VirtualBox

source: vbox/trunk/src/VBox/VMM/PGM.cpp@ 9981

Last change on this file since 9981 was 9981, checked in by vboxsync, 16 years ago

Always switch the PGM mode when we're using VT-x/AMD-v

  • Property svn:eol-style set to native
  • Property svn:keywords set to Id
File size: 193.5 KB
Line 
1/* $Id: PGM.cpp 9981 2008-06-27 08:39:23Z vboxsync $ */
2/** @file
3 * PGM - Page Manager and Monitor. (Mixing stuff here, not good?)
4 */
5
6/*
7 * Copyright (C) 2006-2007 Sun Microsystems, Inc.
8 *
9 * This file is part of VirtualBox Open Source Edition (OSE), as
10 * available from http://www.virtualbox.org. This file is free software;
11 * you can redistribute it and/or modify it under the terms of the GNU
12 * General Public License (GPL) as published by the Free Software
13 * Foundation, in version 2 as it comes in the "COPYING" file of the
14 * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15 * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16 *
17 * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa
18 * Clara, CA 95054 USA or visit http://www.sun.com if you need
19 * additional information or have any questions.
20 */
21
22
23/** @page pg_pgm PGM - The Page Manager and Monitor
24 *
25 *
26 *
27 * @section sec_pgm_modes Paging Modes
28 *
29 * There are three memory contexts: Host Context (HC), Guest Context (GC)
30 * and intermediate context. When talking about paging HC can also be refered to
31 * as "host paging", and GC refered to as "shadow paging".
32 *
33 * We define three basic paging modes: 32-bit, PAE and AMD64. The host paging mode
34 * is defined by the host operating system. The mode used in the shadow paging mode
35 * depends on the host paging mode and what the mode the guest is currently in. The
36 * following relation between the two is defined:
37 *
38 * @verbatim
39 Host > 32-bit | PAE | AMD64 |
40 Guest | | | |
41 ==v================================
42 32-bit 32-bit PAE PAE
43 -------|--------|--------|--------|
44 PAE PAE PAE PAE
45 -------|--------|--------|--------|
46 AMD64 AMD64 AMD64 AMD64
47 -------|--------|--------|--------| @endverbatim
48 *
49 * All configuration except those in the diagonal (upper left) are expected to
50 * require special effort from the switcher (i.e. a bit slower).
51 *
52 *
53 *
54 *
55 * @section sec_pgm_shw The Shadow Memory Context
56 *
57 *
58 * [..]
59 *
60 * Because of guest context mappings requires PDPT and PML4 entries to allow
61 * writing on AMD64, the two upper levels will have fixed flags whatever the
62 * guest is thinking of using there. So, when shadowing the PD level we will
63 * calculate the effective flags of PD and all the higher levels. In legacy
64 * PAE mode this only applies to the PWT and PCD bits (the rest are
65 * ignored/reserved/MBZ). We will ignore those bits for the present.
66 *
67 *
68 *
69 * @section sec_pgm_int The Intermediate Memory Context
70 *
71 * The world switch goes thru an intermediate memory context which purpose it is
72 * to provide different mappings of the switcher code. All guest mappings are also
73 * present in this context.
74 *
75 * The switcher code is mapped at the same location as on the host, at an
76 * identity mapped location (physical equals virtual address), and at the
77 * hypervisor location.
78 *
79 * PGM maintain page tables for 32-bit, PAE and AMD64 paging modes. This
80 * simplifies switching guest CPU mode and consistency at the cost of more
81 * code to do the work. All memory use for those page tables is located below
82 * 4GB (this includes page tables for guest context mappings).
83 *
84 *
85 * @subsection subsec_pgm_int_gc Guest Context Mappings
86 *
87 * During assignment and relocation of a guest context mapping the intermediate
88 * memory context is used to verify the new location.
89 *
90 * Guest context mappings are currently restricted to below 4GB, for reasons
91 * of simplicity. This may change when we implement AMD64 support.
92 *
93 *
94 *
95 *
96 * @section sec_pgm_misc Misc
97 *
98 * @subsection subsec_pgm_misc_diff Differences Between Legacy PAE and Long Mode PAE
99 *
100 * The differences between legacy PAE and long mode PAE are:
101 * -# PDPE bits 1, 2, 5 and 6 are defined differently. In leagcy mode they are
102 * all marked down as must-be-zero, while in long mode 1, 2 and 5 have the
103 * usual meanings while 6 is ignored (AMD). This means that upon switching to
104 * legacy PAE mode we'll have to clear these bits and when going to long mode
105 * they must be set. This applies to both intermediate and shadow contexts,
106 * however we don't need to do it for the intermediate one since we're
107 * executing with CR0.WP at that time.
108 * -# CR3 allows a 32-byte aligned address in legacy mode, while in long mode
109 * a page aligned one is required.
110 *
111 *
112 * @section sec_pgm_handlers Access Handlers
113 *
114 * Placeholder.
115 *
116 *
117 * @subsection sec_pgm_handlers_virt Virtual Access Handlers
118 *
119 * Placeholder.
120 *
121 *
122 * @subsection sec_pgm_handlers_virt Virtual Access Handlers
123 *
124 * We currently implement three types of virtual access handlers: ALL, WRITE
125 * and HYPERVISOR (WRITE). See PGMVIRTHANDLERTYPE for some more details.
126 *
127 * The HYPERVISOR access handlers is kept in a separate tree since it doesn't apply
128 * to physical pages (PGMTREES::HyperVirtHandlers) and only needs to be consulted in
129 * a special \#PF case. The ALL and WRITE are in the PGMTREES::VirtHandlers tree, the
130 * rest of this section is going to be about these handlers.
131 *
132 * We'll go thru the life cycle of a handler and try make sense of it all, don't know
133 * how successfull this is gonna be...
134 *
135 * 1. A handler is registered thru the PGMR3HandlerVirtualRegister and
136 * PGMHandlerVirtualRegisterEx APIs. We check for conflicting virtual handlers
137 * and create a new node that is inserted into the AVL tree (range key). Then
138 * a full PGM resync is flagged (clear pool, sync cr3, update virtual bit of PGMPAGE).
139 *
140 * 2. The following PGMSyncCR3/SyncCR3 operation will first make invoke HandlerVirtualUpdate.
141 *
142 * 2a. HandlerVirtualUpdate will will lookup all the pages covered by virtual handlers
143 * via the current guest CR3 and update the physical page -> virtual handler
144 * translation. Needless to say, this doesn't exactly scale very well. If any changes
145 * are detected, it will flag a virtual bit update just like we did on registration.
146 * PGMPHYS pages with changes will have their virtual handler state reset to NONE.
147 *
148 * 2b. The virtual bit update process will iterate all the pages covered by all the
149 * virtual handlers and update the PGMPAGE virtual handler state to the max of all
150 * virtual handlers on that page.
151 *
152 * 2c. Back in SyncCR3 we will now flush the entire shadow page cache to make sure
153 * we don't miss any alias mappings of the monitored pages.
154 *
155 * 2d. SyncCR3 will then proceed with syncing the CR3 table.
156 *
157 * 3. \#PF(np,read) on a page in the range. This will cause it to be synced
158 * read-only and resumed if it's a WRITE handler. If it's an ALL handler we
159 * will call the handlers like in the next step. If the physical mapping has
160 * changed we will - some time in the future - perform a handler callback
161 * (optional) and update the physical -> virtual handler cache.
162 *
163 * 4. \#PF(,write) on a page in the range. This will cause the handler to
164 * be invoked.
165 *
166 * 5. The guest invalidates the page and changes the physical backing or
167 * unmaps it. This should cause the invalidation callback to be invoked
168 * (it might not yet be 100% perfect). Exactly what happens next... is
169 * this where we mess up and end up out of sync for a while?
170 *
171 * 6. The handler is deregistered by the client via PGMHandlerVirtualDeregister.
172 * We will then set all PGMPAGEs in the physical -> virtual handler cache for
173 * this handler to NONE and trigger a full PGM resync (basically the same
174 * as int step 1). Which means 2 is executed again.
175 *
176 *
177 * @subsubsection sub_sec_pgm_handler_virt_todo TODOs
178 *
179 * There is a bunch of things that needs to be done to make the virtual handlers
180 * work 100% correctly and work more efficiently.
181 *
182 * The first bit hasn't been implemented yet because it's going to slow the
183 * whole mess down even more, and besides it seems to be working reliably for
184 * our current uses. OTOH, some of the optimizations might end up more or less
185 * implementing the missing bits, so we'll see.
186 *
187 * On the optimization side, the first thing to do is to try avoid unnecessary
188 * cache flushing. Then try team up with the shadowing code to track changes
189 * in mappings by means of access to them (shadow in), updates to shadows pages,
190 * invlpg, and shadow PT discarding (perhaps).
191 *
192 * Some idea that have popped up for optimization for current and new features:
193 * - bitmap indicating where there are virtual handlers installed.
194 * (4KB => 2**20 pages, page 2**12 => covers 32-bit address space 1:1!)
195 * - Further optimize this by min/max (needs min/max avl getters).
196 * - Shadow page table entry bit (if any left)?
197 *
198 */
199
200
201/** @page pg_pgmPhys PGMPhys - Physical Guest Memory Management.
202 *
203 *
204 * Objectives:
205 * - Guest RAM over-commitment using memory ballooning,
206 * zero pages and general page sharing.
207 * - Moving or mirroring a VM onto a different physical machine.
208 *
209 *
210 * @subsection subsec_pgmPhys_Definitions Definitions
211 *
212 * Allocation chunk - A RTR0MemObjAllocPhysNC object and the tracking
213 * machinery assoicated with it.
214 *
215 *
216 *
217 *
218 * @subsection subsec_pgmPhys_AllocPage Allocating a page.
219 *
220 * Initially we map *all* guest memory to the (per VM) zero page, which
221 * means that none of the read functions will cause pages to be allocated.
222 *
223 * Exception, access bit in page tables that have been shared. This must
224 * be handled, but we must also make sure PGMGst*Modify doesn't make
225 * unnecessary modifications.
226 *
227 * Allocation points:
228 * - PGMPhysWriteGCPhys and PGMPhysWrite.
229 * - Replacing a zero page mapping at \#PF.
230 * - Replacing a shared page mapping at \#PF.
231 * - ROM registration (currently MMR3RomRegister).
232 * - VM restore (pgmR3Load).
233 *
234 * For the first three it would make sense to keep a few pages handy
235 * until we've reached the max memory commitment for the VM.
236 *
237 * For the ROM registration, we know exactly how many pages we need
238 * and will request these from ring-0. For restore, we will save
239 * the number of non-zero pages in the saved state and allocate
240 * them up front. This would allow the ring-0 component to refuse
241 * the request if the isn't sufficient memory available for VM use.
242 *
243 * Btw. for both ROM and restore allocations we won't be requiring
244 * zeroed pages as they are going to be filled instantly.
245 *
246 *
247 * @subsection subsec_pgmPhys_FreePage Freeing a page
248 *
249 * There are a few points where a page can be freed:
250 * - After being replaced by the zero page.
251 * - After being replaced by a shared page.
252 * - After being ballooned by the guest additions.
253 * - At reset.
254 * - At restore.
255 *
256 * When freeing one or more pages they will be returned to the ring-0
257 * component and replaced by the zero page.
258 *
259 * The reasoning for clearing out all the pages on reset is that it will
260 * return us to the exact same state as on power on, and may thereby help
261 * us reduce the memory load on the system. Further it might have a
262 * (temporary) positive influence on memory fragmentation (@see subsec_pgmPhys_Fragmentation).
263 *
264 * On restore, as mention under the allocation topic, pages should be
265 * freed / allocated depending on how many is actually required by the
266 * new VM state. The simplest approach is to do like on reset, and free
267 * all non-ROM pages and then allocate what we need.
268 *
269 * A measure to prevent some fragmentation, would be to let each allocation
270 * chunk have some affinity towards the VM having allocated the most pages
271 * from it. Also, try make sure to allocate from allocation chunks that
272 * are almost full. Admittedly, both these measures might work counter to
273 * our intentions and its probably not worth putting a lot of effort,
274 * cpu time or memory into this.
275 *
276 *
277 * @subsection subsec_pgmPhys_SharePage Sharing a page
278 *
279 * The basic idea is that there there will be a idle priority kernel
280 * thread walking the non-shared VM pages hashing them and looking for
281 * pages with the same checksum. If such pages are found, it will compare
282 * them byte-by-byte to see if they actually are identical. If found to be
283 * identical it will allocate a shared page, copy the content, check that
284 * the page didn't change while doing this, and finally request both the
285 * VMs to use the shared page instead. If the page is all zeros (special
286 * checksum and byte-by-byte check) it will request the VM that owns it
287 * to replace it with the zero page.
288 *
289 * To make this efficient, we will have to make sure not to try share a page
290 * that will change its contents soon. This part requires the most work.
291 * A simple idea would be to request the VM to write monitor the page for
292 * a while to make sure it isn't modified any time soon. Also, it may
293 * make sense to skip pages that are being write monitored since this
294 * information is readily available to the thread if it works on the
295 * per-VM guest memory structures (presently called PGMRAMRANGE).
296 *
297 *
298 * @subsection subsec_pgmPhys_Fragmentation Fragmentation Concerns and Counter Measures
299 *
300 * The pages are organized in allocation chunks in ring-0, this is a necessity
301 * if we wish to have an OS agnostic approach to this whole thing. (On Linux we
302 * could easily work on a page-by-page basis if we liked. Whether this is possible
303 * or efficient on NT I don't quite know.) Fragmentation within these chunks may
304 * become a problem as part of the idea here is that we wish to return memory to
305 * the host system.
306 *
307 * For instance, starting two VMs at the same time, they will both allocate the
308 * guest memory on-demand and if permitted their page allocations will be
309 * intermixed. Shut down one of the two VMs and it will be difficult to return
310 * any memory to the host system because the page allocation for the two VMs are
311 * mixed up in the same allocation chunks.
312 *
313 * To further complicate matters, when pages are freed because they have been
314 * ballooned or become shared/zero the whole idea is that the page is supposed
315 * to be reused by another VM or returned to the host system. This will cause
316 * allocation chunks to contain pages belonging to different VMs and prevent
317 * returning memory to the host when one of those VM shuts down.
318 *
319 * The only way to really deal with this problem is to move pages. This can
320 * either be done at VM shutdown and or by the idle priority worker thread
321 * that will be responsible for finding sharable/zero pages. The mechanisms
322 * involved for coercing a VM to move a page (or to do it for it) will be
323 * the same as when telling it to share/zero a page.
324 *
325 *
326 * @subsection subsec_pgmPhys_Tracking Tracking Structures And Their Cost
327 *
328 * There's a difficult balance between keeping the per-page tracking structures
329 * (global and guest page) easy to use and keeping them from eating too much
330 * memory. We have limited virtual memory resources available when operating in
331 * 32-bit kernel space (on 64-bit there'll it's quite a different story). The
332 * tracking structures will be attemted designed such that we can deal with up
333 * to 32GB of memory on a 32-bit system and essentially unlimited on 64-bit ones.
334 *
335 *
336 * @subsubsection subsubsec_pgmPhys_Tracking_Kernel Kernel Space
337 *
338 * @see pg_GMM
339 *
340 * @subsubsection subsubsec_pgmPhys_Tracking_PerVM Per-VM
341 *
342 * Fixed info is the physical address of the page (HCPhys) and the page id
343 * (described above). Theoretically we'll need 48(-12) bits for the HCPhys part.
344 * Today we've restricting ourselves to 40(-12) bits because this is the current
345 * restrictions of all AMD64 implementations (I think Barcelona will up this
346 * to 48(-12) bits, not that it really matters) and I needed the bits for
347 * tracking mappings of a page. 48-12 = 36. That leaves 28 bits, which means a
348 * decent range for the page id: 2^(28+12) = 1024TB.
349 *
350 * In additions to these, we'll have to keep maintaining the page flags as we
351 * currently do. Although it wouldn't harm to optimize these quite a bit, like
352 * for instance the ROM shouldn't depend on having a write handler installed
353 * in order for it to become read-only. A RO/RW bit should be considered so
354 * that the page syncing code doesn't have to mess about checking multiple
355 * flag combinations (ROM || RW handler || write monitored) in order to
356 * figure out how to setup a shadow PTE. But this of course, is second
357 * priority at present. Current this requires 12 bits, but could probably
358 * be optimized to ~8.
359 *
360 * Then there's the 24 bits used to track which shadow page tables are
361 * currently mapping a page for the purpose of speeding up physical
362 * access handlers, and thereby the page pool cache. More bit for this
363 * purpose wouldn't hurt IIRC.
364 *
365 * Then there is a new bit in which we need to record what kind of page
366 * this is, shared, zero, normal or write-monitored-normal. This'll
367 * require 2 bits. One bit might be needed for indicating whether a
368 * write monitored page has been written to. And yet another one or
369 * two for tracking migration status. 3-4 bits total then.
370 *
371 * Whatever is left will can be used to record the sharabilitiy of a
372 * page. The page checksum will not be stored in the per-VM table as
373 * the idle thread will not be permitted to do modifications to it.
374 * It will instead have to keep its own working set of potentially
375 * shareable pages and their check sums and stuff.
376 *
377 * For the present we'll keep the current packing of the
378 * PGMRAMRANGE::aHCPhys to keep the changes simple, only of course,
379 * we'll have to change it to a struct with a total of 128-bits at
380 * our disposal.
381 *
382 * The initial layout will be like this:
383 * @verbatim
384 RTHCPHYS HCPhys; The current stuff.
385 63:40 Current shadow PT tracking stuff.
386 39:12 The physical page frame number.
387 11:0 The current flags.
388 uint32_t u28PageId : 28; The page id.
389 uint32_t u2State : 2; The page state { zero, shared, normal, write monitored }.
390 uint32_t fWrittenTo : 1; Whether a write monitored page was written to.
391 uint32_t u1Reserved : 1; Reserved for later.
392 uint32_t u32Reserved; Reserved for later, mostly sharing stats.
393 @endverbatim
394 *
395 * The final layout will be something like this:
396 * @verbatim
397 RTHCPHYS HCPhys; The current stuff.
398 63:48 High page id (12+).
399 47:12 The physical page frame number.
400 11:0 Low page id.
401 uint32_t fReadOnly : 1; Whether it's readonly page (rom or monitored in some way).
402 uint32_t u3Type : 3; The page type {RESERVED, MMIO, MMIO2, ROM, shadowed ROM, RAM}.
403 uint32_t u2PhysMon : 2; Physical access handler type {none, read, write, all}.
404 uint32_t u2VirtMon : 2; Virtual access handler type {none, read, write, all}..
405 uint32_t u2State : 2; The page state { zero, shared, normal, write monitored }.
406 uint32_t fWrittenTo : 1; Whether a write monitored page was written to.
407 uint32_t u20Reserved : 20; Reserved for later, mostly sharing stats.
408 uint32_t u32Tracking; The shadow PT tracking stuff, roughly.
409 @endverbatim
410 *
411 * Cost wise, this means we'll double the cost for guest memory. There isn't anyway
412 * around that I'm afraid. It means that the cost of dealing out 32GB of memory
413 * to one or more VMs is: (32GB >> PAGE_SHIFT) * 16 bytes, or 128MBs. Or another
414 * example, the VM heap cost when assigning 1GB to a VM will be: 4MB.
415 *
416 * A couple of cost examples for the total cost per-VM + kernel.
417 * 32-bit Windows and 32-bit linux:
418 * 1GB guest ram, 256K pages: 4MB + 2MB(+) = 6MB
419 * 4GB guest ram, 1M pages: 16MB + 8MB(+) = 24MB
420 * 32GB guest ram, 8M pages: 128MB + 64MB(+) = 192MB
421 * 64-bit Windows and 64-bit linux:
422 * 1GB guest ram, 256K pages: 4MB + 3MB(+) = 7MB
423 * 4GB guest ram, 1M pages: 16MB + 12MB(+) = 28MB
424 * 32GB guest ram, 8M pages: 128MB + 96MB(+) = 224MB
425 *
426 * UPDATE - 2007-09-27:
427 * Will need a ballooned flag/state too because we cannot
428 * trust the guest 100% and reporting the same page as ballooned more
429 * than once will put the GMM off balance.
430 *
431 *
432 * @subsection subsec_pgmPhys_Serializing Serializing Access
433 *
434 * Initially, we'll try a simple scheme:
435 *
436 * - The per-VM RAM tracking structures (PGMRAMRANGE) is only modified
437 * by the EMT thread of that VM while in the pgm critsect.
438 * - Other threads in the VM process that needs to make reliable use of
439 * the per-VM RAM tracking structures will enter the critsect.
440 * - No process external thread or kernel thread will ever try enter
441 * the pgm critical section, as that just won't work.
442 * - The idle thread (and similar threads) doesn't not need 100% reliable
443 * data when performing it tasks as the EMT thread will be the one to
444 * do the actual changes later anyway. So, as long as it only accesses
445 * the main ram range, it can do so by somehow preventing the VM from
446 * being destroyed while it works on it...
447 *
448 * - The over-commitment management, including the allocating/freeing
449 * chunks, is serialized by a ring-0 mutex lock (a fast one since the
450 * more mundane mutex implementation is broken on Linux).
451 * - A separeate mutex is protecting the set of allocation chunks so
452 * that pages can be shared or/and freed up while some other VM is
453 * allocating more chunks. This mutex can be take from under the other
454 * one, but not the otherway around.
455 *
456 *
457 * @subsection subsec_pgmPhys_Request VM Request interface
458 *
459 * When in ring-0 it will become necessary to send requests to a VM so it can
460 * for instance move a page while defragmenting during VM destroy. The idle
461 * thread will make use of this interface to request VMs to setup shared
462 * pages and to perform write monitoring of pages.
463 *
464 * I would propose an interface similar to the current VMReq interface, similar
465 * in that it doesn't require locking and that the one sending the request may
466 * wait for completion if it wishes to. This shouldn't be very difficult to
467 * realize.
468 *
469 * The requests themselves are also pretty simple. They are basically:
470 * -# Check that some precondition is still true.
471 * -# Do the update.
472 * -# Update all shadow page tables involved with the page.
473 *
474 * The 3rd step is identical to what we're already doing when updating a
475 * physical handler, see pgmHandlerPhysicalSetRamFlagsAndFlushShadowPTs.
476 *
477 *
478 *
479 * @section sec_pgmPhys_MappingCaches Mapping Caches
480 *
481 * In order to be able to map in and out memory and to be able to support
482 * guest with more RAM than we've got virtual address space, we'll employing
483 * a mapping cache. There is already a tiny one for GC (see PGMGCDynMapGCPageEx)
484 * and we'll create a similar one for ring-0 unless we decide to setup a dedicate
485 * memory context for the HWACCM execution.
486 *
487 *
488 * @subsection subsec_pgmPhys_MappingCaches_R3 Ring-3
489 *
490 * We've considered implementing the ring-3 mapping cache page based but found
491 * that this was bother some when one had to take into account TLBs+SMP and
492 * portability (missing the necessary APIs on several platforms). There were
493 * also some performance concerns with this approach which hadn't quite been
494 * worked out.
495 *
496 * Instead, we'll be mapping allocation chunks into the VM process. This simplifies
497 * matters greatly quite a bit since we don't need to invent any new ring-0 stuff,
498 * only some minor RTR0MEMOBJ mapping stuff. The main concern here is that mapping
499 * compared to the previous idea is that mapping or unmapping a 1MB chunk is more
500 * costly than a single page, although how much more costly is uncertain. We'll
501 * try address this by using a very big cache, preferably bigger than the actual
502 * VM RAM size if possible. The current VM RAM sizes should give some idea for
503 * 32-bit boxes, while on 64-bit we can probably get away with employing an
504 * unlimited cache.
505 *
506 * The cache have to parts, as already indicated, the ring-3 side and the
507 * ring-0 side.
508 *
509 * The ring-0 will be tied to the page allocator since it will operate on the
510 * memory objects it contains. It will therefore require the first ring-0 mutex
511 * discussed in @ref subsec_pgmPhys_Serializing. We
512 * some double house keeping wrt to who has mapped what I think, since both
513 * VMMR0.r0 and RTR0MemObj will keep track of mapping relataions
514 *
515 * The ring-3 part will be protected by the pgm critsect. For simplicity, we'll
516 * require anyone that desires to do changes to the mapping cache to do that
517 * from within this critsect. Alternatively, we could employ a separate critsect
518 * for serializing changes to the mapping cache as this would reduce potential
519 * contention with other threads accessing mappings unrelated to the changes
520 * that are in process. We can see about this later, contention will show
521 * up in the statistics anyway, so it'll be simple to tell.
522 *
523 * The organization of the ring-3 part will be very much like how the allocation
524 * chunks are organized in ring-0, that is in an AVL tree by chunk id. To avoid
525 * having to walk the tree all the time, we'll have a couple of lookaside entries
526 * like in we do for I/O ports and MMIO in IOM.
527 *
528 * The simplified flow of a PGMPhysRead/Write function:
529 * -# Enter the PGM critsect.
530 * -# Lookup GCPhys in the ram ranges and get the Page ID.
531 * -# Calc the Allocation Chunk ID from the Page ID.
532 * -# Check the lookaside entries and then the AVL tree for the Chunk ID.
533 * If not found in cache:
534 * -# Call ring-0 and request it to be mapped and supply
535 * a chunk to be unmapped if the cache is maxed out already.
536 * -# Insert the new mapping into the AVL tree (id + R3 address).
537 * -# Update the relevant lookaside entry and return the mapping address.
538 * -# Do the read/write according to monitoring flags and everything.
539 * -# Leave the critsect.
540 *
541 *
542 * @section sec_pgmPhys_Fallback Fallback
543 *
544 * Current all the "second tier" hosts will not support the RTR0MemObjAllocPhysNC
545 * API and thus require a fallback.
546 *
547 * So, when RTR0MemObjAllocPhysNC returns VERR_NOT_SUPPORTED the page allocator
548 * will return to the ring-3 caller (and later ring-0) and asking it to seed
549 * the page allocator with some fresh pages (VERR_GMM_SEED_ME). Ring-3 will
550 * then perform an SUPPageAlloc(cbChunk >> PAGE_SHIFT) call and make a
551 * "SeededAllocPages" call to ring-0.
552 *
553 * The first time ring-0 sees the VERR_NOT_SUPPORTED failure it will disable
554 * all page sharing (zero page detection will continue). It will also force
555 * all allocations to come from the VM which seeded the page. Both these
556 * measures are taken to make sure that there will never be any need for
557 * mapping anything into ring-3 - everything will be mapped already.
558 *
559 * Whether we'll continue to use the current MM locked memory management
560 * for this I don't quite know (I'd prefer not to and just ditch that all
561 * togther), we'll see what's simplest to do.
562 *
563 *
564 *
565 * @section sec_pgmPhys_Changes Changes
566 *
567 * Breakdown of the changes involved?
