HostDrivers, Runtime, Devices, Additions: TSC delta measurement and other changes resulting from bumping supdrv major version. TSC delta measurement currently disabled.
VBoxNetFlt-darwin.cpp: Don't quit early just because we fail to spot p_count once, since we're potentially subject to a race with others changing the promiscous mode settings.
VBoxNetFlt-darwin.cpp: style + only detect p_count offset once.
Netflt: ifnet::if_pcount offset detection for Darwin (#7279)
OS X host: stuck-in-dock: move include/VBox/VBoxNetSend.h to src/VBox/HostDrivers/darwin/VBoxNetSend.h.
VBoxNet*darwin: Corrected two function delcarations (use DECL macros!), header guard and a few other nits. VBoxNetSend.h should move into src/VBox/HostDrivers/darwin/, btw.
VboxNetSendDummy() -> VBoxNetSendDummy().
Mac OS X host: prevent stuck-in-dock: r91968 is substituted with the new solution: send dummy packet in a context of a kernel thread in order to associate interface's receive thread w/ our new kernel thread instead of VirtualBoxVM process.
Mac OS X host: Networking: prevent stuck-in-dock problem if cable is not attached to a network card device.
header (C) fixes
bugref..
RTSpinlock: Redid the interface, eliminating NoInts and Tmp. Whether a spinlock is interrupt safe or not is now defined at creation time, preventing stupid bugs arrising from calling the wrong acquire and/or release methods somewhere. The saved flags are stored in the spinlock strucutre, eliminating the annoying Tmp variable. Needs testing on each platform before fixing the build burn.
Merge comment blocks.
vboxnetflt: guest to host packet capture fix (#5817)
vboxnetflt: partial BPF support for bridged interfaces (#5817)
vboxnetadp: BPF support for vboxnetX devices (#5817)
scm cleanup run.
*: spelling fixes, thanks Timeless!
vboxNetFltDarwinMBufFromSG: Removed obsolete assertion (triggers when intnetR0TrunkIfSendGsoFallback sends stuff).
iprt/asm.h,*: Added ASMAtomicWriteNullPtr and ASMAtomicUoWriteNullPtr to better deal with NULL being 0 in C++.
iprt/asm.h,*: Revised the ASMAtomic*Ptr functions and macros. The new saves lots of unsafe (void * volatile *) casts as well as adding some type safety when using GCC (typeof rulez).
IntNet: Added Interface, Interface private data passing for per-interface based VBoxNetFlt.
IntNet: added MAC address notification and connect/disconnect interface callbacks.
IntNet,VBoxNetFlt: Cleaned up the locking protocol between IntNet and NetFlt. Eleminated the out-bound trunk lock that IntNet always took when calling NetFlt.
Automated rebranding to Oracle copyright/license strings via filemuncher
IntNet/NetFlt: Added INTNETTRUNKSWPORT::pfnReportGsoCapabilities and enabled ring-0 transmission for linux hosts and wires.
intnetinline.h: Changed the prefix to IntNet.
SrvIntNet,VBoxNetFlt: Changing from getting to reporting (VBoxNetFlt reports stuff to SrvIntNet).
build fix.
VBoxNetFlt: Deal with multi segment SG buffers. TODO: One weird windows code path.
intnet, VBoxNet*, network device & drivers: GSO preps.
*: s/RT_\(BEGIN|END\)_DECLS/RT_C_DECLS_\1/g
HostDrivers: __BEGIN_DECLS -> RT_BEGIN_DECLS; __END_DECLS -> RT_END_DECLS.
VBoxNetFlt/darwin: corrected comment and made use of 'else'.
VBoxNetFlt: Fixed vboxNetFltOsInitInstance which had gotten out of sync.
VBoxNetFlt: review of core changes, several review comments needing follow up. untested
#3419: No promisc for WiFi on Darwin either.
fixed OSE headers for linux/darwin/solaris netfilter code
All: license header changes for 2.0 (OSE headers, add Sun GPL/LGPL disclaimer)
filemuncher spaces
VBoxNetFlt for Darwin and Solaris to OSE