== Comment: #0 - SEETEENA THOUFEEK <sthoufee.com> - 2025-05-23 00:00:35 == +++ This bug was initially created as a clone of Bug #208894 +++ This is identified as part of https://bugzilla.linux.ibm.com/show_bug.cgi?id=208629, where below two issues have been identified: 1. Libvirt needs support for P11 added 2. Guest when booted with Libvirt is not able to identify any disks and drops into dracut emergency shell while the same QCOW2 boots fine with a direct QEMU command line. The issues have been explained to Narayana offline and further details can be found in the comments for BZ208629. The following changes to Libvirt code fixing the issue and with custom build qemu-system-ppc64 (with patch series https://lore.kernel.org/all/20240731055022.696051-1-adityag@linux.ibm.com/ applied ), booting the guest : 1. /*The git diff output */ diff --git a/src/cpu/cpu_ppc64.c b/src/cpu/cpu_ppc64.c index 13f5fc9c2c..8673d4f759 100644 --- a/src/cpu/cpu_ppc64.c +++ b/src/cpu/cpu_ppc64.c @@ -93,22 +93,22 @@ ppc64CheckCompatibilityMode(const char *host_model, if (!compat_mode) return VIR_CPU_COMPARE_IDENTICAL; - /* Valid host CPUs: POWER6, POWER7, POWER8, POWER9, POWER10 */ + /* Valid host CPUs: POWER6, POWER7, POWER8, POWER9, POWER10, POWER11 */ if (!STRPREFIX(host_model, "POWER") || !(tmp = (char *) host_model + strlen("POWER")) || virStrToLong_i(tmp, NULL, 10, &host) < 0 || - host < 6 || host > 10) { + host < 6 || host > 11) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("Host CPU does not support compatibility modes")); return VIR_CPU_COMPARE_ERROR; } - /* Valid compatibility modes: power6, power7, power8, power9, power10 */ + /* Valid compatibility modes: power6, power7, power8, power9, power10, power11 */ if (!STRPREFIX(compat_mode, "power") || !(tmp = (char *) compat_mode + strlen("power")) || virStrToLong_i(tmp, NULL, 10, &compat) < 0 || - compat < 6 || compat > 10) { + compat < 6 || compat > 11) { virReportError(VIR_ERR_INTERNAL_ERROR, _("Unknown compatibility mode %1$s"), compat_mode); -------------------- 2. The guest console output for lscpu: ??[root@localhost]?[~] ???? # lscpu Architecture: ppc64le Byte Order: Little Endian CPU(s): 4 On-line CPU(s) list: 0-3 Model name: POWER10 (architected), altivec supported Model: 18.0 (pvr 0082 1200) Thread(s) per core: 1 Core(s) per socket: 1 Socket(s): 4 Virtualization features: Hypervisor vendor: KVM Virtualization type: para Caches (sum of all): L1d: 128 KiB (4 instances) L1i: 128 KiB (4 instances) NUMA: NUMA node(s): 1 NUMA node0 CPU(s): 0-3 Vulnerabilities: Gather data sampling: Not affected Itlb multihit: Not affected L1tf: Mitigation; RFI Flush, L1D private per thread Mds: Not affected Meltdown: Mitigation; RFI Flush, L1D private per thread Mmio stale data: Not affected Reg file data sampling: Not affected Retbleed: Not affected Spec rstack overflow: Not affected Spec store bypass: Mitigation; Kernel entry/exit barrier (eieio) Spectre v1: Mitigation; __user pointer sanitization Spectre v2: Mitigation; Software count cache flush (hardware accel erated), Software link stack flush Srbds: Not affected Tsx async abort: Not affected New patch for libvirt support for p11 and test capabilities (including fixes for compatibility mode test xml) are posted in up stream ( https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/ICUSILEYQYFTZMFTVUL2Q3H3LKIVSKXI/). The Main change in the patch is cpu-model name will be in lowercase (power11) similar to qemu as suggested by the maintainers. please integrate this patches The patches accepted upstream 7b0888b4ef cpu_ppc64: Add POWER11 host-model support e17aa7408c cpu_map: Add POWER11 CPU model support a714d53f74 tests: Add capabilities for QEMU 10.0.0 on ppc64 1c077f6a73 tests: qemuhotplugtest: Set the cpu version at source for PPC64 tests 0742b6487a tests: Pin pseries-2.7 tests to the version 7.0
Backporting these changes is kinda messy because of all the capabilities differences. Since this sounds like effectively new feature support, I think it's better to just say 'power 11 isn't supported until libvirt 11.4.0 which you can get from virt-preview repo' and fedora 43 libvirt will have native support. is that sufficient?
------- Comment From Narayana.Murty.Nagarampalli 2025-06-12 22:02 EDT------- Which RHEL version could we expect POWER11 supported on libvirt-11.4?
(In reply to IBM Bug Proxy from comment #3) > ------- Comment From Narayana.Murty.Nagarampalli 2025-06-12 22:02 > EDT------- > Which RHEL version could we expect POWER11 supported on libvirt-11.4? This is a Fedora bug tracker. If you have requests/expectations for RHEL, in general, you need to file a RHEL request. Note, however, that support for the ppc64 architecture was dropped from QEMU/KVM RHEL-9
------- Comment From Narayana.Murty.Nagarampalli 2025-09-09 07:39 EDT------- (In reply to comment #4) > Backporting these changes is kinda messy because of all the capabilities > differences. Since this sounds like effectively new feature support, I think > it's better to just say 'power 11 isn't supported until libvirt 11.4.0 which > you can get from virt-preview repo' and fedora 43 libvirt will have native > support. is that sufficient? I agree with this approach. Backporting the changes is indeed complicated due to capability differences, and given that this effectively introduces new feature support, it makes sense to state that Power11 support will come with libvirt 11.4.0 available via the virt-preview repository. Including this natively in Fedora 43 should be sufficient. No need to backport to earlier releases.
------- Comment From sthoufee.com 2025-09-22 04:13 EDT------- the status of RHEL bug is closed as WONT FIX. Is this changes would be integrated in F43 ?