568 */
569
570
571/** Saved state data unit version. */
572#define PGM_SAVED_STATE_VERSION 6
573
574/*******************************************************************************
575* Header Files *
576*******************************************************************************/
577#define LOG_GROUP LOG_GROUP_PGM
578#include <VBox/dbgf.h>
579#include <VBox/pgm.h>
580#include <VBox/cpum.h>
581#include <VBox/iom.h>
582#include <VBox/sup.h>
583#include <VBox/mm.h>
584#include <VBox/em.h>
585#include <VBox/stam.h>
586#include <VBox/rem.h>
587#include <VBox/dbgf.h>
588#include <VBox/rem.h>
589#include <VBox/selm.h>
590#include <VBox/ssm.h>
591#include "PGMInternal.h"
592#include <VBox/vm.h>
593#include <VBox/dbg.h>
594#include <VBox/hwaccm.h>
595
596#include <iprt/assert.h>
597#include <iprt/alloc.h>
598#include <iprt/asm.h>
599#include <iprt/thread.h>
600#include <iprt/string.h>
601#include <VBox/param.h>
602#include <VBox/err.h>
603
604
605
606/*******************************************************************************
607* Internal Functions *
608*******************************************************************************/
609static int pgmR3InitPaging(PVM pVM);
610static DECLCALLBACK(void) pgmR3PhysInfo(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
611static DECLCALLBACK(void) pgmR3InfoMode(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
612static DECLCALLBACK(void) pgmR3InfoCr3(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
613static DECLCALLBACK(int) pgmR3RelocatePhysHandler(PAVLROGCPHYSNODECORE pNode, void *pvUser);
614static DECLCALLBACK(int) pgmR3RelocateVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser);
615static DECLCALLBACK(int) pgmR3RelocateHyperVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser);
616#ifdef VBOX_STRICT
617static DECLCALLBACK(void) pgmR3ResetNoMorePhysWritesFlag(PVM pVM, VMSTATE enmState, VMSTATE enmOldState, void *pvUser);
618#endif
619static DECLCALLBACK(int) pgmR3Save(PVM pVM, PSSMHANDLE pSSM);
620static DECLCALLBACK(int) pgmR3Load(PVM pVM, PSSMHANDLE pSSM, uint32_t u32Version);
621static int pgmR3ModeDataInit(PVM pVM, bool fResolveGCAndR0);
622static void pgmR3ModeDataSwitch(PVM pVM, PGMMODE enmShw, PGMMODE enmGst);
623static PGMMODE pgmR3CalcShadowMode(PVM pVM, PGMMODE enmGuestMode, SUPPAGINGMODE enmHostMode, PGMMODE enmShadowMode, VMMSWITCHER *penmSwitcher);
624
625#ifdef VBOX_WITH_STATISTICS
626static void pgmR3InitStats(PVM pVM);
627#endif
628
629#ifdef VBOX_WITH_DEBUGGER
630/** @todo all but the two last commands must be converted to 'info'. */
631static DECLCALLBACK(int) pgmR3CmdRam(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
632static DECLCALLBACK(int) pgmR3CmdMap(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
633static DECLCALLBACK(int) pgmR3CmdSync(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
634static DECLCALLBACK(int) pgmR3CmdSyncAlways(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
635#endif
636
637
638/*******************************************************************************
639* Global Variables *
640*******************************************************************************/
641#ifdef VBOX_WITH_DEBUGGER
642/** Command descriptors. */
643static const DBGCCMD g_aCmds[] =
644{
645 /* pszCmd, cArgsMin, cArgsMax, paArgDesc, cArgDescs, pResultDesc, fFlags, pfnHandler pszSyntax, ....pszDescription */
646 { "pgmram", 0, 0, NULL, 0, NULL, 0, pgmR3CmdRam, "", "Display the ram ranges." },
647 { "pgmmap", 0, 0, NULL, 0, NULL, 0, pgmR3CmdMap, "", "Display the mapping ranges." },
648 { "pgmsync", 0, 0, NULL, 0, NULL, 0, pgmR3CmdSync, "", "Sync the CR3 page." },
649 { "pgmsyncalways", 0, 0, NULL, 0, NULL, 0, pgmR3CmdSyncAlways, "", "Toggle permanent CR3 syncing." },
650};
651#endif
652
653
654
655
656/*
657 * Shadow - 32-bit mode
658 */
659#define PGM_SHW_TYPE PGM_TYPE_32BIT
660#define PGM_SHW_NAME(name) PGM_SHW_NAME_32BIT(name)
661#define PGM_SHW_NAME_GC_STR(name) PGM_SHW_NAME_GC_32BIT_STR(name)
662#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_32BIT_STR(name)
663#include "PGMShw.h"
664
665/* Guest - real mode */
666#define PGM_GST_TYPE PGM_TYPE_REAL
667#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
668#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_REAL_STR(name)
669#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
670#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_REAL(name)
671#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_32BIT_REAL_STR(name)
672#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_REAL_STR(name)
673#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_PHYS
674#include "PGMGst.h"
675#include "PGMBth.h"
676#undef BTH_PGMPOOLKIND_PT_FOR_PT
677#undef PGM_BTH_NAME
678#undef PGM_BTH_NAME_GC_STR
679#undef PGM_BTH_NAME_R0_STR
680#undef PGM_GST_TYPE
681#undef PGM_GST_NAME
682#undef PGM_GST_NAME_GC_STR
683#undef PGM_GST_NAME_R0_STR
684
685/* Guest - protected mode */
686#define PGM_GST_TYPE PGM_TYPE_PROT
687#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
688#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_PROT_STR(name)
689#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
690#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_PROT(name)
691#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_32BIT_PROT_STR(name)
692#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_PROT_STR(name)
693#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_PHYS
694#include "PGMGst.h"
695#include "PGMBth.h"
696#undef BTH_PGMPOOLKIND_PT_FOR_PT
697#undef PGM_BTH_NAME
698#undef PGM_BTH_NAME_GC_STR
699#undef PGM_BTH_NAME_R0_STR
700#undef PGM_GST_TYPE
701#undef PGM_GST_NAME
702#undef PGM_GST_NAME_GC_STR
703#undef PGM_GST_NAME_R0_STR
704
705/* Guest - 32-bit mode */
706#define PGM_GST_TYPE PGM_TYPE_32BIT
707#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
708#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_32BIT_STR(name)
709#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
710#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_32BIT(name)
711#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_32BIT_32BIT_STR(name)
712#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_32BIT_STR(name)
713#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_32BIT_PT
714#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_32BIT_PT_FOR_32BIT_4MB
715#include "PGMGst.h"
716#include "PGMBth.h"
717#undef BTH_PGMPOOLKIND_PT_FOR_BIG
718#undef BTH_PGMPOOLKIND_PT_FOR_PT
719#undef PGM_BTH_NAME
720#undef PGM_BTH_NAME_GC_STR
721#undef PGM_BTH_NAME_R0_STR
722#undef PGM_GST_TYPE
723#undef PGM_GST_NAME
724#undef PGM_GST_NAME_GC_STR
725#undef PGM_GST_NAME_R0_STR
726
727#undef PGM_SHW_TYPE
728#undef PGM_SHW_NAME
729#undef PGM_SHW_NAME_GC_STR
730#undef PGM_SHW_NAME_R0_STR
731
732
733/*
734 * Shadow - PAE mode
735 */
736#define PGM_SHW_TYPE PGM_TYPE_PAE
737#define PGM_SHW_NAME(name) PGM_SHW_NAME_PAE(name)
738#define PGM_SHW_NAME_GC_STR(name) PGM_SHW_NAME_GC_PAE_STR(name)
739#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_PAE_STR(name)
740#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_REAL(name)
741#include "PGMShw.h"
742
743/* Guest - real mode */
744#define PGM_GST_TYPE PGM_TYPE_REAL
745#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
746#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_REAL_STR(name)
747#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
748#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_REAL(name)
749#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_PAE_REAL_STR(name)
750#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_REAL_STR(name)
751#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
752#include "PGMBth.h"
753#undef BTH_PGMPOOLKIND_PT_FOR_PT
754#undef PGM_BTH_NAME
755#undef PGM_BTH_NAME_GC_STR
756#undef PGM_BTH_NAME_R0_STR
757#undef PGM_GST_TYPE
758#undef PGM_GST_NAME
759#undef PGM_GST_NAME_GC_STR
760#undef PGM_GST_NAME_R0_STR
761
762/* Guest - protected mode */
763#define PGM_GST_TYPE PGM_TYPE_PROT
764#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
765#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_PROT_STR(name)
766#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
767#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_PROT(name)
768#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_PAE_PROT_STR(name)
769#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_PROT_STR(name)
770#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
771#include "PGMBth.h"
772#undef BTH_PGMPOOLKIND_PT_FOR_PT
773#undef PGM_BTH_NAME
774#undef PGM_BTH_NAME_GC_STR
775#undef PGM_BTH_NAME_R0_STR
776#undef PGM_GST_TYPE
777#undef PGM_GST_NAME
778#undef PGM_GST_NAME_GC_STR
779#undef PGM_GST_NAME_R0_STR
780
781/* Guest - 32-bit mode */
782#define PGM_GST_TYPE PGM_TYPE_32BIT
783#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
784#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_32BIT_STR(name)
785#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
786#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_32BIT(name)
787#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_PAE_32BIT_STR(name)
788#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_32BIT_STR(name)
789#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_32BIT_PT
790#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_32BIT_4MB
791#include "PGMBth.h"
792#undef BTH_PGMPOOLKIND_PT_FOR_BIG
793#undef BTH_PGMPOOLKIND_PT_FOR_PT
794#undef PGM_BTH_NAME
795#undef PGM_BTH_NAME_GC_STR
796#undef PGM_BTH_NAME_R0_STR
797#undef PGM_GST_TYPE
798#undef PGM_GST_NAME
799#undef PGM_GST_NAME_GC_STR
800#undef PGM_GST_NAME_R0_STR
801
802/* Guest - PAE mode */
803#define PGM_GST_TYPE PGM_TYPE_PAE
804#define PGM_GST_NAME(name) PGM_GST_NAME_PAE(name)
805#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_PAE_STR(name)
806#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PAE_STR(name)
807#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_PAE(name)
808#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_PAE_PAE_STR(name)
809#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_PAE_STR(name)
810#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
811#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
812#include "PGMGst.h"
813#include "PGMBth.h"
814#undef BTH_PGMPOOLKIND_PT_FOR_BIG
815#undef BTH_PGMPOOLKIND_PT_FOR_PT
816#undef PGM_BTH_NAME
817#undef PGM_BTH_NAME_GC_STR
818#undef PGM_BTH_NAME_R0_STR
819#undef PGM_GST_TYPE
820#undef PGM_GST_NAME
821#undef PGM_GST_NAME_GC_STR
822#undef PGM_GST_NAME_R0_STR
823
824#undef PGM_SHW_TYPE
825#undef PGM_SHW_NAME
826#undef PGM_SHW_NAME_GC_STR
827#undef PGM_SHW_NAME_R0_STR
828
829
830/*
831 * Shadow - AMD64 mode
832 */
833#define PGM_SHW_TYPE PGM_TYPE_AMD64
834#define PGM_SHW_NAME(name) PGM_SHW_NAME_AMD64(name)
835#define PGM_SHW_NAME_GC_STR(name) PGM_SHW_NAME_GC_AMD64_STR(name)
836#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_AMD64_STR(name)
837#include "PGMShw.h"
838
839/* Guest - AMD64 mode */
840#define PGM_GST_TYPE PGM_TYPE_AMD64
841#define PGM_GST_NAME(name) PGM_GST_NAME_AMD64(name)
842#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_AMD64_STR(name)
843#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_AMD64_STR(name)
844#define PGM_BTH_NAME(name) PGM_BTH_NAME_AMD64_AMD64(name)
845#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_AMD64_AMD64_STR(name)
846#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_AMD64_AMD64_STR(name)
847#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
848#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
849#include "PGMGst.h"
850#include "PGMBth.h"
851#undef BTH_PGMPOOLKIND_PT_FOR_BIG
852#undef BTH_PGMPOOLKIND_PT_FOR_PT
853#undef PGM_BTH_NAME
854#undef PGM_BTH_NAME_GC_STR
855#undef PGM_BTH_NAME_R0_STR
856#undef PGM_GST_TYPE
857#undef PGM_GST_NAME
858#undef PGM_GST_NAME_GC_STR
859#undef PGM_GST_NAME_R0_STR
860
861#undef PGM_SHW_TYPE
862#undef PGM_SHW_NAME
863#undef PGM_SHW_NAME_GC_STR
864#undef PGM_SHW_NAME_R0_STR
865
866/*
867 * Shadow - Nested paging mode
868 */
869#define PGM_SHW_TYPE PGM_TYPE_NESTED
870#define PGM_SHW_NAME(name) PGM_SHW_NAME_NESTED(name)
871#define PGM_SHW_NAME_GC_STR(name) PGM_SHW_NAME_GC_NESTED_STR(name)
872#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_NESTED_STR(name)
873#include "PGMShw.h"
874
875/* Guest - real mode */
876#define PGM_GST_TYPE PGM_TYPE_REAL
877#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
878#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_REAL_STR(name)
879#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
880#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_REAL(name)
881#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_NESTED_REAL_STR(name)
882#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_REAL_STR(name)
883#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
884#include "PGMBth.h"
885#undef BTH_PGMPOOLKIND_PT_FOR_PT
886#undef PGM_BTH_NAME
887#undef PGM_BTH_NAME_GC_STR
888#undef PGM_BTH_NAME_R0_STR
889#undef PGM_GST_TYPE
890#undef PGM_GST_NAME
891#undef PGM_GST_NAME_GC_STR
892#undef PGM_GST_NAME_R0_STR
893
894/* Guest - protected mode */
895#define PGM_GST_TYPE PGM_TYPE_PROT
896#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
897#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_PROT_STR(name)
898#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
899#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_PROT(name)
900#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_NESTED_PROT_STR(name)
901#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_PROT_STR(name)
902#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
903#include "PGMBth.h"
904#undef BTH_PGMPOOLKIND_PT_FOR_PT
905#undef PGM_BTH_NAME
906#undef PGM_BTH_NAME_GC_STR
907#undef PGM_BTH_NAME_R0_STR
908#undef PGM_GST_TYPE
909#undef PGM_GST_NAME
910#undef PGM_GST_NAME_GC_STR
911#undef PGM_GST_NAME_R0_STR
912
913/* Guest - 32-bit mode */
914#define PGM_GST_TYPE PGM_TYPE_32BIT
915#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
916#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_32BIT_STR(name)
917#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
918#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_32BIT(name)
919#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_NESTED_32BIT_STR(name)
920#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_32BIT_STR(name)
921#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_32BIT_PT
922#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_32BIT_4MB
923#include "PGMBth.h"
924#undef BTH_PGMPOOLKIND_PT_FOR_BIG
925#undef BTH_PGMPOOLKIND_PT_FOR_PT
926#undef PGM_BTH_NAME
927#undef PGM_BTH_NAME_GC_STR
928#undef PGM_BTH_NAME_R0_STR
929#undef PGM_GST_TYPE
930#undef PGM_GST_NAME
931#undef PGM_GST_NAME_GC_STR
932#undef PGM_GST_NAME_R0_STR
933
934/* Guest - PAE mode */
935#define PGM_GST_TYPE PGM_TYPE_PAE
936#define PGM_GST_NAME(name) PGM_GST_NAME_PAE(name)
937#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_PAE_STR(name)
938#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PAE_STR(name)
939#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_PAE(name)
940#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_NESTED_PAE_STR(name)
941#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_PAE_STR(name)
942#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
943#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
944#include "PGMBth.h"
945#undef BTH_PGMPOOLKIND_PT_FOR_BIG
946#undef BTH_PGMPOOLKIND_PT_FOR_PT
947#undef PGM_BTH_NAME
948#undef PGM_BTH_NAME_GC_STR
949#undef PGM_BTH_NAME_R0_STR
950#undef PGM_GST_TYPE
951#undef PGM_GST_NAME
952#undef PGM_GST_NAME_GC_STR
953#undef PGM_GST_NAME_R0_STR
954
955/* Guest - AMD64 mode */
956#define PGM_GST_TYPE PGM_TYPE_AMD64
957#define PGM_GST_NAME(name) PGM_GST_NAME_AMD64(name)
958#define PGM_GST_NAME_GC_STR(name) PGM_GST_NAME_GC_AMD64_STR(name)
959#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_AMD64_STR(name)
960#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_AMD64(name)
961#define PGM_BTH_NAME_GC_STR(name) PGM_BTH_NAME_GC_NESTED_AMD64_STR(name)
962#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_AMD64_STR(name)
963#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
964#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
965#include "PGMBth.h"
966#undef BTH_PGMPOOLKIND_PT_FOR_BIG
967#undef BTH_PGMPOOLKIND_PT_FOR_PT
968#undef PGM_BTH_NAME
969#undef PGM_BTH_NAME_GC_STR
970#undef PGM_BTH_NAME_R0_STR
971#undef PGM_GST_TYPE
972#undef PGM_GST_NAME
973#undef PGM_GST_NAME_GC_STR
974#undef PGM_GST_NAME_R0_STR
975
976#undef PGM_SHW_TYPE
977#undef PGM_SHW_NAME
978#undef PGM_SHW_NAME_GC_STR
979#undef PGM_SHW_NAME_R0_STR
980
981
982/**
983 * Initiates the paging of VM.
984 *
985 * @returns VBox status code.
986 * @param pVM Pointer to VM structure.
987 */
988PGMR3DECL(int) PGMR3Init(PVM pVM)
989{
990 LogFlow(("PGMR3Init:\n"));
991
992 /*
993 * Assert alignment and sizes.
994 */
995 AssertRelease(sizeof(pVM->pgm.s) <= sizeof(pVM->pgm.padding));
996
997 /*
998 * Init the structure.
999 */
1000 pVM->pgm.s.offVM = RT_OFFSETOF(VM, pgm.s);
1001 pVM->pgm.s.enmShadowMode = PGMMODE_INVALID;
1002 pVM->pgm.s.enmGuestMode = PGMMODE_INVALID;
1003 pVM->pgm.s.enmHostMode = SUPPAGINGMODE_INVALID;
1004 pVM->pgm.s.GCPhysCR3 = NIL_RTGCPHYS;
1005 pVM->pgm.s.GCPhysGstCR3Monitored = NIL_RTGCPHYS;
1006 pVM->pgm.s.fA20Enabled = true;
1007 pVM->pgm.s.pGstPaePDPTHC = NULL;
1008 pVM->pgm.s.pGstPaePDPTGC = 0;
1009 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.apGstPaePDsHC); i++)
1010 {
1011 pVM->pgm.s.apGstPaePDsHC[i] = NULL;
1012 pVM->pgm.s.apGstPaePDsGC[i] = 0;
1013 pVM->pgm.s.aGCPhysGstPaePDs[i] = NIL_RTGCPHYS;
1014 pVM->pgm.s.aGCPhysGstPaePDsMonitored[i] = NIL_RTGCPHYS;
1015 }
1016
1017#ifdef VBOX_STRICT
1018 VMR3AtStateRegister(pVM, pgmR3ResetNoMorePhysWritesFlag, NULL);
1019#endif
1020
1021 /*
1022 * Get the configured RAM size - to estimate saved state size.
1023 */
1024 uint64_t cbRam;
1025 int rc = CFGMR3QueryU64(CFGMR3GetRoot(pVM), "RamSize", &cbRam);
1026 if (rc == VERR_CFGM_VALUE_NOT_FOUND)
1027 cbRam = pVM->pgm.s.cbRamSize = 0;
1028 else if (VBOX_SUCCESS(rc))
1029 {
1030 if (cbRam < PAGE_SIZE)
1031 cbRam = 0;
1032 cbRam = RT_ALIGN_64(cbRam, PAGE_SIZE);
1033 pVM->pgm.s.cbRamSize = (RTUINT)cbRam;
1034 }
1035 else
1036 {
1037 AssertMsgFailed(("Configuration error: Failed to query integer \"RamSize\", rc=%Vrc.\n", rc));
1038 return rc;
1039 }
1040
1041 /*
1042 * Register saved state data unit.
1043 */
1044 rc = SSMR3RegisterInternal(pVM, "pgm", 1, PGM_SAVED_STATE_VERSION, (size_t)cbRam + sizeof(PGM),
1045 NULL, pgmR3Save, NULL,
1046 NULL, pgmR3Load, NULL);
1047 if (VBOX_FAILURE(rc))
1048 return rc;
1049
1050 /*
1051 * Initialize the PGM critical section and flush the phys TLBs
1052 */
1053 rc = PDMR3CritSectInit(pVM, &pVM->pgm.s.CritSect, "PGM");
1054 AssertRCReturn(rc, rc);
1055
1056 PGMR3PhysChunkInvalidateTLB(pVM);
1057 PGMPhysInvalidatePageR3MapTLB(pVM);
1058 PGMPhysInvalidatePageR0MapTLB(pVM);
1059 PGMPhysInvalidatePageGCMapTLB(pVM);
1060
1061 /*
1062 * Trees
1063 */
1064 rc = MMHyperAlloc(pVM, sizeof(PGMTREES), 0, MM_TAG_PGM, (void **)&pVM->pgm.s.pTreesHC);
1065 if (VBOX_SUCCESS(rc))
1066 {
1067 pVM->pgm.s.pTreesGC = MMHyperHC2GC(pVM, pVM->pgm.s.pTreesHC);
1068
1069 /*
1070 * Alocate the zero page.
1071 */
1072 rc = MMHyperAlloc(pVM, PAGE_SIZE, PAGE_SIZE, MM_TAG_PGM, &pVM->pgm.s.pvZeroPgR3);
1073 }
1074 if (VBOX_SUCCESS(rc))
1075 {
1076 pVM->pgm.s.pvZeroPgGC = MMHyperR3ToGC(pVM, pVM->pgm.s.pvZeroPgR3);
1077 pVM->pgm.s.pvZeroPgR0 = MMHyperR3ToR0(pVM, pVM->pgm.s.pvZeroPgR3);
1078 AssertRelease(pVM->pgm.s.pvZeroPgR0 != NIL_RTHCPHYS);
1079 pVM->pgm.s.HCPhysZeroPg = MMR3HyperHCVirt2HCPhys(pVM, pVM->pgm.s.pvZeroPgR3);
1080 AssertRelease(pVM->pgm.s.HCPhysZeroPg != NIL_RTHCPHYS);
1081
1082 /*
1083 * Init the paging.
1084 */
1085 rc = pgmR3InitPaging(pVM);
1086 }
1087 if (VBOX_SUCCESS(rc))
1088 {
1089 /*
1090 * Init the page pool.
1091 */
1092 rc = pgmR3PoolInit(pVM);
1093 }
1094 if (VBOX_SUCCESS(rc))
1095 {
1096 /*
1097 * Info & statistics
1098 */
1099 DBGFR3InfoRegisterInternal(pVM, "mode",
1100 "Shows the current paging mode. "
1101 "Recognizes 'all', 'guest', 'shadow' and 'host' as arguments, defaulting to 'all' if nothing's given.",
1102 pgmR3InfoMode);
1103 DBGFR3InfoRegisterInternal(pVM, "pgmcr3",
1104 "Dumps all the entries in the top level paging table. No arguments.",
1105 pgmR3InfoCr3);
1106 DBGFR3InfoRegisterInternal(pVM, "phys",
1107 "Dumps all the physical address ranges. No arguments.",
1108 pgmR3PhysInfo);
1109 DBGFR3InfoRegisterInternal(pVM, "handlers",
1110 "Dumps physical, virtual and hyper virtual handlers. "
1111 "Pass 'phys', 'virt', 'hyper' as argument if only one kind is wanted."
1112 "Add 'nost' if the statistics are unwanted, use together with 'all' or explicit selection.",
1113 pgmR3InfoHandlers);
1114 DBGFR3InfoRegisterInternal(pVM, "mappings",
1115 "Dumps guest mappings.",
1116 pgmR3MapInfo);
1117
1118 STAM_REL_REG(pVM, &pVM->pgm.s.cGuestModeChanges, STAMTYPE_COUNTER, "/PGM/cGuestModeChanges", STAMUNIT_OCCURENCES, "Number of guest mode changes.");
1119#ifdef VBOX_WITH_STATISTICS
1120 pgmR3InitStats(pVM);
1121#endif
1122#ifdef VBOX_WITH_DEBUGGER
1123 /*
1124 * Debugger commands.
1125 */
1126 static bool fRegisteredCmds = false;
1127 if (!fRegisteredCmds)
1128 {
1129 int rc = DBGCRegisterCommands(&g_aCmds[0], ELEMENTS(g_aCmds));
1130 if (VBOX_SUCCESS(rc))
1131 fRegisteredCmds = true;
1132 }
1133#endif
1134 return VINF_SUCCESS;
1135 }
1136
1137 /* Almost no cleanup necessary, MM frees all memory. */
1138 PDMR3CritSectDelete(&pVM->pgm.s.CritSect);
1139
1140 return rc;
1141}
1142
1143
1144/**
1145 * Init paging.
1146 *
1147 * Since we need to check what mode the host is operating in before we can choose
1148 * the right paging functions for the host we have to delay this until R0 has
1149 * been initialized.
1150 *
1151 * @returns VBox status code.
1152 * @param pVM VM handle.
1153 */
1154static int pgmR3InitPaging(PVM pVM)
1155{
1156 /*
1157 * Force a recalculation of modes and switcher so everyone gets notified.
1158 */
1159 pVM->pgm.s.enmShadowMode = PGMMODE_INVALID;
1160 pVM->pgm.s.enmGuestMode = PGMMODE_INVALID;
1161 pVM->pgm.s.enmHostMode = SUPPAGINGMODE_INVALID;
1162
1163 /*
1164 * Allocate static mapping space for whatever the cr3 register
1165 * points to and in the case of PAE mode to the 4 PDs.
1166 */
1167 int rc = MMR3HyperReserve(pVM, PAGE_SIZE * 5, "CR3 mapping", &pVM->pgm.s.GCPtrCR3Mapping);
1168 if (VBOX_FAILURE(rc))
1169 {
1170 AssertMsgFailed(("Failed to reserve two pages for cr mapping in HMA, rc=%Vrc\n", rc));
1171 return rc;
1172 }
1173 MMR3HyperReserve(pVM, PAGE_SIZE, "fence", NULL);
1174
1175 /*
1176 * Allocate pages for the three possible intermediate contexts
1177 * (AMD64, PAE and plain 32-Bit). We maintain all three contexts
1178 * for the sake of simplicity. The AMD64 uses the PAE for the
1179 * lower levels, making the total number of pages 11 (3 + 7 + 1).
1180 *
1181 * We assume that two page tables will be enought for the core code
1182 * mappings (HC virtual and identity).
1183 */
1184 pVM->pgm.s.pInterPD = (PX86PD)MMR3PageAllocLow(pVM);
1185 pVM->pgm.s.apInterPTs[0] = (PX86PT)MMR3PageAllocLow(pVM);
1186 pVM->pgm.s.apInterPTs[1] = (PX86PT)MMR3PageAllocLow(pVM);
1187 pVM->pgm.s.apInterPaePTs[0] = (PX86PTPAE)MMR3PageAlloc(pVM);
1188 pVM->pgm.s.apInterPaePTs[1] = (PX86PTPAE)MMR3PageAlloc(pVM);
1189 pVM->pgm.s.apInterPaePDs[0] = (PX86PDPAE)MMR3PageAlloc(pVM);
1190 pVM->pgm.s.apInterPaePDs[1] = (PX86PDPAE)MMR3PageAlloc(pVM);
1191 pVM->pgm.s.apInterPaePDs[2] = (PX86PDPAE)MMR3PageAlloc(pVM);
1192 pVM->pgm.s.apInterPaePDs[3] = (PX86PDPAE)MMR3PageAlloc(pVM);
1193 pVM->pgm.s.pInterPaePDPT = (PX86PDPT)MMR3PageAllocLow(pVM);
1194 pVM->pgm.s.pInterPaePDPT64 = (PX86PDPT)MMR3PageAllocLow(pVM);
1195 pVM->pgm.s.pInterPaePML4 = (PX86PML4)MMR3PageAllocLow(pVM);
1196 if ( !pVM->pgm.s.pInterPD
1197 || !pVM->pgm.s.apInterPTs[0]
1198 || !pVM->pgm.s.apInterPTs[1]
1199 || !pVM->pgm.s.apInterPaePTs[0]
1200 || !pVM->pgm.s.apInterPaePTs[1]
1201 || !pVM->pgm.s.apInterPaePDs[0]
1202 || !pVM->pgm.s.apInterPaePDs[1]
1203 || !pVM->pgm.s.apInterPaePDs[2]
1204 || !pVM->pgm.s.apInterPaePDs[3]
1205 || !pVM->pgm.s.pInterPaePDPT
1206 || !pVM->pgm.s.pInterPaePDPT64
1207 || !pVM->pgm.s.pInterPaePML4)
1208 {
1209 AssertMsgFailed(("Failed to allocate pages for the intermediate context!\n"));
1210 return VERR_NO_PAGE_MEMORY;
1211 }
1212
1213 pVM->pgm.s.HCPhysInterPD = MMPage2Phys(pVM, pVM->pgm.s.pInterPD);
1214 AssertRelease(pVM->pgm.s.HCPhysInterPD != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPD & PAGE_OFFSET_MASK));
1215 pVM->pgm.s.HCPhysInterPaePDPT = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT);
1216 AssertRelease(pVM->pgm.s.HCPhysInterPaePDPT != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPaePDPT & PAGE_OFFSET_MASK));
1217 pVM->pgm.s.HCPhysInterPaePML4 = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePML4);
1218 AssertRelease(pVM->pgm.s.HCPhysInterPaePML4 != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPaePML4 & PAGE_OFFSET_MASK));
1219
1220 /*
1221 * Initialize the pages, setting up the PML4 and PDPT for repetitive 4GB action.
