75cddd8eb
Denis Vlasenko
2008-02-13
Subject
syslogd: send '\n'-terminated messages over the network. fully closes bug 1574. +8 bytes.
Binary
No binary (build failed)
Architecture & Compiler Flags
—
Build Log
(log yok)
Commit Diff GitHub →
sysklogd/syslogd.c+21 -15
| @@ -370,13 +370,15 @@ static void parse_fac_prio_20(int pri, char *res20) | |
| c_fac = facilitynames; | |
| while (c_fac->c_name) { | |
| if (c_fac->c_val != (LOG_FAC(pri) << 3)) { | |
| - | c_fac++; continue; |
| + | c_fac++; |
| + | continue; |
| } | |
| /* facility is found, look for prio */ | |
| c_pri = prioritynames; | |
| while (c_pri->c_name) { | |
| if (c_pri->c_val != LOG_PRI(pri)) { | |
| - | c_pri++; continue; |
| + | c_pri++; |
| + | continue; |
| } | |
| snprintf(res20, 20, "%s.%s", | |
| c_fac->c_name, c_pri->c_name); | |
| @@ -428,6 +430,9 @@ static void timestamp_and_log_internal(const char *msg) | |
| timestamp_and_log(LOG_SYSLOG | LOG_INFO, (char*)msg, 0); | |
| } | |
| + | /* tmpbuf[len] is a NUL byte (set by caller), but there can be other, |
| + | * embedded NULs. Split messages on each of these NULs, parse prio, |
| + | * escape control chars and log each locally. */ |
| static void split_escape_and_log(char *tmpbuf, int len) | |
| { | |
| char *p = tmpbuf; | |
| @@ -556,30 +561,24 @@ static void do_syslogd(void) | |
| size_t sz; | |
| read_again: | |
| sz = safe_read(sock_fd, G.recvbuf, MAX_READ - 1); | |
| - | if (sz < 0) { |
| + | if (sz < 0) |
| bb_perror_msg_and_die("read from /dev/log"); | |
| - | } |
| - | /* Drop trailing NULs (typically there is one NUL) */ |
| + | /* Drop trailing '\n' and NULs (typically there is one NUL) */ |
| while (1) { | |
| if (sz == 0) | |
| goto read_again; | |
| /* man 3 syslog says: "A trailing newline is added when needed". | |
| * However, neither glibc nor uclibc do this: | |
| * syslog(prio, "test") sends "test\0" to /dev/log, | |
| - | * syslog(prio, "test\n") sends "test\n\0", |
| + | * syslog(prio, "test\n") sends "test\n\0". |
| * IOW: newline is passed verbatim! | |
| * I take it to mean that it's syslogd's job | |
| - | * to make those look identical in the log files */ |
| - | if (G.recvbuf[sz-1] && G.recvbuf[sz-1] != '\n') |
| + | * to make those look identical in the log files. */ |
| + | if (G.recvbuf[sz-1] != '\0' && G.recvbuf[sz-1] != '\n') |
| break; | |
| sz--; | |
| } | |
| - | /* Maybe we need to add '\n' here, not later? |
| - | * It looks like stock syslogd does send '\n' over network, |
| - | * but we do not (see sendto below) */ |
| - | G.recvbuf[sz] = '\0'; /* make sure it *is* NUL terminated */ |
| - | |
| /* TODO: maybe suppress duplicates? */ | |
| #if ENABLE_FEATURE_REMOTE_LOG | |
| /* We are not modifying log messages in any way before send */ | |
| @@ -590,14 +589,21 @@ static void do_syslogd(void) | |
| if (-1 == G.remoteFD) | |
| goto no_luck; | |
| } | |
| + | /* Stock syslogd sends it '\n'-terminated |
| + | * over network, mimic that */ |
| + | G.recvbuf[sz] = '\n'; |
| /* send message to remote logger, ignore possible error */ | |
| - | sendto(G.remoteFD, G.recvbuf, sz, MSG_DONTWAIT, |
| + | /* TODO: on some errors, close and set G.remoteFD to -1 |
| + | * so that DNS resolution and connect is retried? */ |
| + | sendto(G.remoteFD, G.recvbuf, sz+1, MSG_DONTWAIT, |
| &G.remoteAddr->u.sa, G.remoteAddr->len); | |
| no_luck: ; | |
| } | |
| #endif | |
| - | if (!ENABLE_FEATURE_REMOTE_LOG || (option_mask32 & OPT_locallog)) |
| + | if (!ENABLE_FEATURE_REMOTE_LOG || (option_mask32 & OPT_locallog)) { |
| + | G.recvbuf[sz] = '\0'; /* ensure it *is* NUL terminated */ |
| split_escape_and_log(G.recvbuf, sz); | |
| + | } |
| } /* for (;;) */ | |
| } | |