DrChat
a1677d34e1
[CPU] MMIOHandler IsRangeWatched now returns true if the entire range is watched.
2018-02-09 19:23:24 -06:00
DrChat
7da973d9e6
[CPU] Handle cases in access watches where base addresses overlap
2018-02-09 17:21:35 -06:00
DrChat
aaf281351d
Format all code with new clang-format
2017-12-14 20:44:43 -06:00
Dr. Chat
c4b728b121
MMIO Handler: Don't allow more than one watch to occupy the same region (fire old watches if a new one overlaps)
2017-03-11 20:55:01 -06:00
Dr. Chat
0e3c113375
Physical write watches -> access watches (read and/or write watching)
2016-03-17 21:55:16 -05:00
Dr. Chat
529227e4e6
MMIOHandler::InvalidateRange
2015-11-08 18:18:28 -06:00
Dr. Chat
430d4c7182
Fix rare deadlock in the MMIO handler.
2015-11-05 18:13:15 -06:00
Dr. Chat
55e811450d
Lint
2015-10-22 20:22:34 -05:00
Dr. Chat
407d79cf3e
Fix MMIO handler race condition by rechecking page access protections under the global lock.
2015-10-22 20:18:08 -05:00
Ben Vanik
5d033f9cb3
A new debugger.
...
Lots of bugs/rough edges/etc - issues will be filed.
Old-style debugging still works (just use --emit_source_annotations to get
the helpful movs back and --break_on_instruction will still fire).
2015-09-20 21:31:05 -07:00
Dr. Chat
a2bc0443f2
Quick-kill the MMIO handler if the fault address is above the mapping range.
2015-09-07 21:26:33 -05:00
Ben Vanik
3c96b6fa0a
DANGER DANGER. Switching to global critical region.
...
This changes almost all locks held by guest threads to use a single global
critical region. This emulates the behavior on the PPC of disabling
interrupts (by calls like KeRaiseIrqlToDpcLevel or masking interrupts),
and prevents deadlocks from occuring when threads are suspended or
otherwise blocked.
This has performance implications and a pass is needed to ensure the
locking is as granular as possible. It could also break everything
because it's fundamentally unsound. We'll see.
2015-09-06 09:30:54 -07:00
Ben Vanik
5e08889d93
More style cleanup.
2015-08-06 20:17:01 -07:00
Ben Vanik
ec326119cf
Replacing beaengine in mmio handler with custom decoder.
...
Fixes #358 .
2015-08-04 08:25:42 -07:00
Ben Vanik
91d7acf59e
Fixing some particularly bad /W4.
2015-07-18 23:04:21 -07:00
Ben Vanik
58c3a1ba79
VirtualProtect platform abstraction.
2015-07-15 18:20:05 -07:00
Ben Vanik
40621a90bd
Switching to standard mutex for now (but keeping xe::mutex).
2015-07-13 21:49:32 -07:00
Ben Vanik
d335555277
Manually page align address range for write watch.
2015-06-23 17:37:09 -07:00
Ben Vanik
40a6a12800
Attempting to inline a lot of MMIO operations. Works for GPU stuff mainly.
2015-06-02 20:15:43 -07:00
Ben Vanik
8bfc43bb0d
Trying to do MMIO protects on allocation granularity.
2015-05-18 20:47:41 -07:00
Ben Vanik
147a70b9c1
Rewriting memory manager.
2015-05-17 00:16:56 +01:00
Ben Vanik
0c646f4bc2
Prevent some stack overflows in MMIO handler.
2015-05-09 00:53:28 -07:00
Ben Vanik
e3220f7ae6
Code cleanup: moving poly/ into xenia/base/
2015-05-02 03:42:51 -07:00
Ben Vanik
99816056be
Code cleanup: removing poly.h
2015-05-02 02:29:26 -07:00
Ben Vanik
f9a3f0c48e
Fix broken VirtualProtect reset on ClearWriteWatch.
2015-02-02 23:07:17 -08:00
Ben Vanik
00e4a4fe1b
Fix #include format.
2015-01-31 22:49:47 -08:00
Ben Vanik
0529fdb84d
Basic hacky write watching for texture invalidation. Doesn't scale.
2015-01-04 16:28:42 -08:00
Ben Vanik
4ad7662ef7
Fixing MMIO build on Windows.
2014-08-01 20:29:57 -07:00
Ben Vanik
0129a96225
Cross-platformizing MMIO stuff.
...
MSVC build likely needs some fixes.
2014-07-29 22:12:39 -07:00