1222 */
1223 ASMMemZeroPage(pVM->pgm.s.pInterPD);
1224 ASMMemZeroPage(pVM->pgm.s.apInterPTs[0]);
1225 ASMMemZeroPage(pVM->pgm.s.apInterPTs[1]);
1226
1227 ASMMemZeroPage(pVM->pgm.s.apInterPaePTs[0]);
1228 ASMMemZeroPage(pVM->pgm.s.apInterPaePTs[1]);
1229
1230 ASMMemZeroPage(pVM->pgm.s.pInterPaePDPT);
1231 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.apInterPaePDs); i++)
1232 {
1233 ASMMemZeroPage(pVM->pgm.s.apInterPaePDs[i]);
1234 pVM->pgm.s.pInterPaePDPT->a[i].u = X86_PDPE_P | PGM_PLXFLAGS_PERMANENT
1235 | MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[i]);
1236 }
1237
1238 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.pInterPaePDPT64->a); i++)
1239 {
1240 const unsigned iPD = i % ELEMENTS(pVM->pgm.s.apInterPaePDs);
1241 pVM->pgm.s.pInterPaePDPT64->a[i].u = X86_PDPE_P | X86_PDPE_RW | X86_PDPE_US | X86_PDPE_A | PGM_PLXFLAGS_PERMANENT
1242 | MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[iPD]);
1243 }
1244
1245 RTHCPHYS HCPhysInterPaePDPT64 = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT64);
1246 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.pInterPaePML4->a); i++)
1247 pVM->pgm.s.pInterPaePML4->a[i].u = X86_PML4E_P | X86_PML4E_RW | X86_PML4E_US | X86_PML4E_A | PGM_PLXFLAGS_PERMANENT
1248 | HCPhysInterPaePDPT64;
1249
1250 /*
1251 * Allocate pages for the three possible guest contexts (AMD64, PAE and plain 32-Bit).
1252 * We allocate pages for all three posibilities to in order to simplify mappings and
1253 * avoid resource failure during mode switches. So, we need to cover all levels of the
1254 * of the first 4GB down to PD level.
1255 * As with the intermediate context, AMD64 uses the PAE PDPT and PDs.
1256 */
1257 pVM->pgm.s.pHC32BitPD = (PX86PD)MMR3PageAllocLow(pVM);
1258 pVM->pgm.s.apHCPaePDs[0] = (PX86PDPAE)MMR3PageAlloc(pVM);
1259 pVM->pgm.s.apHCPaePDs[1] = (PX86PDPAE)MMR3PageAlloc(pVM);
1260 AssertRelease((uintptr_t)pVM->pgm.s.apHCPaePDs[0] + PAGE_SIZE == (uintptr_t)pVM->pgm.s.apHCPaePDs[1]);
1261 pVM->pgm.s.apHCPaePDs[2] = (PX86PDPAE)MMR3PageAlloc(pVM);
1262 AssertRelease((uintptr_t)pVM->pgm.s.apHCPaePDs[1] + PAGE_SIZE == (uintptr_t)pVM->pgm.s.apHCPaePDs[2]);
1263 pVM->pgm.s.apHCPaePDs[3] = (PX86PDPAE)MMR3PageAlloc(pVM);
1264 AssertRelease((uintptr_t)pVM->pgm.s.apHCPaePDs[2] + PAGE_SIZE == (uintptr_t)pVM->pgm.s.apHCPaePDs[3]);
1265 pVM->pgm.s.pHCPaePDPT = (PX86PDPT)MMR3PageAllocLow(pVM);
1266 if ( !pVM->pgm.s.pHC32BitPD
1267 || !pVM->pgm.s.apHCPaePDs[0]
1268 || !pVM->pgm.s.apHCPaePDs[1]
1269 || !pVM->pgm.s.apHCPaePDs[2]
1270 || !pVM->pgm.s.apHCPaePDs[3]
1271 || !pVM->pgm.s.pHCPaePDPT)
1272 {
1273 AssertMsgFailed(("Failed to allocate pages for the intermediate context!\n"));
1274 return VERR_NO_PAGE_MEMORY;
1275 }
1276
1277 /* get physical addresses. */
1278 pVM->pgm.s.HCPhys32BitPD = MMPage2Phys(pVM, pVM->pgm.s.pHC32BitPD);
1279 Assert(MMPagePhys2Page(pVM, pVM->pgm.s.HCPhys32BitPD) == pVM->pgm.s.pHC32BitPD);
1280 pVM->pgm.s.aHCPhysPaePDs[0] = MMPage2Phys(pVM, pVM->pgm.s.apHCPaePDs[0]);
1281 pVM->pgm.s.aHCPhysPaePDs[1] = MMPage2Phys(pVM, pVM->pgm.s.apHCPaePDs[1]);
1282 pVM->pgm.s.aHCPhysPaePDs[2] = MMPage2Phys(pVM, pVM->pgm.s.apHCPaePDs[2]);
1283 pVM->pgm.s.aHCPhysPaePDs[3] = MMPage2Phys(pVM, pVM->pgm.s.apHCPaePDs[3]);
1284 pVM->pgm.s.HCPhysPaePDPT = MMPage2Phys(pVM, pVM->pgm.s.pHCPaePDPT);
1285
1286 /*
1287 * Initialize the pages, setting up the PML4 and PDPT for action below 4GB.
1288 */
1289 ASMMemZero32(pVM->pgm.s.pHC32BitPD, PAGE_SIZE);
1290
1291 ASMMemZero32(pVM->pgm.s.pHCPaePDPT, PAGE_SIZE);
1292 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.apHCPaePDs); i++)
1293 {
1294 ASMMemZero32(pVM->pgm.s.apHCPaePDs[i], PAGE_SIZE);
1295 pVM->pgm.s.pHCPaePDPT->a[i].u = X86_PDPE_P | PGM_PLXFLAGS_PERMANENT | pVM->pgm.s.aHCPhysPaePDs[i];
1296 /* The flags will be corrected when entering and leaving long mode. */
1297 }
1298
1299 CPUMSetHyperCR3(pVM, (uint32_t)pVM->pgm.s.HCPhys32BitPD);
1300
1301 /*
1302 * Initialize paging workers and mode from current host mode
1303 * and the guest running in real mode.
1304 */
1305 pVM->pgm.s.enmHostMode = SUPGetPagingMode();
1306 switch (pVM->pgm.s.enmHostMode)
1307 {
1308 case SUPPAGINGMODE_32_BIT:
1309 case SUPPAGINGMODE_32_BIT_GLOBAL:
1310 case SUPPAGINGMODE_PAE:
1311 case SUPPAGINGMODE_PAE_GLOBAL:
1312 case SUPPAGINGMODE_PAE_NX:
1313 case SUPPAGINGMODE_PAE_GLOBAL_NX:
1314 break;
1315
1316 case SUPPAGINGMODE_AMD64:
1317 case SUPPAGINGMODE_AMD64_GLOBAL:
1318 case SUPPAGINGMODE_AMD64_NX:
1319 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
1320#ifndef VBOX_WITH_HYBIRD_32BIT_KERNEL
1321 if (ARCH_BITS != 64)
1322 {
1323 AssertMsgFailed(("Host mode %d (64-bit) is not supported by non-64bit builds\n", pVM->pgm.s.enmHostMode));
1324 LogRel(("Host mode %d (64-bit) is not supported by non-64bit builds\n", pVM->pgm.s.enmHostMode));
1325 return VERR_PGM_UNSUPPORTED_HOST_PAGING_MODE;
1326 }
1327#endif
1328 break;
1329 default:
1330 AssertMsgFailed(("Host mode %d is not supported\n", pVM->pgm.s.enmHostMode));
1331 return VERR_PGM_UNSUPPORTED_HOST_PAGING_MODE;
1332 }
1333 rc = pgmR3ModeDataInit(pVM, false /* don't resolve GC and R0 syms yet */);
1334 if (VBOX_SUCCESS(rc))
1335 rc = PGMR3ChangeMode(pVM, PGMMODE_REAL);
1336 if (VBOX_SUCCESS(rc))
1337 {
1338 LogFlow(("pgmR3InitPaging: returns successfully\n"));
1339#if HC_ARCH_BITS == 64
1340LogRel(("Debug: HCPhys32BitPD=%VHp aHCPhysPaePDs={%VHp,%VHp,%VHp,%VHp} HCPhysPaePDPT=%VHp HCPhysPaePML4=%VHp\n",
1341 pVM->pgm.s.HCPhys32BitPD, pVM->pgm.s.aHCPhysPaePDs[0], pVM->pgm.s.aHCPhysPaePDs[1], pVM->pgm.s.aHCPhysPaePDs[2], pVM->pgm.s.aHCPhysPaePDs[3],
1342 pVM->pgm.s.HCPhysPaePDPT, pVM->pgm.s.HCPhysPaePML4));
1343LogRel(("Debug: HCPhysInterPD=%VHp HCPhysInterPaePDPT=%VHp HCPhysInterPaePML4=%VHp\n",
1344 pVM->pgm.s.HCPhysInterPD, pVM->pgm.s.HCPhysInterPaePDPT, pVM->pgm.s.HCPhysInterPaePML4));
1345LogRel(("Debug: apInterPTs={%VHp,%VHp} apInterPaePTs={%VHp,%VHp} apInterPaePDs={%VHp,%VHp,%VHp,%VHp} pInterPaePDPT64=%VHp\n",
1346 MMPage2Phys(pVM, pVM->pgm.s.apInterPTs[0]), MMPage2Phys(pVM, pVM->pgm.s.apInterPTs[1]),
1347 MMPage2Phys(pVM, pVM->pgm.s.apInterPaePTs[0]), MMPage2Phys(pVM, pVM->pgm.s.apInterPaePTs[1]),
1348 MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[0]), MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[1]), MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[2]), MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[3]),
1349 MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT64)));
1350#endif
1351
1352 return VINF_SUCCESS;
1353 }
1354
1355 LogFlow(("pgmR3InitPaging: returns %Vrc\n", rc));
1356 return rc;
1357}
1358
1359
1360#ifdef VBOX_WITH_STATISTICS
1361/**
1362 * Init statistics
1363 */
1364static void pgmR3InitStats(PVM pVM)
1365{
1366 PPGM pPGM = &pVM->pgm.s;
1367 STAM_REG(pVM, &pPGM->StatGCInvalidatePage, STAMTYPE_PROFILE, "/PGM/GC/InvalidatePage", STAMUNIT_TICKS_PER_CALL, "PGMGCInvalidatePage() profiling.");
1368 STAM_REG(pVM, &pPGM->StatGCInvalidatePage4KBPages, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/4KBPages", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for a 4KB page.");
1369 STAM_REG(pVM, &pPGM->StatGCInvalidatePage4MBPages, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/4MBPages", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for a 4MB page.");
1370 STAM_REG(pVM, &pPGM->StatGCInvalidatePage4MBPagesSkip, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/4MBPagesSkip",STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() skipped a 4MB page.");
1371 STAM_REG(pVM, &pPGM->StatGCInvalidatePagePDMappings, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/PDMappings", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for a page directory containing mappings (no conflict).");
1372 STAM_REG(pVM, &pPGM->StatGCInvalidatePagePDNAs, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/PDNAs", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for a not accessed page directory.");
1373 STAM_REG(pVM, &pPGM->StatGCInvalidatePagePDNPs, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/PDNPs", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for a not present page directory.");
1374 STAM_REG(pVM, &pPGM->StatGCInvalidatePagePDOutOfSync, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/PDOutOfSync", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for an out of sync page directory.");
1375 STAM_REG(pVM, &pPGM->StatGCInvalidatePageSkipped, STAMTYPE_COUNTER, "/PGM/GC/InvalidatePage/Skipped", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was skipped due to not present shw or pending pending SyncCR3.");
1376 STAM_REG(pVM, &pPGM->StatGCSyncPT, STAMTYPE_PROFILE, "/PGM/GC/SyncPT", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMGCSyncPT() body.");
1377 STAM_REG(pVM, &pPGM->StatGCAccessedPage, STAMTYPE_COUNTER, "/PGM/GC/AccessedPage", STAMUNIT_OCCURENCES, "The number of pages marked not present for accessed bit emulation.");
1378 STAM_REG(pVM, &pPGM->StatGCDirtyPage, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/Mark", STAMUNIT_OCCURENCES, "The number of pages marked read-only for dirty bit tracking.");
1379 STAM_REG(pVM, &pPGM->StatGCDirtyPageBig, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/MarkBig", STAMUNIT_OCCURENCES, "The number of 4MB pages marked read-only for dirty bit tracking.");
1380 STAM_REG(pVM, &pPGM->StatGCDirtyPageTrap, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/Trap", STAMUNIT_OCCURENCES, "The number of traps generated for dirty bit tracking.");
1381 STAM_REG(pVM, &pPGM->StatGCDirtyPageSkipped, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/Skipped", STAMUNIT_OCCURENCES, "The number of pages already dirty or readonly.");
1382 STAM_REG(pVM, &pPGM->StatGCDirtiedPage, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/SetDirty", STAMUNIT_OCCURENCES, "The number of pages marked dirty because of write accesses.");
1383 STAM_REG(pVM, &pPGM->StatGCDirtyTrackRealPF, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/RealPF", STAMUNIT_OCCURENCES, "The number of real pages faults during dirty bit tracking.");
1384 STAM_REG(pVM, &pPGM->StatGCPageAlreadyDirty, STAMTYPE_COUNTER, "/PGM/GC/DirtyPage/AlreadySet", STAMUNIT_OCCURENCES, "The number of pages already marked dirty because of write accesses.");
1385 STAM_REG(pVM, &pPGM->StatGCDirtyBitTracking, STAMTYPE_PROFILE, "/PGM/GC/DirtyPage", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMTrackDirtyBit() body.");
1386 STAM_REG(pVM, &pPGM->StatGCSyncPTAlloc, STAMTYPE_COUNTER, "/PGM/GC/SyncPT/Alloc", STAMUNIT_OCCURENCES, "The number of times PGMGCSyncPT() needed to allocate page tables.");
1387 STAM_REG(pVM, &pPGM->StatGCSyncPTConflict, STAMTYPE_COUNTER, "/PGM/GC/SyncPT/Conflicts", STAMUNIT_OCCURENCES, "The number of times PGMGCSyncPT() detected conflicts.");
1388 STAM_REG(pVM, &pPGM->StatGCSyncPTFailed, STAMTYPE_COUNTER, "/PGM/GC/SyncPT/Failed", STAMUNIT_OCCURENCES, "The number of times PGMGCSyncPT() failed.");
1389
1390 STAM_REG(pVM, &pPGM->StatGCTrap0e, STAMTYPE_PROFILE, "/PGM/GC/Trap0e", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMGCTrap0eHandler() body.");
1391 STAM_REG(pVM, &pPGM->StatCheckPageFault, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/CheckPageFault", STAMUNIT_TICKS_PER_CALL, "Profiling of checking for dirty/access emulation faults.");
1392 STAM_REG(pVM, &pPGM->StatLazySyncPT, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/SyncPT", STAMUNIT_TICKS_PER_CALL, "Profiling of lazy page table syncing.");
1393 STAM_REG(pVM, &pPGM->StatMapping, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/Mapping", STAMUNIT_TICKS_PER_CALL, "Profiling of checking virtual mappings.");
1394 STAM_REG(pVM, &pPGM->StatOutOfSync, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/OutOfSync", STAMUNIT_TICKS_PER_CALL, "Profiling of out of sync page handling.");
1395 STAM_REG(pVM, &pPGM->StatHandlers, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/Handlers", STAMUNIT_TICKS_PER_CALL, "Profiling of checking handlers.");
1396 STAM_REG(pVM, &pPGM->StatEIPHandlers, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time/EIPHandlers", STAMUNIT_TICKS_PER_CALL, "Profiling of checking eip handlers.");
1397 STAM_REG(pVM, &pPGM->StatTrap0eCSAM, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/CSAM", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is CSAM.");
1398 STAM_REG(pVM, &pPGM->StatTrap0eDirtyAndAccessedBits, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/DirtyAndAccessedBits", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is dirty and/or accessed bit emulation.");
1399 STAM_REG(pVM, &pPGM->StatTrap0eGuestTrap, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/GuestTrap", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is a guest trap.");
1400 STAM_REG(pVM, &pPGM->StatTrap0eHndPhys, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/HandlerPhysical", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is a physical handler.");
1401 STAM_REG(pVM, &pPGM->StatTrap0eHndVirt, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/HandlerVirtual",STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is a virtual handler.");
1402 STAM_REG(pVM, &pPGM->StatTrap0eHndUnhandled, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/HandlerUnhandled", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is access outside the monitored areas of a monitored page.");
1403 STAM_REG(pVM, &pPGM->StatTrap0eMisc, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/Misc", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is not known.");
1404 STAM_REG(pVM, &pPGM->StatTrap0eOutOfSync, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/OutOfSync", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is an out-of-sync page.");
1405 STAM_REG(pVM, &pPGM->StatTrap0eOutOfSyncHndPhys, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/OutOfSyncHndPhys", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is an out-of-sync physical handler page.");
1406 STAM_REG(pVM, &pPGM->StatTrap0eOutOfSyncHndVirt, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/OutOfSyncHndVirt", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is an out-of-sync virtual handler page.");
1407 STAM_REG(pVM, &pPGM->StatTrap0eOutOfSyncObsHnd, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/OutOfSyncObsHnd", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is an obsolete handler page.");
1408 STAM_REG(pVM, &pPGM->StatTrap0eSyncPT, STAMTYPE_PROFILE, "/PGM/GC/Trap0e/Time2/SyncPT", STAMUNIT_TICKS_PER_CALL, "Profiling of the Trap0eHandler body when the cause is lazy syncing of a PT.");
1409
1410 STAM_REG(pVM, &pPGM->StatTrap0eMapHandler, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/Mapping", STAMUNIT_OCCURENCES, "Number of traps due to access handlers in mappings.");
1411 STAM_REG(pVM, &pPGM->StatHandlersOutOfSync, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/OutOfSync", STAMUNIT_OCCURENCES, "Number of traps due to out-of-sync handled pages.");
1412 STAM_REG(pVM, &pPGM->StatHandlersPhysical, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/Physical", STAMUNIT_OCCURENCES, "Number of traps due to physical access handlers.");
1413 STAM_REG(pVM, &pPGM->StatHandlersVirtual, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/Virtual", STAMUNIT_OCCURENCES, "Number of traps due to virtual access handlers.");
1414 STAM_REG(pVM, &pPGM->StatHandlersVirtualByPhys, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/VirtualByPhys", STAMUNIT_OCCURENCES, "Number of traps due to virtual access handlers by physical address.");
1415 STAM_REG(pVM, &pPGM->StatHandlersVirtualUnmarked, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/VirtualUnmarked", STAMUNIT_OCCURENCES,"Number of traps due to virtual access handlers by virtual address (without proper physical flags).");
1416 STAM_REG(pVM, &pPGM->StatHandlersUnhandled, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/Unhandled", STAMUNIT_OCCURENCES, "Number of traps due to access outside range of monitored page(s).");
1417 STAM_REG(pVM, &pPGM->StatHandlersInvalid, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Handlers/Invalid", STAMUNIT_OCCURENCES, "Number of traps due to access to invalid physical memory.");
1418
1419 STAM_REG(pVM, &pPGM->StatGCTrap0eConflicts, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Conflicts", STAMUNIT_OCCURENCES, "The number of times #PF was caused by an undetected conflict.");
1420 STAM_REG(pVM, &pPGM->StatGCTrap0eUSNotPresentRead, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/NPRead", STAMUNIT_OCCURENCES, "Number of user mode not present read page faults.");
1421 STAM_REG(pVM, &pPGM->StatGCTrap0eUSNotPresentWrite, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/NPWrite", STAMUNIT_OCCURENCES, "Number of user mode not present write page faults.");
1422 STAM_REG(pVM, &pPGM->StatGCTrap0eUSWrite, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/Write", STAMUNIT_OCCURENCES, "Number of user mode write page faults.");
1423 STAM_REG(pVM, &pPGM->StatGCTrap0eUSReserved, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/Reserved", STAMUNIT_OCCURENCES, "Number of user mode reserved bit page faults.");
1424 STAM_REG(pVM, &pPGM->StatGCTrap0eUSNXE, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/NXE", STAMUNIT_OCCURENCES, "Number of user mode NXE page faults.");
1425 STAM_REG(pVM, &pPGM->StatGCTrap0eUSRead, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/User/Read", STAMUNIT_OCCURENCES, "Number of user mode read page faults.");
1426
1427 STAM_REG(pVM, &pPGM->StatGCTrap0eSVNotPresentRead, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Supervisor/NPRead", STAMUNIT_OCCURENCES, "Number of supervisor mode not present read page faults.");
1428 STAM_REG(pVM, &pPGM->StatGCTrap0eSVNotPresentWrite, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Supervisor/NPWrite", STAMUNIT_OCCURENCES, "Number of supervisor mode not present write page faults.");
1429 STAM_REG(pVM, &pPGM->StatGCTrap0eSVWrite, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Supervisor/Write", STAMUNIT_OCCURENCES, "Number of supervisor mode write page faults.");
1430 STAM_REG(pVM, &pPGM->StatGCTrap0eSVReserved, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Supervisor/Reserved", STAMUNIT_OCCURENCES, "Number of supervisor mode reserved bit page faults.");
1431 STAM_REG(pVM, &pPGM->StatGCTrap0eSNXE, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/Supervisor/NXE", STAMUNIT_OCCURENCES, "Number of supervisor mode NXE page faults.");
1432 STAM_REG(pVM, &pPGM->StatGCTrap0eUnhandled, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/GuestPF/Unhandled", STAMUNIT_OCCURENCES, "Number of guest real page faults.");
1433 STAM_REG(pVM, &pPGM->StatGCTrap0eMap, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/GuestPF/Map", STAMUNIT_OCCURENCES, "Number of guest page faults due to map accesses.");
1434
1435 STAM_REG(pVM, &pPGM->StatTrap0eWPEmulGC, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/WP/InGC", STAMUNIT_OCCURENCES, "Number of guest page faults due to X86_CR0_WP emulation.");
1436 STAM_REG(pVM, &pPGM->StatTrap0eWPEmulR3, STAMTYPE_COUNTER, "/PGM/GC/Trap0e/WP/ToR3", STAMUNIT_OCCURENCES, "Number of guest page faults due to X86_CR0_WP emulation (forward to R3 for emulation).");
1437
1438 STAM_REG(pVM, &pPGM->StatGCGuestCR3WriteHandled, STAMTYPE_COUNTER, "/PGM/GC/CR3WriteInt", STAMUNIT_OCCURENCES, "The number of times the Guest CR3 change was successfully handled.");
1439 STAM_REG(pVM, &pPGM->StatGCGuestCR3WriteUnhandled, STAMTYPE_COUNTER, "/PGM/GC/CR3WriteEmu", STAMUNIT_OCCURENCES, "The number of times the Guest CR3 change was passed back to the recompiler.");
1440 STAM_REG(pVM, &pPGM->StatGCGuestCR3WriteConflict, STAMTYPE_COUNTER, "/PGM/GC/CR3WriteConflict", STAMUNIT_OCCURENCES, "The number of times the Guest CR3 monitoring detected a conflict.");
1441
1442 STAM_REG(pVM, &pPGM->StatGCPageOutOfSyncSupervisor, STAMTYPE_COUNTER, "/PGM/GC/OutOfSync/SuperVisor", STAMUNIT_OCCURENCES, "Number of traps due to pages out of sync.");
1443 STAM_REG(pVM, &pPGM->StatGCPageOutOfSyncUser, STAMTYPE_COUNTER, "/PGM/GC/OutOfSync/User", STAMUNIT_OCCURENCES, "Number of traps due to pages out of sync.");
1444
1445 STAM_REG(pVM, &pPGM->StatGCGuestROMWriteHandled, STAMTYPE_COUNTER, "/PGM/GC/ROMWriteInt", STAMUNIT_OCCURENCES, "The number of times the Guest ROM change was successfully handled.");
1446 STAM_REG(pVM, &pPGM->StatGCGuestROMWriteUnhandled, STAMTYPE_COUNTER, "/PGM/GC/ROMWriteEmu", STAMUNIT_OCCURENCES, "The number of times the Guest ROM change was passed back to the recompiler.");
1447
1448 STAM_REG(pVM, &pPGM->StatDynMapCacheHits, STAMTYPE_COUNTER, "/PGM/GC/DynMapCache/Hits" , STAMUNIT_OCCURENCES, "Number of dynamic page mapping cache hits.");
1449 STAM_REG(pVM, &pPGM->StatDynMapCacheMisses, STAMTYPE_COUNTER, "/PGM/GC/DynMapCache/Misses" , STAMUNIT_OCCURENCES, "Number of dynamic page mapping cache misses.");
1450
1451 STAM_REG(pVM, &pPGM->StatHCDetectedConflicts, STAMTYPE_COUNTER, "/PGM/HC/DetectedConflicts", STAMUNIT_OCCURENCES, "The number of times PGMR3CheckMappingConflicts() detected a conflict.");
1452 STAM_REG(pVM, &pPGM->StatHCGuestPDWrite, STAMTYPE_COUNTER, "/PGM/HC/PDWrite", STAMUNIT_OCCURENCES, "The total number of times pgmHCGuestPDWriteHandler() was called.");
1453 STAM_REG(pVM, &pPGM->StatHCGuestPDWriteConflict, STAMTYPE_COUNTER, "/PGM/HC/PDWriteConflict", STAMUNIT_OCCURENCES, "The number of times pgmHCGuestPDWriteHandler() detected a conflict.");
1454
1455 STAM_REG(pVM, &pPGM->StatHCInvalidatePage, STAMTYPE_PROFILE, "/PGM/HC/InvalidatePage", STAMUNIT_TICKS_PER_CALL, "PGMHCInvalidatePage() profiling.");
1456 STAM_REG(pVM, &pPGM->StatHCInvalidatePage4KBPages, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/4KBPages", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was called for a 4KB page.");
1457 STAM_REG(pVM, &pPGM->StatHCInvalidatePage4MBPages, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/4MBPages", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was called for a 4MB page.");
1458 STAM_REG(pVM, &pPGM->StatHCInvalidatePage4MBPagesSkip, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/4MBPagesSkip",STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() skipped a 4MB page.");
1459 STAM_REG(pVM, &pPGM->StatHCInvalidatePagePDMappings, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/PDMappings", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was called for a page directory containing mappings (no conflict).");
1460 STAM_REG(pVM, &pPGM->StatHCInvalidatePagePDNAs, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/PDNAs", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was called for a not accessed page directory.");
