Loading...
1/******************************************************************************
2 * xen.h
3 *
4 * Guest OS interface to Xen.
5 *
6 * Copyright (c) 2004, K A Fraser
7 */
8
9#ifndef __XEN_PUBLIC_XEN_H__
10#define __XEN_PUBLIC_XEN_H__
11
12#include <asm/xen/interface.h>
13#include <asm/pvclock-abi.h>
14
15/*
16 * XEN "SYSTEM CALLS" (a.k.a. HYPERCALLS).
17 */
18
19/*
20 * x86_32: EAX = vector; EBX, ECX, EDX, ESI, EDI = args 1, 2, 3, 4, 5.
21 * EAX = return value
22 * (argument registers may be clobbered on return)
23 * x86_64: RAX = vector; RDI, RSI, RDX, R10, R8, R9 = args 1, 2, 3, 4, 5, 6.
24 * RAX = return value
25 * (argument registers not clobbered on return; RCX, R11 are)
26 */
27#define __HYPERVISOR_set_trap_table 0
28#define __HYPERVISOR_mmu_update 1
29#define __HYPERVISOR_set_gdt 2
30#define __HYPERVISOR_stack_switch 3
31#define __HYPERVISOR_set_callbacks 4
32#define __HYPERVISOR_fpu_taskswitch 5
33#define __HYPERVISOR_sched_op_compat 6
34#define __HYPERVISOR_dom0_op 7
35#define __HYPERVISOR_set_debugreg 8
36#define __HYPERVISOR_get_debugreg 9
37#define __HYPERVISOR_update_descriptor 10
38#define __HYPERVISOR_memory_op 12
39#define __HYPERVISOR_multicall 13
40#define __HYPERVISOR_update_va_mapping 14
41#define __HYPERVISOR_set_timer_op 15
42#define __HYPERVISOR_event_channel_op_compat 16
43#define __HYPERVISOR_xen_version 17
44#define __HYPERVISOR_console_io 18
45#define __HYPERVISOR_physdev_op_compat 19
46#define __HYPERVISOR_grant_table_op 20
47#define __HYPERVISOR_vm_assist 21
48#define __HYPERVISOR_update_va_mapping_otherdomain 22
49#define __HYPERVISOR_iret 23 /* x86 only */
50#define __HYPERVISOR_vcpu_op 24
51#define __HYPERVISOR_set_segment_base 25 /* x86/64 only */
52#define __HYPERVISOR_mmuext_op 26
53#define __HYPERVISOR_acm_op 27
54#define __HYPERVISOR_nmi_op 28
55#define __HYPERVISOR_sched_op 29
56#define __HYPERVISOR_callback_op 30
57#define __HYPERVISOR_xenoprof_op 31
58#define __HYPERVISOR_event_channel_op 32
59#define __HYPERVISOR_physdev_op 33
60#define __HYPERVISOR_hvm_op 34
61#define __HYPERVISOR_tmem_op 38
62
63/* Architecture-specific hypercall definitions. */
64#define __HYPERVISOR_arch_0 48
65#define __HYPERVISOR_arch_1 49
66#define __HYPERVISOR_arch_2 50
67#define __HYPERVISOR_arch_3 51
68#define __HYPERVISOR_arch_4 52
69#define __HYPERVISOR_arch_5 53
70#define __HYPERVISOR_arch_6 54
71#define __HYPERVISOR_arch_7 55
72
73/*
74 * VIRTUAL INTERRUPTS
75 *
76 * Virtual interrupts that a guest OS may receive from Xen.
77 */
78#define VIRQ_TIMER 0 /* Timebase update, and/or requested timeout. */
79#define VIRQ_DEBUG 1 /* Request guest to dump debug info. */
80#define VIRQ_CONSOLE 2 /* (DOM0) Bytes received on emergency console. */
81#define VIRQ_DOM_EXC 3 /* (DOM0) Exceptional event for some domain. */
82#define VIRQ_DEBUGGER 6 /* (DOM0) A domain has paused for debugging. */
83
84/* Architecture-specific VIRQ definitions. */
85#define VIRQ_ARCH_0 16
86#define VIRQ_ARCH_1 17
87#define VIRQ_ARCH_2 18
88#define VIRQ_ARCH_3 19
89#define VIRQ_ARCH_4 20
90#define VIRQ_ARCH_5 21
91#define VIRQ_ARCH_6 22
92#define VIRQ_ARCH_7 23
93
94#define NR_VIRQS 24
95/*
96 * MMU-UPDATE REQUESTS
97 *
98 * HYPERVISOR_mmu_update() accepts a list of (ptr, val) pairs.
99 * A foreigndom (FD) can be specified (or DOMID_SELF for none).
100 * Where the FD has some effect, it is described below.
101 * ptr[1:0] specifies the appropriate MMU_* command.
102 *
103 * ptr[1:0] == MMU_NORMAL_PT_UPDATE:
104 * Updates an entry in a page table. If updating an L1 table, and the new
105 * table entry is valid/present, the mapped frame must belong to the FD, if
106 * an FD has been specified. If attempting to map an I/O page then the
107 * caller assumes the privilege of the FD.
108 * FD == DOMID_IO: Permit /only/ I/O mappings, at the priv level of the caller.
109 * FD == DOMID_XEN: Map restricted areas of Xen's heap space.
110 * ptr[:2] -- Machine address of the page-table entry to modify.
111 * val -- Value to write.
112 *
113 * ptr[1:0] == MMU_MACHPHYS_UPDATE:
114 * Updates an entry in the machine->pseudo-physical mapping table.
115 * ptr[:2] -- Machine address within the frame whose mapping to modify.
116 * The frame must belong to the FD, if one is specified.
117 * val -- Value to write into the mapping entry.
118 *
119 * ptr[1:0] == MMU_PT_UPDATE_PRESERVE_AD:
120 * As MMU_NORMAL_PT_UPDATE above, but A/D bits currently in the PTE are ORed
121 * with those in @val.
122 */
123#define MMU_NORMAL_PT_UPDATE 0 /* checked '*ptr = val'. ptr is MA. */
124#define MMU_MACHPHYS_UPDATE 1 /* ptr = MA of frame to modify entry for */
125#define MMU_PT_UPDATE_PRESERVE_AD 2 /* atomically: *ptr = val | (*ptr&(A|D)) */
126
127/*
128 * MMU EXTENDED OPERATIONS
129 *
130 * HYPERVISOR_mmuext_op() accepts a list of mmuext_op structures.
131 * A foreigndom (FD) can be specified (or DOMID_SELF for none).
132 * Where the FD has some effect, it is described below.
133 *
134 * cmd: MMUEXT_(UN)PIN_*_TABLE
135 * mfn: Machine frame number to be (un)pinned as a p.t. page.
136 * The frame must belong to the FD, if one is specified.
137 *
138 * cmd: MMUEXT_NEW_BASEPTR
139 * mfn: Machine frame number of new page-table base to install in MMU.
140 *
141 * cmd: MMUEXT_NEW_USER_BASEPTR [x86/64 only]
142 * mfn: Machine frame number of new page-table base to install in MMU
143 * when in user space.
144 *
145 * cmd: MMUEXT_TLB_FLUSH_LOCAL
146 * No additional arguments. Flushes local TLB.
147 *
148 * cmd: MMUEXT_INVLPG_LOCAL
149 * linear_addr: Linear address to be flushed from the local TLB.
150 *
151 * cmd: MMUEXT_TLB_FLUSH_MULTI
152 * vcpumask: Pointer to bitmap of VCPUs to be flushed.
153 *
154 * cmd: MMUEXT_INVLPG_MULTI
155 * linear_addr: Linear address to be flushed.
