Additions: comment nits.
fixed some Log() statments where we didn't properly display uint64_t or RTUINTPTR variables
Guest Additions: add release logger to VBoxGuest driver (enabled for darwin only); make VBoxGuest.kext workable on 64 bit system.
BUGZ:4686 OS X guest: VBoxGuest device driver + VBoxService and VBoxControl now work.
header (C) fixes
Let the other hosts use page lists again.
VBoxGuest: disable page list HGCM transfers for Locked buffers to prevent crashes on Win guests
Additions/common/VBoxGuestLib: fixed LOG_GROUP define
Additions/common/VBoxGuestLib: logging.
HGCMInternal.cpp: RTR0MemObjSize does the alignment assertions now.
fixed assertion
Additions/common/HGCMInternal: added an assertion
Additions/common/HGCMInternal: typo which caused to allocate a buffer of cb * PAGE_SIZE
*: spelling fixes, thanks Timeless!
Additions/VBoxGuestLib: fix invalid LogRel statement.
Additions/VBoxGuest: limit the occurrence of some more LogRel statements
HGCMInternal.cpp: Avoid unused variable on non-linux platforms.
Additions: fixed the gcc C90 warnings regardings mixed declarations and code
crOpenGL: fixed OpenGL support for 32bit apps under 64bit Windows guests
warning (unused variable)
crOpenGL: fix ETQW failure on level load
Automated rebranding to Oracle copyright/license strings via filemuncher
VBoxGuest: Hrmpf!
VBoxGuest: be less verbose on timeouts
alternative license for VBoxGuestLib is CDDL
Additions/common, Additions/Solaris: Fix logging.
Additions/VBoxGuestLib: removed debug LogRel.
Additions/VBoxGuestLib: Oops.
Additions/VBoxGuest,VBoxGuestLib: fixed HGCM 32-bit calls in 64-bit kernel driver. OpenGL works again.
IPRT,VMM,SUPDrv,VBGLR0: Added a parameter to RTR0MemObjLockUser/Kernel that indicates read/write intent so we can correctly lock readonly memory on Windows and OS/2. (Guest property strings, see #4238.)
crOpenGL: allow to transfer upto 16megs in r3 hgcm calls, fixes compiz etc on linux
HGCMInternal.cpp: missed HGCM parameter type conversion.
page list bufix.
VBoxGuest,VMMDev: New HGCM cancellation request.
VbglR0HGCMInternal.cpp: Enable the physical page list code everywhere.
VbglR0HGCMInternal.cpp: Less wasteful bounce buffering, some bugfixes, logging.
VbglR0HGCMInternal.cpp: Wrote and tested the physical page list code. Implemented bounch buffering for Linux. (both disabled)
bufix.
VMMDevHGCMParmType_PhysAddr is not a valid parameter any longer.
HGCM&Co: physical page list.
HGCMInternal.cpp,SysHlp.cpp: use the kernel/user mode indicator.
VBGL: Renamed the internal HGCM functions and added a kernel/user mode indicator to HGCM the calls. The code is not yet using the indicator for anything.
VBoxGuest.h,VBoxGuestLib: Moved the VbglR3 API out of VBoxGuest.h and did some cleanup.
Todos added
Introduced VbglHCGMCall32 for supporting 32 bits clients on 64 bits guests.
Additions/Guest Library: comments and logging
s/%Vr\([acfs]\)/%Rr\1/g - since I'm upsetting everyone anyway, better make the most of it...
s/VBOX_SUCCESS/RT_SUCCESS/g s/VBOX_FAILURE/RT_FAILURE/g - VBOX_SUCCESS and VBOX_FAILURE have *NOT* been retired (because old habbits die hard) just sligtly deprecated.
Implemented cancelling of HGCM requests from the guest.
The Big Sun Rebranding Header Change
The Giant CDDL Dual-License Header Change.
Guest additions: deactivated some logging which was filling up my release log with a debug build of the additions
gcc warning
Bugfix for type changes
Cleaned up
Extended interface for locked pointers
Backed out 23565: alloca is not present in NT4
Use alloca instead of wasting 1k/2k of kernel stack.
Biggest check-in ever. New source code headers for all (C) innotek files.
Wrong locking for read-only memory in Windows guests.
Added a PORTME notice about a potential problem with kernel/user address space separation and assumptions on the VMMDev side.
Do not use alloca in the guest additions.
prevent warnings if compiled as C files (Linux guest modules)
Lock user buffers in a HGCM call (Windows guest only)
InnoTek -> innotek: all the headers and comments.
some cleanup
fixed includes
Logging update
HGCM: VbglHGCMCall shouldn't return failure if the request produced an error; it should only do so if the request could not be completed
Prevent some warnings as these files are compiled as .c files on Linux now.
Changed the Linux Additions to compile everything from source (no more binaries linked in) and cleaned up those files to compile as plain C.
import