- Convert NDIS_ERROR_CODE from a typedef to a define.
- Add some missing NDIS_STATUS_* definitions.
- Some _ANONYMOUS_UNION/_ANONYMOUS_STRUCT/__GNU_EXTENSION -> __MINGW_EXTENSION conversions.
- Update NDIS_REQUEST_TYPE and NDIS_PER_PACKET_INFO.
- Guard several definitions against NDIS_LEGACY_DRIVER.
- Improve NDIS_REQUEST.
- Add missing CO_ADDRESS_FAMILY_INFINIBAND, Ndis{Set,Get}PacketCancelId, NDIS_PACKET_{FIRST,LAST}_NDIS_BUFFER and NDIS_PACKET_VALID_COUNTS.
- Group related definitions.
- Update several Ndis* function declarations.
- Add missing NdisInterlocked{Push,Pop}EntryList and NdisGetCurrentSystemTime macros.
svn path=/trunk/; revision=48205
[NTOS]: Also stop creating a memory area for the illegal user-mode parts of address space.
[NTOS]: Instead, mark the area between MM_HIGHEST_VAD_ADDRESS and MM_HIGHEST_USER_ADDRESS as being ARM3 as well: this way, ARM3 will get the illegal access (and fault), and it will also get the shared user data page access.
[NTOS]: With the previous commit, ARM3 knows how to handle the shared user data page access, and does so succesfully. End result: two more MAREA types have been removed, and the address space setup code is now much simpler.
svn path=/trunk/; revision=48202
[NTOS]: Implement handling a very special case of "prototype PTE", the one used to map the shared user data for user-mode applications.
[NTOS]: MiCheckVirtualAddress detects this (Windows behavior) and returns a prototype PTE that's marked MM_READONLY. This is our MmSharedUserDataPte from before. This gets sent to MiDispatchFault which calls MiResolveProtoPteFault to handle it. In turn, this calls MiCompleteProtoPteFault. All these code paths have heavy ASSERTions to only allow them to be hit for the shared user data page, however, in the far distant future when prototype PTEs are used for section objects, we'll at least have the right checks and code flow (many of these ASSERTions will then have to be removed).
[NTOS]: The end result is that we also now have STATUS_PAGE_FAULT_TRANSITION, not just STATUS_PAGE_FAULT_DEMAND_ZERO, and that prototype PTEs are somewhat understood and some assumptions have been removed.
svn path=/trunk/; revision=48201
- Add a hack so we wait for TCP/IP to get ready and bind to adapters instead of failing immediately after the first probe
- Live CD and 2nd stage setup now have fully functional networking :)
svn path=/trunk/; revision=48195
[NTOS]: Allocate PEB and TEB using VADs! The user-mode fault handler seems to work fine, and I could find no regressions. This is the beginning of the end for MAREAs for VM allocations (they will remain for sections).
svn path=/trunk/; revision=48194
- Refresh the protocol bindings after a new miniport is added
- Adapters added after boot work now
- One step closer to Live CD networking
svn path=/trunk/; revision=48193
- Start the TCP/IP driver after we install a network adapter
- This fixes the major bug that blocks Live CD networking but we still have some DHCP issues (DHCP starts and fails before the NIC and TCP/IP has been installed) and some other adapter detection issues that I need to look into but this is a good first step
svn path=/trunk/; revision=48188
[NTOS]: Initialize MmTotalCommitLimit, MmTotalCommitLimitMaximum for consistency.
[NTOS]: Initialize and honor MmAllocationFragment value from registry, or from defaults.
[NTOS]: Initialize the system cache working set pushlock.
[NTOS]: Implement Mi(Un)Lock(Process)WorkingSet inlines to handle the various scenarios. Replace broken/hacky code that was in the system loader, fault handler, and VM deletion paths with the correct macros. No locking is done yet as ReactOS' MAREA implementation overloads the Vm field in EPROCESS, but the correct APC blockings and state transitions are done.
[NTOS]: Add another helper from symbols, MI_IS_PAGE_TABLE_OR_HYPER_ADDRESS, and use it instead of the math-by-hand.
[NTOS]: Add MM_IS_ANY_LOCK_HELD and MI_WS_OWNER to help out the new working let locking inlines (names taken from symbols).
[NTOS]: Add _1GB helper for large-memory checks. Also define _1KB as "1024u". The "u" makes GCC not complain when sizes reach >= 2GB, because GCC will interpret the integer literals as "signed", even if the final calculated value is in an unsigned variable.
svn path=/trunk/; revision=48187
- Remove getopt from mingw lib and put it into it's own static library
- change crt="dll" to crt="static" for the mingw startup libs, to avoid things like importing from itself.
svn path=/trunk/; revision=48181
- In the fat bootsector, don't clobber ax, it contains the number of first cluster of freeldr.sys and is used by the fathelper code.
- Fix jump code in the fathelper code, like in the other bootsectors.
- Now booting from fat partitions (looks like that's what sysreg does) works again. It's safe under the condition that the cluster size is at least 4352 bytes, which is true for harddisks of sizes bigger than 272MB. Booting from smaller fat disks, like floppy breaks when freeldr.sys gets fragmented, which should rarely happen.
svn path=/trunk/; revision=48177
[NTOS]: Knowing if this change causes issues will help later in determining regressions due to TEB/PEBs mapped as VADs by ARM3, and regressions simply due to the change in VA layout.
[NTOS]: When implemented, the VAD mapping for PEB/TEB will only use that ~1MB, which yes, will limit ReactOS processes to each have only 256 threads. That is obviously a temporary limitation, one I doubt we'll even hit, but I'm putting it out here so you know.
svn path=/trunk/; revision=48176
[NTOS]: MI_MAKE_HARDWARE_PTE becomes MI_MAKE_HARDWARE_PTE_KERNEL, since it assumed this. MI_MAKE_HARDWARE_PTE is now a "generic" you can use when you don't know what the PTE should be. It uses MiDetermineUserGlobalMask to set the right bits.
[NTOS]: Add two more helpers: MI_IS_PAGE_TABLE_ADDRESS and MI_IS_SYSTEM_PAGE_TABLE_ADDDRESS. One is in the symbols, the other I made up to make things clearer.
[NTOS]: MiResolveDemandZeroFault now knnows how to resolve user-demand-zero-faults.
[NTOS]: Implement MiZeroPfn to do the actual zeroing during user-demand-zero-faults (also later for VAD faults).
svn path=/trunk/; revision=48175
[RTL]: Uncouple generic table from AVL table implementation into its own avltable.c
[RTL]: Get rid of "Austin" and fix prototypes of AVL table functions.
[RTL]: Re-implement AVL table functions, sharing as much code as possible with the SPLAY tree implementation which is pretty decent. Lookup, insert, enumeration are implemented, but not delete.
[RTL]: Make large part of the RTL AVL package into its own "support" file that can work both with MMADDRESS_NODE and RTL_BALANCED_LINKS structures. The former is used by ARM3 for VADs.
[NTOS]: Implement basic VAD AVL tree routines (Insert, LookupEmpty, GetPrevious, CheckForConflict, Locate). This is enough to insert VADs, find a free address range, and locate a VAD by address. No delete yet
Thanks to Timo Kreuzer for some clever definitions, Knuth for his genius, several online C implementations for ideas, the HPI kernel blog for insight on how Windows does it, and others.
svn path=/trunk/; revision=48173
- Add a number of amd64 specific math functions
- add generic c versions of cos and sin
- Remove leading underscores from amd64 symbols
svn path=/trunk/; revision=48161
