Re: [BUG] [powerpc] KVM guest boot failure - hangs on startup after commit 98884e0c

[Misbah Anjum N <misanjum@linux.ibm.com>]

Hi Ani,
I've completed the testing you suggested. Unfortunately, the SMP hang 
still persists with these changes.

Changes made:
As requested, I reverted everything in stubs/kvm.c and the related 
changes in stubs/meson.build, include/system/kvm.h, and 
target/i386/kvm/kvm.c. I also commented out the calls to 
kvm_arch_supports_vmfd_change() and kvm_arch_on_vmfd_change() in 
kvm_reset_vmfd().

Test result:
The issue persists - guests still hang indefinitely during boot when SMP 
is configured.

Git diff:
diff --git a/accel/kvm/kvm-all.c b/accel/kvm/kvm-all.c
index cc5c42ce4d..04b9cbe7c9 100644
--- a/accel/kvm/kvm-all.c
+++ b/accel/kvm/kvm-all.c
@@ -2622,11 +2622,12 @@ static int kvm_reset_vmfd(MachineState *ms)
       * bail if the current architecture does not support VM file
       * descriptor change.
       */
-    if (!kvm_arch_supports_vmfd_change()) {
+    /*if (!kvm_arch_supports_vmfd_change()) {
          error_report("This target architecture does not support KVM VM 
"
                       "file descriptor change.");
          return -EOPNOTSUPP;
      }
+    */

      s = KVM_STATE(ms->accelerator);
      kml = &s->memory_listener;
@@ -2659,10 +2660,10 @@ static int kvm_reset_vmfd(MachineState *ms)
      }
      assert(!err);

-    ret = kvm_arch_on_vmfd_change(ms, s);
+    /*ret = kvm_arch_on_vmfd_change(ms, s);
      if (ret < 0) {
          return ret;
-    }
+    }*/

      if (s->kernel_irqchip_allowed) {
          do_kvm_irqchip_create(s);

diff --git a/include/system/kvm.h b/include/system/kvm.h
index 5fc7251fd9..0dad0079ed 100644
--- a/include/system/kvm.h
+++ b/include/system/kvm.h
@@ -456,8 +456,6 @@ int kvm_physical_memory_addr_from_host(KVMState *s, 
void *ram_addr,

  #endif /* COMPILING_PER_TARGET */

-bool kvm_arch_supports_vmfd_change(void);
-int kvm_arch_on_vmfd_change(MachineState *ms, KVMState *s);

  void kvm_cpu_synchronize_state(CPUState *cpu);

diff --git a/stubs/kvm.c b/stubs/kvm.c
deleted file mode 100644
index 2db61d89a7..0000000000
--- a/stubs/kvm.c
+++ /dev/null
@@ -1,22 +0,0 @@
-/*
- * kvm target arch specific stubs
- *
- * Copyright (c) 2026 Red Hat, Inc.
- *
- * Author:
- *   Ani Sinha <anisinha@redhat.com>
- *
- * SPDX-License-Identifier: GPL-2.0-or-later
- */
-#include "qemu/osdep.h"
-#include "system/kvm.h"
-
-int kvm_arch_on_vmfd_change(MachineState *ms, KVMState *s)
-{
-    abort();
-}
-
-bool kvm_arch_supports_vmfd_change(void)
-{
-    return false;
-}

diff --git a/stubs/meson.build b/stubs/meson.build
index 6ae478bacc..8a07059500 100644
--- a/stubs/meson.build
+++ b/stubs/meson.build
@@ -74,7 +74,6 @@ if have_system
    if igvm.found()
      stub_ss.add(files('igvm.c'))
    endif
-  stub_ss.add(files('kvm.c'))
    stub_ss.add(files('target-get-monitor-def.c'))
    stub_ss.add(files('target-monitor-defs.c'))
    stub_ss.add(files('win32-kbd-hook.c'))

diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 524b5276a6..3dfd9a5974 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -3389,15 +3389,6 @@ static int kvm_vm_enable_energy_msrs(KVMState *s)
      return 0;
  }

-int kvm_arch_on_vmfd_change(MachineState *ms, KVMState *s)
-{
-    abort();
-}
-
-bool kvm_arch_supports_vmfd_change(void)
-{
-    return false;
-}

  int kvm_arch_init(MachineState *ms, KVMState *s)
  {


I've also tested with the latest QEMU build from master, and the issue 
still persists there as well. Could you suggest what additional 
debugging steps I should take to help identify the root cause?

Thanks,
Misbah Anjum N <misanjum@linux.ibm.com>


On 2026-03-18 15:00, Ani Sinha wrote:
> One possible thing to try is:
> 
> Revert everything in stubs/kvm.c and hence changes in
> stubs/meson.build, include/system/kvm.h and in target/i386/kvm/kvm.c
> introduced by 98884e0cc1 .
> You will have to comment out calls to kvm_arch_supports_vmfd_change()
> and kvm_arch_on_vmfd_change() in kvm_reset_vmfd(). Since
> kvm_reset_vmfd() is not called anyway, not should make no difference
> if those calls are commented out.
> 
> Let me know what you get after doing the above.