156 * vcpumask: Pointer to bitmap of VCPUs to be flushed.
157 *
158 * cmd: MMUEXT_TLB_FLUSH_ALL
159 * No additional arguments. Flushes all VCPUs' TLBs.
160 *
161 * cmd: MMUEXT_INVLPG_ALL
162 * linear_addr: Linear address to be flushed from all VCPUs' TLBs.
163 *
164 * cmd: MMUEXT_FLUSH_CACHE
165 * No additional arguments. Writes back and flushes cache contents.
166 *
167 * cmd: MMUEXT_SET_LDT
168 * linear_addr: Linear address of LDT base (NB. must be page-aligned).
169 * nr_ents: Number of entries in LDT.
170 */
171#define MMUEXT_PIN_L1_TABLE 0
172#define MMUEXT_PIN_L2_TABLE 1
173#define MMUEXT_PIN_L3_TABLE 2
174#define MMUEXT_PIN_L4_TABLE 3
175#define MMUEXT_UNPIN_TABLE 4
176#define MMUEXT_NEW_BASEPTR 5
177#define MMUEXT_TLB_FLUSH_LOCAL 6
178#define MMUEXT_INVLPG_LOCAL 7
179#define MMUEXT_TLB_FLUSH_MULTI 8
180#define MMUEXT_INVLPG_MULTI 9
181#define MMUEXT_TLB_FLUSH_ALL 10
182#define MMUEXT_INVLPG_ALL 11
183#define MMUEXT_FLUSH_CACHE 12
184#define MMUEXT_SET_LDT 13
185#define MMUEXT_NEW_USER_BASEPTR 15
186
187#ifndef __ASSEMBLY__
188struct mmuext_op {
189 unsigned int cmd;
190 union {
191 /* [UN]PIN_TABLE, NEW_BASEPTR, NEW_USER_BASEPTR */
192 unsigned long mfn;
193 /* INVLPG_LOCAL, INVLPG_ALL, SET_LDT */
194 unsigned long linear_addr;
195 } arg1;
196 union {
197 /* SET_LDT */
198 unsigned int nr_ents;
199 /* TLB_FLUSH_MULTI, INVLPG_MULTI */
200 void *vcpumask;
201 } arg2;
202};
203DEFINE_GUEST_HANDLE_STRUCT(mmuext_op);
204#endif
205
206/* These are passed as 'flags' to update_va_mapping. They can be ORed. */
207/* When specifying UVMF_MULTI, also OR in a pointer to a CPU bitmap. */
208/* UVMF_LOCAL is merely UVMF_MULTI with a NULL bitmap pointer. */
209#define UVMF_NONE (0UL<<0) /* No flushing at all. */
210#define UVMF_TLB_FLUSH (1UL<<0) /* Flush entire TLB(s). */
211#define UVMF_INVLPG (2UL<<0) /* Flush only one entry. */
212#define UVMF_FLUSHTYPE_MASK (3UL<<0)
213#define UVMF_MULTI (0UL<<2) /* Flush subset of TLBs. */
214#define UVMF_LOCAL (0UL<<2) /* Flush local TLB. */
215#define UVMF_ALL (1UL<<2) /* Flush all TLBs. */
216
217/*
218 * Commands to HYPERVISOR_console_io().
219 */
220#define CONSOLEIO_write 0
221#define CONSOLEIO_read 1
222
223/*
224 * Commands to HYPERVISOR_vm_assist().
225 */
226#define VMASST_CMD_enable 0
227#define VMASST_CMD_disable 1
228#define VMASST_TYPE_4gb_segments 0
229#define VMASST_TYPE_4gb_segments_notify 1
230#define VMASST_TYPE_writable_pagetables 2
231#define VMASST_TYPE_pae_extended_cr3 3
232#define MAX_VMASST_TYPE 3
233
234#ifndef __ASSEMBLY__
235
236typedef uint16_t domid_t;
237
238/* Domain ids >= DOMID_FIRST_RESERVED cannot be used for ordinary domains. */
239#define DOMID_FIRST_RESERVED (0x7FF0U)
240
241/* DOMID_SELF is used in certain contexts to refer to oneself. */
242#define DOMID_SELF (0x7FF0U)
243
244/*
245 * DOMID_IO is used to restrict page-table updates to mapping I/O memory.
246 * Although no Foreign Domain need be specified to map I/O pages, DOMID_IO
247 * is useful to ensure that no mappings to the OS's own heap are accidentally
248 * installed. (e.g., in Linux this could cause havoc as reference counts
249 * aren't adjusted on the I/O-mapping code path).
250 * This only makes sense in MMUEXT_SET_FOREIGNDOM, but in that context can
251 * be specified by any calling domain.
252 */
253#define DOMID_IO (0x7FF1U)
254
255/*
256 * DOMID_XEN is used to allow privileged domains to map restricted parts of
257 * Xen's heap space (e.g., the machine_to_phys table).
258 * This only makes sense in MMUEXT_SET_FOREIGNDOM, and is only permitted if
259 * the caller is privileged.
260 */
261#define DOMID_XEN (0x7FF2U)
262
263/*
264 * Send an array of these to HYPERVISOR_mmu_update().
265 * NB. The fields are natural pointer/address size for this architecture.
266 */
267struct mmu_update {
268 uint64_t ptr; /* Machine address of PTE. */
269 uint64_t val; /* New contents of PTE. */
270};
271DEFINE_GUEST_HANDLE_STRUCT(mmu_update);
272
273/*
274 * Send an array of these to HYPERVISOR_multicall().
275 * NB. The fields are natural register size for this architecture.
276 */
277struct multicall_entry {
278 unsigned long op;
279 long result;
280 unsigned long args[6];
281};
282DEFINE_GUEST_HANDLE_STRUCT(multicall_entry);
283
284/*
285 * Event channel endpoints per domain:
286 * 1024 if a long is 32 bits; 4096 if a long is 64 bits.
287 */
288#define NR_EVENT_CHANNELS (sizeof(unsigned long) * sizeof(unsigned long) * 64)
289
290struct vcpu_time_info {
291 /*
292 * Updates to the following values are preceded and followed
293 * by an increment of 'version'. The guest can therefore
294 * detect updates by looking for changes to 'version'. If the
295 * least-significant bit of the version number is set then an
296 * update is in progress and the guest must wait to read a
297 * consistent set of values. The correct way to interact with
298 * the version number is similar to Linux's seqlock: see the
299 * implementations of read_seqbegin/read_seqretry.
300 */
301 uint32_t version;
302 uint32_t pad0;
303 uint64_t tsc_timestamp; /* TSC at last update of time vals. */
304 uint64_t system_time; /* Time, in nanosecs, since boot. */
305 /*
306 * Current system time:
307 * system_time + ((tsc - tsc_timestamp) << tsc_shift) * tsc_to_system_mul
308 * CPU frequency (Hz):
309 * ((10^9 << 32) / tsc_to_system_mul) >> tsc_shift
310 */
311 uint32_t tsc_to_system_mul;
312 int8_t tsc_shift;
313 int8_t pad1[3];
314}; /* 32 bytes */
315
316struct vcpu_info {
317 /*
318 * 'evtchn_upcall_pending' is written non-zero by Xen to indicate
319 * a pending notification for a particular VCPU. It is then cleared
320 * by the guest OS /before/ checking for pending work, thus avoiding
321 * a set-and-check race. Note that the mask is only accessed by Xen
322 * on the CPU that is currently hosting the VCPU. This means that the
323 * pending and mask flags can be updated by the guest without special
324 * synchronisation (i.e., no need for the x86 LOCK prefix).
