|
|
@16776
|
16 years |
vboxsync |
operator precedence (gcc found this bug)
|
|
|
@16679
|
16 years |
vboxsync |
VBOX_WITH_PGMPOOL_PAGING_ONLY: PGMGetHyperCR3 fix for EPT
|
|
|
@16645
|
16 years |
vboxsync |
Adjusted assertion
|
|
|
@16575
|
16 years |
vboxsync |
VBOX_WITH_PGMPOOL_PAGING_ONLY: update
|
|
|
@16121
|
16 years |
vboxsync |
Another paranoid assertion
|
|
|
@16119
|
16 years |
vboxsync |
Clear CPUM_SYNC_FPU_STATE in CPUMR0SaveGuestFPU.
|
|
|
@16113
|
16 years |
vboxsync |
Save the host FPU/XMM state for the 32/64 case.
|
|
|
@16111
|
16 years |
vboxsync |
cpumR0SaveHostRestoreGuestFPUState sets CPUM_USED_FPU, unnecessary to …
|
|
|
@16110
|
16 years |
vboxsync |
HWCMXR0: debugger (dbgf) breakpoints; DEBUG build only for the time …
|
|
|
@16109
|
16 years |
vboxsync |
CPUMR0: Don't forget to set CPUM_USED_FPU_SINCE_REM.
|
|
|
@16108
|
16 years |
vboxsync |
CPUM: Fixed 64-bit FPU management on the mac.
|
|
|
@16106
|
16 years |
vboxsync |
CPUMR0: forgot to remove a line.
|
|
|
@16105
|
16 years |
vboxsync |
DBGFR0Trap03handler: use rip and not eip, pc is accurate and not …
|
|
|
@16103
|
16 years |
vboxsync |
CPUMR0.cpp: minus RTLogPrintf.
|
|
|
@16102
|
16 years |
vboxsync |
CPUMR0.cpp: VBOX_ENABLE_64_BITS_GUESTS is a HWACCM local macro, use …
|
|
|
@16084
|
16 years |
vboxsync |
Flush the recompiler code cache when switch from real to protected mode. */
|
|
|
@16077
|
16 years |
vboxsync |
Minor logging change
|
|
|
@15989
|
16 years |
vboxsync |
Fixed clearing properly
|
|
|
@15988
|
16 years |
vboxsync |
Restore previous EFER state while checking for SVM presence.
|
|
|
@15981
|
16 years |
vboxsync |
Hidden selector limit must be set to 0xffff in real mode.
|
|
|
@15971
|
16 years |
vboxsync |
VT-x: The limit must correspond to the granularity bit
|
|
|
@15962
|
16 years |
vboxsync |
Added debugging code (inactive) for tracking the state of the 32/64 …
|
|
|
@15931
|
16 years |
vboxsync |
gcc found one of these bugs
|
|
|
@15866
|
16 years |
vboxsync |
Backed out previous experiments.
|
|
|
@15865
|
16 years |
vboxsync |
GVMMR0.cpp: Increased the max VM handles from 127+1 to 1023+1 for …
|
|
|
@15858
|
16 years |
vboxsync |
More statistics
|
|
|
@15857
|
16 years |
vboxsync |
More logging
|
|
|
@15854
|
16 years |
vboxsync |
Compile fixes
|
|
|
@15852
|
16 years |
vboxsync |
Another experiment with limiting the time spent in ring 0.
|
|
|
@15797
|
16 years |
vboxsync |
Redid r41365.
|
|
|
@15789
|
16 years |
vboxsync |
Protect code that relies on not being rescheduled to another cpu. …
|
|
|
@15749
|
16 years |
vboxsync |
Log additional VT-x flags
|
|
|
@15746
|
16 years |
vboxsync |
Intercept rdpmc instructions as well.
|
|
|
@15741
|
16 years |
vboxsync |
Measure time spent in the 32/64 switcher for AMD-V too
|
|
|
@15738
|
16 years |
vboxsync |
Bit 0 is valid
|
|
|
@15736
|
16 years |
vboxsync |
Corrected max time spent in VT-x guest mode
|
|
|
@15735
|
16 years |
vboxsync |
Extra check
|
|
|
@15734
|
16 years |
vboxsync |
Allow up to 1M clock ticks worth of guest execution.
|
|
|
@15733
|
16 years |
vboxsync |
Experiment with the VT-x preemption timer for the 32/64 switcher.
|
|
|
@15729
|
16 years |
vboxsync |
Assertion update
|
|
|
@15728
|
16 years |
vboxsync |
Assertion only valid for windows hosts.
|
|
|
@15724
|
16 years |
vboxsync |
Measure time spent in the 32/64 world switcher.
|
|
|
@15701
|
16 years |
vboxsync |
fixed build
|
|
|
@15696
|
16 years |
vboxsync |
Activate switcher test
|
|
|
@15695
|
16 years |
vboxsync |
Compile fix
|
|
|
@15694
|
16 years |
vboxsync |
Profiling of the 32/64 switcher.
|
|
|
@15679
|
16 years |
vboxsync |
Another paranoid assertion
|
|
|
@15666
|
16 years |
vboxsync |
Correct the guest cr4 value when the host isn't in PAE mode (and the …
|
|
|
@15656
|
16 years |
vboxsync |
Removed obsolete assertion
|
|
|
@15654
|
16 years |
vboxsync |
Turn off interrupts in the 32/64 switcher.
