Before taking any default action if there is no timeout, check whether
the supplied key filter callback function may handle a specific user
keypress. If it does, the menu timeout is cancelled.
This allows e.g. handling F8 press for displaying boot options even when
the timeout is zero.
CORE-14046
We now run the 1st-stage setup with a regular system hive, similarly to
what's done for the LiveCD, or for a regular ROS installation.
The ExpInTextModeSetup hacks I previously removed are now completely unneeded.
svn path=/branches/setup_improvements/; revision=74762
Split the "main" winldr.h header into the one containing global code
that actually doesn't really depend on the "windows" NT loader part,
and one that actually concerns code just for the NT Loader.
The latter goes into "ntldr".
svn path=/branches/setup_improvements/; revision=74743
Volume boot record is also implemented, it supports reading BTRFS tree structures with upto 64k node size.
This support required to change all path in Free Loader to lowercase for better performance.
CORE-13769
This happened for the livecd, where the cdfs driver was inserted twice - once since it is the driver for the boot device, and the 2nd time, because it was specified in the registry - which was then initialized twice by the kernel, leading to a name conflict when trying to create another device object, which resulted in the newly created device object to be dereferenced again and the driver unloaded. This can be seen from the debug message "(ntoskrnl\mm\ARM3\sysldr.c:955) Leaking driver: cdfs.sys"
On x64 we only map 1GB of pages, so adjust MM_MAX_PAGE accordingly and also respect that value when searching for the best location of the page lookup table.
CORE-11048 #resolve
GAS uses ".double" symbol for declaring floating-point constants and
".quad" symbol for declaring 64-bit numbers.
This is not compatible with our macro for MASM and introduces bugs.
Now 64-bit constants are supposed to be declared using ".quad" macro.
- Plus: Add missing email address in 74f92c3d86.
- Remove superfluous blank lines.
- Add 2 TRACE("FunctionName()\n").
- Use "0x%llx", not "%lu", for BaseAddress and Length. (Copypasta in my recent commits.)
- Add "0x" to 1 "%p" and 1 "%lx".
[FREELDR] Update ACPI (Extended Attributes) support in PcMemGetBiosMemoryMap()
CORE-13332
* PcMemGetBiosMemoryMap(): Add Extended Attributes set and check for entry validity and default handling of unexpected case.
* pcbios.h: Rename superceded BIOS_MEMORY_MAP.Reserved. Adapt existing code to new ACPI 6.2-A definitions.
* pcbios.h: Update BIOS_MEMORY_TYPE and BIOS_MEMORY_MAP to ACPI 6.2-A from 1.0+.
* PcMemGetBiosMemoryMap(): Misc fixes, no functional changes.
*Create PcMemCheckUsableMemorySize(), to split unrelated code out.
*Fix a copypasta in 2 output strings from ba9a1c3abb.
*Improve output readability of TRACE("ECX ...", ...).
*Move a TRACE("\n").
*Improve code style a bit.
- We use this as a chance to also update the CMake file in terms of consistency.
- The clang-cl build applies the /Os flag on all the source files that are now using the PCH too.
Cast MAX_BIOS_DESCRIPTORS to explicit ULONG from implicit int.
Comment/Add checks for PcMemoryMap/PcBiosMemoryMap arrays being full and bare handling of error cases.