325 * This may seem suboptimal because if the pending flag is set by
326 * a different CPU then an IPI may be scheduled even when the mask
327 * is set. However, note:
328 * 1. The task of 'interrupt holdoff' is covered by the per-event-
329 * channel mask bits. A 'noisy' event that is continually being
330 * triggered can be masked at source at this very precise
331 * granularity.
332 * 2. The main purpose of the per-VCPU mask is therefore to restrict
333 * reentrant execution: whether for concurrency control, or to
334 * prevent unbounded stack usage. Whatever the purpose, we expect
335 * that the mask will be asserted only for short periods at a time,
336 * and so the likelihood of a 'spurious' IPI is suitably small.
337 * The mask is read before making an event upcall to the guest: a
338 * non-zero mask therefore guarantees that the VCPU will not receive
339 * an upcall activation. The mask is cleared when the VCPU requests
340 * to block: this avoids wakeup-waiting races.
341 */
342 uint8_t evtchn_upcall_pending;
343 uint8_t evtchn_upcall_mask;
344 unsigned long evtchn_pending_sel;
345 struct arch_vcpu_info arch;
346 struct pvclock_vcpu_time_info time;
347}; /* 64 bytes (x86) */
348
349/*
350 * Xen/kernel shared data -- pointer provided in start_info.
351 * NB. We expect that this struct is smaller than a page.
352 */
353struct shared_info {
354 struct vcpu_info vcpu_info[MAX_VIRT_CPUS];
355
356 /*
357 * A domain can create "event channels" on which it can send and receive
358 * asynchronous event notifications. There are three classes of event that
359 * are delivered by this mechanism:
360 * 1. Bi-directional inter- and intra-domain connections. Domains must
361 * arrange out-of-band to set up a connection (usually by allocating
362 * an unbound 'listener' port and avertising that via a storage service
363 * such as xenstore).
364 * 2. Physical interrupts. A domain with suitable hardware-access
365 * privileges can bind an event-channel port to a physical interrupt
366 * source.
367 * 3. Virtual interrupts ('events'). A domain can bind an event-channel
368 * port to a virtual interrupt source, such as the virtual-timer
369 * device or the emergency console.
370 *
371 * Event channels are addressed by a "port index". Each channel is
372 * associated with two bits of information:
373 * 1. PENDING -- notifies the domain that there is a pending notification
374 * to be processed. This bit is cleared by the guest.
375 * 2. MASK -- if this bit is clear then a 0->1 transition of PENDING
376 * will cause an asynchronous upcall to be scheduled. This bit is only
377 * updated by the guest. It is read-only within Xen. If a channel
378 * becomes pending while the channel is masked then the 'edge' is lost
379 * (i.e., when the channel is unmasked, the guest must manually handle
380 * pending notifications as no upcall will be scheduled by Xen).
381 *
382 * To expedite scanning of pending notifications, any 0->1 pending
383 * transition on an unmasked channel causes a corresponding bit in a
384 * per-vcpu selector word to be set. Each bit in the selector covers a
385 * 'C long' in the PENDING bitfield array.
386 */
387 unsigned long evtchn_pending[sizeof(unsigned long) * 8];
388 unsigned long evtchn_mask[sizeof(unsigned long) * 8];
389
390 /*
391 * Wallclock time: updated only by control software. Guests should base
392 * their gettimeofday() syscall on this wallclock-base value.
393 */
394 struct pvclock_wall_clock wc;
395
396 struct arch_shared_info arch;
397
398};
399
400/*
401 * Start-of-day memory layout for the initial domain (DOM0):
402 * 1. The domain is started within contiguous virtual-memory region.
403 * 2. The contiguous region begins and ends on an aligned 4MB boundary.
404 * 3. The region start corresponds to the load address of the OS image.
405 * If the load address is not 4MB aligned then the address is rounded down.
406 * 4. This the order of bootstrap elements in the initial virtual region:
407 * a. relocated kernel image
408 * b. initial ram disk [mod_start, mod_len]
409 * c. list of allocated page frames [mfn_list, nr_pages]
410 * d. start_info_t structure [register ESI (x86)]
411 * e. bootstrap page tables [pt_base, CR3 (x86)]
412 * f. bootstrap stack [register ESP (x86)]
413 * 5. Bootstrap elements are packed together, but each is 4kB-aligned.
414 * 6. The initial ram disk may be omitted.
415 * 7. The list of page frames forms a contiguous 'pseudo-physical' memory
416 * layout for the domain. In particular, the bootstrap virtual-memory
417 * region is a 1:1 mapping to the first section of the pseudo-physical map.
418 * 8. All bootstrap elements are mapped read-writable for the guest OS. The
419 * only exception is the bootstrap page table, which is mapped read-only.
420 * 9. There is guaranteed to be at least 512kB padding after the final
421 * bootstrap element. If necessary, the bootstrap virtual region is
422 * extended by an extra 4MB to ensure this.
423 */
424
425#define MAX_GUEST_CMDLINE 1024
426struct start_info {
427 /* THE FOLLOWING ARE FILLED IN BOTH ON INITIAL BOOT AND ON RESUME. */
428 char magic[32]; /* "xen-<version>-<platform>". */
429 unsigned long nr_pages; /* Total pages allocated to this domain. */
430 unsigned long shared_info; /* MACHINE address of shared info struct. */
431 uint32_t flags; /* SIF_xxx flags. */
432 unsigned long store_mfn; /* MACHINE page number of shared page. */
433 uint32_t store_evtchn; /* Event channel for store communication. */
434 union {
435 struct {
436 unsigned long mfn; /* MACHINE page number of console page. */
437 uint32_t evtchn; /* Event channel for console page. */
438 } domU;
439 struct {
440 uint32_t info_off; /* Offset of console_info struct. */
441 uint32_t info_size; /* Size of console_info struct from start.*/
442 } dom0;
443 } console;
444 /* THE FOLLOWING ARE ONLY FILLED IN ON INITIAL BOOT (NOT RESUME). */
445 unsigned long pt_base; /* VIRTUAL address of page directory. */
446 unsigned long nr_pt_frames; /* Number of bootstrap p.t. frames. */
447 unsigned long mfn_list; /* VIRTUAL address of page-frame list. */
448 unsigned long mod_start; /* VIRTUAL address of pre-loaded module. */
449 unsigned long mod_len; /* Size (bytes) of pre-loaded module. */
450 int8_t cmd_line[MAX_GUEST_CMDLINE];
451};
452
453struct dom0_vga_console_info {
454 uint8_t video_type;
455#define XEN_VGATYPE_TEXT_MODE_3 0x03
456#define XEN_VGATYPE_VESA_LFB 0x23
457
458 union {
459 struct {
460 /* Font height, in pixels. */
461 uint16_t font_height;
462 /* Cursor location (column, row). */
463 uint16_t cursor_x, cursor_y;
464 /* Number of rows and columns (dimensions in characters). */
465 uint16_t rows, columns;
466 } text_mode_3;
467
468 struct {
469 /* Width and height, in pixels. */
470 uint16_t width, height;
471 /* Bytes per scan line. */
472 uint16_t bytes_per_line;
473 /* Bits per pixel. */
474 uint16_t bits_per_pixel;
475 /* LFB physical address, and size (in units of 64kB). */
476 uint32_t lfb_base;
477 uint32_t lfb_size;
478 /* RGB mask offsets and sizes, as defined by VBE 1.2+ */
479 uint8_t red_pos, red_size;
480 uint8_t green_pos, green_size;
481 uint8_t blue_pos, blue_size;
482 uint8_t rsvd_pos, rsvd_size;
483
484 /* VESA capabilities (offset 0xa, VESA command 0x4f00). */
485 uint32_t gbl_caps;
486 /* Mode attributes (offset 0x0, VESA command 0x4f01). */
487 uint16_t mode_attrs;
488 } vesa_lfb;
489 } u;
490};
491
492/* These flags are passed in the 'flags' field of start_info_t. */
493#define SIF_PRIVILEGED (1<<0) /* Is the domain privileged? */
494#define SIF_INITDOMAIN (1<<1) /* Is this the initial control domain? */
495
496typedef uint64_t cpumap_t;
497
498typedef uint8_t xen_domain_handle_t[16];
499
500/* Turn a plain number into a C unsigned long constant. */
501#define __mk_unsigned_long(x) x ## UL
502#define mk_unsigned_long(x) __mk_unsigned_long(x)
503
504#define TMEM_SPEC_VERSION 1
505
506struct tmem_op {
507 uint32_t cmd;
508 int32_t pool_id;
509 union {
510 struct { /* for cmd == TMEM_NEW_POOL */
511 uint64_t uuid[2];
512 uint32_t flags;
513 } new;
514 struct {
515 uint64_t oid[3];
516 uint32_t index;
517 uint32_t tmem_offset;
518 uint32_t pfn_offset;
519 uint32_t len;
520 GUEST_HANDLE(void) gmfn; /* guest machine page frame */
521 } gen;
522 } u;
523};
524
525#else /* __ASSEMBLY__ */
526
527/* In assembly code we cannot use C numeric constant suffixes. */
528#define mk_unsigned_long(x) x
529
530#endif /* !__ASSEMBLY__ */
531
532#endif /* __XEN_PUBLIC_XEN_H__ */
1/******************************************************************************
2 * xen.h
3 *
4 * Guest OS interface to Xen.