1461 STAM_REG(pVM, &pPGM->StatHCInvalidatePagePDNPs, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/PDNPs", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was called for a not present page directory.");
1462 STAM_REG(pVM, &pPGM->StatHCInvalidatePagePDOutOfSync, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/PDOutOfSync", STAMUNIT_OCCURENCES, "The number of times PGMGCInvalidatePage() was called for an out of sync page directory.");
1463 STAM_REG(pVM, &pPGM->StatHCInvalidatePageSkipped, STAMTYPE_COUNTER, "/PGM/HC/InvalidatePage/Skipped", STAMUNIT_OCCURENCES, "The number of times PGMHCInvalidatePage() was skipped due to not present shw or pending pending SyncCR3.");
1464 STAM_REG(pVM, &pPGM->StatHCResolveConflict, STAMTYPE_PROFILE, "/PGM/HC/ResolveConflict", STAMUNIT_TICKS_PER_CALL, "pgmR3SyncPTResolveConflict() profiling (includes the entire relocation).");
1465 STAM_REG(pVM, &pPGM->StatHCPrefetch, STAMTYPE_PROFILE, "/PGM/HC/Prefetch", STAMUNIT_TICKS_PER_CALL, "PGMR3PrefetchPage profiling.");
1466
1467 STAM_REG(pVM, &pPGM->StatHCSyncPT, STAMTYPE_PROFILE, "/PGM/HC/SyncPT", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMR3SyncPT() body.");
1468 STAM_REG(pVM, &pPGM->StatHCAccessedPage, STAMTYPE_COUNTER, "/PGM/HC/AccessedPage", STAMUNIT_OCCURENCES, "The number of pages marked not present for accessed bit emulation.");
1469 STAM_REG(pVM, &pPGM->StatHCDirtyPage, STAMTYPE_COUNTER, "/PGM/HC/DirtyPage/Mark", STAMUNIT_OCCURENCES, "The number of pages marked read-only for dirty bit tracking.");
1470 STAM_REG(pVM, &pPGM->StatHCDirtyPageBig, STAMTYPE_COUNTER, "/PGM/HC/DirtyPage/MarkBig", STAMUNIT_OCCURENCES, "The number of 4MB pages marked read-only for dirty bit tracking.");
1471 STAM_REG(pVM, &pPGM->StatHCDirtyPageTrap, STAMTYPE_COUNTER, "/PGM/HC/DirtyPage/Trap", STAMUNIT_OCCURENCES, "The number of traps generated for dirty bit tracking.");
1472 STAM_REG(pVM, &pPGM->StatHCDirtyPageSkipped, STAMTYPE_COUNTER, "/PGM/HC/DirtyPage/Skipped", STAMUNIT_OCCURENCES, "The number of pages already dirty or readonly.");
1473 STAM_REG(pVM, &pPGM->StatHCDirtyBitTracking, STAMTYPE_PROFILE, "/PGM/HC/DirtyPage", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMTrackDirtyBit() body.");
1474
1475 STAM_REG(pVM, &pPGM->StatGCSyncPagePDNAs, STAMTYPE_COUNTER, "/PGM/GC/SyncPagePDNAs", STAMUNIT_OCCURENCES, "The number of time we've marked a PD not present from SyncPage to virtualize the accessed bit.");
1476 STAM_REG(pVM, &pPGM->StatGCSyncPagePDOutOfSync, STAMTYPE_COUNTER, "/PGM/GC/SyncPagePDOutOfSync", STAMUNIT_OCCURENCES, "The number of time we've encountered an out-of-sync PD in SyncPage.");
1477 STAM_REG(pVM, &pPGM->StatHCSyncPagePDNAs, STAMTYPE_COUNTER, "/PGM/HC/SyncPagePDNAs", STAMUNIT_OCCURENCES, "The number of time we've marked a PD not present from SyncPage to virtualize the accessed bit.");
1478 STAM_REG(pVM, &pPGM->StatHCSyncPagePDOutOfSync, STAMTYPE_COUNTER, "/PGM/HC/SyncPagePDOutOfSync", STAMUNIT_OCCURENCES, "The number of time we've encountered an out-of-sync PD in SyncPage.");
1479
1480 STAM_REG(pVM, &pPGM->StatFlushTLB, STAMTYPE_PROFILE, "/PGM/FlushTLB", STAMUNIT_OCCURENCES, "Profiling of the PGMFlushTLB() body.");
1481 STAM_REG(pVM, &pPGM->StatFlushTLBNewCR3, STAMTYPE_COUNTER, "/PGM/FlushTLB/NewCR3", STAMUNIT_OCCURENCES, "The number of times PGMFlushTLB was called with a new CR3, non-global. (switch)");
1482 STAM_REG(pVM, &pPGM->StatFlushTLBNewCR3Global, STAMTYPE_COUNTER, "/PGM/FlushTLB/NewCR3Global", STAMUNIT_OCCURENCES, "The number of times PGMFlushTLB was called with a new CR3, global. (switch)");
1483 STAM_REG(pVM, &pPGM->StatFlushTLBSameCR3, STAMTYPE_COUNTER, "/PGM/FlushTLB/SameCR3", STAMUNIT_OCCURENCES, "The number of times PGMFlushTLB was called with the same CR3, non-global. (flush)");
1484 STAM_REG(pVM, &pPGM->StatFlushTLBSameCR3Global, STAMTYPE_COUNTER, "/PGM/FlushTLB/SameCR3Global", STAMUNIT_OCCURENCES, "The number of times PGMFlushTLB was called with the same CR3, global. (flush)");
1485
1486 STAM_REG(pVM, &pPGM->StatGCSyncCR3, STAMTYPE_PROFILE, "/PGM/GC/SyncCR3", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMSyncCR3() body.");
1487 STAM_REG(pVM, &pPGM->StatGCSyncCR3Handlers, STAMTYPE_PROFILE, "/PGM/GC/SyncCR3/Handlers", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMSyncCR3() update handler section.");
1488 STAM_REG(pVM, &pPGM->StatGCSyncCR3HandlerVirtualUpdate, STAMTYPE_PROFILE, "/PGM/GC/SyncCR3/Handlers/VirtualUpdate",STAMUNIT_TICKS_PER_CALL, "Profiling of the virtual handler updates.");
1489 STAM_REG(pVM, &pPGM->StatGCSyncCR3HandlerVirtualReset, STAMTYPE_PROFILE, "/PGM/GC/SyncCR3/Handlers/VirtualReset", STAMUNIT_TICKS_PER_CALL, "Profiling of the virtual handler resets.");
1490 STAM_REG(pVM, &pPGM->StatGCSyncCR3Global, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/Global", STAMUNIT_OCCURENCES, "The number of global CR3 syncs.");
1491 STAM_REG(pVM, &pPGM->StatGCSyncCR3NotGlobal, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/NotGlobal", STAMUNIT_OCCURENCES, "The number of non-global CR3 syncs.");
1492 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstCacheHit, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstChacheHit", STAMUNIT_OCCURENCES, "The number of times we got some kind of a cache hit.");
1493 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstFreed, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstFreed", STAMUNIT_OCCURENCES, "The number of times we've had to free a shadow entry.");
1494 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstFreedSrcNP, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstFreedSrcNP", STAMUNIT_OCCURENCES, "The number of times we've had to free a shadow entry for which the source entry was not present.");
1495 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstNotPresent, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstNotPresent", STAMUNIT_OCCURENCES, "The number of times we've encountered a not present shadow entry for a present guest entry.");
1496 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstSkippedGlobalPD, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstSkippedGlobalPD", STAMUNIT_OCCURENCES, "The number of times a global page directory wasn't flushed.");
1497 STAM_REG(pVM, &pPGM->StatGCSyncCR3DstSkippedGlobalPT, STAMTYPE_COUNTER, "/PGM/GC/SyncCR3/DstSkippedGlobalPT", STAMUNIT_OCCURENCES, "The number of times a page table with only global entries wasn't flushed.");
1498
1499 STAM_REG(pVM, &pPGM->StatHCSyncCR3, STAMTYPE_PROFILE, "/PGM/HC/SyncCR3", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMSyncCR3() body.");
1500 STAM_REG(pVM, &pPGM->StatHCSyncCR3Handlers, STAMTYPE_PROFILE, "/PGM/HC/SyncCR3/Handlers", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMSyncCR3() update handler section.");
1501 STAM_REG(pVM, &pPGM->StatHCSyncCR3HandlerVirtualUpdate, STAMTYPE_PROFILE, "/PGM/HC/SyncCR3/Handlers/VirtualUpdate",STAMUNIT_TICKS_PER_CALL, "Profiling of the virtual handler updates.");
1502 STAM_REG(pVM, &pPGM->StatHCSyncCR3HandlerVirtualReset, STAMTYPE_PROFILE, "/PGM/HC/SyncCR3/Handlers/VirtualReset", STAMUNIT_TICKS_PER_CALL, "Profiling of the virtual handler resets.");
1503 STAM_REG(pVM, &pPGM->StatHCSyncCR3Global, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/Global", STAMUNIT_OCCURENCES, "The number of global CR3 syncs.");
1504 STAM_REG(pVM, &pPGM->StatHCSyncCR3NotGlobal, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/NotGlobal", STAMUNIT_OCCURENCES, "The number of non-global CR3 syncs.");
1505 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstCacheHit, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstChacheHit", STAMUNIT_OCCURENCES, "The number of times we got some kind of a cache hit.");
1506 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstFreed, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstFreed", STAMUNIT_OCCURENCES, "The number of times we've had to free a shadow entry.");
1507 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstFreedSrcNP, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstFreedSrcNP", STAMUNIT_OCCURENCES, "The number of times we've had to free a shadow entry for which the source entry was not present.");
1508 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstNotPresent, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstNotPresent", STAMUNIT_OCCURENCES, "The number of times we've encountered a not present shadow entry for a present guest entry.");
1509 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstSkippedGlobalPD, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstSkippedGlobalPD", STAMUNIT_OCCURENCES, "The number of times a global page directory wasn't flushed.");
1510 STAM_REG(pVM, &pPGM->StatHCSyncCR3DstSkippedGlobalPT, STAMTYPE_COUNTER, "/PGM/HC/SyncCR3/DstSkippedGlobalPT", STAMUNIT_OCCURENCES, "The number of times a page table with only global entries wasn't flushed.");
1511
1512 STAM_REG(pVM, &pPGM->StatVirtHandleSearchByPhysGC, STAMTYPE_PROFILE, "/PGM/VirtHandler/SearchByPhys/GC", STAMUNIT_TICKS_PER_CALL, "Profiling of pgmHandlerVirtualFindByPhysAddr in GC.");
1513 STAM_REG(pVM, &pPGM->StatVirtHandleSearchByPhysHC, STAMTYPE_PROFILE, "/PGM/VirtHandler/SearchByPhys/HC", STAMUNIT_TICKS_PER_CALL, "Profiling of pgmHandlerVirtualFindByPhysAddr in HC.");
1514 STAM_REG(pVM, &pPGM->StatHandlePhysicalReset, STAMTYPE_COUNTER, "/PGM/HC/HandlerPhysicalReset", STAMUNIT_OCCURENCES, "The number of times PGMR3HandlerPhysicalReset is called.");
1515
1516 STAM_REG(pVM, &pPGM->StatHCGstModifyPage, STAMTYPE_PROFILE, "/PGM/HC/GstModifyPage", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMGstModifyPage() body.");
1517 STAM_REG(pVM, &pPGM->StatGCGstModifyPage, STAMTYPE_PROFILE, "/PGM/GC/GstModifyPage", STAMUNIT_TICKS_PER_CALL, "Profiling of the PGMGstModifyPage() body.");
1518
1519 STAM_REG(pVM, &pPGM->StatSynPT4kGC, STAMTYPE_COUNTER, "/PGM/GC/SyncPT/4k", STAMUNIT_OCCURENCES, "Nr of 4k PT syncs");
1520 STAM_REG(pVM, &pPGM->StatSynPT4kHC, STAMTYPE_COUNTER, "/PGM/HC/SyncPT/4k", STAMUNIT_OCCURENCES, "Nr of 4k PT syncs");
1521 STAM_REG(pVM, &pPGM->StatSynPT4MGC, STAMTYPE_COUNTER, "/PGM/GC/SyncPT/4M", STAMUNIT_OCCURENCES, "Nr of 4M PT syncs");
1522 STAM_REG(pVM, &pPGM->StatSynPT4MHC, STAMTYPE_COUNTER, "/PGM/HC/SyncPT/4M", STAMUNIT_OCCURENCES, "Nr of 4M PT syncs");
1523
1524 STAM_REG(pVM, &pPGM->StatDynRamTotal, STAMTYPE_COUNTER, "/PGM/RAM/TotalAlloc", STAMUNIT_MEGABYTES, "Allocated mbs of guest ram.");
1525 STAM_REG(pVM, &pPGM->StatDynRamGrow, STAMTYPE_COUNTER, "/PGM/RAM/Grow", STAMUNIT_OCCURENCES, "Nr of pgmr3PhysGrowRange calls.");
1526
1527 STAM_REG(pVM, &pPGM->StatPageHCMapTlbHits, STAMTYPE_COUNTER, "/PGM/PageHCMap/TlbHits", STAMUNIT_OCCURENCES, "TLB hits.");
1528 STAM_REG(pVM, &pPGM->StatPageHCMapTlbMisses, STAMTYPE_COUNTER, "/PGM/PageHCMap/TlbMisses", STAMUNIT_OCCURENCES, "TLB misses.");
1529 STAM_REG(pVM, &pPGM->ChunkR3Map.c, STAMTYPE_U32, "/PGM/ChunkR3Map/c", STAMUNIT_OCCURENCES, "Number of mapped chunks.");
1530 STAM_REG(pVM, &pPGM->ChunkR3Map.cMax, STAMTYPE_U32, "/PGM/ChunkR3Map/cMax", STAMUNIT_OCCURENCES, "Maximum number of mapped chunks.");
1531 STAM_REG(pVM, &pPGM->StatChunkR3MapTlbHits, STAMTYPE_COUNTER, "/PGM/ChunkR3Map/TlbHits", STAMUNIT_OCCURENCES, "TLB hits.");
1532 STAM_REG(pVM, &pPGM->StatChunkR3MapTlbMisses, STAMTYPE_COUNTER, "/PGM/ChunkR3Map/TlbMisses", STAMUNIT_OCCURENCES, "TLB misses.");
1533 STAM_REG(pVM, &pPGM->StatPageReplaceShared, STAMTYPE_COUNTER, "/PGM/Page/ReplacedShared", STAMUNIT_OCCURENCES, "Times a shared page was replaced.");
1534 STAM_REG(pVM, &pPGM->StatPageReplaceZero, STAMTYPE_COUNTER, "/PGM/Page/ReplacedZero", STAMUNIT_OCCURENCES, "Times the zero page was replaced.");
1535 STAM_REG(pVM, &pPGM->StatPageHandyAllocs, STAMTYPE_COUNTER, "/PGM/Page/HandyAllocs", STAMUNIT_OCCURENCES, "Number of times we've allocated more handy pages.");
1536 STAM_REG(pVM, &pPGM->cAllPages, STAMTYPE_U32, "/PGM/Page/cAllPages", STAMUNIT_OCCURENCES, "The total number of pages.");
1537 STAM_REG(pVM, &pPGM->cPrivatePages, STAMTYPE_U32, "/PGM/Page/cPrivatePages", STAMUNIT_OCCURENCES, "The number of private pages.");
1538 STAM_REG(pVM, &pPGM->cSharedPages, STAMTYPE_U32, "/PGM/Page/cSharedPages", STAMUNIT_OCCURENCES, "The number of shared pages.");
1539 STAM_REG(pVM, &pPGM->cZeroPages, STAMTYPE_U32, "/PGM/Page/cZeroPages", STAMUNIT_OCCURENCES, "The number of zero backed pages.");
1540
1541#ifdef PGMPOOL_WITH_GCPHYS_TRACKING
1542 STAM_REG(pVM, &pPGM->StatTrackVirgin, STAMTYPE_COUNTER, "/PGM/Track/Virgin", STAMUNIT_OCCURENCES, "The number of first time shadowings");
1543 STAM_REG(pVM, &pPGM->StatTrackAliased, STAMTYPE_COUNTER, "/PGM/Track/Aliased", STAMUNIT_OCCURENCES, "The number of times switching to cRef2, i.e. the page is being shadowed by two PTs.");
1544 STAM_REG(pVM, &pPGM->StatTrackAliasedMany, STAMTYPE_COUNTER, "/PGM/Track/AliasedMany", STAMUNIT_OCCURENCES, "The number of times we're tracking using cRef2.");
1545 STAM_REG(pVM, &pPGM->StatTrackAliasedLots, STAMTYPE_COUNTER, "/PGM/Track/AliasedLots", STAMUNIT_OCCURENCES, "The number of times we're hitting pages which has overflowed cRef2");
1546 STAM_REG(pVM, &pPGM->StatTrackOverflows, STAMTYPE_COUNTER, "/PGM/Track/Overflows", STAMUNIT_OCCURENCES, "The number of times the extent list grows to long.");
1547 STAM_REG(pVM, &pPGM->StatTrackDeref, STAMTYPE_PROFILE, "/PGM/Track/Deref", STAMUNIT_OCCURENCES, "Profiling of SyncPageWorkerTrackDeref (expensive).");
1548#endif
1549
1550 for (unsigned i = 0; i < X86_PG_ENTRIES; i++)
1551 {
1552 /** @todo r=bird: We need a STAMR3RegisterF()! */
1553 char szName[32];
1554
1555 RTStrPrintf(szName, sizeof(szName), "/PGM/GC/PD/Trap0e/%04X", i);
1556 int rc = STAMR3Register(pVM, &pPGM->StatGCTrap0ePD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, szName, STAMUNIT_OCCURENCES, "The number of traps in page directory n.");
1557 AssertRC(rc);
1558
1559 RTStrPrintf(szName, sizeof(szName), "/PGM/GC/PD/SyncPt/%04X", i);
1560 rc = STAMR3Register(pVM, &pPGM->StatGCSyncPtPD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, szName, STAMUNIT_OCCURENCES, "The number of syncs per PD n.");
1561 AssertRC(rc);
1562
1563 RTStrPrintf(szName, sizeof(szName), "/PGM/GC/PD/SyncPage/%04X", i);
1564 rc = STAMR3Register(pVM, &pPGM->StatGCSyncPagePD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, szName, STAMUNIT_OCCURENCES, "The number of out of sync pages per page directory n.");
1565 AssertRC(rc);
1566 }
1567}
1568#endif /* VBOX_WITH_STATISTICS */
1569
1570/**
1571 * Init the PGM bits that rely on VMMR0 and MM to be fully initialized.
1572 *
1573 * The dynamic mapping area will also be allocated and initialized at this
1574 * time. We could allocate it during PGMR3Init of course, but the mapping
1575 * wouldn't be allocated at that time preventing us from setting up the
1576 * page table entries with the dummy page.
1577 *
1578 * @returns VBox status code.
1579 * @param pVM VM handle.
1580 */
1581PGMR3DECL(int) PGMR3InitDynMap(PVM pVM)
1582{
1583 RTGCPTR GCPtr;
1584 /*
1585 * Reserve space for mapping the paging pages into guest context.
1586 */
1587 int rc = MMR3HyperReserve(pVM, PAGE_SIZE * (2 + ELEMENTS(pVM->pgm.s.apHCPaePDs) + 1 + 2 + 2), "Paging", &GCPtr);
1588 AssertRCReturn(rc, rc);
1589 pVM->pgm.s.pGC32BitPD = GCPtr;
1590 MMR3HyperReserve(pVM, PAGE_SIZE, "fence", NULL);
1591
1592 /*
1593 * Reserve space for the dynamic mappings.
1594 */
1595 /** @todo r=bird: Need to verify that the checks for crossing PTs are correct here. They seems to be assuming 4MB PTs.. */
1596 rc = MMR3HyperReserve(pVM, MM_HYPER_DYNAMIC_SIZE, "Dynamic mapping", &GCPtr);
1597 if (VBOX_SUCCESS(rc))
1598 pVM->pgm.s.pbDynPageMapBaseGC = GCPtr;
1599
1600 if ( VBOX_SUCCESS(rc)
1601 && (pVM->pgm.s.pbDynPageMapBaseGC >> X86_PD_SHIFT) != ((pVM->pgm.s.pbDynPageMapBaseGC + MM_HYPER_DYNAMIC_SIZE - 1) >> X86_PD_SHIFT))
1602 {
1603 rc = MMR3HyperReserve(pVM, MM_HYPER_DYNAMIC_SIZE, "Dynamic mapping not crossing", &GCPtr);
1604 if (VBOX_SUCCESS(rc))
1605 pVM->pgm.s.pbDynPageMapBaseGC = GCPtr;
1606 }
1607 if (VBOX_SUCCESS(rc))
1608 {
1609 AssertRelease((pVM->pgm.s.pbDynPageMapBaseGC >> X86_PD_SHIFT) == ((pVM->pgm.s.pbDynPageMapBaseGC + MM_HYPER_DYNAMIC_SIZE - 1) >> X86_PD_SHIFT));
1610 MMR3HyperReserve(pVM, PAGE_SIZE, "fence", NULL);
1611 }
1612 return rc;
1613}
1614
1615
1616/**
1617 * Ring-3 init finalizing.
1618 *
1619 * @returns VBox status code.
1620 * @param pVM The VM handle.
1621 */
1622PGMR3DECL(int) PGMR3InitFinalize(PVM pVM)
1623{
1624 /*
1625 * Map the paging pages into the guest context.
1626 */
1627 RTGCPTR GCPtr = pVM->pgm.s.pGC32BitPD;
1628 AssertReleaseReturn(GCPtr, VERR_INTERNAL_ERROR);
1629
1630 int rc = PGMMap(pVM, GCPtr, pVM->pgm.s.HCPhys32BitPD, PAGE_SIZE, 0);
1631 AssertRCReturn(rc, rc);
1632 pVM->pgm.s.pGC32BitPD = GCPtr;
1633 GCPtr += PAGE_SIZE;
1634 GCPtr += PAGE_SIZE; /* reserved page */
1635
1636 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.apHCPaePDs); i++)
1637 {
1638 rc = PGMMap(pVM, GCPtr, pVM->pgm.s.aHCPhysPaePDs[i], PAGE_SIZE, 0);
1639 AssertRCReturn(rc, rc);
1640 pVM->pgm.s.apGCPaePDs[i] = GCPtr;
1641 GCPtr += PAGE_SIZE;
1642 }
1643 /* A bit of paranoia is justified. */
1644 AssertRelease((RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[0] + PAGE_SIZE == (RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[1]);
1645 AssertRelease((RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[1] + PAGE_SIZE == (RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[2]);
1646 AssertRelease((RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[2] + PAGE_SIZE == (RTGCUINTPTR)pVM->pgm.s.apGCPaePDs[3]);
1647 GCPtr += PAGE_SIZE; /* reserved page */
1648
1649 rc = PGMMap(pVM, GCPtr, pVM->pgm.s.HCPhysPaePDPT, PAGE_SIZE, 0);
1650 AssertRCReturn(rc, rc);
1651 pVM->pgm.s.pGCPaePDPT = GCPtr;
1652 GCPtr += PAGE_SIZE;
1653 GCPtr += PAGE_SIZE; /* reserved page */
1654
1655
1656 /*
1657 * Reserve space for the dynamic mappings.
1658 * Initialize the dynamic mapping pages with dummy pages to simply the cache.
1659 */
1660 /* get the pointer to the page table entries. */
1661 PPGMMAPPING pMapping = pgmGetMapping(pVM, pVM->pgm.s.pbDynPageMapBaseGC);
1662 AssertRelease(pMapping);
1663 const uintptr_t off = pVM->pgm.s.pbDynPageMapBaseGC - pMapping->GCPtr;
1664 const unsigned iPT = off >> X86_PD_SHIFT;
1665 const unsigned iPG = (off >> X86_PT_SHIFT) & X86_PT_MASK;
1666 pVM->pgm.s.paDynPageMap32BitPTEsGC = pMapping->aPTs[iPT].pPTGC + iPG * sizeof(pMapping->aPTs[0].pPTR3->a[0]);
1667 pVM->pgm.s.paDynPageMapPaePTEsGC = pMapping->aPTs[iPT].paPaePTsGC + iPG * sizeof(pMapping->aPTs[0].paPaePTsR3->a[0]);
1668
1669 /* init cache */
1670 RTHCPHYS HCPhysDummy = MMR3PageDummyHCPhys(pVM);
1671 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.aHCPhysDynPageMapCache); i++)
1672 pVM->pgm.s.aHCPhysDynPageMapCache[i] = HCPhysDummy;
1673
1674 for (unsigned i = 0; i < MM_HYPER_DYNAMIC_SIZE; i += PAGE_SIZE)
1675 {
1676 rc = PGMMap(pVM, pVM->pgm.s.pbDynPageMapBaseGC + i, HCPhysDummy, PAGE_SIZE, 0);
1677 AssertRCReturn(rc, rc);
1678 }
1679
1680 return rc;
1681}
1682
1683
1684/**
1685 * Applies relocations to data and code managed by this
1686 * component. This function will be called at init and
1687 * whenever the VMM need to relocate it self inside the GC.
1688 *
1689 * @param pVM The VM.
1690 * @param offDelta Relocation delta relative to old location.
1691 */
1692PGMR3DECL(void) PGMR3Relocate(PVM pVM, RTGCINTPTR offDelta)
1693{
1694 LogFlow(("PGMR3Relocate\n"));
1695
1696 /*
1697 * Paging stuff.
1698 */
1699 pVM->pgm.s.GCPtrCR3Mapping += offDelta;
1700 /** @todo move this into shadow and guest specific relocation functions. */
1701 AssertMsg(pVM->pgm.s.pGC32BitPD, ("Init order, no relocation before paging is initialized!\n"));
1702 pVM->pgm.s.pGC32BitPD += offDelta;
1703 pVM->pgm.s.pGuestPDGC += offDelta;
1704 AssertCompile(ELEMENTS(pVM->pgm.s.apGCPaePDs) == ELEMENTS(pVM->pgm.s.apGstPaePDsGC));
1705 for (unsigned i = 0; i < ELEMENTS(pVM->pgm.s.apGCPaePDs); i++)
1706 {
1707 pVM->pgm.s.apGCPaePDs[i] += offDelta;
1708 pVM->pgm.s.apGstPaePDsGC[i] += offDelta;
1709 }
1710 pVM->pgm.s.pGstPaePDPTGC += offDelta;
1711 pVM->pgm.s.pGCPaePDPT += offDelta;
1712
1713 pgmR3ModeDataInit(pVM, true /* resolve GC/R0 symbols */);
1714 pgmR3ModeDataSwitch(pVM, pVM->pgm.s.enmShadowMode, pVM->pgm.s.enmGuestMode);
1715
1716 PGM_SHW_PFN(Relocate, pVM)(pVM, offDelta);
1717 PGM_GST_PFN(Relocate, pVM)(pVM, offDelta);
1718 PGM_BTH_PFN(Relocate, pVM)(pVM, offDelta);
1719
1720 /*
1721 * Trees.
1722 */
1723 pVM->pgm.s.pTreesGC = MMHyperHC2GC(pVM, pVM->pgm.s.pTreesHC);
1724
1725 /*
1726 * Ram ranges.
1727 */
1728 if (pVM->pgm.s.pRamRangesR3)
1729 {
1730 pVM->pgm.s.pRamRangesGC = MMHyperHC2GC(pVM, pVM->pgm.s.pRamRangesR3);
1731 for (PPGMRAMRANGE pCur = pVM->pgm.s.pRamRangesR3; pCur->pNextR3; pCur = pCur->pNextR3)
1732#ifdef VBOX_WITH_NEW_PHYS_CODE
1733 pCur->pNextGC = MMHyperR3ToGC(pVM, pCur->pNextR3);
1734#else
1735 {
1736 pCur->pNextGC = MMHyperR3ToGC(pVM, pCur->pNextR3);
1737 if (pCur->pavHCChunkGC)
1738 pCur->pavHCChunkGC = MMHyperHC2GC(pVM, pCur->pavHCChunkHC);
1739 }
1740#endif
1741 }
1742
1743 /*
1744 * Update the two page directories with all page table mappings.
1745 * (One or more of them have changed, that's why we're here.)
1746 */
1747 pVM->pgm.s.pMappingsGC = MMHyperHC2GC(pVM, pVM->pgm.s.pMappingsR3);
1748 for (PPGMMAPPING pCur = pVM->pgm.s.pMappingsR3; pCur->pNextR3; pCur = pCur->pNextR3)
1749 pCur->pNextGC = MMHyperHC2GC(pVM, pCur->pNextR3);
1750
1751 /* Relocate GC addresses of Page Tables. */
1752 for (PPGMMAPPING pCur = pVM->pgm.s.pMappingsR3; pCur; pCur = pCur->pNextR3)
1753 {
1754 for (RTHCUINT i = 0; i < pCur->cPTs; i++)
1755 {
1756 pCur->aPTs[i].pPTGC = MMHyperR3ToGC(pVM, pCur->aPTs[i].pPTR3);
1757 pCur->aPTs[i].paPaePTsGC = MMHyperR3ToGC(pVM, pCur->aPTs[i].paPaePTsR3);
1758 }
1759 }
1760
1761 /*
1762 * Dynamic page mapping area.
1763 */
1764 pVM->pgm.s.paDynPageMap32BitPTEsGC += offDelta;
1765 pVM->pgm.s.paDynPageMapPaePTEsGC += offDelta;
1766 pVM->pgm.s.pbDynPageMapBaseGC += offDelta;
1767
1768 /*
1769 * The Zero page.
1770 */
1771 pVM->pgm.s.pvZeroPgR0 = MMHyperR3ToR0(pVM, pVM->pgm.s.pvZeroPgR3);
1772 AssertRelease(pVM->pgm.s.pvZeroPgR0);
1773
1774 /*
1775 * Physical and virtual handlers.
1776 */
1777 RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesHC->PhysHandlers, true, pgmR3RelocatePhysHandler, &offDelta);
1778 RTAvlroGCPtrDoWithAll(&pVM->pgm.s.pTreesHC->VirtHandlers, true, pgmR3RelocateVirtHandler, &offDelta);
1779 RTAvlroGCPtrDoWithAll(&pVM->pgm.s.pTreesHC->HyperVirtHandlers, true, pgmR3RelocateHyperVirtHandler, &offDelta);
1780
1781 /*
1782 * The page pool.
