RosMm uses MEMORY_AREAs to map sections and VACBs in system space. The previous implementation allowed them to be allocated everywhere, except for regions already used otherwise. This makes sense for x86, where the address space is limited, but we don't need that on x64.
This allows to place all other VA regions where we want without the need to allocate memory areas for completely unused regions that do not even have PXEs mapped.
This is required for both x86 and x64. On x86 we only got away without, because the page fault handler is buggy and considers a fault on kernel PTE addresses as a user mode fault and makes the PDE valid for us. On x64 this is not enough, because it only works for invalid PDEs, not for invalid PPEs and we only got away with this, because RosMm sections are allocated from the first range available for memory areas, which happens to be the system cache WS following directly after the shared user page, which already has a PPE mapped.
The bug in the fault handler needs to stay for now, since ARM3 also depends on it.
CORE-17604
- Disable some functions for ARM32;
- Remove some link options not used on ARM;
- Add get _controlfp() to LIBCNTPR to link properly on ARM;
- Unify Freeldr UI Drawing on ARM;
- Add qemu UART debugging for ARM32/ARM64.
- Use GetKeyboardLayoutList to get the keyboard list instead of using Preload registry key.
- Get the special IDs from registry to handle special HKLs in newly-added LoadSpecialIds function.
- Add GetKLIDFromHKL, GetHKLFromLayoutNum, UpdateLayoutList, and GetKLIDFromLayoutNum helper functions.
CORE-13145, CORE-10667, CORE-18924
The minimized non-task windows were not usable due to the bugs. In some situations, the system will restore the minimized non-task windows.
- Add IsTaskWnd helper function.
- Add SendPulseToTray function to send a pulse to the tray window.
- At some shell hook handlings, send a pulse to the tray window.
- Add IgnorePulse flag to control the timing of restoring.
- Add a timer to reset IgnorePulse flag.
- If the pulse has come and IgnorePulse flag is false, then restore the minimized non-task windows.
- Modify the rundll32 window procedure.
- Use WINDOWPLACEMENT to restore the minimized windows.
CORE-13895, CORE-18350
NOTE: There are no new features in this PR, we stay with upstream version 0.69.
CORE-18645
- Add ifdefs to distinguish ReactOS code/fixes from the upstream code;
- If upstream code was overriden, reintroduce it in #else section;
- Re-align .rc version info with upstream (0.69).
As on Windows, gives precedence to the user's extension in file names if there is one.
Otherwise appends the selected extension. If not selected, append default extension.
Correctly display the composition window and the candidate window.
- At update_caret function, set the position and font of the composition window.
- We don't use internal composition string. Rely on the composition window.
- Improve WM_IME_SETCONTEXT, WM_IME_CONTROL, WM_IME_SELECT, WM_IME_STARTCOMPOSITION, WM_IME_COMPOSITION and WM_IME_ENDCOMPOSITION message handlings.
CORE-11700
The data has to be written into ObjectTypeInfo based on the return length,
not only what is provided by the input buffer length. Fix suggested by
Hermès.
On a x86 system aligning the return length pointer to a 4-byte boundary
works best since pointers in general are 4-byte aligned for x86 systems.
However, what happens on a AMD64 system is that we still align this pointer
to 4-byte, ObjectTypeInfo is a 8-byte pointer and we might write into
the return length past the 4-byte boundary.
If one were to allocate a pool of memory with that length and query all
the object types info and free the said pool of memory thereafter, the
system will crash with BAD_POOL_HEADER because ObQueryTypeInfo overwrote
the return length past the 4-byte boundary length therefore leading up with
corrupted memory blocks in the pool header.
This symptom of BAD_POOL_HEADER happens exactly the same in Windows Server
2003 x64 Edition. Newer versions of Windows like 10 aren't affected.
But, Windows has another bug where they are using MaximumLength for the
calculation of the needed length to be returned to caller. MaximumLength
does not guarantee you that it includes the NULL-terminator in the length
and that potentially leads to a buffer overrun.
Also annotate the ObQueryTypeInfo function with SAL2.
https://processhacker.sourceforge.io/doc/object_8c_source.html (read the
comment in KphObjectTypeInformation).
- Flush Map registers once the DMA completes
- Free Map registers once the DMA completes
- Add support for SGL allocated from NonPagedPool
Test:
Force Allocations of SGL from Non Paged Pool and ensure OS boots and functions properly
Test Logs:
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
DHCPCSVC: Adapter Name: [{7cd69ac0-dabb-410a-b927-cb3961d174da}] (dynamic)
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
WARNING: HalCalculateScatterGatherListSize at hal\halx86\generic\dma.c:1168 is UNIMPLEMENTED!
SpiAdapterControlFORCING ALLOCATION FROM SGPOOL
SpiAdapterControlFORCING ALLOCATION FROM SGPOO
Second parameter is optional, so mark it as such and check whether it was passed. Fixes a sporadic 0x24 bugcheck caused by access violation when running ReactOS on NTFS volume with WinXP ntfs.sys.
