[PATCH v7 8/8] KVM: selftests: Skip hardware_disable_test on powerpc pseries LPAR
flat view
COOLING5d
IN LINUX-NEXT
From: "Ritesh Harjani (IBM)" <ritesh.list@gmail.com>
Date: 2026-10-06 10:30:28
Also in:
kvm, lkml
Subsystem:
kernel selftest framework, kernel virtual machine (kvm), the rest · Maintainers:
Shuah Khan, Shuah Khan, Paolo Bonzini, Linus Torvalds
Revision v7 of 4 in this series; queued in linux-next as 23140d6e54ac on 2026-10-13.
Revisions (4)
- v3 [diff vs current]
- v3 [diff vs current]
- v5 [diff vs current]
- v7 current
The test forks 512 children, each of which does KVM_CREATE_VM and then SIGKILLs without waiting. On a PAPR LPAR that ioctl is H_GUEST_CREATE and unreaped guests may still occupy L0 Guest Management Space, which means create can fail with -ENOMEM error. Hence, skip the test when we are running on powerpc pseries LPAR. Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com> --- tools/testing/selftests/kvm/Makefile.kvm | 1 + tools/testing/selftests/kvm/hardware_disable_test.c | 3 +++ 2 files changed, 4 insertions(+)
diff --git a/tools/testing/selftests/kvm/Makefile.kvm b/tools/testing/selftests/kvm/Makefile.kvm
index 692bbae44a65..38c3900cac68 100644
--- a/tools/testing/selftests/kvm/Makefile.kvm
+++ b/tools/testing/selftests/kvm/Makefile.kvm@@ -258,6 +258,7 @@ TEST_GEN_PROGS_loongarch += steal_time TEST_GEN_PROGS_powerpc = $(TEST_GEN_PROGS_COMMON) TEST_GEN_PROGS_powerpc += access_tracking_perf_test +TEST_GEN_PROGS_powerpc += hardware_disable_test SPLIT_TESTS += arch_timer SPLIT_TESTS += get-reg-list
diff --git a/tools/testing/selftests/kvm/hardware_disable_test.c b/tools/testing/selftests/kvm/hardware_disable_test.c
index 43a36ef3ead8..0104b7e840b5 100644
--- a/tools/testing/selftests/kvm/hardware_disable_test.c
+++ b/tools/testing/selftests/kvm/hardware_disable_test.c@@ -130,6 +130,9 @@ int main(int argc, char **argv) int s, r, cpu, i; pid_t pid; + /* PowerPC PAPR L0 cannot guarantee 512 overlapping H_GUEST_CREATE from fork+SIGKILL. */ + TEST_REQUIRE(!kvm_is_ppc_pseries_lpar()); + kvm_sched_getaffinity(0, sizeof(cpu_set_t), &allowed_cpu_set); for (i = 0; i < NR_VCPUS && CPU_COUNT(&allowed_cpu_set); i++) {
--
2.39.5