- There is no reason to implement the Interlocked* APIs in assembly in the kernel -- just have them call the intrinsics.
- Also fix the way we were overloading the exported APIs, we now use the intrinsics all the time.
- Convert loader block address to virtual, and parse the addresses inside it and convert them to virtual too, after we don't use them in FreeLDR anymore -- the kernel will delete the p->v identity mapping, and eventually still need to touch the loader block.
- Implement MmGetPhysicalAddress and MiGetUserPageDirectoryCount for ARM.
- We now reach the point where we start initializing the boot drivers' memory information.
svn path=/trunk/; revision=32654
- Fix incorrect definition of ARM_CONTROL_REGISTER, which was stopping us from correctly enabling High Vectors.
- The RtlCaptureStackLimits and RtlWakeChainFrame routines in libsupp.c cannot possibly be portable across all architectures -- separated ARM versions in an arm directory.
- Setup the panic stack for abort/undefined exceptions, and the interrupt stack for FIQs.
- Implement support for exceptions! We can now display the address which caused a data abort, and begin handling exceptions.
- Implement all the HAL Spinlock/IRQL functions except KfRaise/LowerIrql which actually do the work.
- We're booting all the way to setting up the user_shared_data memory area.
svn path=/trunk/; revision=32653
- Implemented and defined the MMU-OS architecture for the ARM port. The details are too long for a commit message, but we have decided to replicate the x86 NT memory manager layout. We've defined a PTE_BASE at 0xC0000000 just like on x86, and we use a PDE_BASE at 0xC1000000. Unlike the x86, we can't use PDE-PTE self-mapping because ARM has different formats (and sizes!) for PDE vs PTEs! We emulate the behavior however (which adds a small performance hit) and the Mm porting is thus at least 10 times easier.
- Moved serial port to 0xE0000000 for now.
- We now parse the board memory map from u-boot.
- Added memory allocation code to FreeLDR -- we now build a full ARC memory map for the kernel.
- FreeLDR allocates page tables and sets up the initial support for our memory layout (see comments for some lengthier explenations)
- Allocations made by FreeLDR for loading ReactOS are now made from a "shared heap" page that's also marked in the memory map.
- Registry and NLS data are now being put into the loader block.
- We now create a loader entry for the kernel (but not anything else -- we'll have to parse the list properly later).
- Defined correct _HARDWARE_PTE_ARM and _MMPTE_HARDWARE for ARM.
- ARM_COARSE_PAGE_TABLE is now 4KB instead of 1KB, going against the architecture! We do this for proper OS support of the PTE_BASE.
- Fixed build due to KiSystemStartulReal change.
- Fixed a bug on the x86 build when creating memory allocation descriptors. Memory corruption could occur in certain scenarios.
- Implemented significant portions of the ARM memory manager code in the kernel:
- MmGetPageDirectory.
- MmDeletePageTable (for the kernel address space only).
- MmIsPagePresent (for the kernel address space only).
- MmCreateVirtualMappingForKernel.
- MmCreateVirtualMapping (calls MmCreateVirtualMappingUnsafe).
- MmCreateVirtualMappingUnsafe (for the kernel address space only).
- MmSetPageProtect (unused on ARM).
- MmCreateHyperspaceMapping.
- MmDeleteHyperspaceMapping.
- MmInitGlobalKernelPageDirectory.
- MmInitPageDirectoryMap.
- With the above, this means we now go well inside MmInit1: the PFN database is setup and works, memory areas are functional, and non-paged pool is fully working.
- We currently hit a data abort during paged pool setup -- this is to be expected, since we don't have any exception handlers yet. These are coming up next -- we have to start handling crashes (and page faults).
svn path=/trunk/; revision=32640
implemented command DRAFTMODE, FLUSHOUTPUT, SETCOLORTABLE for they are not longer supported in windows XP/2003
implemented command SETABORTPROC it is still stub in gdi32/misc/stubs.c
svn path=/trunk/; revision=32635
- Add two new fields into PHYSICAL_PAGE structure, they will be used (in the future new pool implementation) to find end/start of a non-paged pool allocation without storing the actual pointers (based on NTs own PFN entry optimization for this) - thanks to Alex for providing this great idea!
svn path=/trunk/; revision=32630
implemented command GETCOLORTABLE and GETEXTENDEDTEXTMETRICS in Escape
MS have releae docs over this api in msdn, u can also found docs in nt4 ddk and win98 ddk how this api works.
and thanks to Filip and Blight that have explain how this api works time to time, in gdi32
svn path=/trunk/; revision=32628
Remove the TempName variable from the previous cabman commit and pass the temp file directly again.
- Add the same "const"'s to sysreg, so that it compiles with gcc >= 4.2 as well
svn path=/trunk/; revision=32619
- Assume VACB_MAPPING_GRANULARITY as a default cache segment size.
- Use FSD callbacks instead of direct accessing FCB's MainResource via FileObject/FsContext. This allows removing of ObReferenceObject(FileObject) which I had to add some time ago in order to keep FileObject alive for the lazy writer thread.
- IMPORTANT: Testers should thoroughly try this revision in order to see if any new problems appear. If they do appear, this revision will most probably be reverted by me.
svn path=/trunk/; revision=32612
- TODO: Provide callbacks for lazy writer and read ahead when calling CcInitializeCacheMap. Could be done similar to r32607 for CDFS.
svn path=/trunk/; revision=32610
