1.. SPDX-License-Identifier: GPL-2.0
2
3==============================================================
4ARM Virtual Generic Interrupt Controller v3 and later (VGICv3)
5==============================================================
6
7
8Device types supported:
9  - KVM_DEV_TYPE_ARM_VGIC_V3     ARM Generic Interrupt Controller v3.0
10
11Only one VGIC instance may be instantiated through this API.  The created VGIC
12will act as the VM interrupt controller, requiring emulated user-space devices
13to inject interrupts to the VGIC instead of directly to CPUs.  It is not
14possible to create both a GICv3 and GICv2 on the same VM.
15
16Creating a guest GICv3 device requires a host GICv3 as well.
17
18
19Groups:
20  KVM_DEV_ARM_VGIC_GRP_ADDR
21   Attributes:
22
23    KVM_VGIC_V3_ADDR_TYPE_DIST (rw, 64-bit)
24      Base address in the guest physical address space of the GICv3 distributor
25      register mappings. Only valid for KVM_DEV_TYPE_ARM_VGIC_V3.
26      This address needs to be 64K aligned and the region covers 64 KByte.
27
28    KVM_VGIC_V3_ADDR_TYPE_REDIST (rw, 64-bit)
29      Base address in the guest physical address space of the GICv3
30      redistributor register mappings. There are two 64K pages for each
31      VCPU and all of the redistributor pages are contiguous.
32      Only valid for KVM_DEV_TYPE_ARM_VGIC_V3.
33      This address needs to be 64K aligned.
34
35    KVM_VGIC_V3_ADDR_TYPE_REDIST_REGION (rw, 64-bit)
36      The attribute data pointed to by kvm_device_attr.addr is a __u64 value::
37
38        bits:     | 63   ....  52  |  51   ....   16 | 15 - 12  |11 - 0
39        values:   |     count      |       base      |  flags   | index
40
41      - index encodes the unique redistributor region index
42      - flags: reserved for future use, currently 0
43      - base field encodes bits [51:16] of the guest physical base address
44        of the first redistributor in the region.
45      - count encodes the number of redistributors in the region. Must be
46        greater than 0.
47
48      There are two 64K pages for each redistributor in the region and
49      redistributors are laid out contiguously within the region. Regions
50      are filled with redistributors in the index order. The sum of all
51      region count fields must be greater than or equal to the number of
52      VCPUs. Redistributor regions must be registered in the incremental
53      index order, starting from index 0.
54
55      The characteristics of a specific redistributor region can be read
56      by presetting the index field in the attr data.
57      Only valid for KVM_DEV_TYPE_ARM_VGIC_V3.
58
59  It is invalid to mix calls with KVM_VGIC_V3_ADDR_TYPE_REDIST and
60  KVM_VGIC_V3_ADDR_TYPE_REDIST_REGION attributes.
61
62  Errors:
63
64    =======  =============================================================
65    -E2BIG   Address outside of addressable IPA range
66    -EINVAL  Incorrectly aligned address, bad redistributor region
67             count/index, mixed redistributor region attribute usage
68    -EEXIST  Address already configured
69    -ENOENT  Attempt to read the characteristics of a non existing
70             redistributor region
71    -ENXIO   The group or attribute is unknown/unsupported for this device
72             or hardware support is missing.
73    -EFAULT  Invalid user pointer for attr->addr.
74    =======  =============================================================
75
76
77  KVM_DEV_ARM_VGIC_GRP_DIST_REGS, KVM_DEV_ARM_VGIC_GRP_REDIST_REGS
78   Attributes:
79
80    The attr field of kvm_device_attr encodes two values::
81
82      bits:     | 63   ....  32  |  31   ....    0 |
83      values:   |      mpidr     |      offset     |
84
85    All distributor regs are (rw, 32-bit) and kvm_device_attr.addr points to a
86    __u32 value.  64-bit registers must be accessed by separately accessing the
87    lower and higher word.
88
89    Writes to read-only registers are ignored by the kernel.
90
91    KVM_DEV_ARM_VGIC_GRP_DIST_REGS accesses the main distributor registers.
92    KVM_DEV_ARM_VGIC_GRP_REDIST_REGS accesses the redistributor of the CPU
93    specified by the mpidr.
