task switching - We now are using software switching to be consistant with the rest of the world because all of this open source freaks gave me a hard time about something I liked. There doesn't seem to be any gain in performance but it is working fine and flawlessly

TCA Is A BONER!!!!

updates: converted to software task switching however it is not working yet

Changes

adjustCowCounter: we no longer need to debug unhandled adjustments they are normal situations now

make sure it still works before I goto bed

Major: ...

Pages now have strict security enforcement.
  Many null dereferences have been resolved.
  When apps loaded permissions set for pages rw and ro

bah

Major: ...

Disconnected page 0x0 from the system... Unfortunately this broke many things
  all of which have been fixed. This was good because nothing deferences NULL
  any more.

Things affected:
  malloc,kmalloc,getfreepage,getfreevirtualpage,pagefault,fork,exec,ld,ld.so,exec,file

chg: fixed marks bug readFile did a lookup which is why it looked like it was loopping so much

chg: made page 0x0 writeable should boot now

Put in a test for 0x0

chg: fixed ubu's typo

Try now TCA

test code, fixes and the like

Last of my fixes for the morning

ok last of my commits for the night I'm going to bed

Oh yeah try this one

Changes... mark let me know if you fault at pid 185 when you type stress

Lots of changes to the vmm subsystem.... Page faults have been adjust to now be blocking on a per thread basis not system wide. This has resulted in no more deadlocks.. also the addition of per thread locking has removed segfaults as a result of COW in which two tasks fault the same COW page and try to modify it.

vmm_pageFault: deadlock resolved thanks to a propper solution suggested by geist

I broke it hopefully

assert: remade assert

CFLAGS: I updated cflags so optimizations are consistant on all files different optimization levels maybe the source of some of our problems?

cleaned out some debug code

vmmPageFault: had a potential memory leak here for one page it was still using sysID on certain COW scenarios

memory.c: added spinlocks to vmmFindFreePage and vmmFreePage to prevent two tasks from possibly allocating the same page

vmmCopyVirtualSpace: cleaned up one full page memory leak we were still using old sysID over pid

kmalloc: added assert() ...

bioscall: fixed some potential problem by not making 16bit code
paging: added assert()

Updating Initialization Procedures