Message ID | 20240318165545.3898-34-anthony.perard@citrix.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | Switch to Debian Bookworm | expand |
On Mon, Mar 18, 2024 at 04:55:42PM +0000, Anthony PERARD wrote: > When starting the installation of the L2 guest, L0 kills L1. Switching > the L2 guest back to Debian Buster works fine, so do that to prevent > regression in the test. > > Part of the logs from the host L0: > > > domain_crash called from arch/x86/hvm/vmx/vvmx.c:2770 > > Domain 3 (vcpu#0) crashed on cpu#4: > > d3v0 vmentry failure (reason 0x80000021): Invalid guest state (2) Hm, I guess we need this as otherwise the changes themselves won't past the self-push gate, but this is just masking a real issue. I would be tempted to (after the switch to bookworm has gone in) revert this and force-push it into osstest, so that the failure is visible. Thanks, Roger.
On Wed, Mar 20, 2024 at 06:24:18PM +0100, Roger Pau Monné wrote: > On Mon, Mar 18, 2024 at 04:55:42PM +0000, Anthony PERARD wrote: > > When starting the installation of the L2 guest, L0 kills L1. Switching > > the L2 guest back to Debian Buster works fine, so do that to prevent > > regression in the test. > > > > Part of the logs from the host L0: > > > > > domain_crash called from arch/x86/hvm/vmx/vvmx.c:2770 > > > Domain 3 (vcpu#0) crashed on cpu#4: > > > d3v0 vmentry failure (reason 0x80000021): Invalid guest state (2) > > Hm, I guess we need this as otherwise the changes themselves won't > past the self-push gate, but this is just masking a real issue. I > would be tempted to (after the switch to bookworm has gone in) revert > this and force-push it into osstest, so that the failure is visible. That means that we might have to force-push on every branches. Or fix Xen. Thanks,
diff --git a/make-flight b/make-flight index 155a0c1f..d0c950bc 100755 --- a/make-flight +++ b/make-flight @@ -360,6 +360,19 @@ do_hvm_debian_nested_tests () { xen-4.3-testing) return;; esac + local l2_runvar + case $guestsuite in + bookworm) + # Bookworm install image lead to a crash of l1, so keep using + # Buster's image. + l2_runvar=(l2_suite=buster) + l2_runvar+=(l2_image=$(usual_debianhvm_image amd64 buster)) + ;; + *) + l2_runvar=(l2_image=$(usual_debianhvm_image amd64)) + ;; + esac + for cpuvendor in amd intel; do job_create_test test-$xenarch$kern-$dom0arch$qemuu_suffix-nested-$cpuvendor \ @@ -368,7 +381,7 @@ do_hvm_debian_nested_tests () { l1_vifmodel='e1000' \ l1_memsize='3072' \ l1_enable_nestedhvm='true' \ - l2_image=$(usual_debianhvm_image amd64) \ + ${l2_runvar[@]} \ bios=$bios \ all_hostflags=$most_hostflags,hvm-$cpuvendor diff --git a/mfi-common b/mfi-common index 6dc39422..bbe714bf 100644 --- a/mfi-common +++ b/mfi-common @@ -547,18 +547,19 @@ job_create_test () { usual_debianhvm_image () { local arch=$1; shift + local suite=${1:-$guestsuite} if [ -n "$DEBIAN_IMAGE_FILE" ]; then echo $DEBIAN_IMAGE_FILE return fi - local file=`getconfig DebianImageFile_${guestsuite}_${arch}` + local file=`getconfig DebianImageFile_${suite}_${arch}` if [ -n "$file" ]; then echo $file return fi local ver=$DEBIAN_IMAGE_VERSION if [ -z "$ver" ] ; then - ver=`getconfig DebianImageVersion_$guestsuite` + ver=`getconfig DebianImageVersion_$suite` fi echo debian-$ver-$arch-CD-1.iso }