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining a copy
7 * of this software and associated documentation files (the "Software"), to
8 * deal in the Software without restriction, including without limitation the
9 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10 * sell copies of the Software, and to permit persons to whom the Software is
11 * furnished to do so, subject to the following conditions:
12 *
13 * The above copyright notice and this permission notice shall be included in
14 * all copies or substantial portions of the Software.
15 *
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
21 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
22 * DEALINGS IN THE SOFTWARE.
23 *
24 * Copyright (c) 2004, K A Fraser
25 */
26
27#ifndef __XEN_PUBLIC_XEN_H__
28#define __XEN_PUBLIC_XEN_H__
29
30#include <asm/xen/interface.h>
31
32/*
33 * XEN "SYSTEM CALLS" (a.k.a. HYPERCALLS).
34 */
35
36/*
37 * x86_32: EAX = vector; EBX, ECX, EDX, ESI, EDI = args 1, 2, 3, 4, 5.
38 * EAX = return value
39 * (argument registers may be clobbered on return)
40 * x86_64: RAX = vector; RDI, RSI, RDX, R10, R8, R9 = args 1, 2, 3, 4, 5, 6.
41 * RAX = return value
42 * (argument registers not clobbered on return; RCX, R11 are)
43 */
44#define __HYPERVISOR_set_trap_table 0
45#define __HYPERVISOR_mmu_update 1
46#define __HYPERVISOR_set_gdt 2
47#define __HYPERVISOR_stack_switch 3
48#define __HYPERVISOR_set_callbacks 4
49#define __HYPERVISOR_fpu_taskswitch 5
50#define __HYPERVISOR_sched_op_compat 6
51#define __HYPERVISOR_platform_op 7
52#define __HYPERVISOR_set_debugreg 8
53#define __HYPERVISOR_get_debugreg 9
54#define __HYPERVISOR_update_descriptor 10
55#define __HYPERVISOR_memory_op 12
56#define __HYPERVISOR_multicall 13
57#define __HYPERVISOR_update_va_mapping 14
58#define __HYPERVISOR_set_timer_op 15
59#define __HYPERVISOR_event_channel_op_compat 16
60#define __HYPERVISOR_xen_version 17
61#define __HYPERVISOR_console_io 18
62#define __HYPERVISOR_physdev_op_compat 19
63#define __HYPERVISOR_grant_table_op 20
64#define __HYPERVISOR_vm_assist 21
65#define __HYPERVISOR_update_va_mapping_otherdomain 22
66#define __HYPERVISOR_iret 23 /* x86 only */
67#define __HYPERVISOR_vcpu_op 24
68#define __HYPERVISOR_set_segment_base 25 /* x86/64 only */
69#define __HYPERVISOR_mmuext_op 26
70#define __HYPERVISOR_xsm_op 27
71#define __HYPERVISOR_nmi_op 28
72#define __HYPERVISOR_sched_op 29
73#define __HYPERVISOR_callback_op 30
74#define __HYPERVISOR_xenoprof_op 31
75#define __HYPERVISOR_event_channel_op 32
76#define __HYPERVISOR_physdev_op 33
77#define __HYPERVISOR_hvm_op 34
78#define __HYPERVISOR_sysctl 35
79#define __HYPERVISOR_domctl 36
80#define __HYPERVISOR_kexec_op 37
81#define __HYPERVISOR_tmem_op 38
82#define __HYPERVISOR_xc_reserved_op 39 /* reserved for XenClient */
83#define __HYPERVISOR_xenpmu_op 40
84#define __HYPERVISOR_dm_op 41
85
86/* Architecture-specific hypercall definitions. */
87#define __HYPERVISOR_arch_0 48
88#define __HYPERVISOR_arch_1 49
89#define __HYPERVISOR_arch_2 50
90#define __HYPERVISOR_arch_3 51
91#define __HYPERVISOR_arch_4 52
92#define __HYPERVISOR_arch_5 53
93#define __HYPERVISOR_arch_6 54
94#define __HYPERVISOR_arch_7 55
95
96/*
97 * VIRTUAL INTERRUPTS
98 *
99 * Virtual interrupts that a guest OS may receive from Xen.
100 * In the side comments, 'V.' denotes a per-VCPU VIRQ while 'G.' denotes a
101 * global VIRQ. The former can be bound once per VCPU and cannot be re-bound.
102 * The latter can be allocated only once per guest: they must initially be
103 * allocated to VCPU0 but can subsequently be re-bound.
104 */
105#define VIRQ_TIMER 0 /* V. Timebase update, and/or requested timeout. */
106#define VIRQ_DEBUG 1 /* V. Request guest to dump debug info. */
107#define VIRQ_CONSOLE 2 /* G. (DOM0) Bytes received on emergency console. */
108#define VIRQ_DOM_EXC 3 /* G. (DOM0) Exceptional event for some domain. */
109#define VIRQ_TBUF 4 /* G. (DOM0) Trace buffer has records available. */
110#define VIRQ_DEBUGGER 6 /* G. (DOM0) A domain has paused for debugging. */
111#define VIRQ_XENOPROF 7 /* V. XenOprofile interrupt: new sample available */
112#define VIRQ_CON_RING 8 /* G. (DOM0) Bytes received on console */
113#define VIRQ_PCPU_STATE 9 /* G. (DOM0) PCPU state changed */
114#define VIRQ_MEM_EVENT 10 /* G. (DOM0) A memory event has occured */
115#define VIRQ_XC_RESERVED 11 /* G. Reserved for XenClient */
116#define VIRQ_ENOMEM 12 /* G. (DOM0) Low on heap memory */
117#define VIRQ_XENPMU 13 /* PMC interrupt */
118
119/* Architecture-specific VIRQ definitions. */
120#define VIRQ_ARCH_0 16
121#define VIRQ_ARCH_1 17
122#define VIRQ_ARCH_2 18
123#define VIRQ_ARCH_3 19
124#define VIRQ_ARCH_4 20
125#define VIRQ_ARCH_5 21
126#define VIRQ_ARCH_6 22
127#define VIRQ_ARCH_7 23
128
129#define NR_VIRQS 24
130
131/*
132 * enum neg_errnoval HYPERVISOR_mmu_update(const struct mmu_update reqs[],
133 * unsigned count, unsigned *done_out,
134 * unsigned foreigndom)
135 * @reqs is an array of mmu_update_t structures ((ptr, val) pairs).