94
95    The offset is relative to the "[Re]Distributor base address" as defined
96    in the GICv3/4 specs.  Getting or setting such a register has the same
97    effect as reading or writing the register on real hardware, except for the
98    following registers: GICD_STATUSR, GICR_STATUSR, GICD_ISPENDR,
99    GICR_ISPENDR0, GICD_ICPENDR, and GICR_ICPENDR0.  These registers behave
100    differently when accessed via this interface compared to their
101    architecturally defined behavior to allow software a full view of the
102    VGIC's internal state.
103
104    The mpidr field is used to specify which
105    redistributor is accessed.  The mpidr is ignored for the distributor.
106
107    The mpidr encoding is based on the affinity information in the
108    architecture defined MPIDR, and the field is encoded as follows::
109
110      | 63 .... 56 | 55 .... 48 | 47 .... 40 | 39 .... 32 |
111      |    Aff3    |    Aff2    |    Aff1    |    Aff0    |
112
113    Note that distributor fields are not banked, but return the same value
114    regardless of the mpidr used to access the register.
115
116    GICD_IIDR.Revision is updated when the KVM implementation is changed in a
117    way directly observable by the guest or userspace.  Userspace should read
118    GICD_IIDR from KVM and write back the read value to confirm its expected
119    behavior is aligned with the KVM implementation.  Userspace should set
120    GICD_IIDR before setting any other registers to ensure the expected
121    behavior.
122
123
124    The GICD_STATUSR and GICR_STATUSR registers are architecturally defined such
125    that a write of a clear bit has no effect, whereas a write with a set bit
126    clears that value.  To allow userspace to freely set the values of these two
127    registers, setting the attributes with the register offsets for these two
128    registers simply sets the non-reserved bits to the value written.
129
130
131    Accesses (reads and writes) to the GICD_ISPENDR register region and
132    GICR_ISPENDR0 registers get/set the value of the latched pending state for
133    the interrupts.
134
135    This is identical to the value returned by a guest read from ISPENDR for an
136    edge triggered interrupt, but may differ for level triggered interrupts.
137    For edge triggered interrupts, once an interrupt becomes pending (whether
138    because of an edge detected on the input line or because of a guest write
139    to ISPENDR) this state is "latched", and only cleared when either the
140    interrupt is activated or when the guest writes to ICPENDR. A level
141    triggered interrupt may be pending either because the level input is held
142    high by a device, or because of a guest write to the ISPENDR register. Only
143    ISPENDR writes are latched; if the device lowers the line level then the
144    interrupt is no longer pending unless the guest also wrote to ISPENDR, and
145    conversely writes to ICPENDR or activations of the interrupt do not clear
146    the pending status if the line level is still being held high.  (These
147    rules are documented in the GICv3 specification descriptions of the ICPENDR
148    and ISPENDR registers.) For a level triggered interrupt the value accessed
149    here is that of the latch which is set by ISPENDR and cleared by ICPENDR or
150    interrupt activation, whereas the value returned by a guest read from
151    ISPENDR is the logical OR of the latch value and the input line level.
152
153    Raw access to the latch state is provided to userspace so that it can save
154    and restore the entire GIC internal state (which is defined by the
155    combination of the current input line level and the latch state, and cannot
156    be deduced from purely the line level and the value of the ISPENDR
157    registers).
158
159    Accesses to GICD_ICPENDR register region and GICR_ICPENDR0 registers have
160    RAZ/WI semantics, meaning that reads always return 0 and writes are always
161    ignored.
162
163  Errors:
164
165    ======  =====================================================
166    -ENXIO  Getting or setting this register is not yet supported
167    -EBUSY  One or more VCPUs are running
168    ======  =====================================================
169
170
171  KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS
172   Attributes:
173
174    The attr field of kvm_device_attr encodes two values::
175
176      bits:     | 63      ....       32 | 31  ....  16 | 15  ....  0 |
177      values:   |         mpidr         |      RES     |    instr    |
178
179    The mpidr field encodes the CPU ID based on the affinity information in the
180    architecture defined MPIDR, and the field is encoded as follows::
181
182      | 63 .... 56 | 55 .... 48 | 47 .... 40 | 39 .... 32 |
183      |    Aff3    |    Aff2    |    Aff1    |    Aff0    |
184
185    The instr field encodes the system register to access based on the fields
186    defined in the A64 instruction set encoding for system register access
187    (RES means the bits are reserved for future use and should be zero)::
188
189      | 15 ... 14 | 13 ... 11 | 10 ... 7 | 6 ... 3 | 2 ... 0 |
190      |   Op 0    |    Op1    |    CRn   |   CRm   |   Op2   |
191
192    All system regs accessed through this API are (rw, 64-bit) and
193    kvm_device_attr.addr points to a __u64 value.
