Message ID | 20190523164309.13345-1-thuth@redhat.com (mailing list archive) |
---|---|
Headers | show |
Series | KVM selftests for s390x | expand |
I do get [10400.440298] kvm-s390: failed to commit memory region [10400.508723] kvm-s390: failed to commit memory region when running the tests. Will have a look. On 23.05.19 18:43, Thomas Huth wrote: > This patch series enables the KVM selftests for s390x. As a first > test, the sync_regs from x86 has been adapted to s390x, and after > a fix for KVM_CAP_MAX_VCPU_ID on s390x, the kvm_create_max_vcpus > is now enabled here, too. > > Please note that the ucall() interface is not used yet - since > s390x neither has PIO nor MMIO, this needs some more work first > before it becomes usable (we likely should use a DIAG hypercall > here, which is what the sync_reg test is currently using, too... > I started working on that topic, but did not finish that work > yet, so I decided to not include it yet). > > RFC -> v1: > - Rebase, needed to add the first patch for vcpu_nested_state_get/set > - Added patch to introduce VM_MODE_DEFAULT macro > - Improved/cleaned up the code in processor.c > - Added patch to fix KVM_CAP_MAX_VCPU_ID on s390x > - Added patch to enable the kvm_create_max_vcpus on s390x and aarch64 > > Andrew Jones (1): > kvm: selftests: aarch64: fix default vm mode > > Thomas Huth (8): > KVM: selftests: Wrap vcpu_nested_state_get/set functions with x86 > guard > KVM: selftests: Guard struct kvm_vcpu_events with > __KVM_HAVE_VCPU_EVENTS > KVM: selftests: Introduce a VM_MODE_DEFAULT macro for the default bits > KVM: selftests: Align memory region addresses to 1M on s390x > KVM: selftests: Add processor code for s390x > KVM: selftests: Add the sync_regs test for s390x > KVM: s390: Do not report unusabled IDs via KVM_CAP_MAX_VCPU_ID > KVM: selftests: Move kvm_create_max_vcpus test to generic code > > MAINTAINERS | 2 + > arch/mips/kvm/mips.c | 3 + > arch/powerpc/kvm/powerpc.c | 3 + > arch/s390/kvm/kvm-s390.c | 1 + > arch/x86/kvm/x86.c | 3 + > tools/testing/selftests/kvm/Makefile | 7 +- > .../testing/selftests/kvm/include/kvm_util.h | 10 + > .../selftests/kvm/include/s390x/processor.h | 22 ++ > .../kvm/{x86_64 => }/kvm_create_max_vcpus.c | 3 +- > .../selftests/kvm/lib/aarch64/processor.c | 2 +- > tools/testing/selftests/kvm/lib/kvm_util.c | 25 +- > .../selftests/kvm/lib/s390x/processor.c | 286 ++++++++++++++++++ > .../selftests/kvm/lib/x86_64/processor.c | 2 +- > .../selftests/kvm/s390x/sync_regs_test.c | 151 +++++++++ > virt/kvm/arm/arm.c | 3 + > virt/kvm/kvm_main.c | 2 - > 16 files changed, 514 insertions(+), 11 deletions(-) > create mode 100644 tools/testing/selftests/kvm/include/s390x/processor.h > rename tools/testing/selftests/kvm/{x86_64 => }/kvm_create_max_vcpus.c (93%) > create mode 100644 tools/testing/selftests/kvm/lib/s390x/processor.c > create mode 100644 tools/testing/selftests/kvm/s390x/sync_regs_test.c >
On 24.05.19 13:11, Christian Borntraeger wrote: > I do get > > [10400.440298] kvm-s390: failed to commit memory region > [10400.508723] kvm-s390: failed to commit memory region > > when running the tests. Will have a look. It comes from kvm_vm_free. This calls KVM_SET_USER_MEMORY_REGION with size 0, which the s390 code does not like.