136 * @count is the length of the above array.
137 * @pdone is an output parameter indicating number of completed operations
138 * @foreigndom[15:0]: FD, the expected owner of data pages referenced in this
139 * hypercall invocation. Can be DOMID_SELF.
140 * @foreigndom[31:16]: PFD, the expected owner of pagetable pages referenced
141 * in this hypercall invocation. The value of this field
142 * (x) encodes the PFD as follows:
143 * x == 0 => PFD == DOMID_SELF
144 * x != 0 => PFD == x - 1
145 *
146 * Sub-commands: ptr[1:0] specifies the appropriate MMU_* command.
147 * -------------
148 * ptr[1:0] == MMU_NORMAL_PT_UPDATE:
149 * Updates an entry in a page table belonging to PFD. If updating an L1 table,
150 * and the new table entry is valid/present, the mapped frame must belong to
151 * FD. If attempting to map an I/O page then the caller assumes the privilege
152 * of the FD.
153 * FD == DOMID_IO: Permit /only/ I/O mappings, at the priv level of the caller.
154 * FD == DOMID_XEN: Map restricted areas of Xen's heap space.
155 * ptr[:2] -- Machine address of the page-table entry to modify.
156 * val -- Value to write.
157 *
158 * There also certain implicit requirements when using this hypercall. The
159 * pages that make up a pagetable must be mapped read-only in the guest.
160 * This prevents uncontrolled guest updates to the pagetable. Xen strictly
161 * enforces this, and will disallow any pagetable update which will end up
162 * mapping pagetable page RW, and will disallow using any writable page as a
163 * pagetable. In practice it means that when constructing a page table for a
164 * process, thread, etc, we MUST be very dilligient in following these rules:
165 * 1). Start with top-level page (PGD or in Xen language: L4). Fill out
166 * the entries.
167 * 2). Keep on going, filling out the upper (PUD or L3), and middle (PMD
168 * or L2).
169 * 3). Start filling out the PTE table (L1) with the PTE entries. Once
170 * done, make sure to set each of those entries to RO (so writeable bit
171 * is unset). Once that has been completed, set the PMD (L2) for this
172 * PTE table as RO.
173 * 4). When completed with all of the PMD (L2) entries, and all of them have
174 * been set to RO, make sure to set RO the PUD (L3). Do the same
175 * operation on PGD (L4) pagetable entries that have a PUD (L3) entry.
176 * 5). Now before you can use those pages (so setting the cr3), you MUST also
177 * pin them so that the hypervisor can verify the entries. This is done
178 * via the HYPERVISOR_mmuext_op(MMUEXT_PIN_L4_TABLE, guest physical frame
179 * number of the PGD (L4)). And this point the HYPERVISOR_mmuext_op(
180 * MMUEXT_NEW_BASEPTR, guest physical frame number of the PGD (L4)) can be
181 * issued.
182 * For 32-bit guests, the L4 is not used (as there is less pagetables), so
183 * instead use L3.
184 * At this point the pagetables can be modified using the MMU_NORMAL_PT_UPDATE
185 * hypercall. Also if so desired the OS can also try to write to the PTE
186 * and be trapped by the hypervisor (as the PTE entry is RO).
187 *
188 * To deallocate the pages, the operations are the reverse of the steps
189 * mentioned above. The argument is MMUEXT_UNPIN_TABLE for all levels and the
190 * pagetable MUST not be in use (meaning that the cr3 is not set to it).
191 *
192 * ptr[1:0] == MMU_MACHPHYS_UPDATE:
193 * Updates an entry in the machine->pseudo-physical mapping table.
194 * ptr[:2] -- Machine address within the frame whose mapping to modify.
195 * The frame must belong to the FD, if one is specified.
196 * val -- Value to write into the mapping entry.
197 *
198 * ptr[1:0] == MMU_PT_UPDATE_PRESERVE_AD:
199 * As MMU_NORMAL_PT_UPDATE above, but A/D bits currently in the PTE are ORed
200 * with those in @val.
201 *
202 * @val is usually the machine frame number along with some attributes.
203 * The attributes by default follow the architecture defined bits. Meaning that
204 * if this is a X86_64 machine and four page table layout is used, the layout
205 * of val is:
206 * - 63 if set means No execute (NX)
207 * - 46-13 the machine frame number
208 * - 12 available for guest
209 * - 11 available for guest
210 * - 10 available for guest
211 * - 9 available for guest
212 * - 8 global
213 * - 7 PAT (PSE is disabled, must use hypercall to make 4MB or 2MB pages)
214 * - 6 dirty
215 * - 5 accessed
216 * - 4 page cached disabled
217 * - 3 page write through
218 * - 2 userspace accessible
219 * - 1 writeable
220 * - 0 present
221 *
222 * The one bits that does not fit with the default layout is the PAGE_PSE
223 * also called PAGE_PAT). The MMUEXT_[UN]MARK_SUPER arguments to the
224 * HYPERVISOR_mmuext_op serve as mechanism to set a pagetable to be 4MB
225 * (or 2MB) instead of using the PAGE_PSE bit.
226 *
227 * The reason that the PAGE_PSE (bit 7) is not being utilized is due to Xen
228 * using it as the Page Attribute Table (PAT) bit - for details on it please
229 * refer to Intel SDM 10.12. The PAT allows to set the caching attributes of
230 * pages instead of using MTRRs.
231 *
232 * The PAT MSR is as follows (it is a 64-bit value, each entry is 8 bits):
233 * PAT4 PAT0
234 * +-----+-----+----+----+----+-----+----+----+
235 * | UC | UC- | WC | WB | UC | UC- | WC | WB | <= Linux
236 * +-----+-----+----+----+----+-----+----+----+
237 * | UC | UC- | WT | WB | UC | UC- | WT | WB | <= BIOS (default when machine boots)
238 * +-----+-----+----+----+----+-----+----+----+
239 * | rsv | rsv | WP | WC | UC | UC- | WT | WB | <= Xen
240 * +-----+-----+----+----+----+-----+----+----+
241 *
242 * The lookup of this index table translates to looking up
243 * Bit 7, Bit 4, and Bit 3 of val entry:
244 *
245 * PAT/PSE (bit 7) ... PCD (bit 4) .. PWT (bit 3).
246 *
247 * If all bits are off, then we are using PAT0. If bit 3 turned on,
248 * then we are using PAT1, if bit 3 and bit 4, then PAT2..
249 *
250 * As you can see, the Linux PAT1 translates to PAT4 under Xen. Which means
251 * that if a guest that follows Linux's PAT setup and would like to set Write
252 * Combined on pages it MUST use PAT4 entry. Meaning that Bit 7 (PAGE_PAT) is
253 * set. For example, under Linux it only uses PAT0, PAT1, and PAT2 for the
254 * caching as:
255 *
256 * WB = none (so PAT0)
257 * WC = PWT (bit 3 on)
258 * UC = PWT | PCD (bit 3 and 4 are on).
259 *
260 * To make it work with Xen, it needs to translate the WC bit as so:
261 *
262 * PWT (so bit 3 on) --> PAT (so bit 7 is on) and clear bit 3
263 *
264 * And to translate back it would:
265 *
266 * PAT (bit 7 on) --> PWT (bit 3 on) and clear bit 7.
