On 12/06/2024 10:40 am, Damien Thenot wrote: > Hello, > > A XCP-ng 8.3 user that use Blue Iris Software encountered a crash with > Xen upgraded to version 4.17. > It worked correctly when XCP-ng 8.3 used Xen 4.13. > It is happening on Intel Xeon E-2378 CPU @ 2.60GHz CPUs and it seems > more recent Intel CPUs. > His guests are Windows with a NVIDIA GPU given to the guest. > > The user added: > > On an older box with i9-9900K CPU it does not happen and the VM works > as expected. Also working on an older > > Xeon Intel Xeon E-2146G and a E-2276G. Anything newer than that > however the VM will just BSOD. > > You can find more information in the XCP-ng forum post: > https://xcp-ng.org/forum/topic/8873/windows-blue-iris-xcp-ng-8-3 > > The user tried enabling `msr-relaxed` following notes in Xen 4.15 > documentation. > But it didn't change the behavior and the guest still crashes. > > Has someone else observed such behavior? > > Here is a Xen dmesg output with the error that the user was able to obtain: > > ``` > > (d1) [ 132.028963] xen|BUGCHECK: ====> > (d1) [ 132.029008] xen|BUGCHECK: SYSTEM_SERVICE_EXCEPTION: > 00000000C0000096 FFFFF80418A21E27 FFFFAC009F27B900 0000000000000000 > (d1) [ 132.029057] xen|BUGCHECK: EXCEPTION (FFFFF80418A21E27): > (d1) [ 132.029096] xen|BUGCHECK: - Code = 8589320F > (d1) [ 132.029134] xen|BUGCHECK: - Flags = 000000A8 > (d1) [ 132.029174] xen|BUGCHECK: - Address = 8589320F008EECE9 > (d1) [ 132.029214] xen|BUGCHECK: - Parameter[0] = 0F000001D9B90000 > (d1) [ 132.029255] xen|BUGCHECK: - Parameter[1] = 4166300FFCE08332 > (d1) [ 132.029297] xen|BUGCHECK: - Parameter[2] = 0355000002C881F7 > (d1) [ 132.029338] xen|BUGCHECK: - Parameter[3] = 0002A0818B497B74 > (d1) [ 132.029379] xen|BUGCHECK: - Parameter[4] = 000002A8918B4900 > (d1) [ 132.029421] xen|BUGCHECK: - Parameter[5] = 8B49CA230FC0230F > (d1) [ 132.029462] xen|BUGCHECK: - Parameter[6] = 918B49000002B081 > (d1) [ 132.029502] xen|BUGCHECK: - Parameter[7] = 0FD0230F000002B8 > (d1) [ 132.029544] xen|BUGCHECK: - Parameter[8] = 0002C8918B49DA23 > (d1) [ 132.029585] xen|BUGCHECK: - Parameter[9] = 230FF0230FC03300 > (d1) [ 132.029626] xen|BUGCHECK: - Parameter[10] = 008AE22504F665FA > (d1) [ 132.029667] xen|BUGCHECK: - Parameter[11] = 00C2F76639740200 > (d1) [ 132.029709] xen|BUGCHECK: - Parameter[12] = 81342405F7327403 > (d1) [ 132.029753] xen|BUGCHECK: - Parameter[13] = 6626750000000200 > (d1) [ 132.029794] xen|BUGCHECK: - Parameter[14] = C88303740200C2F7 > (d1) [ 132.029835] xen|BUGCHECK: EXCEPTION (0D8B000000AC9589): > (d1) [ 132.029888] xen|BUGCHECK: CONTEXT (FFFFAC009F27B900): > (d1) [ 132.029925] xen|BUGCHECK: - GS = 002B > (d1) [ 132.029959] xen|BUGCHECK: - FS = 0053 > (d1) [ 132.029994] xen|BUGCHECK: - ES = 002B > (d1) [ 132.030028] xen|BUGCHECK: - DS = 002B > (d1) [ 132.030056] xen|BUGCHECK: - SS = 0018 > (d1) [ 132.030089] xen|BUGCHECK: - CS = 0010 > (d1) [ 132.030123] xen|BUGCHECK: - EFLAGS = 00040046 > (d1) [ 132.030160] xen|BUGCHECK: - RDI = 00000000000002C4 > (d1) [ 132.030199] xen|BUGCHECK: - RSI = 00000000005FFA48 > (d1) [ 132.030237] xen|BUGCHECK: - RBX = 00000000426ED080 > (d1) [ 132.030275] xen|BUGCHECK: - RDX = 0000000000000000 > (d1) [ 132.030312] xen|BUGCHECK: - RCX = 00000000000001DD > (d1) [ 132.030349] xen|BUGCHECK: - RAX = 0000000000000000 > (d1) [ 132.030386] xen|BUGCHECK: - RBP = 00000000E4427520 > (d1) [ 132.030424] xen|BUGCHECK: - RIP = 0000000018A21E27 > (d1) [ 132.030463] xen|BUGCHECK: - RSP = 00000000E4427498 > (d1) [ 132.030504] xen|BUGCHECK: - R8 = 0000000000000000 > (d1) [ 132.030543] xen|BUGCHECK: - R9 = 000000009F25A000 > (d1) [ 132.030580] xen|BUGCHECK: - R10 = 00000000000002C4 > (d1) [ 132.030618] xen|BUGCHECK: - R11 = 0000000000000246 > (d1) [ 132.030657] xen|BUGCHECK: - R12 = 0000000002CEB528 > (d1) [ 132.030696] xen|BUGCHECK: - R13 = 00000000A81E3A80 > (d1) [ 132.030735] xen|BUGCHECK: - R14 = 00000000000002C4 > (d1) [ 132.030775] xen|BUGCHECK: - R15 = 0000000000000000 > (d1) [ 132.030812] xen|BUGCHECK: STACK: > (d1) [ 132.030858] xen|BUGCHECK: 00000000E44274A0: (00000000426ED080 > 00000000005FF898 0000000000000000 0000000000000000) ntoskrnl.exe + > 000000000043667A > (d1) [ 132.030935] xen|BUGCHECK: 00000000005FFA18: (00000000A74AD76E > 0000000000000000 00000000A81E3A80 000000000000000E) 00007FFBA9BFF4D4 > (d1) [ 132.031010] xen|BUGCHECK: 00000000005FFA20: (0000000000000000 > 00000000A81E3A80 000000000000000E 0000000000000003) 00007FFBA74AD76E > (d1) [ 132.031086] xen|BUGCHECK: 00000000005FFA28: (00000000A81E3A80 > 000000000000000E 0000000000000003 00000000005FFB10) 0000000000000000 > (d1) [ 132.031151] xen|BUGCHECK: <==== > (XEN) [ 132.040828] memory_map:remove: dom1 gfn=f60c0 mfn=a3080 nr=4 > (XEN) [ 132.040975] memory_map:remove: dom1 gfn=f5000 mfn=a2000 nr=1000 > (XEN) [ 132.041124] memory_map:remove: dom1 gfn=e0000 mfn=4000000 nr=10000 > (XEN) [ 132.041463] memory_map:remove: dom1 gfn=e8000 mfn=4008000 nr=8000 > (XEN) [ 132.041905] memory_map:remove: dom1 gfn=f8000 mfn=4010000 nr=2000 > (XEN) [ 132.042072] ioport_map:remove: dom1 gport=c100 mport=6000 nr=80 > ```
There was a #GP fault accessing MSR 0x1DD You'll need to investigate how that MSR is supposed to behave on real hardware, and see if Xen's behaviour matches or not. ~Andrew