1783 */
1784 pgmR3PoolRelocate(pVM);
1785}
1786
1787
1788/**
1789 * Callback function for relocating a physical access handler.
1790 *
1791 * @returns 0 (continue enum)
1792 * @param pNode Pointer to a PGMPHYSHANDLER node.
1793 * @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
1794 * not certain the delta will fit in a void pointer for all possible configs.
1795 */
1796static DECLCALLBACK(int) pgmR3RelocatePhysHandler(PAVLROGCPHYSNODECORE pNode, void *pvUser)
1797{
1798 PPGMPHYSHANDLER pHandler = (PPGMPHYSHANDLER)pNode;
1799 RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
1800 if (pHandler->pfnHandlerGC)
1801 pHandler->pfnHandlerGC += offDelta;
1802 if ((RTGCUINTPTR)pHandler->pvUserGC >= 0x10000)
1803 pHandler->pvUserGC += offDelta;
1804 return 0;
1805}
1806
1807
1808/**
1809 * Callback function for relocating a virtual access handler.
1810 *
1811 * @returns 0 (continue enum)
1812 * @param pNode Pointer to a PGMVIRTHANDLER node.
1813 * @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
1814 * not certain the delta will fit in a void pointer for all possible configs.
1815 */
1816static DECLCALLBACK(int) pgmR3RelocateVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser)
1817{
1818 PPGMVIRTHANDLER pHandler = (PPGMVIRTHANDLER)pNode;
1819 RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
1820 Assert( pHandler->enmType == PGMVIRTHANDLERTYPE_ALL
1821 || pHandler->enmType == PGMVIRTHANDLERTYPE_WRITE);
1822 Assert(pHandler->pfnHandlerGC);
1823 pHandler->pfnHandlerGC += offDelta;
1824 return 0;
1825}
1826
1827
1828/**
1829 * Callback function for relocating a virtual access handler for the hypervisor mapping.
1830 *
1831 * @returns 0 (continue enum)
1832 * @param pNode Pointer to a PGMVIRTHANDLER node.
1833 * @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
1834 * not certain the delta will fit in a void pointer for all possible configs.
1835 */
1836static DECLCALLBACK(int) pgmR3RelocateHyperVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser)
1837{
1838 PPGMVIRTHANDLER pHandler = (PPGMVIRTHANDLER)pNode;
1839 RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
1840 Assert(pHandler->enmType == PGMVIRTHANDLERTYPE_HYPERVISOR);
1841 Assert(pHandler->pfnHandlerGC);
1842 pHandler->pfnHandlerGC += offDelta;
1843 return 0;
1844}
1845
1846
1847/**
1848 * The VM is being reset.
1849 *
1850 * For the PGM component this means that any PD write monitors
1851 * needs to be removed.
1852 *
1853 * @param pVM VM handle.
1854 */
1855PGMR3DECL(void) PGMR3Reset(PVM pVM)
1856{
1857 LogFlow(("PGMR3Reset:\n"));
1858 VM_ASSERT_EMT(pVM);
1859
1860 pgmLock(pVM);
1861
1862 /*
1863 * Unfix any fixed mappings and disable CR3 monitoring.
1864 */
1865 pVM->pgm.s.fMappingsFixed = false;
1866 pVM->pgm.s.GCPtrMappingFixed = 0;
1867 pVM->pgm.s.cbMappingFixed = 0;
1868
1869 int rc = PGM_GST_PFN(UnmonitorCR3, pVM)(pVM);
1870 AssertRC(rc);
1871#ifdef DEBUG
1872 DBGFR3InfoLog(pVM, "mappings", NULL);
1873 DBGFR3InfoLog(pVM, "handlers", "all nostat");
1874#endif
1875
1876 /*
1877 * Reset the shadow page pool.
1878 */
1879 pgmR3PoolReset(pVM);
1880
1881 /*
1882 * Re-init other members.
1883 */
1884 pVM->pgm.s.fA20Enabled = true;
1885
1886 /*
1887 * Clear the FFs PGM owns.
1888 */
1889 VM_FF_CLEAR(pVM, VM_FF_PGM_SYNC_CR3);
1890 VM_FF_CLEAR(pVM, VM_FF_PGM_SYNC_CR3_NON_GLOBAL);
1891
1892 /*
1893 * Reset (zero) RAM pages.
1894 */
1895 rc = pgmR3PhysRamReset(pVM);
1896 if (RT_SUCCESS(rc))
1897 {
1898#ifdef VBOX_WITH_NEW_PHYS_CODE
1899 /*
1900 * Reset (zero) shadow ROM pages.
1901 */
1902 rc = pgmR3PhysRomReset(pVM);
1903#endif
1904 if (RT_SUCCESS(rc))
1905 {
1906 /*
1907 * Switch mode back to real mode.
1908 */
1909 rc = PGMR3ChangeMode(pVM, PGMMODE_REAL);
1910 STAM_REL_COUNTER_RESET(&pVM->pgm.s.cGuestModeChanges);
1911 }
1912 }
1913
1914 pgmUnlock(pVM);
1915 //return rc;
1916 AssertReleaseRC(rc);
1917}
1918
1919
1920#ifdef VBOX_STRICT
1921/**
1922 * VM state change callback for clearing fNoMorePhysWrites after
1923 * a snapshot has been created.
1924 */
1925static DECLCALLBACK(void) pgmR3ResetNoMorePhysWritesFlag(PVM pVM, VMSTATE enmState, VMSTATE enmOldState, void *pvUser)
1926{
1927 if (enmState == VMSTATE_RUNNING)
1928 pVM->pgm.s.fNoMorePhysWrites = false;
1929}
1930#endif
1931
1932
1933/**
1934 * Terminates the PGM.
1935 *
1936 * @returns VBox status code.
1937 * @param pVM Pointer to VM structure.
1938 */
1939PGMR3DECL(int) PGMR3Term(PVM pVM)
1940{
1941 return PDMR3CritSectDelete(&pVM->pgm.s.CritSect);
1942}
1943
1944
1945/**
1946 * Execute state save operation.
1947 *
1948 * @returns VBox status code.
1949 * @param pVM VM Handle.
1950 * @param pSSM SSM operation handle.
1951 */
1952static DECLCALLBACK(int) pgmR3Save(PVM pVM, PSSMHANDLE pSSM)
1953{
1954 PPGM pPGM = &pVM->pgm.s;
1955
1956 /* No more writes to physical memory after this point! */
1957 pVM->pgm.s.fNoMorePhysWrites = true;
1958
1959 /*
1960 * Save basic data (required / unaffected by relocation).
1961 */
1962#if 1
1963 SSMR3PutBool(pSSM, pPGM->fMappingsFixed);
1964#else
1965 SSMR3PutUInt(pSSM, pPGM->fMappingsFixed);
1966#endif
1967 SSMR3PutGCPtr(pSSM, pPGM->GCPtrMappingFixed);
1968 SSMR3PutU32(pSSM, pPGM->cbMappingFixed);
1969 SSMR3PutUInt(pSSM, pPGM->cbRamSize);
1970 SSMR3PutGCPhys(pSSM, pPGM->GCPhysA20Mask);
1971 SSMR3PutUInt(pSSM, pPGM->fA20Enabled);
1972 SSMR3PutUInt(pSSM, pPGM->fSyncFlags);
1973 SSMR3PutUInt(pSSM, pPGM->enmGuestMode);
1974 SSMR3PutU32(pSSM, ~0); /* Separator. */
1975
1976 /*
1977 * The guest mappings.
1978 */
1979 uint32_t i = 0;
1980 for (PPGMMAPPING pMapping = pPGM->pMappingsR3; pMapping; pMapping = pMapping->pNextR3, i++)
1981 {
1982 SSMR3PutU32(pSSM, i);
1983 SSMR3PutStrZ(pSSM, pMapping->pszDesc); /* This is the best unique id we have... */
1984 SSMR3PutGCPtr(pSSM, pMapping->GCPtr);
1985 SSMR3PutGCUIntPtr(pSSM, pMapping->cPTs);
1986 /* flags are done by the mapping owners! */
1987 }
1988 SSMR3PutU32(pSSM, ~0); /* terminator. */
1989
1990 /*
1991 * Ram range flags and bits.
1992 */
1993 i = 0;
1994 for (PPGMRAMRANGE pRam = pPGM->pRamRangesR3; pRam; pRam = pRam->pNextR3, i++)
1995 {
1996 /** @todo MMIO ranges may move (PCI reconfig), we currently assume they don't. */
1997
1998 SSMR3PutU32(pSSM, i);
1999 SSMR3PutGCPhys(pSSM, pRam->GCPhys);
2000 SSMR3PutGCPhys(pSSM, pRam->GCPhysLast);
2001 SSMR3PutGCPhys(pSSM, pRam->cb);
2002 SSMR3PutU8(pSSM, !!pRam->pvHC); /* boolean indicating memory or not. */
2003
2004 /* Flags. */
2005 const unsigned cPages = pRam->cb >> PAGE_SHIFT;
2006 for (unsigned iPage = 0; iPage < cPages; iPage++)
2007 SSMR3PutU16(pSSM, (uint16_t)(pRam->aPages[iPage].HCPhys & ~X86_PTE_PAE_PG_MASK)); /** @todo PAGE FLAGS */
2008
2009 /* any memory associated with the range. */
2010 if (pRam->fFlags & MM_RAM_FLAGS_DYNAMIC_ALLOC)
2011 {
2012 for (unsigned iChunk = 0; iChunk < (pRam->cb >> PGM_DYNAMIC_CHUNK_SHIFT); iChunk++)
2013 {
2014 if (pRam->pavHCChunkHC[iChunk])
2015 {
2016 SSMR3PutU8(pSSM, 1); /* chunk present */
2017 SSMR3PutMem(pSSM, pRam->pavHCChunkHC[iChunk], PGM_DYNAMIC_CHUNK_SIZE);
2018 }
2019 else
2020 SSMR3PutU8(pSSM, 0); /* no chunk present */
2021 }
2022 }
2023 else if (pRam->pvHC)
2024 {
2025 int rc = SSMR3PutMem(pSSM, pRam->pvHC, pRam->cb);
2026 if (VBOX_FAILURE(rc))
2027 {
2028 Log(("pgmR3Save: SSMR3PutMem(, %p, %#x) -> %Vrc\n", pRam->pvHC, pRam->cb, rc));
2029 return rc;
2030 }
2031 }
2032 }
2033 return SSMR3PutU32(pSSM, ~0); /* terminator. */
2034}
2035
2036
2037/**
2038 * Execute state load operation.
2039 *
2040 * @returns VBox status code.
2041 * @param pVM VM Handle.
2042 * @param pSSM SSM operation handle.
2043 * @param u32Version Data layout version.
2044 */
2045static DECLCALLBACK(int) pgmR3Load(PVM pVM, PSSMHANDLE pSSM, uint32_t u32Version)
2046{
2047 /*
2048 * Validate version.
2049 */
2050 if (u32Version != PGM_SAVED_STATE_VERSION)
2051 {
2052 Log(("pgmR3Load: Invalid version u32Version=%d (current %d)!\n", u32Version, PGM_SAVED_STATE_VERSION));
2053 return VERR_SSM_UNSUPPORTED_DATA_UNIT_VERSION;
2054 }
2055
2056 /*
2057 * Call the reset function to make sure all the memory is cleared.
2058 */
2059 PGMR3Reset(pVM);
2060
2061 /*
2062 * Load basic data (required / unaffected by relocation).
2063 */
2064 PPGM pPGM = &pVM->pgm.s;
2065#if 1
2066 SSMR3GetBool(pSSM, &pPGM->fMappingsFixed);
2067#else
2068 uint32_t u;
2069 SSMR3GetU32(pSSM, &u);
2070 pPGM->fMappingsFixed = u;
2071#endif
2072 SSMR3GetGCPtr(pSSM, &pPGM->GCPtrMappingFixed);
2073 SSMR3GetU32(pSSM, &pPGM->cbMappingFixed);
2074
2075 RTUINT cbRamSize;
2076 int rc = SSMR3GetU32(pSSM, &cbRamSize);
2077 if (VBOX_FAILURE(rc))
2078 return rc;
2079 if (cbRamSize != pPGM->cbRamSize)
2080 return VERR_SSM_LOAD_MEMORY_SIZE_MISMATCH;
2081 SSMR3GetGCPhys(pSSM, &pPGM->GCPhysA20Mask);
2082 SSMR3GetUInt(pSSM, &pPGM->fA20Enabled);
2083 SSMR3GetUInt(pSSM, &pPGM->fSyncFlags);
2084 RTUINT uGuestMode;
2085 SSMR3GetUInt(pSSM, &uGuestMode);
2086 pPGM->enmGuestMode = (PGMMODE)uGuestMode;
2087
2088 /* check separator. */
2089 uint32_t u32Sep;
2090 SSMR3GetU32(pSSM, &u32Sep);
2091 if (VBOX_FAILURE(rc))
2092 return rc;
2093 if (u32Sep != (uint32_t)~0)
2094 {
2095 AssertMsgFailed(("u32Sep=%#x (first)\n", u32Sep));
2096 return VERR_SSM_DATA_UNIT_FORMAT_CHANGED;
2097 }
2098
2099 /*
2100 * The guest mappings.
2101 */
2102 uint32_t i = 0;
2103 for (;; i++)
2104 {
2105 /* Check the seqence number / separator. */
2106 rc = SSMR3GetU32(pSSM, &u32Sep);
2107 if (VBOX_FAILURE(rc))
2108 return rc;
2109 if (u32Sep == ~0U)
2110 break;
2111 if (u32Sep != i)
2112 {
2113 AssertMsgFailed(("u32Sep=%#x (last)\n", u32Sep));
2114 return VERR_SSM_DATA_UNIT_FORMAT_CHANGED;
2115 }
2116
2117 /* get the mapping details. */
2118 char szDesc[256];
2119 szDesc[0] = '\0';
2120 rc = SSMR3GetStrZ(pSSM, szDesc, sizeof(szDesc));
2121 if (VBOX_FAILURE(rc))
2122 return rc;
2123 RTGCPTR GCPtr;
2124 SSMR3GetGCPtr(pSSM, &GCPtr);
2125 RTGCUINTPTR cPTs;
2126 rc = SSMR3GetGCUIntPtr(pSSM, &cPTs);
2127 if (VBOX_FAILURE(rc))
2128 return rc;
2129
2130 /* find matching range. */
2131 PPGMMAPPING pMapping;
2132 for (pMapping = pPGM->pMappingsR3; pMapping; pMapping = pMapping->pNextR3)
2133 if ( pMapping->cPTs == cPTs
2134 && !strcmp(pMapping->pszDesc, szDesc))
2135 break;
2136 if (!pMapping)
2137 {
2138 LogRel(("Couldn't find mapping: cPTs=%#x szDesc=%s (GCPtr=%VGv)\n",
2139 cPTs, szDesc, GCPtr));
2140 AssertFailed();
2141 return VERR_SSM_LOAD_CONFIG_MISMATCH;
2142 }
2143
2144 /* relocate it. */
2145 if (pMapping->GCPtr != GCPtr)
2146 {
2147 AssertMsg((GCPtr >> X86_PD_SHIFT << X86_PD_SHIFT) == GCPtr, ("GCPtr=%VGv\n", GCPtr));
2148#if HC_ARCH_BITS == 64
2149LogRel(("Mapping: %VGv -> %VGv %s\n", pMapping->GCPtr, GCPtr, pMapping->pszDesc));
2150#endif
2151 pgmR3MapRelocate(pVM, pMapping, pMapping->GCPtr, GCPtr);
2152 }
2153 else
2154 Log(("pgmR3Load: '%s' needed no relocation (%VGv)\n", szDesc, GCPtr));
2155 }
2156
2157 /*
2158 * Ram range flags and bits.
2159 */
2160 i = 0;
2161 for (PPGMRAMRANGE pRam = pPGM->pRamRangesR3; pRam; pRam = pRam->pNextR3, i++)
2162 {
2163 /** @todo MMIO ranges may move (PCI reconfig), we currently assume they don't. */
2164 /* Check the seqence number / separator. */
2165 rc = SSMR3GetU32(pSSM, &u32Sep);
2166 if (VBOX_FAILURE(rc))
2167 return rc;
2168 if (u32Sep == ~0U)
2169 break;
2170 if (u32Sep != i)
2171 {
2172 AssertMsgFailed(("u32Sep=%#x (last)\n", u32Sep));
2173 return VERR_SSM_DATA_UNIT_FORMAT_CHANGED;
2174 }
2175
2176 /* Get the range details. */
2177 RTGCPHYS GCPhys;
2178 SSMR3GetGCPhys(pSSM, &GCPhys);
2179 RTGCPHYS GCPhysLast;
2180 SSMR3GetGCPhys(pSSM, &GCPhysLast);
2181 RTGCPHYS cb;
2182 SSMR3GetGCPhys(pSSM, &cb);
2183 uint8_t fHaveBits;
2184 rc = SSMR3GetU8(pSSM, &fHaveBits);
2185 if (VBOX_FAILURE(rc))
2186 return rc;
2187 if (fHaveBits & ~1)
2188 {
2189 AssertMsgFailed(("u32Sep=%#x (last)\n", u32Sep));
2190 return VERR_SSM_DATA_UNIT_FORMAT_CHANGED;
2191 }
2192
2193 /* Match it up with the current range. */
2194 if ( GCPhys != pRam->GCPhys
2195 || GCPhysLast != pRam->GCPhysLast
2196 || cb != pRam->cb
2197 || fHaveBits != !!pRam->pvHC)
2198 {
2199 LogRel(("Ram range: %VGp-%VGp %VGp bytes %s\n"
2200 "State : %VGp-%VGp %VGp bytes %s\n",
2201 pRam->GCPhys, pRam->GCPhysLast, pRam->cb, pRam->pvHC ? "bits" : "nobits",
2202 GCPhys, GCPhysLast, cb, fHaveBits ? "bits" : "nobits"));
2203 /*
2204 * If we're loading a state for debugging purpose, don't make a fuss if
2205 * the MMIO[2] and ROM stuff isn't 100% right, just skip the mismatches.
2206 */
2207 if ( SSMR3HandleGetAfter(pSSM) != SSMAFTER_DEBUG_IT
2208 || GCPhys < 8 * _1M)
2209 AssertFailedReturn(VERR_SSM_LOAD_CONFIG_MISMATCH);
2210
2211 RTGCPHYS cPages = ((GCPhysLast - GCPhys) + 1) >> PAGE_SHIFT;
2212 while (cPages-- > 0)
2213 {
2214 uint16_t u16Ignore;
2215 SSMR3GetU16(pSSM, &u16Ignore);
2216 }
2217 continue;
2218 }
2219
2220 /* Flags. */
2221 const unsigned cPages = pRam->cb >> PAGE_SHIFT;
2222 for (unsigned iPage = 0; iPage < cPages; iPage++)
2223 {
2224 uint16_t u16 = 0;
2225 SSMR3GetU16(pSSM, &u16);
2226 u16 &= PAGE_OFFSET_MASK & ~( RT_BIT(4) | RT_BIT(5) | RT_BIT(6)
2227 | RT_BIT(7) | RT_BIT(8) | RT_BIT(9) | RT_BIT(10) );
2228 // &= MM_RAM_FLAGS_DYNAMIC_ALLOC | MM_RAM_FLAGS_RESERVED | MM_RAM_FLAGS_ROM | MM_RAM_FLAGS_MMIO | MM_RAM_FLAGS_MMIO2
2229 pRam->aPages[iPage].HCPhys = PGM_PAGE_GET_HCPHYS(&pRam->aPages[iPage]) | (RTHCPHYS)u16; /** @todo PAGE FLAGS */
2230 }
2231
2232 /* any memory associated with the range. */
2233 if (pRam->fFlags & MM_RAM_FLAGS_DYNAMIC_ALLOC)
2234 {
2235 for (unsigned iChunk = 0; iChunk < (pRam->cb >> PGM_DYNAMIC_CHUNK_SHIFT); iChunk++)
2236 {
2237 uint8_t fValidChunk;
2238
2239 rc = SSMR3GetU8(pSSM, &fValidChunk);
2240 if (VBOX_FAILURE(rc))
2241 return rc;
2242 if (fValidChunk > 1)
2243 return VERR_SSM_DATA_UNIT_FORMAT_CHANGED;
2244
2245 if (fValidChunk)
2246 {
2247 if (!pRam->pavHCChunkHC[iChunk])
2248 {
2249 rc = pgmr3PhysGrowRange(pVM, pRam->GCPhys + iChunk * PGM_DYNAMIC_CHUNK_SIZE);
2250 if (VBOX_FAILURE(rc))
2251 return rc;
2252 }
2253 Assert(pRam->pavHCChunkHC[iChunk]);
2254
2255 SSMR3GetMem(pSSM, pRam->pavHCChunkHC[iChunk], PGM_DYNAMIC_CHUNK_SIZE);
2256 }
2257 /* else nothing to do */
2258 }
2259 }
2260 else if (pRam->pvHC)
2261 {
2262 int rc = SSMR3GetMem(pSSM, pRam->pvHC, pRam->cb);
2263 if (VBOX_FAILURE(rc))
2264 {
2265 Log(("pgmR3Save: SSMR3GetMem(, %p, %#x) -> %Vrc\n", pRam->pvHC, pRam->cb, rc));
2266 return rc;
2267 }
2268 }
2269 }
2270
2271 /*
2272 * We require a full resync now.
2273 */
2274 VM_FF_SET(pVM, VM_FF_PGM_SYNC_CR3_NON_GLOBAL);
2275 VM_FF_SET(pVM, VM_FF_PGM_SYNC_CR3);
2276 pPGM->fSyncFlags |= PGM_SYNC_UPDATE_PAGE_BIT_VIRTUAL;
2277 pPGM->fPhysCacheFlushPending = true;
2278 pgmR3HandlerPhysicalUpdateAll(pVM);
2279
2280 /*
2281 * Change the paging mode.
2282 */
2283 return PGMR3ChangeMode(pVM, pPGM->enmGuestMode);
2284}
2285
2286
2287/**
2288 * Show paging mode.
2289 *
2290 * @param pVM VM Handle.
2291 * @param pHlp The info helpers.
2292 * @param pszArgs "all" (default), "guest", "shadow" or "host".
2293 */
2294static DECLCALLBACK(void) pgmR3InfoMode(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2295{
2296 /* digest argument. */
2297 bool fGuest, fShadow, fHost;
2298 if (pszArgs)
2299 pszArgs = RTStrStripL(pszArgs);
2300 if (!pszArgs || !*pszArgs || strstr(pszArgs, "all"))
2301 fShadow = fHost = fGuest = true;
2302 else
2303 {
2304 fShadow = fHost = fGuest = false;
2305 if (strstr(pszArgs, "guest"))
2306 fGuest = true;
2307 if (strstr(pszArgs, "shadow"))
2308 fShadow = true;
2309 if (strstr(pszArgs, "host"))
2310 fHost = true;
2311 }
2312
2313 /* print info. */
2314 if (fGuest)
2315 pHlp->pfnPrintf(pHlp, "Guest paging mode: %s, changed %RU64 times, A20 %s\n",
2316 PGMGetModeName(pVM->pgm.s.enmGuestMode), pVM->pgm.s.cGuestModeChanges.c,
2317 pVM->pgm.s.fA20Enabled ? "enabled" : "disabled");
2318 if (fShadow)
2319 pHlp->pfnPrintf(pHlp, "Shadow paging mode: %s\n", PGMGetModeName(pVM->pgm.s.enmShadowMode));
2320 if (fHost)
2321 {
2322 const char *psz;
2323 switch (pVM->pgm.s.enmHostMode)
2324 {
2325 case SUPPAGINGMODE_INVALID: psz = "invalid"; break;
2326 case SUPPAGINGMODE_32_BIT: psz = "32-bit"; break;
2327 case SUPPAGINGMODE_32_BIT_GLOBAL: psz = "32-bit+G"; break;
2328 case SUPPAGINGMODE_PAE: psz = "PAE"; break;
2329 case SUPPAGINGMODE_PAE_GLOBAL: psz = "PAE+G"; break;
2330 case SUPPAGINGMODE_PAE_NX: psz = "PAE+NX"; break;
2331 case SUPPAGINGMODE_PAE_GLOBAL_NX: psz = "PAE+G+NX"; break;
2332 case SUPPAGINGMODE_AMD64: psz = "AMD64"; break;
2333 case SUPPAGINGMODE_AMD64_GLOBAL: psz = "AMD64+G"; break;
2334 case SUPPAGINGMODE_AMD64_NX: psz = "AMD64+NX"; break;
2335 case SUPPAGINGMODE_AMD64_GLOBAL_NX: psz = "AMD64+G+NX"; break;
2336 default: psz = "unknown"; break;
2337 }
2338 pHlp->pfnPrintf(pHlp, "Host paging mode: %s\n", psz);
2339 }
2340}
2341
2342
2343/**
2344 * Dump registered MMIO ranges to the log.
2345 *
2346 * @param pVM VM Handle.
2347 * @param pHlp The info helpers.
2348 * @param pszArgs Arguments, ignored.
2349 */
2350static DECLCALLBACK(void) pgmR3PhysInfo(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2351{
2352 NOREF(pszArgs);
2353 pHlp->pfnPrintf(pHlp,
2354 "RAM ranges (pVM=%p)\n"
2355 "%.*s %.*s\n",
2356 pVM,
2357 sizeof(RTGCPHYS) * 4 + 1, "GC Phys Range ",
2358 sizeof(RTHCPTR) * 2, "pvHC ");
2359
2360 for (PPGMRAMRANGE pCur = pVM->pgm.s.pRamRangesR3; pCur; pCur = pCur->pNextR3)
2361 pHlp->pfnPrintf(pHlp,
2362 "%RGp-%RGp %RHv %s\n",
2363 pCur->GCPhys,
2364 pCur->GCPhysLast,
2365 pCur->pvHC,
2366 pCur->pszDesc);
2367}
2368
2369/**
2370 * Dump the page directory to the log.
2371 *
2372 * @param pVM VM Handle.
2373 * @param pHlp The info helpers.
2374 * @param pszArgs Arguments, ignored.
2375 */
2376static DECLCALLBACK(void) pgmR3InfoCr3(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2377{
2378/** @todo fix this! Convert the PGMR3DumpHierarchyHC functions to do guest stuff. */
2379 /* Big pages supported? */
2380 const bool fPSE = !!(CPUMGetGuestCR4(pVM) & X86_CR4_PSE);
2381
2382 /* Global pages supported? */
2383 const bool fPGE = !!(CPUMGetGuestCR4(pVM) & X86_CR4_PGE);
2384
2385 NOREF(pszArgs);
2386
2387 /*
2388 * Get page directory addresses.
2389 */
2390 PX86PD pPDSrc = pVM->pgm.s.pGuestPDHC;
2391 Assert(pPDSrc);
2392 Assert(MMPhysGCPhys2HCVirt(pVM, (RTGCPHYS)(CPUMGetGuestCR3(pVM) & X86_CR3_PAGE_MASK), sizeof(*pPDSrc)) == pPDSrc);
2393
2394 /*
2395 * Iterate the page directory.
2396 */
2397 for (unsigned iPD = 0; iPD < ELEMENTS(pPDSrc->a); iPD++)
2398 {
2399 X86PDE PdeSrc = pPDSrc->a[iPD];
2400 if (PdeSrc.n.u1Present)
2401 {
2402 if (PdeSrc.b.u1Size && fPSE)
2403 {
2404 pHlp->pfnPrintf(pHlp,
2405 "%04X - %VGp P=%d U=%d RW=%d G=%d - BIG\n",
2406 iPD,
2407 PdeSrc.u & X86_PDE_PG_MASK,
2408 PdeSrc.b.u1Present, PdeSrc.b.u1User, PdeSrc.b.u1Write, PdeSrc.b.u1Global && fPGE);
2409 }
2410 else
2411 {
2412 pHlp->pfnPrintf(pHlp,
2413 "%04X - %VGp P=%d U=%d RW=%d [G=%d]\n",
2414 iPD,
2415 PdeSrc.u & X86_PDE4M_PG_MASK,
2416 PdeSrc.n.u1Present, PdeSrc.n.u1User, PdeSrc.n.u1Write, PdeSrc.b.u1Global && fPGE);
2417 }
2418 }
2419 }
2420}
2421
2422
2423/**
2424 * Serivce a VMMCALLHOST_PGM_LOCK call.
