PGM.cpp@ 24760

Last change on this file since 24760 was 24723, checked in by vboxsync, 15 years ago
Introducing PGMPhysInvalidatePageMapTLBEntry
Property svn:eol-style set to `native` Property svn:keywords set to `Id`
File size: 213.0 KB

Line
1	/* $Id: PGM.cpp 24723 2009-11-17 14:09:41Z 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	* @see grp_pgm,
26	* @ref pg_pgm_pool,
27	* @ref pg_pgm_phys.
28	*
29	*
30	* @section sec_pgm_modes Paging Modes
31	*
32	* There are three memory contexts: Host Context (HC), Guest Context (GC)
33	* and intermediate context. When talking about paging HC can also be refered to
34	* as "host paging", and GC refered to as "shadow paging".
35	*
36	* We define three basic paging modes: 32-bit, PAE and AMD64. The host paging mode
37	* is defined by the host operating system. The mode used in the shadow paging mode
38	* depends on the host paging mode and what the mode the guest is currently in. The
39	* following relation between the two is defined:
40	*
41	* @verbatim
42	Host > 32-bit \| PAE \| AMD64 \|
43	Guest \| \| \| \|
44	==v================================
45	32-bit 32-bit PAE PAE
46	-------\|--------\|--------\|--------\|
47	PAE PAE PAE PAE
48	-------\|--------\|--------\|--------\|
49	AMD64 AMD64 AMD64 AMD64
50	-------\|--------\|--------\|--------\| @endverbatim
51	*
52	* All configuration except those in the diagonal (upper left) are expected to
53	* require special effort from the switcher (i.e. a bit slower).
54	*
55	*
56	*
57	*
58	* @section sec_pgm_shw The Shadow Memory Context
59	*
60	*
61	* [..]
62	*
63	* Because of guest context mappings requires PDPT and PML4 entries to allow
64	* writing on AMD64, the two upper levels will have fixed flags whatever the
65	* guest is thinking of using there. So, when shadowing the PD level we will
66	* calculate the effective flags of PD and all the higher levels. In legacy
67	* PAE mode this only applies to the PWT and PCD bits (the rest are
68	* ignored/reserved/MBZ). We will ignore those bits for the present.
69	*
70	*
71	*
72	* @section sec_pgm_int The Intermediate Memory Context
73	*
74	* The world switch goes thru an intermediate memory context which purpose it is
75	* to provide different mappings of the switcher code. All guest mappings are also
76	* present in this context.
77	*
78	* The switcher code is mapped at the same location as on the host, at an
79	* identity mapped location (physical equals virtual address), and at the
80	* hypervisor location. The identity mapped location is for when the world
81	* switches that involves disabling paging.
82	*
83	* PGM maintain page tables for 32-bit, PAE and AMD64 paging modes. This
84	* simplifies switching guest CPU mode and consistency at the cost of more
85	* code to do the work. All memory use for those page tables is located below
86	* 4GB (this includes page tables for guest context mappings).
87	*
88	*
89	* @subsection subsec_pgm_int_gc Guest Context Mappings
90	*
91	* During assignment and relocation of a guest context mapping the intermediate
92	* memory context is used to verify the new location.
93	*
94	* Guest context mappings are currently restricted to below 4GB, for reasons
95	* of simplicity. This may change when we implement AMD64 support.
96	*
97	*
98	*
99	*
100	* @section sec_pgm_misc Misc
101	*
102	* @subsection subsec_pgm_misc_diff Differences Between Legacy PAE and Long Mode PAE
103	*
104	* The differences between legacy PAE and long mode PAE are:
105	* -# PDPE bits 1, 2, 5 and 6 are defined differently. In leagcy mode they are
106	* all marked down as must-be-zero, while in long mode 1, 2 and 5 have the
107	* usual meanings while 6 is ignored (AMD). This means that upon switching to
108	* legacy PAE mode we'll have to clear these bits and when going to long mode
109	* they must be set. This applies to both intermediate and shadow contexts,
110	* however we don't need to do it for the intermediate one since we're
111	* executing with CR0.WP at that time.
112	* -# CR3 allows a 32-byte aligned address in legacy mode, while in long mode
113	* a page aligned one is required.
114	*
115	*
116	* @section sec_pgm_handlers Access Handlers
117	*
118	* Placeholder.
119	*
120	*
121	* @subsection sec_pgm_handlers_virt Virtual Access Handlers
122	*
123	* Placeholder.
124	*
125	*
126	* @subsection sec_pgm_handlers_virt Virtual Access Handlers
127	*
128	* We currently implement three types of virtual access handlers: ALL, WRITE
129	* and HYPERVISOR (WRITE). See PGMVIRTHANDLERTYPE for some more details.
130	*
131	* The HYPERVISOR access handlers is kept in a separate tree since it doesn't apply
132	* to physical pages (PGMTREES::HyperVirtHandlers) and only needs to be consulted in
133	* a special \#PF case. The ALL and WRITE are in the PGMTREES::VirtHandlers tree, the
134	* rest of this section is going to be about these handlers.
135	*
136	* We'll go thru the life cycle of a handler and try make sense of it all, don't know
137	* how successfull this is gonna be...
138	*
139	* 1. A handler is registered thru the PGMR3HandlerVirtualRegister and
140	* PGMHandlerVirtualRegisterEx APIs. We check for conflicting virtual handlers
141	* and create a new node that is inserted into the AVL tree (range key). Then
142	* a full PGM resync is flagged (clear pool, sync cr3, update virtual bit of PGMPAGE).
143	*
144	* 2. The following PGMSyncCR3/SyncCR3 operation will first make invoke HandlerVirtualUpdate.
145	*
146	* 2a. HandlerVirtualUpdate will will lookup all the pages covered by virtual handlers
147	* via the current guest CR3 and update the physical page -> virtual handler
148	* translation. Needless to say, this doesn't exactly scale very well. If any changes
149	* are detected, it will flag a virtual bit update just like we did on registration.
150	* PGMPHYS pages with changes will have their virtual handler state reset to NONE.
151	*
152	* 2b. The virtual bit update process will iterate all the pages covered by all the
153	* virtual handlers and update the PGMPAGE virtual handler state to the max of all
154	* virtual handlers on that page.
155	*
156	* 2c. Back in SyncCR3 we will now flush the entire shadow page cache to make sure
157	* we don't miss any alias mappings of the monitored pages.
158	*
159	* 2d. SyncCR3 will then proceed with syncing the CR3 table.
160	*
161	* 3. \#PF(np,read) on a page in the range. This will cause it to be synced
162	* read-only and resumed if it's a WRITE handler. If it's an ALL handler we
163	* will call the handlers like in the next step. If the physical mapping has
164	* changed we will - some time in the future - perform a handler callback
165	* (optional) and update the physical -> virtual handler cache.
166	*
167	* 4. \#PF(,write) on a page in the range. This will cause the handler to
168	* be invoked.
169	*
170	* 5. The guest invalidates the page and changes the physical backing or
171	* unmaps it. This should cause the invalidation callback to be invoked
172	* (it might not yet be 100% perfect). Exactly what happens next... is
173	* this where we mess up and end up out of sync for a while?
174	*
175	* 6. The handler is deregistered by the client via PGMHandlerVirtualDeregister.
176	* We will then set all PGMPAGEs in the physical -> virtual handler cache for
177	* this handler to NONE and trigger a full PGM resync (basically the same
178	* as int step 1). Which means 2 is executed again.
179	*
180	*
181	* @subsubsection sub_sec_pgm_handler_virt_todo TODOs
182	*
183	* There is a bunch of things that needs to be done to make the virtual handlers
184	* work 100% correctly and work more efficiently.
185	*
186	* The first bit hasn't been implemented yet because it's going to slow the
187	* whole mess down even more, and besides it seems to be working reliably for
188	* our current uses. OTOH, some of the optimizations might end up more or less
189	* implementing the missing bits, so we'll see.
190	*
191	* On the optimization side, the first thing to do is to try avoid unnecessary
192	* cache flushing. Then try team up with the shadowing code to track changes
193	* in mappings by means of access to them (shadow in), updates to shadows pages,
194	* invlpg, and shadow PT discarding (perhaps).
195	*
196	* Some idea that have popped up for optimization for current and new features:
197	* - bitmap indicating where there are virtual handlers installed.
198	* (4KB => 220 pages, page 212 => covers 32-bit address space 1:1!)
199	* - Further optimize this by min/max (needs min/max avl getters).
200	* - Shadow page table entry bit (if any left)?
201	*
202	*/
203
204
205	/** @page pg_pgm_phys PGM Physical Guest Memory Management
206	*
207	*
208	* Objectives:
209	* - Guest RAM over-commitment using memory ballooning,
210	* zero pages and general page sharing.
211	* - Moving or mirroring a VM onto a different physical machine.
212	*
213	*
214	* @subsection subsec_pgmPhys_Definitions Definitions
215	*
216	* Allocation chunk - A RTR0MemObjAllocPhysNC object and the tracking
217	* machinery assoicated with it.
218	*
219	*
220	*
221	*
222	* @subsection subsec_pgmPhys_AllocPage Allocating a page.
223	*
224	* Initially we map all guest memory to the (per VM) zero page, which
225	* means that none of the read functions will cause pages to be allocated.
226	*
227	* Exception, access bit in page tables that have been shared. This must
228	* be handled, but we must also make sure PGMGst*Modify doesn't make
229	* unnecessary modifications.
230	*
231	* Allocation points:
232	* - PGMPhysSimpleWriteGCPhys and PGMPhysWrite.
233	* - Replacing a zero page mapping at \#PF.
234	* - Replacing a shared page mapping at \#PF.
235	* - ROM registration (currently MMR3RomRegister).
236	* - VM restore (pgmR3Load).
237	*
238	* For the first three it would make sense to keep a few pages handy
239	* until we've reached the max memory commitment for the VM.
240	*
241	* For the ROM registration, we know exactly how many pages we need
242	* and will request these from ring-0. For restore, we will save
243	* the number of non-zero pages in the saved state and allocate
244	* them up front. This would allow the ring-0 component to refuse
245	* the request if the isn't sufficient memory available for VM use.
246	*
247	* Btw. for both ROM and restore allocations we won't be requiring
248	* zeroed pages as they are going to be filled instantly.
249	*
250	*
251	* @subsection subsec_pgmPhys_FreePage Freeing a page
252	*
253	* There are a few points where a page can be freed:
254	* - After being replaced by the zero page.
255	* - After being replaced by a shared page.
256	* - After being ballooned by the guest additions.
257	* - At reset.
258	* - At restore.
259	*
260	* When freeing one or more pages they will be returned to the ring-0
261	* component and replaced by the zero page.
262	*
263	* The reasoning for clearing out all the pages on reset is that it will
264	* return us to the exact same state as on power on, and may thereby help
265	* us reduce the memory load on the system. Further it might have a
266	* (temporary) positive influence on memory fragmentation (@see subsec_pgmPhys_Fragmentation).
267	*
268	* On restore, as mention under the allocation topic, pages should be
269	* freed / allocated depending on how many is actually required by the
270	* new VM state. The simplest approach is to do like on reset, and free
271	* all non-ROM pages and then allocate what we need.
272	*
273	* A measure to prevent some fragmentation, would be to let each allocation
274	* chunk have some affinity towards the VM having allocated the most pages
275	* from it. Also, try make sure to allocate from allocation chunks that
276	* are almost full. Admittedly, both these measures might work counter to
277	* our intentions and its probably not worth putting a lot of effort,
278	* cpu time or memory into this.
279	*
280	*
281	* @subsection subsec_pgmPhys_SharePage Sharing a page
282	*
283	* The basic idea is that there there will be a idle priority kernel
284	* thread walking the non-shared VM pages hashing them and looking for
285	* pages with the same checksum. If such pages are found, it will compare
286	* them byte-by-byte to see if they actually are identical. If found to be
287	* identical it will allocate a shared page, copy the content, check that
288	* the page didn't change while doing this, and finally request both the
289	* VMs to use the shared page instead. If the page is all zeros (special
290	* checksum and byte-by-byte check) it will request the VM that owns it
291	* to replace it with the zero page.
292	*
293	* To make this efficient, we will have to make sure not to try share a page
294	* that will change its contents soon. This part requires the most work.
295	* A simple idea would be to request the VM to write monitor the page for
296	* a while to make sure it isn't modified any time soon. Also, it may
297	* make sense to skip pages that are being write monitored since this
298	* information is readily available to the thread if it works on the
299	* per-VM guest memory structures (presently called PGMRAMRANGE).
300	*
301	*
302	* @subsection subsec_pgmPhys_Fragmentation Fragmentation Concerns and Counter Measures
303	*
304	* The pages are organized in allocation chunks in ring-0, this is a necessity
305	* if we wish to have an OS agnostic approach to this whole thing. (On Linux we
306	* could easily work on a page-by-page basis if we liked. Whether this is possible
307	* or efficient on NT I don't quite know.) Fragmentation within these chunks may
308	* become a problem as part of the idea here is that we wish to return memory to
309	* the host system.
310	*
311	* For instance, starting two VMs at the same time, they will both allocate the
312	* guest memory on-demand and if permitted their page allocations will be
313	* intermixed. Shut down one of the two VMs and it will be difficult to return
314	* any memory to the host system because the page allocation for the two VMs are
315	* mixed up in the same allocation chunks.
316	*
317	* To further complicate matters, when pages are freed because they have been
318	* ballooned or become shared/zero the whole idea is that the page is supposed
319	* to be reused by another VM or returned to the host system. This will cause
320	* allocation chunks to contain pages belonging to different VMs and prevent
321	* returning memory to the host when one of those VM shuts down.
322	*
323	* The only way to really deal with this problem is to move pages. This can
324	* either be done at VM shutdown and or by the idle priority worker thread
325	* that will be responsible for finding sharable/zero pages. The mechanisms
326	* involved for coercing a VM to move a page (or to do it for it) will be
327	* the same as when telling it to share/zero a page.
328	*
329	*
330	* @subsection subsec_pgmPhys_Tracking Tracking Structures And Their Cost
331	*
332	* There's a difficult balance between keeping the per-page tracking structures
333	* (global and guest page) easy to use and keeping them from eating too much
334	* memory. We have limited virtual memory resources available when operating in
335	* 32-bit kernel space (on 64-bit there'll it's quite a different story). The
336	* tracking structures will be attemted designed such that we can deal with up
337	* to 32GB of memory on a 32-bit system and essentially unlimited on 64-bit ones.
338	*
339	*
340	* @subsubsection subsubsec_pgmPhys_Tracking_Kernel Kernel Space
341	*
342	* @see pg_GMM
343	*
344	* @subsubsection subsubsec_pgmPhys_Tracking_PerVM Per-VM
345	*
346	* Fixed info is the physical address of the page (HCPhys) and the page id
347	* (described above). Theoretically we'll need 48(-12) bits for the HCPhys part.
348	* Today we've restricting ourselves to 40(-12) bits because this is the current
349	* restrictions of all AMD64 implementations (I think Barcelona will up this
350	* to 48(-12) bits, not that it really matters) and I needed the bits for
351	* tracking mappings of a page. 48-12 = 36. That leaves 28 bits, which means a
352	* decent range for the page id: 2^(28+12) = 1024TB.
353	*
354	* In additions to these, we'll have to keep maintaining the page flags as we
355	* currently do. Although it wouldn't harm to optimize these quite a bit, like
356	* for instance the ROM shouldn't depend on having a write handler installed
357	* in order for it to become read-only. A RO/RW bit should be considered so
358	* that the page syncing code doesn't have to mess about checking multiple
359	* flag combinations (ROM \|\| RW handler \|\| write monitored) in order to
360	* figure out how to setup a shadow PTE. But this of course, is second
361	* priority at present. Current this requires 12 bits, but could probably
362	* be optimized to ~8.
363	*
364	* Then there's the 24 bits used to track which shadow page tables are
365	* currently mapping a page for the purpose of speeding up physical
366	* access handlers, and thereby the page pool cache. More bit for this
367	* purpose wouldn't hurt IIRC.
368	*
369	* Then there is a new bit in which we need to record what kind of page
370	* this is, shared, zero, normal or write-monitored-normal. This'll
371	* require 2 bits. One bit might be needed for indicating whether a
372	* write monitored page has been written to. And yet another one or
373	* two for tracking migration status. 3-4 bits total then.
374	*
375	* Whatever is left will can be used to record the sharabilitiy of a
376	* page. The page checksum will not be stored in the per-VM table as
377	* the idle thread will not be permitted to do modifications to it.
378	* It will instead have to keep its own working set of potentially
379	* shareable pages and their check sums and stuff.
380	*
381	* For the present we'll keep the current packing of the
382	* PGMRAMRANGE::aHCPhys to keep the changes simple, only of course,
383	* we'll have to change it to a struct with a total of 128-bits at
384	* our disposal.
385	*
386	* The initial layout will be like this:
387	* @verbatim
388	RTHCPHYS HCPhys; The current stuff.
389	63:40 Current shadow PT tracking stuff.
390	39:12 The physical page frame number.
391	11:0 The current flags.
392	uint32_t u28PageId : 28; The page id.
393	uint32_t u2State : 2; The page state { zero, shared, normal, write monitored }.
394	uint32_t fWrittenTo : 1; Whether a write monitored page was written to.
395	uint32_t u1Reserved : 1; Reserved for later.
396	uint32_t u32Reserved; Reserved for later, mostly sharing stats.
397	@endverbatim
398	*
399	* The final layout will be something like this:
400	* @verbatim
401	RTHCPHYS HCPhys; The current stuff.
402	63:48 High page id (12+).
403	47:12 The physical page frame number.
404	11:0 Low page id.
405	uint32_t fReadOnly : 1; Whether it's readonly page (rom or monitored in some way).
406	uint32_t u3Type : 3; The page type {RESERVED, MMIO, MMIO2, ROM, shadowed ROM, RAM}.
407	uint32_t u2PhysMon : 2; Physical access handler type {none, read, write, all}.
408	uint32_t u2VirtMon : 2; Virtual access handler type {none, read, write, all}..
409	uint32_t u2State : 2; The page state { zero, shared, normal, write monitored }.
410	uint32_t fWrittenTo : 1; Whether a write monitored page was written to.
411	uint32_t u20Reserved : 20; Reserved for later, mostly sharing stats.
412	uint32_t u32Tracking; The shadow PT tracking stuff, roughly.
413	@endverbatim
414	*
415	* Cost wise, this means we'll double the cost for guest memory. There isn't anyway
416	* around that I'm afraid. It means that the cost of dealing out 32GB of memory
417	* to one or more VMs is: (32GB >> PAGE_SHIFT) * 16 bytes, or 128MBs. Or another
418	* example, the VM heap cost when assigning 1GB to a VM will be: 4MB.
419	*
420	* A couple of cost examples for the total cost per-VM + kernel.
421	* 32-bit Windows and 32-bit linux:
422	* 1GB guest ram, 256K pages: 4MB + 2MB(+) = 6MB
423	* 4GB guest ram, 1M pages: 16MB + 8MB(+) = 24MB
424	* 32GB guest ram, 8M pages: 128MB + 64MB(+) = 192MB
425	* 64-bit Windows and 64-bit linux:
426	* 1GB guest ram, 256K pages: 4MB + 3MB(+) = 7MB
427	* 4GB guest ram, 1M pages: 16MB + 12MB(+) = 28MB
428	* 32GB guest ram, 8M pages: 128MB + 96MB(+) = 224MB
429	*
430	* UPDATE - 2007-09-27:
431	* Will need a ballooned flag/state too because we cannot
432	* trust the guest 100% and reporting the same page as ballooned more
433	* than once will put the GMM off balance.
434	*
435	*
436	* @subsection subsec_pgmPhys_Serializing Serializing Access
437	*
438	* Initially, we'll try a simple scheme:
439	*
440	* - The per-VM RAM tracking structures (PGMRAMRANGE) is only modified
441	* by the EMT thread of that VM while in the pgm critsect.
442	* - Other threads in the VM process that needs to make reliable use of
443	* the per-VM RAM tracking structures will enter the critsect.
444	* - No process external thread or kernel thread will ever try enter
445	* the pgm critical section, as that just won't work.
446	* - The idle thread (and similar threads) doesn't not need 100% reliable
447	* data when performing it tasks as the EMT thread will be the one to
448	* do the actual changes later anyway. So, as long as it only accesses
449	* the main ram range, it can do so by somehow preventing the VM from
450	* being destroyed while it works on it...
451	*
452	* - The over-commitment management, including the allocating/freeing
453	* chunks, is serialized by a ring-0 mutex lock (a fast one since the
454	* more mundane mutex implementation is broken on Linux).
455	* - A separeate mutex is protecting the set of allocation chunks so
456	* that pages can be shared or/and freed up while some other VM is
457	* allocating more chunks. This mutex can be take from under the other
458	* one, but not the otherway around.
459	*
460	*
461	* @subsection subsec_pgmPhys_Request VM Request interface
462	*
463	* When in ring-0 it will become necessary to send requests to a VM so it can
464	* for instance move a page while defragmenting during VM destroy. The idle
465	* thread will make use of this interface to request VMs to setup shared
466	* pages and to perform write monitoring of pages.
467	*
468	* I would propose an interface similar to the current VMReq interface, similar
469	* in that it doesn't require locking and that the one sending the request may
470	* wait for completion if it wishes to. This shouldn't be very difficult to
471	* realize.
472	*
473	* The requests themselves are also pretty simple. They are basically:
474	* -# Check that some precondition is still true.
475	* -# Do the update.
476	* -# Update all shadow page tables involved with the page.
477	*
478	* The 3rd step is identical to what we're already doing when updating a
479	* physical handler, see pgmHandlerPhysicalSetRamFlagsAndFlushShadowPTs.
480	*
481	*
482	*
483	* @section sec_pgmPhys_MappingCaches Mapping Caches
484	*
485	* In order to be able to map in and out memory and to be able to support
486	* guest with more RAM than we've got virtual address space, we'll employing
487	* a mapping cache. There is already a tiny one for GC (see PGMGCDynMapGCPageEx)
488	* and we'll create a similar one for ring-0 unless we decide to setup a dedicate
489	* memory context for the HWACCM execution.
490	*
491	*
492	* @subsection subsec_pgmPhys_MappingCaches_R3 Ring-3
493	*
494	* We've considered implementing the ring-3 mapping cache page based but found
495	* that this was bother some when one had to take into account TLBs+SMP and
496	* portability (missing the necessary APIs on several platforms). There were
497	* also some performance concerns with this approach which hadn't quite been
498	* worked out.
499	*
500	* Instead, we'll be mapping allocation chunks into the VM process. This simplifies
501	* matters greatly quite a bit since we don't need to invent any new ring-0 stuff,
502	* only some minor RTR0MEMOBJ mapping stuff. The main concern here is that mapping
503	* compared to the previous idea is that mapping or unmapping a 1MB chunk is more
504	* costly than a single page, although how much more costly is uncertain. We'll
505	* try address this by using a very big cache, preferably bigger than the actual
506	* VM RAM size if possible. The current VM RAM sizes should give some idea for
507	* 32-bit boxes, while on 64-bit we can probably get away with employing an
508	* unlimited cache.
509	*
510	* The cache have to parts, as already indicated, the ring-3 side and the
511	* ring-0 side.
512	*
513	* The ring-0 will be tied to the page allocator since it will operate on the
514	* memory objects it contains. It will therefore require the first ring-0 mutex
515	* discussed in @ref subsec_pgmPhys_Serializing. We
516	* some double house keeping wrt to who has mapped what I think, since both
517	* VMMR0.r0 and RTR0MemObj will keep track of mapping relataions
518	*
519	* The ring-3 part will be protected by the pgm critsect. For simplicity, we'll
520	* require anyone that desires to do changes to the mapping cache to do that
521	* from within this critsect. Alternatively, we could employ a separate critsect
522	* for serializing changes to the mapping cache as this would reduce potential
523	* contention with other threads accessing mappings unrelated to the changes
524	* that are in process. We can see about this later, contention will show
525	* up in the statistics anyway, so it'll be simple to tell.
526	*
527	* The organization of the ring-3 part will be very much like how the allocation
528	* chunks are organized in ring-0, that is in an AVL tree by chunk id. To avoid
529	* having to walk the tree all the time, we'll have a couple of lookaside entries
530	* like in we do for I/O ports and MMIO in IOM.
531	*
532	* The simplified flow of a PGMPhysRead/Write function:
533	* -# Enter the PGM critsect.
534	* -# Lookup GCPhys in the ram ranges and get the Page ID.
535	* -# Calc the Allocation Chunk ID from the Page ID.
536	* -# Check the lookaside entries and then the AVL tree for the Chunk ID.
537	* If not found in cache:
538	* -# Call ring-0 and request it to be mapped and supply
539	* a chunk to be unmapped if the cache is maxed out already.
540	* -# Insert the new mapping into the AVL tree (id + R3 address).
541	* -# Update the relevant lookaside entry and return the mapping address.
542	* -# Do the read/write according to monitoring flags and everything.
543	* -# Leave the critsect.
544	*
545	*
546	* @section sec_pgmPhys_Fallback Fallback
547	*
548	* Current all the "second tier" hosts will not support the RTR0MemObjAllocPhysNC
549	* API and thus require a fallback.
550	*
551	* So, when RTR0MemObjAllocPhysNC returns VERR_NOT_SUPPORTED the page allocator
552	* will return to the ring-3 caller (and later ring-0) and asking it to seed
553	* the page allocator with some fresh pages (VERR_GMM_SEED_ME). Ring-3 will
554	* then perform an SUPR3PageAlloc(cbChunk >> PAGE_SHIFT) call and make a
555	* "SeededAllocPages" call to ring-0.
556	*
557	* The first time ring-0 sees the VERR_NOT_SUPPORTED failure it will disable
558	* all page sharing (zero page detection will continue). It will also force
559	* all allocations to come from the VM which seeded the page. Both these
560	* measures are taken to make sure that there will never be any need for
561	* mapping anything into ring-3 - everything will be mapped already.
562	*
563	* Whether we'll continue to use the current MM locked memory management
564	* for this I don't quite know (I'd prefer not to and just ditch that all
565	* togther), we'll see what's simplest to do.
566	*
567	*
568	*
569	* @section sec_pgmPhys_Changes Changes
570	*
571	* Breakdown of the changes involved?