On 24.05.19 14:17, Christian Borntraeger wrote: > > > On 24.05.19 13:11, Christian Borntraeger wrote: >> I do get >> >> [10400.440298] kvm-s390: failed to commit memory region >> [10400.508723] kvm-s390: failed to commit memory region >> >> when running the tests. Will have a look. > > It comes from kvm_vm_free. This calls KVM_SET_USER_MEMORY_REGION with size 0, > which the s390 code does not like. > The doc says about KVM_SET_USER_MEMORY_REGION: This ioctl allows the user to create or modify a guest physical memory slot. When changing an existing slot, it may be moved in the guest physical memory space, or its flags may be modified. --> It may not be resized. <---- $ strace -f -e trace=ioctl tools/testing/selftests/kvm/s390x/sync_regs_test ioctl(3, KVM_CHECK_EXTENSION, KVM_CAP_SYNC_REGS) = 1 ioctl(4, KVM_CHECK_EXTENSION, KVM_CAP_IMMEDIATE_EXIT) = 1 ioctl(3, KVM_CREATE_VM, 0) = 4 ioctl(4, KVM_SET_USER_MEMORY_REGION, {slot=0, flags=0, guest_phys_addr=0, memory_size=2097152, userspace_addr=0x3ffac500000}) = 0 ioctl(4, KVM_CREATE_VCPU, 5) = 7 ioctl(8, KVM_GET_VCPU_MMAP_SIZE, 0) = 4096 ioctl(8, KVM_GET_VCPU_MMAP_SIZE, 0) = 4096 ioctl(7, KVM_GET_SREGS, 0x3ffef0fdb90) = 0 ioctl(7, KVM_SET_SREGS, 0x3ffef0fdb90) = 0 ioctl(7, KVM_GET_REGS, 0x3ffef0fdcf8) = 0 ioctl(7, KVM_SET_REGS, 0x3ffef0fdcf8) = 0 ioctl(7, KVM_GET_SREGS, 0x3ffef0fdd78) = 0 ioctl(7, KVM_SET_SREGS, 0x3ffef0fdd78) = 0 ioctl(7, KVM_RUN, 0) = 0 ioctl(7, KVM_GET_REGS, 0x3ffef0fdf90) = 0 ioctl(7, KVM_GET_SREGS, 0x3ffef0fe010) = 0 ioctl(7, KVM_RUN, 0) = 0 ioctl(7, KVM_GET_REGS, 0x3ffef0fdf90) = 0 ioctl(7, KVM_GET_SREGS, 0x3ffef0fe010) = 0 ioctl(7, KVM_RUN, 0) = 0 ioctl(4, KVM_SET_USER_MEMORY_REGION, {slot=0, flags=0, guest_phys_addr=0, memory_size=0, userspace_addr=0x3ffac500000}) = 0 +++ exited with 0 +++ So the testcase is wrong? (I think the s390 code is also not fully correct will double check)
On 24.05.19 14:29, Christian Borntraeger wrote: > > > On 24.05.19 14:17, Christian Borntraeger wrote: >> >> >> On 24.05.19 13:11, Christian Borntraeger wrote: >>> I do get >>> >>> [10400.440298] kvm-s390: failed to commit memory region >>> [10400.508723] kvm-s390: failed to commit memory region >>> >>> when running the tests. Will have a look. >> >> It comes from kvm_vm_free. This calls KVM_SET_USER_MEMORY_REGION with size 0, >> which the s390 code does not like. >> > > The doc says about KVM_SET_USER_MEMORY_REGION: > > This ioctl allows the user to create or modify a guest physical memory > slot. When changing an existing slot, it may be moved in the guest > physical memory space, or its flags may be modified. --> It may not be > resized. <---- Size 0 is deleting, not resizing AFAIK.
On 24.05.19 14:36, David Hildenbrand wrote: > On 24.05.19 14:29, Christian Borntraeger wrote: >> >> >> On 24.05.19 14:17, Christian Borntraeger wrote: >>> >>> >>> On 24.05.19 13:11, Christian Borntraeger wrote: >>>> I do get >>>> >>>> [10400.440298] kvm-s390: failed to commit memory region >>>> [10400.508723] kvm-s390: failed to commit memory region >>>> >>>> when running the tests. Will have a look. >>> >>> It comes from kvm_vm_free. This calls KVM_SET_USER_MEMORY_REGION with size 0, >>> which the s390 code does not like. >>> >> >> The doc says about KVM_SET_USER_MEMORY_REGION: >> >> This ioctl allows the user to create or modify a guest physical memory >> slot. When changing an existing slot, it may be moved in the guest >> physical memory space, or its flags may be modified. --> It may not be >> resized. <---- > > Size 0 is deleting, not resizing AFAIK. Right this seems to translate to KVM_MR_DELETE, which the s390 code does not handle (we will simply deliver a page fault as we share the last page table level). I will have a look at implementing KVM_MR_DELETE and KVM_MR_MOVE. In fact, we should have a testcase for that as well.