2425 *
2426 * @returns VBox status code.
2427 * @param pVM The VM handle.
2428 */
2429PDMR3DECL(int) PGMR3LockCall(PVM pVM)
2430{
2431 int rc = PDMR3CritSectEnterEx(&pVM->pgm.s.CritSect, true /* fHostCall */);
2432 AssertRC(rc);
2433 return rc;
2434}
2435
2436
2437/**
2438 * Converts a PGMMODE value to a PGM_TYPE_* \#define.
2439 *
2440 * @returns PGM_TYPE_*.
2441 * @param pgmMode The mode value to convert.
2442 */
2443DECLINLINE(unsigned) pgmModeToType(PGMMODE pgmMode)
2444{
2445 switch (pgmMode)
2446 {
2447 case PGMMODE_REAL: return PGM_TYPE_REAL;
2448 case PGMMODE_PROTECTED: return PGM_TYPE_PROT;
2449 case PGMMODE_32_BIT: return PGM_TYPE_32BIT;
2450 case PGMMODE_PAE:
2451 case PGMMODE_PAE_NX: return PGM_TYPE_PAE;
2452 case PGMMODE_AMD64:
2453 case PGMMODE_AMD64_NX: return PGM_TYPE_AMD64;
2454 case PGMMODE_NESTED: return PGM_TYPE_NESTED;
2455 default:
2456 AssertFatalMsgFailed(("pgmMode=%d\n", pgmMode));
2457 }
2458}
2459
2460
2461/**
2462 * Gets the index into the paging mode data array of a SHW+GST mode.
2463 *
2464 * @returns PGM::paPagingData index.
2465 * @param uShwType The shadow paging mode type.
2466 * @param uGstType The guest paging mode type.
2467 */
2468DECLINLINE(unsigned) pgmModeDataIndex(unsigned uShwType, unsigned uGstType)
2469{
2470 Assert(uShwType >= PGM_TYPE_32BIT && uShwType <= PGM_TYPE_NESTED);
2471 Assert(uGstType >= PGM_TYPE_REAL && uGstType <= PGM_TYPE_AMD64);
2472 return (uShwType - PGM_TYPE_32BIT) * (PGM_TYPE_AMD64 - PGM_TYPE_REAL + 1)
2473 + (uGstType - PGM_TYPE_REAL);
2474}
2475
2476
2477/**
2478 * Gets the index into the paging mode data array of a SHW+GST mode.
2479 *
2480 * @returns PGM::paPagingData index.
2481 * @param enmShw The shadow paging mode.
2482 * @param enmGst The guest paging mode.
2483 */
2484DECLINLINE(unsigned) pgmModeDataIndexByMode(PGMMODE enmShw, PGMMODE enmGst)
2485{
2486 Assert(enmShw >= PGMMODE_32_BIT && enmShw <= PGMMODE_MAX);
2487 Assert(enmGst > PGMMODE_INVALID && enmGst < PGMMODE_MAX);
2488 return pgmModeDataIndex(pgmModeToType(enmShw), pgmModeToType(enmGst));
2489}
2490
2491
2492/**
2493 * Calculates the max data index.
2494 * @returns The number of entries in the paging data array.
2495 */
2496DECLINLINE(unsigned) pgmModeDataMaxIndex(void)
2497{
2498 return pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_AMD64) + 1;
2499}
2500
2501
2502/**
2503 * Initializes the paging mode data kept in PGM::paModeData.
2504 *
2505 * @param pVM The VM handle.
2506 * @param fResolveGCAndR0 Indicate whether or not GC and Ring-0 symbols can be resolved now.
2507 * This is used early in the init process to avoid trouble with PDM
2508 * not being initialized yet.
2509 */
2510static int pgmR3ModeDataInit(PVM pVM, bool fResolveGCAndR0)
2511{
2512 PPGMMODEDATA pModeData;
2513 int rc;
2514
2515 /*
2516 * Allocate the array on the first call.
2517 */
2518 if (!pVM->pgm.s.paModeData)
2519 {
2520 pVM->pgm.s.paModeData = (PPGMMODEDATA)MMR3HeapAllocZ(pVM, MM_TAG_PGM, sizeof(PGMMODEDATA) * pgmModeDataMaxIndex());
2521 AssertReturn(pVM->pgm.s.paModeData, VERR_NO_MEMORY);
2522 }
2523
2524 /*
2525 * Initialize the array entries.
2526 */
2527 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGM_TYPE_REAL)];
2528 pModeData->uShwType = PGM_TYPE_32BIT;
2529 pModeData->uGstType = PGM_TYPE_REAL;
2530 rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2531 rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2532 rc = PGM_BTH_NAME_32BIT_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2533
2534 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGMMODE_PROTECTED)];
2535 pModeData->uShwType = PGM_TYPE_32BIT;
2536 pModeData->uGstType = PGM_TYPE_PROT;
2537 rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2538 rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2539 rc = PGM_BTH_NAME_32BIT_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2540
2541 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGM_TYPE_32BIT)];
2542 pModeData->uShwType = PGM_TYPE_32BIT;
2543 pModeData->uGstType = PGM_TYPE_32BIT;
2544 rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2545 rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2546 rc = PGM_BTH_NAME_32BIT_32BIT(InitData)(pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2547
2548 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_REAL)];
2549 pModeData->uShwType = PGM_TYPE_PAE;
2550 pModeData->uGstType = PGM_TYPE_REAL;
2551 rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2552 rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2553 rc = PGM_BTH_NAME_PAE_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2554
2555 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_PROT)];
2556 pModeData->uShwType = PGM_TYPE_PAE;
2557 pModeData->uGstType = PGM_TYPE_PROT;
2558 rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2559 rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2560 rc = PGM_BTH_NAME_PAE_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2561
2562 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_32BIT)];
2563 pModeData->uShwType = PGM_TYPE_PAE;
2564 pModeData->uGstType = PGM_TYPE_32BIT;
2565 rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2566 rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2567 rc = PGM_BTH_NAME_PAE_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2568
2569 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_PAE)];
2570 pModeData->uShwType = PGM_TYPE_PAE;
2571 pModeData->uGstType = PGM_TYPE_PAE;
2572 rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2573 rc = PGM_GST_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2574 rc = PGM_BTH_NAME_PAE_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2575
2576 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_AMD64, PGM_TYPE_AMD64)];
2577 pModeData->uShwType = PGM_TYPE_AMD64;
2578 pModeData->uGstType = PGM_TYPE_AMD64;
2579 rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2580 rc = PGM_GST_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2581 rc = PGM_BTH_NAME_AMD64_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2582
2583 /* The nested paging mode. */
2584 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_REAL)];
2585 pModeData->uShwType = PGM_TYPE_NESTED;
2586 pModeData->uGstType = PGM_TYPE_REAL;
2587 rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2588 rc = PGM_BTH_NAME_NESTED_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2589
2590 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGMMODE_PROTECTED)];
2591 pModeData->uShwType = PGM_TYPE_NESTED;
2592 pModeData->uGstType = PGM_TYPE_PROT;
2593 rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2594 rc = PGM_BTH_NAME_NESTED_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2595
2596 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_32BIT)];
2597 pModeData->uShwType = PGM_TYPE_NESTED;
2598 pModeData->uGstType = PGM_TYPE_32BIT;
2599 rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2600 rc = PGM_BTH_NAME_NESTED_32BIT(InitData)(pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2601
2602 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_PAE)];
2603 pModeData->uShwType = PGM_TYPE_NESTED;
2604 pModeData->uGstType = PGM_TYPE_PAE;
2605 rc = PGM_GST_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2606 rc = PGM_BTH_NAME_NESTED_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2607
2608 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_AMD64)];
2609 pModeData->uShwType = PGM_TYPE_NESTED;
2610 pModeData->uGstType = PGM_TYPE_AMD64;
2611 rc = PGM_GST_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2612 rc = PGM_BTH_NAME_NESTED_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2613
2614 /* The shadow part of the nested callback mode depends on the host paging mode (AMD-V only). */
2615 switch(pVM->pgm.s.enmHostMode)
2616 {
2617 case SUPPAGINGMODE_32_BIT:
2618 case SUPPAGINGMODE_32_BIT_GLOBAL:
2619 for (unsigned i=PGM_TYPE_REAL;i<=PGM_TYPE_AMD64;i++)
2620 {
2621 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2622 rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2623 }
2624 break;
2625
2626 case SUPPAGINGMODE_PAE:
2627 case SUPPAGINGMODE_PAE_NX:
2628 case SUPPAGINGMODE_PAE_GLOBAL:
2629 case SUPPAGINGMODE_PAE_GLOBAL_NX:
2630 for (unsigned i=PGM_TYPE_REAL;i<=PGM_TYPE_AMD64;i++)
2631 {
2632 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2633 rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2634 }
2635 break;
2636
2637 case SUPPAGINGMODE_AMD64:
2638 case SUPPAGINGMODE_AMD64_GLOBAL:
2639 case SUPPAGINGMODE_AMD64_NX:
2640 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2641 for (unsigned i=PGM_TYPE_REAL;i<=PGM_TYPE_AMD64;i++)
2642 {
2643 pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2644 rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2645 }
2646 break;
2647 default:
2648 AssertFailed();
2649 break;
2650 }
2651 return VINF_SUCCESS;
2652}
2653
2654
2655/**
2656 * Switch to different (or relocated in the relocate case) mode data.
2657 *
2658 * @param pVM The VM handle.
2659 * @param enmShw The the shadow paging mode.
2660 * @param enmGst The the guest paging mode.
2661 */
2662static void pgmR3ModeDataSwitch(PVM pVM, PGMMODE enmShw, PGMMODE enmGst)
2663{
2664 PPGMMODEDATA pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndexByMode(enmShw, enmGst)];
2665
2666 Assert(pModeData->uGstType == pgmModeToType(enmGst));
2667 Assert(pModeData->uShwType == pgmModeToType(enmShw));
2668
2669 /* shadow */
2670 pVM->pgm.s.pfnR3ShwRelocate = pModeData->pfnR3ShwRelocate;
2671 pVM->pgm.s.pfnR3ShwExit = pModeData->pfnR3ShwExit;
2672 pVM->pgm.s.pfnR3ShwGetPage = pModeData->pfnR3ShwGetPage;
2673 Assert(pVM->pgm.s.pfnR3ShwGetPage);
2674 pVM->pgm.s.pfnR3ShwModifyPage = pModeData->pfnR3ShwModifyPage;
2675
2676 pVM->pgm.s.pfnGCShwGetPage = pModeData->pfnGCShwGetPage;
2677 pVM->pgm.s.pfnGCShwModifyPage = pModeData->pfnGCShwModifyPage;
2678
2679 pVM->pgm.s.pfnR0ShwGetPage = pModeData->pfnR0ShwGetPage;
2680 pVM->pgm.s.pfnR0ShwModifyPage = pModeData->pfnR0ShwModifyPage;
2681
2682
2683 /* guest */
2684 pVM->pgm.s.pfnR3GstRelocate = pModeData->pfnR3GstRelocate;
2685 pVM->pgm.s.pfnR3GstExit = pModeData->pfnR3GstExit;
2686 pVM->pgm.s.pfnR3GstGetPage = pModeData->pfnR3GstGetPage;
2687 Assert(pVM->pgm.s.pfnR3GstGetPage);
2688 pVM->pgm.s.pfnR3GstModifyPage = pModeData->pfnR3GstModifyPage;
2689 pVM->pgm.s.pfnR3GstGetPDE = pModeData->pfnR3GstGetPDE;
2690 pVM->pgm.s.pfnR3GstMonitorCR3 = pModeData->pfnR3GstMonitorCR3;
2691 pVM->pgm.s.pfnR3GstUnmonitorCR3 = pModeData->pfnR3GstUnmonitorCR3;
2692 pVM->pgm.s.pfnR3GstMapCR3 = pModeData->pfnR3GstMapCR3;
2693 pVM->pgm.s.pfnR3GstUnmapCR3 = pModeData->pfnR3GstUnmapCR3;
2694 pVM->pgm.s.pfnR3GstWriteHandlerCR3 = pModeData->pfnR3GstWriteHandlerCR3;
2695 pVM->pgm.s.pszR3GstWriteHandlerCR3 = pModeData->pszR3GstWriteHandlerCR3;
2696 pVM->pgm.s.pfnR3GstPAEWriteHandlerCR3 = pModeData->pfnR3GstPAEWriteHandlerCR3;
2697 pVM->pgm.s.pszR3GstPAEWriteHandlerCR3 = pModeData->pszR3GstPAEWriteHandlerCR3;
2698
2699 pVM->pgm.s.pfnGCGstGetPage = pModeData->pfnGCGstGetPage;
2700 pVM->pgm.s.pfnGCGstModifyPage = pModeData->pfnGCGstModifyPage;
2701 pVM->pgm.s.pfnGCGstGetPDE = pModeData->pfnGCGstGetPDE;
2702 pVM->pgm.s.pfnGCGstMonitorCR3 = pModeData->pfnGCGstMonitorCR3;
2703 pVM->pgm.s.pfnGCGstUnmonitorCR3 = pModeData->pfnGCGstUnmonitorCR3;
2704 pVM->pgm.s.pfnGCGstMapCR3 = pModeData->pfnGCGstMapCR3;
2705 pVM->pgm.s.pfnGCGstUnmapCR3 = pModeData->pfnGCGstUnmapCR3;
2706 pVM->pgm.s.pfnGCGstWriteHandlerCR3 = pModeData->pfnGCGstWriteHandlerCR3;
2707 pVM->pgm.s.pfnGCGstPAEWriteHandlerCR3 = pModeData->pfnGCGstPAEWriteHandlerCR3;
2708
2709 pVM->pgm.s.pfnR0GstGetPage = pModeData->pfnR0GstGetPage;
2710 pVM->pgm.s.pfnR0GstModifyPage = pModeData->pfnR0GstModifyPage;
2711 pVM->pgm.s.pfnR0GstGetPDE = pModeData->pfnR0GstGetPDE;
2712 pVM->pgm.s.pfnR0GstMonitorCR3 = pModeData->pfnR0GstMonitorCR3;
2713 pVM->pgm.s.pfnR0GstUnmonitorCR3 = pModeData->pfnR0GstUnmonitorCR3;
2714 pVM->pgm.s.pfnR0GstMapCR3 = pModeData->pfnR0GstMapCR3;
2715 pVM->pgm.s.pfnR0GstUnmapCR3 = pModeData->pfnR0GstUnmapCR3;
2716 pVM->pgm.s.pfnR0GstWriteHandlerCR3 = pModeData->pfnR0GstWriteHandlerCR3;
2717 pVM->pgm.s.pfnR0GstPAEWriteHandlerCR3 = pModeData->pfnR0GstPAEWriteHandlerCR3;
2718
2719
2720 /* both */
2721 pVM->pgm.s.pfnR3BthRelocate = pModeData->pfnR3BthRelocate;
2722 pVM->pgm.s.pfnR3BthTrap0eHandler = pModeData->pfnR3BthTrap0eHandler;
2723 pVM->pgm.s.pfnR3BthInvalidatePage = pModeData->pfnR3BthInvalidatePage;
2724 pVM->pgm.s.pfnR3BthSyncCR3 = pModeData->pfnR3BthSyncCR3;
2725 Assert(pVM->pgm.s.pfnR3BthSyncCR3);
2726 pVM->pgm.s.pfnR3BthSyncPage = pModeData->pfnR3BthSyncPage;
2727 pVM->pgm.s.pfnR3BthPrefetchPage = pModeData->pfnR3BthPrefetchPage;
2728 pVM->pgm.s.pfnR3BthVerifyAccessSyncPage = pModeData->pfnR3BthVerifyAccessSyncPage;
2729#ifdef VBOX_STRICT
2730 pVM->pgm.s.pfnR3BthAssertCR3 = pModeData->pfnR3BthAssertCR3;
2731#endif
2732
2733 pVM->pgm.s.pfnGCBthTrap0eHandler = pModeData->pfnGCBthTrap0eHandler;
2734 pVM->pgm.s.pfnGCBthInvalidatePage = pModeData->pfnGCBthInvalidatePage;
2735 pVM->pgm.s.pfnGCBthSyncCR3 = pModeData->pfnGCBthSyncCR3;
2736 pVM->pgm.s.pfnGCBthSyncPage = pModeData->pfnGCBthSyncPage;
2737 pVM->pgm.s.pfnGCBthPrefetchPage = pModeData->pfnGCBthPrefetchPage;
2738 pVM->pgm.s.pfnGCBthVerifyAccessSyncPage = pModeData->pfnGCBthVerifyAccessSyncPage;
2739#ifdef VBOX_STRICT
2740 pVM->pgm.s.pfnGCBthAssertCR3 = pModeData->pfnGCBthAssertCR3;
2741#endif
2742
2743 pVM->pgm.s.pfnR0BthTrap0eHandler = pModeData->pfnR0BthTrap0eHandler;
2744 pVM->pgm.s.pfnR0BthInvalidatePage = pModeData->pfnR0BthInvalidatePage;
2745 pVM->pgm.s.pfnR0BthSyncCR3 = pModeData->pfnR0BthSyncCR3;
2746 pVM->pgm.s.pfnR0BthSyncPage = pModeData->pfnR0BthSyncPage;
2747 pVM->pgm.s.pfnR0BthPrefetchPage = pModeData->pfnR0BthPrefetchPage;
2748 pVM->pgm.s.pfnR0BthVerifyAccessSyncPage = pModeData->pfnR0BthVerifyAccessSyncPage;
2749#ifdef VBOX_STRICT
2750 pVM->pgm.s.pfnR0BthAssertCR3 = pModeData->pfnR0BthAssertCR3;
2751#endif
2752}
2753
2754
2755#ifdef DEBUG_bird
2756#include <stdlib.h> /* getenv() remove me! */
2757#endif
2758
2759/**
2760 * Calculates the shadow paging mode.
2761 *
2762 * @returns The shadow paging mode.
2763 * @param pVM VM handle.
2764 * @param enmGuestMode The guest mode.
2765 * @param enmHostMode The host mode.
2766 * @param enmShadowMode The current shadow mode.
2767 * @param penmSwitcher Where to store the switcher to use.
2768 * VMMSWITCHER_INVALID means no change.
2769 */
2770static PGMMODE pgmR3CalcShadowMode(PVM pVM, PGMMODE enmGuestMode, SUPPAGINGMODE enmHostMode, PGMMODE enmShadowMode, VMMSWITCHER *penmSwitcher)
2771{
2772 VMMSWITCHER enmSwitcher = VMMSWITCHER_INVALID;
2773 switch (enmGuestMode)
2774 {
2775 /*
2776 * When switching to real or protected mode we don't change
2777 * anything since it's likely that we'll switch back pretty soon.
2778 *
2779 * During pgmR3InitPaging we'll end up here with PGMMODE_INVALID
2780 * and is supposed to determine which shadow paging and switcher to
2781 * use during init.
2782 */
2783 case PGMMODE_REAL:
2784 case PGMMODE_PROTECTED:
2785 if ( enmShadowMode != PGMMODE_INVALID
2786 && !HWACCMIsEnabled(pVM) /* always switch in hwaccm mode! */)
2787 break; /* (no change) */
2788
2789 switch (enmHostMode)
2790 {
2791 case SUPPAGINGMODE_32_BIT:
2792 case SUPPAGINGMODE_32_BIT_GLOBAL:
2793 enmShadowMode = PGMMODE_32_BIT;
2794 enmSwitcher = VMMSWITCHER_32_TO_32;
2795 break;
2796
2797 case SUPPAGINGMODE_PAE:
2798 case SUPPAGINGMODE_PAE_NX:
2799 case SUPPAGINGMODE_PAE_GLOBAL:
2800 case SUPPAGINGMODE_PAE_GLOBAL_NX:
2801 enmShadowMode = PGMMODE_PAE;
2802 enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2803#ifdef DEBUG_bird
2804if (getenv("VBOX_32BIT"))
2805{
2806 enmShadowMode = PGMMODE_32_BIT;
2807 enmSwitcher = VMMSWITCHER_PAE_TO_32;
2808}
2809#endif
2810 break;
2811
2812 case SUPPAGINGMODE_AMD64:
2813 case SUPPAGINGMODE_AMD64_GLOBAL:
2814 case SUPPAGINGMODE_AMD64_NX:
2815 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2816 enmShadowMode = PGMMODE_PAE;
2817 enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2818 break;
2819
2820 default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2821 }
2822 break;
2823
2824 case PGMMODE_32_BIT:
2825 switch (enmHostMode)
2826 {
2827 case SUPPAGINGMODE_32_BIT:
2828 case SUPPAGINGMODE_32_BIT_GLOBAL:
2829 enmShadowMode = PGMMODE_32_BIT;
2830 enmSwitcher = VMMSWITCHER_32_TO_32;
2831 break;
2832
2833 case SUPPAGINGMODE_PAE:
2834 case SUPPAGINGMODE_PAE_NX:
2835 case SUPPAGINGMODE_PAE_GLOBAL:
2836 case SUPPAGINGMODE_PAE_GLOBAL_NX:
2837 enmShadowMode = PGMMODE_PAE;
2838 enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2839#ifdef DEBUG_bird
2840if (getenv("VBOX_32BIT"))
2841{
2842 enmShadowMode = PGMMODE_32_BIT;
2843 enmSwitcher = VMMSWITCHER_PAE_TO_32;
2844}
2845#endif
2846 break;
2847
2848 case SUPPAGINGMODE_AMD64:
2849 case SUPPAGINGMODE_AMD64_GLOBAL:
2850 case SUPPAGINGMODE_AMD64_NX:
2851 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2852 enmShadowMode = PGMMODE_PAE;
2853 enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2854 break;
2855
2856 default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2857 }
2858 break;
2859
2860 case PGMMODE_PAE:
2861 case PGMMODE_PAE_NX: /** @todo This might require more switchers and guest+both modes. */
2862 switch (enmHostMode)
2863 {
2864 case SUPPAGINGMODE_32_BIT:
2865 case SUPPAGINGMODE_32_BIT_GLOBAL:
2866 enmShadowMode = PGMMODE_PAE;
2867 enmSwitcher = VMMSWITCHER_32_TO_PAE;
2868 break;
2869
2870 case SUPPAGINGMODE_PAE:
2871 case SUPPAGINGMODE_PAE_NX:
2872 case SUPPAGINGMODE_PAE_GLOBAL:
2873 case SUPPAGINGMODE_PAE_GLOBAL_NX:
2874 enmShadowMode = PGMMODE_PAE;
2875 enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2876 break;
2877
2878 case SUPPAGINGMODE_AMD64:
2879 case SUPPAGINGMODE_AMD64_GLOBAL:
2880 case SUPPAGINGMODE_AMD64_NX:
2881 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2882 enmShadowMode = PGMMODE_PAE;
2883 enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2884 break;
2885
2886 default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2887 }
2888 break;
2889
2890 case PGMMODE_AMD64:
2891 case PGMMODE_AMD64_NX:
2892 switch (enmHostMode)
2893 {
2894 case SUPPAGINGMODE_32_BIT:
2895 case SUPPAGINGMODE_32_BIT_GLOBAL:
2896 enmShadowMode = PGMMODE_PAE;
2897 enmSwitcher = VMMSWITCHER_32_TO_AMD64;
2898 break;
2899
2900 case SUPPAGINGMODE_PAE:
2901 case SUPPAGINGMODE_PAE_NX:
2902 case SUPPAGINGMODE_PAE_GLOBAL:
2903 case SUPPAGINGMODE_PAE_GLOBAL_NX:
2904 enmShadowMode = PGMMODE_PAE;
2905 enmSwitcher = VMMSWITCHER_PAE_TO_AMD64;
2906 break;
2907
2908 case SUPPAGINGMODE_AMD64:
2909 case SUPPAGINGMODE_AMD64_GLOBAL:
2910 case SUPPAGINGMODE_AMD64_NX:
2911 case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2912 enmShadowMode = PGMMODE_AMD64;
2913 enmSwitcher = VMMSWITCHER_AMD64_TO_AMD64;
2914 break;
2915
2916 default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2917 }
2918 break;
2919
2920
2921 default:
2922 AssertReleaseMsgFailed(("enmGuestMode=%d\n", enmGuestMode));
2923 return PGMMODE_INVALID;
2924 }
2925 /* Override the shadow mode is nested paging is active. */
2926 if (HWACCMIsNestedPagingActive(pVM))
2927 enmShadowMode = PGMMODE_NESTED;
2928
2929 *penmSwitcher = enmSwitcher;
2930 return enmShadowMode;
2931}
2932
2933#ifdef LOG_ENABLED
2934/**
2935 * Return the string corresponding to the guest mode
2936 *
2937 * @returns string
2938 * @param enmGuestMode The guest mode.
2939 */
2940const char *pgmr3GuestModeString(PGMMODE enmGuestMode)
2941{
2942 switch(enmGuestMode)
2943 {
2944 case PGMMODE_REAL:
2945 return "Real mode";
2946
2947 case PGMMODE_PROTECTED:
2948 return "Protected mode without paging";
2949
2950 case PGMMODE_32_BIT:
2951 return "32 bits protected mode";
2952
2953 case PGMMODE_PAE:
2954 return "PAE";
2955
2956 case PGMMODE_PAE_NX:
2957 return "PAE + NX";
2958
2959 case PGMMODE_AMD64:
2960 return "AMD64";
2961
2962 case PGMMODE_AMD64_NX:
2963 return "AMD64 + NX";
2964
2965 default:
2966 return "Unknown";
2967 }
2968}
2969#endif
2970
2971/**
2972 * Performs the actual mode change.
2973 * This is called by PGMChangeMode and pgmR3InitPaging().
2974 *
2975 * @returns VBox status code.
2976 * @param pVM VM handle.
2977 * @param enmGuestMode The new guest mode. This is assumed to be different from
2978 * the current mode.
2979 */
2980PGMR3DECL(int) PGMR3ChangeMode(PVM pVM, PGMMODE enmGuestMode)
2981{
2982 LogFlow(("PGMR3ChangeMode: Guest mode: %s -> %s\n", pgmr3GuestModeString(pVM->pgm.s.enmGuestMode), pgmr3GuestModeString(enmGuestMode)));
2983 STAM_REL_COUNTER_INC(&pVM->pgm.s.cGuestModeChanges);
2984
2985 /*
2986 * Calc the shadow mode and switcher.
2987 */
2988 VMMSWITCHER enmSwitcher;
2989 PGMMODE enmShadowMode = pgmR3CalcShadowMode(pVM, enmGuestMode, pVM->pgm.s.enmHostMode, pVM->pgm.s.enmShadowMode, &enmSwitcher);
2990 if (enmSwitcher != VMMSWITCHER_INVALID)
2991 {
2992 /*
2993 * Select new switcher.
2994 */
2995 int rc = VMMR3SelectSwitcher(pVM, enmSwitcher);
2996 if (VBOX_FAILURE(rc))
2997 {
2998 AssertReleaseMsgFailed(("VMMR3SelectSwitcher(%d) -> %Vrc\n", enmSwitcher, rc));
2999 return rc;
3000 }
3001 }
3002
3003 /*
3004 * Exit old mode(s).
3005 */
3006 /* shadow */
3007 if (enmShadowMode != pVM->pgm.s.enmShadowMode)
3008 {
3009 LogFlow(("PGMR3ChangeMode: Shadow mode: %d -> %d\n", pVM->pgm.s.enmShadowMode, enmShadowMode));
3010 if (PGM_SHW_PFN(Exit, pVM))
3011 {
3012 int rc = PGM_SHW_PFN(Exit, pVM)(pVM);
3013 if (VBOX_FAILURE(rc))
3014 {
3015 AssertMsgFailed(("Exit failed for shadow mode %d: %Vrc\n", pVM->pgm.s.enmShadowMode, rc));
3016 return rc;
3017 }
3018 }
3019
3020 }
3021
3022 /* guest */
3023 if (PGM_GST_PFN(Exit, pVM))
3024 {
3025 int rc = PGM_GST_PFN(Exit, pVM)(pVM);
3026 if (VBOX_FAILURE(rc))
3027 {
3028 AssertMsgFailed(("Exit failed for guest mode %d: %Vrc\n", pVM->pgm.s.enmGuestMode, rc));
3029 return rc;
3030 }
3031 }
3032
3033 /*
3034 * Load new paging mode data.
3035 */
3036 pgmR3ModeDataSwitch(pVM, enmShadowMode, enmGuestMode);
3037
3038 /*
3039 * Enter new shadow mode (if changed).