572	*/
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/selm.h>
588	#include <VBox/ssm.h>
589	#include <VBox/hwaccm.h>
590	#include "PGMInternal.h"
591	#include <VBox/vm.h>
592
593	#include <VBox/dbg.h>
594	#include <VBox/param.h>
595	#include <VBox/err.h>
596
597	#include <iprt/asm.h>
598	#include <iprt/assert.h>
599	#include <iprt/env.h>
600	#include <iprt/mem.h>
601	#include <iprt/file.h>
602	#include <iprt/string.h>
603	#include <iprt/thread.h>
604
605
606	/*******************************************************************************
607	* Defined Constants And Macros *
608	*******************************************************************************/
609	/** Saved state data unit version for 2.5.x and later. */
610	#define PGM_SAVED_STATE_VERSION 9
611	/** Saved state data unit version for 2.2.2 and later. */
612	#define PGM_SAVED_STATE_VERSION_2_2_2 8
613	/** Saved state data unit version for 2.2.0. */
614	#define PGM_SAVED_STATE_VERSION_RR_DESC 7
615	/** Saved state data unit version. */
616	#define PGM_SAVED_STATE_VERSION_OLD_PHYS_CODE 6
617
618
619	/*******************************************************************************
620	* Internal Functions *
621	*******************************************************************************/
622	static int pgmR3InitPaging(PVM pVM);
623	static void pgmR3InitStats(PVM pVM);
624	static DECLCALLBACK(void) pgmR3PhysInfo(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
625	static DECLCALLBACK(void) pgmR3InfoMode(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
626	static DECLCALLBACK(void) pgmR3InfoCr3(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs);
627	static DECLCALLBACK(int) pgmR3RelocatePhysHandler(PAVLROGCPHYSNODECORE pNode, void *pvUser);
628	static DECLCALLBACK(int) pgmR3RelocateVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser);
629	static DECLCALLBACK(int) pgmR3RelocateHyperVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser);
630	#ifdef VBOX_STRICT
631	static DECLCALLBACK(void) pgmR3ResetNoMorePhysWritesFlag(PVM pVM, VMSTATE enmState, VMSTATE enmOldState, void *pvUser);
632	#endif
633	static int pgmR3ModeDataInit(PVM pVM, bool fResolveGCAndR0);
634	static void pgmR3ModeDataSwitch(PVM pVM, PVMCPU pVCpu, PGMMODE enmShw, PGMMODE enmGst);
635	static PGMMODE pgmR3CalcShadowMode(PVM pVM, PGMMODE enmGuestMode, SUPPAGINGMODE enmHostMode, PGMMODE enmShadowMode, VMMSWITCHER *penmSwitcher);
636
637	#ifdef VBOX_WITH_DEBUGGER
638	/** @todo Convert the first two commands to 'info' items. */
639	static DECLCALLBACK(int) pgmR3CmdRam(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
640	static DECLCALLBACK(int) pgmR3CmdMap(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
641	static DECLCALLBACK(int) pgmR3CmdError(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
642	static DECLCALLBACK(int) pgmR3CmdSync(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
643	static DECLCALLBACK(int) pgmR3CmdSyncAlways(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
644	# ifdef VBOX_STRICT
645	static DECLCALLBACK(int) pgmR3CmdAssertCR3(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
646	# endif
647	static DECLCALLBACK(int) pgmR3CmdPhysToFile(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult);
648	#endif
649
650
651	/*******************************************************************************
652	* Global Variables *
653	*******************************************************************************/
654	#ifdef VBOX_WITH_DEBUGGER
655	/** Argument descriptors for '.pgmerror' and '.pgmerroroff'. */
656	static const DBGCVARDESC g_aPgmErrorArgs[] =
657	{
658	/* cTimesMin, cTimesMax, enmCategory, fFlags, pszName, pszDescription */
659	{ 0, 1, DBGCVAR_CAT_STRING, 0, "where", "Error injection location." },
660	};
661
662	static const DBGCVARDESC g_aPgmPhysToFileArgs[] =
663	{
664	/* cTimesMin, cTimesMax, enmCategory, fFlags, pszName, pszDescription */
665	{ 1, 1, DBGCVAR_CAT_STRING, 0, "file", "The file name." },
666	{ 0, 1, DBGCVAR_CAT_STRING, 0, "nozero", "If present, zero pages are skipped." },
667	};
668
669	/** Command descriptors. */
670	static const DBGCCMD g_aCmds[] =
671	{
672	/* pszCmd, cArgsMin, cArgsMax, paArgDesc, cArgDescs, pResultDesc, fFlags, pfnHandler pszSyntax, ....pszDescription */
673	{ "pgmram", 0, 0, NULL, 0, NULL, 0, pgmR3CmdRam, "", "Display the ram ranges." },
674	{ "pgmmap", 0, 0, NULL, 0, NULL, 0, pgmR3CmdMap, "", "Display the mapping ranges." },
675	{ "pgmsync", 0, 0, NULL, 0, NULL, 0, pgmR3CmdSync, "", "Sync the CR3 page." },
676	{ "pgmerror", 0, 1, &g_aPgmErrorArgs[0], 1, NULL, 0, pgmR3CmdError, "", "Enables inject runtime of errors into parts of PGM." },
677	{ "pgmerroroff", 0, 1, &g_aPgmErrorArgs[0], 1, NULL, 0, pgmR3CmdError, "", "Disables inject runtime errors into parts of PGM." },
678	#ifdef VBOX_STRICT
679	{ "pgmassertcr3", 0, 0, NULL, 0, NULL, 0, pgmR3CmdAssertCR3, "", "Check the shadow CR3 mapping." },
680	#endif
681	{ "pgmsyncalways", 0, 0, NULL, 0, NULL, 0, pgmR3CmdSyncAlways, "", "Toggle permanent CR3 syncing." },
682	{ "pgmphystofile", 1, 2, &g_aPgmPhysToFileArgs[0], 2, NULL, 0, pgmR3CmdPhysToFile, "", "Save the physical memory to file." },
683	};
684	#endif
685
686
687
688
689	/*
690	* Shadow - 32-bit mode
691	*/
692	#define PGM_SHW_TYPE PGM_TYPE_32BIT
693	#define PGM_SHW_NAME(name) PGM_SHW_NAME_32BIT(name)
694	#define PGM_SHW_NAME_RC_STR(name) PGM_SHW_NAME_RC_32BIT_STR(name)
695	#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_32BIT_STR(name)
696	#include "PGMShw.h"
697
698	/* Guest - real mode */
699	#define PGM_GST_TYPE PGM_TYPE_REAL
700	#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
701	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_REAL_STR(name)
702	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
703	#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_REAL(name)
704	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_32BIT_REAL_STR(name)
705	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_REAL_STR(name)
706	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_PHYS
707	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_32BIT_PD_PHYS
708	#include "PGMBth.h"
709	#include "PGMGstDefs.h"
710	#include "PGMGst.h"
711	#undef BTH_PGMPOOLKIND_PT_FOR_PT
712	#undef BTH_PGMPOOLKIND_ROOT
713	#undef PGM_BTH_NAME
714	#undef PGM_BTH_NAME_RC_STR
715	#undef PGM_BTH_NAME_R0_STR
716	#undef PGM_GST_TYPE
717	#undef PGM_GST_NAME
718	#undef PGM_GST_NAME_RC_STR
719	#undef PGM_GST_NAME_R0_STR
720
721	/* Guest - protected mode */
722	#define PGM_GST_TYPE PGM_TYPE_PROT
723	#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
724	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PROT_STR(name)
725	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
726	#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_PROT(name)
727	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_32BIT_PROT_STR(name)
728	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_PROT_STR(name)
729	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_PHYS
730	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_32BIT_PD_PHYS
731	#include "PGMBth.h"
732	#include "PGMGstDefs.h"
733	#include "PGMGst.h"
734	#undef BTH_PGMPOOLKIND_PT_FOR_PT
735	#undef BTH_PGMPOOLKIND_ROOT
736	#undef PGM_BTH_NAME
737	#undef PGM_BTH_NAME_RC_STR
738	#undef PGM_BTH_NAME_R0_STR
739	#undef PGM_GST_TYPE
740	#undef PGM_GST_NAME
741	#undef PGM_GST_NAME_RC_STR
742	#undef PGM_GST_NAME_R0_STR
743
744	/* Guest - 32-bit mode */
745	#define PGM_GST_TYPE PGM_TYPE_32BIT
746	#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
747	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_32BIT_STR(name)
748	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
749	#define PGM_BTH_NAME(name) PGM_BTH_NAME_32BIT_32BIT(name)
750	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_32BIT_32BIT_STR(name)
751	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_32BIT_32BIT_STR(name)
752	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_32BIT_PT_FOR_32BIT_PT
753	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_32BIT_PT_FOR_32BIT_4MB
754	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_32BIT_PD
755	#include "PGMBth.h"
756	#include "PGMGstDefs.h"
757	#include "PGMGst.h"
758	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
759	#undef BTH_PGMPOOLKIND_PT_FOR_PT
760	#undef BTH_PGMPOOLKIND_ROOT
761	#undef PGM_BTH_NAME
762	#undef PGM_BTH_NAME_RC_STR
763	#undef PGM_BTH_NAME_R0_STR
764	#undef PGM_GST_TYPE
765	#undef PGM_GST_NAME
766	#undef PGM_GST_NAME_RC_STR
767	#undef PGM_GST_NAME_R0_STR
768
769	#undef PGM_SHW_TYPE
770	#undef PGM_SHW_NAME
771	#undef PGM_SHW_NAME_RC_STR
772	#undef PGM_SHW_NAME_R0_STR
773
774
775	/*
776	* Shadow - PAE mode
777	*/
778	#define PGM_SHW_TYPE PGM_TYPE_PAE
779	#define PGM_SHW_NAME(name) PGM_SHW_NAME_PAE(name)
780	#define PGM_SHW_NAME_RC_STR(name) PGM_SHW_NAME_RC_PAE_STR(name)
781	#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_PAE_STR(name)
782	#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_REAL(name)
783	#include "PGMShw.h"
784
785	/* Guest - real mode */
786	#define PGM_GST_TYPE PGM_TYPE_REAL
787	#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
788	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_REAL_STR(name)
789	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
790	#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_REAL(name)
791	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_PAE_REAL_STR(name)
792	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_REAL_STR(name)
793	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
794	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_PAE_PDPT_PHYS
795	#include "PGMGstDefs.h"
796	#include "PGMBth.h"
797	#undef BTH_PGMPOOLKIND_PT_FOR_PT
798	#undef BTH_PGMPOOLKIND_ROOT
799	#undef PGM_BTH_NAME
800	#undef PGM_BTH_NAME_RC_STR
801	#undef PGM_BTH_NAME_R0_STR
802	#undef PGM_GST_TYPE
803	#undef PGM_GST_NAME
804	#undef PGM_GST_NAME_RC_STR
805	#undef PGM_GST_NAME_R0_STR
806
807	/* Guest - protected mode */
808	#define PGM_GST_TYPE PGM_TYPE_PROT
809	#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
810	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PROT_STR(name)
811	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
812	#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_PROT(name)
813	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_PAE_PROT_STR(name)
814	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_PROT_STR(name)
815	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
816	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_PAE_PDPT_PHYS
817	#include "PGMGstDefs.h"
818	#include "PGMBth.h"
819	#undef BTH_PGMPOOLKIND_PT_FOR_PT
820	#undef BTH_PGMPOOLKIND_ROOT
821	#undef PGM_BTH_NAME
822	#undef PGM_BTH_NAME_RC_STR
823	#undef PGM_BTH_NAME_R0_STR
824	#undef PGM_GST_TYPE
825	#undef PGM_GST_NAME
826	#undef PGM_GST_NAME_RC_STR
827	#undef PGM_GST_NAME_R0_STR
828
829	/* Guest - 32-bit mode */
830	#define PGM_GST_TYPE PGM_TYPE_32BIT
831	#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
832	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_32BIT_STR(name)
833	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
834	#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_32BIT(name)
835	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_PAE_32BIT_STR(name)
836	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_32BIT_STR(name)
837	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_32BIT_PT
838	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_32BIT_4MB
839	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_PAE_PDPT_FOR_32BIT
840	#include "PGMGstDefs.h"
841	#include "PGMBth.h"
842	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
843	#undef BTH_PGMPOOLKIND_PT_FOR_PT
844	#undef BTH_PGMPOOLKIND_ROOT
845	#undef PGM_BTH_NAME
846	#undef PGM_BTH_NAME_RC_STR
847	#undef PGM_BTH_NAME_R0_STR
848	#undef PGM_GST_TYPE
849	#undef PGM_GST_NAME
850	#undef PGM_GST_NAME_RC_STR
851	#undef PGM_GST_NAME_R0_STR
852
853	/* Guest - PAE mode */
854	#define PGM_GST_TYPE PGM_TYPE_PAE
855	#define PGM_GST_NAME(name) PGM_GST_NAME_PAE(name)
856	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PAE_STR(name)
857	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PAE_STR(name)
858	#define PGM_BTH_NAME(name) PGM_BTH_NAME_PAE_PAE(name)
859	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_PAE_PAE_STR(name)
860	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_PAE_PAE_STR(name)
861	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
862	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
863	#define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_PAE_PDPT
864	#include "PGMBth.h"
865	#include "PGMGstDefs.h"
866	#include "PGMGst.h"
867	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
868	#undef BTH_PGMPOOLKIND_PT_FOR_PT
869	#undef BTH_PGMPOOLKIND_ROOT
870	#undef PGM_BTH_NAME
871	#undef PGM_BTH_NAME_RC_STR
872	#undef PGM_BTH_NAME_R0_STR
873	#undef PGM_GST_TYPE
874	#undef PGM_GST_NAME
875	#undef PGM_GST_NAME_RC_STR
876	#undef PGM_GST_NAME_R0_STR
877
878	#undef PGM_SHW_TYPE
879	#undef PGM_SHW_NAME
880	#undef PGM_SHW_NAME_RC_STR
881	#undef PGM_SHW_NAME_R0_STR
882
883
884	/*
885	* Shadow - AMD64 mode
886	*/
887	#define PGM_SHW_TYPE PGM_TYPE_AMD64
888	#define PGM_SHW_NAME(name) PGM_SHW_NAME_AMD64(name)
889	#define PGM_SHW_NAME_RC_STR(name) PGM_SHW_NAME_RC_AMD64_STR(name)
890	#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_AMD64_STR(name)
891	#include "PGMShw.h"
892
893	#ifdef VBOX_WITH_64_BITS_GUESTS
894	/* Guest - AMD64 mode */
895	# define PGM_GST_TYPE PGM_TYPE_AMD64
896	# define PGM_GST_NAME(name) PGM_GST_NAME_AMD64(name)
897	# define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_AMD64_STR(name)
898	# define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_AMD64_STR(name)
899	# define PGM_BTH_NAME(name) PGM_BTH_NAME_AMD64_AMD64(name)
900	# define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_AMD64_AMD64_STR(name)
901	# define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_AMD64_AMD64_STR(name)
902	# define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
903	# define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
904	# define BTH_PGMPOOLKIND_ROOT PGMPOOLKIND_64BIT_PML4
905	# include "PGMBth.h"
906	# include "PGMGstDefs.h"
907	# include "PGMGst.h"
908	# undef BTH_PGMPOOLKIND_PT_FOR_BIG
909	# undef BTH_PGMPOOLKIND_PT_FOR_PT
910	# undef BTH_PGMPOOLKIND_ROOT
911	# undef PGM_BTH_NAME
912	# undef PGM_BTH_NAME_RC_STR
913	# undef PGM_BTH_NAME_R0_STR
914	# undef PGM_GST_TYPE
915	# undef PGM_GST_NAME
916	# undef PGM_GST_NAME_RC_STR
917	# undef PGM_GST_NAME_R0_STR
918	#endif /* VBOX_WITH_64_BITS_GUESTS */
919
920	#undef PGM_SHW_TYPE
921	#undef PGM_SHW_NAME
922	#undef PGM_SHW_NAME_RC_STR
923	#undef PGM_SHW_NAME_R0_STR
924
925
926	/*
927	* Shadow - Nested paging mode
928	*/
929	#define PGM_SHW_TYPE PGM_TYPE_NESTED
930	#define PGM_SHW_NAME(name) PGM_SHW_NAME_NESTED(name)
931	#define PGM_SHW_NAME_RC_STR(name) PGM_SHW_NAME_RC_NESTED_STR(name)
932	#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_NESTED_STR(name)
933	#include "PGMShw.h"
934
935	/* Guest - real mode */
936	#define PGM_GST_TYPE PGM_TYPE_REAL
937	#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
938	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_REAL_STR(name)
939	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
940	#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_REAL(name)
941	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_NESTED_REAL_STR(name)
942	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_REAL_STR(name)
943	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
944	#include "PGMGstDefs.h"
945	#include "PGMBth.h"
946	#undef BTH_PGMPOOLKIND_PT_FOR_PT
947	#undef PGM_BTH_NAME
948	#undef PGM_BTH_NAME_RC_STR
949	#undef PGM_BTH_NAME_R0_STR
950	#undef PGM_GST_TYPE
951	#undef PGM_GST_NAME
952	#undef PGM_GST_NAME_RC_STR
953	#undef PGM_GST_NAME_R0_STR
954
955	/* Guest - protected mode */
956	#define PGM_GST_TYPE PGM_TYPE_PROT
957	#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
958	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PROT_STR(name)
959	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
960	#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_PROT(name)
961	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_NESTED_PROT_STR(name)
962	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_PROT_STR(name)
963	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
964	#include "PGMGstDefs.h"
965	#include "PGMBth.h"
966	#undef BTH_PGMPOOLKIND_PT_FOR_PT
967	#undef PGM_BTH_NAME
968	#undef PGM_BTH_NAME_RC_STR
969	#undef PGM_BTH_NAME_R0_STR
970	#undef PGM_GST_TYPE
971	#undef PGM_GST_NAME
972	#undef PGM_GST_NAME_RC_STR
973	#undef PGM_GST_NAME_R0_STR
974
975	/* Guest - 32-bit mode */
976	#define PGM_GST_TYPE PGM_TYPE_32BIT
977	#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
978	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_32BIT_STR(name)
979	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
980	#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_32BIT(name)
981	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_NESTED_32BIT_STR(name)
982	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_32BIT_STR(name)
983	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_32BIT_PT
984	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_32BIT_4MB
985	#include "PGMGstDefs.h"
986	#include "PGMBth.h"
987	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
988	#undef BTH_PGMPOOLKIND_PT_FOR_PT
989	#undef PGM_BTH_NAME
990	#undef PGM_BTH_NAME_RC_STR
991	#undef PGM_BTH_NAME_R0_STR
992	#undef PGM_GST_TYPE
993	#undef PGM_GST_NAME
994	#undef PGM_GST_NAME_RC_STR
995	#undef PGM_GST_NAME_R0_STR
996
997	/* Guest - PAE mode */
998	#define PGM_GST_TYPE PGM_TYPE_PAE
999	#define PGM_GST_NAME(name) PGM_GST_NAME_PAE(name)
1000	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PAE_STR(name)
1001	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PAE_STR(name)
1002	#define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_PAE(name)
1003	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_NESTED_PAE_STR(name)
1004	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_PAE_STR(name)
1005	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
1006	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
1007	#include "PGMGstDefs.h"
1008	#include "PGMBth.h"
1009	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
1010	#undef BTH_PGMPOOLKIND_PT_FOR_PT
1011	#undef PGM_BTH_NAME
1012	#undef PGM_BTH_NAME_RC_STR
1013	#undef PGM_BTH_NAME_R0_STR
1014	#undef PGM_GST_TYPE
1015	#undef PGM_GST_NAME
1016	#undef PGM_GST_NAME_RC_STR
1017	#undef PGM_GST_NAME_R0_STR
1018
1019	#ifdef VBOX_WITH_64_BITS_GUESTS
1020	/* Guest - AMD64 mode */
1021	# define PGM_GST_TYPE PGM_TYPE_AMD64
1022	# define PGM_GST_NAME(name) PGM_GST_NAME_AMD64(name)
1023	# define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_AMD64_STR(name)
1024	# define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_AMD64_STR(name)
1025	# define PGM_BTH_NAME(name) PGM_BTH_NAME_NESTED_AMD64(name)
1026	# define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_NESTED_AMD64_STR(name)
1027	# define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_NESTED_AMD64_STR(name)
1028	# define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
1029	# define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
1030	# include "PGMGstDefs.h"
1031	# include "PGMBth.h"
1032	# undef BTH_PGMPOOLKIND_PT_FOR_BIG
1033	# undef BTH_PGMPOOLKIND_PT_FOR_PT
1034	# undef PGM_BTH_NAME
1035	# undef PGM_BTH_NAME_RC_STR
1036	# undef PGM_BTH_NAME_R0_STR
1037	# undef PGM_GST_TYPE
1038	# undef PGM_GST_NAME
1039	# undef PGM_GST_NAME_RC_STR
1040	# undef PGM_GST_NAME_R0_STR
1041	#endif /* VBOX_WITH_64_BITS_GUESTS */
1042
1043	#undef PGM_SHW_TYPE
1044	#undef PGM_SHW_NAME
1045	#undef PGM_SHW_NAME_RC_STR
1046	#undef PGM_SHW_NAME_R0_STR
1047
1048
1049	/*
1050	* Shadow - EPT
1051	*/
1052	#define PGM_SHW_TYPE PGM_TYPE_EPT
1053	#define PGM_SHW_NAME(name) PGM_SHW_NAME_EPT(name)
1054	#define PGM_SHW_NAME_RC_STR(name) PGM_SHW_NAME_RC_EPT_STR(name)
1055	#define PGM_SHW_NAME_R0_STR(name) PGM_SHW_NAME_R0_EPT_STR(name)
1056	#include "PGMShw.h"
1057
1058	/* Guest - real mode */
1059	#define PGM_GST_TYPE PGM_TYPE_REAL
1060	#define PGM_GST_NAME(name) PGM_GST_NAME_REAL(name)
1061	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_REAL_STR(name)
1062	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_REAL_STR(name)
1063	#define PGM_BTH_NAME(name) PGM_BTH_NAME_EPT_REAL(name)
1064	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_EPT_REAL_STR(name)
1065	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_EPT_REAL_STR(name)
1066	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
1067	#include "PGMGstDefs.h"
1068	#include "PGMBth.h"
1069	#undef BTH_PGMPOOLKIND_PT_FOR_PT
1070	#undef PGM_BTH_NAME
1071	#undef PGM_BTH_NAME_RC_STR
1072	#undef PGM_BTH_NAME_R0_STR
1073	#undef PGM_GST_TYPE
1074	#undef PGM_GST_NAME
1075	#undef PGM_GST_NAME_RC_STR
1076	#undef PGM_GST_NAME_R0_STR
1077
1078	/* Guest - protected mode */
1079	#define PGM_GST_TYPE PGM_TYPE_PROT
1080	#define PGM_GST_NAME(name) PGM_GST_NAME_PROT(name)
1081	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PROT_STR(name)
1082	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PROT_STR(name)
1083	#define PGM_BTH_NAME(name) PGM_BTH_NAME_EPT_PROT(name)
1084	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_EPT_PROT_STR(name)
1085	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_EPT_PROT_STR(name)
1086	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PHYS
1087	#include "PGMGstDefs.h"
1088	#include "PGMBth.h"
1089	#undef BTH_PGMPOOLKIND_PT_FOR_PT
1090	#undef PGM_BTH_NAME
1091	#undef PGM_BTH_NAME_RC_STR
1092	#undef PGM_BTH_NAME_R0_STR
1093	#undef PGM_GST_TYPE
1094	#undef PGM_GST_NAME
1095	#undef PGM_GST_NAME_RC_STR
1096	#undef PGM_GST_NAME_R0_STR
1097
1098	/* Guest - 32-bit mode */
1099	#define PGM_GST_TYPE PGM_TYPE_32BIT
1100	#define PGM_GST_NAME(name) PGM_GST_NAME_32BIT(name)
1101	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_32BIT_STR(name)
1102	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_32BIT_STR(name)
1103	#define PGM_BTH_NAME(name) PGM_BTH_NAME_EPT_32BIT(name)
1104	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_EPT_32BIT_STR(name)
1105	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_EPT_32BIT_STR(name)
1106	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_32BIT_PT
1107	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_32BIT_4MB
1108	#include "PGMGstDefs.h"
1109	#include "PGMBth.h"
1110	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
1111	#undef BTH_PGMPOOLKIND_PT_FOR_PT
1112	#undef PGM_BTH_NAME
1113	#undef PGM_BTH_NAME_RC_STR
1114	#undef PGM_BTH_NAME_R0_STR
1115	#undef PGM_GST_TYPE
1116	#undef PGM_GST_NAME
1117	#undef PGM_GST_NAME_RC_STR
1118	#undef PGM_GST_NAME_R0_STR
1119
1120	/* Guest - PAE mode */
1121	#define PGM_GST_TYPE PGM_TYPE_PAE
1122	#define PGM_GST_NAME(name) PGM_GST_NAME_PAE(name)
1123	#define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_PAE_STR(name)
1124	#define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_PAE_STR(name)
1125	#define PGM_BTH_NAME(name) PGM_BTH_NAME_EPT_PAE(name)
1126	#define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_EPT_PAE_STR(name)
1127	#define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_EPT_PAE_STR(name)
1128	#define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
1129	#define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
1130	#include "PGMGstDefs.h"
1131	#include "PGMBth.h"
1132	#undef BTH_PGMPOOLKIND_PT_FOR_BIG
1133	#undef BTH_PGMPOOLKIND_PT_FOR_PT
1134	#undef PGM_BTH_NAME
1135	#undef PGM_BTH_NAME_RC_STR
1136	#undef PGM_BTH_NAME_R0_STR
1137	#undef PGM_GST_TYPE
1138	#undef PGM_GST_NAME
1139	#undef PGM_GST_NAME_RC_STR
1140	#undef PGM_GST_NAME_R0_STR
1141
1142	#ifdef VBOX_WITH_64_BITS_GUESTS
1143	/* Guest - AMD64 mode */
1144	# define PGM_GST_TYPE PGM_TYPE_AMD64
1145	# define PGM_GST_NAME(name) PGM_GST_NAME_AMD64(name)
1146	# define PGM_GST_NAME_RC_STR(name) PGM_GST_NAME_RC_AMD64_STR(name)
1147	# define PGM_GST_NAME_R0_STR(name) PGM_GST_NAME_R0_AMD64_STR(name)
1148	# define PGM_BTH_NAME(name) PGM_BTH_NAME_EPT_AMD64(name)
1149	# define PGM_BTH_NAME_RC_STR(name) PGM_BTH_NAME_RC_EPT_AMD64_STR(name)
1150	# define PGM_BTH_NAME_R0_STR(name) PGM_BTH_NAME_R0_EPT_AMD64_STR(name)
1151	# define BTH_PGMPOOLKIND_PT_FOR_PT PGMPOOLKIND_PAE_PT_FOR_PAE_PT
1152	# define BTH_PGMPOOLKIND_PT_FOR_BIG PGMPOOLKIND_PAE_PT_FOR_PAE_2MB
1153	# include "PGMGstDefs.h"
1154	# include "PGMBth.h"
1155	# undef BTH_PGMPOOLKIND_PT_FOR_BIG
1156	# undef BTH_PGMPOOLKIND_PT_FOR_PT
1157	# undef PGM_BTH_NAME
1158	# undef PGM_BTH_NAME_RC_STR
1159	# undef PGM_BTH_NAME_R0_STR
1160	# undef PGM_GST_TYPE
1161	# undef PGM_GST_NAME
1162	# undef PGM_GST_NAME_RC_STR
1163	# undef PGM_GST_NAME_R0_STR
1164	#endif /* VBOX_WITH_64_BITS_GUESTS */
1165
1166	#undef PGM_SHW_TYPE
1167	#undef PGM_SHW_NAME
1168	#undef PGM_SHW_NAME_RC_STR
1169	#undef PGM_SHW_NAME_R0_STR
1170
1171
1172
1173	/**
1174	* Initiates the paging of VM.
1175	*
1176	* @returns VBox status code.
1177	* @param pVM Pointer to VM structure.
1178	*/
1179	VMMR3DECL(int) PGMR3Init(PVM pVM)
1180	{
1181	LogFlow(("PGMR3Init:\n"));
1182	PCFGMNODE pCfgPGM = CFGMR3GetChild(CFGMR3GetRoot(pVM), "/PGM");
1183	int rc;
1184
1185	/*
1186	* Assert alignment and sizes.
1187	*/
1188	AssertCompile(sizeof(pVM->pgm.s) <= sizeof(pVM->pgm.padding));
1189	AssertCompileMemberAlignment(PGM, CritSect, sizeof(uintptr_t));
1190
1191	/*
1192	* Init the structure.
1193	*/
1194	pVM->pgm.s.offVM = RT_OFFSETOF(VM, pgm.s);
1195	pVM->pgm.s.offVCpuPGM = RT_OFFSETOF(VMCPU, pgm.s);
1196
1197	/* Init the per-CPU part. */
1198	for (VMCPUID i = 0; i < pVM->cCpus; i++)
1199	{
1200	PVMCPU pVCpu = &pVM->aCpus[i];
1201	PPGMCPU pPGM = &pVCpu->pgm.s;
1202
1203	pPGM->offVM = (uintptr_t)&pVCpu->pgm.s - (uintptr_t)pVM;
1204	pPGM->offVCpu = RT_OFFSETOF(VMCPU, pgm.s);
1205	pPGM->offPGM = (uintptr_t)&pVCpu->pgm.s - (uintptr_t)&pVM->pgm.s;
1206
1207	pPGM->enmShadowMode = PGMMODE_INVALID;
1208	pPGM->enmGuestMode = PGMMODE_INVALID;
1209
1210	pPGM->GCPhysCR3 = NIL_RTGCPHYS;
1211
1212	pPGM->pGstPaePdptR3 = NULL;
1213	#ifndef VBOX_WITH_2X_4GB_ADDR_SPACE
1214	pPGM->pGstPaePdptR0 = NIL_RTR0PTR;
1215	#endif
1216	pPGM->pGstPaePdptRC = NIL_RTRCPTR;
1217	for (unsigned i = 0; i < RT_ELEMENTS(pVCpu->pgm.s.apGstPaePDsR3); i++)
1218	{
1219	pPGM->apGstPaePDsR3[i] = NULL;
1220	#ifndef VBOX_WITH_2X_4GB_ADDR_SPACE
1221	pPGM->apGstPaePDsR0[i] = NIL_RTR0PTR;
1222	#endif
1223	pPGM->apGstPaePDsRC[i] = NIL_RTRCPTR;
1224	pPGM->aGCPhysGstPaePDs[i] = NIL_RTGCPHYS;
1225	pPGM->aGCPhysGstPaePDsMonitored[i] = NIL_RTGCPHYS;
1226	}
1227
1228	pPGM->fA20Enabled = true;
1229	}
1230
1231	pVM->pgm.s.enmHostMode = SUPPAGINGMODE_INVALID;
1232	pVM->pgm.s.GCPhys4MBPSEMask = RT_BIT_64(32) - 1; /* default; checked later */
1233	pVM->pgm.s.GCPtrPrevRamRangeMapping = MM_HYPER_AREA_ADDRESS;
1234
1235	rc = CFGMR3QueryBoolDef(CFGMR3GetRoot(pVM), "RamPreAlloc", &pVM->pgm.s.fRamPreAlloc,
1236	#ifdef VBOX_WITH_PREALLOC_RAM_BY_DEFAULT
1237	true
1238	#else
1239	false
1240	#endif
1241	);
1242	AssertLogRelRCReturn(rc, rc);
1243
1244	#if HC_ARCH_BITS == 64 \|\| 1 /** @todo 4GB/32-bit: remove \|\| 1 later and adjust the limit. */
1245	rc = CFGMR3QueryU32Def(pCfgPGM, "MaxRing3Chunks", &pVM->pgm.s.ChunkR3Map.cMax, UINT32_MAX);
1246	#else
1247	rc = CFGMR3QueryU32Def(pCfgPGM, "MaxRing3Chunks", &pVM->pgm.s.ChunkR3Map.cMax, _1G / GMM_CHUNK_SIZE);
1248	#endif
1249	AssertLogRelRCReturn(rc, rc);
1250	for (uint32_t i = 0; i < RT_ELEMENTS(pVM->pgm.s.ChunkR3Map.Tlb.aEntries); i++)
1251	pVM->pgm.s.ChunkR3Map.Tlb.aEntries[i].idChunk = NIL_GMM_CHUNKID;
1252
1253	/*
1254	* Get the configured RAM size - to estimate saved state size.
1255	*/
1256	uint64_t cbRam;
1257	rc = CFGMR3QueryU64(CFGMR3GetRoot(pVM), "RamSize", &cbRam);
1258	if (rc == VERR_CFGM_VALUE_NOT_FOUND)
1259	cbRam = 0;
1260	else if (RT_SUCCESS(rc))
1261	{
1262	if (cbRam < PAGE_SIZE)
1263	cbRam = 0;
1264	cbRam = RT_ALIGN_64(cbRam, PAGE_SIZE);
1265	}
1266	else
1267	{
1268	AssertMsgFailed(("Configuration error: Failed to query integer \"RamSize\", rc=%Rrc.\n", rc));
1269	return rc;
1270	}
1271
1272	/*
1273	* Register callbacks, string formatters and the saved state data unit.
1274	*/
1275	#ifdef VBOX_STRICT
1276	VMR3AtStateRegister(pVM, pgmR3ResetNoMorePhysWritesFlag, NULL);
1277	#endif
1278	PGMRegisterStringFormatTypes();
1279
1280	rc = pgmR3InitSavedState(pVM, cbRam);
1281	if (RT_FAILURE(rc))
1282	return rc;
1283
1284	/*
1285	* Initialize the PGM critical section and flush the phys TLBs
1286	*/
1287	rc = PDMR3CritSectInit(pVM, &pVM->pgm.s.CritSect, "PGM");
1288	AssertRCReturn(rc, rc);
1289
1290	PGMR3PhysChunkInvalidateTLB(pVM);
1291	PGMPhysInvalidatePageMapTLB(pVM);
1292
1293	/*
1294	* For the time being we sport a full set of handy pages in addition to the base
1295	* memory to simplify things.
1296	*/
1297	rc = MMR3ReserveHandyPages(pVM, RT_ELEMENTS(pVM->pgm.s.aHandyPages)); /** @todo this should be changed to PGM_HANDY_PAGES_MIN but this needs proper testing... */
1298	AssertRCReturn(rc, rc);
1299
1300	/*
1301	* Trees
1302	*/
1303	rc = MMHyperAlloc(pVM, sizeof(PGMTREES), 0, MM_TAG_PGM, (void **)&pVM->pgm.s.pTreesR3);
1304	if (RT_SUCCESS(rc))
1305	{
1306	pVM->pgm.s.pTreesR0 = MMHyperR3ToR0(pVM, pVM->pgm.s.pTreesR3);
1307	pVM->pgm.s.pTreesRC = MMHyperR3ToRC(pVM, pVM->pgm.s.pTreesR3);
1308
1309	/*
1310	* Alocate the zero page.
1311	*/
1312	rc = MMHyperAlloc(pVM, PAGE_SIZE, PAGE_SIZE, MM_TAG_PGM, &pVM->pgm.s.pvZeroPgR3);
1313	}
1314	if (RT_SUCCESS(rc))
1315	{
1316	pVM->pgm.s.pvZeroPgRC = MMHyperR3ToRC(pVM, pVM->pgm.s.pvZeroPgR3);
1317	pVM->pgm.s.pvZeroPgR0 = MMHyperR3ToR0(pVM, pVM->pgm.s.pvZeroPgR3);
1318	pVM->pgm.s.HCPhysZeroPg = MMR3HyperHCVirt2HCPhys(pVM, pVM->pgm.s.pvZeroPgR3);
1319	AssertRelease(pVM->pgm.s.HCPhysZeroPg != NIL_RTHCPHYS);
1320
1321	/*
1322	* Init the paging.
1323	*/
1324	rc = pgmR3InitPaging(pVM);
1325	}
1326	if (RT_SUCCESS(rc))
1327	{
1328	/*
1329	* Init the page pool.
1330	*/
1331	rc = pgmR3PoolInit(pVM);
1332	}
1333	if (RT_SUCCESS(rc))
1334	{
1335	for (VMCPUID i = 0; i < pVM->cCpus; i++)
1336	{
1337	PVMCPU pVCpu = &pVM->aCpus[i];
1338	rc = PGMR3ChangeMode(pVM, pVCpu, PGMMODE_REAL);
1339	if (RT_FAILURE(rc))
1340	break;
1341	}
1342	}
1343
1344	if (RT_SUCCESS(rc))
1345	{
1346	/*
1347	* Info & statistics
1348	*/
1349	DBGFR3InfoRegisterInternal(pVM, "mode",
1350	"Shows the current paging mode. "
1351	"Recognizes 'all', 'guest', 'shadow' and 'host' as arguments, defaulting to 'all' if nothing's given.",
1352	pgmR3InfoMode);
1353	DBGFR3InfoRegisterInternal(pVM, "pgmcr3",
1354	"Dumps all the entries in the top level paging table. No arguments.",
1355	pgmR3InfoCr3);
1356	DBGFR3InfoRegisterInternal(pVM, "phys",
1357	"Dumps all the physical address ranges. No arguments.",
1358	pgmR3PhysInfo);
1359	DBGFR3InfoRegisterInternal(pVM, "handlers",
1360	"Dumps physical, virtual and hyper virtual handlers. "
1361	"Pass 'phys', 'virt', 'hyper' as argument if only one kind is wanted."
