ReactOS Portable Systems Group 2e1b82cf72 NMI Support Patch 1: ... [HAL]: The I/O Permissions Map on a typical Privileged Mode x86 OS is all 0xFF's, so it's quite wasteful to copy-in/out the entire map each single BIOS Call. As an optimization, only save and restore non-0xFF entries. [NTOS]: Define and use constants for different IOPM values, instead of using hardcoded values. svn path=/trunk/; revision=44841		2009-12-31 23:51:26 +00:00
..
arm	- Add KCALLOUT_FRAME to define the kernel callback stack layout during user mode callbacks. Define the NextCallback and FramePointer members of the debugger data block more portably with this.	2009-11-09 23:19:03 +00:00
i386	NMI Support Patch 1:	2009-12-31 23:51:26 +00:00
powerpc	- Don't use KeBugCheck(0) -- 0 is an invalid bugcode. Use a proper bugcode, DbgBreakPoint or UNIMPLEMENTED instead depending on how the old call was used.	2009-10-21 17:52:11 +00:00
apc.c
balmgr.c	- Remove svn:needs-lock, svn:eol-type, and svn:eol-tyle properties.	2009-10-12 03:35:35 +00:00
bug.c	Uncomment SYSTEM_THREAD_EXCEPTION_NOT_HANDLED case.	2009-12-10 08:54:53 +00:00
clock.c
config.c	- Fix a critical bug in KeFindConfigurationEntry; it passed NULL to KeFindConfigurationNextEntry instead of passing a pointer containing NULL. KeFindConfigurationNextEntry dereferenced this and, because the old bootloader maps NULL, read some bogus value. KeFindConfigurationNextEntry would then try to find something in the configuration tree that matched this, but failed. This was no problem because we currently have no callers of those routines in ReactOS. However, the kdcom.dll from Windows 2003 calls KeFindConfigurationEntry to find COM port entries, and this would have crashed if FreeLdr hadn't mapped it. As it is, it didn't, and kdcom fell back to default values for the COM port (which worked). WinLdr doesn't map NULL and this resulted in a crash during kdcom initialization. Bug introduced in revision 15911 over 4 years ago. KD64 now works when booting with the new boot method.	2009-10-11 21:57:52 +00:00
devqueue.c
dpc.c	ReactOS SMP Bringup Lite:	2009-09-30 20:30:57 +00:00
eventobj.c
except.c	- Fix a bug in KiRaiseException from 42923 -- a _SEH2_LEAVE wasn't converted to _SEH2_YIELD	2009-08-26 14:52:00 +00:00
freeldr.c	- Disable the MiCacheImageSymbols call in MmLoadSystemImage for KD too as this hack is required for rossym rather than KDBG. Fixes detection of non-boot driver images by WinDbg.	2009-11-08 01:13:49 +00:00
freeze.c	- Correct author of freeze.c (code was copied from cpu.c).	2009-11-10 16:42:31 +00:00
gate.c
gmutex.c
ipi.c
krnlinit.c
mutex.c
procobj.c	- Get rid of the concept and tracking of "mapped pages". The MapCount was never checked anywhere in terms of page accounting -- only the reference count is.	2009-07-26 08:20:29 +00:00
profobj.c	- INITIAL_STALL_COUNT makes more sense in decimal (100) than hex (0x64) so define it that way.	2009-10-29 19:58:41 +00:00
queue.c
semphobj.c
spinlock.c	ReactOS SMP Bringup Lite:	2009-09-30 20:30:57 +00:00
thrdobj.c	- Change CPUID to match the old Ki386Cpuid and take 4 output arguments instead of an array. This way we save some stack when using a dummy cpuid for synchronization and can query only the registers we want in the case we don't want all 4.	2009-09-27 10:09:38 +00:00
thrdschd.c	ReactOS SMP Bringup Lite:	2009-09-30 20:30:57 +00:00
timerobj.c	[3DTEXT, FREELDR, HAL, MINGW_COMMON, MMDRV, MSGINA, NTOSKRNL] Add extern.	2009-10-18 18:52:56 +00:00
wait.c	- Use _SEH2_YIELD when returning from an exception instead of returning outside the SEH block. Avoids unnecessary status checks for the most common case (no exception). Move the cleanup code into the handler too in favor of the no-exception case. Futhermore, don't call ExSystemExceptionFilter when we know we are called from user mode. Finally, only enter SEH if we need to do any probing.	2009-08-24 19:58:15 +00:00