267 */
268#define MMU_NORMAL_PT_UPDATE 0 /* checked '*ptr = val'. ptr is MA. */
269#define MMU_MACHPHYS_UPDATE 1 /* ptr = MA of frame to modify entry for */
270#define MMU_PT_UPDATE_PRESERVE_AD 2 /* atomically: *ptr = val | (*ptr&(A|D)) */
271#define MMU_PT_UPDATE_NO_TRANSLATE 3 /* checked '*ptr = val'. ptr is MA. */
272
273/*
274 * MMU EXTENDED OPERATIONS
275 *
276 * enum neg_errnoval HYPERVISOR_mmuext_op(mmuext_op_t uops[],
277 * unsigned int count,
278 * unsigned int *pdone,
279 * unsigned int foreigndom)
280 */
281/* HYPERVISOR_mmuext_op() accepts a list of mmuext_op structures.
282 * A foreigndom (FD) can be specified (or DOMID_SELF for none).
283 * Where the FD has some effect, it is described below.
284 *
285 * cmd: MMUEXT_(UN)PIN_*_TABLE
286 * mfn: Machine frame number to be (un)pinned as a p.t. page.
287 * The frame must belong to the FD, if one is specified.
288 *
289 * cmd: MMUEXT_NEW_BASEPTR
290 * mfn: Machine frame number of new page-table base to install in MMU.
291 *
292 * cmd: MMUEXT_NEW_USER_BASEPTR [x86/64 only]
293 * mfn: Machine frame number of new page-table base to install in MMU
294 * when in user space.
295 *
296 * cmd: MMUEXT_TLB_FLUSH_LOCAL
297 * No additional arguments. Flushes local TLB.
298 *
299 * cmd: MMUEXT_INVLPG_LOCAL
300 * linear_addr: Linear address to be flushed from the local TLB.
301 *
302 * cmd: MMUEXT_TLB_FLUSH_MULTI
303 * vcpumask: Pointer to bitmap of VCPUs to be flushed.
304 *
305 * cmd: MMUEXT_INVLPG_MULTI
306 * linear_addr: Linear address to be flushed.
307 * vcpumask: Pointer to bitmap of VCPUs to be flushed.
308 *
309 * cmd: MMUEXT_TLB_FLUSH_ALL
310 * No additional arguments. Flushes all VCPUs' TLBs.
311 *
312 * cmd: MMUEXT_INVLPG_ALL
313 * linear_addr: Linear address to be flushed from all VCPUs' TLBs.
314 *
315 * cmd: MMUEXT_FLUSH_CACHE
316 * No additional arguments. Writes back and flushes cache contents.
317 *
318 * cmd: MMUEXT_FLUSH_CACHE_GLOBAL
319 * No additional arguments. Writes back and flushes cache contents
320 * on all CPUs in the system.
321 *
322 * cmd: MMUEXT_SET_LDT
323 * linear_addr: Linear address of LDT base (NB. must be page-aligned).
324 * nr_ents: Number of entries in LDT.
325 *
326 * cmd: MMUEXT_CLEAR_PAGE
327 * mfn: Machine frame number to be cleared.
328 *
329 * cmd: MMUEXT_COPY_PAGE
330 * mfn: Machine frame number of the destination page.
331 * src_mfn: Machine frame number of the source page.
332 *
333 * cmd: MMUEXT_[UN]MARK_SUPER
334 * mfn: Machine frame number of head of superpage to be [un]marked.
335 */
336#define MMUEXT_PIN_L1_TABLE 0
337#define MMUEXT_PIN_L2_TABLE 1
338#define MMUEXT_PIN_L3_TABLE 2
339#define MMUEXT_PIN_L4_TABLE 3
340#define MMUEXT_UNPIN_TABLE 4
341#define MMUEXT_NEW_BASEPTR 5
342#define MMUEXT_TLB_FLUSH_LOCAL 6
343#define MMUEXT_INVLPG_LOCAL 7
344#define MMUEXT_TLB_FLUSH_MULTI 8
345#define MMUEXT_INVLPG_MULTI 9
346#define MMUEXT_TLB_FLUSH_ALL 10
347#define MMUEXT_INVLPG_ALL 11
348#define MMUEXT_FLUSH_CACHE 12
349#define MMUEXT_SET_LDT 13
350#define MMUEXT_NEW_USER_BASEPTR 15
351#define MMUEXT_CLEAR_PAGE 16
352#define MMUEXT_COPY_PAGE 17
353#define MMUEXT_FLUSH_CACHE_GLOBAL 18
354#define MMUEXT_MARK_SUPER 19
355#define MMUEXT_UNMARK_SUPER 20
356
357#ifndef __ASSEMBLY__
358struct mmuext_op {
359 unsigned int cmd;
360 union {
361 /* [UN]PIN_TABLE, NEW_BASEPTR, NEW_USER_BASEPTR
362 * CLEAR_PAGE, COPY_PAGE, [UN]MARK_SUPER */
363 xen_pfn_t mfn;
364 /* INVLPG_LOCAL, INVLPG_ALL, SET_LDT */
365 unsigned long linear_addr;
366 } arg1;
367 union {
368 /* SET_LDT */
369 unsigned int nr_ents;
370 /* TLB_FLUSH_MULTI, INVLPG_MULTI */
371 void *vcpumask;
372 /* COPY_PAGE */
373 xen_pfn_t src_mfn;
374 } arg2;
375};
376DEFINE_GUEST_HANDLE_STRUCT(mmuext_op);
377#endif
378
379/* These are passed as 'flags' to update_va_mapping. They can be ORed. */
380/* When specifying UVMF_MULTI, also OR in a pointer to a CPU bitmap. */
381/* UVMF_LOCAL is merely UVMF_MULTI with a NULL bitmap pointer. */
382#define UVMF_NONE (0UL<<0) /* No flushing at all. */
383#define UVMF_TLB_FLUSH (1UL<<0) /* Flush entire TLB(s). */
384#define UVMF_INVLPG (2UL<<0) /* Flush only one entry. */
385#define UVMF_FLUSHTYPE_MASK (3UL<<0)
386#define UVMF_MULTI (0UL<<2) /* Flush subset of TLBs. */
387#define UVMF_LOCAL (0UL<<2) /* Flush local TLB. */
388#define UVMF_ALL (1UL<<2) /* Flush all TLBs. */
389
390/*
391 * Commands to HYPERVISOR_console_io().
392 */
393#define CONSOLEIO_write 0
394#define CONSOLEIO_read 1
395
396/*
397 * Commands to HYPERVISOR_vm_assist().
398 */
399#define VMASST_CMD_enable 0
400#define VMASST_CMD_disable 1
401
402/* x86/32 guests: simulate full 4GB segment limits. */
403#define VMASST_TYPE_4gb_segments 0
404
405/* x86/32 guests: trap (vector 15) whenever above vmassist is used. */
406#define VMASST_TYPE_4gb_segments_notify 1
407
408/*
409 * x86 guests: support writes to bottom-level PTEs.
410 * NB1. Page-directory entries cannot be written.
411 * NB2. Guest must continue to remove all writable mappings of PTEs.
412 */
413#define VMASST_TYPE_writable_pagetables 2
414
415/* x86/PAE guests: support PDPTs above 4GB. */
416#define VMASST_TYPE_pae_extended_cr3 3
417
418/*
419 * x86 guests: Sane behaviour for virtual iopl
420 * - virtual iopl updated from do_iret() hypercalls.
421 * - virtual iopl reported in bounce frames.
422 * - guest kernels assumed to be level 0 for the purpose of iopl checks.
423 */
424#define VMASST_TYPE_architectural_iopl 4
425
426/*
427 * All guests: activate update indicator in vcpu_runstate_info
428 * Enable setting the XEN_RUNSTATE_UPDATE flag in guest memory mapped
429 * vcpu_runstate_info during updates of the runstate information.