194
195    KVM_DEV_ARM_VGIC_GRP_CPU_SYSREGS accesses the CPU interface registers for the
196    CPU specified by the mpidr field.
197
198    CPU interface registers access is not implemented for AArch32 mode.
199    Error -ENXIO is returned when accessed in AArch32 mode.
200
201  Errors:
202
203    =======  =====================================================
204    -ENXIO   Getting or setting this register is not yet supported
205    -EBUSY   VCPU is running
206    -EINVAL  Invalid mpidr or register value supplied
207    =======  =====================================================
208
209
210  KVM_DEV_ARM_VGIC_GRP_NR_IRQS
211   Attributes:
212
213    A value describing the number of interrupts (SGI, PPI and SPI) for
214    this GIC instance, ranging from 64 to 1024, in increments of 32.
215
216    kvm_device_attr.addr points to a __u32 value.
217
218  Errors:
219
220    =======  ======================================
221    -EINVAL  Value set is out of the expected range
222    -EBUSY   Value has already be set.
223    =======  ======================================
224
225
226  KVM_DEV_ARM_VGIC_GRP_CTRL
227   Attributes:
228
229    KVM_DEV_ARM_VGIC_CTRL_INIT
230      request the initialization of the VGIC, no additional parameter in
231      kvm_device_attr.addr. Must be called after all VCPUs have been created.
232    KVM_DEV_ARM_VGIC_SAVE_PENDING_TABLES
233      save all LPI pending bits into guest RAM pending tables.
234
235      The first kB of the pending table is not altered by this operation.
236
237  Errors:
238
239    =======  ========================================================
240    -ENXIO   VGIC not properly configured as required prior to calling
241             this attribute
242    -ENODEV  no online VCPU
243    -ENOMEM  memory shortage when allocating vgic internal data
244    -EFAULT  Invalid guest ram access
245    -EBUSY   One or more VCPUS are running
246    =======  ========================================================
247
248
249  KVM_DEV_ARM_VGIC_GRP_LEVEL_INFO
250   Attributes:
251
252    The attr field of kvm_device_attr encodes the following values::
253
254      bits:     | 63      ....       32 | 31   ....    10 | 9  ....  0 |
255      values:   |         mpidr         |      info       |   vINTID   |
256
257    The vINTID specifies which set of IRQs is reported on.
258
259    The info field specifies which information userspace wants to get or set
260    using this interface.  Currently we support the following info values:
261
262      VGIC_LEVEL_INFO_LINE_LEVEL:
263	Get/Set the input level of the IRQ line for a set of 32 contiguously
264	numbered interrupts.
265
266	vINTID must be a multiple of 32.
267
268	kvm_device_attr.addr points to a __u32 value which will contain a
269	bitmap where a set bit means the interrupt level is asserted.
270
271	Bit[n] indicates the status for interrupt vINTID + n.
272
273    SGIs and any interrupt with a higher ID than the number of interrupts
274    supported, will be RAZ/WI.  LPIs are always edge-triggered and are
275    therefore not supported by this interface.
276
277    PPIs are reported per VCPU as specified in the mpidr field, and SPIs are
278    reported with the same value regardless of the mpidr specified.
279
280    The mpidr field encodes the CPU ID based on the affinity information in the
281    architecture defined MPIDR, and the field is encoded as follows::
282
283      | 63 .... 56 | 55 .... 48 | 47 .... 40 | 39 .... 32 |
284      |    Aff3    |    Aff2    |    Aff1    |    Aff0    |
285
286  Errors:
287
288    =======  =============================================
289    -EINVAL  vINTID is not multiple of 32 or info field is
290	     not VGIC_LEVEL_INFO_LINE_LEVEL
291    =======  =============================================
292