1362	"Add 'nost' if the statistics are unwanted, use together with 'all' or explicit selection.",
1363	pgmR3InfoHandlers);
1364	DBGFR3InfoRegisterInternal(pVM, "mappings",
1365	"Dumps guest mappings.",
1366	pgmR3MapInfo);
1367
1368	pgmR3InitStats(pVM);
1369
1370	#ifdef VBOX_WITH_DEBUGGER
1371	/*
1372	* Debugger commands.
1373	*/
1374	static bool s_fRegisteredCmds = false;
1375	if (!s_fRegisteredCmds)
1376	{
1377	int rc = DBGCRegisterCommands(&g_aCmds[0], RT_ELEMENTS(g_aCmds));
1378	if (RT_SUCCESS(rc))
1379	s_fRegisteredCmds = true;
1380	}
1381	#endif
1382	return VINF_SUCCESS;
1383	}
1384
1385	/* Almost no cleanup necessary, MM frees all memory. */
1386	PDMR3CritSectDelete(&pVM->pgm.s.CritSect);
1387
1388	return rc;
1389	}
1390
1391
1392	/**
1393	* Initializes the per-VCPU PGM.
1394	*
1395	* @returns VBox status code.
1396	* @param pVM The VM to operate on.
1397	*/
1398	VMMR3DECL(int) PGMR3InitCPU(PVM pVM)
1399	{
1400	LogFlow(("PGMR3InitCPU\n"));
1401	return VINF_SUCCESS;
1402	}
1403
1404
1405	/**
1406	* Init paging.
1407	*
1408	* Since we need to check what mode the host is operating in before we can choose
1409	* the right paging functions for the host we have to delay this until R0 has
1410	* been initialized.
1411	*
1412	* @returns VBox status code.
1413	* @param pVM VM handle.
1414	*/
1415	static int pgmR3InitPaging(PVM pVM)
1416	{
1417	/*
1418	* Force a recalculation of modes and switcher so everyone gets notified.
1419	*/
1420	for (VMCPUID i = 0; i < pVM->cCpus; i++)
1421	{
1422	PVMCPU pVCpu = &pVM->aCpus[i];
1423
1424	pVCpu->pgm.s.enmShadowMode = PGMMODE_INVALID;
1425	pVCpu->pgm.s.enmGuestMode = PGMMODE_INVALID;
1426	}
1427
1428	pVM->pgm.s.enmHostMode = SUPPAGINGMODE_INVALID;
1429
1430	/*
1431	* Allocate static mapping space for whatever the cr3 register
1432	* points to and in the case of PAE mode to the 4 PDs.
1433	*/
1434	int rc = MMR3HyperReserve(pVM, PAGE_SIZE * 5, "CR3 mapping", &pVM->pgm.s.GCPtrCR3Mapping);
1435	if (RT_FAILURE(rc))
1436	{
1437	AssertMsgFailed(("Failed to reserve two pages for cr mapping in HMA, rc=%Rrc\n", rc));
1438	return rc;
1439	}
1440	MMR3HyperReserve(pVM, PAGE_SIZE, "fence", NULL);
1441
1442	/*
1443	* Allocate pages for the three possible intermediate contexts
1444	* (AMD64, PAE and plain 32-Bit). We maintain all three contexts
1445	* for the sake of simplicity. The AMD64 uses the PAE for the
1446	* lower levels, making the total number of pages 11 (3 + 7 + 1).
1447	*
1448	* We assume that two page tables will be enought for the core code
1449	* mappings (HC virtual and identity).
1450	*/
1451	pVM->pgm.s.pInterPD = (PX86PD)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.pInterPD, VERR_NO_PAGE_MEMORY);
1452	pVM->pgm.s.apInterPTs[0] = (PX86PT)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.apInterPTs[0], VERR_NO_PAGE_MEMORY);
1453	pVM->pgm.s.apInterPTs[1] = (PX86PT)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.apInterPTs[1], VERR_NO_PAGE_MEMORY);
1454	pVM->pgm.s.apInterPaePTs[0] = (PX86PTPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePTs[0], VERR_NO_PAGE_MEMORY);
1455	pVM->pgm.s.apInterPaePTs[1] = (PX86PTPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePTs[1], VERR_NO_PAGE_MEMORY);
1456	pVM->pgm.s.apInterPaePDs[0] = (PX86PDPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePDs[0], VERR_NO_PAGE_MEMORY);
1457	pVM->pgm.s.apInterPaePDs[1] = (PX86PDPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePDs[1], VERR_NO_PAGE_MEMORY);
1458	pVM->pgm.s.apInterPaePDs[2] = (PX86PDPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePDs[2], VERR_NO_PAGE_MEMORY);
1459	pVM->pgm.s.apInterPaePDs[3] = (PX86PDPAE)MMR3PageAlloc(pVM); AssertReturn(pVM->pgm.s.apInterPaePDs[3], VERR_NO_PAGE_MEMORY);
1460	pVM->pgm.s.pInterPaePDPT = (PX86PDPT)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.pInterPaePDPT, VERR_NO_PAGE_MEMORY);
1461	pVM->pgm.s.pInterPaePDPT64 = (PX86PDPT)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.pInterPaePDPT64, VERR_NO_PAGE_MEMORY);
1462	pVM->pgm.s.pInterPaePML4 = (PX86PML4)MMR3PageAllocLow(pVM); AssertReturn(pVM->pgm.s.pInterPaePML4, VERR_NO_PAGE_MEMORY);
1463
1464	pVM->pgm.s.HCPhysInterPD = MMPage2Phys(pVM, pVM->pgm.s.pInterPD);
1465	AssertRelease(pVM->pgm.s.HCPhysInterPD != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPD & PAGE_OFFSET_MASK));
1466	pVM->pgm.s.HCPhysInterPaePDPT = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT);
1467	AssertRelease(pVM->pgm.s.HCPhysInterPaePDPT != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPaePDPT & PAGE_OFFSET_MASK));
1468	pVM->pgm.s.HCPhysInterPaePML4 = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePML4);
1469	AssertRelease(pVM->pgm.s.HCPhysInterPaePML4 != NIL_RTHCPHYS && !(pVM->pgm.s.HCPhysInterPaePML4 & PAGE_OFFSET_MASK) && pVM->pgm.s.HCPhysInterPaePML4 < 0xffffffff);
1470
1471	/*
1472	* Initialize the pages, setting up the PML4 and PDPT for repetitive 4GB action.
1473	*/
1474	ASMMemZeroPage(pVM->pgm.s.pInterPD);
1475	ASMMemZeroPage(pVM->pgm.s.apInterPTs[0]);
1476	ASMMemZeroPage(pVM->pgm.s.apInterPTs[1]);
1477
1478	ASMMemZeroPage(pVM->pgm.s.apInterPaePTs[0]);
1479	ASMMemZeroPage(pVM->pgm.s.apInterPaePTs[1]);
1480
1481	ASMMemZeroPage(pVM->pgm.s.pInterPaePDPT);
1482	for (unsigned i = 0; i < RT_ELEMENTS(pVM->pgm.s.apInterPaePDs); i++)
1483	{
1484	ASMMemZeroPage(pVM->pgm.s.apInterPaePDs[i]);
1485	pVM->pgm.s.pInterPaePDPT->a[i].u = X86_PDPE_P \| PGM_PLXFLAGS_PERMANENT
1486	\| MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[i]);
1487	}
1488
1489	for (unsigned i = 0; i < RT_ELEMENTS(pVM->pgm.s.pInterPaePDPT64->a); i++)
1490	{
1491	const unsigned iPD = i % RT_ELEMENTS(pVM->pgm.s.apInterPaePDs);
1492	pVM->pgm.s.pInterPaePDPT64->a[i].u = X86_PDPE_P \| X86_PDPE_RW \| X86_PDPE_US \| X86_PDPE_A \| PGM_PLXFLAGS_PERMANENT
1493	\| MMPage2Phys(pVM, pVM->pgm.s.apInterPaePDs[iPD]);
1494	}
1495
1496	RTHCPHYS HCPhysInterPaePDPT64 = MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT64);
1497	for (unsigned i = 0; i < RT_ELEMENTS(pVM->pgm.s.pInterPaePML4->a); i++)
1498	pVM->pgm.s.pInterPaePML4->a[i].u = X86_PML4E_P \| X86_PML4E_RW \| X86_PML4E_US \| X86_PML4E_A \| PGM_PLXFLAGS_PERMANENT
1499	\| HCPhysInterPaePDPT64;
1500
1501	/*
1502	* Initialize paging workers and mode from current host mode
1503	* and the guest running in real mode.
1504	*/
1505	pVM->pgm.s.enmHostMode = SUPR3GetPagingMode();
1506	switch (pVM->pgm.s.enmHostMode)
1507	{
1508	case SUPPAGINGMODE_32_BIT:
1509	case SUPPAGINGMODE_32_BIT_GLOBAL:
1510	case SUPPAGINGMODE_PAE:
1511	case SUPPAGINGMODE_PAE_GLOBAL:
1512	case SUPPAGINGMODE_PAE_NX:
1513	case SUPPAGINGMODE_PAE_GLOBAL_NX:
1514	break;
1515
1516	case SUPPAGINGMODE_AMD64:
1517	case SUPPAGINGMODE_AMD64_GLOBAL:
1518	case SUPPAGINGMODE_AMD64_NX:
1519	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
1520	#ifndef VBOX_WITH_HYBRID_32BIT_KERNEL
1521	if (ARCH_BITS != 64)
1522	{
1523	AssertMsgFailed(("Host mode %d (64-bit) is not supported by non-64bit builds\n", pVM->pgm.s.enmHostMode));
1524	LogRel(("Host mode %d (64-bit) is not supported by non-64bit builds\n", pVM->pgm.s.enmHostMode));
1525	return VERR_PGM_UNSUPPORTED_HOST_PAGING_MODE;
1526	}
1527	#endif
1528	break;
1529	default:
1530	AssertMsgFailed(("Host mode %d is not supported\n", pVM->pgm.s.enmHostMode));
1531	return VERR_PGM_UNSUPPORTED_HOST_PAGING_MODE;
1532	}
1533	rc = pgmR3ModeDataInit(pVM, false /* don't resolve GC and R0 syms yet */);
1534	if (RT_SUCCESS(rc))
1535	{
1536	LogFlow(("pgmR3InitPaging: returns successfully\n"));
1537	#if HC_ARCH_BITS == 64
1538	LogRel(("Debug: HCPhysInterPD=%RHp HCPhysInterPaePDPT=%RHp HCPhysInterPaePML4=%RHp\n",
1539	pVM->pgm.s.HCPhysInterPD, pVM->pgm.s.HCPhysInterPaePDPT, pVM->pgm.s.HCPhysInterPaePML4));
1540	LogRel(("Debug: apInterPTs={%RHp,%RHp} apInterPaePTs={%RHp,%RHp} apInterPaePDs={%RHp,%RHp,%RHp,%RHp} pInterPaePDPT64=%RHp\n",
1541	MMPage2Phys(pVM, pVM->pgm.s.apInterPTs[0]), MMPage2Phys(pVM, pVM->pgm.s.apInterPTs[1]),
1542	MMPage2Phys(pVM, pVM->pgm.s.apInterPaePTs[0]), MMPage2Phys(pVM, pVM->pgm.s.apInterPaePTs[1]),
1543	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]),
1544	MMPage2Phys(pVM, pVM->pgm.s.pInterPaePDPT64)));
1545	#endif
1546
1547	return VINF_SUCCESS;
1548	}
1549
1550	LogFlow(("pgmR3InitPaging: returns %Rrc\n", rc));
1551	return rc;
1552	}
1553
1554
1555	/**
1556	* Init statistics
1557	*/
1558	static void pgmR3InitStats(PVM pVM)
1559	{
1560	PPGM pPGM = &pVM->pgm.s;
1561	int rc;
1562
1563	/* Common - misc variables */
1564	STAM_REL_REG(pVM, &pPGM->cAllPages, STAMTYPE_U32, "/PGM/Page/cAllPages", STAMUNIT_OCCURENCES, "The total number of pages.");
1565	STAM_REL_REG(pVM, &pPGM->cPrivatePages, STAMTYPE_U32, "/PGM/Page/cPrivatePages", STAMUNIT_OCCURENCES, "The number of private pages.");
1566	STAM_REL_REG(pVM, &pPGM->cSharedPages, STAMTYPE_U32, "/PGM/Page/cSharedPages", STAMUNIT_OCCURENCES, "The number of shared pages.");
1567	STAM_REL_REG(pVM, &pPGM->cZeroPages, STAMTYPE_U32, "/PGM/Page/cZeroPages", STAMUNIT_OCCURENCES, "The number of zero backed pages.");
1568	STAM_REL_REG(pVM, &pPGM->cMonitoredPages, STAMTYPE_U32, "/PGM/Page/cMonitoredPages", STAMUNIT_OCCURENCES, "The number of write monitored pages.");
1569	STAM_REL_REG(pVM, &pPGM->cWrittenToPages, STAMTYPE_U32, "/PGM/Page/cWrittenToPages", STAMUNIT_OCCURENCES, "The number of previously write monitored pages that have been written to.");
1570	STAM_REL_REG(pVM, &pPGM->cWriteLockedPages, STAMTYPE_U32, "/PGM/Page/cWriteLockedPages", STAMUNIT_OCCURENCES, "The number of write(/read) locked pages.");
1571	STAM_REL_REG(pVM, &pPGM->cReadLockedPages, STAMTYPE_U32, "/PGM/Page/cReadLockedPages", STAMUNIT_OCCURENCES, "The number of read (only) locked pages.");
1572	STAM_REL_REG(pVM, &pPGM->cHandyPages, STAMTYPE_U32, "/PGM/Page/cHandyPages", STAMUNIT_OCCURENCES, "The number of handy pages (not included in cAllPages).");
1573	STAM_REL_REG(pVM, &pPGM->cRelocations, STAMTYPE_COUNTER, "/PGM/cRelocations", STAMUNIT_OCCURENCES, "Number of hypervisor relocations.");
1574	STAM_REL_REG(pVM, &pPGM->ChunkR3Map.c, STAMTYPE_U32, "/PGM/ChunkR3Map/c", STAMUNIT_OCCURENCES, "Number of mapped chunks.");
1575	STAM_REL_REG(pVM, &pPGM->ChunkR3Map.cMax, STAMTYPE_U32, "/PGM/ChunkR3Map/cMax", STAMUNIT_OCCURENCES, "Maximum number of mapped chunks.");
1576
1577	#ifdef VBOX_WITH_STATISTICS
1578
1579	# define PGM_REG_COUNTER(a, b, c) \
1580	rc = STAMR3RegisterF(pVM, a, STAMTYPE_COUNTER, STAMVISIBILITY_ALWAYS, STAMUNIT_OCCURENCES, c, b); \
1581	AssertRC(rc);
1582
1583	# define PGM_REG_COUNTER_BYTES(a, b, c) \
1584	rc = STAMR3RegisterF(pVM, a, STAMTYPE_COUNTER, STAMVISIBILITY_ALWAYS, STAMUNIT_BYTES, c, b); \
1585	AssertRC(rc);
1586
1587	# define PGM_REG_PROFILE(a, b, c) \
1588	rc = STAMR3RegisterF(pVM, a, STAMTYPE_PROFILE, STAMVISIBILITY_ALWAYS, STAMUNIT_TICKS_PER_CALL, c, b); \
1589	AssertRC(rc);
1590
1591	PGM_REG_COUNTER(&pPGM->StatR3DetectedConflicts, "/PGM/R3/DetectedConflicts", "The number of times PGMR3CheckMappingConflicts() detected a conflict.");
1592	PGM_REG_PROFILE(&pPGM->StatR3ResolveConflict, "/PGM/R3/ResolveConflict", "pgmR3SyncPTResolveConflict() profiling (includes the entire relocation).");
1593	PGM_REG_COUNTER(&pPGM->StatR3PhysRead, "/PGM/R3/Phys/Read", "The number of times PGMPhysRead was called.");
1594	PGM_REG_COUNTER_BYTES(&pPGM->StatR3PhysReadBytes, "/PGM/R3/Phys/Read/Bytes", "The number of bytes read by PGMPhysRead.");
1595	PGM_REG_COUNTER(&pPGM->StatR3PhysWrite, "/PGM/R3/Phys/Write", "The number of times PGMPhysWrite was called.");
1596	PGM_REG_COUNTER_BYTES(&pPGM->StatR3PhysWriteBytes, "/PGM/R3/Phys/Write/Bytes", "The number of bytes written by PGMPhysWrite.");
1597	PGM_REG_COUNTER(&pPGM->StatR3PhysSimpleRead, "/PGM/R3/Phys/Simple/Read", "The number of times PGMPhysSimpleReadGCPtr was called.");
1598	PGM_REG_COUNTER_BYTES(&pPGM->StatR3PhysSimpleReadBytes, "/PGM/R3/Phys/Simple/Read/Bytes", "The number of bytes read by PGMPhysSimpleReadGCPtr.");
1599	PGM_REG_COUNTER(&pPGM->StatR3PhysSimpleWrite, "/PGM/R3/Phys/Simple/Write", "The number of times PGMPhysSimpleWriteGCPtr was called.");
1600	PGM_REG_COUNTER_BYTES(&pPGM->StatR3PhysSimpleWriteBytes, "/PGM/R3/Phys/Simple/Write/Bytes", "The number of bytes written by PGMPhysSimpleWriteGCPtr.");
1601
1602	PGM_REG_COUNTER(&pPGM->StatRZChunkR3MapTlbHits, "/PGM/ChunkR3Map/TlbHitsRZ", "TLB hits.");
1603	PGM_REG_COUNTER(&pPGM->StatRZChunkR3MapTlbMisses, "/PGM/ChunkR3Map/TlbMissesRZ", "TLB misses.");
1604	PGM_REG_COUNTER(&pPGM->StatRZPageMapTlbHits, "/PGM/RZ/Page/MapTlbHits", "TLB hits.");
1605	PGM_REG_COUNTER(&pPGM->StatRZPageMapTlbMisses, "/PGM/RZ/Page/MapTlbMisses", "TLB misses.");
1606	PGM_REG_COUNTER(&pPGM->StatR3ChunkR3MapTlbHits, "/PGM/ChunkR3Map/TlbHitsR3", "TLB hits.");
1607	PGM_REG_COUNTER(&pPGM->StatR3ChunkR3MapTlbMisses, "/PGM/ChunkR3Map/TlbMissesR3", "TLB misses.");
1608	PGM_REG_COUNTER(&pPGM->StatR3PageMapTlbHits, "/PGM/R3/Page/MapTlbHits", "TLB hits.");
1609	PGM_REG_COUNTER(&pPGM->StatR3PageMapTlbMisses, "/PGM/R3/Page/MapTlbMisses", "TLB misses.");
1610	PGM_REG_COUNTER(&pPGM->StatPageMapTlbFlushes, "/PGM/R3/Page/MapTlbFlushes", "TLB flushes (all contexts).");
1611	PGM_REG_COUNTER(&pPGM->StatPageMapTlbFlushEntry, "/PGM/R3/Page/MapTlbFlushEntry", "TLB entry flushes (all contexts).");
1612
1613	PGM_REG_PROFILE(&pPGM->StatRZSyncCR3HandlerVirtualUpdate, "/PGM/RZ/SyncCR3/Handlers/VirtualUpdate", "Profiling of the virtual handler updates.");
1614	PGM_REG_PROFILE(&pPGM->StatRZSyncCR3HandlerVirtualReset, "/PGM/RZ/SyncCR3/Handlers/VirtualReset", "Profiling of the virtual handler resets.");
1615	PGM_REG_PROFILE(&pPGM->StatR3SyncCR3HandlerVirtualUpdate, "/PGM/R3/SyncCR3/Handlers/VirtualUpdate", "Profiling of the virtual handler updates.");
1616	PGM_REG_PROFILE(&pPGM->StatR3SyncCR3HandlerVirtualReset, "/PGM/R3/SyncCR3/Handlers/VirtualReset", "Profiling of the virtual handler resets.");
1617
1618	PGM_REG_COUNTER(&pPGM->StatRZPhysHandlerReset, "/PGM/RZ/PhysHandlerReset", "The number of times PGMHandlerPhysicalReset is called.");
1619	PGM_REG_COUNTER(&pPGM->StatR3PhysHandlerReset, "/PGM/R3/PhysHandlerReset", "The number of times PGMHandlerPhysicalReset is called.");
1620	PGM_REG_PROFILE(&pPGM->StatRZVirtHandlerSearchByPhys, "/PGM/RZ/VirtHandlerSearchByPhys", "Profiling of pgmHandlerVirtualFindByPhysAddr.");
1621	PGM_REG_PROFILE(&pPGM->StatR3VirtHandlerSearchByPhys, "/PGM/R3/VirtHandlerSearchByPhys", "Profiling of pgmHandlerVirtualFindByPhysAddr.");
1622
1623	PGM_REG_COUNTER(&pPGM->StatRZPageReplaceShared, "/PGM/RZ/Page/ReplacedShared", "Times a shared page was replaced.");
1624	PGM_REG_COUNTER(&pPGM->StatRZPageReplaceZero, "/PGM/RZ/Page/ReplacedZero", "Times the zero page was replaced.");
1625	/// @todo PGM_REG_COUNTER(&pPGM->StatRZPageHandyAllocs, "/PGM/RZ/Page/HandyAllocs", "Number of times we've allocated more handy pages.");
1626	PGM_REG_COUNTER(&pPGM->StatR3PageReplaceShared, "/PGM/R3/Page/ReplacedShared", "Times a shared page was replaced.");
1627	PGM_REG_COUNTER(&pPGM->StatR3PageReplaceZero, "/PGM/R3/Page/ReplacedZero", "Times the zero page was replaced.");
1628	/// @todo PGM_REG_COUNTER(&pPGM->StatR3PageHandyAllocs, "/PGM/R3/Page/HandyAllocs", "Number of times we've allocated more handy pages.");
1629
1630	PGM_REG_COUNTER(&pPGM->StatRZPhysRead, "/PGM/RZ/Phys/Read", "The number of times PGMPhysRead was called.");
1631	PGM_REG_COUNTER_BYTES(&pPGM->StatRZPhysReadBytes, "/PGM/RZ/Phys/Read/Bytes", "The number of bytes read by PGMPhysRead.");
1632	PGM_REG_COUNTER(&pPGM->StatRZPhysWrite, "/PGM/RZ/Phys/Write", "The number of times PGMPhysWrite was called.");
1633	PGM_REG_COUNTER_BYTES(&pPGM->StatRZPhysWriteBytes, "/PGM/RZ/Phys/Write/Bytes", "The number of bytes written by PGMPhysWrite.");
1634	PGM_REG_COUNTER(&pPGM->StatRZPhysSimpleRead, "/PGM/RZ/Phys/Simple/Read", "The number of times PGMPhysSimpleReadGCPtr was called.");
1635	PGM_REG_COUNTER_BYTES(&pPGM->StatRZPhysSimpleReadBytes, "/PGM/RZ/Phys/Simple/Read/Bytes", "The number of bytes read by PGMPhysSimpleReadGCPtr.");
1636	PGM_REG_COUNTER(&pPGM->StatRZPhysSimpleWrite, "/PGM/RZ/Phys/Simple/Write", "The number of times PGMPhysSimpleWriteGCPtr was called.");
1637	PGM_REG_COUNTER_BYTES(&pPGM->StatRZPhysSimpleWriteBytes, "/PGM/RZ/Phys/Simple/Write/Bytes", "The number of bytes written by PGMPhysSimpleWriteGCPtr.");
1638
1639	/* GC only: */
1640	PGM_REG_COUNTER(&pPGM->StatRCDynMapCacheHits, "/PGM/RC/DynMapCache/Hits" , "Number of dynamic page mapping cache hits.");
1641	PGM_REG_COUNTER(&pPGM->StatRCDynMapCacheMisses, "/PGM/RC/DynMapCache/Misses" , "Number of dynamic page mapping cache misses.");
1642	PGM_REG_COUNTER(&pPGM->StatRCInvlPgConflict, "/PGM/RC/InvlPgConflict", "Number of times PGMInvalidatePage() detected a mapping conflict.");
1643	PGM_REG_COUNTER(&pPGM->StatRCInvlPgSyncMonCR3, "/PGM/RC/InvlPgSyncMonitorCR3", "Number of times PGMInvalidatePage() ran into PGM_SYNC_MONITOR_CR3.");
1644
1645	PGM_REG_COUNTER(&pPGM->StatRCPhysRead, "/PGM/RC/Phys/Read", "The number of times PGMPhysRead was called.");
1646	PGM_REG_COUNTER_BYTES(&pPGM->StatRCPhysReadBytes, "/PGM/RC/Phys/Read/Bytes", "The number of bytes read by PGMPhysRead.");
1647	PGM_REG_COUNTER(&pPGM->StatRCPhysWrite, "/PGM/RC/Phys/Write", "The number of times PGMPhysWrite was called.");
1648	PGM_REG_COUNTER_BYTES(&pPGM->StatRCPhysWriteBytes, "/PGM/RC/Phys/Write/Bytes", "The number of bytes written by PGMPhysWrite.");
1649	PGM_REG_COUNTER(&pPGM->StatRCPhysSimpleRead, "/PGM/RC/Phys/Simple/Read", "The number of times PGMPhysSimpleReadGCPtr was called.");
1650	PGM_REG_COUNTER_BYTES(&pPGM->StatRCPhysSimpleReadBytes, "/PGM/RC/Phys/Simple/Read/Bytes", "The number of bytes read by PGMPhysSimpleReadGCPtr.");
1651	PGM_REG_COUNTER(&pPGM->StatRCPhysSimpleWrite, "/PGM/RC/Phys/Simple/Write", "The number of times PGMPhysSimpleWriteGCPtr was called.");
1652	PGM_REG_COUNTER_BYTES(&pPGM->StatRCPhysSimpleWriteBytes, "/PGM/RC/Phys/Simple/Write/Bytes", "The number of bytes written by PGMPhysSimpleWriteGCPtr.");
1653
1654	# ifdef PGMPOOL_WITH_GCPHYS_TRACKING
1655	PGM_REG_COUNTER(&pPGM->StatTrackVirgin, "/PGM/Track/Virgin", "The number of first time shadowings");
1656	PGM_REG_COUNTER(&pPGM->StatTrackAliased, "/PGM/Track/Aliased", "The number of times switching to cRef2, i.e. the page is being shadowed by two PTs.");
1657	PGM_REG_COUNTER(&pPGM->StatTrackAliasedMany, "/PGM/Track/AliasedMany", "The number of times we're tracking using cRef2.");
1658	PGM_REG_COUNTER(&pPGM->StatTrackAliasedLots, "/PGM/Track/AliasedLots", "The number of times we're hitting pages which has overflowed cRef2");
1659	PGM_REG_COUNTER(&pPGM->StatTrackOverflows, "/PGM/Track/Overflows", "The number of times the extent list grows too long.");
1660	PGM_REG_PROFILE(&pPGM->StatTrackDeref, "/PGM/Track/Deref", "Profiling of SyncPageWorkerTrackDeref (expensive).");
1661	# endif
1662
1663	# undef PGM_REG_COUNTER
1664	# undef PGM_REG_PROFILE
1665	#endif
1666
1667	/*
1668	* Note! The layout below matches the member layout exactly!