|
|
|
@15605
|
16 years |
vboxsync |
#3281: Making #undef HWACCM_VMX_EMULATE_REALMODE work again, looks …
|
|
|
@15587
|
16 years |
vboxsync |
#3281: Two more places to check for pRealModeTSS.
|
|
|
@15576
|
16 years |
vboxsync |
More paranoid checks
|
|
|
@15555
|
16 years |
vboxsync |
Clear PAE bit when in 32 bits shadow mode
|
|
|
@15545
|
16 years |
vboxsync |
More paranoia.
|
|
|
@15507
|
16 years |
vboxsync |
VT-x: keep track of missed paging mode changes
|
|
|
@15502
|
16 years |
vboxsync |
Cleaned up
|
|
|
@15490
|
16 years |
vboxsync |
Disabled VMX_VMCS_GUEST_EFER_FULL code. (crashes my Nehalem host)
|
|
|
@15440
|
16 years |
vboxsync |
Sync CR2 properly for VT-x EPT (unable to access the 64 bits cr2 from …
|
|
|
@15439
|
16 years |
vboxsync |
Enable 64 bits guest support on 32 bits hosts. Only use rem64 if the …
|
|
|
@15436
|
16 years |
vboxsync |
PGMR0DynMap: Corrected PGMR0DYNMAP_PAGES_PER_CPU_MIN (>= max set …
|
|
|
@15434
|
16 years |
vboxsync |
PGMR0DynMap: Tuning and logging.
|
|
|
@15429
|
16 years |
vboxsync |
PGM: Don't use PGMDynMapHCPage in R0/darwin.
|
|
|
@15427
|
16 years |
vboxsync |
PGMR0DynMap: PGMR0DynMap never returns anything but VINF_SUCCESS (no …
|
|
|
@15423
|
16 years |
vboxsync |
PGMR0DynMap: bail out on failure, don't crash the host.
|
|
|
@15417
|
16 years |
vboxsync |
CPUM: save and restore 64-bit debug registers on darwin (#3202).
|
|
|
@15416
|
16 years |
vboxsync |
CPUM: hybrid 32-bit kernel FPU mess.
|
|
|
@15415
|
16 years |
vboxsync |
HWACCMR0.asm: strimmed down the three other Run wrappers.
|
|
|
@15414
|
16 years |
vboxsync |
Corrected VBOX_WITH_HYBIRD_32BIT_KERNEL to …
|
|
|
@15411
|
16 years |
vboxsync |
VMM: Working around set overflows caused by the page pool.
|
|
|
@15410
|
16 years |
vboxsync |
PGM: Deal with pgmPoolAlloc failure in MapCR3 without relying on …
|
|
|
@15409
|
16 years |
vboxsync |
Made exitting on the wrong cpu fatal in strict builds but not when …
|
|
|
@15404
|
16 years |
vboxsync |
#3202: 64-bit guest support on the mac.
|
|
|
@15402
|
16 years |
vboxsync |
PGMDynMapMigrateAutoSet: Fixed incorrect updating of PGMMAPSET::iCpu.
|
|
|
@15395
|
16 years |
vboxsync |
spaces
|
|
|
@15394
|
16 years |
vboxsync |
Workaround for guest EFER load regression, see defect comment 36.
|
|
|
@15389
|
16 years |
vboxsync |
Update the guest EFER VMCS field too.
|
|
|
@15386
|
16 years |
vboxsync |
Missing cases
|
|
|
@15385
|
16 years |
vboxsync |
Cache only those writes which require special attention in 32 bits …
|
|
|
@15377
|
16 years |
vboxsync |
Program entry and exit controls on each guest state sync.
|
|
|
@15371
|
16 years |
vboxsync |
More paranoia
|
|
|
@15368
|
16 years |
vboxsync |
Paranoia
|
|
|
@15364
|
16 years |
vboxsync |
Cleanup + extra check
|
|
|
@15363
|
16 years |
vboxsync |
Type corrections for logging.
|
|
|
@15362
|
16 years |
vboxsync |
Don't use RTCCUINTREG.
|
|
|
@15360
|
16 years |
vboxsync |
Comment
|
|
|
@15359
|
16 years |
vboxsync |
Makes no difference.
|
|
|
@15358
|
16 years |
vboxsync |
Resync the whole guest state after each world switch.
|
|
|
@15347
|
16 years |
vboxsync |
PGMR0DynMap: Fixed testcase (missing migration).
|
|
|
@15346
|
16 years |
vboxsync |
Avoid a few calls to RTMpCpuId() and RTMCpuIdToSetIndex().
|
|
|
@15344
|
16 years |
vboxsync |
#3202: Optimizations of the dynamic page mapping code (ring-0). Do …
|
|
|
@15307
|
16 years |
vboxsync |
VT-x switcher update
|
|
|
@15296
|
16 years |
vboxsync |
Update the 64 bits variant too
|
|
|
@15295
|
16 years |
vboxsync |
Update
|
|
|
@15294
|
16 years |
vboxsync |
Write order is important.
|
|
|
@15292
|
16 years |
vboxsync |
better assertion
|
|
|
@15291
|
16 years |
vboxsync |
Flush pending writes when leaving ring 0.
|
|
|
@15289
|
16 years |
vboxsync |
Flush the queued writes first before checking the values. (debug only)
|
|
|
@15288
|
16 years |
vboxsync |
Exclusive
|
|
|
@15276
|
16 years |
vboxsync |
Fix
|
|
|