430 */
431#define VMASST_TYPE_runstate_update_flag 5
432
433#define MAX_VMASST_TYPE 5
434
435#ifndef __ASSEMBLY__
436
437typedef uint16_t domid_t;
438
439/* Domain ids >= DOMID_FIRST_RESERVED cannot be used for ordinary domains. */
440#define DOMID_FIRST_RESERVED (0x7FF0U)
441
442/* DOMID_SELF is used in certain contexts to refer to oneself. */
443#define DOMID_SELF (0x7FF0U)
444
445/*
446 * DOMID_IO is used to restrict page-table updates to mapping I/O memory.
447 * Although no Foreign Domain need be specified to map I/O pages, DOMID_IO
448 * is useful to ensure that no mappings to the OS's own heap are accidentally
449 * installed. (e.g., in Linux this could cause havoc as reference counts
450 * aren't adjusted on the I/O-mapping code path).
451 * This only makes sense in MMUEXT_SET_FOREIGNDOM, but in that context can
452 * be specified by any calling domain.
453 */
454#define DOMID_IO (0x7FF1U)
455
456/*
457 * DOMID_XEN is used to allow privileged domains to map restricted parts of
458 * Xen's heap space (e.g., the machine_to_phys table).
459 * This only makes sense in MMUEXT_SET_FOREIGNDOM, and is only permitted if
460 * the caller is privileged.
461 */
462#define DOMID_XEN (0x7FF2U)
463
464/* DOMID_COW is used as the owner of sharable pages */
465#define DOMID_COW (0x7FF3U)
466
467/* DOMID_INVALID is used to identify pages with unknown owner. */
468#define DOMID_INVALID (0x7FF4U)
469
470/* Idle domain. */
471#define DOMID_IDLE (0x7FFFU)
472
473/*
474 * Send an array of these to HYPERVISOR_mmu_update().
475 * NB. The fields are natural pointer/address size for this architecture.
476 */
477struct mmu_update {
478 uint64_t ptr; /* Machine address of PTE. */
479 uint64_t val; /* New contents of PTE. */
480};
481DEFINE_GUEST_HANDLE_STRUCT(mmu_update);
482
483/*
484 * Send an array of these to HYPERVISOR_multicall().
485 * NB. The fields are logically the natural register size for this
486 * architecture. In cases where xen_ulong_t is larger than this then
487 * any unused bits in the upper portion must be zero.
488 */
489struct multicall_entry {
490 xen_ulong_t op;
491 xen_long_t result;
492 xen_ulong_t args[6];
493};
494DEFINE_GUEST_HANDLE_STRUCT(multicall_entry);
495
496struct vcpu_time_info {
497 /*
498 * Updates to the following values are preceded and followed
499 * by an increment of 'version'. The guest can therefore
500 * detect updates by looking for changes to 'version'. If the
501 * least-significant bit of the version number is set then an
502 * update is in progress and the guest must wait to read a
503 * consistent set of values. The correct way to interact with
504 * the version number is similar to Linux's seqlock: see the
505 * implementations of read_seqbegin/read_seqretry.
506 */
507 uint32_t version;
508 uint32_t pad0;
509 uint64_t tsc_timestamp; /* TSC at last update of time vals. */
510 uint64_t system_time; /* Time, in nanosecs, since boot. */
511 /*
512 * Current system time:
513 * system_time + ((tsc - tsc_timestamp) << tsc_shift) * tsc_to_system_mul
514 * CPU frequency (Hz):
515 * ((10^9 << 32) / tsc_to_system_mul) >> tsc_shift
516 */
517 uint32_t tsc_to_system_mul;
518 int8_t tsc_shift;
519 int8_t pad1[3];
520}; /* 32 bytes */
521
522struct vcpu_info {
523 /*
524 * 'evtchn_upcall_pending' is written non-zero by Xen to indicate
525 * a pending notification for a particular VCPU. It is then cleared
526 * by the guest OS /before/ checking for pending work, thus avoiding
527 * a set-and-check race. Note that the mask is only accessed by Xen
528 * on the CPU that is currently hosting the VCPU. This means that the
529 * pending and mask flags can be updated by the guest without special
530 * synchronisation (i.e., no need for the x86 LOCK prefix).
531 * This may seem suboptimal because if the pending flag is set by
532 * a different CPU then an IPI may be scheduled even when the mask
533 * is set. However, note:
534 * 1. The task of 'interrupt holdoff' is covered by the per-event-
535 * channel mask bits. A 'noisy' event that is continually being
536 * triggered can be masked at source at this very precise
537 * granularity.
538 * 2. The main purpose of the per-VCPU mask is therefore to restrict
539 * reentrant execution: whether for concurrency control, or to
540 * prevent unbounded stack usage. Whatever the purpose, we expect
541 * that the mask will be asserted only for short periods at a time,
542 * and so the likelihood of a 'spurious' IPI is suitably small.
543 * The mask is read before making an event upcall to the guest: a
544 * non-zero mask therefore guarantees that the VCPU will not receive
545 * an upcall activation. The mask is cleared when the VCPU requests
546 * to block: this avoids wakeup-waiting races.
547 */
548 uint8_t evtchn_upcall_pending;
549 uint8_t evtchn_upcall_mask;
550 xen_ulong_t evtchn_pending_sel;
551 struct arch_vcpu_info arch;
552 struct pvclock_vcpu_time_info time;
553}; /* 64 bytes (x86) */
554
555/*
556 * Xen/kernel shared data -- pointer provided in start_info.
557 * NB. We expect that this struct is smaller than a page.
558 */
559struct shared_info {
560 struct vcpu_info vcpu_info[MAX_VIRT_CPUS];
561
562 /*
563 * A domain can create "event channels" on which it can send and receive
564 * asynchronous event notifications. There are three classes of event that
565 * are delivered by this mechanism:
566 * 1. Bi-directional inter- and intra-domain connections. Domains must
567 * arrange out-of-band to set up a connection (usually by allocating
568 * an unbound 'listener' port and avertising that via a storage service
569 * such as xenstore).
570 * 2. Physical interrupts. A domain with suitable hardware-access
571 * privileges can bind an event-channel port to a physical interrupt
572 * source.
573 * 3. Virtual interrupts ('events'). A domain can bind an event-channel
574 * port to a virtual interrupt source, such as the virtual-timer
575 * device or the emergency console.
576 *
577 * Event channels are addressed by a "port index". Each channel is
578 * associated with two bits of information:
579 * 1. PENDING -- notifies the domain that there is a pending notification
580 * to be processed. This bit is cleared by the guest.
581 * 2. MASK -- if this bit is clear then a 0->1 transition of PENDING
582 * will cause an asynchronous upcall to be scheduled. This bit is only
583 * updated by the guest. It is read-only within Xen. If a channel
584 * becomes pending while the channel is masked then the 'edge' is lost
585 * (i.e., when the channel is unmasked, the guest must manually handle
586 * pending notifications as no upcall will be scheduled by Xen).
587 *
588 * To expedite scanning of pending notifications, any 0->1 pending
589 * transition on an unmasked channel causes a corresponding bit in a
590 * per-vcpu selector word to be set. Each bit in the selector covers a
591 * 'C long' in the PENDING bitfield array.
592 */
593 xen_ulong_t evtchn_pending[sizeof(xen_ulong_t) * 8];
594 xen_ulong_t evtchn_mask[sizeof(xen_ulong_t) * 8];
595
596 /*
597 * Wallclock time: updated only by control software. Guests should base
598 * their gettimeofday() syscall on this wallclock-base value.
599 */
600 struct pvclock_wall_clock wc;
601
602 struct arch_shared_info arch;
603
604};
605
606/*
607 * Start-of-day memory layout
608 *
609 * 1. The domain is started within contiguous virtual-memory region.