1669	*/
1670
1671	/*
1672	* Common - stats
1673	*/
1674	for (VMCPUID i = 0; i < pVM->cCpus; i++)
1675	{
1676	PVMCPU pVCpu = &pVM->aCpus[i];
1677	PPGMCPU pPGM = &pVCpu->pgm.s;
1678
1679	#define PGM_REG_COUNTER(a, b, c) \
1680	rc = STAMR3RegisterF(pVM, a, STAMTYPE_COUNTER, STAMVISIBILITY_ALWAYS, STAMUNIT_OCCURENCES, c, b, i); \
1681	AssertRC(rc);
1682	#define PGM_REG_PROFILE(a, b, c) \
1683	rc = STAMR3RegisterF(pVM, a, STAMTYPE_PROFILE, STAMVISIBILITY_ALWAYS, STAMUNIT_TICKS_PER_CALL, c, b, i); \
1684	AssertRC(rc);
1685
1686	PGM_REG_COUNTER(&pPGM->cGuestModeChanges, "/PGM/CPU%d/cGuestModeChanges", "Number of guest mode changes.");
1687
1688	#ifdef VBOX_WITH_STATISTICS
1689
1690	# if 0 /* rarely useful; leave for debugging. */
1691	for (unsigned j = 0; j < RT_ELEMENTS(pPGM->StatSyncPtPD); j++)
1692	STAMR3RegisterF(pVM, &pPGM->StatSyncPtPD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, STAMUNIT_OCCURENCES,
1693	"The number of SyncPT per PD n.", "/PGM/CPU%d/PDSyncPT/%04X", i, j);
1694	for (unsigned j = 0; j < RT_ELEMENTS(pPGM->StatSyncPagePD); j++)
1695	STAMR3RegisterF(pVM, &pPGM->StatSyncPagePD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, STAMUNIT_OCCURENCES,
1696	"The number of SyncPage per PD n.", "/PGM/CPU%d/PDSyncPage/%04X", i, j);
1697	# endif
1698	/* R0 only: */
1699	PGM_REG_COUNTER(&pPGM->StatR0DynMapMigrateInvlPg, "/PGM/CPU%d/R0/DynMapMigrateInvlPg", "invlpg count in PGMDynMapMigrateAutoSet.");
1700	PGM_REG_PROFILE(&pPGM->StatR0DynMapGCPageInl, "/PGM/CPU%d/R0/DynMapPageGCPageInl", "Calls to pgmR0DynMapGCPageInlined.");
1701	PGM_REG_COUNTER(&pPGM->StatR0DynMapGCPageInlHits, "/PGM/CPU%d/R0/DynMapPageGCPageInl/Hits", "Hash table lookup hits.");
1702	PGM_REG_COUNTER(&pPGM->StatR0DynMapGCPageInlMisses, "/PGM/CPU%d/R0/DynMapPageGCPageInl/Misses", "Misses that falls back to code common with PGMDynMapHCPage.");
1703	PGM_REG_COUNTER(&pPGM->StatR0DynMapGCPageInlRamHits, "/PGM/CPU%d/R0/DynMapPageGCPageInl/RamHits", "1st ram range hits.");
1704	PGM_REG_COUNTER(&pPGM->StatR0DynMapGCPageInlRamMisses, "/PGM/CPU%d/R0/DynMapPageGCPageInl/RamMisses", "1st ram range misses, takes slow path.");
1705	PGM_REG_PROFILE(&pPGM->StatR0DynMapHCPageInl, "/PGM/CPU%d/R0/DynMapPageHCPageInl", "Calls to pgmR0DynMapHCPageInlined.");
1706	PGM_REG_COUNTER(&pPGM->StatR0DynMapHCPageInlHits, "/PGM/CPU%d/R0/DynMapPageHCPageInl/Hits", "Hash table lookup hits.");
1707	PGM_REG_COUNTER(&pPGM->StatR0DynMapHCPageInlMisses, "/PGM/CPU%d/R0/DynMapPageHCPageInl/Misses", "Misses that falls back to code common with PGMDynMapHCPage.");
1708	PGM_REG_COUNTER(&pPGM->StatR0DynMapPage, "/PGM/CPU%d/R0/DynMapPage", "Calls to pgmR0DynMapPage");
1709	PGM_REG_COUNTER(&pPGM->StatR0DynMapSetOptimize, "/PGM/CPU%d/R0/DynMapPage/SetOptimize", "Calls to pgmDynMapOptimizeAutoSet.");
1710	PGM_REG_COUNTER(&pPGM->StatR0DynMapSetSearchFlushes, "/PGM/CPU%d/R0/DynMapPage/SetSearchFlushes","Set search restorting to subset flushes.");
1711	PGM_REG_COUNTER(&pPGM->StatR0DynMapSetSearchHits, "/PGM/CPU%d/R0/DynMapPage/SetSearchHits", "Set search hits.");
1712	PGM_REG_COUNTER(&pPGM->StatR0DynMapSetSearchMisses, "/PGM/CPU%d/R0/DynMapPage/SetSearchMisses", "Set search misses.");
1713	PGM_REG_PROFILE(&pPGM->StatR0DynMapHCPage, "/PGM/CPU%d/R0/DynMapPage/HCPage", "Calls to PGMDynMapHCPage (ring-0).");
1714	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageHits0, "/PGM/CPU%d/R0/DynMapPage/Hits0", "Hits at iPage+0");
1715	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageHits1, "/PGM/CPU%d/R0/DynMapPage/Hits1", "Hits at iPage+1");
1716	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageHits2, "/PGM/CPU%d/R0/DynMapPage/Hits2", "Hits at iPage+2");
1717	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageInvlPg, "/PGM/CPU%d/R0/DynMapPage/InvlPg", "invlpg count in pgmR0DynMapPageSlow.");
1718	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageSlow, "/PGM/CPU%d/R0/DynMapPage/Slow", "Calls to pgmR0DynMapPageSlow - subtract this from pgmR0DynMapPage to get 1st level hits.");
1719	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageSlowLoopHits, "/PGM/CPU%d/R0/DynMapPage/SlowLoopHits" , "Hits in the loop path.");
1720	PGM_REG_COUNTER(&pPGM->StatR0DynMapPageSlowLoopMisses, "/PGM/CPU%d/R0/DynMapPage/SlowLoopMisses", "Misses in the loop path. NonLoopMisses = Slow - SlowLoopHit - SlowLoopMisses");
1721	//PGM_REG_COUNTER(&pPGM->StatR0DynMapPageSlowLostHits, "/PGM/CPU%d/R0/DynMapPage/SlowLostHits", "Lost hits.");
1722	PGM_REG_COUNTER(&pPGM->StatR0DynMapSubsets, "/PGM/CPU%d/R0/Subsets", "Times PGMDynMapPushAutoSubset was called.");
1723	PGM_REG_COUNTER(&pPGM->StatR0DynMapPopFlushes, "/PGM/CPU%d/R0/SubsetPopFlushes", "Times PGMDynMapPopAutoSubset flushes the subset.");
1724	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[0], "/PGM/CPU%d/R0/SetSize000..09", "00-09% filled");
1725	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[1], "/PGM/CPU%d/R0/SetSize010..19", "10-19% filled");
1726	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[2], "/PGM/CPU%d/R0/SetSize020..29", "20-29% filled");
1727	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[3], "/PGM/CPU%d/R0/SetSize030..39", "30-39% filled");
1728	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[4], "/PGM/CPU%d/R0/SetSize040..49", "40-49% filled");
1729	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[5], "/PGM/CPU%d/R0/SetSize050..59", "50-59% filled");
1730	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[6], "/PGM/CPU%d/R0/SetSize060..69", "60-69% filled");
1731	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[7], "/PGM/CPU%d/R0/SetSize070..79", "70-79% filled");
1732	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[8], "/PGM/CPU%d/R0/SetSize080..89", "80-89% filled");
1733	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[9], "/PGM/CPU%d/R0/SetSize090..99", "90-99% filled");
1734	PGM_REG_COUNTER(&pPGM->aStatR0DynMapSetSize[10], "/PGM/CPU%d/R0/SetSize100", "100% filled");
1735
1736	/* RZ only: */
1737	PGM_REG_PROFILE(&pPGM->StatRZTrap0e, "/PGM/CPU%d/RZ/Trap0e", "Profiling of the PGMTrap0eHandler() body.");
1738	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTimeCheckPageFault, "/PGM/CPU%d/RZ/Trap0e/Time/CheckPageFault", "Profiling of checking for dirty/access emulation faults.");
1739	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTimeSyncPT, "/PGM/CPU%d/RZ/Trap0e/Time/SyncPT", "Profiling of lazy page table syncing.");
1740	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTimeMapping, "/PGM/CPU%d/RZ/Trap0e/Time/Mapping", "Profiling of checking virtual mappings.");
1741	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTimeOutOfSync, "/PGM/CPU%d/RZ/Trap0e/Time/OutOfSync", "Profiling of out of sync page handling.");
1742	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTimeHandlers, "/PGM/CPU%d/RZ/Trap0e/Time/Handlers", "Profiling of checking handlers.");
1743	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2CSAM, "/PGM/CPU%d/RZ/Trap0e/Time2/CSAM", "Profiling of the Trap0eHandler body when the cause is CSAM.");
1744	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2DirtyAndAccessed, "/PGM/CPU%d/RZ/Trap0e/Time2/DirtyAndAccessedBits", "Profiling of the Trap0eHandler body when the cause is dirty and/or accessed bit emulation.");
1745	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2GuestTrap, "/PGM/CPU%d/RZ/Trap0e/Time2/GuestTrap", "Profiling of the Trap0eHandler body when the cause is a guest trap.");
1746	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2HndPhys, "/PGM/CPU%d/RZ/Trap0e/Time2/HandlerPhysical", "Profiling of the Trap0eHandler body when the cause is a physical handler.");
1747	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2HndVirt, "/PGM/CPU%d/RZ/Trap0e/Time2/HandlerVirtual", "Profiling of the Trap0eHandler body when the cause is a virtual handler.");
1748	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2HndUnhandled, "/PGM/CPU%d/RZ/Trap0e/Time2/HandlerUnhandled", "Profiling of the Trap0eHandler body when the cause is access outside the monitored areas of a monitored page.");
1749	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2Misc, "/PGM/CPU%d/RZ/Trap0e/Time2/Misc", "Profiling of the Trap0eHandler body when the cause is not known.");
1750	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2OutOfSync, "/PGM/CPU%d/RZ/Trap0e/Time2/OutOfSync", "Profiling of the Trap0eHandler body when the cause is an out-of-sync page.");
1751	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2OutOfSyncHndPhys, "/PGM/CPU%d/RZ/Trap0e/Time2/OutOfSyncHndPhys", "Profiling of the Trap0eHandler body when the cause is an out-of-sync physical handler page.");
1752	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2OutOfSyncHndVirt, "/PGM/CPU%d/RZ/Trap0e/Time2/OutOfSyncHndVirt", "Profiling of the Trap0eHandler body when the cause is an out-of-sync virtual handler page.");
1753	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2OutOfSyncHndObs, "/PGM/CPU%d/RZ/Trap0e/Time2/OutOfSyncObsHnd", "Profiling of the Trap0eHandler body when the cause is an obsolete handler page.");
1754	PGM_REG_PROFILE(&pPGM->StatRZTrap0eTime2SyncPT, "/PGM/CPU%d/RZ/Trap0e/Time2/SyncPT", "Profiling of the Trap0eHandler body when the cause is lazy syncing of a PT.");
1755	PGM_REG_COUNTER(&pPGM->StatRZTrap0eConflicts, "/PGM/CPU%d/RZ/Trap0e/Conflicts", "The number of times #PF was caused by an undetected conflict.");
1756	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersMapping, "/PGM/CPU%d/RZ/Trap0e/Handlers/Mapping", "Number of traps due to access handlers in mappings.");
1757	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersOutOfSync, "/PGM/CPU%d/RZ/Trap0e/Handlers/OutOfSync", "Number of traps due to out-of-sync handled pages.");
1758	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersPhysical, "/PGM/CPU%d/RZ/Trap0e/Handlers/Physical", "Number of traps due to physical access handlers.");
1759	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersVirtual, "/PGM/CPU%d/RZ/Trap0e/Handlers/Virtual", "Number of traps due to virtual access handlers.");
1760	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersVirtualByPhys, "/PGM/CPU%d/RZ/Trap0e/Handlers/VirtualByPhys", "Number of traps due to virtual access handlers by physical address.");
1761	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersVirtualUnmarked,"/PGM/CPU%d/RZ/Trap0e/Handlers/VirtualUnmarked","Number of traps due to virtual access handlers by virtual address (without proper physical flags).");
1762	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersUnhandled, "/PGM/CPU%d/RZ/Trap0e/Handlers/Unhandled", "Number of traps due to access outside range of monitored page(s).");
1763	PGM_REG_COUNTER(&pPGM->StatRZTrap0eHandlersInvalid, "/PGM/CPU%d/RZ/Trap0e/Handlers/Invalid", "Number of traps due to access to invalid physical memory.");
1764	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSNotPresentRead, "/PGM/CPU%d/RZ/Trap0e/Err/User/NPRead", "Number of user mode not present read page faults.");
1765	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSNotPresentWrite, "/PGM/CPU%d/RZ/Trap0e/Err/User/NPWrite", "Number of user mode not present write page faults.");
1766	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSWrite, "/PGM/CPU%d/RZ/Trap0e/Err/User/Write", "Number of user mode write page faults.");
1767	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSReserved, "/PGM/CPU%d/RZ/Trap0e/Err/User/Reserved", "Number of user mode reserved bit page faults.");
1768	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSNXE, "/PGM/CPU%d/RZ/Trap0e/Err/User/NXE", "Number of user mode NXE page faults.");
1769	PGM_REG_COUNTER(&pPGM->StatRZTrap0eUSRead, "/PGM/CPU%d/RZ/Trap0e/Err/User/Read", "Number of user mode read page faults.");
1770	PGM_REG_COUNTER(&pPGM->StatRZTrap0eSVNotPresentRead, "/PGM/CPU%d/RZ/Trap0e/Err/Supervisor/NPRead", "Number of supervisor mode not present read page faults.");
1771	PGM_REG_COUNTER(&pPGM->StatRZTrap0eSVNotPresentWrite, "/PGM/CPU%d/RZ/Trap0e/Err/Supervisor/NPWrite", "Number of supervisor mode not present write page faults.");
1772	PGM_REG_COUNTER(&pPGM->StatRZTrap0eSVWrite, "/PGM/CPU%d/RZ/Trap0e/Err/Supervisor/Write", "Number of supervisor mode write page faults.");
1773	PGM_REG_COUNTER(&pPGM->StatRZTrap0eSVReserved, "/PGM/CPU%d/RZ/Trap0e/Err/Supervisor/Reserved", "Number of supervisor mode reserved bit page faults.");
1774	PGM_REG_COUNTER(&pPGM->StatRZTrap0eSNXE, "/PGM/CPU%d/RZ/Trap0e/Err/Supervisor/NXE", "Number of supervisor mode NXE page faults.");
1775	PGM_REG_COUNTER(&pPGM->StatRZTrap0eGuestPF, "/PGM/CPU%d/RZ/Trap0e/GuestPF", "Number of real guest page faults.");
1776	PGM_REG_COUNTER(&pPGM->StatRZTrap0eGuestPFUnh, "/PGM/CPU%d/RZ/Trap0e/GuestPF/Unhandled", "Number of real guest page faults from the 'unhandled' case.");
1777	PGM_REG_COUNTER(&pPGM->StatRZTrap0eGuestPFMapping, "/PGM/CPU%d/RZ/Trap0e/GuestPF/InMapping", "Number of real guest page faults in a mapping.");
1778	PGM_REG_COUNTER(&pPGM->StatRZTrap0eWPEmulInRZ, "/PGM/CPU%d/RZ/Trap0e/WP/InRZ", "Number of guest page faults due to X86_CR0_WP emulation.");
1779	PGM_REG_COUNTER(&pPGM->StatRZTrap0eWPEmulToR3, "/PGM/CPU%d/RZ/Trap0e/WP/ToR3", "Number of guest page faults due to X86_CR0_WP emulation (forward to R3 for emulation).");
1780	#if 0 /* rarely useful; leave for debugging. */
1781	for (unsigned j = 0; j < RT_ELEMENTS(pPGM->StatRZTrap0ePD); j++)
1782	STAMR3RegisterF(pVM, &pPGM->StatRZTrap0ePD[i], STAMTYPE_COUNTER, STAMVISIBILITY_USED, STAMUNIT_OCCURENCES,
1783	"The number of traps in page directory n.", "/PGM/CPU%d/RZ/Trap0e/PD/%04X", i, j);
1784	#endif
1785	PGM_REG_COUNTER(&pPGM->StatRZGuestCR3WriteHandled, "/PGM/CPU%d/RZ/CR3WriteHandled", "The number of times the Guest CR3 change was successfully handled.");
1786	PGM_REG_COUNTER(&pPGM->StatRZGuestCR3WriteUnhandled, "/PGM/CPU%d/RZ/CR3WriteUnhandled", "The number of times the Guest CR3 change was passed back to the recompiler.");
1787	PGM_REG_COUNTER(&pPGM->StatRZGuestCR3WriteConflict, "/PGM/CPU%d/RZ/CR3WriteConflict", "The number of times the Guest CR3 monitoring detected a conflict.");
1788	PGM_REG_COUNTER(&pPGM->StatRZGuestROMWriteHandled, "/PGM/CPU%d/RZ/ROMWriteHandled", "The number of times the Guest ROM change was successfully handled.");
1789	PGM_REG_COUNTER(&pPGM->StatRZGuestROMWriteUnhandled, "/PGM/CPU%d/RZ/ROMWriteUnhandled", "The number of times the Guest ROM change was passed back to the recompiler.");
1790
1791	/* HC only: */
1792
1793	/* RZ & R3: */
1794	PGM_REG_PROFILE(&pPGM->StatRZSyncCR3, "/PGM/CPU%d/RZ/SyncCR3", "Profiling of the PGMSyncCR3() body.");
1795	PGM_REG_PROFILE(&pPGM->StatRZSyncCR3Handlers, "/PGM/CPU%d/RZ/SyncCR3/Handlers", "Profiling of the PGMSyncCR3() update handler section.");
1796	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3Global, "/PGM/CPU%d/RZ/SyncCR3/Global", "The number of global CR3 syncs.");
1797	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3NotGlobal, "/PGM/CPU%d/RZ/SyncCR3/NotGlobal", "The number of non-global CR3 syncs.");
1798	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstCacheHit, "/PGM/CPU%d/RZ/SyncCR3/DstChacheHit", "The number of times we got some kind of a cache hit.");
1799	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstFreed, "/PGM/CPU%d/RZ/SyncCR3/DstFreed", "The number of times we've had to free a shadow entry.");
1800	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstFreedSrcNP, "/PGM/CPU%d/RZ/SyncCR3/DstFreedSrcNP", "The number of times we've had to free a shadow entry for which the source entry was not present.");
1801	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstNotPresent, "/PGM/CPU%d/RZ/SyncCR3/DstNotPresent", "The number of times we've encountered a not present shadow entry for a present guest entry.");
1802	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstSkippedGlobalPD, "/PGM/CPU%d/RZ/SyncCR3/DstSkippedGlobalPD", "The number of times a global page directory wasn't flushed.");
1803	PGM_REG_COUNTER(&pPGM->StatRZSyncCR3DstSkippedGlobalPT, "/PGM/CPU%d/RZ/SyncCR3/DstSkippedGlobalPT", "The number of times a page table with only global entries wasn't flushed.");
1804	PGM_REG_PROFILE(&pPGM->StatRZSyncPT, "/PGM/CPU%d/RZ/SyncPT", "Profiling of the pfnSyncPT() body.");
1805	PGM_REG_COUNTER(&pPGM->StatRZSyncPTFailed, "/PGM/CPU%d/RZ/SyncPT/Failed", "The number of times pfnSyncPT() failed.");
1806	PGM_REG_COUNTER(&pPGM->StatRZSyncPT4K, "/PGM/CPU%d/RZ/SyncPT/4K", "Nr of 4K PT syncs");
1807	PGM_REG_COUNTER(&pPGM->StatRZSyncPT4M, "/PGM/CPU%d/RZ/SyncPT/4M", "Nr of 4M PT syncs");
1808	PGM_REG_COUNTER(&pPGM->StatRZSyncPagePDNAs, "/PGM/CPU%d/RZ/SyncPagePDNAs", "The number of time we've marked a PD not present from SyncPage to virtualize the accessed bit.");
1809	PGM_REG_COUNTER(&pPGM->StatRZSyncPagePDOutOfSync, "/PGM/CPU%d/RZ/SyncPagePDOutOfSync", "The number of time we've encountered an out-of-sync PD in SyncPage.");
1810	PGM_REG_COUNTER(&pPGM->StatRZAccessedPage, "/PGM/CPU%d/RZ/AccessedPage", "The number of pages marked not present for accessed bit emulation.");
1811	PGM_REG_PROFILE(&pPGM->StatRZDirtyBitTracking, "/PGM/CPU%d/RZ/DirtyPage", "Profiling the dirty bit tracking in CheckPageFault().");
1812	PGM_REG_COUNTER(&pPGM->StatRZDirtyPage, "/PGM/CPU%d/RZ/DirtyPage/Mark", "The number of pages marked read-only for dirty bit tracking.");
1813	PGM_REG_COUNTER(&pPGM->StatRZDirtyPageBig, "/PGM/CPU%d/RZ/DirtyPage/MarkBig", "The number of 4MB pages marked read-only for dirty bit tracking.");
1814	PGM_REG_COUNTER(&pPGM->StatRZDirtyPageSkipped, "/PGM/CPU%d/RZ/DirtyPage/Skipped", "The number of pages already dirty or readonly.");
1815	PGM_REG_COUNTER(&pPGM->StatRZDirtyPageTrap, "/PGM/CPU%d/RZ/DirtyPage/Trap", "The number of traps generated for dirty bit tracking.");
1816	PGM_REG_COUNTER(&pPGM->StatRZDirtyPageStale, "/PGM/CPU%d/RZ/DirtyPage/Stale", "The number of traps generated for dirty bit tracking (stale tlb entries).");
1817	PGM_REG_COUNTER(&pPGM->StatRZDirtiedPage, "/PGM/CPU%d/RZ/DirtyPage/SetDirty", "The number of pages marked dirty because of write accesses.");
1818	PGM_REG_COUNTER(&pPGM->StatRZDirtyTrackRealPF, "/PGM/CPU%d/RZ/DirtyPage/RealPF", "The number of real pages faults during dirty bit tracking.");
1819	PGM_REG_COUNTER(&pPGM->StatRZPageAlreadyDirty, "/PGM/CPU%d/RZ/DirtyPage/AlreadySet", "The number of pages already marked dirty because of write accesses.");
1820	PGM_REG_PROFILE(&pPGM->StatRZInvalidatePage, "/PGM/CPU%d/RZ/InvalidatePage", "PGMInvalidatePage() profiling.");
1821	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePage4KBPages, "/PGM/CPU%d/RZ/InvalidatePage/4KBPages", "The number of times PGMInvalidatePage() was called for a 4KB page.");
1822	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePage4MBPages, "/PGM/CPU%d/RZ/InvalidatePage/4MBPages", "The number of times PGMInvalidatePage() was called for a 4MB page.");
1823	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePage4MBPagesSkip, "/PGM/CPU%d/RZ/InvalidatePage/4MBPagesSkip","The number of times PGMInvalidatePage() skipped a 4MB page.");
1824	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePagePDMappings, "/PGM/CPU%d/RZ/InvalidatePage/PDMappings", "The number of times PGMInvalidatePage() was called for a page directory containing mappings (no conflict).");
1825	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePagePDNAs, "/PGM/CPU%d/RZ/InvalidatePage/PDNAs", "The number of times PGMInvalidatePage() was called for a not accessed page directory.");
1826	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePagePDNPs, "/PGM/CPU%d/RZ/InvalidatePage/PDNPs", "The number of times PGMInvalidatePage() was called for a not present page directory.");
1827	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePagePDOutOfSync, "/PGM/CPU%d/RZ/InvalidatePage/PDOutOfSync", "The number of times PGMInvalidatePage() was called for an out of sync page directory.");
1828	PGM_REG_COUNTER(&pPGM->StatRZInvalidatePageSkipped, "/PGM/CPU%d/RZ/InvalidatePage/Skipped", "The number of times PGMInvalidatePage() was skipped due to not present shw or pending pending SyncCR3.");
1829	PGM_REG_COUNTER(&pPGM->StatRZPageOutOfSyncSupervisor, "/PGM/CPU%d/RZ/OutOfSync/SuperVisor", "Number of traps due to pages out of sync (P) and times VerifyAccessSyncPage calls SyncPage.");
1830	PGM_REG_COUNTER(&pPGM->StatRZPageOutOfSyncUser, "/PGM/CPU%d/RZ/OutOfSync/User", "Number of traps due to pages out of sync (P) and times VerifyAccessSyncPage calls SyncPage.");
1831	PGM_REG_COUNTER(&pPGM->StatRZPageOutOfSyncSupervisorWrite,"/PGM/CPU%d/RZ/OutOfSync/SuperVisorWrite", "Number of traps due to pages out of sync (RW) and times VerifyAccessSyncPage calls SyncPage.");
1832	PGM_REG_COUNTER(&pPGM->StatRZPageOutOfSyncUserWrite, "/PGM/CPU%d/RZ/OutOfSync/UserWrite", "Number of traps due to pages out of sync (RW) and times VerifyAccessSyncPage calls SyncPage.");
1833	PGM_REG_PROFILE(&pPGM->StatRZPrefetch, "/PGM/CPU%d/RZ/Prefetch", "PGMPrefetchPage profiling.");
1834	PGM_REG_PROFILE(&pPGM->StatRZFlushTLB, "/PGM/CPU%d/RZ/FlushTLB", "Profiling of the PGMFlushTLB() body.");
1835	PGM_REG_COUNTER(&pPGM->StatRZFlushTLBNewCR3, "/PGM/CPU%d/RZ/FlushTLB/NewCR3", "The number of times PGMFlushTLB was called with a new CR3, non-global. (switch)");
1836	PGM_REG_COUNTER(&pPGM->StatRZFlushTLBNewCR3Global, "/PGM/CPU%d/RZ/FlushTLB/NewCR3Global", "The number of times PGMFlushTLB was called with a new CR3, global. (switch)");
1837	PGM_REG_COUNTER(&pPGM->StatRZFlushTLBSameCR3, "/PGM/CPU%d/RZ/FlushTLB/SameCR3", "The number of times PGMFlushTLB was called with the same CR3, non-global. (flush)");
1838	PGM_REG_COUNTER(&pPGM->StatRZFlushTLBSameCR3Global, "/PGM/CPU%d/RZ/FlushTLB/SameCR3Global", "The number of times PGMFlushTLB was called with the same CR3, global. (flush)");
1839	PGM_REG_PROFILE(&pPGM->StatRZGstModifyPage, "/PGM/CPU%d/RZ/GstModifyPage", "Profiling of the PGMGstModifyPage() body.");
1840
1841	PGM_REG_PROFILE(&pPGM->StatR3SyncCR3, "/PGM/CPU%d/R3/SyncCR3", "Profiling of the PGMSyncCR3() body.");
1842	PGM_REG_PROFILE(&pPGM->StatR3SyncCR3Handlers, "/PGM/CPU%d/R3/SyncCR3/Handlers", "Profiling of the PGMSyncCR3() update handler section.");
1843	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3Global, "/PGM/CPU%d/R3/SyncCR3/Global", "The number of global CR3 syncs.");
1844	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3NotGlobal, "/PGM/CPU%d/R3/SyncCR3/NotGlobal", "The number of non-global CR3 syncs.");
1845	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstCacheHit, "/PGM/CPU%d/R3/SyncCR3/DstChacheHit", "The number of times we got some kind of a cache hit.");
1846	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstFreed, "/PGM/CPU%d/R3/SyncCR3/DstFreed", "The number of times we've had to free a shadow entry.");
1847	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstFreedSrcNP, "/PGM/CPU%d/R3/SyncCR3/DstFreedSrcNP", "The number of times we've had to free a shadow entry for which the source entry was not present.");
1848	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstNotPresent, "/PGM/CPU%d/R3/SyncCR3/DstNotPresent", "The number of times we've encountered a not present shadow entry for a present guest entry.");
1849	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstSkippedGlobalPD, "/PGM/CPU%d/R3/SyncCR3/DstSkippedGlobalPD", "The number of times a global page directory wasn't flushed.");
1850	PGM_REG_COUNTER(&pPGM->StatR3SyncCR3DstSkippedGlobalPT, "/PGM/CPU%d/R3/SyncCR3/DstSkippedGlobalPT", "The number of times a page table with only global entries wasn't flushed.");
1851	PGM_REG_PROFILE(&pPGM->StatR3SyncPT, "/PGM/CPU%d/R3/SyncPT", "Profiling of the pfnSyncPT() body.");
1852	PGM_REG_COUNTER(&pPGM->StatR3SyncPTFailed, "/PGM/CPU%d/R3/SyncPT/Failed", "The number of times pfnSyncPT() failed.");
1853	PGM_REG_COUNTER(&pPGM->StatR3SyncPT4K, "/PGM/CPU%d/R3/SyncPT/4K", "Nr of 4K PT syncs");
1854	PGM_REG_COUNTER(&pPGM->StatR3SyncPT4M, "/PGM/CPU%d/R3/SyncPT/4M", "Nr of 4M PT syncs");
1855	PGM_REG_COUNTER(&pPGM->StatR3SyncPagePDNAs, "/PGM/CPU%d/R3/SyncPagePDNAs", "The number of time we've marked a PD not present from SyncPage to virtualize the accessed bit.");
1856	PGM_REG_COUNTER(&pPGM->StatR3SyncPagePDOutOfSync, "/PGM/CPU%d/R3/SyncPagePDOutOfSync", "The number of time we've encountered an out-of-sync PD in SyncPage.");
1857	PGM_REG_COUNTER(&pPGM->StatR3AccessedPage, "/PGM/CPU%d/R3/AccessedPage", "The number of pages marked not present for accessed bit emulation.");
1858	PGM_REG_PROFILE(&pPGM->StatR3DirtyBitTracking, "/PGM/CPU%d/R3/DirtyPage", "Profiling the dirty bit tracking in CheckPageFault().");
1859	PGM_REG_COUNTER(&pPGM->StatR3DirtyPage, "/PGM/CPU%d/R3/DirtyPage/Mark", "The number of pages marked read-only for dirty bit tracking.");
1860	PGM_REG_COUNTER(&pPGM->StatR3DirtyPageBig, "/PGM/CPU%d/R3/DirtyPage/MarkBig", "The number of 4MB pages marked read-only for dirty bit tracking.");
1861	PGM_REG_COUNTER(&pPGM->StatR3DirtyPageSkipped, "/PGM/CPU%d/R3/DirtyPage/Skipped", "The number of pages already dirty or readonly.");
1862	PGM_REG_COUNTER(&pPGM->StatR3DirtyPageTrap, "/PGM/CPU%d/R3/DirtyPage/Trap", "The number of traps generated for dirty bit tracking.");
1863	PGM_REG_COUNTER(&pPGM->StatR3DirtiedPage, "/PGM/CPU%d/R3/DirtyPage/SetDirty", "The number of pages marked dirty because of write accesses.");
1864	PGM_REG_COUNTER(&pPGM->StatR3DirtyTrackRealPF, "/PGM/CPU%d/R3/DirtyPage/RealPF", "The number of real pages faults during dirty bit tracking.");
1865	PGM_REG_COUNTER(&pPGM->StatR3PageAlreadyDirty, "/PGM/CPU%d/R3/DirtyPage/AlreadySet", "The number of pages already marked dirty because of write accesses.");
1866	PGM_REG_PROFILE(&pPGM->StatR3InvalidatePage, "/PGM/CPU%d/R3/InvalidatePage", "PGMInvalidatePage() profiling.");
1867	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePage4KBPages, "/PGM/CPU%d/R3/InvalidatePage/4KBPages", "The number of times PGMInvalidatePage() was called for a 4KB page.");
1868	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePage4MBPages, "/PGM/CPU%d/R3/InvalidatePage/4MBPages", "The number of times PGMInvalidatePage() was called for a 4MB page.");
1869	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePage4MBPagesSkip, "/PGM/CPU%d/R3/InvalidatePage/4MBPagesSkip","The number of times PGMInvalidatePage() skipped a 4MB page.");
1870	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePagePDMappings, "/PGM/CPU%d/R3/InvalidatePage/PDMappings", "The number of times PGMInvalidatePage() was called for a page directory containing mappings (no conflict).");
1871	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePagePDNAs, "/PGM/CPU%d/R3/InvalidatePage/PDNAs", "The number of times PGMInvalidatePage() was called for a not accessed page directory.");
1872	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePagePDNPs, "/PGM/CPU%d/R3/InvalidatePage/PDNPs", "The number of times PGMInvalidatePage() was called for a not present page directory.");
1873	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePagePDOutOfSync, "/PGM/CPU%d/R3/InvalidatePage/PDOutOfSync", "The number of times PGMInvalidatePage() was called for an out of sync page directory.");
1874	PGM_REG_COUNTER(&pPGM->StatR3InvalidatePageSkipped, "/PGM/CPU%d/R3/InvalidatePage/Skipped", "The number of times PGMInvalidatePage() was skipped due to not present shw or pending pending SyncCR3.");
1875	PGM_REG_COUNTER(&pPGM->StatR3PageOutOfSyncSupervisor, "/PGM/CPU%d/R3/OutOfSync/SuperVisor", "Number of traps due to pages out of sync and times VerifyAccessSyncPage calls SyncPage.");
1876	PGM_REG_COUNTER(&pPGM->StatR3PageOutOfSyncUser, "/PGM/CPU%d/R3/OutOfSync/User", "Number of traps due to pages out of sync and times VerifyAccessSyncPage calls SyncPage.");
1877	PGM_REG_PROFILE(&pPGM->StatR3Prefetch, "/PGM/CPU%d/R3/Prefetch", "PGMPrefetchPage profiling.");
1878	PGM_REG_PROFILE(&pPGM->StatR3FlushTLB, "/PGM/CPU%d/R3/FlushTLB", "Profiling of the PGMFlushTLB() body.");
1879	PGM_REG_COUNTER(&pPGM->StatR3FlushTLBNewCR3, "/PGM/CPU%d/R3/FlushTLB/NewCR3", "The number of times PGMFlushTLB was called with a new CR3, non-global. (switch)");
1880	PGM_REG_COUNTER(&pPGM->StatR3FlushTLBNewCR3Global, "/PGM/CPU%d/R3/FlushTLB/NewCR3Global", "The number of times PGMFlushTLB was called with a new CR3, global. (switch)");
1881	PGM_REG_COUNTER(&pPGM->StatR3FlushTLBSameCR3, "/PGM/CPU%d/R3/FlushTLB/SameCR3", "The number of times PGMFlushTLB was called with the same CR3, non-global. (flush)");
1882	PGM_REG_COUNTER(&pPGM->StatR3FlushTLBSameCR3Global, "/PGM/CPU%d/R3/FlushTLB/SameCR3Global", "The number of times PGMFlushTLB was called with the same CR3, global. (flush)");
1883	PGM_REG_PROFILE(&pPGM->StatR3GstModifyPage, "/PGM/CPU%d/R3/GstModifyPage", "Profiling of the PGMGstModifyPage() body.");
1884	#endif /* VBOX_WITH_STATISTICS */
1885
1886	#undef PGM_REG_PROFILE
1887	#undef PGM_REG_COUNTER
1888
1889	}
1890	}
1891
1892
1893	/**
1894	* Init the PGM bits that rely on VMMR0 and MM to be fully initialized.
1895	*
1896	* The dynamic mapping area will also be allocated and initialized at this
1897	* time. We could allocate it during PGMR3Init of course, but the mapping
1898	* wouldn't be allocated at that time preventing us from setting up the
1899	* page table entries with the dummy page.
1900	*
1901	* @returns VBox status code.
1902	* @param pVM VM handle.
1903	*/
1904	VMMR3DECL(int) PGMR3InitDynMap(PVM pVM)
1905	{
1906	RTGCPTR GCPtr;
1907	int rc;
1908
1909	/*
1910	* Reserve space for the dynamic mappings.
