blob: 85f3280d7ef6f2a2d450ae7aa835ba1a58ca20af [file] [log] [blame]
Carsten Otte5ecee4b2008-03-25 18:47:38 +01001*** BIG FAT WARNING ***
2The kvm module is currently in EXPERIMENTAL state for s390. This means that
3the interface to the module is not yet considered to remain stable. Thus, be
4prepared that we keep breaking your userspace application and guest
5compatibility over and over again until we feel happy with the result. Make sure
6your guest kernel, your host kernel, and your userspace launcher are in a
7consistent state.
8
9This Documentation describes the unique ioctl calls to /dev/kvm, the resulting
10kvm-vm file descriptors, and the kvm-vcpu file descriptors that differ from x86.
11
121. ioctl calls to /dev/kvm
13KVM does support the following ioctls on s390 that are common with other
14architectures and do behave the same:
15KVM_GET_API_VERSION
16KVM_CREATE_VM (*) see note
17KVM_CHECK_EXTENSION
18KVM_GET_VCPU_MMAP_SIZE
19
20Notes:
21* KVM_CREATE_VM may fail on s390, if the calling process has multiple
22threads and has not called KVM_S390_ENABLE_SIE before.
23
24In addition, on s390 the following architecture specific ioctls are supported:
25ioctl: KVM_S390_ENABLE_SIE
26args: none
27see also: include/linux/kvm.h
28This call causes the kernel to switch on PGSTE in the user page table. This
29operation is needed in order to run a virtual machine, and it requires the
30calling process to be single-threaded. Note that the first call to KVM_CREATE_VM
31will implicitly try to switch on PGSTE if the user process has not called
32KVM_S390_ENABLE_SIE before. User processes that want to launch multiple threads
33before creating a virtual machine have to call KVM_S390_ENABLE_SIE, or will
34observe an error calling KVM_CREATE_VM. Switching on PGSTE is a one-time
35operation, is not reversible, and will persist over the entire lifetime of
36the calling process. It does not have any user-visible effect other than a small
37performance penalty.
38
392. ioctl calls to the kvm-vm file descriptor
40KVM does support the following ioctls on s390 that are common with other
41architectures and do behave the same:
42KVM_CREATE_VCPU
43KVM_SET_USER_MEMORY_REGION (*) see note
44KVM_GET_DIRTY_LOG (**) see note
45
46Notes:
47* kvm does only allow exactly one memory slot on s390, which has to start
48 at guest absolute address zero and at a user address that is aligned on any
49 page boundary. This hardware "limitation" allows us to have a few unique
50 optimizations. The memory slot doesn't have to be filled
51 with memory actually, it may contain sparse holes. That said, with different
52 user memory layout this does still allow a large flexibility when
53 doing the guest memory setup.
54** KVM_GET_DIRTY_LOG doesn't work properly yet. The user will receive an empty
55log. This ioctl call is only needed for guest migration, and we intend to
56implement this one in the future.
57
58In addition, on s390 the following architecture specific ioctls for the kvm-vm
59file descriptor are supported:
60ioctl: KVM_S390_INTERRUPT
61args: struct kvm_s390_interrupt *
62see also: include/linux/kvm.h
63This ioctl is used to submit a floating interrupt for a virtual machine.
64Floating interrupts may be delivered to any virtual cpu in the configuration.
65Only some interrupt types defined in include/linux/kvm.h make sense when
66submitted as floating interrupts. The following interrupts are not considered
67to be useful as floating interrupts, and a call to inject them will result in
68-EINVAL error code: program interrupts and interprocessor signals. Valid
69floating interrupts are:
70KVM_S390_INT_VIRTIO
71KVM_S390_INT_SERVICE
72
733. ioctl calls to the kvm-vcpu file descriptor
74KVM does support the following ioctls on s390 that are common with other
75architectures and do behave the same:
76KVM_RUN
77KVM_GET_REGS
78KVM_SET_REGS
79KVM_GET_SREGS
80KVM_SET_SREGS
81KVM_GET_FPU
82KVM_SET_FPU
83
84In addition, on s390 the following architecture specific ioctls for the
85kvm-vcpu file descriptor are supported:
86ioctl: KVM_S390_INTERRUPT
87args: struct kvm_s390_interrupt *
88see also: include/linux/kvm.h
89This ioctl is used to submit an interrupt for a specific virtual cpu.
90Only some interrupt types defined in include/linux/kvm.h make sense when
91submitted for a specific cpu. The following interrupts are not considered
92to be useful, and a call to inject them will result in -EINVAL error code:
93service processor calls and virtio interrupts. Valid interrupt types are:
94KVM_S390_PROGRAM_INT
95KVM_S390_SIGP_STOP
96KVM_S390_RESTART
97KVM_S390_SIGP_SET_PREFIX
98KVM_S390_INT_EMERGENCY
99
100ioctl: KVM_S390_STORE_STATUS
101args: unsigned long
102see also: include/linux/kvm.h
103This ioctl stores the state of the cpu at the guest real address given as
104argument, unless one of the following values defined in include/linux/kvm.h
Daniel Mack3ad2f3f2010-02-03 08:01:28 +0800105is given as argument:
Carsten Otte5ecee4b2008-03-25 18:47:38 +0100106KVM_S390_STORE_STATUS_NOADDR - the CPU stores its status to the save area in
107absolute lowcore as defined by the principles of operation
108KVM_S390_STORE_STATUS_PREFIXED - the CPU stores its status to the save area in
109its prefix page just like the dump tool that comes with zipl. This is useful
110to create a system dump for use with lkcdutils or crash.
111
112ioctl: KVM_S390_SET_INITIAL_PSW
113args: struct kvm_s390_psw *
114see also: include/linux/kvm.h
115This ioctl can be used to set the processor status word (psw) of a stopped cpu
116prior to running it with KVM_RUN. Note that this call is not required to modify
117the psw during sie intercepts that fall back to userspace because struct kvm_run
118does contain the psw, and this value is evaluated during reentry of KVM_RUN
119after the intercept exit was recognized.
120
121ioctl: KVM_S390_INITIAL_RESET
122args: none
123see also: include/linux/kvm.h
124This ioctl can be used to perform an initial cpu reset as defined by the
125principles of operation. The target cpu has to be in stopped state.