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