1911	*/
1912	rc = MMR3HyperReserve(pVM, MM_HYPER_DYNAMIC_SIZE, "Dynamic mapping", &GCPtr);
1913	if (RT_SUCCESS(rc))
1914	pVM->pgm.s.pbDynPageMapBaseGC = GCPtr;
1915
1916	if ( RT_SUCCESS(rc)
1917	&& (pVM->pgm.s.pbDynPageMapBaseGC >> X86_PD_PAE_SHIFT) != ((pVM->pgm.s.pbDynPageMapBaseGC + MM_HYPER_DYNAMIC_SIZE - 1) >> X86_PD_PAE_SHIFT))
1918	{
1919	rc = MMR3HyperReserve(pVM, MM_HYPER_DYNAMIC_SIZE, "Dynamic mapping not crossing", &GCPtr);
1920	if (RT_SUCCESS(rc))
1921	pVM->pgm.s.pbDynPageMapBaseGC = GCPtr;
1922	}
1923	if (RT_SUCCESS(rc))
1924	{
1925	AssertRelease((pVM->pgm.s.pbDynPageMapBaseGC >> X86_PD_PAE_SHIFT) == ((pVM->pgm.s.pbDynPageMapBaseGC + MM_HYPER_DYNAMIC_SIZE - 1) >> X86_PD_PAE_SHIFT));
1926	MMR3HyperReserve(pVM, PAGE_SIZE, "fence", NULL);
1927	}
1928	return rc;
1929	}
1930
1931
1932	/**
1933	* Ring-3 init finalizing.
1934	*
1935	* @returns VBox status code.
1936	* @param pVM The VM handle.
1937	*/
1938	VMMR3DECL(int) PGMR3InitFinalize(PVM pVM)
1939	{
1940	int rc;
1941
1942	/*
1943	* Reserve space for the dynamic mappings.
1944	* Initialize the dynamic mapping pages with dummy pages to simply the cache.
1945	*/
1946	/* get the pointer to the page table entries. */
1947	PPGMMAPPING pMapping = pgmGetMapping(pVM, pVM->pgm.s.pbDynPageMapBaseGC);
1948	AssertRelease(pMapping);
1949	const uintptr_t off = pVM->pgm.s.pbDynPageMapBaseGC - pMapping->GCPtr;
1950	const unsigned iPT = off >> X86_PD_SHIFT;
1951	const unsigned iPG = (off >> X86_PT_SHIFT) & X86_PT_MASK;
1952	pVM->pgm.s.paDynPageMap32BitPTEsGC = pMapping->aPTs[iPT].pPTRC + iPG * sizeof(pMapping->aPTs[0].pPTR3->a[0]);
1953	pVM->pgm.s.paDynPageMapPaePTEsGC = pMapping->aPTs[iPT].paPaePTsRC + iPG * sizeof(pMapping->aPTs[0].paPaePTsR3->a[0]);
1954
1955	/* init cache */
1956	RTHCPHYS HCPhysDummy = MMR3PageDummyHCPhys(pVM);
1957	for (unsigned i = 0; i < RT_ELEMENTS(pVM->pgm.s.aHCPhysDynPageMapCache); i++)
1958	pVM->pgm.s.aHCPhysDynPageMapCache[i] = HCPhysDummy;
1959
1960	for (unsigned i = 0; i < MM_HYPER_DYNAMIC_SIZE; i += PAGE_SIZE)
1961	{
1962	rc = PGMMap(pVM, pVM->pgm.s.pbDynPageMapBaseGC + i, HCPhysDummy, PAGE_SIZE, 0);
1963	AssertRCReturn(rc, rc);
1964	}
1965
1966	/*
1967	* Note that AMD uses all the 8 reserved bits for the address (so 40 bits in total);
1968	* Intel only goes up to 36 bits, so we stick to 36 as well.
1969	*/
1970	/** @todo How to test for the 40 bits support? Long mode seems to be the test criterium. */
1971	uint32_t u32Dummy, u32Features;
1972	CPUMGetGuestCpuId(VMMGetCpu(pVM), 1, &u32Dummy, &u32Dummy, &u32Dummy, &u32Features);
1973
1974	if (u32Features & X86_CPUID_FEATURE_EDX_PSE36)
1975	pVM->pgm.s.GCPhys4MBPSEMask = RT_BIT_64(36) - 1;
1976	else
1977	pVM->pgm.s.GCPhys4MBPSEMask = RT_BIT_64(32) - 1;
1978
1979	/*
1980	* Allocate memory if we're supposed to do that.
1981	*/
1982	if (pVM->pgm.s.fRamPreAlloc)
1983	rc = pgmR3PhysRamPreAllocate(pVM);
1984
1985	LogRel(("PGMR3InitFinalize: 4 MB PSE mask %RGp\n", pVM->pgm.s.GCPhys4MBPSEMask));
1986	return rc;
1987	}
1988
1989
1990	/**
1991	* Applies relocations to data and code managed by this component.
1992	*
1993	* This function will be called at init and whenever the VMM need to relocate it
1994	* self inside the GC.
1995	*
1996	* @param pVM The VM.
1997	* @param offDelta Relocation delta relative to old location.
1998	*/
1999	VMMR3DECL(void) PGMR3Relocate(PVM pVM, RTGCINTPTR offDelta)
2000	{
2001	LogFlow(("PGMR3Relocate %RGv to %RGv\n", pVM->pgm.s.GCPtrCR3Mapping, pVM->pgm.s.GCPtrCR3Mapping + offDelta));
2002
2003	/*
2004	* Paging stuff.
2005	*/
2006	pVM->pgm.s.GCPtrCR3Mapping += offDelta;
2007
2008	pgmR3ModeDataInit(pVM, true /* resolve GC/R0 symbols */);
2009
2010	/* Shadow, guest and both mode switch & relocation for each VCPU. */
2011	for (VMCPUID i = 0; i < pVM->cCpus; i++)
2012	{
2013	PVMCPU pVCpu = &pVM->aCpus[i];
2014
2015	pgmR3ModeDataSwitch(pVM, pVCpu, pVCpu->pgm.s.enmShadowMode, pVCpu->pgm.s.enmGuestMode);
2016
2017	PGM_SHW_PFN(Relocate, pVCpu)(pVCpu, offDelta);
2018	PGM_GST_PFN(Relocate, pVCpu)(pVCpu, offDelta);
2019	PGM_BTH_PFN(Relocate, pVCpu)(pVCpu, offDelta);
2020	}
2021
2022	/*
2023	* Trees.
2024	*/
2025	pVM->pgm.s.pTreesRC = MMHyperR3ToRC(pVM, pVM->pgm.s.pTreesR3);
2026
2027	/*
2028	* Ram ranges.
2029	*/
2030	if (pVM->pgm.s.pRamRangesR3)
2031	{
2032	/* Update the pSelfRC pointers and relink them. */
2033	for (PPGMRAMRANGE pCur = pVM->pgm.s.pRamRangesR3; pCur; pCur = pCur->pNextR3)
2034	if (!(pCur->fFlags & PGM_RAM_RANGE_FLAGS_FLOATING))
2035	pCur->pSelfRC = MMHyperCCToRC(pVM, pCur);
2036	pgmR3PhysRelinkRamRanges(pVM);
2037	}
2038
2039	/*
2040	* Update the pSelfRC pointer of the MMIO2 ram ranges since they might not
2041	* be mapped and thus not included in the above exercise.
2042	*/
2043	for (PPGMMMIO2RANGE pCur = pVM->pgm.s.pMmio2RangesR3; pCur; pCur = pCur->pNextR3)
2044	if (!(pCur->RamRange.fFlags & PGM_RAM_RANGE_FLAGS_FLOATING))
2045	pCur->RamRange.pSelfRC = MMHyperCCToRC(pVM, &pCur->RamRange);
2046
2047	/*
2048	* Update the two page directories with all page table mappings.
2049	* (One or more of them have changed, that's why we're here.)
2050	*/
2051	pVM->pgm.s.pMappingsRC = MMHyperR3ToRC(pVM, pVM->pgm.s.pMappingsR3);
2052	for (PPGMMAPPING pCur = pVM->pgm.s.pMappingsR3; pCur->pNextR3; pCur = pCur->pNextR3)
2053	pCur->pNextRC = MMHyperR3ToRC(pVM, pCur->pNextR3);
2054
2055	/* Relocate GC addresses of Page Tables. */
2056	for (PPGMMAPPING pCur = pVM->pgm.s.pMappingsR3; pCur; pCur = pCur->pNextR3)
2057	{
2058	for (RTHCUINT i = 0; i < pCur->cPTs; i++)
2059	{
2060	pCur->aPTs[i].pPTRC = MMHyperR3ToRC(pVM, pCur->aPTs[i].pPTR3);
2061	pCur->aPTs[i].paPaePTsRC = MMHyperR3ToRC(pVM, pCur->aPTs[i].paPaePTsR3);
2062	}
2063	}
2064
2065	/*
2066	* Dynamic page mapping area.
2067	*/
2068	pVM->pgm.s.paDynPageMap32BitPTEsGC += offDelta;
2069	pVM->pgm.s.paDynPageMapPaePTEsGC += offDelta;
2070	pVM->pgm.s.pbDynPageMapBaseGC += offDelta;
2071
2072	/*
2073	* The Zero page.
2074	*/
2075	pVM->pgm.s.pvZeroPgR0 = MMHyperR3ToR0(pVM, pVM->pgm.s.pvZeroPgR3);
2076	#ifdef VBOX_WITH_2X_4GB_ADDR_SPACE
2077	AssertRelease(pVM->pgm.s.pvZeroPgR0 != NIL_RTR0PTR \|\| !VMMIsHwVirtExtForced(pVM));
2078	#else
2079	AssertRelease(pVM->pgm.s.pvZeroPgR0 != NIL_RTR0PTR);
2080	#endif
2081
2082	/*
2083	* Physical and virtual handlers.
2084	*/
2085	RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesR3->PhysHandlers, true, pgmR3RelocatePhysHandler, &offDelta);
2086	RTAvlroGCPtrDoWithAll(&pVM->pgm.s.pTreesR3->VirtHandlers, true, pgmR3RelocateVirtHandler, &offDelta);
2087	RTAvlroGCPtrDoWithAll(&pVM->pgm.s.pTreesR3->HyperVirtHandlers, true, pgmR3RelocateHyperVirtHandler, &offDelta);
2088
2089	/*
2090	* The page pool.
2091	*/
2092	pgmR3PoolRelocate(pVM);
2093	}
2094
2095
2096	/**
2097	* Callback function for relocating a physical access handler.
2098	*
2099	* @returns 0 (continue enum)
2100	* @param pNode Pointer to a PGMPHYSHANDLER node.
2101	* @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
2102	* not certain the delta will fit in a void pointer for all possible configs.
2103	*/
2104	static DECLCALLBACK(int) pgmR3RelocatePhysHandler(PAVLROGCPHYSNODECORE pNode, void *pvUser)
2105	{
2106	PPGMPHYSHANDLER pHandler = (PPGMPHYSHANDLER)pNode;
2107	RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
2108	if (pHandler->pfnHandlerRC)
2109	pHandler->pfnHandlerRC += offDelta;
2110	if (pHandler->pvUserRC >= 0x10000)
2111	pHandler->pvUserRC += offDelta;
2112	return 0;
2113	}
2114
2115
2116	/**
2117	* Callback function for relocating a virtual access handler.
2118	*
2119	* @returns 0 (continue enum)
2120	* @param pNode Pointer to a PGMVIRTHANDLER node.
2121	* @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
2122	* not certain the delta will fit in a void pointer for all possible configs.
2123	*/
2124	static DECLCALLBACK(int) pgmR3RelocateVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser)
2125	{
2126	PPGMVIRTHANDLER pHandler = (PPGMVIRTHANDLER)pNode;
2127	RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
2128	Assert( pHandler->enmType == PGMVIRTHANDLERTYPE_ALL
2129	\|\| pHandler->enmType == PGMVIRTHANDLERTYPE_WRITE);
2130	Assert(pHandler->pfnHandlerRC);
2131	pHandler->pfnHandlerRC += offDelta;
2132	return 0;
2133	}
2134
2135
2136	/**
2137	* Callback function for relocating a virtual access handler for the hypervisor mapping.
2138	*
2139	* @returns 0 (continue enum)
2140	* @param pNode Pointer to a PGMVIRTHANDLER node.
2141	* @param pvUser Pointer to the offDelta. This is a pointer to the delta since we're
2142	* not certain the delta will fit in a void pointer for all possible configs.
2143	*/
2144	static DECLCALLBACK(int) pgmR3RelocateHyperVirtHandler(PAVLROGCPTRNODECORE pNode, void *pvUser)
2145	{
2146	PPGMVIRTHANDLER pHandler = (PPGMVIRTHANDLER)pNode;
2147	RTGCINTPTR offDelta = *(PRTGCINTPTR)pvUser;
2148	Assert(pHandler->enmType == PGMVIRTHANDLERTYPE_HYPERVISOR);
2149	Assert(pHandler->pfnHandlerRC);
2150	pHandler->pfnHandlerRC += offDelta;
2151	return 0;
2152	}
2153
2154
2155	/**
2156	* The VM is being reset.
2157	*
2158	* For the PGM component this means that any PD write monitors
2159	* needs to be removed.
2160	*
2161	* @param pVM VM handle.
2162	*/
2163	VMMR3DECL(void) PGMR3Reset(PVM pVM)
2164	{
2165	int rc;
2166
2167	LogFlow(("PGMR3Reset:\n"));
2168	VM_ASSERT_EMT(pVM);
2169
2170	pgmLock(pVM);
2171
2172	/*
2173	* Unfix any fixed mappings and disable CR3 monitoring.
2174	*/
2175	pVM->pgm.s.fMappingsFixed = false;
2176	pVM->pgm.s.GCPtrMappingFixed = 0;
2177	pVM->pgm.s.cbMappingFixed = 0;
2178
2179	/* Exit the guest paging mode before the pgm pool gets reset.
2180	* Important to clean up the amd64 case.
2181	*/
2182	for (VMCPUID i = 0; i < pVM->cCpus; i++)
2183	{
2184	PVMCPU pVCpu = &pVM->aCpus[i];
2185	rc = PGM_GST_PFN(Exit, pVCpu)(pVCpu);
2186	AssertRC(rc);
2187	}
2188
2189	#ifdef DEBUG
2190	DBGFR3InfoLog(pVM, "mappings", NULL);
2191	DBGFR3InfoLog(pVM, "handlers", "all nostat");
2192	#endif
2193
2194	/*
2195	* Switch mode back to real mode. (before resetting the pgm pool!)
2196	*/
2197	for (VMCPUID i = 0; i < pVM->cCpus; i++)
2198	{
2199	PVMCPU pVCpu = &pVM->aCpus[i];
2200
2201	rc = PGMR3ChangeMode(pVM, pVCpu, PGMMODE_REAL);
2202	AssertRC(rc);
2203
2204	STAM_REL_COUNTER_RESET(&pVCpu->pgm.s.cGuestModeChanges);
2205	}
2206
2207	/*
2208	* Reset the shadow page pool.
2209	*/
2210	pgmR3PoolReset(pVM);
2211
2212	for (VMCPUID i = 0; i < pVM->cCpus; i++)
2213	{
2214	PVMCPU pVCpu = &pVM->aCpus[i];
2215
2216	/*
2217	* Re-init other members.
2218	*/
2219	pVCpu->pgm.s.fA20Enabled = true;
2220
2221	/*
2222	* Clear the FFs PGM owns.
2223	*/
2224	VMCPU_FF_CLEAR(pVCpu, VMCPU_FF_PGM_SYNC_CR3);
2225	VMCPU_FF_CLEAR(pVCpu, VMCPU_FF_PGM_SYNC_CR3_NON_GLOBAL);
2226	}
2227
2228	/*
2229	* Reset (zero) RAM pages.
2230	*/
2231	rc = pgmR3PhysRamReset(pVM);
2232	if (RT_SUCCESS(rc))
2233	{
2234	/*
2235	* Reset (zero) shadow ROM pages.
2236	*/
2237	rc = pgmR3PhysRomReset(pVM);
2238	}
2239
2240	pgmUnlock(pVM);
2241	//return rc;
2242	AssertReleaseRC(rc);
2243	}
2244
2245
2246	#ifdef VBOX_STRICT
2247	/**
2248	* VM state change callback for clearing fNoMorePhysWrites after
2249	* a snapshot has been created.
2250	*/
2251	static DECLCALLBACK(void) pgmR3ResetNoMorePhysWritesFlag(PVM pVM, VMSTATE enmState, VMSTATE enmOldState, void *pvUser)
2252	{
2253	if ( enmState == VMSTATE_RUNNING
2254	\|\| enmState == VMSTATE_RESUMING)
2255	pVM->pgm.s.fNoMorePhysWrites = false;
2256	}
2257	#endif
2258
2259
2260	/**
2261	* Terminates the PGM.
2262	*
2263	* @returns VBox status code.
2264	* @param pVM Pointer to VM structure.
2265	*/
2266	VMMR3DECL(int) PGMR3Term(PVM pVM)
2267	{
2268	PGMDeregisterStringFormatTypes();
2269	return PDMR3CritSectDelete(&pVM->pgm.s.CritSect);
2270	}
2271
2272
2273	/**
2274	* Terminates the per-VCPU PGM.
2275	*
2276	* Termination means cleaning up and freeing all resources,
2277	* the VM it self is at this point powered off or suspended.
2278	*
2279	* @returns VBox status code.
2280	* @param pVM The VM to operate on.
2281	*/
2282	VMMR3DECL(int) PGMR3TermCPU(PVM pVM)
2283	{
2284	return 0;
2285	}
2286
2287
2288	/**
2289	* Show paging mode.
2290	*
2291	* @param pVM VM Handle.
2292	* @param pHlp The info helpers.
2293	* @param pszArgs "all" (default), "guest", "shadow" or "host".
2294	*/
2295	static DECLCALLBACK(void) pgmR3InfoMode(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2296	{
2297	/* digest argument. */
2298	bool fGuest, fShadow, fHost;
2299	if (pszArgs)
2300	pszArgs = RTStrStripL(pszArgs);
2301	if (!pszArgs \|\| !*pszArgs \|\| strstr(pszArgs, "all"))
2302	fShadow = fHost = fGuest = true;
2303	else
2304	{
2305	fShadow = fHost = fGuest = false;
2306	if (strstr(pszArgs, "guest"))
2307	fGuest = true;
2308	if (strstr(pszArgs, "shadow"))
2309	fShadow = true;
2310	if (strstr(pszArgs, "host"))
2311	fHost = true;
2312	}
2313
2314	/** @todo SMP support! */
2315	/* print info. */
2316	if (fGuest)
2317	pHlp->pfnPrintf(pHlp, "Guest paging mode: %s, changed %RU64 times, A20 %s\n",
2318	PGMGetModeName(pVM->aCpus[0].pgm.s.enmGuestMode), pVM->aCpus[0].pgm.s.cGuestModeChanges.c,
2319	pVM->aCpus[0].pgm.s.fA20Enabled ? "enabled" : "disabled");
2320	if (fShadow)
2321	pHlp->pfnPrintf(pHlp, "Shadow paging mode: %s\n", PGMGetModeName(pVM->aCpus[0].pgm.s.enmShadowMode));
2322	if (fHost)
2323	{
2324	const char *psz;
2325	switch (pVM->pgm.s.enmHostMode)
2326	{
2327	case SUPPAGINGMODE_INVALID: psz = "invalid"; break;
2328	case SUPPAGINGMODE_32_BIT: psz = "32-bit"; break;
2329	case SUPPAGINGMODE_32_BIT_GLOBAL: psz = "32-bit+G"; break;
2330	case SUPPAGINGMODE_PAE: psz = "PAE"; break;
2331	case SUPPAGINGMODE_PAE_GLOBAL: psz = "PAE+G"; break;
2332	case SUPPAGINGMODE_PAE_NX: psz = "PAE+NX"; break;
2333	case SUPPAGINGMODE_PAE_GLOBAL_NX: psz = "PAE+G+NX"; break;
2334	case SUPPAGINGMODE_AMD64: psz = "AMD64"; break;
2335	case SUPPAGINGMODE_AMD64_GLOBAL: psz = "AMD64+G"; break;
2336	case SUPPAGINGMODE_AMD64_NX: psz = "AMD64+NX"; break;
2337	case SUPPAGINGMODE_AMD64_GLOBAL_NX: psz = "AMD64+G+NX"; break;
2338	default: psz = "unknown"; break;
2339	}
2340	pHlp->pfnPrintf(pHlp, "Host paging mode: %s\n", psz);
2341	}
2342	}
2343
2344
2345	/**
2346	* Dump registered MMIO ranges to the log.
2347	*
2348	* @param pVM VM Handle.
2349	* @param pHlp The info helpers.
2350	* @param pszArgs Arguments, ignored.
2351	*/
2352	static DECLCALLBACK(void) pgmR3PhysInfo(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2353	{
2354	NOREF(pszArgs);
2355	pHlp->pfnPrintf(pHlp,
2356	"RAM ranges (pVM=%p)\n"
2357	"%.s %.s\n",
2358	pVM,
2359	sizeof(RTGCPHYS) * 4 + 1, "GC Phys Range ",
2360	sizeof(RTHCPTR) * 2, "pvHC ");
2361
2362	for (PPGMRAMRANGE pCur = pVM->pgm.s.pRamRangesR3; pCur; pCur = pCur->pNextR3)
2363	pHlp->pfnPrintf(pHlp,
2364	"%RGp-%RGp %RHv %s\n",
2365	pCur->GCPhys,
2366	pCur->GCPhysLast,
2367	pCur->pvR3,
2368	pCur->pszDesc);
2369	}
2370
2371	/**
2372	* Dump the page directory to the log.
2373	*
2374	* @param pVM VM Handle.
2375	* @param pHlp The info helpers.
2376	* @param pszArgs Arguments, ignored.
2377	*/
2378	static DECLCALLBACK(void) pgmR3InfoCr3(PVM pVM, PCDBGFINFOHLP pHlp, const char *pszArgs)
2379	{
2380	/** @todo SMP support!! */
2381	PVMCPU pVCpu = &pVM->aCpus[0];
2382
2383	/** @todo fix this! Convert the PGMR3DumpHierarchyHC functions to do guest stuff. */
2384	/* Big pages supported? */
2385	const bool fPSE = !!(CPUMGetGuestCR4(pVCpu) & X86_CR4_PSE);
2386
2387	/* Global pages supported? */
2388	const bool fPGE = !!(CPUMGetGuestCR4(pVCpu) & X86_CR4_PGE);
2389
2390	NOREF(pszArgs);
2391
2392	/*
2393	* Get page directory addresses.
2394	*/
2395	PX86PD pPDSrc = pgmGstGet32bitPDPtr(&pVCpu->pgm.s);
2396	Assert(pPDSrc);
2397	Assert(PGMPhysGCPhys2R3PtrAssert(pVM, (RTGCPHYS)(CPUMGetGuestCR3(pVCpu) & X86_CR3_PAGE_MASK), sizeof(*pPDSrc)) == pPDSrc);
2398
2399	/*
2400	* Iterate the page directory.
2401	*/
2402	for (unsigned iPD = 0; iPD < RT_ELEMENTS(pPDSrc->a); iPD++)
2403	{
2404	X86PDE PdeSrc = pPDSrc->a[iPD];
2405	if (PdeSrc.n.u1Present)
2406	{
2407	if (PdeSrc.b.u1Size && fPSE)
2408	pHlp->pfnPrintf(pHlp,
2409	"%04X - %RGp P=%d U=%d RW=%d G=%d - BIG\n",
2410	iPD,
2411	pgmGstGet4MBPhysPage(&pVM->pgm.s, PdeSrc),
2412	PdeSrc.b.u1Present, PdeSrc.b.u1User, PdeSrc.b.u1Write, PdeSrc.b.u1Global && fPGE);
2413	else
2414	pHlp->pfnPrintf(pHlp,
2415	"%04X - %RGp P=%d U=%d RW=%d [G=%d]\n",
2416	iPD,
2417	(RTGCPHYS)(PdeSrc.u & X86_PDE_PG_MASK),
2418	PdeSrc.n.u1Present, PdeSrc.n.u1User, PdeSrc.n.u1Write, PdeSrc.b.u1Global && fPGE);
2419	}
2420	}
2421	}
2422
2423
2424	/**
2425	* Service a VMMCALLRING3_PGM_LOCK call.
2426	*
2427	* @returns VBox status code.
2428	* @param pVM The VM handle.
2429	*/
2430	VMMR3DECL(int) PGMR3LockCall(PVM pVM)
2431	{
2432	int rc = PDMR3CritSectEnterEx(&pVM->pgm.s.CritSect, true /* fHostCall */);
2433	AssertRC(rc);
2434	return rc;
2435	}
2436
2437
2438	/**
2439	* Converts a PGMMODE value to a PGM_TYPE_* \#define.
2440	*
2441	* @returns PGM_TYPE_*.
2442	* @param pgmMode The mode value to convert.
2443	*/
2444	DECLINLINE(unsigned) pgmModeToType(PGMMODE pgmMode)
2445	{
2446	switch (pgmMode)
2447	{
2448	case PGMMODE_REAL: return PGM_TYPE_REAL;
2449	case PGMMODE_PROTECTED: return PGM_TYPE_PROT;
2450	case PGMMODE_32_BIT: return PGM_TYPE_32BIT;
2451	case PGMMODE_PAE:
2452	case PGMMODE_PAE_NX: return PGM_TYPE_PAE;
2453	case PGMMODE_AMD64:
2454	case PGMMODE_AMD64_NX: return PGM_TYPE_AMD64;
2455	case PGMMODE_NESTED: return PGM_TYPE_NESTED;
2456	case PGMMODE_EPT: return PGM_TYPE_EPT;
2457	default:
2458	AssertFatalMsgFailed(("pgmMode=%d\n", pgmMode));
2459	}
2460	}
2461
2462
2463	/**
2464	* Gets the index into the paging mode data array of a SHW+GST mode.
2465	*
2466	* @returns PGM::paPagingData index.
2467	* @param uShwType The shadow paging mode type.
2468	* @param uGstType The guest paging mode type.
2469	*/
2470	DECLINLINE(unsigned) pgmModeDataIndex(unsigned uShwType, unsigned uGstType)
2471	{
2472	Assert(uShwType >= PGM_TYPE_32BIT && uShwType <= PGM_TYPE_MAX);
2473	Assert(uGstType >= PGM_TYPE_REAL && uGstType <= PGM_TYPE_AMD64);
2474	return (uShwType - PGM_TYPE_32BIT) * (PGM_TYPE_AMD64 - PGM_TYPE_REAL + 1)
2475	+ (uGstType - PGM_TYPE_REAL);
2476	}
2477
2478
2479	/**
2480	* Gets the index into the paging mode data array of a SHW+GST mode.
2481	*
2482	* @returns PGM::paPagingData index.
2483	* @param enmShw The shadow paging mode.
2484	* @param enmGst The guest paging mode.
2485	*/
2486	DECLINLINE(unsigned) pgmModeDataIndexByMode(PGMMODE enmShw, PGMMODE enmGst)
2487	{
2488	Assert(enmShw >= PGMMODE_32_BIT && enmShw <= PGMMODE_MAX);
2489	Assert(enmGst > PGMMODE_INVALID && enmGst < PGMMODE_MAX);
2490	return pgmModeDataIndex(pgmModeToType(enmShw), pgmModeToType(enmGst));
2491	}
2492
2493
2494	/**
2495	* Calculates the max data index.
2496	* @returns The number of entries in the paging data array.
2497	*/
2498	DECLINLINE(unsigned) pgmModeDataMaxIndex(void)
2499	{
2500	return pgmModeDataIndex(PGM_TYPE_MAX, PGM_TYPE_AMD64) + 1;
2501	}
2502
2503
2504	/**
2505	* Initializes the paging mode data kept in PGM::paModeData.
2506	*
2507	* @param pVM The VM handle.
2508	* @param fResolveGCAndR0 Indicate whether or not GC and Ring-0 symbols can be resolved now.
2509	* This is used early in the init process to avoid trouble with PDM
2510	* not being initialized yet.
2511	*/
2512	static int pgmR3ModeDataInit(PVM pVM, bool fResolveGCAndR0)
2513	{
2514	PPGMMODEDATA pModeData;
2515	int rc;
2516
2517	/*
2518	* Allocate the array on the first call.
2519	*/
2520	if (!pVM->pgm.s.paModeData)
2521	{
2522	pVM->pgm.s.paModeData = (PPGMMODEDATA)MMR3HeapAllocZ(pVM, MM_TAG_PGM, sizeof(PGMMODEDATA) * pgmModeDataMaxIndex());
2523	AssertReturn(pVM->pgm.s.paModeData, VERR_NO_MEMORY);
2524	}
2525
2526	/*
2527	* Initialize the array entries.
