From: Lennart Poettering Date: Fri, 26 Oct 2012 18:05:19 +0000 (+0200) Subject: journal: provide an API that allows client to figure out whether they need to recheck... X-Git-Tag: v196~217 X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=commitdiff_plain;h=85210bffd8363e491b4c31f2d09404f9869ad0c7;hp=e9f600f2fb4b0df55c7a8fb4b4d09f9979997223 journal: provide an API that allows client to figure out whether they need to recheck the journal manually for changes in regular intervals Network file systems generally do not offer inotify() that would work across the network. We hence cannot rely on inotify() exclusiely in those case. Provide an API to determine these cases, and suggest doing manual regular rechecks. Note that this is not complete yet, as we need to rescan journal dirs on network file systems explicitly to find new/removed files --- diff --git a/Makefile.am b/Makefile.am index 0bcde9c58..634c67d13 100644 --- a/Makefile.am +++ b/Makefile.am @@ -576,6 +576,7 @@ MANPAGES_ALIAS = \ man/SD_JOURNAL_FOREACH_DATA.3 \ man/sd_journal_get_monotonic_usec.3 \ man/sd_journal_get_cutoff_monotonic_usec.3 \ + man/sd_journal_reliable_fd.3 \ man/sd_journal_process.3 \ man/sd_journal_wait.3 \ man/SD_JOURNAL_NOP.3 \ @@ -649,6 +650,7 @@ man/sd_journal_restart_data.3: man/sd_journal_get_data.3 man/SD_JOURNAL_FOREACH_DATA.3: man/sd_journal_get_data.3 man/sd_journal_get_monotonic_usec.3: man/sd_journal_get_realtime_usec.3 man/sd_journal_get_cutoff_monotonic_usec.3: man/sd_journal_get_cutoff_realtime_usec.3 +man/sd_journal_reliable_fd.3: man/sd_journal_get_fd.3 man/sd_journal_process.3: man/sd_journal_get_fd.3 man/sd_journal_wait.3: man/sd_journal_get_fd.3 man/SD_JOURNAL_NOP.3: man/sd_journal_get_fd.3 diff --git a/man/sd_journal_get_fd.xml b/man/sd_journal_get_fd.xml index e3b5eaa13..189d21352 100644 --- a/man/sd_journal_get_fd.xml +++ b/man/sd_journal_get_fd.xml @@ -44,6 +44,7 @@ sd_journal_get_fd + sd_journal_reliable_fd sd_journal_process sd_journal_wait SD_JOURNAL_NOP @@ -62,6 +63,11 @@ sd_journal* j + + int sd_journal_reliable_fd + sd_journal* j + + int sd_journal_process sd_journal* j @@ -82,12 +88,36 @@ sd_journal_get_fd() returns a file descriptor that may be asynchronously polled in an external event loop and is signaled readable as - soon as the journal changes, for example because new - entries were added. The file descriptor is suitable - for usage in + soon as the journal changes, because new entries or + files were added, rotation took place, or files have + been deleted, and similar. The file descriptor is + suitable for usage in poll2 - where it will yield POLLIN on all changes. The call - takes one argument: the journal context object. + where it will yield POLLIN on changes. The call takes + one argument: the journal context object. Note that + not all file systems are capable of generating the + necessary events for wakeups from this file descriptor + to be enirely reliable. In particular network files + systems do not generate suitable file change events in + all cases. In such a case an application should not + rely alone on wake-ups from this file descriptor but + wake up and recheck the journal in regular time + intervals, for example every 2s. To detect + cases where this is necessary, use + sd_journal_reliable_fd(), + below. + + sd_journal_reliable_fd() + may be used to check whether the wakeup events from + the file descriptor returned by + sd_journal_get_fd are sufficient + to track changes to the journal. If this call returns + 0, it is necessary to regularly recheck for journal + changes (suggestion: every 2s). If this call returns a + positive integer this is not necessary, and wakeups + from the file descriptor returned by + sd_journal_get_fd() are + sufficient as only source for wake-ups. After each POLLIN wake-up sd_journal_process() needs to be @@ -97,14 +127,19 @@ that spurious wake-ups are possible). A synchronous alternative for using - sd_journal_get_fd() and + sd_journal_get_fd(), + sd_journal_reliable_fd() and sd_journal_process() is sd_journal_wait(). It will - synchronously wait until the journal gets changed up - to a certain time-out as specified in its second - argument. Pass (uint64_t) -1 to - wait indefinitely. Internally this call simply - combines sd_journal_get_fd(), + synchronously wait until the journal gets changed, + possibly using a 2s time-out if this is necessary (see + above). In either way the maximum time this call + sleeps may be controlled with the + timeout_usec parameter. Pass + (uint64_t) -1 to wait + indefinitely. Internally this call simply combines + sd_journal_get_fd(), + sd_journal_reliable_fd(), poll() and sd_journal_process() into one. @@ -117,6 +152,15 @@ sd_journal_get_fd() returns a valid file descriptor on success or a negative errno-style error code. + sd_journal_reliable_fd() + returns a positive integer if the file descriptor + returned by sd_journal_get_fd() + is sufficient as sole wake-up source for journal + change events. Returns 0 if it is not sufficient and + the journal needs to be checked manually in regular + time intervals for changes. Returns a negative + errno-style error code on failure. + sd_journal_process() and sd_journal_wait() return one of SD_JOURNAL_NOP, @@ -140,6 +184,7 @@ Notes The sd_journal_get_fd(), + sd_journal_reliable_fd(), sd_journal_process() and sd_journal_wait() interfaces are available as shared library, which can be compiled and @@ -149,6 +194,69 @@ file. + + Examples + + Iterating through the journal, in a live view tracking all changes: + + #include <stdio.h> +#include <string.h> +#include <systemd/sd-journal.h> + +int main(int argc, char *argv[]) { + int r; + sd_journal *j; + r = sd_journal_open(&j, SD_JOURNAL_LOCAL_ONLY); + if (r < 0) { + fprintf(stderr, "Failed to open journal: %s\n", strerror(-r)); + return 1; + } + for (;;) { + const char *d; + size_t l; + r = sd_journal_next(j); + if (r < 0) { + fprintf(stderr, "Failed to iterate to next entry: %s\n", strerror(-r)); + break; + } + if (r == 0) { + /* Reached the end, let's wait for changes, and try again */ + r = sd_journal_wait(j, (uint64_t) -1); + if (r < 0) { + fprintf(stderr, "Failed to wait for changes: %s\n", strerror(-r)); + break; + } + continue; + } + r = sd_journal_get_data(j, "MESSAGE", &d, &l); + if (r < 0) { + fprintf(stderr, "Failed to read message field: %s\n", strerror(-r)); + continue; + } + printf("%.*s\n", (int) l, d); + } + sd_journal_close(j); + return 0; +} + + Waiting with poll() (this + example lacks all error checking for the sake of + simplicity): + + #include <sys/poll.h> +#include <systemd/sd-journal.h> + +int wait_for_changes(sd_journal *j) { + struct pollfd pollfd; + pollfd.fd = sd_journal_get_fd(); + pollfd.events = POLLIN; + poll(&pollfd, 1, sd_journal_reliable_fd() > 0 ? -1 : 2000); + return sd_journal_process(j); +} + + + + See Also @@ -156,7 +264,8 @@ systemd1, sd-journal3, sd_journal_open3, - sd_journal_next3 + sd_journal_next3, + poll2 diff --git a/src/journal/journal-internal.h b/src/journal/journal-internal.h index f68ca996a..75a4129e5 100644 --- a/src/journal/journal-internal.h +++ b/src/journal/journal-internal.h @@ -119,6 +119,8 @@ struct sd_journal { char *unique_field; JournalFile *unique_file; uint64_t unique_offset; + + bool on_network; }; char *journal_make_match_string(sd_journal *j); diff --git a/src/journal/libsystemd-journal.sym b/src/journal/libsystemd-journal.sym index 96293b8a2..ad78fcc74 100644 --- a/src/journal/libsystemd-journal.sym +++ b/src/journal/libsystemd-journal.sym @@ -80,3 +80,8 @@ global: sd_journal_enumerate_unique; sd_journal_restart_unique; } LIBSYSTEMD_JOURNAL_190; + +LIBSYSTEMD_JOURNAL_196 { +global: + sd_journal_fd_reliable; +} LIBSYSTEMD_JOURNAL_195; diff --git a/src/journal/sd-journal.c b/src/journal/sd-journal.c index d5d2d78f1..a346691e2 100644 --- a/src/journal/sd-journal.c +++ b/src/journal/sd-journal.c @@ -25,6 +25,8 @@ #include #include #include +#include +#include #include "sd-journal.h" #include "journal-def.h" @@ -35,9 +37,12 @@ #include "lookup3.h" #include "compress.h" #include "journal-internal.h" +#include "missing.h" #define JOURNAL_FILES_MAX 1024 +#define JOURNAL_FILES_RECHECK_USEC (2 * USEC_PER_SEC) + static void detach_location(sd_journal *j) { Iterator i; JournalFile *f; @@ -1184,6 +1189,25 @@ _public_ int sd_journal_seek_tail(sd_journal *j) { return 0; } +static void check_network(sd_journal *j, int fd) { + struct statfs sfs; + + assert(j); + + if (j->on_network) + return; + + if (fstatfs(fd, &sfs) < 0) + return; + + j->on_network = + sfs.f_type == CIFS_MAGIC_NUMBER || + sfs.f_type == CODA_SUPER_MAGIC || + sfs.f_type == NCP_SUPER_MAGIC || + sfs.f_type == NFS_SUPER_MAGIC || + sfs.f_type == SMB_SUPER_MAGIC; +} + static int add_file(sd_journal *j, const char *prefix, const char *filename) { char *path; int r; @@ -1233,6 +1257,8 @@ static int add_file(sd_journal *j, const char *prefix, const char *filename) { return r; } + check_network(j, f->fd); + j->current_invalidate_counter ++; log_debug("File %s got added.", f->path); @@ -1366,6 +1392,8 @@ static int add_directory(sd_journal *j, const char *prefix, const char *dirname) } } + check_network(j, dirfd(d)); + closedir(d); return 0; @@ -1453,6 +1481,8 @@ static int add_root_directory(sd_journal *j, const char *p) { } } + check_network(j, dirfd(d)); + closedir(d); return 0; @@ -2079,6 +2109,14 @@ _public_ int sd_journal_wait(sd_journal *j, uint64_t timeout_usec) { return determine_change(j); } + if (j->on_network) { + /* If we are on the network we need to regularly check + * for changes manually */ + + if (timeout_usec == (uint64_t) -1 || timeout_usec > JOURNAL_FILES_RECHECK_USEC) + timeout_usec = JOURNAL_FILES_RECHECK_USEC; + } + do { r = fd_wait_for_event(j->inotify_fd, POLLIN, timeout_usec); } while (r == -EINTR); @@ -2344,3 +2382,10 @@ _public_ void sd_journal_restart_unique(sd_journal *j) { j->unique_file = NULL; j->unique_offset = 0; } + +_public_ int sd_journal_reliable_fd(sd_journal *j) { + if (!j) + return -EINVAL; + + return !j->on_network; +} diff --git a/src/shared/missing.h b/src/shared/missing.h index f58a609c7..3777cf69b 100644 --- a/src/shared/missing.h +++ b/src/shared/missing.h @@ -249,3 +249,7 @@ static inline int name_to_handle_at(int fd, const char *name, struct file_handle # error neither secure_getenv nor __secure_getenv are available # endif #endif + +#ifndef CIFS_MAGIC_NUMBER +#define CIFS_MAGIC_NUMBER 0xFF534D42 +#endif diff --git a/src/systemd/sd-journal.h b/src/systemd/sd-journal.h index 55281b134..724117303 100644 --- a/src/systemd/sd-journal.h +++ b/src/systemd/sd-journal.h @@ -124,6 +124,7 @@ int sd_journal_enumerate_unique(sd_journal *j, const void **data, size_t *l); void sd_journal_restart_unique(sd_journal *j); int sd_journal_get_fd(sd_journal *j); +int sd_journal_reliable_fd(sd_journal *j); int sd_journal_process(sd_journal *j); int sd_journal_wait(sd_journal *j, uint64_t timeout_usec);