I've been having problems getting basic power management working on my laptop. Turning SMP support off causes a compilation error, and turning it on disables some of the APIC features (and various documented sources note problems between AP/APIC and SMP enabled kernels). Can gentoo provide a How-To or FAQ note on how to get APM/APIC working under gentoo? Thanks, EBo --
You mean APM/ACPI? Regarding kernel compilation errors, that's a kernelsource bug...
I think John mis-spelled it. gentoo-sources/gs-sources are working fine with most laptops. I'm using gs-sources on my vaio (acpi + firewire + usb2) btw, what source are you using?
ummm... I wrote a reply and it looks like it did not comit... Yes, it was a spelling/typing error (I come from the school for the spelling impaired ;-) The source I was originally using was pre 2.4.20 kernel. The new source mostly works. The current issues are the KDE interface reports the correct % battery left but not the remaining time. Since the low battery and critical low poweroff is triggered by time left and not % left, it does not work for me. It is possible that something is not correctly configured, but the functionality appears to be a bug. The little documentation I have seen on this claims it is a KDE bug. I do not think it is entire so for the following reason: the "present rate" in /proc/acpi/battery/BAT0/state always remains as 0 mW. The one time I tried to track down the problem failed. I will try to break out some time and take another look into this again.
OK. I took part of the morning out to deal with the battery monitor issue... What I found is that the klaptop deamon has the following comments in the code: // ACPI (unlike nice systems like some APM implementations) doesn't // tell us how much battery TIME we have left - here we // do a weighted average of the discharge rate (in mW) and // figure out how long we have left by dividing it into the // remaining capacity // // because some ACPI implementations return bogus // rates when charging we can't estimate the battery life // so we only collect discharge rate data when we're actually // discharging The program goes on to use the rates anyway. On my laptop (an HPN5210), the rate is always set to 0 mW regardless of it charging, discharging, or fully charged. So when the remaining time is calculated [ap->battery_time = (rate==0||part==0?0 : 60*part/rate);] it always evaluates to 0... While the following hack fixes my problem with the klaptopdaemon, it is likely that the real source of the problem is either a bug or configuration problem in ACPI or the kernel. I hacked the following into klaptopdaemon (in kdeutils-3.1.4) to attempt to solve this probloem. Remember this is only a hack. It appears to fuction reasonably well at this point. The following is a diff of portable.cpp 334,408c334 < < // HACK -- The rate of discharge is mucked up so estimate it < // using the remaining charge and dt. Later in the daemon the < // rate is averaged so we will ignore the subtlities for < // now... < if (0 == rate) < { < // FIXME: should go at top on in included .h files. Placed here for < // now to keep all hack changes as local as possible... < #include <time.h> < < static time_t saved_time=0, cur_time; < static int saved_part=part; < static int ave_rate=1, num_ave=0; < static const int MAX_AVE=7; < static int chached_rates[MAX_AVE]; < < // the rate is assumed to be per second or some such... < // So, we need to know the elapse time to calc the rate... < time (&cur_time); < < rate = ((saved_part - part) * 3600)/(cur_time-saved_time); < < // the rate will return a negitive number when charging. < // By forcing it positive, we estimate the duration based < // on (dis)charge rate. < if (0 > rate) rate *= -1; < < if (0 == saved_time) < { < saved_time = cur_time; < saved_part = part; < } < < // Turns out that the resolution of the battery monitoring < // results in the rate being equal to 0 much of the time. < // So I also average the rates here. < if (0 < rate) < { < int i; < < if (MAX_AVE > num_ave) num_ave++; < < // cache the rate and average it. < for (i=num_ave-1; 0<=i; i--) < chached_rates[i] = chached_rates[i-1]; < chached_rates[0] = rate; < < for (i=num_ave-1,ave_rate=0; 0<=i; i--) < ave_rate += chached_rates[i]; < ave_rate /= num_ave; < < < /* reality check -- for debuging */ < if (0) < { < FILE *fout = fopen ("/tmp/klaptop_debug", "a+"); < fprintf (fout, "dt = %d\n",(int)(cur_time-saved_time)); < fprintf (fout, "dp = %d\n",(saved_part - part)); < fprintf (fout, "raw_rt = %d\n",rate); < fprintf (fout, "ave_rt = %d\n",ave_rate); < fprintf (fout, "num_ave = %d\n",num_ave); < fprintf (fout, "\n"); < fclose (fout); < } < < // cache the time and the remaining battery life. < saved_time = cur_time; < saved_part = part; < } < < // use the averaged rate < rate = ave_rate; < } < ---
wow, this is assigned to the kde team? I'm using an nforce1 motherboard... and I'm using genkernel's power management settings... and I don't have any power management... my bios turns off the hard disks when they arn't in use, but my monitor absolutely will not turn off.. when I run apmd it says I don't have apm support in the kernel WTH???
closing as is. if you have further requests, please resubmit a new bug to the docs-team.