2528	*/
2529	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGM_TYPE_REAL)];
2530	pModeData->uShwType = PGM_TYPE_32BIT;
2531	pModeData->uGstType = PGM_TYPE_REAL;
2532	rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2533	rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2534	rc = PGM_BTH_NAME_32BIT_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2535
2536	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGMMODE_PROTECTED)];
2537	pModeData->uShwType = PGM_TYPE_32BIT;
2538	pModeData->uGstType = PGM_TYPE_PROT;
2539	rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2540	rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2541	rc = PGM_BTH_NAME_32BIT_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2542
2543	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_32BIT, PGM_TYPE_32BIT)];
2544	pModeData->uShwType = PGM_TYPE_32BIT;
2545	pModeData->uGstType = PGM_TYPE_32BIT;
2546	rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2547	rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2548	rc = PGM_BTH_NAME_32BIT_32BIT(InitData)(pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2549
2550	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_REAL)];
2551	pModeData->uShwType = PGM_TYPE_PAE;
2552	pModeData->uGstType = PGM_TYPE_REAL;
2553	rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2554	rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2555	rc = PGM_BTH_NAME_PAE_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2556
2557	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_PROT)];
2558	pModeData->uShwType = PGM_TYPE_PAE;
2559	pModeData->uGstType = PGM_TYPE_PROT;
2560	rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2561	rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2562	rc = PGM_BTH_NAME_PAE_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2563
2564	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_32BIT)];
2565	pModeData->uShwType = PGM_TYPE_PAE;
2566	pModeData->uGstType = PGM_TYPE_32BIT;
2567	rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2568	rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2569	rc = PGM_BTH_NAME_PAE_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2570
2571	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_PAE, PGM_TYPE_PAE)];
2572	pModeData->uShwType = PGM_TYPE_PAE;
2573	pModeData->uGstType = PGM_TYPE_PAE;
2574	rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2575	rc = PGM_GST_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2576	rc = PGM_BTH_NAME_PAE_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2577
2578	#ifdef VBOX_WITH_64_BITS_GUESTS
2579	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_AMD64, PGM_TYPE_AMD64)];
2580	pModeData->uShwType = PGM_TYPE_AMD64;
2581	pModeData->uGstType = PGM_TYPE_AMD64;
2582	rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2583	rc = PGM_GST_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2584	rc = PGM_BTH_NAME_AMD64_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2585	#endif
2586
2587	/* The nested paging mode. */
2588	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_REAL)];
2589	pModeData->uShwType = PGM_TYPE_NESTED;
2590	pModeData->uGstType = PGM_TYPE_REAL;
2591	rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2592	rc = PGM_BTH_NAME_NESTED_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2593
2594	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGMMODE_PROTECTED)];
2595	pModeData->uShwType = PGM_TYPE_NESTED;
2596	pModeData->uGstType = PGM_TYPE_PROT;
2597	rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2598	rc = PGM_BTH_NAME_NESTED_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2599
2600	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_32BIT)];
2601	pModeData->uShwType = PGM_TYPE_NESTED;
2602	pModeData->uGstType = PGM_TYPE_32BIT;
2603	rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2604	rc = PGM_BTH_NAME_NESTED_32BIT(InitData)(pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2605
2606	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_PAE)];
2607	pModeData->uShwType = PGM_TYPE_NESTED;
2608	pModeData->uGstType = PGM_TYPE_PAE;
2609	rc = PGM_GST_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2610	rc = PGM_BTH_NAME_NESTED_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2611
2612	#ifdef VBOX_WITH_64_BITS_GUESTS
2613	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_AMD64)];
2614	pModeData->uShwType = PGM_TYPE_NESTED;
2615	pModeData->uGstType = PGM_TYPE_AMD64;
2616	rc = PGM_GST_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2617	rc = PGM_BTH_NAME_NESTED_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2618	#endif
2619
2620	/* The shadow part of the nested callback mode depends on the host paging mode (AMD-V only). */
2621	switch (pVM->pgm.s.enmHostMode)
2622	{
2623	#if HC_ARCH_BITS == 32
2624	case SUPPAGINGMODE_32_BIT:
2625	case SUPPAGINGMODE_32_BIT_GLOBAL:
2626	for (unsigned i = PGM_TYPE_REAL; i <= PGM_TYPE_PAE; i++)
2627	{
2628	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2629	rc = PGM_SHW_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2630	}
2631	# ifdef VBOX_WITH_64_BITS_GUESTS
2632	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_AMD64)];
2633	rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2634	# endif
2635	break;
2636
2637	case SUPPAGINGMODE_PAE:
2638	case SUPPAGINGMODE_PAE_NX:
2639	case SUPPAGINGMODE_PAE_GLOBAL:
2640	case SUPPAGINGMODE_PAE_GLOBAL_NX:
2641	for (unsigned i = PGM_TYPE_REAL; i <= PGM_TYPE_PAE; i++)
2642	{
2643	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2644	rc = PGM_SHW_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2645	}
2646	# ifdef VBOX_WITH_64_BITS_GUESTS
2647	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, PGM_TYPE_AMD64)];
2648	rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2649	# endif
2650	break;
2651	#endif /* HC_ARCH_BITS == 32 */
2652
2653	#if HC_ARCH_BITS == 64 \|\| defined(RT_OS_DARWIN)
2654	case SUPPAGINGMODE_AMD64:
2655	case SUPPAGINGMODE_AMD64_GLOBAL:
2656	case SUPPAGINGMODE_AMD64_NX:
2657	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2658	# ifdef VBOX_WITH_64_BITS_GUESTS
2659	for (unsigned i = PGM_TYPE_REAL; i <= PGM_TYPE_AMD64; i++)
2660	# else
2661	for (unsigned i = PGM_TYPE_REAL; i <= PGM_TYPE_PAE; i++)
2662	# endif
2663	{
2664	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_NESTED, i)];
2665	rc = PGM_SHW_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2666	}
2667	break;
2668	#endif /* HC_ARCH_BITS == 64 \|\| RT_OS_DARWIN */
2669
2670	default:
2671	AssertFailed();
2672	break;
2673	}
2674
2675	/* Extended paging (EPT) / Intel VT-x */
2676	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_EPT, PGM_TYPE_REAL)];
2677	pModeData->uShwType = PGM_TYPE_EPT;
2678	pModeData->uGstType = PGM_TYPE_REAL;
2679	rc = PGM_SHW_NAME_EPT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2680	rc = PGM_GST_NAME_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2681	rc = PGM_BTH_NAME_EPT_REAL(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2682
2683	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_EPT, PGM_TYPE_PROT)];
2684	pModeData->uShwType = PGM_TYPE_EPT;
2685	pModeData->uGstType = PGM_TYPE_PROT;
2686	rc = PGM_SHW_NAME_EPT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2687	rc = PGM_GST_NAME_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2688	rc = PGM_BTH_NAME_EPT_PROT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2689
2690	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_EPT, PGM_TYPE_32BIT)];
2691	pModeData->uShwType = PGM_TYPE_EPT;
2692	pModeData->uGstType = PGM_TYPE_32BIT;
2693	rc = PGM_SHW_NAME_EPT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2694	rc = PGM_GST_NAME_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2695	rc = PGM_BTH_NAME_EPT_32BIT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2696
2697	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_EPT, PGM_TYPE_PAE)];
2698	pModeData->uShwType = PGM_TYPE_EPT;
2699	pModeData->uGstType = PGM_TYPE_PAE;
2700	rc = PGM_SHW_NAME_EPT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2701	rc = PGM_GST_NAME_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2702	rc = PGM_BTH_NAME_EPT_PAE(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2703
2704	#ifdef VBOX_WITH_64_BITS_GUESTS
2705	pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndex(PGM_TYPE_EPT, PGM_TYPE_AMD64)];
2706	pModeData->uShwType = PGM_TYPE_EPT;
2707	pModeData->uGstType = PGM_TYPE_AMD64;
2708	rc = PGM_SHW_NAME_EPT(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2709	rc = PGM_GST_NAME_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2710	rc = PGM_BTH_NAME_EPT_AMD64(InitData)( pVM, pModeData, fResolveGCAndR0); AssertRCReturn(rc, rc);
2711	#endif
2712	return VINF_SUCCESS;
2713	}
2714
2715
2716	/**
2717	* Switch to different (or relocated in the relocate case) mode data.
2718	*
2719	* @param pVM The VM handle.
2720	* @param pVCpu The VMCPU to operate on.
2721	* @param enmShw The the shadow paging mode.
2722	* @param enmGst The the guest paging mode.
2723	*/
2724	static void pgmR3ModeDataSwitch(PVM pVM, PVMCPU pVCpu, PGMMODE enmShw, PGMMODE enmGst)
2725	{
2726	PPGMMODEDATA pModeData = &pVM->pgm.s.paModeData[pgmModeDataIndexByMode(enmShw, enmGst)];
2727
2728	Assert(pModeData->uGstType == pgmModeToType(enmGst));
2729	Assert(pModeData->uShwType == pgmModeToType(enmShw));
2730
2731	/* shadow */
2732	pVCpu->pgm.s.pfnR3ShwRelocate = pModeData->pfnR3ShwRelocate;
2733	pVCpu->pgm.s.pfnR3ShwExit = pModeData->pfnR3ShwExit;
2734	pVCpu->pgm.s.pfnR3ShwGetPage = pModeData->pfnR3ShwGetPage;
2735	Assert(pVCpu->pgm.s.pfnR3ShwGetPage);
2736	pVCpu->pgm.s.pfnR3ShwModifyPage = pModeData->pfnR3ShwModifyPage;
2737
2738	pVCpu->pgm.s.pfnRCShwGetPage = pModeData->pfnRCShwGetPage;
2739	pVCpu->pgm.s.pfnRCShwModifyPage = pModeData->pfnRCShwModifyPage;
2740
2741	pVCpu->pgm.s.pfnR0ShwGetPage = pModeData->pfnR0ShwGetPage;
2742	pVCpu->pgm.s.pfnR0ShwModifyPage = pModeData->pfnR0ShwModifyPage;
2743
2744
2745	/* guest */
2746	pVCpu->pgm.s.pfnR3GstRelocate = pModeData->pfnR3GstRelocate;
2747	pVCpu->pgm.s.pfnR3GstExit = pModeData->pfnR3GstExit;
2748	pVCpu->pgm.s.pfnR3GstGetPage = pModeData->pfnR3GstGetPage;
2749	Assert(pVCpu->pgm.s.pfnR3GstGetPage);
2750	pVCpu->pgm.s.pfnR3GstModifyPage = pModeData->pfnR3GstModifyPage;
2751	pVCpu->pgm.s.pfnR3GstGetPDE = pModeData->pfnR3GstGetPDE;
2752	pVCpu->pgm.s.pfnRCGstGetPage = pModeData->pfnRCGstGetPage;
2753	pVCpu->pgm.s.pfnRCGstModifyPage = pModeData->pfnRCGstModifyPage;
2754	pVCpu->pgm.s.pfnRCGstGetPDE = pModeData->pfnRCGstGetPDE;
2755	pVCpu->pgm.s.pfnR0GstGetPage = pModeData->pfnR0GstGetPage;
2756	pVCpu->pgm.s.pfnR0GstModifyPage = pModeData->pfnR0GstModifyPage;
2757	pVCpu->pgm.s.pfnR0GstGetPDE = pModeData->pfnR0GstGetPDE;
2758
2759	/* both */
2760	pVCpu->pgm.s.pfnR3BthRelocate = pModeData->pfnR3BthRelocate;
2761	pVCpu->pgm.s.pfnR3BthInvalidatePage = pModeData->pfnR3BthInvalidatePage;
2762	pVCpu->pgm.s.pfnR3BthSyncCR3 = pModeData->pfnR3BthSyncCR3;
2763	Assert(pVCpu->pgm.s.pfnR3BthSyncCR3);
2764	pVCpu->pgm.s.pfnR3BthSyncPage = pModeData->pfnR3BthSyncPage;
2765	pVCpu->pgm.s.pfnR3BthPrefetchPage = pModeData->pfnR3BthPrefetchPage;
2766	pVCpu->pgm.s.pfnR3BthVerifyAccessSyncPage = pModeData->pfnR3BthVerifyAccessSyncPage;
2767	#ifdef VBOX_STRICT
2768	pVCpu->pgm.s.pfnR3BthAssertCR3 = pModeData->pfnR3BthAssertCR3;
2769	#endif
2770	pVCpu->pgm.s.pfnR3BthMapCR3 = pModeData->pfnR3BthMapCR3;
2771	pVCpu->pgm.s.pfnR3BthUnmapCR3 = pModeData->pfnR3BthUnmapCR3;
2772
2773	pVCpu->pgm.s.pfnRCBthTrap0eHandler = pModeData->pfnRCBthTrap0eHandler;
2774	pVCpu->pgm.s.pfnRCBthInvalidatePage = pModeData->pfnRCBthInvalidatePage;
2775	pVCpu->pgm.s.pfnRCBthSyncCR3 = pModeData->pfnRCBthSyncCR3;
2776	pVCpu->pgm.s.pfnRCBthSyncPage = pModeData->pfnRCBthSyncPage;
2777	pVCpu->pgm.s.pfnRCBthPrefetchPage = pModeData->pfnRCBthPrefetchPage;
2778	pVCpu->pgm.s.pfnRCBthVerifyAccessSyncPage = pModeData->pfnRCBthVerifyAccessSyncPage;
2779	#ifdef VBOX_STRICT
2780	pVCpu->pgm.s.pfnRCBthAssertCR3 = pModeData->pfnRCBthAssertCR3;
2781	#endif
2782	pVCpu->pgm.s.pfnRCBthMapCR3 = pModeData->pfnRCBthMapCR3;
2783	pVCpu->pgm.s.pfnRCBthUnmapCR3 = pModeData->pfnRCBthUnmapCR3;
2784
2785	pVCpu->pgm.s.pfnR0BthTrap0eHandler = pModeData->pfnR0BthTrap0eHandler;
2786	pVCpu->pgm.s.pfnR0BthInvalidatePage = pModeData->pfnR0BthInvalidatePage;
2787	pVCpu->pgm.s.pfnR0BthSyncCR3 = pModeData->pfnR0BthSyncCR3;
2788	pVCpu->pgm.s.pfnR0BthSyncPage = pModeData->pfnR0BthSyncPage;
2789	pVCpu->pgm.s.pfnR0BthPrefetchPage = pModeData->pfnR0BthPrefetchPage;
2790	pVCpu->pgm.s.pfnR0BthVerifyAccessSyncPage = pModeData->pfnR0BthVerifyAccessSyncPage;
2791	#ifdef VBOX_STRICT
2792	pVCpu->pgm.s.pfnR0BthAssertCR3 = pModeData->pfnR0BthAssertCR3;
2793	#endif
2794	pVCpu->pgm.s.pfnR0BthMapCR3 = pModeData->pfnR0BthMapCR3;
2795	pVCpu->pgm.s.pfnR0BthUnmapCR3 = pModeData->pfnR0BthUnmapCR3;
2796	}
2797
2798
2799	/**
2800	* Calculates the shadow paging mode.
2801	*
2802	* @returns The shadow paging mode.
2803	* @param pVM VM handle.
2804	* @param enmGuestMode The guest mode.
2805	* @param enmHostMode The host mode.
2806	* @param enmShadowMode The current shadow mode.
2807	* @param penmSwitcher Where to store the switcher to use.
2808	* VMMSWITCHER_INVALID means no change.
2809	*/
2810	static PGMMODE pgmR3CalcShadowMode(PVM pVM, PGMMODE enmGuestMode, SUPPAGINGMODE enmHostMode, PGMMODE enmShadowMode, VMMSWITCHER *penmSwitcher)
2811	{
2812	VMMSWITCHER enmSwitcher = VMMSWITCHER_INVALID;
2813	switch (enmGuestMode)
2814	{
2815	/*
2816	* When switching to real or protected mode we don't change
2817	* anything since it's likely that we'll switch back pretty soon.
2818	*
2819	* During pgmR3InitPaging we'll end up here with PGMMODE_INVALID
2820	* and is supposed to determine which shadow paging and switcher to
2821	* use during init.
2822	*/
2823	case PGMMODE_REAL:
2824	case PGMMODE_PROTECTED:
2825	if ( enmShadowMode != PGMMODE_INVALID
2826	&& !HWACCMIsEnabled(pVM) /* always switch in hwaccm mode! */)
2827	break; /* (no change) */
2828
2829	switch (enmHostMode)
2830	{
2831	case SUPPAGINGMODE_32_BIT:
2832	case SUPPAGINGMODE_32_BIT_GLOBAL:
2833	enmShadowMode = PGMMODE_32_BIT;
2834	enmSwitcher = VMMSWITCHER_32_TO_32;
2835	break;
2836
2837	case SUPPAGINGMODE_PAE:
2838	case SUPPAGINGMODE_PAE_NX:
2839	case SUPPAGINGMODE_PAE_GLOBAL:
2840	case SUPPAGINGMODE_PAE_GLOBAL_NX:
2841	enmShadowMode = PGMMODE_PAE;
2842	enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2843	#ifdef DEBUG_bird
2844	if (RTEnvExist("VBOX_32BIT"))
2845	{
2846	enmShadowMode = PGMMODE_32_BIT;
2847	enmSwitcher = VMMSWITCHER_PAE_TO_32;
2848	}
2849	#endif
2850	break;
2851
2852	case SUPPAGINGMODE_AMD64:
2853	case SUPPAGINGMODE_AMD64_GLOBAL:
2854	case SUPPAGINGMODE_AMD64_NX:
2855	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2856	enmShadowMode = PGMMODE_PAE;
2857	enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2858	#ifdef DEBUG_bird
2859	if (RTEnvExist("VBOX_32BIT"))
2860	{
2861	enmShadowMode = PGMMODE_32_BIT;
2862	enmSwitcher = VMMSWITCHER_AMD64_TO_32;
2863	}
2864	#endif
2865	break;
2866
2867	default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2868	}
2869	break;
2870
2871	case PGMMODE_32_BIT:
2872	switch (enmHostMode)
2873	{
2874	case SUPPAGINGMODE_32_BIT:
2875	case SUPPAGINGMODE_32_BIT_GLOBAL:
2876	enmShadowMode = PGMMODE_32_BIT;
2877	enmSwitcher = VMMSWITCHER_32_TO_32;
2878	break;
2879
2880	case SUPPAGINGMODE_PAE:
2881	case SUPPAGINGMODE_PAE_NX:
2882	case SUPPAGINGMODE_PAE_GLOBAL:
2883	case SUPPAGINGMODE_PAE_GLOBAL_NX:
2884	enmShadowMode = PGMMODE_PAE;
2885	enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2886	#ifdef DEBUG_bird
2887	if (RTEnvExist("VBOX_32BIT"))
2888	{
2889	enmShadowMode = PGMMODE_32_BIT;
2890	enmSwitcher = VMMSWITCHER_PAE_TO_32;
2891	}
2892	#endif
2893	break;
2894
2895	case SUPPAGINGMODE_AMD64:
2896	case SUPPAGINGMODE_AMD64_GLOBAL:
2897	case SUPPAGINGMODE_AMD64_NX:
2898	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2899	enmShadowMode = PGMMODE_PAE;
2900	enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2901	#ifdef DEBUG_bird
2902	if (RTEnvExist("VBOX_32BIT"))
2903	{
2904	enmShadowMode = PGMMODE_32_BIT;
2905	enmSwitcher = VMMSWITCHER_AMD64_TO_32;
2906	}
2907	#endif
2908	break;
2909
2910	default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2911	}
2912	break;
2913
2914	case PGMMODE_PAE:
2915	case PGMMODE_PAE_NX: /** @todo This might require more switchers and guest+both modes. */
2916	switch (enmHostMode)
2917	{
2918	case SUPPAGINGMODE_32_BIT:
2919	case SUPPAGINGMODE_32_BIT_GLOBAL:
2920	enmShadowMode = PGMMODE_PAE;
2921	enmSwitcher = VMMSWITCHER_32_TO_PAE;
2922	break;
2923
2924	case SUPPAGINGMODE_PAE:
2925	case SUPPAGINGMODE_PAE_NX:
2926	case SUPPAGINGMODE_PAE_GLOBAL:
2927	case SUPPAGINGMODE_PAE_GLOBAL_NX:
2928	enmShadowMode = PGMMODE_PAE;
2929	enmSwitcher = VMMSWITCHER_PAE_TO_PAE;
2930	break;
2931
2932	case SUPPAGINGMODE_AMD64:
2933	case SUPPAGINGMODE_AMD64_GLOBAL:
2934	case SUPPAGINGMODE_AMD64_NX:
2935	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2936	enmShadowMode = PGMMODE_PAE;
2937	enmSwitcher = VMMSWITCHER_AMD64_TO_PAE;
2938	break;
2939
2940	default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2941	}
2942	break;
2943
2944	case PGMMODE_AMD64:
2945	case PGMMODE_AMD64_NX:
2946	switch (enmHostMode)
2947	{
2948	case SUPPAGINGMODE_32_BIT:
2949	case SUPPAGINGMODE_32_BIT_GLOBAL:
2950	enmShadowMode = PGMMODE_AMD64;
2951	enmSwitcher = VMMSWITCHER_32_TO_AMD64;
2952	break;
2953
2954	case SUPPAGINGMODE_PAE:
2955	case SUPPAGINGMODE_PAE_NX:
2956	case SUPPAGINGMODE_PAE_GLOBAL:
2957	case SUPPAGINGMODE_PAE_GLOBAL_NX:
2958	enmShadowMode = PGMMODE_AMD64;
2959	enmSwitcher = VMMSWITCHER_PAE_TO_AMD64;
2960	break;
2961
2962	case SUPPAGINGMODE_AMD64:
2963	case SUPPAGINGMODE_AMD64_GLOBAL:
2964	case SUPPAGINGMODE_AMD64_NX:
2965	case SUPPAGINGMODE_AMD64_GLOBAL_NX:
2966	enmShadowMode = PGMMODE_AMD64;
2967	enmSwitcher = VMMSWITCHER_AMD64_TO_AMD64;
2968	break;
2969
2970	default: AssertMsgFailed(("enmHostMode=%d\n", enmHostMode)); break;
2971	}
2972	break;
2973
2974
2975	default:
2976	AssertReleaseMsgFailed(("enmGuestMode=%d\n", enmGuestMode));
2977	return PGMMODE_INVALID;
2978	}
2979	/* Override the shadow mode is nested paging is active. */
2980	if (HWACCMIsNestedPagingActive(pVM))
2981	enmShadowMode = HWACCMGetShwPagingMode(pVM);
2982
2983	*penmSwitcher = enmSwitcher;
2984	return enmShadowMode;
2985	}
2986
2987
2988	/**
2989	* Performs the actual mode change.
2990	* This is called by PGMChangeMode and pgmR3InitPaging().
2991	*
2992	* @returns VBox status code. May suspend or power off the VM on error, but this
2993	* will trigger using FFs and not status codes.
2994	*
2995	* @param pVM VM handle.
2996	* @param pVCpu The VMCPU to operate on.
2997	* @param enmGuestMode The new guest mode. This is assumed to be different from
2998	* the current mode.
2999	*/
3000	VMMR3DECL(int) PGMR3ChangeMode(PVM pVM, PVMCPU pVCpu, PGMMODE enmGuestMode)
3001	{
3002	bool fIsOldGuestPagingMode64Bits = (pVCpu->pgm.s.enmGuestMode >= PGMMODE_AMD64);
3003	bool fIsNewGuestPagingMode64Bits = (enmGuestMode >= PGMMODE_AMD64);
3004
3005	Log(("PGMR3ChangeMode: Guest mode: %s -> %s\n", PGMGetModeName(pVCpu->pgm.s.enmGuestMode), PGMGetModeName(enmGuestMode)));
3006	STAM_REL_COUNTER_INC(&pVCpu->pgm.s.cGuestModeChanges);
3007
3008	/*
3009	* Calc the shadow mode and switcher.
3010	*/
3011	VMMSWITCHER enmSwitcher;
3012	PGMMODE enmShadowMode = pgmR3CalcShadowMode(pVM, enmGuestMode, pVM->pgm.s.enmHostMode, pVCpu->pgm.s.enmShadowMode, &enmSwitcher);
3013	if (enmSwitcher != VMMSWITCHER_INVALID)
3014	{
3015	/*
3016	* Select new switcher.
3017	*/
3018	int rc = VMMR3SelectSwitcher(pVM, enmSwitcher);
3019	if (RT_FAILURE(rc))
3020	{
3021	AssertReleaseMsgFailed(("VMMR3SelectSwitcher(%d) -> %Rrc\n", enmSwitcher, rc));
3022	return rc;
3023	}
3024	}
3025
3026	/*
3027	* Exit old mode(s).
3028	*/
3029	#if HC_ARCH_BITS == 32
3030	/* The nested shadow paging mode for AMD-V does change when running 64 bits guests on 32 bits hosts; typically PAE <-> AMD64 */
3031	const bool fForceShwEnterExit = ( fIsOldGuestPagingMode64Bits != fIsNewGuestPagingMode64Bits
3032	&& enmShadowMode == PGMMODE_NESTED);
3033	#else
3034	const bool fForceShwEnterExit = false;
3035	#endif
3036	/* shadow */
3037	if ( enmShadowMode != pVCpu->pgm.s.enmShadowMode
3038	\|\| fForceShwEnterExit)
3039	{
3040	LogFlow(("PGMR3ChangeMode: Shadow mode: %s -> %s\n", PGMGetModeName(pVCpu->pgm.s.enmShadowMode), PGMGetModeName(enmShadowMode)));
3041	if (PGM_SHW_PFN(Exit, pVCpu))
3042	{
3043	int rc = PGM_SHW_PFN(Exit, pVCpu)(pVCpu);
3044	if (RT_FAILURE(rc))
3045	{
3046	AssertMsgFailed(("Exit failed for shadow mode %d: %Rrc\n", pVCpu->pgm.s.enmShadowMode, rc));
3047	return rc;
3048	}
3049	}
3050
3051	}
3052	else
3053	LogFlow(("PGMR3ChangeMode: Shadow mode remains: %s\n", PGMGetModeName(pVCpu->pgm.s.enmShadowMode)));
3054
3055	/* guest */
3056	if (PGM_GST_PFN(Exit, pVCpu))
3057	{
3058	int rc = PGM_GST_PFN(Exit, pVCpu)(pVCpu);
3059	if (RT_FAILURE(rc))
3060	{
3061	AssertMsgFailed(("Exit failed for guest mode %d: %Rrc\n", pVCpu->pgm.s.enmGuestMode, rc));
3062	return rc;
3063	}
3064	}
3065
3066	/*
3067	* Load new paging mode data.
3068	*/
3069	pgmR3ModeDataSwitch(pVM, pVCpu, enmShadowMode, enmGuestMode);
3070
3071	/*
3072	* Enter new shadow mode (if changed).
3073	*/
3074	if ( enmShadowMode != pVCpu->pgm.s.enmShadowMode
3075	\|\| fForceShwEnterExit)
3076	{
3077	int rc;
3078	pVCpu->pgm.s.enmShadowMode = enmShadowMode;
3079	switch (enmShadowMode)
3080	{
3081	case PGMMODE_32_BIT:
3082	rc = PGM_SHW_NAME_32BIT(Enter)(pVCpu, false);
3083	break;
3084	case PGMMODE_PAE:
3085	case PGMMODE_PAE_NX:
3086	rc = PGM_SHW_NAME_PAE(Enter)(pVCpu, false);
3087	break;
3088	case PGMMODE_AMD64:
3089	case PGMMODE_AMD64_NX:
3090	rc = PGM_SHW_NAME_AMD64(Enter)(pVCpu, fIsNewGuestPagingMode64Bits);
3091	break;
3092	case PGMMODE_NESTED:
3093	rc = PGM_SHW_NAME_NESTED(Enter)(pVCpu, fIsNewGuestPagingMode64Bits);
3094	break;
3095	case PGMMODE_EPT:
3096	rc = PGM_SHW_NAME_EPT(Enter)(pVCpu, fIsNewGuestPagingMode64Bits);
3097	break;
3098	case PGMMODE_REAL:
3099	case PGMMODE_PROTECTED:
3100	default:
3101	AssertReleaseMsgFailed(("enmShadowMode=%d\n", enmShadowMode));
3102	return VERR_INTERNAL_ERROR;
3103	}
3104	if (RT_FAILURE(rc))
3105	{
3106	AssertReleaseMsgFailed(("Entering enmShadowMode=%d failed: %Rrc\n", enmShadowMode, rc));
3107	pVCpu->pgm.s.enmShadowMode = PGMMODE_INVALID;
3108	return rc;
3109	}
3110	}
3111
3112	/*
3113	* Always flag the necessary updates
3114	*/
3115	VMCPU_FF_SET(pVCpu, VMCPU_FF_PGM_SYNC_CR3);
3116
3117	/*
3118	* Enter the new guest and shadow+guest modes.
3119	*/
3120	int rc = -1;
3121	int rc2 = -1;
3122	RTGCPHYS GCPhysCR3 = NIL_RTGCPHYS;
3123	pVCpu->pgm.s.enmGuestMode = enmGuestMode;
3124	switch (enmGuestMode)
3125	{
3126	case PGMMODE_REAL:
3127	rc = PGM_GST_NAME_REAL(Enter)(pVCpu, NIL_RTGCPHYS);
3128	switch (pVCpu->pgm.s.enmShadowMode)
3129	{
3130	case PGMMODE_32_BIT:
3131	rc2 = PGM_BTH_NAME_32BIT_REAL(Enter)(pVCpu, NIL_RTGCPHYS);
3132	break;
3133	case PGMMODE_PAE:
3134	case PGMMODE_PAE_NX:
3135	rc2 = PGM_BTH_NAME_PAE_REAL(Enter)(pVCpu, NIL_RTGCPHYS);
3136	break;
3137	case PGMMODE_NESTED:
3138	rc2 = PGM_BTH_NAME_NESTED_REAL(Enter)(pVCpu, NIL_RTGCPHYS);
3139	break;
3140	case PGMMODE_EPT:
3141	rc2 = PGM_BTH_NAME_EPT_REAL(Enter)(pVCpu, NIL_RTGCPHYS);
3142	break;
3143	case PGMMODE_AMD64:
3144	case PGMMODE_AMD64_NX:
3145	AssertMsgFailed(("Should use PAE shadow mode!\n"));
3146	default: AssertFailed(); break;
3147	}
3148	break;
3149
3150	case PGMMODE_PROTECTED:
3151	rc = PGM_GST_NAME_PROT(Enter)(pVCpu, NIL_RTGCPHYS);
3152	switch (pVCpu->pgm.s.enmShadowMode)
3153	{
3154	case PGMMODE_32_BIT:
3155	rc2 = PGM_BTH_NAME_32BIT_PROT(Enter)(pVCpu, NIL_RTGCPHYS);
3156	break;
3157	case PGMMODE_PAE:
3158	case PGMMODE_PAE_NX:
3159	rc2 = PGM_BTH_NAME_PAE_PROT(Enter)(pVCpu, NIL_RTGCPHYS);
3160	break;
3161	case PGMMODE_NESTED:
3162	rc2 = PGM_BTH_NAME_NESTED_PROT(Enter)(pVCpu, NIL_RTGCPHYS);
3163	break;
3164	case PGMMODE_EPT:
3165	rc2 = PGM_BTH_NAME_EPT_PROT(Enter)(pVCpu, NIL_RTGCPHYS);
3166	break;
3167	case PGMMODE_AMD64:
3168	case PGMMODE_AMD64_NX:
3169	AssertMsgFailed(("Should use PAE shadow mode!\n"));
3170	default: AssertFailed(); break;
3171	}
3172	break;
3173
3174	case PGMMODE_32_BIT:
3175	GCPhysCR3 = CPUMGetGuestCR3(pVCpu) & X86_CR3_PAGE_MASK;
3176	rc = PGM_GST_NAME_32BIT(Enter)(pVCpu, GCPhysCR3);
3177	switch (pVCpu->pgm.s.enmShadowMode)
3178	{
3179	case PGMMODE_32_BIT:
3180	rc2 = PGM_BTH_NAME_32BIT_32BIT(Enter)(pVCpu, GCPhysCR3);
3181	break;
3182	case PGMMODE_PAE:
3183	case PGMMODE_PAE_NX:
3184	rc2 = PGM_BTH_NAME_PAE_32BIT(Enter)(pVCpu, GCPhysCR3);
3185	break;
3186	case PGMMODE_NESTED:
3187	rc2 = PGM_BTH_NAME_NESTED_32BIT(Enter)(pVCpu, GCPhysCR3);
3188	break;
3189	case PGMMODE_EPT:
3190	rc2 = PGM_BTH_NAME_EPT_32BIT(Enter)(pVCpu, GCPhysCR3);
3191	break;
3192	case PGMMODE_AMD64:
3193	case PGMMODE_AMD64_NX:
3194	AssertMsgFailed(("Should use PAE shadow mode!\n"));
3195	default: AssertFailed(); break;
3196	}
3197	break;
3198
3199	case PGMMODE_PAE_NX:
3200	case PGMMODE_PAE:
3201	{
3202	uint32_t u32Dummy, u32Features;
3203
3204	CPUMGetGuestCpuId(pVCpu, 1, &u32Dummy, &u32Dummy, &u32Dummy, &u32Features);
3205	if (!(u32Features & X86_CPUID_FEATURE_EDX_PAE))
3206	return VMSetRuntimeError(pVM, VMSETRTERR_FLAGS_FATAL, "PAEmode",
3207	N_("The guest is trying to switch to the PAE mode which is currently disabled by default in VirtualBox. PAE support can be enabled using the VM settings (General/Advanced)"));
3208
3209	GCPhysCR3 = CPUMGetGuestCR3(pVCpu) & X86_CR3_PAE_PAGE_MASK;
3210	rc = PGM_GST_NAME_PAE(Enter)(pVCpu, GCPhysCR3);
3211	switch (pVCpu->pgm.s.enmShadowMode)
3212	{
3213	case PGMMODE_PAE:
3214	case PGMMODE_PAE_NX:
3215	rc2 = PGM_BTH_NAME_PAE_PAE(Enter)(pVCpu, GCPhysCR3);
3216	break;
3217	case PGMMODE_NESTED:
3218	rc2 = PGM_BTH_NAME_NESTED_PAE(Enter)(pVCpu, GCPhysCR3);
3219	break;
3220	case PGMMODE_EPT:
3221	rc2 = PGM_BTH_NAME_EPT_PAE(Enter)(pVCpu, GCPhysCR3);
3222	break;
3223	case PGMMODE_32_BIT:
3224	case PGMMODE_AMD64:
3225	case PGMMODE_AMD64_NX:
3226	AssertMsgFailed(("Should use PAE shadow mode!\n"));
3227	default: AssertFailed(); break;
3228	}
3229	break;
3230	}
3231
3232	#ifdef VBOX_WITH_64_BITS_GUESTS
3233	case PGMMODE_AMD64_NX:
3234	case PGMMODE_AMD64:
3235	GCPhysCR3 = CPUMGetGuestCR3(pVCpu) & UINT64_C(0xfffffffffffff000); /** @todo define this mask! */
3236	rc = PGM_GST_NAME_AMD64(Enter)(pVCpu, GCPhysCR3);
3237	switch (pVCpu->pgm.s.enmShadowMode)
3238	{
3239	case PGMMODE_AMD64:
3240	case PGMMODE_AMD64_NX:
3241	rc2 = PGM_BTH_NAME_AMD64_AMD64(Enter)(pVCpu, GCPhysCR3);
3242	break;
3243	case PGMMODE_NESTED:
3244	rc2 = PGM_BTH_NAME_NESTED_AMD64(Enter)(pVCpu, GCPhysCR3);
3245	break;
3246	case PGMMODE_EPT:
3247	rc2 = PGM_BTH_NAME_EPT_AMD64(Enter)(pVCpu, GCPhysCR3);
3248	break;
3249	case PGMMODE_32_BIT:
3250	case PGMMODE_PAE:
3251	case PGMMODE_PAE_NX:
3252	AssertMsgFailed(("Should use AMD64 shadow mode!\n"));
3253	default: AssertFailed(); break;
3254	}
3255	break;
3256	#endif
3257
3258	default:
3259	AssertReleaseMsgFailed(("enmGuestMode=%d\n", enmGuestMode));
3260	rc = VERR_NOT_IMPLEMENTED;
3261	break;
3262	}
3263
3264	/* status codes. */
3265	AssertRC(rc);
3266	AssertRC(rc2);
3267	if (RT_SUCCESS(rc))
3268	{
3269	rc = rc2;
3270	if (RT_SUCCESS(rc)) /* no informational status codes. */
3271	rc = VINF_SUCCESS;
3272	}
3273
3274	/* Notify HWACCM as well. */
3275	HWACCMR3PagingModeChanged(pVM, pVCpu, pVCpu->pgm.s.enmShadowMode, pVCpu->pgm.s.enmGuestMode);
3276	return rc;
3277	}
3278
3279	/**
3280	* Release the pgm lock if owned by the current VCPU
3281	*
3282	* @param pVM The VM to operate on.
