…n information to the README; added some minor files to the repo for preserveration: commands.txt contains commonly used commands & tests, fastboot_instructions.txt has the commands used when loading a saved image to an Android device, fix_server.sh is a small script to restart the adb server when deploying to Android, stack_template.cfg is an example stack containing only the template & logger modules.
… the workshop handout to the README while the wiki is down; added logger_iperf module, a version of the logger module specialized for receiving iperf UDP traffic directly from the interface module (as is done in Experiment 2 of the TMC paper); iperf_udp_client.c is a generator that creates traffic similar to iperf but also provides the ability to set the data length or set the ethernet-level throughput, while iperf_udp_server.c is a simple traffic sink for verifyin iperf_udp_client over the traditional stack; added client_udp_test.c in accordance with previous commits; closed a couple rare and minor memory leaks in the TCP module; changed TCP_PROTOCOL / UDP_PROTOCOL / ICMP_PROTOCOL constants in most modules to be module specific to maintain module compartmentalization; added code for capturing frames from the local loopback of the traditional stack (currently commented out) in ethermod.c for those wanting to do so; minor debug changes.
…e wedge, PRINT_INFO() & PRINT_WARN(), changed many of the previously PRINT_IMPORTANT output to PRINT_INFO to reduce the amount of output from the wedge & increase performance; removed debug code left in udpHandling; minor debug changes to udpHandling/icmpHandling/tcpHandling.
… fixed bugs with poll & TCP sockets; plugged a few minor memory leaks related to poll; created a patch for when binding or autobinding to INADDR_ANY (may change in the future such that the TCP module receives INADDR_ANY & matches to multiple addreses); changed library functions in finstypes & metadata from <function_name>_full() to <function_name>_macro() to better reflect that they are meant to be used through the secure_<function_name>() macro and prevent confusion with operators such as list_full(); added more content & comments to the template module to better show how to add a new module & examples of functionality provided in libs
… wedge, PRINT_INFO() & PRINT_WARN(), changed many of the previously PRINT_IMPORTANT output to PRINT_INFO to reduce the amount of output from the wedge & increase performance; removed debug code left in udpHandling; minor debug changes to udpHandling/icmpHandling/tcpHandling.
…fixed bugs with poll & TCP sockets; plugged a few minor memory leaks related to poll; created a patch for when binding or autobinding to INADDR_ANY (may change in the future such that the TCP module receives INADDR_ANY & matches to multiple addreses); changed library functions in finstypes & metadata from <function_name>_full() to <function_name>_macro() to better reflect that they are meant to be used through the secure_<function_name>() macro and prevent confusion with operators such as list_full(); added more content & comments to the template module to better show how to add a new module & examples of functionality provided in libs
…erouted internal errors through ICMP module & IPv4; fixed bug in Daemon when flow[x]==0; added template module as example of adding module; added secure_config_setting_add() to simplify init_knobs(); working on problem with poll() in tcpHandling.c; renamed some constants
…andlers and TCP
…rnel 3.8.0; fixed medium Daemon poll bug where waiting poll calls wouldn't be triggered; fixed Bind issue with address (0.0.0.0:0); included changes to handle MAC address unresolved errors; simple testing with dig/firefox, simple tests suggest work but not full functionality
…ow compilation of socket interceptor for Linux kernel versions 2.6.38 to 3.8
…; minor renaming of a couple of functions in TCP; small changes to output for the wedge, converted 2 print lines to PRINT_IMPORTANT for better tracking; minor changes to output in interface, daemon, icmp, and tcp
…ive opening on listen); removed start up of accept thread in TCP when get accept() from Daemon; changed conn_stub to a conn & merged conn_stub_list_sem into conn_list_sem; patched some memory leaks
…f NewReno that FR for partial recoveries; fixed creation of duplicate ACKs and improved when they are created
… changes to debug output; small tweaks to TCP
…d install directory to allow for editing by Android text apps; changed some PRINT_ERRORs to PRINT_WARN and improved the output message; small improvement to recvmsg to return msg_flags
…aries in favor of static linked libraries, after extensive testing this was found as the cause of the thread scheduling problems that caused interface to stop and in turn capturer; small changes to placement of code (i.e. execute before or after capture/inject fork) to help stop creating zombie/defunct capture process on core termination; minor changes to speed_udp to better print at high throughput/lengths
…ucture, added fins_module_flow struct in the table, updated module data's & the get/set for flow/link/dual to match; added timestamps to PRINT_*() macros, NOTE: not garuanteed order; changed output, especially for startup; optimized capturer & removed output for better throughput, still haven't solved interface scheduling issue
… SET_PARAM one) to update the flows & links; removed valgrind flags from CCOPTS in settings.finsmk; improved speed_udp applications; improved some debug, especially related to start-up & common errors; fixed some udp/icmp recv behavior incorrectly implemented (rare case); fixed incorrect usage of totalReceived in udp.c; minor changes
… applications; NOTE: last commit removed most of the old prototypes & tests, these are planned to move to 'prototypes' branch
…uding 1-2 close bugs, a window size bug, and other edge cases, changed behavior of sem_to_handler; changed naming scheme of some #defines to be more aligned, in general is <THIS_MODULE>_<CTRL_OP>_<PARAM_ID> (e.g. ARP_EXEC_GET_ADDR); added support for shutdown() system call for SOCK_STREAM sockets, moved tcp<->daemon poll/shutdown event calls from exec to alert; added speed_tcp to tests as a simple tcp throughput tester; minor edit to build files
…caught strange edge case in daemon & fixed (delt with irregular sized packets and recvmsg buffers); renamed: tcp_connection->tcp_conn, tcp_connection_stub->tcp_conn_stub, tcp_segment->tcp_seg; added per connection/stub stats tracking, no RTM access yet though
…unnecessary test.cfg file; added 'note...' to stack.cfg
…nit() to not include flow and changed update to dual from links; implemented module pushing data to RTM and then to a listener application, modded console to act as either console (request/reply) or listener (reads events); added makefile to obj that relays every command to <FINS_ROOT>; integrated wedge makefile into overall system; improved Launcher's handling of Android events; minor edits to core to remove testing code; fixed minor bugs/memory leaks
…parately as a persistant process (will change to daemon, though can use & for now); this change allows core to be run without sudo for Abdallah's driver & running on Android; increased functionality of make install so that it correctly puts executables/so's for Android and installs the wedge for Ubuntu; minor name changes of nl header structures; fixed some minor TCP bugs, close bug still occurs, and added MSS parameter for RTM/TCP; transitioned some output to correct debug level
…s in daemon returning to wedge & having no buffer causing errors; fixed edge case so that if another core connects to the wedge, the wedge will correctly clear arrays; fixed edge case so that if the core dies & doesn't disconnect properly, wedge will fix itself correctly when a new core connects; expanded debug macros to 5 levels (debug, info, warn, important, error) with specific console colors for each; moved dynamic loading functions to common; added useful scripts in <FINS_ROOT>/obj; improved makesystem slightly with 'make install' for android; reordered route table so most specific first, least specific last, and to allow for local routing
…ed promiscuous mode in capturer; small edits to comments/debug in other files; moved testing executable test.c
… randomly occuring because of memory mismanagement from using the skb pointer; moved netlink related functions to new file in daemon
… stub to platform independent; minor other changes to wedge to try fix randomly occuring error
…ome scripts for common usage; changed some debug statements