diff --git a/tap-rpcstat.c b/tap-rpcstat.c index 9a0fb04024..3f7245f425 100644 --- a/tap-rpcstat.c +++ b/tap-rpcstat.c @@ -23,7 +23,7 @@ */ /* This module provides rpc call/reply SRT statistics to tshark. - * It is only used by tshark and not wireshark + * It is only used by tshark and not wireshark. * * It serves as an example on how to use the tap api. */ @@ -72,9 +72,10 @@ typedef struct _rpcstat_t { * When registering below, we could just have left this function as NULL. * * When used by wireshark, this function will be called whenever we would need - * to reset all state. Such as when wireshark opens a new file, when it + * to reset all state, such as when wireshark opens a new file, when it * starts a new capture, when it rescans the packetlist after some prefs have * changed etc. + * * So if your application has some state it needs to clean up in those * situations, here is a good place to put that code. */ @@ -96,27 +97,26 @@ rpcstat_reset(void *prs) } -/* This callback is invoked whenever the tap system has seen a packet - * we might be interested in. - * The function is to be used to only update internal state information - * in the *tapdata structure, and if there were state changes which requires - * the window to be redrawn, return 1 and (*draw) will be called sometime - * later. +/* This callback is invoked whenever the tap system has seen a packet we might + * be interested in. The function is to be used to only update internal state + * information in the *tapdata structure, and if there were state changes which + * requires the window to be redrawn, return 1 and (*draw) will be called + * sometime later. * - * This function should be as lightweight as possible since it executes together - * with the normal wireshark dissectors. Try to push as much processing as - * possible into (*draw) instead since that function executes asynchronously - * and does not affect the main threads performance. + * This function should be as lightweight as possible since it executes + * together with the normal wireshark dissectors. Try to push as much + * processing as possible into (*draw) instead since that function executes + * asynchronously and does not affect the main thread's performance. * * If it is possible, try to do all "filtering" explicitly as we do below in * this example since you will get MUCH better performance than applying * a similar display-filter in the register call. * - * The third parameter is tap dependant. Since we register this one to the "rpc" - * tap the third parameter type is rpc_call_info_value. + * The third parameter is tap dependent. Since we register this one to the + * "rpc" tap, the third parameter type is rpc_call_info_value. * - * The filtering we do is just to check the rpc_call_info_value struct that - * we were called for the proper program and version. We didn't apply a filter + * The filtering we do is just to check the rpc_call_info_value struct that we + * were called for the proper program and version. We didn't apply a filter * when we registered so we will be called for ALL rpc packets and not just * the ones we are collecting stats for. * @@ -186,16 +186,15 @@ rpcstat_packet(void *prs, packet_info *pinfo, epan_dissect_t *edt _U_, const voi return 1; } -/* This callback is used when tshark wants us to draw/update our - * data to the output device. Since this is tshark only output is - * stdout. - * TShark will only call this callback once, which is when tshark has - * finished reading all packets and exits. +/* This callback is used when tshark wants us to draw/update our data to the + * output device. Since this is tshark, the only output is stdout. + * TShark will only call this callback once, which is when tshark has finished + * reading all packets and exits. * If used with wireshark this may be called any time, perhaps once every 3 * seconds or so. - * This function may even be called in parallell with (*reset) or (*draw) - * so make sure there are no races. The data in the rpcstat_t can thus change - * beneath us. Beware. + * This function may even be called in parallel with (*reset) or (*draw), so + * make sure there are no races. The data in the rpcstat_t can thus change + * beneath us. Beware! */ static void rpcstat_draw(void *prs) @@ -259,11 +258,13 @@ rpcstat_find_procs(gpointer *key, gpointer *value _U_, gpointer *user_data _U_) /* When called, this function will create a new instance of rpcstat. - * program and version are which onc-rpc program/version we want to - * collect statistics for. - * This function is called from tshark when it parses the -z rpc, arguments - * and it creates a new instance to store statistics in and registers this - * new instance for the rpc tap. + * + * program and version are which onc-rpc program/version we want to collect + * statistics for. + * + * This function is called from tshark when it parses the -z rpc, arguments and + * it creates a new instance to store statistics in and registers this new + * instance for the rpc tap. */ static void rpcstat_init(const char *optarg, void* userdata _U_) @@ -320,14 +321,15 @@ rpcstat_init(const char *optarg, void* userdata _U_) rs->procedures[i].tot.nsecs=0; } -/* It is possible to create a filter and attach it to the callbacks. Then the +/* It is possible to create a filter and attach it to the callbacks. Then the * callbacks would only be invoked if the filter matched. - * Evaluating filters is expensive and if we can avoid it and not use them - * we gain performance. - * In this case we do the filtering for protocol and version inside the + * + * Evaluating filters is expensive and if we can avoid it and not use them, + * then we gain performance. + * + * In this case, we do the filtering for protocol and version inside the * callback itself but use whatever filter the user provided. * (Perhaps the user only wants the stats for nis+ traffic for certain objects?) - * */ error_string=register_tap_listener("rpc", rs, filter, 0, rpcstat_reset, rpcstat_packet, rpcstat_draw);