On 5/24/2025 10:09 AM, Manuel Bouyer wrote: > On Sat, May 24, 2025 at 08:16:42AM -0400, Chuck Zmudzinski wrote: >> > >> > N.B. My cpu has 14 cores and 20 threads because there are 6 cores that >> > have hyperthreading - so it hangs where the hyperthreading starts. Last >> > thing to try is with smt=false... >> >> >> Adding smt=false to the Xen command line gets me a little farther but it >> still hangs: >> >> snip ... >> >> [ 1.0000000] vcpu13 at hypervisor0 >> >> [ 1.0000000] vcpu13: Intel(R) Core(TM) i5-14500, id 0xb06f2 >> >> [ 1.0000000] vcpu13: node 0, package 0, core 0, smt 0 >> >> [ 1.0000000] xenbus0 at hypervisor0: Xen Virtual Bus Interface >> >> [ 1.0000000] xencons0 at hypervisor0: Xen Virtual Console Driver >> >> [ 1.0000000] acpi0 at hypervisor0: Intel ACPICA 20221020 >> >> [ 1.0000000] ACPI: Dynamic OEM Table Load: >> >> [ 1.0000000] ACPI: SSDT 0xFFFFAE80049B3808 000394 (v02 PmRef Cpu0Cst >> 00003001 INTL 20200717) >> >> [ 1.0000000] ACPI: Dynamic OEM Table Load: >> >> [ 1.0000000] ACPI: SSDT 0xFFFFAE80049B900 >> >> Then it hangs... > > Could you try with > conring_size=4096k sync_console > on the Xen command line, with or without smt ? > the fact that it stops in the middle of a line with smt makes me wonder if > we're really seeing the whole messages > >
I can try that in several hours I think I was getting lots of extra data in the buffer because of those (XEN) altcall messages - I am away from the box right now but I will try it when I get back home. I also plan to compare Fedora's Xen build options with the pkgsrc Xen build options. I think Fedora 40 is using Xen 4.18.x - I am curious why the pkgsrc build of Xen 4.18 crashes unless I set the cet=no-ibt option, but the Fedora build of Xen 4.18 used to boot fine on the same box without needing to adjust the cet setting when I was running Fedora 40 with Xen 4.18.x on the same box.