I have now queued every patch in the kselftest branch on kvms390 at kernel.org. I will push out for next as soon as I have some ack/nacks on the "KVM: s390: fix memory slot handling for KVM_SET_USER_MEMORY_REGION" patch. On 23.05.19 18:43, Thomas Huth wrote: > This patch series enables the KVM selftests for s390x. As a first > test, the sync_regs from x86 has been adapted to s390x, and after > a fix for KVM_CAP_MAX_VCPU_ID on s390x, the kvm_create_max_vcpus > is now enabled here, too. > > Please note that the ucall() interface is not used yet - since > s390x neither has PIO nor MMIO, this needs some more work first > before it becomes usable (we likely should use a DIAG hypercall > here, which is what the sync_reg test is currently using, too... > I started working on that topic, but did not finish that work > yet, so I decided to not include it yet). > > RFC -> v1: > - Rebase, needed to add the first patch for vcpu_nested_state_get/set > - Added patch to introduce VM_MODE_DEFAULT macro > - Improved/cleaned up the code in processor.c > - Added patch to fix KVM_CAP_MAX_VCPU_ID on s390x > - Added patch to enable the kvm_create_max_vcpus on s390x and aarch64 > > Andrew Jones (1): > kvm: selftests: aarch64: fix default vm mode > > Thomas Huth (8): > KVM: selftests: Wrap vcpu_nested_state_get/set functions with x86 > guard > KVM: selftests: Guard struct kvm_vcpu_events with > __KVM_HAVE_VCPU_EVENTS > KVM: selftests: Introduce a VM_MODE_DEFAULT macro for the default bits > KVM: selftests: Align memory region addresses to 1M on s390x > KVM: selftests: Add processor code for s390x > KVM: selftests: Add the sync_regs test for s390x > KVM: s390: Do not report unusabled IDs via KVM_CAP_MAX_VCPU_ID > KVM: selftests: Move kvm_create_max_vcpus test to generic code > > MAINTAINERS | 2 + > arch/mips/kvm/mips.c | 3 + > arch/powerpc/kvm/powerpc.c | 3 + > arch/s390/kvm/kvm-s390.c | 1 + > arch/x86/kvm/x86.c | 3 + > tools/testing/selftests/kvm/Makefile | 7 +- > .../testing/selftests/kvm/include/kvm_util.h | 10 + > .../selftests/kvm/include/s390x/processor.h | 22 ++ > .../kvm/{x86_64 => }/kvm_create_max_vcpus.c | 3 +- > .../selftests/kvm/lib/aarch64/processor.c | 2 +- > tools/testing/selftests/kvm/lib/kvm_util.c | 25 +- > .../selftests/kvm/lib/s390x/processor.c | 286 ++++++++++++++++++ > .../selftests/kvm/lib/x86_64/processor.c | 2 +- > .../selftests/kvm/s390x/sync_regs_test.c | 151 +++++++++ > virt/kvm/arm/arm.c | 3 + > virt/kvm/kvm_main.c | 2 - > 16 files changed, 514 insertions(+), 11 deletions(-) > create mode 100644 tools/testing/selftests/kvm/include/s390x/processor.h > rename tools/testing/selftests/kvm/{x86_64 => }/kvm_create_max_vcpus.c (93%) > create mode 100644 tools/testing/selftests/kvm/lib/s390x/processor.c > create mode 100644 tools/testing/selftests/kvm/s390x/sync_regs_test.c >
On 23/05/19 18:43, Thomas Huth wrote: > This patch series enables the KVM selftests for s390x. As a first > test, the sync_regs from x86 has been adapted to s390x, and after > a fix for KVM_CAP_MAX_VCPU_ID on s390x, the kvm_create_max_vcpus > is now enabled here, too. > > Please note that the ucall() interface is not used yet - since > s390x neither has PIO nor MMIO, this needs some more work first > before it becomes usable (we likely should use a DIAG hypercall > here, which is what the sync_reg test is currently using, too... > I started working on that topic, but did not finish that work > yet, so I decided to not include it yet). Christian, please include this in your tree (rebasing on top of kvm/next as soon as I push it). Note that Thomas is away for about a month. Paolo > RFC -> v1: > - Rebase, needed to add the first patch for vcpu_nested_state_get/set > - Added patch to introduce VM_MODE_DEFAULT macro > - Improved/cleaned up the code in processor.c > - Added patch to fix KVM_CAP_MAX_VCPU_ID on s390x > - Added patch to enable the kvm_create_max_vcpus on s390x and aarch64 > > Andrew Jones (1): > kvm: selftests: aarch64: fix default vm mode > > Thomas Huth (8): > KVM: selftests: Wrap vcpu_nested_state_get/set functions with x86 > guard > KVM: selftests: Guard struct kvm_vcpu_events with > __KVM_HAVE_VCPU_EVENTS > KVM: selftests: Introduce a VM_MODE_DEFAULT macro for the default bits > KVM: