[PATCH] target/i386: clear CPU_INTERRUPT_SIPI for all accelerators

Paolo Bonzini posted 1 patch 3 days, 1 hour ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20251024071918.228749-1-pbonzini@redhat.com
Maintainers: Cameron Esfahani <dirty@apple.com>, Roman Bolshakov <rbolshakov@ddn.com>, Phil Dennis-Jordan <phil@philjordan.eu>, Paolo Bonzini <pbonzini@redhat.com>, Marcelo Tosatti <mtosatti@redhat.com>, Reinoud Zandijk <reinoud@netbsd.org>, Sunil Muthuswamy <sunilmut@microsoft.com>
target/i386/hvf/x86hvf.c    | 1 +
target/i386/kvm/kvm.c       | 1 +
target/i386/nvmm/nvmm-all.c | 1 +
target/i386/whpx/whpx-all.c | 1 +
4 files changed, 4 insertions(+)
[PATCH] target/i386: clear CPU_INTERRUPT_SIPI for all accelerators
Posted by Paolo Bonzini 3 days, 1 hour ago
Similar to what commit df32e5c5 did for TCG; fixes boot with multiple
processors on WHPX and probably more accelerators

Fixes: df32e5c568c ("i386/cpu: Prevent delivering SIPI during SMM in TCG mode", 2025-10-14)
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/3178
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 target/i386/hvf/x86hvf.c    | 1 +
 target/i386/kvm/kvm.c       | 1 +
 target/i386/nvmm/nvmm-all.c | 1 +
 target/i386/whpx/whpx-all.c | 1 +
 4 files changed, 4 insertions(+)

diff --git a/target/i386/hvf/x86hvf.c b/target/i386/hvf/x86hvf.c
index a502437c303..3838c9f5a6f 100644
--- a/target/i386/hvf/x86hvf.c
+++ b/target/i386/hvf/x86hvf.c
@@ -446,6 +446,7 @@ int hvf_process_events(CPUState *cs)
         cs->halted = 0;
     }
     if (cpu_test_interrupt(cs, CPU_INTERRUPT_SIPI)) {
+        cpu_reset_interrupt(cs, CPU_INTERRUPT_SIPI);
         cpu_synchronize_state(cs);
         do_cpu_sipi(cpu);
     }
diff --git a/target/i386/kvm/kvm.c b/target/i386/kvm/kvm.c
index 3095625b525..23a0329d7a5 100644
--- a/target/i386/kvm/kvm.c
+++ b/target/i386/kvm/kvm.c
@@ -5659,6 +5659,7 @@ int kvm_arch_process_async_events(CPUState *cs)
         cs->halted = 0;
     }
     if (cpu_test_interrupt(cs, CPU_INTERRUPT_SIPI)) {
+        cpu_reset_interrupt(cs, CPU_INTERRUPT_SIPI);
         kvm_cpu_synchronize_state(cs);
         do_cpu_sipi(cpu);
     }
diff --git a/target/i386/nvmm/nvmm-all.c b/target/i386/nvmm/nvmm-all.c
index 2e442baf4b7..60c9b975562 100644
--- a/target/i386/nvmm/nvmm-all.c
+++ b/target/i386/nvmm/nvmm-all.c
@@ -709,6 +709,7 @@ nvmm_vcpu_loop(CPUState *cpu)
         cpu->halted = false;
     }
     if (cpu_test_interrupt(cpu, CPU_INTERRUPT_SIPI)) {
+        cpu_reset_interrupt(cs, CPU_INTERRUPT_SIPI);
         nvmm_cpu_synchronize_state(cpu);
         do_cpu_sipi(x86_cpu);
     }
diff --git a/target/i386/whpx/whpx-all.c b/target/i386/whpx/whpx-all.c
index 256761834c9..1cb33001cc4 100644
--- a/target/i386/whpx/whpx-all.c
+++ b/target/i386/whpx/whpx-all.c
@@ -1621,6 +1621,7 @@ static void whpx_vcpu_process_async_events(CPUState *cpu)
     }
 
     if (cpu_test_interrupt(cpu, CPU_INTERRUPT_SIPI)) {
+        cpu_reset_interrupt(cs, CPU_INTERRUPT_SIPI);
         whpx_cpu_synchronize_state(cpu);
         do_cpu_sipi(x86_cpu);
     }
-- 
2.51.0
Re: [PATCH] target/i386: clear CPU_INTERRUPT_SIPI for all accelerators
Posted by Peter Maydell 2 days, 22 hours ago
On Fri, 24 Oct 2025 at 08:19, Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> Similar to what commit df32e5c5 did for TCG; fixes boot with multiple
> processors on WHPX and probably more accelerators
>
> Fixes: df32e5c568c ("i386/cpu: Prevent delivering SIPI during SMM in TCG mode", 2025-10-14)
> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/3178
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

This should be
Cc: qemu-stable@nongnu.org

because although df32e5c568c itself has not yet been in a
mainline release, it was cherry-picked back to the 10.1 and
10.0 branches.

thanks
-- PMM