- With these APIs we now have a skeleton NTLPC that is fully NT-compatible. Thanks to ea's (Emanuele)'s smss work which he strived to make portable, building with NTLPC = 1 will also build a compatible smss that now works.
- CSRSS requires some changes since it was not written to be NTLPC portable.
svn path=/trunk/; revision=24673
- Kernel support is complete (except temporary LPC hack), and CSRSS properly receives the event and sends it off to win32csr's hard error handler (much like on NT). However, the handler doesn't call NtUserHardErrorControl nor do much of the complex sequence of events that NT does, instead, we only display a hello-world style message box. Anyone up for implementing the rest is welcome.
svn path=/trunk/; revision=24665
- Fix various LPC structures in the NDK. Fix some prototypes too.
- Don't export PEPORT or PLPCP_PORT_OBJECT outside of /lpc and use PVOID isntead.
- Add a new RBUILD setting called NTLPC, which enables NTLPC when turned on.
svn path=/trunk/; revision=24664
-removed CRT_SECURE_NO_DEPRECATE because it is declared in warning.h
-set warning level to /W3
-set calling convention to stdcall
-use nasmw for asm files
svn path=/trunk/; revision=24662
- Stubplement the System Hard Error handler (it simply bugchecks)
- Fully implement ExRaiseHardError when called from user-mode so that the strings and parameters are properly captured into user-mode memory.
- Almost completely implement ExpRaiseHardError which is responsible for actually doing the error. However, ROS LPC is screwed up so sending the LPC Message crashes since the port isn't in a listening state or something.
- Would appreciate some advice on how to get the LPC message to CSRSS.
- Fix Get/SetErrorMode. The Native API uses an inverted value of SEM_FAILCRITICALERRORS, so this must be handled in win32.
- Cleanup formatting of RaiseException, and only normalize the paramter count if we're actualy going to use it.
- Implement LpcRequestWaitReplyPort as a big hack, but it should work.
svn path=/trunk/; revision=24658
- Get rid of FLUSH_TLB macros and use KeFlushCurrentTb and the MSVC_compatible instrinsic __invlpg;
- Fix some prototypes.
- Fixed unsigned/signed mismatches.
- Fix SEH usage in ExAllocatePoolWithQuotaTag.
svn path=/trunk/; revision=24652
- Add MSVC versions for some of them (not yet complete).
- Fix the fact that KeGetLocalDescriptorTable was setting the LDT instead of retrieving it.
- Fix bug in LIST_FOR_EACH and LIST_FOR_EACH_SAFE which was setting the flink to NULL instead of checking if the flink is NULL. One more reason these damned macros should've never been used.
- Use MSVC-intrinsics when applicable (_disable/_enable, etc).
- Fix JOB_SET_ARRAY problems.
- Fix buffer overflow in SystemProcessInformation QSI_DEF.
- Fix some broken compares/arithmetic to due to lack of parens.
- Add some ASSERTS to some unknown functions that make pointer assumptions.
svn path=/trunk/; revision=24650
- Add all the Nt functions to sysfuncs.lst and fixup some entries which were out-of-order. Unless I skipped anything (might've skipped 1 or 2 APIs), we should now have the same system call table as Windows 2003 SP1 (useful for my testing of ntdll.dll inside Windows).
svn path=/trunk/; revision=24638
- Add NtAddDriverEntry, NtDeleteDriverEntry, NtEnumerateDriverEntries, NtModifyBootEntry, NtModifyDriverEntry, NtQueryDriverEntryOrder, NtSetDriverEntryOrder.
- These are for EFI (or pseudo-EFI if we ever emulate Vista).
svn path=/trunk/; revision=24627