3283	*/
3284	VMMR3DECL(void) PGMR3ReleaseOwnedLocks(PVM pVM)
3285	{
3286	while (PDMCritSectIsOwner(&pVM->pgm.s.CritSect))
3287	PDMCritSectLeave(&pVM->pgm.s.CritSect);
3288	}
3289
3290	/**
3291	* Called by pgmPoolFlushAllInt prior to flushing the pool.
3292	*
3293	* @returns VBox status code, fully asserted.
3294	* @param pVM The VM handle.
3295	* @param pVCpu The VMCPU to operate on.
3296	*/
3297	int pgmR3ExitShadowModeBeforePoolFlush(PVM pVM, PVMCPU pVCpu)
3298	{
3299	/* Unmap the old CR3 value before flushing everything. */
3300	int rc = PGM_BTH_PFN(UnmapCR3, pVCpu)(pVCpu);
3301	AssertRC(rc);
3302
3303	/* Exit the current shadow paging mode as well; nested paging and EPT use a root CR3 which will get flushed here. */
3304	rc = PGM_SHW_PFN(Exit, pVCpu)(pVCpu);
3305	AssertRC(rc);
3306	Assert(pVCpu->pgm.s.pShwPageCR3R3 == NULL);
3307	return rc;
3308	}
3309
3310
3311	/**
3312	* Called by pgmPoolFlushAllInt after flushing the pool.
3313	*
3314	* @returns VBox status code, fully asserted.
3315	* @param pVM The VM handle.
3316	* @param pVCpu The VMCPU to operate on.
3317	*/
3318	int pgmR3ReEnterShadowModeAfterPoolFlush(PVM pVM, PVMCPU pVCpu)
3319	{
3320	pVCpu->pgm.s.enmShadowMode = PGMMODE_INVALID;
3321	int rc = PGMR3ChangeMode(pVM, pVCpu, PGMGetGuestMode(pVCpu));
3322	Assert(VMCPU_FF_ISSET(pVCpu, VMCPU_FF_PGM_SYNC_CR3));
3323	AssertRCReturn(rc, rc);
3324	AssertRCSuccessReturn(rc, VERR_IPE_UNEXPECTED_INFO_STATUS);
3325
3326	Assert(pVCpu->pgm.s.pShwPageCR3R3 != NULL);
3327	AssertMsg( pVCpu->pgm.s.enmShadowMode >= PGMMODE_NESTED
3328	\|\| CPUMGetHyperCR3(pVCpu) == PGMGetHyperCR3(pVCpu),
3329	("%RHp != %RHp %s\n", (RTHCPHYS)CPUMGetHyperCR3(pVCpu), PGMGetHyperCR3(pVCpu), PGMGetModeName(pVCpu->pgm.s.enmShadowMode)));
3330	return rc;
3331	}
3332
3333
3334	/**
3335	* Dumps a PAE shadow page table.
3336	*
3337	* @returns VBox status code (VINF_SUCCESS).
3338	* @param pVM The VM handle.
3339	* @param pPT Pointer to the page table.
3340	* @param u64Address The virtual address of the page table starts.
3341	* @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3342	* @param cMaxDepth The maxium depth.
3343	* @param pHlp Pointer to the output functions.
3344	*/
3345	static int pgmR3DumpHierarchyHCPaePT(PVM pVM, PX86PTPAE pPT, uint64_t u64Address, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3346	{
3347	for (unsigned i = 0; i < RT_ELEMENTS(pPT->a); i++)
3348	{
3349	X86PTEPAE Pte = pPT->a[i];
3350	if (Pte.n.u1Present)
3351	{
3352	pHlp->pfnPrintf(pHlp,
3353	fLongMode /P R S A D G WT CD AT NX 4M a p ? /
3354	? "%016llx 3 \| P %c %c %c %c %c %s %s %s %s 4K %c%c%c %016llx\n"
3355	: "%08llx 2 \| P %c %c %c %c %c %s %s %s %s 4K %c%c%c %016llx\n",
3356	u64Address + ((uint64_t)i << X86_PT_PAE_SHIFT),
3357	Pte.n.u1Write ? 'W' : 'R',
3358	Pte.n.u1User ? 'U' : 'S',
3359	Pte.n.u1Accessed ? 'A' : '-',
3360	Pte.n.u1Dirty ? 'D' : '-',
3361	Pte.n.u1Global ? 'G' : '-',
3362	Pte.n.u1WriteThru ? "WT" : "--",
3363	Pte.n.u1CacheDisable? "CD" : "--",
3364	Pte.n.u1PAT ? "AT" : "--",
3365	Pte.n.u1NoExecute ? "NX" : "--",
3366	Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3367	Pte.u & RT_BIT(10) ? '1' : '0',
3368	Pte.u & PGM_PTFLAGS_CSAM_VALIDATED? 'v' : '-',
3369	Pte.u & X86_PTE_PAE_PG_MASK);
3370	}
3371	}
3372	return VINF_SUCCESS;
3373	}
3374
3375
3376	/**
3377	* Dumps a PAE shadow page directory table.
3378	*
3379	* @returns VBox status code (VINF_SUCCESS).
3380	* @param pVM The VM handle.
3381	* @param HCPhys The physical address of the page directory table.
3382	* @param u64Address The virtual address of the page table starts.
3383	* @param cr4 The CR4, PSE is currently used.
3384	* @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3385	* @param cMaxDepth The maxium depth.
3386	* @param pHlp Pointer to the output functions.
3387	*/
3388	static int pgmR3DumpHierarchyHCPaePD(PVM pVM, RTHCPHYS HCPhys, uint64_t u64Address, uint32_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3389	{
3390	PX86PDPAE pPD = (PX86PDPAE)MMPagePhys2Page(pVM, HCPhys);
3391	if (!pPD)
3392	{
3393	pHlp->pfnPrintf(pHlp, "%0*llx error! Page directory at HCPhys=%RHp was not found in the page pool!\n",
3394	fLongMode ? 16 : 8, u64Address, HCPhys);
3395	return VERR_INVALID_PARAMETER;
3396	}
3397	const bool fBigPagesSupported = fLongMode \|\| !!(cr4 & X86_CR4_PSE);
3398
3399	int rc = VINF_SUCCESS;
3400	for (unsigned i = 0; i < RT_ELEMENTS(pPD->a); i++)
3401	{
3402	X86PDEPAE Pde = pPD->a[i];
3403	if (Pde.n.u1Present)
3404	{
3405	if (fBigPagesSupported && Pde.b.u1Size)
3406	pHlp->pfnPrintf(pHlp,
3407	fLongMode /P R S A D G WT CD AT NX 4M a p ? /
3408	? "%016llx 2 \| P %c %c %c %c %c %s %s %s %s 4M %c%c%c %016llx\n"
3409	: "%08llx 1 \| P %c %c %c %c %c %s %s %s %s 4M %c%c%c %016llx\n",
3410	u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT),
3411	Pde.b.u1Write ? 'W' : 'R',
3412	Pde.b.u1User ? 'U' : 'S',
3413	Pde.b.u1Accessed ? 'A' : '-',
3414	Pde.b.u1Dirty ? 'D' : '-',
3415	Pde.b.u1Global ? 'G' : '-',
3416	Pde.b.u1WriteThru ? "WT" : "--",
3417	Pde.b.u1CacheDisable? "CD" : "--",
3418	Pde.b.u1PAT ? "AT" : "--",
3419	Pde.b.u1NoExecute ? "NX" : "--",
3420	Pde.u & RT_BIT_64(9) ? '1' : '0',
3421	Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3422	Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3423	Pde.u & X86_PDE_PAE_PG_MASK);
3424	else
3425	{
3426	pHlp->pfnPrintf(pHlp,
3427	fLongMode /P R S A D G WT CD AT NX 4M a p ? /
3428	? "%016llx 2 \| P %c %c %c %c %c %s %s .. %s 4K %c%c%c %016llx\n"
3429	: "%08llx 1 \| P %c %c %c %c %c %s %s .. %s 4K %c%c%c %016llx\n",
3430	u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT),
3431	Pde.n.u1Write ? 'W' : 'R',
3432	Pde.n.u1User ? 'U' : 'S',
3433	Pde.n.u1Accessed ? 'A' : '-',
3434	Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3435	Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3436	Pde.n.u1WriteThru ? "WT" : "--",
3437	Pde.n.u1CacheDisable? "CD" : "--",
3438	Pde.n.u1NoExecute ? "NX" : "--",
3439	Pde.u & RT_BIT_64(9) ? '1' : '0',
3440	Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3441	Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3442	Pde.u & X86_PDE_PAE_PG_MASK);
3443	if (cMaxDepth >= 1)
3444	{
3445	/** @todo what about using the page pool for mapping PTs? */
3446	uint64_t u64AddressPT = u64Address + ((uint64_t)i << X86_PD_PAE_SHIFT);
3447	RTHCPHYS HCPhysPT = Pde.u & X86_PDE_PAE_PG_MASK;
3448	PX86PTPAE pPT = NULL;
3449	if (!(Pde.u & PGM_PDFLAGS_MAPPING))
3450	pPT = (PX86PTPAE)MMPagePhys2Page(pVM, HCPhysPT);
3451	else
3452	{
3453	for (PPGMMAPPING pMap = pVM->pgm.s.pMappingsR3; pMap; pMap = pMap->pNextR3)
3454	{
3455	uint64_t off = u64AddressPT - pMap->GCPtr;
3456	if (off < pMap->cb)
3457	{
3458	const int iPDE = (uint32_t)(off >> X86_PD_SHIFT);
3459	const int iSub = (int)((off >> X86_PD_PAE_SHIFT) & 1); /* MSC is a pain sometimes */
3460	if ((iSub ? pMap->aPTs[iPDE].HCPhysPaePT1 : pMap->aPTs[iPDE].HCPhysPaePT0) != HCPhysPT)
3461	pHlp->pfnPrintf(pHlp, "%0*llx error! Mapping error! PT %d has HCPhysPT=%RHp not %RHp is in the PD.\n",
3462	fLongMode ? 16 : 8, u64AddressPT, iPDE,
3463	iSub ? pMap->aPTs[iPDE].HCPhysPaePT1 : pMap->aPTs[iPDE].HCPhysPaePT0, HCPhysPT);
3464	pPT = &pMap->aPTs[iPDE].paPaePTsR3[iSub];
3465	}
3466	}
3467	}
3468	int rc2 = VERR_INVALID_PARAMETER;
3469	if (pPT)
3470	rc2 = pgmR3DumpHierarchyHCPaePT(pVM, pPT, u64AddressPT, fLongMode, cMaxDepth - 1, pHlp);
3471	else
3472	pHlp->pfnPrintf(pHlp, "%0*llx error! Page table at HCPhys=%RHp was not found in the page pool!\n",
3473	fLongMode ? 16 : 8, u64AddressPT, HCPhysPT);
3474	if (rc2 < rc && RT_SUCCESS(rc))
3475	rc = rc2;
3476	}
3477	}
3478	}
3479	}
3480	return rc;
3481	}
3482
3483
3484	/**
3485	* Dumps a PAE shadow page directory pointer table.
3486	*
3487	* @returns VBox status code (VINF_SUCCESS).
3488	* @param pVM The VM handle.
3489	* @param HCPhys The physical address of the page directory pointer table.
3490	* @param u64Address The virtual address of the page table starts.
3491	* @param cr4 The CR4, PSE is currently used.
3492	* @param fLongMode Set if this a long mode table; clear if it's a legacy mode table.
3493	* @param cMaxDepth The maxium depth.
3494	* @param pHlp Pointer to the output functions.
3495	*/
3496	static int pgmR3DumpHierarchyHCPaePDPT(PVM pVM, RTHCPHYS HCPhys, uint64_t u64Address, uint32_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3497	{
3498	PX86PDPT pPDPT = (PX86PDPT)MMPagePhys2Page(pVM, HCPhys);
3499	if (!pPDPT)
3500	{
3501	pHlp->pfnPrintf(pHlp, "%0*llx error! Page directory pointer table at HCPhys=%RHp was not found in the page pool!\n",
3502	fLongMode ? 16 : 8, u64Address, HCPhys);
3503	return VERR_INVALID_PARAMETER;
3504	}
3505
3506	int rc = VINF_SUCCESS;
3507	const unsigned c = fLongMode ? RT_ELEMENTS(pPDPT->a) : X86_PG_PAE_PDPE_ENTRIES;
3508	for (unsigned i = 0; i < c; i++)
3509	{
3510	X86PDPE Pdpe = pPDPT->a[i];
3511	if (Pdpe.n.u1Present)
3512	{
3513	if (fLongMode)
3514	pHlp->pfnPrintf(pHlp, /P R S A D G WT CD AT NX 4M a p ? /
3515	"%016llx 1 \| P %c %c %c %c %c %s %s %s %s .. %c%c%c %016llx\n",
3516	u64Address + ((uint64_t)i << X86_PDPT_SHIFT),
3517	Pdpe.lm.u1Write ? 'W' : 'R',
3518	Pdpe.lm.u1User ? 'U' : 'S',
3519	Pdpe.lm.u1Accessed ? 'A' : '-',
3520	Pdpe.lm.u3Reserved & 1? '?' : '.', /* ignored */
3521	Pdpe.lm.u3Reserved & 4? '!' : '.', /* mbz */
3522	Pdpe.lm.u1WriteThru ? "WT" : "--",
3523	Pdpe.lm.u1CacheDisable? "CD" : "--",
3524	Pdpe.lm.u3Reserved & 2? "!" : "..",/* mbz */
3525	Pdpe.lm.u1NoExecute ? "NX" : "--",
3526	Pdpe.u & RT_BIT(9) ? '1' : '0',
3527	Pdpe.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3528	Pdpe.u & RT_BIT(11) ? '1' : '0',
3529	Pdpe.u & X86_PDPE_PG_MASK);
3530	else
3531	pHlp->pfnPrintf(pHlp, /P G WT CD AT NX 4M a p ? /
3532	"%08x 0 \| P %c %s %s %s %s .. %c%c%c %016llx\n",
3533	i << X86_PDPT_SHIFT,
3534	Pdpe.n.u4Reserved & 1? '!' : '.', /* mbz */
3535	Pdpe.n.u4Reserved & 4? '!' : '.', /* mbz */
3536	Pdpe.n.u1WriteThru ? "WT" : "--",
3537	Pdpe.n.u1CacheDisable? "CD" : "--",
3538	Pdpe.n.u4Reserved & 2? "!" : "..",/* mbz */
3539	Pdpe.u & RT_BIT(9) ? '1' : '0',
3540	Pdpe.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3541	Pdpe.u & RT_BIT(11) ? '1' : '0',
3542	Pdpe.u & X86_PDPE_PG_MASK);
3543	if (cMaxDepth >= 1)
3544	{
3545	int rc2 = pgmR3DumpHierarchyHCPaePD(pVM, Pdpe.u & X86_PDPE_PG_MASK, u64Address + ((uint64_t)i << X86_PDPT_SHIFT),
3546	cr4, fLongMode, cMaxDepth - 1, pHlp);
3547	if (rc2 < rc && RT_SUCCESS(rc))
3548	rc = rc2;
3549	}
3550	}
3551	}
3552	return rc;
3553	}
3554
3555
3556	/**
3557	* Dumps a 32-bit shadow page table.
3558	*
3559	* @returns VBox status code (VINF_SUCCESS).
3560	* @param pVM The VM handle.
3561	* @param HCPhys The physical address of the table.
3562	* @param cr4 The CR4, PSE is currently used.
3563	* @param cMaxDepth The maxium depth.
3564	* @param pHlp Pointer to the output functions.
3565	*/
3566	static int pgmR3DumpHierarchyHcPaePML4(PVM pVM, RTHCPHYS HCPhys, uint32_t cr4, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3567	{
3568	PX86PML4 pPML4 = (PX86PML4)MMPagePhys2Page(pVM, HCPhys);
3569	if (!pPML4)
3570	{
3571	pHlp->pfnPrintf(pHlp, "Page map level 4 at HCPhys=%RHp was not found in the page pool!\n", HCPhys);
3572	return VERR_INVALID_PARAMETER;
3573	}
3574
3575	int rc = VINF_SUCCESS;
3576	for (unsigned i = 0; i < RT_ELEMENTS(pPML4->a); i++)
3577	{
3578	X86PML4E Pml4e = pPML4->a[i];
3579	if (Pml4e.n.u1Present)
3580	{
3581	uint64_t u64Address = ((uint64_t)i << X86_PML4_SHIFT) \| (((uint64_t)i >> (X86_PML4_SHIFT - X86_PDPT_SHIFT - 1)) * 0xffff000000000000ULL);
3582	pHlp->pfnPrintf(pHlp, /P R S A D G WT CD AT NX 4M a p ? /
3583	"%016llx 0 \| P %c %c %c %c %c %s %s %s %s .. %c%c%c %016llx\n",
3584	u64Address,
3585	Pml4e.n.u1Write ? 'W' : 'R',
3586	Pml4e.n.u1User ? 'U' : 'S',
3587	Pml4e.n.u1Accessed ? 'A' : '-',
3588	Pml4e.n.u3Reserved & 1? '?' : '.', /* ignored */
3589	Pml4e.n.u3Reserved & 4? '!' : '.', /* mbz */
3590	Pml4e.n.u1WriteThru ? "WT" : "--",
3591	Pml4e.n.u1CacheDisable? "CD" : "--",
3592	Pml4e.n.u3Reserved & 2? "!" : "..",/* mbz */
3593	Pml4e.n.u1NoExecute ? "NX" : "--",
3594	Pml4e.u & RT_BIT(9) ? '1' : '0',
3595	Pml4e.u & PGM_PLXFLAGS_PERMANENT ? 'p' : '-',
3596	Pml4e.u & RT_BIT(11) ? '1' : '0',
3597	Pml4e.u & X86_PML4E_PG_MASK);
3598
3599	if (cMaxDepth >= 1)
3600	{
3601	int rc2 = pgmR3DumpHierarchyHCPaePDPT(pVM, Pml4e.u & X86_PML4E_PG_MASK, u64Address, cr4, true, cMaxDepth - 1, pHlp);
3602	if (rc2 < rc && RT_SUCCESS(rc))
3603	rc = rc2;
3604	}
3605	}
3606	}
3607	return rc;
3608	}
3609
3610
3611	/**
3612	* Dumps a 32-bit shadow page table.
3613	*
3614	* @returns VBox status code (VINF_SUCCESS).
3615	* @param pVM The VM handle.
3616	* @param pPT Pointer to the page table.
3617	* @param u32Address The virtual address this table starts at.
3618	* @param pHlp Pointer to the output functions.
3619	*/
3620	int pgmR3DumpHierarchyHC32BitPT(PVM pVM, PX86PT pPT, uint32_t u32Address, PCDBGFINFOHLP pHlp)
3621	{
3622	for (unsigned i = 0; i < RT_ELEMENTS(pPT->a); i++)
3623	{
3624	X86PTE Pte = pPT->a[i];
3625	if (Pte.n.u1Present)
3626	{
3627	pHlp->pfnPrintf(pHlp, /P R S A D G WT CD AT NX 4M a m d /
3628	"%08x 1 \| P %c %c %c %c %c %s %s %s .. 4K %c%c%c %08x\n",
3629	u32Address + (i << X86_PT_SHIFT),
3630	Pte.n.u1Write ? 'W' : 'R',
3631	Pte.n.u1User ? 'U' : 'S',
3632	Pte.n.u1Accessed ? 'A' : '-',
3633	Pte.n.u1Dirty ? 'D' : '-',
3634	Pte.n.u1Global ? 'G' : '-',
3635	Pte.n.u1WriteThru ? "WT" : "--",
3636	Pte.n.u1CacheDisable? "CD" : "--",
3637	Pte.n.u1PAT ? "AT" : "--",
3638	Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3639	Pte.u & RT_BIT(10) ? '1' : '0',
3640	Pte.u & PGM_PTFLAGS_CSAM_VALIDATED ? 'v' : '-',
3641	Pte.u & X86_PDE_PG_MASK);
3642	}
3643	}
3644	return VINF_SUCCESS;
3645	}
3646
3647
3648	/**
3649	* Dumps a 32-bit shadow page directory and page tables.
3650	*
3651	* @returns VBox status code (VINF_SUCCESS).
3652	* @param pVM The VM handle.
3653	* @param cr3 The root of the hierarchy.
3654	* @param cr4 The CR4, PSE is currently used.
3655	* @param cMaxDepth How deep into the hierarchy the dumper should go.
3656	* @param pHlp Pointer to the output functions.
3657	*/
3658	int pgmR3DumpHierarchyHC32BitPD(PVM pVM, uint32_t cr3, uint32_t cr4, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3659	{
3660	PX86PD pPD = (PX86PD)MMPagePhys2Page(pVM, cr3 & X86_CR3_PAGE_MASK);
3661	if (!pPD)
3662	{
3663	pHlp->pfnPrintf(pHlp, "Page directory at %#x was not found in the page pool!\n", cr3 & X86_CR3_PAGE_MASK);
3664	return VERR_INVALID_PARAMETER;
3665	}
3666
3667	int rc = VINF_SUCCESS;
3668	for (unsigned i = 0; i < RT_ELEMENTS(pPD->a); i++)
3669	{
3670	X86PDE Pde = pPD->a[i];
3671	if (Pde.n.u1Present)
3672	{
3673	const uint32_t u32Address = i << X86_PD_SHIFT;
3674	if ((cr4 & X86_CR4_PSE) && Pde.b.u1Size)
3675	pHlp->pfnPrintf(pHlp, /P R S A D G WT CD AT NX 4M a m d /
3676	"%08x 0 \| P %c %c %c %c %c %s %s %s .. 4M %c%c%c %08x\n",
3677	u32Address,
3678	Pde.b.u1Write ? 'W' : 'R',
3679	Pde.b.u1User ? 'U' : 'S',
3680	Pde.b.u1Accessed ? 'A' : '-',
3681	Pde.b.u1Dirty ? 'D' : '-',
3682	Pde.b.u1Global ? 'G' : '-',
3683	Pde.b.u1WriteThru ? "WT" : "--",
3684	Pde.b.u1CacheDisable? "CD" : "--",
3685	Pde.b.u1PAT ? "AT" : "--",
3686	Pde.u & RT_BIT_64(9) ? '1' : '0',
3687	Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3688	Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3689	Pde.u & X86_PDE4M_PG_MASK);
3690	else
3691	{
3692	pHlp->pfnPrintf(pHlp, /P R S A D G WT CD AT NX 4M a m d /
3693	"%08x 0 \| P %c %c %c %c %c %s %s .. .. 4K %c%c%c %08x\n",
3694	u32Address,
3695	Pde.n.u1Write ? 'W' : 'R',
3696	Pde.n.u1User ? 'U' : 'S',
3697	Pde.n.u1Accessed ? 'A' : '-',
3698	Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3699	Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3700	Pde.n.u1WriteThru ? "WT" : "--",
3701	Pde.n.u1CacheDisable? "CD" : "--",
3702	Pde.u & RT_BIT_64(9) ? '1' : '0',
3703	Pde.u & PGM_PDFLAGS_MAPPING ? 'm' : '-',
3704	Pde.u & PGM_PDFLAGS_TRACK_DIRTY ? 'd' : '-',
3705	Pde.u & X86_PDE_PG_MASK);
3706	if (cMaxDepth >= 1)
3707	{
3708	/** @todo what about using the page pool for mapping PTs? */
3709	RTHCPHYS HCPhys = Pde.u & X86_PDE_PG_MASK;
3710	PX86PT pPT = NULL;
3711	if (!(Pde.u & PGM_PDFLAGS_MAPPING))
3712	pPT = (PX86PT)MMPagePhys2Page(pVM, HCPhys);
3713	else
3714	{
3715	for (PPGMMAPPING pMap = pVM->pgm.s.pMappingsR3; pMap; pMap = pMap->pNextR3)
3716	if (u32Address - pMap->GCPtr < pMap->cb)
3717	{
3718	int iPDE = (u32Address - pMap->GCPtr) >> X86_PD_SHIFT;
3719	if (pMap->aPTs[iPDE].HCPhysPT != HCPhys)
3720	pHlp->pfnPrintf(pHlp, "%08x error! Mapping error! PT %d has HCPhysPT=%RHp not %RHp is in the PD.\n",
3721	u32Address, iPDE, pMap->aPTs[iPDE].HCPhysPT, HCPhys);
3722	pPT = pMap->aPTs[iPDE].pPTR3;
3723	}
3724	}
3725	int rc2 = VERR_INVALID_PARAMETER;
3726	if (pPT)
3727	rc2 = pgmR3DumpHierarchyHC32BitPT(pVM, pPT, u32Address, pHlp);
3728	else
3729	pHlp->pfnPrintf(pHlp, "%08x error! Page table at %#x was not found in the page pool!\n", u32Address, HCPhys);
3730	if (rc2 < rc && RT_SUCCESS(rc))
3731	rc = rc2;
3732	}
3733	}
3734	}
3735	}
3736
3737	return rc;
3738	}
3739
3740
3741	/**
3742	* Dumps a 32-bit shadow page table.
3743	*
3744	* @returns VBox status code (VINF_SUCCESS).
3745	* @param pVM The VM handle.
3746	* @param pPT Pointer to the page table.
3747	* @param u32Address The virtual address this table starts at.
3748	* @param PhysSearch Address to search for.
3749	*/
3750	int pgmR3DumpHierarchyGC32BitPT(PVM pVM, PX86PT pPT, uint32_t u32Address, RTGCPHYS PhysSearch)
3751	{
3752	for (unsigned i = 0; i < RT_ELEMENTS(pPT->a); i++)
3753	{
3754	X86PTE Pte = pPT->a[i];
3755	if (Pte.n.u1Present)
3756	{
3757	Log(( /P R S A D G WT CD AT NX 4M a m d /
3758	"%08x 1 \| P %c %c %c %c %c %s %s %s .. 4K %c%c%c %08x\n",
3759	u32Address + (i << X86_PT_SHIFT),
3760	Pte.n.u1Write ? 'W' : 'R',
3761	Pte.n.u1User ? 'U' : 'S',
3762	Pte.n.u1Accessed ? 'A' : '-',
3763	Pte.n.u1Dirty ? 'D' : '-',
3764	Pte.n.u1Global ? 'G' : '-',
3765	Pte.n.u1WriteThru ? "WT" : "--",
3766	Pte.n.u1CacheDisable? "CD" : "--",
3767	Pte.n.u1PAT ? "AT" : "--",
3768	Pte.u & PGM_PTFLAGS_TRACK_DIRTY ? 'd' : '-',
3769	Pte.u & RT_BIT(10) ? '1' : '0',
3770	Pte.u & PGM_PTFLAGS_CSAM_VALIDATED ? 'v' : '-',
3771	Pte.u & X86_PDE_PG_MASK));
3772
3773	if ((Pte.u & X86_PDE_PG_MASK) == PhysSearch)
3774	{
3775	uint64_t fPageShw = 0;
3776	RTHCPHYS pPhysHC = 0;
3777
3778	/** @todo SMP support!! */
3779	PGMShwGetPage(&pVM->aCpus[0], (RTGCPTR)(u32Address + (i << X86_PT_SHIFT)), &fPageShw, &pPhysHC);
3780	Log(("Found %RGp at %RGv -> flags=%llx\n", PhysSearch, (RTGCPTR)(u32Address + (i << X86_PT_SHIFT)), fPageShw));
3781	}
3782	}
3783	}
3784	return VINF_SUCCESS;
3785	}
3786
3787
3788	/**
3789	* Dumps a 32-bit guest page directory and page tables.
3790	*
3791	* @returns VBox status code (VINF_SUCCESS).
3792	* @param pVM The VM handle.
3793	* @param cr3 The root of the hierarchy.
3794	* @param cr4 The CR4, PSE is currently used.
3795	* @param PhysSearch Address to search for.