610 * 2. The contiguous region begins and ends on an aligned 4MB boundary.
611 * 3. This the order of bootstrap elements in the initial virtual region:
612 * a. relocated kernel image
613 * b. initial ram disk [mod_start, mod_len]
614 * (may be omitted)
615 * c. list of allocated page frames [mfn_list, nr_pages]
616 * (unless relocated due to XEN_ELFNOTE_INIT_P2M)
617 * d. start_info_t structure [register ESI (x86)]
618 * in case of dom0 this page contains the console info, too
619 * e. unless dom0: xenstore ring page
620 * f. unless dom0: console ring page
621 * g. bootstrap page tables [pt_base, CR3 (x86)]
622 * h. bootstrap stack [register ESP (x86)]
623 * 4. Bootstrap elements are packed together, but each is 4kB-aligned.
624 * 5. The list of page frames forms a contiguous 'pseudo-physical' memory
625 * layout for the domain. In particular, the bootstrap virtual-memory
626 * region is a 1:1 mapping to the first section of the pseudo-physical map.
627 * 6. All bootstrap elements are mapped read-writable for the guest OS. The
628 * only exception is the bootstrap page table, which is mapped read-only.
629 * 7. There is guaranteed to be at least 512kB padding after the final
630 * bootstrap element. If necessary, the bootstrap virtual region is
631 * extended by an extra 4MB to ensure this.
632 */
633
634#define MAX_GUEST_CMDLINE 1024
635struct start_info {
636 /* THE FOLLOWING ARE FILLED IN BOTH ON INITIAL BOOT AND ON RESUME. */
637 char magic[32]; /* "xen-<version>-<platform>". */
638 unsigned long nr_pages; /* Total pages allocated to this domain. */
639 unsigned long shared_info; /* MACHINE address of shared info struct. */
640 uint32_t flags; /* SIF_xxx flags. */
641 xen_pfn_t store_mfn; /* MACHINE page number of shared page. */
642 uint32_t store_evtchn; /* Event channel for store communication. */
643 union {
644 struct {
645 xen_pfn_t mfn; /* MACHINE page number of console page. */
646 uint32_t evtchn; /* Event channel for console page. */
647 } domU;
648 struct {
649 uint32_t info_off; /* Offset of console_info struct. */
650 uint32_t info_size; /* Size of console_info struct from start.*/
651 } dom0;
652 } console;
653 /* THE FOLLOWING ARE ONLY FILLED IN ON INITIAL BOOT (NOT RESUME). */
654 unsigned long pt_base; /* VIRTUAL address of page directory. */
655 unsigned long nr_pt_frames; /* Number of bootstrap p.t. frames. */
656 unsigned long mfn_list; /* VIRTUAL address of page-frame list. */
657 unsigned long mod_start; /* VIRTUAL address of pre-loaded module. */
658 unsigned long mod_len; /* Size (bytes) of pre-loaded module. */
659 int8_t cmd_line[MAX_GUEST_CMDLINE];
660 /* The pfn range here covers both page table and p->m table frames. */
661 unsigned long first_p2m_pfn;/* 1st pfn forming initial P->M table. */
662 unsigned long nr_p2m_frames;/* # of pfns forming initial P->M table. */
663};
664
665/* These flags are passed in the 'flags' field of start_info_t. */
666#define SIF_PRIVILEGED (1<<0) /* Is the domain privileged? */
667#define SIF_INITDOMAIN (1<<1) /* Is this the initial control domain? */
668#define SIF_MULTIBOOT_MOD (1<<2) /* Is mod_start a multiboot module? */
669#define SIF_MOD_START_PFN (1<<3) /* Is mod_start a PFN? */
670#define SIF_VIRT_P2M_4TOOLS (1<<4) /* Do Xen tools understand a virt. mapped */
671 /* P->M making the 3 level tree obsolete? */
672#define SIF_PM_MASK (0xFF<<8) /* reserve 1 byte for xen-pm options */
673
674/*
675 * A multiboot module is a package containing modules very similar to a
676 * multiboot module array. The only differences are:
677 * - the array of module descriptors is by convention simply at the beginning
678 * of the multiboot module,
679 * - addresses in the module descriptors are based on the beginning of the
680 * multiboot module,
681 * - the number of modules is determined by a termination descriptor that has
682 * mod_start == 0.
683 *
684 * This permits to both build it statically and reference it in a configuration
685 * file, and let the PV guest easily rebase the addresses to virtual addresses
686 * and at the same time count the number of modules.
687 */
688struct xen_multiboot_mod_list {
689 /* Address of first byte of the module */
690 uint32_t mod_start;
691 /* Address of last byte of the module (inclusive) */
692 uint32_t mod_end;
693 /* Address of zero-terminated command line */
694 uint32_t cmdline;
695 /* Unused, must be zero */
696 uint32_t pad;
697};
698/*
699 * The console structure in start_info.console.dom0
700 *
701 * This structure includes a variety of information required to
702 * have a working VGA/VESA console.
703 */
704struct dom0_vga_console_info {
705 uint8_t video_type;
706#define XEN_VGATYPE_TEXT_MODE_3 0x03
707#define XEN_VGATYPE_VESA_LFB 0x23
708#define XEN_VGATYPE_EFI_LFB 0x70
709
710 union {
711 struct {
712 /* Font height, in pixels. */
713 uint16_t font_height;
714 /* Cursor location (column, row). */
715 uint16_t cursor_x, cursor_y;
716 /* Number of rows and columns (dimensions in characters). */
717 uint16_t rows, columns;
718 } text_mode_3;
719
720 struct {
721 /* Width and height, in pixels. */
722 uint16_t width, height;
723 /* Bytes per scan line. */
724 uint16_t bytes_per_line;
725 /* Bits per pixel. */
726 uint16_t bits_per_pixel;
727 /* LFB physical address, and size (in units of 64kB). */
728 uint32_t lfb_base;
729 uint32_t lfb_size;
730 /* RGB mask offsets and sizes, as defined by VBE 1.2+ */
731 uint8_t red_pos, red_size;
732 uint8_t green_pos, green_size;
733 uint8_t blue_pos, blue_size;
734 uint8_t rsvd_pos, rsvd_size;
735
736 /* VESA capabilities (offset 0xa, VESA command 0x4f00). */
737 uint32_t gbl_caps;
738 /* Mode attributes (offset 0x0, VESA command 0x4f01). */
739 uint16_t mode_attrs;
740 } vesa_lfb;
741 } u;
742};
743
744typedef uint64_t cpumap_t;
745
746typedef uint8_t xen_domain_handle_t[16];
747
748/* Turn a plain number into a C unsigned long constant. */
749#define __mk_unsigned_long(x) x ## UL
750#define mk_unsigned_long(x) __mk_unsigned_long(x)
751
752#define TMEM_SPEC_VERSION 1
753
754struct tmem_op {
755 uint32_t cmd;
756 int32_t pool_id;
757 union {
758 struct { /* for cmd == TMEM_NEW_POOL */
759 uint64_t uuid[2];
760 uint32_t flags;
761 } new;
762 struct {
763 uint64_t oid[3];
764 uint32_t index;
765 uint32_t tmem_offset;
766 uint32_t pfn_offset;
767 uint32_t len;
768 GUEST_HANDLE(void) gmfn; /* guest machine page frame */
769 } gen;
770 } u;
771};
772
773DEFINE_GUEST_HANDLE(u64);
774
775#else /* __ASSEMBLY__ */
776
777/* In assembly code we cannot use C numeric constant suffixes. */
778#define mk_unsigned_long(x) x
779
780#endif /* !__ASSEMBLY__ */
781
782#endif /* __XEN_PUBLIC_XEN_H__ */