing list vfio-users@redhat.com https://www.redhat.com/mailman/listinfo/vfio-users </div></div></blockquote></div> Hi, I see you have this line in dmesg:[ 184.082279] kvm [557]: vcpu1 unhandled rdmsr: 0x641I think It’s related to the guest trying to read frequency scaling info from the CPU which it cannot.Have you tried to enable the ignore_msrs kvm kernel module parameter?<div><p style="color: rgb(0, 0, 0); font-variant: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;">echo 1 > /sys/module/kvm/parameters/ignore_msrs</div>Or you can also set it up in /etc/modprobe.d Best regards,-- Okky Hendriansyah<div></div></body></html>