selftests: Align memory region addresses to 1M on s390x > KVM: selftests: Add processor code for s390x > KVM: selftests: Add the sync_regs test for s390x > KVM: s390: Do not report unusabled IDs via KVM_CAP_MAX_VCPU_ID > KVM: selftests: Move kvm_create_max_vcpus test to generic code > > MAINTAINERS | 2 + > arch/mips/kvm/mips.c | 3 + > arch/powerpc/kvm/powerpc.c | 3 + > arch/s390/kvm/kvm-s390.c | 1 + > arch/x86/kvm/x86.c | 3 + > tools/testing/selftests/kvm/Makefile | 7 +- > .../testing/selftests/kvm/include/kvm_util.h | 10 + > .../selftests/kvm/include/s390x/processor.h | 22 ++ > .../kvm/{x86_64 => }/kvm_create_max_vcpus.c | 3 +- > .../selftests/kvm/lib/aarch64/processor.c | 2 +- > tools/testing/selftests/kvm/lib/kvm_util.c | 25 +- > .../selftests/kvm/lib/s390x/processor.c | 286 ++++++++++++++++++ > .../selftests/kvm/lib/x86_64/processor.c | 2 +- > .../selftests/kvm/s390x/sync_regs_test.c | 151 +++++++++ > virt/kvm/arm/arm.c | 3 + > virt/kvm/kvm_main.c | 2 - > 16 files changed, 514 insertions(+), 11 deletions(-) > create mode 100644 tools/testing/selftests/kvm/include/s390x/processor.h > rename tools/testing/selftests/kvm/{x86_64 => }/kvm_create_max_vcpus.c (93%) > create mode 100644 tools/testing/selftests/kvm/lib/s390x/processor.c > create mode 100644 tools/testing/selftests/kvm/s390x/sync_regs_test.c >
On 04.06.19 19:19, Paolo Bonzini wrote: > On 23/05/19 18:43, Thomas Huth wrote: >> This patch series enables the KVM selftests for s390x. As a first >> test, the sync_regs from x86 has been adapted to s390x, and after >> a fix for KVM_CAP_MAX_VCPU_ID on s390x, the kvm_create_max_vcpus >> is now enabled here, too. >> >> Please note that the ucall() interface is not used yet - since >> s390x neither has PIO nor MMIO, this needs some more work first >> before it becomes usable (we likely should use a DIAG hypercall >> here, which is what the sync_reg test is currently using, too... >> I started working on that topic, but did not finish that work >> yet, so I decided to not include it yet). > > Christian, please include this in your tree (rebasing on top of kvm/next > as soon as I push it). Note that Thomas is away for about a month. Will do. Right now it is part of my next tree on top of rc3. > > Paolo > >> RFC -> v1: >> - Rebase, needed to add the first patch for vcpu_nested_state_get/set >> - Added patch to introduce VM_MODE_DEFAULT macro >> - Improved/cleaned up the code in processor.c >> - Added patch to fix KVM_CAP_MAX_VCPU_ID on s390x >> - Added patch to enable the kvm_create_max_vcpus on s390x and aarch64 >> >> Andrew Jones (1): >> kvm: selftests: aarch64: fix default vm mode >> >> Thomas Huth (8): >> KVM: selftests: Wrap vcpu_nested_state_get/set functions with x86 >> guard >> KVM: selftests: Guard struct kvm_vcpu_events with >> __KVM_HAVE_VCPU_EVENTS >> KVM: selftests: Introduce a VM_MODE_DEFAULT macro for the default bits >> KVM: selftests: Align memory region addresses to 1M on s390x >> KVM: selftests: Add processor code for s390x >> KVM: selftests: Add the sync_regs test for s390x >> KVM: s390: Do not report unusabled IDs via KVM_CAP_MAX_VCPU_ID >> KVM: selftests: Move kvm_create_max_vcpus test to generic code >> >> MAINTAINERS | 2 + >> arch/mips/kvm/mips.c | 3 + >> arch/powerpc/kvm/powerpc.c | 3 + >> arch/s390/kvm/kvm-s390.c | 1 + >> arch/x86/kvm/x86.c | 3 + >> tools/testing/selftests/kvm/Makefile | 7 +- >> .../testing/selftests/kvm/include/kvm_util.h | 10 + >> .../selftests/kvm/include/s390x/processor.h | 22 ++ >> .../kvm/{x86_64 => }/kvm_create_max_vcpus.c | 3 +- >> .../selftests/kvm/lib/aarch64/processor.c | 2 +- >> tools/testing/selftests/kvm/lib/kvm_util.c | 25 +- >> .../selftests/kvm/lib/s390x/processor.c | 286 ++++++++++++++++++ >> .../selftests/kvm/lib/x86_64/processor.c | 2 +- >> .../selftests/kvm/s390x/sync_regs_test.c | 151 +++++++++ >> virt/kvm/arm/arm.c | 3 + >> virt/kvm/kvm_main.c | 2 - >> 16 files changed, 514 insertions(+), 11 deletions(-) >> create mode 100644 tools/testing/selftests/kvm/include/s390x/processor.h >> rename tools/testing/selftests/kvm/{x86_64 => }/kvm_create_max_vcpus.c (93%) >> create mode 100644 tools/testing/selftests/kvm/lib/s390x/processor.c >> create mode 100644 tools/testing/selftests/kvm/s390x/sync_regs_test.c >> >