3796	*/
3797	VMMR3DECL(int) PGMR3DumpHierarchyGC(PVM pVM, uint64_t cr3, uint64_t cr4, RTGCPHYS PhysSearch)
3798	{
3799	bool fLongMode = false;
3800	const unsigned cch = fLongMode ? 16 : 8; NOREF(cch);
3801	PX86PD pPD = 0;
3802
3803	int rc = PGM_GCPHYS_2_PTR(pVM, cr3 & X86_CR3_PAGE_MASK, &pPD);
3804	if (RT_FAILURE(rc) \|\| !pPD)
3805	{
3806	Log(("Page directory at %#x was not found in the page pool!\n", cr3 & X86_CR3_PAGE_MASK));
3807	return VERR_INVALID_PARAMETER;
3808	}
3809
3810	Log(("cr3=%08x cr4=%08x%s\n"
3811	"%-*s P - Present\n"
3812	"%-*s \| R/W - Read (0) / Write (1)\n"
3813	"%-*s \| \| U/S - User (1) / Supervisor (0)\n"
3814	"%-*s \| \| \| A - Accessed\n"
3815	"%-*s \| \| \| \| D - Dirty\n"
3816	"%-*s \| \| \| \| \| G - Global\n"
3817	"%-*s \| \| \| \| \| \| WT - Write thru\n"
3818	"%-*s \| \| \| \| \| \| \| CD - Cache disable\n"
3819	"%-*s \| \| \| \| \| \| \| \| AT - Attribute table (PAT)\n"
3820	"%-*s \| \| \| \| \| \| \| \| \| NX - No execute (K8)\n"
3821	"%-*s \| \| \| \| \| \| \| \| \| \| 4K/4M/2M - Page size.\n"
3822	"%-*s \| \| \| \| \| \| \| \| \| \| \| AVL - a=allocated; m=mapping; d=track dirty;\n"
3823	"%-*s \| \| \| \| \| \| \| \| \| \| \| \| p=permanent; v=validated;\n"
3824	"%-*s Level \| \| \| \| \| \| \| \| \| \| \| \| Page\n"
3825	/* xxxx n **** P R S A D G WT CD AT NX 4M AVL xxxxxxxxxxxxx
3826	- W U - - - -- -- -- -- -- 010 */
3827	, cr3, cr4, fLongMode ? " Long Mode" : "",
3828	cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "",
3829	cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "Address"));
3830
3831	for (unsigned i = 0; i < RT_ELEMENTS(pPD->a); i++)
3832	{
3833	X86PDE Pde = pPD->a[i];
3834	if (Pde.n.u1Present)
3835	{
3836	const uint32_t u32Address = i << X86_PD_SHIFT;
3837
3838	if ((cr4 & X86_CR4_PSE) && Pde.b.u1Size)
3839	Log(( /P R S A D G WT CD AT NX 4M a m d /
3840	"%08x 0 \| P %c %c %c %c %c %s %s %s .. 4M %c%c%c %08x\n",
3841	u32Address,
3842	Pde.b.u1Write ? 'W' : 'R',
3843	Pde.b.u1User ? 'U' : 'S',
3844	Pde.b.u1Accessed ? 'A' : '-',
3845	Pde.b.u1Dirty ? 'D' : '-',
3846	Pde.b.u1Global ? 'G' : '-',
3847	Pde.b.u1WriteThru ? "WT" : "--",
3848	Pde.b.u1CacheDisable? "CD" : "--",
3849	Pde.b.u1PAT ? "AT" : "--",
3850	Pde.u & RT_BIT(9) ? '1' : '0',
3851	Pde.u & RT_BIT(10) ? '1' : '0',
3852	Pde.u & RT_BIT(11) ? '1' : '0',
3853	pgmGstGet4MBPhysPage(&pVM->pgm.s, Pde)));
3854	/** @todo PhysSearch */
3855	else
3856	{
3857	Log(( /P R S A D G WT CD AT NX 4M a m d /
3858	"%08x 0 \| P %c %c %c %c %c %s %s .. .. 4K %c%c%c %08x\n",
3859	u32Address,
3860	Pde.n.u1Write ? 'W' : 'R',
3861	Pde.n.u1User ? 'U' : 'S',
3862	Pde.n.u1Accessed ? 'A' : '-',
3863	Pde.n.u1Reserved0 ? '?' : '.', /* ignored */
3864	Pde.n.u1Reserved1 ? '?' : '.', /* ignored */
3865	Pde.n.u1WriteThru ? "WT" : "--",
3866	Pde.n.u1CacheDisable? "CD" : "--",
3867	Pde.u & RT_BIT(9) ? '1' : '0',
3868	Pde.u & RT_BIT(10) ? '1' : '0',
3869	Pde.u & RT_BIT(11) ? '1' : '0',
3870	Pde.u & X86_PDE_PG_MASK));
3871	////if (cMaxDepth >= 1)
3872	{
3873	/** @todo what about using the page pool for mapping PTs? */
3874	RTGCPHYS GCPhys = Pde.u & X86_PDE_PG_MASK;
3875	PX86PT pPT = NULL;
3876
3877	rc = PGM_GCPHYS_2_PTR(pVM, GCPhys, &pPT);
3878
3879	int rc2 = VERR_INVALID_PARAMETER;
3880	if (pPT)
3881	rc2 = pgmR3DumpHierarchyGC32BitPT(pVM, pPT, u32Address, PhysSearch);
3882	else
3883	Log(("%08x error! Page table at %#x was not found in the page pool!\n", u32Address, GCPhys));
3884	if (rc2 < rc && RT_SUCCESS(rc))
3885	rc = rc2;
3886	}
3887	}
3888	}
3889	}
3890
3891	return rc;
3892	}
3893
3894
3895	/**
3896	* Dumps a page table hierarchy use only physical addresses and cr4/lm flags.
3897	*
3898	* @returns VBox status code (VINF_SUCCESS).
3899	* @param pVM The VM handle.
3900	* @param cr3 The root of the hierarchy.
3901	* @param cr4 The cr4, only PAE and PSE is currently used.
3902	* @param fLongMode Set if long mode, false if not long mode.
3903	* @param cMaxDepth Number of levels to dump.
3904	* @param pHlp Pointer to the output functions.
3905	*/
3906	VMMR3DECL(int) PGMR3DumpHierarchyHC(PVM pVM, uint64_t cr3, uint64_t cr4, bool fLongMode, unsigned cMaxDepth, PCDBGFINFOHLP pHlp)
3907	{
3908	if (!pHlp)
3909	pHlp = DBGFR3InfoLogHlp();
3910	if (!cMaxDepth)
3911	return VINF_SUCCESS;
3912	const unsigned cch = fLongMode ? 16 : 8;
3913	pHlp->pfnPrintf(pHlp,
3914	"cr3=%08x cr4=%08x%s\n"
3915	"%-*s P - Present\n"
3916	"%-*s \| R/W - Read (0) / Write (1)\n"
3917	"%-*s \| \| U/S - User (1) / Supervisor (0)\n"
3918	"%-*s \| \| \| A - Accessed\n"
3919	"%-*s \| \| \| \| D - Dirty\n"
3920	"%-*s \| \| \| \| \| G - Global\n"
3921	"%-*s \| \| \| \| \| \| WT - Write thru\n"
3922	"%-*s \| \| \| \| \| \| \| CD - Cache disable\n"
3923	"%-*s \| \| \| \| \| \| \| \| AT - Attribute table (PAT)\n"
3924	"%-*s \| \| \| \| \| \| \| \| \| NX - No execute (K8)\n"
3925	"%-*s \| \| \| \| \| \| \| \| \| \| 4K/4M/2M - Page size.\n"
3926	"%-*s \| \| \| \| \| \| \| \| \| \| \| AVL - a=allocated; m=mapping; d=track dirty;\n"
3927	"%-*s \| \| \| \| \| \| \| \| \| \| \| \| p=permanent; v=validated;\n"
3928	"%-*s Level \| \| \| \| \| \| \| \| \| \| \| \| Page\n"
3929	/* xxxx n **** P R S A D G WT CD AT NX 4M AVL xxxxxxxxxxxxx
3930	- W U - - - -- -- -- -- -- 010 */
3931	, cr3, cr4, fLongMode ? " Long Mode" : "",
3932	cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "",
3933	cch, "", cch, "", cch, "", cch, "", cch, "", cch, "", cch, "Address");
3934	if (cr4 & X86_CR4_PAE)
3935	{
3936	if (fLongMode)
3937	return pgmR3DumpHierarchyHcPaePML4(pVM, cr3 & X86_CR3_PAGE_MASK, cr4, cMaxDepth, pHlp);
3938	return pgmR3DumpHierarchyHCPaePDPT(pVM, cr3 & X86_CR3_PAE_PAGE_MASK, 0, cr4, false, cMaxDepth, pHlp);
3939	}
3940	return pgmR3DumpHierarchyHC32BitPD(pVM, cr3 & X86_CR3_PAGE_MASK, cr4, cMaxDepth, pHlp);
3941	}
3942
3943	#ifdef VBOX_WITH_DEBUGGER
3944
3945	/**
3946	* The '.pgmram' command.
3947	*
3948	* @returns VBox status.
3949	* @param pCmd Pointer to the command descriptor (as registered).
3950	* @param pCmdHlp Pointer to command helper functions.
3951	* @param pVM Pointer to the current VM (if any).
3952	* @param paArgs Pointer to (readonly) array of arguments.
3953	* @param cArgs Number of arguments in the array.
3954	*/
3955	static DECLCALLBACK(int) pgmR3CmdRam(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3956	{
3957	/*
3958	* Validate input.
3959	*/
3960	if (!pVM)
3961	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
3962	if (!pVM->pgm.s.pRamRangesRC)
3963	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Sorry, no Ram is registered.\n");
3964
3965	/*
3966	* Dump the ranges.
3967	*/
3968	int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, "From - To (incl) pvHC\n");
3969	PPGMRAMRANGE pRam;
3970	for (pRam = pVM->pgm.s.pRamRangesR3; pRam; pRam = pRam->pNextR3)
3971	{
3972	rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL,
3973	"%RGp - %RGp %p\n",
3974	pRam->GCPhys, pRam->GCPhysLast, pRam->pvR3);
3975	if (RT_FAILURE(rc))
3976	return rc;
3977	}
3978
3979	return VINF_SUCCESS;
3980	}
3981
3982
3983	/**
3984	* The '.pgmmap' command.
3985	*
3986	* @returns VBox status.
3987	* @param pCmd Pointer to the command descriptor (as registered).
3988	* @param pCmdHlp Pointer to command helper functions.
3989	* @param pVM Pointer to the current VM (if any).
3990	* @param paArgs Pointer to (readonly) array of arguments.
3991	* @param cArgs Number of arguments in the array.
3992	*/
3993	static DECLCALLBACK(int) pgmR3CmdMap(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
3994	{
3995	/*
3996	* Validate input.
3997	*/
3998	if (!pVM)
3999	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4000	if (!pVM->pgm.s.pMappingsR3)
4001	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Sorry, no mappings are registered.\n");
4002
4003	/*
4004	* Print message about the fixedness of the mappings.
4005	*/
4006	int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, pVM->pgm.s.fMappingsFixed ? "The mappings are FIXED.\n" : "The mappings are FLOATING.\n");
4007	if (RT_FAILURE(rc))
4008	return rc;
4009
4010	/*
4011	* Dump the ranges.
4012	*/
4013	PPGMMAPPING pCur;
4014	for (pCur = pVM->pgm.s.pMappingsR3; pCur; pCur = pCur->pNextR3)
4015	{
4016	rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL,
4017	"%08x - %08x %s\n",
4018	pCur->GCPtr, pCur->GCPtrLast, pCur->pszDesc);
4019	if (RT_FAILURE(rc))
4020	return rc;
4021	}
4022
4023	return VINF_SUCCESS;
4024	}
4025
4026
4027	/**
4028	* The '.pgmerror' and '.pgmerroroff' commands.
4029	*
4030	* @returns VBox status.
4031	* @param pCmd Pointer to the command descriptor (as registered).
4032	* @param pCmdHlp Pointer to command helper functions.
4033	* @param pVM Pointer to the current VM (if any).
4034	* @param paArgs Pointer to (readonly) array of arguments.
4035	* @param cArgs Number of arguments in the array.
4036	*/
4037	static DECLCALLBACK(int) pgmR3CmdError(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
4038	{
4039	/*
4040	* Validate input.
4041	*/
4042	if (!pVM)
4043	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4044	AssertReturn(cArgs == 0 \|\| (cArgs == 1 && paArgs[0].enmType == DBGCVAR_TYPE_STRING),
4045	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: Hit bug in the parser.\n"));
4046
4047	if (!cArgs)
4048	{
4049	/*
4050	* Print the list of error injection locations with status.
4051	*/
4052	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "PGM error inject locations:\n");
4053	pCmdHlp->pfnPrintf(pCmdHlp, NULL, " handy - %RTbool\n", pVM->pgm.s.fErrInjHandyPages);
4054	}
4055	else
4056	{
4057
4058	/*
4059	* String switch on where to inject the error.
4060	*/
4061	bool const fNewState = !strcmp(pCmd->pszCmd, "pgmerror");
4062	const char *pszWhere = paArgs[0].u.pszString;
4063	if (!strcmp(pszWhere, "handy"))
4064	ASMAtomicWriteBool(&pVM->pgm.s.fErrInjHandyPages, fNewState);
4065	else
4066	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: Invalid 'where' value: %s.\n", pszWhere);
4067	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "done\n");
4068	}
4069	return VINF_SUCCESS;
4070	}
4071
4072
4073	/**
4074	* The '.pgmsync' command.
4075	*
4076	* @returns VBox status.
4077	* @param pCmd Pointer to the command descriptor (as registered).
4078	* @param pCmdHlp Pointer to command helper functions.
4079	* @param pVM Pointer to the current VM (if any).
4080	* @param paArgs Pointer to (readonly) array of arguments.
4081	* @param cArgs Number of arguments in the array.
4082	*/
4083	static DECLCALLBACK(int) pgmR3CmdSync(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
4084	{
4085	/** @todo SMP support */
4086	PVMCPU pVCpu = &pVM->aCpus[0];
4087
4088	/*
4089	* Validate input.
4090	*/
4091	if (!pVM)
4092	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4093
4094	/*
4095	* Force page directory sync.
4096	*/
4097	VMCPU_FF_SET(pVCpu, VMCPU_FF_PGM_SYNC_CR3);
4098
4099	int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Forcing page directory sync.\n");
4100	if (RT_FAILURE(rc))
4101	return rc;
4102
4103	return VINF_SUCCESS;
4104	}
4105
4106
4107	#ifdef VBOX_STRICT
4108	/**
4109	* The '.pgmassertcr3' command.
4110	*
4111	* @returns VBox status.
4112	* @param pCmd Pointer to the command descriptor (as registered).
4113	* @param pCmdHlp Pointer to command helper functions.
4114	* @param pVM Pointer to the current VM (if any).
4115	* @param paArgs Pointer to (readonly) array of arguments.
4116	* @param cArgs Number of arguments in the array.
4117	*/
4118	static DECLCALLBACK(int) pgmR3CmdAssertCR3(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
4119	{
4120	/** @todo SMP support!! */
4121	PVMCPU pVCpu = &pVM->aCpus[0];
4122
4123	/*
4124	* Validate input.
4125	*/
4126	if (!pVM)
4127	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4128
4129	int rc = pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Checking shadow CR3 page tables for consistency.\n");
4130	if (RT_FAILURE(rc))
4131	return rc;
4132
4133	PGMAssertCR3(pVM, pVCpu, CPUMGetGuestCR3(pVCpu), CPUMGetGuestCR4(pVCpu));
4134
4135	return VINF_SUCCESS;
4136	}
4137	#endif /* VBOX_STRICT */
4138
4139
4140	/**
4141	* The '.pgmsyncalways' command.
4142	*
4143	* @returns VBox status.
4144	* @param pCmd Pointer to the command descriptor (as registered).
4145	* @param pCmdHlp Pointer to command helper functions.
4146	* @param pVM Pointer to the current VM (if any).
4147	* @param paArgs Pointer to (readonly) array of arguments.
4148	* @param cArgs Number of arguments in the array.
4149	*/
4150	static DECLCALLBACK(int) pgmR3CmdSyncAlways(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
4151	{
4152	/** @todo SMP support!! */
4153	PVMCPU pVCpu = &pVM->aCpus[0];
4154
4155	/*
4156	* Validate input.
4157	*/
4158	if (!pVM)
4159	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4160
4161	/*
4162	* Force page directory sync.
4163	*/
4164	if (pVCpu->pgm.s.fSyncFlags & PGM_SYNC_ALWAYS)
4165	{
4166	ASMAtomicAndU32(&pVCpu->pgm.s.fSyncFlags, ~PGM_SYNC_ALWAYS);
4167	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Disabled permanent forced page directory syncing.\n");
4168	}
4169	else
4170	{
4171	ASMAtomicOrU32(&pVCpu->pgm.s.fSyncFlags, PGM_SYNC_ALWAYS);
4172	VMCPU_FF_SET(pVCpu, VMCPU_FF_PGM_SYNC_CR3);
4173	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Enabled permanent forced page directory syncing.\n");
4174	}
4175	}
4176
4177
4178	/**
4179	* The '.pgmsyncalways' command.
4180	*
4181	* @returns VBox status.
4182	* @param pCmd Pointer to the command descriptor (as registered).
4183	* @param pCmdHlp Pointer to command helper functions.
4184	* @param pVM Pointer to the current VM (if any).
4185	* @param paArgs Pointer to (readonly) array of arguments.
4186	* @param cArgs Number of arguments in the array.
4187	*/
4188	static DECLCALLBACK(int) pgmR3CmdPhysToFile(PCDBGCCMD pCmd, PDBGCCMDHLP pCmdHlp, PVM pVM, PCDBGCVAR paArgs, unsigned cArgs, PDBGCVAR pResult)
4189	{
4190	/*
4191	* Validate input.
4192	*/
4193	if (!pVM)
4194	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: The command requires a VM to be selected.\n");
4195	if ( cArgs < 1
4196	\|\| cArgs > 2
4197	\|\| paArgs[0].enmType != DBGCVAR_TYPE_STRING
4198	\|\| ( cArgs > 1
4199	&& paArgs[1].enmType != DBGCVAR_TYPE_STRING))
4200	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: parser error, invalid arguments.\n");
4201	if ( cArgs >= 2
4202	&& strcmp(paArgs[1].u.pszString, "nozero"))
4203	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: Invalid 2nd argument '%s', must be 'nozero'.\n", paArgs[1].u.pszString);
4204	bool fIncZeroPgs = cArgs < 2;
4205
4206	/*
4207	* Open the output file and get the ram parameters.
4208	*/
4209	RTFILE hFile;
4210	int rc = RTFileOpen(&hFile, paArgs[0].u.pszString, RTFILE_O_WRITE \| RTFILE_O_CREATE_REPLACE \| RTFILE_O_DENY_WRITE);
4211	if (RT_FAILURE(rc))
4212	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: RTFileOpen(,'%s',) -> %Rrc.\n", paArgs[0].u.pszString, rc);
4213
4214	uint32_t cbRamHole = 0;
4215	CFGMR3QueryU32Def(CFGMR3GetRoot(pVM), "RamHoleSize", &cbRamHole, MM_RAM_HOLE_SIZE_DEFAULT);
4216	uint64_t cbRam = 0;
4217	CFGMR3QueryU64Def(CFGMR3GetRoot(pVM), "RamSize", &cbRam, 0);
4218	RTGCPHYS GCPhysEnd = cbRam + cbRamHole;
4219
4220	/*
4221	* Dump the physical memory, page by page.
4222	*/
4223	RTGCPHYS GCPhys = 0;
4224	char abZeroPg[PAGE_SIZE];
4225	RT_ZERO(abZeroPg);
4226
4227	pgmLock(pVM);
4228	for (PPGMRAMRANGE pRam = pVM->pgm.s.pRamRangesR3;
4229	pRam && pRam->GCPhys < GCPhysEnd && RT_SUCCESS(rc);
4230	pRam = pRam->pNextR3)
4231	{
4232	/* fill the gap */
4233	if (pRam->GCPhys > GCPhys && fIncZeroPgs)
4234	{
4235	while (pRam->GCPhys > GCPhys && RT_SUCCESS(rc))
4236	{
4237	rc = RTFileWrite(hFile, abZeroPg, PAGE_SIZE, NULL);
4238	GCPhys += PAGE_SIZE;
4239	}
4240	}
4241
4242	PCPGMPAGE pPage = &pRam->aPages[0];
4243	while (GCPhys < pRam->GCPhysLast && RT_SUCCESS(rc))
4244	{
4245	if (PGM_PAGE_IS_ZERO(pPage))
4246	{
4247	if (fIncZeroPgs)
4248	{
4249	rc = RTFileWrite(hFile, abZeroPg, PAGE_SIZE, NULL);
4250	if (RT_FAILURE(rc))
4251	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: RTFileWrite -> %Rrc at GCPhys=%RGp.\n", rc, GCPhys);
4252	}
4253	}
4254	else
4255	{
4256	switch (PGM_PAGE_GET_TYPE(pPage))
4257	{
4258	case PGMPAGETYPE_RAM:
4259	case PGMPAGETYPE_ROM_SHADOW: /* trouble?? */
4260	case PGMPAGETYPE_ROM:
4261	case PGMPAGETYPE_MMIO2:
4262	{
4263	void const *pvPage;
4264	PGMPAGEMAPLOCK Lock;
4265	rc = PGMPhysGCPhys2CCPtrReadOnly(pVM, GCPhys, &pvPage, &Lock);
4266	if (RT_SUCCESS(rc))
4267	{
4268	rc = RTFileWrite(hFile, pvPage, PAGE_SIZE, NULL);
4269	PGMPhysReleasePageMappingLock(pVM, &Lock);
4270	if (RT_FAILURE(rc))
4271	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: RTFileWrite -> %Rrc at GCPhys=%RGp.\n", rc, GCPhys);
4272	}
4273	else
4274	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: PGMPhysGCPhys2CCPtrReadOnly -> %Rrc at GCPhys=%RGp.\n", rc, GCPhys);
4275	break;
4276	}
4277
4278	default:
4279	AssertFailed();
4280	case PGMPAGETYPE_MMIO2_ALIAS_MMIO:
4281	case PGMPAGETYPE_MMIO:
4282	if (fIncZeroPgs)
4283	{
4284	rc = RTFileWrite(hFile, abZeroPg, PAGE_SIZE, NULL);
4285	if (RT_FAILURE(rc))
4286	pCmdHlp->pfnPrintf(pCmdHlp, NULL, "error: RTFileWrite -> %Rrc at GCPhys=%RGp.\n", rc, GCPhys);
4287	}
4288	break;
4289	}
4290	}
4291
4292
4293	/* advance */
4294	GCPhys += PAGE_SIZE;
4295	pPage++;
4296	}
4297	}
4298	pgmUnlock(pVM);
4299
4300	RTFileClose(hFile);
4301	if (RT_SUCCESS(rc))
4302	return pCmdHlp->pfnPrintf(pCmdHlp, NULL, "Successfully saved physical memory to '%s'.\n", paArgs[0].u.pszString);
4303	return VINF_SUCCESS;
4304	}
4305
4306	#endif /* VBOX_WITH_DEBUGGER */
4307
4308	/**
4309	* pvUser argument of the pgmR3CheckIntegrity*Node callbacks.
4310	*/
4311	typedef struct PGMCHECKINTARGS
4312	{
4313	bool fLeftToRight; /*< true: left-to-right; false: right-to-left. /
4314	PPGMPHYSHANDLER pPrevPhys;
4315	PPGMVIRTHANDLER pPrevVirt;
4316	PPGMPHYS2VIRTHANDLER pPrevPhys2Virt;
4317	PVM pVM;
4318	} PGMCHECKINTARGS, *PPGMCHECKINTARGS;
4319
4320	/**
4321	* Validate a node in the physical handler tree.
4322	*
4323	* @returns 0 on if ok, other wise 1.
4324	* @param pNode The handler node.
4325	* @param pvUser pVM.
4326	*/
4327	static DECLCALLBACK(int) pgmR3CheckIntegrityPhysHandlerNode(PAVLROGCPHYSNODECORE pNode, void *pvUser)
4328	{
4329	PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4330	PPGMPHYSHANDLER pCur = (PPGMPHYSHANDLER)pNode;
4331	AssertReleaseReturn(!((uintptr_t)pCur & 7), 1);
4332	AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %RGp-%RGp %s\n", pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4333	AssertReleaseMsg( !pArgs->pPrevPhys
4334	\|\| (pArgs->fLeftToRight ? pArgs->pPrevPhys->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys->Core.KeyLast > pCur->Core.Key),
4335	("pPrevPhys=%p %RGp-%RGp %s\n"
4336	" pCur=%p %RGp-%RGp %s\n",
4337	pArgs->pPrevPhys, pArgs->pPrevPhys->Core.Key, pArgs->pPrevPhys->Core.KeyLast, pArgs->pPrevPhys->pszDesc,
4338	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4339	pArgs->pPrevPhys = pCur;
4340	return 0;
4341	}
4342
4343
4344	/**
4345	* Validate a node in the virtual handler tree.
4346	*
4347	* @returns 0 on if ok, other wise 1.
4348	* @param pNode The handler node.
4349	* @param pvUser pVM.
4350	*/
4351	static DECLCALLBACK(int) pgmR3CheckIntegrityVirtHandlerNode(PAVLROGCPTRNODECORE pNode, void *pvUser)
4352	{
4353	PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4354	PPGMVIRTHANDLER pCur = (PPGMVIRTHANDLER)pNode;
4355	AssertReleaseReturn(!((uintptr_t)pCur & 7), 1);
4356	AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %RGv-%RGv %s\n", pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4357	AssertReleaseMsg( !pArgs->pPrevVirt
4358	\|\| (pArgs->fLeftToRight ? pArgs->pPrevVirt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevVirt->Core.KeyLast > pCur->Core.Key),
4359	("pPrevVirt=%p %RGv-%RGv %s\n"
4360	" pCur=%p %RGv-%RGv %s\n",
4361	pArgs->pPrevVirt, pArgs->pPrevVirt->Core.Key, pArgs->pPrevVirt->Core.KeyLast, pArgs->pPrevVirt->pszDesc,
4362	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc));
4363	for (unsigned iPage = 0; iPage < pCur->cPages; iPage++)
4364	{
4365	AssertReleaseMsg(pCur->aPhysToVirt[iPage].offVirtHandler == -RT_OFFSETOF(PGMVIRTHANDLER, aPhysToVirt[iPage]),
4366	("pCur=%p %RGv-%RGv %s\n"
4367	"iPage=%d offVirtHandle=%#x expected %#x\n",
4368	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->pszDesc,
4369	iPage, pCur->aPhysToVirt[iPage].offVirtHandler, -RT_OFFSETOF(PGMVIRTHANDLER, aPhysToVirt[iPage])));
4370	}
4371	pArgs->pPrevVirt = pCur;
4372	return 0;
4373	}
4374
4375
4376	/**
4377	* Validate a node in the virtual handler tree.
4378	*
4379	* @returns 0 on if ok, other wise 1.
4380	* @param pNode The handler node.
4381	* @param pvUser pVM.
4382	*/
4383	static DECLCALLBACK(int) pgmR3CheckIntegrityPhysToVirtHandlerNode(PAVLROGCPHYSNODECORE pNode, void *pvUser)
4384	{
4385	PPGMCHECKINTARGS pArgs = (PPGMCHECKINTARGS)pvUser;
4386	PPGMPHYS2VIRTHANDLER pCur = (PPGMPHYS2VIRTHANDLER)pNode;
4387	AssertReleaseMsgReturn(!((uintptr_t)pCur & 3), ("\n"), 1);
4388	AssertReleaseMsgReturn(!(pCur->offVirtHandler & 3), ("\n"), 1);
4389	AssertReleaseMsg(pCur->Core.Key <= pCur->Core.KeyLast,("pCur=%p %RGp-%RGp\n", pCur, pCur->Core.Key, pCur->Core.KeyLast));
4390	AssertReleaseMsg( !pArgs->pPrevPhys2Virt
4391	\|\| (pArgs->fLeftToRight ? pArgs->pPrevPhys2Virt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys2Virt->Core.KeyLast > pCur->Core.Key),
4392	("pPrevPhys2Virt=%p %RGp-%RGp\n"
4393	" pCur=%p %RGp-%RGp\n",
4394	pArgs->pPrevPhys2Virt, pArgs->pPrevPhys2Virt->Core.Key, pArgs->pPrevPhys2Virt->Core.KeyLast,
4395	pCur, pCur->Core.Key, pCur->Core.KeyLast));
4396	AssertReleaseMsg( !pArgs->pPrevPhys2Virt
4397	\|\| (pArgs->fLeftToRight ? pArgs->pPrevPhys2Virt->Core.KeyLast < pCur->Core.Key : pArgs->pPrevPhys2Virt->Core.KeyLast > pCur->Core.Key),
4398	("pPrevPhys2Virt=%p %RGp-%RGp\n"
4399	" pCur=%p %RGp-%RGp\n",
4400	pArgs->pPrevPhys2Virt, pArgs->pPrevPhys2Virt->Core.Key, pArgs->pPrevPhys2Virt->Core.KeyLast,
4401	pCur, pCur->Core.Key, pCur->Core.KeyLast));
4402	AssertReleaseMsg((pCur->offNextAlias & (PGMPHYS2VIRTHANDLER_IN_TREE \| PGMPHYS2VIRTHANDLER_IS_HEAD)) == (PGMPHYS2VIRTHANDLER_IN_TREE \| PGMPHYS2VIRTHANDLER_IS_HEAD),
4403	("pCur=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4404	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias));
4405	if (pCur->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK)
4406	{
4407	PPGMPHYS2VIRTHANDLER pCur2 = pCur;
4408	for (;;)
4409	{
4410	pCur2 = (PPGMPHYS2VIRTHANDLER)((intptr_t)pCur + (pCur->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK));
4411	AssertReleaseMsg(pCur2 != pCur,
4412	(" pCur=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4413	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias));
4414	AssertReleaseMsg((pCur2->offNextAlias & (PGMPHYS2VIRTHANDLER_IN_TREE \| PGMPHYS2VIRTHANDLER_IS_HEAD)) == PGMPHYS2VIRTHANDLER_IN_TREE,
4415	(" pCur=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4416	"pCur2=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4417	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4418	pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4419	AssertReleaseMsg((pCur2->Core.Key ^ pCur->Core.Key) < PAGE_SIZE,
4420	(" pCur=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4421	"pCur2=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4422	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4423	pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4424	AssertReleaseMsg((pCur2->Core.KeyLast ^ pCur->Core.KeyLast) < PAGE_SIZE,
4425	(" pCur=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n"
4426	"pCur2=%p:{.Core.Key=%RGp, .Core.KeyLast=%RGp, .offVirtHandler=%#RX32, .offNextAlias=%#RX32}\n",
4427	pCur, pCur->Core.Key, pCur->Core.KeyLast, pCur->offVirtHandler, pCur->offNextAlias,
4428	pCur2, pCur2->Core.Key, pCur2->Core.KeyLast, pCur2->offVirtHandler, pCur2->offNextAlias));
4429	if (!(pCur2->offNextAlias & PGMPHYS2VIRTHANDLER_OFF_MASK))
4430	break;
4431	}
4432	}
4433
4434	pArgs->pPrevPhys2Virt = pCur;
4435	return 0;
4436	}
4437
4438
4439	/**
4440	* Perform an integrity check on the PGM component.
4441	*
4442	* @returns VINF_SUCCESS if everything is fine.
4443	* @returns VBox error status after asserting on integrity breach.
4444	* @param pVM The VM handle.
4445	*/
4446	VMMR3DECL(int) PGMR3CheckIntegrity(PVM pVM)
4447	{
4448	AssertReleaseReturn(pVM->pgm.s.offVM, VERR_INTERNAL_ERROR);
4449
4450	/*
4451	* Check the trees.
4452	*/
4453	int cErrors = 0;
4454	const static PGMCHECKINTARGS s_LeftToRight = { true, NULL, NULL, NULL, pVM };
4455	const static PGMCHECKINTARGS s_RightToLeft = { false, NULL, NULL, NULL, pVM };
4456	PGMCHECKINTARGS Args = s_LeftToRight;
4457	cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesR3->PhysHandlers, true, pgmR3CheckIntegrityPhysHandlerNode, &Args);
4458	Args = s_RightToLeft;
4459	cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesR3->PhysHandlers, false, pgmR3CheckIntegrityPhysHandlerNode, &Args);
4460	Args = s_LeftToRight;
4461	cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesR3->VirtHandlers, true, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4462	Args = s_RightToLeft;
4463	cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesR3->VirtHandlers, false, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4464	Args = s_LeftToRight;
4465	cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesR3->HyperVirtHandlers, true, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4466	Args = s_RightToLeft;
4467	cErrors += RTAvlroGCPtrDoWithAll( &pVM->pgm.s.pTreesR3->HyperVirtHandlers, false, pgmR3CheckIntegrityVirtHandlerNode, &Args);
4468	Args = s_LeftToRight;
4469	cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesR3->PhysToVirtHandlers, true, pgmR3CheckIntegrityPhysToVirtHandlerNode, &Args);
4470	Args = s_RightToLeft;
4471	cErrors += RTAvlroGCPhysDoWithAll(&pVM->pgm.s.pTreesR3->PhysToVirtHandlers, false, pgmR3CheckIntegrityPhysToVirtHandlerNode, &Args);
4472
4473	return !cErrors ? VINF_SUCCESS : VERR_INTERNAL_ERROR;
4474	}
4475
4476

Note: See TracBrowser for help on using the repository browser.

source: vbox/trunk/src/VBox/VMM/PGM.cpp@ 24760

Download in other formats: