I've submitted this to acpid.sourceforge.net as feature request 1786726. Here's a paste of the description: I'm currently using acpid on an HP Turion ML40 based laptop under gentoo as the acpid-1.0.6 ebuild. This uses the apcid-1.0.6.tar.gz tarball as is without any patches. I'm not sure if this is a quirk of the ML40, but acpid handles a CPU0 processor event on the order of one every second or more. This results in about 3 LOG_INFO level messages going to the syslog, and the current software doesn't allow the logging level to be turned off or tuned. The resulting datastorm rapidly fills both /var/log/syslog and /var/log/messages and prevents any sort of battery mode conservation by spinning down the hard drive. I've hacked my config as follows in order to provide a new q switch for quiet mode that only logs LOG_NOTICE or lower when enabled. I've also changed a couple of logging calls from LOG_INFO to LOG_NOTICE so that they would still appear. Logging runs as before if the switch is not supplied on startup. I actually think that LOG_INFO level reports from events should be removed in favor of having the user do something with the triggers and scripts in /etc/acpid instead, but I guess you wouldn't see what a CLIENT may be up to. Reproducible: Always Steps to Reproduce: 1.Start acpid on a laptop that generates a lot of processor events 2. 3. Actual Results: As I mentioned, I don't know if this is a quirk of my laptop's processor or is a fairly common thing. However your syslog and messages will grow rapidly Expected Results: Being able to use the syslog and messages files as useful logs. Here's the diffs I posted of the two files acpid.c and event.c ---------------------- diffs to acpid.c 58a59 > static int quiet = 0; 161c162,163 < acpid_log(LOG_INFO, "starting up\n"); --- > acpid_log(LOG_NOTICE, "starting up\n"); > if (quiet) acpid_log(LOG_NOTICE, "starting in quiet mode\n"); 272a275 > {"quiet", 0, 0, 'q'}, 297c300 < i = getopt_long(*argc, *argv, "c:de:fg:m:s:Svh", opts, NULL); --- > i = getopt_long(*argc, *argv, "c:de:fg:m:q:s:Svh", opts, NULL); 320a324,326 > case 'q': > quiet = 1; > break; 456c462,463 < vsyslog(level, fmt, args); --- > if ((level < LOG_INFO) || (!quiet)) > vsyslog(level, fmt, args); ------------------ diffs to event.c 157c157 < acpid_log(LOG_INFO, "%d rule%s loaded\n", --- > acpid_log(LOG_NOTICE, "%d rule%s loaded\n", 343c343 < acpid_log(LOG_INFO, "%d client rule%s loaded\n", --- > acpid_log(LOG_NOTICE, "%d client rule%s loaded\n"
(In reply to comment #0) > I've submitted this to acpid.sourceforge.net as feature request 1786726. Yeah, that's where it belongs. Reopen if/once this patch has been accepted. Thanks.