3040 */
3041 if (enmShadowMode != pVM->pgm.s.enmShadowMode)
3042 {
3043 int rc;
3044 pVM->pgm.s.enmShadowMode = enmShadowMode;
3045 switch (enmShadowMode)
3046 {
3047 case PGMMODE_32_BIT:
3048 rc = PGM_SHW_NAME_32BIT(Enter)(pVM);
3049 break;
3050 case PGMMODE_PAE:
3051 case PGMMODE_PAE_NX:
3052 rc = PGM_SHW_NAME_PAE(Enter)(pVM);
3053 break;
3054 case PGMMODE_AMD64:
3055 case PGMMODE_AMD64_NX:
3056 rc = PGM_SHW_NAME_AMD64(Enter)(pVM);
3057 break;
3058 case PGMMODE_NESTED:
3059 rc = PGM_SHW_NAME_NESTED(Enter)(pVM);
3060 break;
3061 case PGMMODE_REAL:
3062 case PGMMODE_PROTECTED:
3063 default:
3064 AssertReleaseMsgFailed(("enmShadowMode=%d\n", enmShadowMode));
3065 return VERR_INTERNAL_ERROR;
3066 }
3067 if (VBOX_FAILURE(rc))
3068 {
3069 AssertReleaseMsgFailed(("Entering enmShadowMode=%d failed: %Vrc\n", enmShadowMode, rc));
3070 pVM->pgm.s.enmShadowMode = PGMMODE_INVALID;
3071 return rc;
3072 }
3073 }
3074
3075 /*
3076 * Enter the new guest and shadow+guest modes.
3077 */
3078 int rc = -1;
3079 int rc2 = -1;
3080 RTGCPHYS GCPhysCR3 = NIL_RTGCPHYS;
3081 pVM->pgm.s.enmGuestMode = enmGuestMode;
3082 switch (enmGuestMode)
3083 {
3084 case PGMMODE_REAL:
3085 rc = PGM_GST_NAME_REAL(Enter)(pVM, NIL_RTGCPHYS);
3086 switch (pVM->pgm.s.enmShadowMode)
3087 {
3088 case PGMMODE_32_BIT:
3089 rc2 = PGM_BTH_NAME_32BIT_REAL(Enter)(pVM, NIL_RTGCPHYS);
3090 break;
3091 case PGMMODE_PAE:
3092 case PGMMODE_PAE_NX:
3093 rc2 = PGM_BTH_NAME_PAE_REAL(Enter)(pVM, NIL_RTGCPHYS);
3094 break;
3095 case PGMMODE_NESTED:
3096 rc2 = PGM_BTH_NAME_NESTED_REAL(Enter)(pVM, NIL_RTGCPHYS);
3097 break;
3098 case PGMMODE_AMD64:
3099 case PGMMODE_AMD64_NX:
3100 AssertMsgFailed(("Should use PAE shadow mode!\n"));
3101 default: AssertFailed(); break;
3102 }
3103 break;
3104
3105 case PGMMODE_PROTECTED:
3106 rc = PGM_GST_NAME_PROT(Enter)(pVM, NIL_RTGCPHYS);
3107 switch (pVM->pgm.s.enmShadowMode)
3108 {
3109 case PGMMODE_32_BIT:
3110 rc2 = PGM_BTH_NAME_32BIT_PROT(Enter)(pVM, NIL_RTGCPHYS);
3111 break;
3112 case PGMMODE_PAE:
3113 case PGMMODE_PAE_NX:
3114 rc2 = PGM_BTH_NAME_PAE_PROT(Enter)(pVM, NIL_RTGCPHYS);
3115 break;
3116 case PGMMODE_NESTED:
3117 rc2 = PGM_BTH_NAME_NESTED_PROT(Enter)(pVM, NIL_RTGCPHYS);
3118 break;
3119 case PGMMODE_AMD64:
3120 case PGMMODE_AMD64_NX:
3121 AssertMsgFailed(("Should use PAE shadow mode!\n"));
3122 default: AssertFailed(); break;
3123 }
3124 break;
3125
3126 case PGMMODE_32_BIT:
3127 GCPhysCR3 = CPUMGetGuestCR3(pVM) & X86_CR3_PAGE_MASK;
3128 rc = PGM_GST_NAME_32BIT(Enter)(pVM, GCPhysCR3);
3129 switch (pVM->pgm.s.enmShadowMode)
3130 {
3131 case PGMMODE_32_BIT:
3132 rc2 = PGM_BTH_NAME_32BIT_32BIT(Enter)(pVM, GCPhysCR3);
3133 break;
3134 case PGMMODE_PAE:
3135 case PGMMODE_PAE_NX:
3136 rc2 = PGM_BTH_NAME_PAE_32BIT(Enter)(pVM, GCPhysCR3);
3137 break;
3138 case PGMMODE_NESTED:
3139 rc2 = PGM_BTH_NAME_NESTED_32BIT(Enter)(pVM, GCPhysCR3);
3140 break;
3141 case PGMMODE_AMD64:
3142 case PGMMODE_AMD64_NX:
3143 AssertMsgFailed(("Should use PAE shadow mode!\n"));
3144 default: AssertFailed(); break;
3145 }
3146 break;
3147
3148 //case PGMMODE_PAE_NX:
3149 case PGMMODE_PAE:
3150 {
3151 uint32_t u32Dummy, u32Features;
3152
3153 CPUMGetGuestCpuId(pVM, 1, &u32Dummy, &u32Dummy, &u32Dummy, &u32Features);
3154 if (!(u32Features & X86_CPUID_FEATURE_EDX_PAE))
3155 {
3156 /* Pause first, then inform Main. */
3157 rc = VMR3SuspendNoSave(pVM);
3158 AssertRC(rc);
3159
3160 VMSetRuntimeError(pVM, true, "PAEmode",
3161 N_("The guest is trying to switch to the PAE mode which is currently disabled by default in VirtualBox. Experimental PAE support can be enabled using the -pae option with VBoxManage."));
3162 /* we must return TRUE here otherwise the recompiler will assert */
3163 return VINF_SUCCESS;
3164 }
3165 GCPhysCR3 = CPUMGetGuestCR3(pVM) & X86_CR3_PAE_PAGE_MASK;
3166 rc = PGM_GST_NAME_PAE(Enter)(pVM, GCPhysCR3);
3167 switch (pVM->pgm.s.enmShadowMode)
3168 {
3169 case PGMMODE_PAE:
3170 case PGMMODE_PAE_NX:
3171 rc2 = PGM_BTH_NAME_PAE_PAE(Enter)(pVM, GCPhysCR3);
3172 break;
3173 case PGMMODE_NESTED:
3174 rc2 = PGM_BTH_NAME_NESTED_PAE(Enter)(pVM, GCPhysCR3);
3175 break;
3176 case PGMMODE_32_BIT:
3177 case PGMMODE_AMD64:
3178 case PGMMODE_AMD64_NX:
3179 AssertMsgFailed(("Should use PAE shadow mode!\n"));
3180 default: AssertFailed(); break;
3181 }
3182 break;
3183 }
3184
3185 case PGMMODE_AMD64_NX:
3186 case PGMMODE_AMD64:
3187 GCPhysCR3 = CPUMGetGuestCR3(pVM) & 0xfffffffffffff000ULL; /** @todo define this mask! */
3188 rc = PGM_GST_NAME_AMD64(Enter)(pVM, GCPhysCR3);
3189 switch (pVM->pgm.s.enmShadowMode)
3190 {
3191 case PGMMODE_AMD64:
3192 case PGMMODE_AMD64_NX:
3193 rc2 = PGM_BTH_NAME_AMD64_AMD64(Enter)(pVM, GCPhysCR3);
3194 break;
3195 case PGMMODE_NESTED:
3196 rc2 = PGM_BTH_NAME_NESTED_AMD64(Enter)(pVM, GCPhysCR3);
3197 break;
3198 case PGMMODE_32_BIT:
3199 case PGMMODE_PAE:
3200 case PGMMODE_PAE_NX:
3201 AssertMsgFailed(("Should use AMD64 shadow mode!\n"));
3202 default: AssertFailed(); break;
3203 }
3204 break;
3205
3206 default:
3207 AssertReleaseMsgFailed(("enmGuestMode=%d\n", enmGuestMode));
3208 rc = VERR_NOT_IMPLEMENTED;
3209 break;
3210 }
3211
3212 /* status codes. */
3213 AssertRC(rc);
3214 AssertRC(rc2);
3215 if (VBOX_SUCCESS(rc))
3216 {
3217 rc = rc2;
3218 if (VBOX_SUCCESS(rc)) /* no informational status codes. */
3219 rc = VINF_SUCCESS;
3220 }
3221
3222 /*
3223 * Notify SELM so it can update the TSSes with correct CR3s.
3224 */
3225 SELMR3PagingModeChanged(pVM);
3226
3227 /* Notify HWACCM as well. */
3228 HWACCMR3PagingModeChanged(pVM, pVM->pgm.s.enmShadowMode);
3229 return rc;
3230}
3231
3232
3233/**
3234 * Dumps a PAE shadow page table.
3235 *
3236 * @returns VBox status code (VINF_SUCCESS).
3237 * @param pVM The VM handle.
3238 * @param pPT Pointer to the page table.
3239 * @param u64Address The virtual address of the page table starts.
3240 * @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3241 * @param cMaxDepth The maxium depth.
3242 * @param pHlp Pointer to the output functions.
3243 */
3244static int pgmR3DumpHierarchyHCPaePT(PVM pVM, PX86PTPAE pPT, uint64_t u64Address, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3245{
3246 for (unsigned i = 0; i < ELEMENTS(pPT->a); i++)
3247 {
3248 X86PTEPAE Pte = pPT->a[i];
3249 if (Pte.n.u1Present)
3250 {
3251 pHlp->pfnPrintf(pHlp,
3252 fLongMode /*P R S A D G WT CD AT NX 4M a p ? */
3253 ? "%016llx 3 | P %c %c %c %c %c %s %s %s %s 4K %c%c%c %016llx\n"
3254 : "%08llx 2 | P %c %c %c %c %c %s %s %s %s 4K %c%c%c %016llx\n",
3255 u64Address + ((uint64_t)i << X86_PT_PAE_SHIFT),
3256 Pte.n.u1Write ? 'W' : 'R',
3257 Pte.n.u1User ? 'U' : 'S',
3258 Pte.n.u1Accessed ? 'A' : '-',
3259 Pte.n.u1Dirty ? 'D' : '-',
3260 Pte.n.u1Global ? 'G' : '-',
3261 Pte.n.u1WriteThru ? "WT" : "--",
3262 Pte.n.u1CacheDisable? "CD" : "--",
3263 Pte.n.u1PAT ? "AT" : "--",
3264 Pte.n.u1NoExecute ? "NX" : "--",
3265 Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3266 Pte.u & RT_BIT(10) ? '1' : '0',
3267 Pte.u & PGM_PTFLAGS_CSAM_VALIDATED? 'v' : '-',
3268 Pte.u & X86_PTE_PAE_PG_MASK);
3269 }
3270 }
3271 return VINF_SUCCESS;
3272}
3273
3274
3275/**
3276 * Dumps a PAE shadow page directory table.
3277 *
3278 * @returns VBox status code (VINF_SUCCESS).
3279 * @param pVM The VM handle.
3280 * @param HCPhys The physical address of the page directory table.
3281 * @param u64Address The virtual address of the page table starts.
3282 * @param cr4 The CR4, PSE is currently used.
3283 * @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3284 * @param cMaxDepth The maxium depth.
3285 * @param pHlp Pointer to the output functions.
3286 */
3287static int pgmR3DumpHierarchyHCPaePD(PVM pVM, RTHCPHYS HCPhys, uint64_t u64Address, uint32_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3288{
3289 PX86PDPAE pPD = (PX86PDPAE)MMPagePhys2Page(pVM, HCPhys);
3290 if (!pPD)
3291 {
3292 pHlp->pfnPrintf(pHlp, "%0*llx error! Page directory at HCPhys=%#VHp was not found in the page pool!\n",
3293 fLongMode ? 16 : 8, u64Address, HCPhys);
3294 return VERR_INVALID_PARAMETER;
3295 }
3296 const bool fBigPagesSupported = fLongMode || !!(cr4 & X86_CR4_PSE);
3297
3298 int rc = VINF_SUCCESS;
3299 for (unsigned i = 0; i < ELEMENTS(pPD->a); i++)
3300 {
3301 X86PDEPAE Pde = pPD->a[i];
3302 if (Pde.n.u1Present)
3303 {
3304 if (fBigPagesSupported && Pde.b.u1Size)
3305 pHlp->pfnPrintf(pHlp,
3306 fLongMode /*P R S A D G WT CD AT NX 4M a p ? */
3307 ? "%016llx 2 | P %c %c %c %c %c %s %s %s %s 4M %c%c%c %016llx\n"
3308 : "%08llx 1 | P %c %c %c %c %c %s %s %s %s 4M %c%c%c %016llx\n",
3309 u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT),
3310 Pde.b.u1Write ? 'W' : 'R',
3311 Pde.b.u1User ? 'U' : 'S',
3312 Pde.b.u1Accessed ? 'A' : '-',
3313 Pde.b.u1Dirty ? 'D' : '-',
3314 Pde.b.u1Global ? 'G' : '-',
3315 Pde.b.u1WriteThru ? "WT" : "--",
3316 Pde.b.u1CacheDisable? "CD" : "--",
3317 Pde.b.u1PAT ? "AT" : "--",
3318 Pde.b.u1NoExecute ? "NX" : "--",
3319 Pde.u & RT_BIT_64(9) ? '1' : '0',
3320 Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3321 Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3322 Pde.u & X86_PDE_PAE_PG_MASK);
3323 else
3324 {
3325 pHlp->pfnPrintf(pHlp,
3326 fLongMode /*P R S A D G WT CD AT NX 4M a p ? */
3327 ? "%016llx 2 | P %c %c %c %c %c %s %s .. %s 4K %c%c%c %016llx\n"
3328 : "%08llx 1 | P %c %c %c %c %c %s %s .. %s 4K %c%c%c %016llx\n",
3329 u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT),
3330 Pde.n.u1Write ? 'W' : 'R',
3331 Pde.n.u1User ? 'U' : 'S',
3332 Pde.n.u1Accessed ? 'A' : '-',
3333 Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3334 Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3335 Pde.n.u1WriteThru ? "WT" : "--",
3336 Pde.n.u1CacheDisable? "CD" : "--",
3337 Pde.n.u1NoExecute ? "NX" : "--",
3338 Pde.u & RT_BIT_64(9) ? '1' : '0',
3339 Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3340 Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3341 Pde.u & X86_PDE_PAE_PG_MASK);
3342 if (cMaxDepth >= 1)
3343 {
3344 /** @todo what about using the page pool for mapping PTs? */
3345 uint64_t u64AddressPT = u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT);
3346 RTHCPHYS HCPhysPT = Pde.u & X86_PDE_PAE_PG_MASK;
3347 PX86PTPAE pPT = NULL;
3348 if (!(Pde.u & PGM_PDFLAGS_MAPPING))
3349 pPT = (PX86PTPAE)MMPagePhys2Page(pVM, HCPhysPT);
3350 else
3351 {
3352 for (PPGMMAPPING pMap = pVM->pgm.s.pMappingsR3; pMap; pMap = pMap->pNextR3)
3353 {
3354 uint64_t off = u64AddressPT - pMap->GCPtr;
3355 if (off < pMap->cb)
3356 {
3357 const int iPDE = (uint32_t)(off >> X86_PD_SHIFT);
3358 const int iSub = (int)((off >> X86_PD_PAE_SHIFT) & 1); /* MSC is a pain sometimes */
3359 if ((iSub ? pMap->aPTs[iPDE].HCPhysPaePT1 : pMap->aPTs[iPDE].HCPhysPaePT0) != HCPhysPT)
3360 pHlp->pfnPrintf(pHlp, "%0*llx error! Mapping error! PT %d has HCPhysPT=%VHp not %VHp is in the PD.\n",
3361 fLongMode ? 16 : 8, u64AddressPT, iPDE,
3362 iSub ? pMap->aPTs[iPDE].HCPhysPaePT1 : pMap->aPTs[iPDE].HCPhysPaePT0, HCPhysPT);
3363 pPT = &pMap->aPTs[iPDE].paPaePTsR3[iSub];
3364 }
3365 }
3366 }
3367 int rc2 = VERR_INVALID_PARAMETER;
3368 if (pPT)
3369 rc2 = pgmR3DumpHierarchyHCPaePT(pVM, pPT, u64AddressPT, fLongMode, cMaxDepth - 1, pHlp);
3370 else
3371 pHlp->pfnPrintf(pHlp, "%0*llx error! Page table at HCPhys=%#VHp was not found in the page pool!\n",
3372 fLongMode ? 16 : 8, u64AddressPT, HCPhysPT);
3373 if (rc2 < rc && VBOX_SUCCESS(rc))
3374 rc = rc2;
3375 }
3376 }
3377 }
3378 }
3379 return rc;
3380}
3381
3382
3383/**
3384 * Dumps a PAE shadow page directory pointer table.
3385 *
3386 * @returns VBox status code (VINF_SUCCESS).
3387 * @param pVM The VM handle.
3388 * @param HCPhys The physical address of the page directory pointer table.
3389 * @param u64Address The virtual address of the page table starts.
3390 * @param cr4 The CR4, PSE is currently used.
3391 * @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3392 * @param cMaxDepth The maxium depth.
3393 * @param pHlp Pointer to the output functions.
3394 */
3395static int pgmR3DumpHierarchyHCPaePDPT(PVM pVM, RTHCPHYS HCPhys, uint64_t u64Address, uint32_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3396{
3397 PX86PDPT pPDPT = (PX86PDPT)MMPagePhys2Page(pVM, HCPhys);
3398 if (!pPDPT)
3399 {
3400 pHlp->pfnPrintf(pHlp, "%0*llx error! Page directory pointer table at HCPhys=%#VHp was not found in the page pool!\n",
3401 fLongMode ? 16 : 8, u64Address, HCPhys);
3402 return VERR_INVALID_PARAMETER;
3403 }
3404
3405 int rc = VINF_SUCCESS;
3406 const unsigned c = fLongMode ? ELEMENTS(pPDPT->a) : X86_PG_PAE_PDPE_ENTRIES;
3407 for (unsigned i = 0; i < c; i++)
3408 {
3409 X86PDPE Pdpe = pPDPT->a[i];
3410 if (Pdpe.n.u1Present)
3411 {
3412 if (fLongMode)
3413 pHlp->pfnPrintf(pHlp, /*P R S A D G WT CD AT NX 4M a p ? */
3414 "%016llx 1 | P %c %c %c %c %c %s %s %s %s .. %c%c%c %016llx\n",
3415 u64Address + ((uint64_t)i << X86_PDPT_SHIFT),
3416 Pdpe.lm.u1Write ? 'W' : 'R',
3417 Pdpe.lm.u1User ? 'U' : 'S',
3418 Pdpe.lm.u1Accessed ? 'A' : '-',
3419 Pdpe.lm.u3Reserved & 1? '?' : '.', /* ignored */
3420 Pdpe.lm.u3Reserved & 4? '!' : '.', /* mbz */
3421 Pdpe.lm.u1WriteThru ? "WT" : "--",
3422 Pdpe.lm.u1CacheDisable? "CD" : "--",
3423 Pdpe.lm.u3Reserved & 2? "!" : "..",/* mbz */
3424 Pdpe.lm.u1NoExecute ? "NX" : "--",
3425 Pdpe.u & RT_BIT(9) ? '1' : '0',
3426 Pdpe.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3427 Pdpe.u & RT_BIT(11) ? '1' : '0',
3428 Pdpe.u & X86_PDPE_PG_MASK);
3429 else
3430 pHlp->pfnPrintf(pHlp, /*P G WT CD AT NX 4M a p ? */
3431 "%08x 0 | P %c %s %s %s %s .. %c%c%c %016llx\n",
3432 i << X86_PDPT_SHIFT,
3433 Pdpe.n.u4Reserved & 1? '!' : '.', /* mbz */
3434 Pdpe.n.u4Reserved & 4? '!' : '.', /* mbz */
3435 Pdpe.n.u1WriteThru ? "WT" : "--",
3436 Pdpe.n.u1CacheDisable? "CD" : "--",
3437 Pdpe.n.u4Reserved & 2? "!" : "..",/* mbz */
3438 Pdpe.u & RT_BIT(9) ? '1' : '0',
3439 Pdpe.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3440 Pdpe.u & RT_BIT(11) ? '1' : '0',
3441 Pdpe.u & X86_PDPE_PG_MASK);
3442 if (cMaxDepth >= 1)
3443 {
3444 int rc2 = pgmR3DumpHierarchyHCPaePD(pVM, Pdpe.u & X86_PDPE_PG_MASK, u64Address + ((uint64_t)i << X86_PDPT_SHIFT),
3445 cr4, fLongMode, cMaxDepth - 1, pHlp);
3446 if (rc2 < rc && VBOX_SUCCESS(rc))
3447 rc = rc2;
3448 }
3449 }
3450 }
3451 return rc;
3452}
3453
3454
3455/**
3456 * Dumps a 32-bit shadow page table.
3457 *
3458 * @returns VBox status code (VINF_SUCCESS).
3459 * @param pVM The VM handle.
3460 * @param HCPhys The physical address of the table.
3461 * @param cr4 The CR4, PSE is currently used.
3462 * @param cMaxDepth The maxium depth.
3463 * @param pHlp Pointer to the output functions.
3464 */
3465static int pgmR3DumpHierarchyHcPaePML4(PVM pVM, RTHCPHYS HCPhys, uint32_t cr4, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3466{
3467 PX86PML4 pPML4 = (PX86PML4)MMPagePhys2Page(pVM, HCPhys);
3468 if (!pPML4)
3469 {
3470 pHlp->pfnPrintf(pHlp, "Page map level 4 at HCPhys=%#VHp was not found in the page pool!\n", HCPhys);
3471 return VERR_INVALID_PARAMETER;
3472 }
3473
3474 int rc = VINF_SUCCESS;
3475 for (unsigned i = 0; i < ELEMENTS(pPML4->a); i++)
3476 {
3477 X86PML4E Pml4e = pPML4->a[i];
3478 if (Pml4e.n.u1Present)
3479 {
3480 uint64_t u64Address = ((uint64_t)i << X86_PML4_SHIFT) | (((uint64_t)i >> (X86_PML4_SHIFT - X86_PDPT_SHIFT - 1)) * 0xffff000000000000ULL);
3481 pHlp->pfnPrintf(pHlp, /*P R S A D G WT CD AT NX 4M a p ? */
3482 "%016llx 0 | P %c %c %c %c %c %s %s %s %s .. %c%c%c %016llx\n",
3483 u64Address,
3484 Pml4e.n.u1Write ? 'W' : 'R',
3485 Pml4e.n.u1User ? 'U' : 'S',
3486 Pml4e.n.u1Accessed ? 'A' : '-',
3487 Pml4e.n.u3Reserved & 1? '?' : '.', /* ignored */
3488 Pml4e.n.u3Reserved & 4? '!' : '.', /* mbz */
3489 Pml4e.n.u1WriteThru ? "WT" : "--",
3490 Pml4e.n.u1CacheDisable? "CD" : "--",
3491 Pml4e.n.u3Reserved & 2? "!" : "..",/* mbz */
3492 Pml4e.n.u1NoExecute ? "NX" : "--",
3493 Pml4e.u & RT_BIT(9) ? '1' : '0',
3494 Pml4e.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3495 Pml4e.u & RT_BIT(11) ? '1' : '0',
3496 Pml4e.u & X86_PML4E_PG_MASK);
3497
3498 if (cMaxDepth >= 1)
3499 {
3500 int rc2 = pgmR3DumpHierarchyHCPaePDPT(pVM, Pml4e.u & X86_PML4E_PG_MASK, u64Address, cr4, true, cMaxDepth - 1, pHlp);
3501 if (rc2 < rc && VBOX_SUCCESS(rc))
3502 rc = rc2;
3503 }
3504 }
3505 }
3506 return rc;
3507}
3508
3509
3510/**
3511 * Dumps a 32-bit shadow page table.
3512 *
3513 * @returns VBox status code (VINF_SUCCESS).
3514 * @param pVM The VM handle.
3515 * @param pPT Pointer to the page table.
3516 * @param u32Address The virtual address this table starts at.
3517 * @param pHlp Pointer to the output functions.
3518 */
3519int pgmR3DumpHierarchyHC32BitPT(PVM pVM, PX86PT pPT, uint32_t u32Address, PCDBGFINFOHLP pHlp)
3520{
3521 for (unsigned i = 0; i < ELEMENTS(pPT->a); i++)
3522 {
3523 X86PTE Pte = pPT->a[i];
3524 if (Pte.n.u1Present)
3525 {
3526 pHlp->pfnPrintf(pHlp, /*P R S A D G WT CD AT NX 4M a m d */
3527 "%08x 1 | P %c %c %c %c %c %s %s %s .. 4K %c%c%c %08x\n",
3528 u32Address + (i << X86_PT_SHIFT),
3529 Pte.n.u1Write ? 'W' : 'R',
3530 Pte.n.u1User ? 'U' : 'S',
3531 Pte.n.u1Accessed ? 'A' : '-',
3532 Pte.n.u1Dirty ? 'D' : '-',
3533 Pte.n.u1Global ? 'G' : '-',
3534 Pte.n.u1WriteThru ? "WT" : "--",
3535 Pte.n.u1CacheDisable? "CD" : "--",
3536 Pte.n.u1PAT ? "AT" : "--",
3537 Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3538 Pte.u & RT_BIT(10) ? '1' : '0',
3539 Pte.u & PGM_PTFLAGS_CSAM_VALIDATED ? 'v' : '-',
3540 Pte.u & X86_PDE_PG_MASK);
3541 }
3542 }
3543 return VINF_SUCCESS;
3544}
3545
3546
3547/**
3548 * Dumps a 32-bit shadow page directory and page tables.
3549 *
3550 * @returns VBox status code (VINF_SUCCESS).
3551 * @param pVM The VM handle.
3552 * @param cr3 The root of the hierarchy.
3553 * @param cr4 The CR4, PSE is currently used.
3554 * @param cMaxDepth How deep into the hierarchy the dumper should go.
3555 * @param pHlp Pointer to the output functions.
3556 */
3557int pgmR3DumpHierarchyHC32BitPD(PVM pVM, uint32_t cr3, uint32_t cr4, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3558{
3559 PX86PD pPD = (PX86PD)MMPagePhys2Page(pVM, cr3 & X86_CR3_PAGE_MASK);
3560 if (!pPD)
3561 {
3562 pHlp->pfnPrintf(pHlp, "Page directory at %#x was not found in the page pool!\n", cr3 & X86_CR3_PAGE_MASK);
3563 return VERR_INVALID_PARAMETER;
3564 }
3565
3566 int rc = VINF_SUCCESS;
3567 for (unsigned i = 0; i < ELEMENTS(pPD->a); i++)
3568 {
3569 X86PDE Pde = pPD->a[i];
3570 if (Pde.n.u1Present)
3571 {
3572 const uint32_t u32Address = i << X86_PD_SHIFT;
3573 if ((cr4 & X86_CR4_PSE) && Pde.b.u1Size)
3574 pHlp->pfnPrintf(pHlp, /*P R S A D G WT CD AT NX 4M a m d */
3575 "%08x 0 | P %c %c %c %c %c %s %s %s .. 4M %c%c%c %08x\n",
3576 u32Address,
3577 Pde.b.u1Write ? 'W' : 'R',
3578 Pde.b.u1User ? 'U' : 'S',
3579 Pde.b.u1Accessed ? 'A' : '-',
3580 Pde.b.u1Dirty ? 'D' : '-',
3581 Pde.b.u1Global ? 'G' : '-',
3582 Pde.b.u1WriteThru ? "WT" : "--",
3583 Pde.b.u1CacheDisable? "CD" : "--",
3584 Pde.b.u1PAT ? "AT" : "--",
3585 Pde.u & RT_BIT_64(9) ? '1' : '0',
3586 Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3587 Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3588 Pde.u & X86_PDE4M_PG_MASK);
3589 else
3590 {
3591 pHlp->pfnPrintf(pHlp, /*P R S A D G WT CD AT NX 4M a m d */
3592 "%08x 0 | P %c %c %c %c %c %s %s .. .. 4K %c%c%c %08x\n",
3593 u32Address,
3594 Pde.n.u1Write ? 'W' : 'R',
3595 Pde.n.u1User ? 'U' : 'S',
3596 Pde.n.u1Accessed ? 'A' : '-',
3597 Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3598 Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3599 Pde.n.u1WriteThru ? "WT" : "--",
3600 Pde.n.u1CacheDisable? "CD" : "--",
3601 Pde.u & RT_BIT_64(9) ? '1' : '0',
3602 Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3603 Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3604 Pde.u & X86_PDE_PG_MASK);
3605 if (cMaxDepth >= 1)
3606 {
3607 /** @todo what about using the page pool for mapping PTs? */
3608 RTHCPHYS HCPhys = Pde.u & X86_PDE_PG_MASK;
3609 PX86PT pPT = NULL;
3610 if (!(Pde.u & PGM_PDFLAGS_MAPPING))
3611 pPT = (PX86PT)MMPagePhys2Page(pVM, HCPhys);
3612 else
3613 {
3614 for (PPGMMAPPING pMap = pVM->pgm.s.pMappingsR3; pMap; pMap = pMap->pNextR3)
3615 if (u32Address - pMap->GCPtr < pMap->cb)
3616 {
3617 int iPDE = (u32Address - pMap->GCPtr) >> X86_PD_SHIFT;
3618 if (pMap->aPTs[iPDE].HCPhysPT != HCPhys)
3619 pHlp->pfnPrintf(pHlp, "%08x error! Mapping error! PT %d has HCPhysPT=%VHp not %VHp is in the PD.\n",
3620 u32Address, iPDE, pMap->aPTs[iPDE].HCPhysPT, HCPhys);
3621 pPT = pMap->aPTs[iPDE].pPTR3;
3622 }
3623 }
3624 int rc2 = VERR_INVALID_PARAMETER;
3625 if (pPT)
3626 rc2 = pgmR3DumpHierarchyHC32BitPT(pVM, pPT, u32Address, pHlp);
3627 else
3628 pHlp->pfnPrintf(pHlp, "%08x error! Page table at %#x was not found in the page pool!\n", u32Address, HCPhys);
3629 if (rc2 < rc && VBOX_SUCCESS(rc))
3630 rc = rc2;
3631 }
3632 }
3633 }
3634 }
3635
3636 return rc;
3637}
3638
3639
3640/**
3641 * Dumps a 32-bit shadow page table.
3642 *
3643 * @returns VBox status code (VINF_SUCCESS).
3644 * @param pVM The VM handle.
3645 * @param pPT Pointer to the page table.
3646 * @param u32Address The virtual address this table starts at.
3647 * @param PhysSearch Address to search for.
3648 */
3649int pgmR3DumpHierarchyGC32BitPT(PVM pVM, PX86PT pPT, uint32_t u32Address, RTGCPHYS PhysSearch)
3650{
3651 for (unsigned i = 0; i < ELEMENTS(pPT->a); i++)
3652 {
3653 X86PTE Pte = pPT->a[i];
3654 if (Pte.n.u1Present)
3655 {
3656 Log(( /*P R S A D G WT CD AT NX 4M a m d */
3657 "%08x 1 | P %c %c %c %c %c %s %s %s .. 4K %c%c%c %08x\n",
3658 u32Address + (i << X86_PT_SHIFT),
3659 Pte.n.u1Write ? 'W' : 'R',
3660 Pte.n.u1User ? 'U' : 'S',
3661 Pte.n.u1Accessed ? 'A' : '-',
3662 Pte.n.u1Dirty ? 'D' : '-',
3663 Pte.n.u1Global ? 'G' : '-',
3664 Pte.n.u1WriteThru ? "WT" : "--",
3665 Pte.n.u1CacheDisable? "CD" : "--",
3666 Pte.n.u1PAT ? "AT" : "--",
3667 Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3668 Pte.u & RT_BIT(10) ? '1' : '0',
3669 Pte.u & PGM_PTFLAGS_CSAM_VALIDATED ? 'v' : '-',
3670 Pte.u & X86_PDE_PG_MASK));
3671
3672 if ((Pte.u & X86_PDE_PG_MASK) == PhysSearch)
3673 {
3674 uint64_t fPageShw = 0;
3675 RTHCPHYS pPhysHC = 0;
3676
3677 PGMShwGetPage(pVM, (RTGCPTR)(u32Address + (i << X86_PT_SHIFT)), &fPageShw, &pPhysHC);
3678 Log(("Found %VGp at %VGv -> flags=%llx\n", PhysSearch, (RTGCPTR)(u32Address + (i << X86_PT_SHIFT)), fPageShw));
3679 }
3680 }
3681 }
3682 return VINF_SUCCESS;
3683}
3684
3685
3686/**
3687 * Dumps a 32-bit guest page directory and page tables.
3688 *
3689 * @returns VBox status code (VINF_SUCCESS).
3690 * @param pVM The VM handle.
3691 * @param cr3 The root of the hierarchy.
3692 * @param cr4 The CR4, PSE is currently used.
3693 * @param PhysSearch Address to search for.
3694 */
3695PGMR3DECL(int) PGMR3DumpHierarchyGC(PVM pVM, uint64_t cr3, uint64_t cr4, RTGCPHYS PhysSearch)
3696{
3697 bool fLongMode = false;
3698 const unsigned cch = fLongMode ? 16 : 8; NOREF(cch);
3699 PX86PD pPD = 0;
3700
3701 int rc = PGM_GCPHYS_2_PTR(pVM, cr3 & X86_CR3_PAGE_MASK, &pPD);
3702 if (VBOX_FAILURE(rc) || !pPD)
3703 {
3704 Log(("Page directory at %#x was not found in the page pool!\n", cr3 & X86_CR3_PAGE_MASK));
3705 return VERR_INVALID_PARAMETER;
3706 }
3707
3708 Log(("cr3=%08x cr4=%08x%s\n"
3709 "%-*s P - Present\n"
3710 "%-*s | R/W - Read (0) / Write (1)\n"
3711 "%-*s | | U/S - User (1) / Supervisor (0)\n"
3712 "%-*s | | | A - Accessed\n"
3713 "%-*s | | | | D - Dirty\n"
3714 "%-*s | | | | | G - Global\n"
3715 "%-*s | | | | | | WT - Write thru\n"
3716 "%-*s | | | | | | | CD - Cache disable\n"
3717 "%-*s | | | | | | | | AT - Attribute table (PAT)\n"
3718 "%-*s | | | | | | | | | NX - No execute (K8)\n"
3719 "%-*s | | | | | | | | | | 4K/4M/2M - Page size.\n"
3720 "%-*s | | | | | | | | | | | AVL - a=allocated; m=mapping; d=track dirty;\n"
3721 "%-*s | | | | | | | | | | | | p=permanent; v=validated;\n"
3722 "%-*s Level | | | | | | | | | | | | Page\n"
3723 /* xxxx n **** P R S A D G WT CD AT NX 4M AVL xxxxxxxxxxxxx
3724 - W U - - - -- -- -- -- -- 010 */
3725 , cr3, cr4, fLongMode ? " Long Mode" : "",
3726 cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "",
3727 cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "Address"));
3728
3729 for (unsigned i = 0; i < ELEMENTS(pPD->a); i++)
3730 {
3731 X86PDE Pde = pPD->a[i];
3732 if (Pde.n.u1Present)
3733 {
3734 const uint32_t u32Address = i << X86_PD_SHIFT;
3735
3736 if ((cr4 & X86_CR4_PSE) && Pde.b.u1Size)
3737 Log(( /*P R S A D G WT CD AT NX 4M a m d */
3738 "%08x 0 | P %c %c %c %c %c %s %s %s .. 4M %c%c%c %08x\n",
3739 u32Address,
3740 Pde.b.u1Write ? 'W' : 'R',
3741 Pde.b.u1User ? 'U' : 'S',
3742 Pde.b.u1Accessed ? 'A' : '-',
3743 Pde.b.u1Dirty ? 'D' : '-',
3744 Pde.b.u1Global ? 'G' : '-',
3745 Pde.b.u1WriteThru ? "WT" : "--",
3746 Pde.b.u1CacheDisable? "CD" : "--",
3747 Pde.b.u1PAT ? "AT" : "--",
3748 Pde.u & RT_BIT(9) ? '1' : '0',
3749 Pde.u & RT_BIT(10) ? '1' : '0',
3750 Pde.u & RT_BIT(11) ? '1' : '0',
3751 Pde.u & X86_PDE4M_PG_MASK));
3752 /** @todo PhysSearch */
3753 else
3754 {
3755 Log(( /*P R S A D G WT CD AT NX 4M a m d */
3756 "%08x 0 | P %c %c %c %c %c %s %s .. .. 4K %c%c%c %08x\n",
3757 u32Address,
3758 Pde.n.u1Write ? 'W' : 'R',
3759 Pde.n.u1User ? 'U' : 'S',
3760 Pde.n.u1Accessed ? 'A' : '-',
3761 Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3762 Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3763 Pde.n.u1WriteThru ? "WT" : "--",
3764 Pde.n.u1CacheDisable? "CD" : "--",
3765 Pde.u & RT_BIT(9) ? '1' : '0',
3766 Pde.u & RT_BIT(10) ? '1' : '0',
3767 Pde.u & RT_BIT(11) ? '1' : '0',
3768 Pde.u & X86_PDE_PG_MASK));
3769 ////if (cMaxDepth >= 1)
3770 {
3771 /** @todo what about using the page pool for mapping PTs? */
3772 RTGCPHYS GCPhys = Pde.u & X86_PDE_PG_MASK;
3773 PX86PT pPT = NULL;
3774
3775 rc = PGM_GCPHYS_2_PTR(pVM, GCPhys, &pPT);
3776
3777 int rc2 = VERR_INVALID_PARAMETER;
3778 if (pPT)
3779 rc2 = pgmR3DumpHierarchyGC32BitPT(pVM, pPT, u32Address, PhysSearch);
3780 else
3781 Log(("%08x error! Page table at %#x was not found in the page pool!\n", u32Address, GCPhys));
3782 if (rc2 < rc && VBOX_SUCCESS(rc))
3783 rc = rc2;
3784 }
3785 }
3786 }
3787 }
3788
3789 return rc;
3790}
3791
3792
3793/**
3794 * Dumps a page table hierarchy use only physical addresses and cr4/lm flags.
3795 *
3796 * @returns VBox status code (VINF_SUCCESS).
3797 * @param pVM The VM handle.
3798 * @param cr3 The root of the hierarchy.
3799 * @param cr4 The cr4, only PAE and PSE is currently used.
3800 * @param fLongMode Set if long mode, false if not long mode.
3801 * @param cMaxDepth Number of levels to dump.
3802 * @param pHlp Pointer to the output functions.
3803 */
3804PGMR3DECL(int) PGMR3DumpHierarchyHC(PVM pVM, uint64_t cr3, uint64_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3805{
3806 if (!pHlp)
3807 pHlp = DBGFR3InfoLogHlp();
3808 if (!cMaxDepth)
3809 return VINF_SUCCESS;
3810 const unsigned cch = fLongMode ? 16 : 8;
3811 pHlp->pfnPrintf(pHlp,
3812 "cr3=%08x cr4=%08x%s\n"
3813 "%-*s P - Present\n"
3814 "%-*s | R/W - Read (0) / Write (1)\n"
3815 "%-*s | | U/S - User (1) / Supervisor (0)\n"
3816 "%-*s | | | A - Accessed\n"
3817 "%-*s | | | | D - Dirty\n"
3818 "%-*s | | | | | G - Global\n"
3819 "%-*s | | | | | | WT - Write thru\n"
3820 "%-*s | | | | | | | CD - Cache disable\n"
3821 "%-*s | | | | | | | | AT - Attribute table (PAT)\n"
3822 "%-*s | | | | | | | | | NX - No execute (K8)\n"
3823 "%-*s | | | | | | | | | | 4K/4M/2M - Page size.\n"
3824 "%-*s | | | | | | | | | | | AVL - a=allocated; m=mapping; d=track dirty;\n"
3825 "%-*s | | | | | | | | | | | | p=permanent; v=validated;\n"
3826 "%-*s Level | | | | | | | | | | | | Page\n"
3827 /* xxxx n **** P R S A D G WT CD AT NX 4M AVL xxxxxxxxxxxxx
3828 - W U - - - -- -- -- -- -- 010 */
3829 , cr3, cr4, fLongMode ? " Long Mode" : "",
3830 cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "",
3831 cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "Address");
3832 if (cr4 & X86_CR4_PAE)
3833 {
3834 if (fLongMode)
3835 return pgmR3DumpHierarchyHcPaePML4(pVM, cr3 & X86_CR3_PAGE_MASK, cr4, cMaxDepth, pHlp);
3836 return pgmR3DumpHierarchyHCPaePDPT(pVM, cr3 & X86_CR3_PAE_PAGE_MASK, 0, cr4, false, cMaxDepth, pHlp);
3837 }
3838 return pgmR3DumpHierarchyHC32BitPD(pVM, cr3 & X86_CR3_PAGE_MASK, cr4, cMaxDepth, pHlp);
3839}
3840
3841
3842
3843#ifdef VBOX_WITH_DEBUGGER
3844/**
3845 * The '.pgmram' command.
3846 *
3847 * @returns VBox status.
3848 * @param pCmd Pointer to the command descriptor (as registered).
3849 * @param pCmdHlp Pointer to command helper functions.
3850 * @param pVM Pointer to the current VM (if any).
3851 * @param paArgs Pointer to (readonly) array of arguments.
3852 * @param cArgs Number of arguments in the array.
3853 */
3854static DECLCALLBACK(int) pgmR3CmdRam(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3855{
3856 /*
3857 * Validate input.
3858 */
3859 if (!pVM)
3860 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires VM to be selected.\n");
3861 if (!pVM->pgm.s.pRamRangesGC)
3862 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Sorry, no Ram is registered.\n");
3863
3864 /*
3865 * Dump the ranges.
3866 */
3867 int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, "From - To (incl) pvHC\n");
3868 PPGMRAMRANGE pRam;
3869 for (pRam = pVM->pgm.s.pRamRangesR3; pRam; pRam = pRam->pNextR3)
3870 {
3871 rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL,
3872 "%VGp - %VGp %p\n",
3873 pRam->GCPhys, pRam->GCPhysLast, pRam->pvHC);
3874 if (VBOX_FAILURE(rc))
3875 return rc;
3876 }
3877
3878 return VINF_SUCCESS;
3879}
3880
3881
3882/**
3883 * The '.pgmmap' command.
3884 *
3885 * @returns VBox status.
3886 * @param pCmd Pointer to the command descriptor (as registered).
3887 * @param pCmdHlp Pointer to command helper functions.
3888 * @param pVM Pointer to the current VM (if any).
3889 * @param paArgs Pointer to (readonly) array of arguments.
3890 * @param cArgs Number of arguments in the array.
3891 */
3892static DECLCALLBACK(int) pgmR3CmdMap(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3893{
3894 /*
3895 * Validate input.
3896 */
3897 if (!pVM)
3898 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires VM to be selected.\n");
3899 if (!pVM->pgm.s.pMappingsR3)
3900 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Sorry, no mappings are registered.\n");
3901
3902 /*
3903 * Print message about the fixedness of the mappings.
3904 */
3905 int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, pVM->pgm.s.fMappingsFixed ? "The mappings are FIXED.\n" : "The mappings are FLOATING.\n");
3906 if (VBOX_FAILURE(rc))
3907 return rc;
3908
3909 /*
3910 * Dump the ranges.
3911 */
3912 PPGMMAPPING pCur;
3913 for (pCur = pVM->pgm.s.pMappingsR3; pCur; pCur = pCur->pNextR3)
3914 {
3915 rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL,
3916 "%08x - %08x %s\n",
3917 pCur->GCPtr, pCur->GCPtrLast, pCur->pszDesc);
3918 if (VBOX_FAILURE(rc))
3919 return rc;
3920 }
3921
3922 return VINF_SUCCESS;
3923}
3924
3925
3926/**
3927 * The '.pgmsync' command.
3928 *
3929 * @returns VBox status.
3930 * @param pCmd Pointer to the command descriptor (as registered).
3931 * @param pCmdHlp Pointer to command helper functions.
3932 * @param pVM Pointer to the current VM (if any).
3933 * @param paArgs Pointer to (readonly) array of arguments.
3934 * @param cArgs Number of arguments in the array.
3935 */
3936static DECLCALLBACK(int) pgmR3CmdSync(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3937{
3938 /*
3939 * Validate input.
3940 */
3941 if (!pVM)
3942 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires VM to be selected.\n");
3943
3944 /*
3945 * Force page directory sync.
3946 */
3947 VM_FF_SET(pVM, VM_FF_PGM_SYNC_CR3);
3948
3949 int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Forcing page directory sync.\n");
3950 if (VBOX_FAILURE(rc))
3951 return rc;
3952
3953 return VINF_SUCCESS;
3954}
3955
3956
3957/**
3958 * The '.pgmsyncalways' command.
3959 *
3960 * @returns VBox status.
3961 * @param pCmd Pointer to the command descriptor (as registered).
3962 * @param pCmdHlp Pointer to command helper functions.
3963 * @param pVM Pointer to the current VM (if any).
3964 * @param paArgs Pointer to (readonly) array of arguments.
3965 * @param cArgs Number of arguments in the array.
3966 */
3967static DECLCALLBACK(int) pgmR3CmdSyncAlways(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3968{
3969 /*
3970 * Validate input.
3971 */
3972 if (!pVM)
3973 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires VM to be selected.\n");
3974
3975 /*
3976 * Force page directory sync.
3977 */
3978 if (pVM->pgm.s.fSyncFlags & PGM_SYNC_ALWAYS)
3979 {
3980 ASMAtomicAndU32(&pVM->pgm.s.fSyncFlags, ~PGM_SYNC_ALWAYS);
3981 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Disabled permanent forced page directory syncing.\n");
3982 }
3983 else
3984 {
3985 ASMAtomicOrU32(&pVM->pgm.s.fSyncFlags, PGM_SYNC_ALWAYS);
3986 VM_FF_SET(pVM, VM_FF_PGM_SYNC_CR3);
3987 return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Enabled permanent forced page directory syncing.\n");
3988 }
3989}
3990
3991#endif
3992
3993/**
3994 * pvUser argument of the pgmR3CheckIntegrity*Node callbacks.
3995 */
3996typedef struct PGMCHECKINTARGS
3997{
3998 bool fLeftToRight; /**< true: left-to-right; false: right-to-left. */
3999 PPGMPHYSHANDLER pPrevPhys;
4000 PPGMVIRTHANDLER pPrevVirt;
4001 PPGMPHYS2VIRTHANDLER pPrevPhys2Virt;
4002 PVM pVM;
4003} PGMCHECKINTARGS, *PPGMCHECKINTARGS;
4004
4005/**
4006 * Validate a node in the physical handler tree.
4007 *
4008 * @returns 0 on if ok, other wise 1.
4009 * @param pNode The handler node.
4010 * @param pvUser pVM.
4011 */
4012static DECLCALLBACK(int) pgmR3CheckIntegrityPhysHandlerNode(PAVLROGCPHYSNODECORE pNode, void *pvUser)
4013{
4014 PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4015 PPGMPHYSHANDLER pCur = (PPGMPHYSHANDLER)pNode;
4016 AssertReleaseReturn(!((uintptr_t)pCur & 7), 1);
4017 AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %VGp-%VGp %s\n", pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4018 AssertReleaseMsg( !pArgs->pPrevPhys
4019 || (pArgs->fLeftToRight ? pArgs->pPrevPhys->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys->Core.KeyLast > pCur->Core.Key),
4020 ("pPrevPhys=%p %VGp-%VGp %s\n"
4021 " pCur=%p %VGp-%VGp %s\n",
4022 pArgs->pPrevPhys, pArgs->pPrevPhys->Core.Key, pArgs->pPrevPhys->Core.KeyLast, pArgs->pPrevPhys->pszDesc,
4023 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4024 pArgs->pPrevPhys = pCur;
4025 return 0;
4026}
4027
4028
4029/**
4030 * Validate a node in the virtual handler tree.
4031 *
4032 * @returns 0 on if ok, other wise 1.
4033 * @param pNode The handler node.
4034 * @param pvUser pVM.
4035 */
4036static DECLCALLBACK(int) pgmR3CheckIntegrityVirtHandlerNode(PAVLROGCPTRNODECORE pNode, void *pvUser)
4037{
4038 PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4039 PPGMVIRTHANDLER pCur = (PPGMVIRTHANDLER)pNode;
4040 AssertReleaseReturn(!((uintptr_t)pCur & 7), 1);
4041 AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %VGv-%VGv %s\n", pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4042 AssertReleaseMsg( !pArgs->pPrevVirt
4043 || (pArgs->fLeftToRight ? pArgs->pPrevVirt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevVirt->Core.KeyLast > pCur->Core.Key),
4044 ("pPrevVirt=%p %VGv-%VGv %s\n"
4045 " pCur=%p %VGv-%VGv %s\n",
4046 pArgs->pPrevVirt, pArgs->pPrevVirt->Core.Key, pArgs->pPrevVirt->Core.KeyLast, pArgs->pPrevVirt->pszDesc,
4047 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4048 for (unsigned iPage = 0; iPage < pCur->cPages; iPage++)
4049 {
4050 AssertReleaseMsg(pCur->aPhysToVirt[iPage].offVirtHandler == -RT_OFFSETOF(PGMVIRTHANDLER, aPhysToVirt[iPage]),
4051 ("pCur=%p %VGv-%VGv %s\n"
4052 "iPage=%d offVirtHandle=%#x expected %#x\n",
4053 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc,
4054 iPage, pCur->aPhysToVirt[iPage].offVirtHandler, -RT_OFFSETOF(PGMVIRTHANDLER, aPhysToVirt[iPage])));
4055 }
4056 pArgs->pPrevVirt = pCur;
4057 return 0;
4058}
4059
4060
4061/**
4062 * Validate a node in the virtual handler tree.
4063 *
4064 * @returns 0 on if ok, other wise 1.
4065 * @param pNode The handler node.
4066 * @param pvUser pVM.
4067 */
4068static DECLCALLBACK(int) pgmR3CheckIntegrityPhysToVirtHandlerNode(PAVLROGCPHYSNODECORE pNode, void *pvUser)
4069{
4070 PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4071 PPGMPHYS2VIRTHANDLER pCur = (PPGMPHYS2VIRTHANDLER)pNode;
4072 AssertReleaseMsgReturn(!((uintptr_t)pCur & 3), ("\n"), 1);
4073 AssertReleaseMsgReturn(!(pCur->offVirtHandler & 3), ("\n"), 1);
4074 AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %VGp-%VGp\n", pCur, pCur->Core.Key, pCur->Core.KeyLast));
4075 AssertReleaseMsg( !pArgs->pPrevPhys2Virt
4076 || (pArgs->fLeftToRight ? pArgs->pPrevPhys2Virt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys2Virt->Core.KeyLast > pCur->Core.Key),
4077 ("pPrevPhys2Virt=%p %VGp-%VGp\n"
4078 " pCur=%p %VGp-%VGp\n",
4079 pArgs->pPrevPhys2Virt, pArgs->pPrevPhys2Virt->Core.Key, pArgs->pPrevPhys2Virt->Core.KeyLast,
4080 pCur, pCur->Core.Key, pCur->Core.KeyLast));
4081 AssertReleaseMsg( !pArgs->pPrevPhys2Virt
4082 || (pArgs->fLeftToRight ? pArgs->pPrevPhys2Virt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys2Virt->Core.KeyLast > pCur->Core.Key),
4083 ("pPrevPhys2Virt=%p %VGp-%VGp\n"
4084 " pCur=%p %VGp-%VGp\n",
4085 pArgs->pPrevPhys2Virt, pArgs->pPrevPhys2Virt->Core.Key, pArgs->pPrevPhys2Virt->Core.KeyLast,
4086 pCur, pCur->Core.Key, pCur->Core.KeyLast));
4087 AssertReleaseMsg((pCur->offNextAlias & (PGMPHYS2VIRTHANDLER_IN_TREE | PGMPHYS2VIRTHANDLER_IS_HEAD)) == (PGMPHYS2VIRTHANDLER_IN_TREE | PGMPHYS2VIRTHANDLER_IS_HEAD),
4088 ("pCur=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4089 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias));
4090 if (pCur->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK)
4091 {
4092 PPGMPHYS2VIRTHANDLER pCur2 = pCur;
4093 for (;;)
4094 {
4095 pCur2 = (PPGMPHYS2VIRTHANDLER)((intptr_t)pCur + (pCur->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK));
4096 AssertReleaseMsg(pCur2 != pCur,
4097 (" pCur=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4098 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias));
4099 AssertReleaseMsg((pCur2->offNextAlias & (PGMPHYS2VIRTHANDLER_IN_TREE | PGMPHYS2VIRTHANDLER_IS_HEAD)) == PGMPHYS2VIRTHANDLER_IN_TREE,
4100 (" pCur=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4101 "pCur2=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4102 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4103 pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4104 AssertReleaseMsg((pCur2->Core.Key ^ pCur->Core.Key) < PAGE_SIZE,
4105 (" pCur=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4106 "pCur2=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4107 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4108 pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4109 AssertReleaseMsg((pCur2->Core.KeyLast ^ pCur->Core.KeyLast) < PAGE_SIZE,
4110 (" pCur=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4111 "pCur2=%p:{.Core.Key=%VGp, .Core.KeyLast=%VGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4112 pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4113 pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4114 if (!(pCur2->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK))
4115 break;
4116 }
4117 }
4118
4119 pArgs->pPrevPhys2Virt = pCur;
4120 return 0;
4121}
4122
4123
4124/**
4125 * Perform an integrity check on the PGM component.
4126 *
4127 * @returns VINF_SUCCESS if everything is fine.
4128 * @returns VBox error status after asserting on integrity breach.
4129 * @param pVM The VM handle.
4130 */
4131PDMR3DECL(int) PGMR3CheckIntegrity(PVM pVM)
4132{
4133 AssertReleaseReturn(pVM->pgm.s.offVM, VERR_INTERNAL_ERROR);
4134
4135 /*
4136 * Check the trees.
4137 */
4138 int cErrors = 0;
4139 const static PGMCHECKINTARGS s_LeftToRight = { true, NULL, NULL, NULL, pVM };
4140 const static PGMCHECKINTARGS s_RightToLeft = { false, NULL, NULL, NULL, pVM };
4141 PGMCHECKINTARGS Args = s_LeftToRight;
4142 cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesHC->PhysHandlers, true, pgmR3CheckIntegrityPhysHandlerNode, &Args);
4143 Args = s_RightToLeft;
4144 cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesHC->PhysHandlers, false, pgmR3CheckIntegrityPhysHandlerNode, &Args);
4145 Args = s_LeftToRight;
4146 cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesHC->VirtHandlers, true, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4147 Args = s_RightToLeft;
4148 cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesHC->VirtHandlers, false, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4149 Args = s_LeftToRight;
4150 cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesHC->HyperVirtHandlers, true, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4151 Args = s_RightToLeft;
4152 cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesHC->HyperVirtHandlers, false, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4153 Args = s_LeftToRight;
4154 cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesHC->PhysToVirtHandlers, true, pgmR3CheckIntegrityPhysToVirtHandlerNode, &Args);
4155 Args = s_RightToLeft;
4156 cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesHC->PhysToVirtHandlers, false, pgmR3CheckIntegrityPhysToVirtHandlerNode, &Args);
4157
4158 return !cErrors ? VINF_SUCCESS : VERR_INTERNAL_ERROR;
4159}
4160
4161
4162/**
4163 * Inform PGM if we want all mappings to be put into the shadow page table. (necessary for e.g. VMX)
4164 *
4165 * @returns VBox status code.
4166 * @param pVM VM handle.
4167 * @param fEnable Enable or disable shadow mappings
4168 */
4169PGMR3DECL(int) PGMR3ChangeShwPDMappings(PVM pVM, bool fEnable)
4170{
4171 pVM->pgm.s.fDisableMappings = !fEnable;
4172
4173 uint32_t cb;
4174 int rc = PGMR3MappingsSize(pVM, &cb);
4175 AssertRCReturn(rc, rc);
4176
4177 /* Pretend the mappings are now fixed; to force a refresh of the reserved PDEs. */
4178 rc = PGMR3MappingsFix(pVM, MM_HYPER_AREA_ADDRESS, cb);
4179 AssertRCReturn(rc, rc);
4180
4181 return VINF_SUCCESS;
4182}
Note: See TracBrowser for help on using the repository browser.

© 2024 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette