pmcputemp crashes JWM in Bionic 32 (SOLVED)

[amethyst]

I have this random problem using Bionic 32. Everything boots fine and pmcputemp works correctly. Then sometime during a session a warning message for pmcputemp pops up, the pmcputemp icon is still in the tray but there's a new gap between the items in the tray but no new icon. This is a random issue and seems to happen more readily after some heavy use. Closing the warning message, kills pmcputemp and causes JWM to crash, icons change to gibberish, can't start an application from menu, can't reboot or restart jwm or X. Basically frozen, have to switch off machine. Is this a known issue?

[Phoenix]

If you open a terminal and attempt executing a executable that is not cached, does it have an I/O error? Also try ctrl+alt+backspace, the terminal might have some kernel errors.

[cobaka]

@amethyst
Below you see a screen-shot from the Task Manager. (Available in uPupBB32 at the top of my monitor screen.)
I cannot understand why an application that monitors the temperature sensor in my system needs/uses/reports 616MiB of virtual memory.
Clearly I don't understand something - but what?

Cobaka

pmcputemp.jpg (41.49 KiB) Viewed 604 times

[JakeSFR]

I cannot understand why an application that monitors the temperature sensor in my system needs/uses/reports 616MiB of virtual memory.
Clearly I don't understand something - but what?

The explanation is simple - it leaks memory, most likely.
I just checked and in Fatdog we have a patch that fixes two memleaks and I just found another one.
Here's unified patch:

Code: Select all

diff -ur pmcputemp-082016_old/pmcputemp.c pmcputemp-082016_new/pmcputemp.c
--- pmcputemp-082016_old/pmcputemp.c	2015-05-06 08:48:00.000000000 +0200
+++ pmcputemp-082016_new/pmcputemp.c	2022-02-07 13:28:58.250017485 +0100
@@ -234,6 +234,8 @@
 	cairo_set_source_rgb(c, fr, fg, fb);
 	cairo_show_text(c, deg);
 	cairo_surface_write_to_png (cs, temp_icon);
+	cairo_pattern_destroy(linear);
+	cairo_surface_destroy(cs);
 	cairo_destroy(c);
 	return temp;
 }
@@ -308,6 +310,7 @@
 #endif /* HAVE_GTK3 */
 	temp_pixbuf = gdk_pixbuf_new_from_file(temp_icon,&gerror);
 	gtk_status_icon_set_from_pixbuf(tray_icon,temp_pixbuf);
+	g_object_unref(temp_pixbuf);
 	g_free(tooltip_out);
 	return 1;
 }

EDIT: all these memleaks are already fixed upstream https://github.com/01micko/pmcputemp, as of version 0.71. Bionicpup 8.0 comes with 0.63.

Greetings!

[amethyst]

Guys, I'm testing this. I just wonder maybe if this is not kernel related somehow. I was using an older kernel because the newer kernels have some hang time on this machine at bootup. I've been using the original kernel since this morning and haven't had issues yet. Too early to report so it's a work in progress. I'm going to disable the thing if the problem persists. BTW - I haven't had issues with this application with other Puppys. Could it be a flaw with this specific version?

[amethyst]

@amethyst
Below you see a screen-shot from the Task Manager. (Available in uPupBB32 at the top of my monitor screen.)
I cannot understand why an application that monitors the temperature sensor in my system needs/uses/reports 616MiB of virtual memory.
Clearly I don't understand something - but what?

Cobaka
pmcputemp.jpg

Interesting. I see mine is about 18MB right now but worthwhile to keep a look at it. Thanks. What is your uptime? Actually - this won't be a problem for me because I have a script that clears the vm-caches every 20 minutes.

[cobaka]

@amethyst
Your Q: what is my up-time?
PupSysinfo (one of the most useful 'thingies' in Puppy) says: Pup-SysInfo Hardware Report (Summary), Tue 8 Feb 2022
Current Time: 07:36:48 System Uptime: 6d 16h 34m Load Average: 0.38 0.36 0.19 Processes: 228 total, 2 running

<Now I re-boot. Pls wait ... OK - rebooted>
Task Manager report is: memory used by pmCPUtemp is below (see image). Total memory 8075 MiB available, 401 MiB used (fr. bar across the top of Task Manager - next to CPU use.)

Now, for your pleasure and enjoyment (below) I present a fragment of the Task Manager dealing with cpu temperature.

Task_Manager+cpu_temp.jpg (13.83 KiB) Viewed 577 times

(Later) Having eaten breakfast I note that memory usage has increased to 19%.
Looking at "top" in the terminal: Cannot see pmcputemp.

Cobaka

My puppy is: Kernel: 4.9.163-lxpup-32-pae (i686), Version: #1 SMP Thu Mar 14 15:41:19 GMT 2019
Build GCC: 8.3.0 PAE Enabled: Yes

Kernel Command Line:
pdrv=029012e5-8ee3-4fc0-93b2-03af0bee2d42 psubdir=/OS_BOOT pmedia=atahd pfix=fsck

Distro: BionicPup32 19.03
Window Manager: JWM v2.3.7
Desktop Start: xwin jwm

[amethyst]

@cobaka

You should reboot more often! Anyways - I wouldn't be too worried about those memory readings. I've just rebooted and started running Palemoon and Java (JRE). Both are reporting VM-sizes of more than 700MB but the actual memory usage of the whole system is only 350MB.

[amethyst]

Back to the topic. I still have this issue. Difficult to pinpoint as it comes and goes. Don't know what is triggering this sometimes. Have now just disabled the bloody thing.

[bigpup]

all these memleaks in pmcputemp are already fixed upstream https://github.com/01micko/pmcputemp, as of version 0.71. Bionicpup 8.0 comes with 0.63.

Did you try and replace with this newer version of pmcputemp?

[wizard]

Is there a pet of version 0.71?

Thanks
wizard

[amethyst]

I think I've managed to fix this but since other users do not seem to have encountered the same issue, I'm not going to explain my solution here since it may lead to confusion for some.

[ozsouth]

@wizard - here is another option: viewtopic.php?p=4697#p4697

[amethyst]

@wizard - here is another option: viewtopic.php?p=4697#p4697

Thanks, I'll try it if I encounter a problem again.

[amethyst]

Now that I've fixed this problem, I was thinking about the usefulness of this application. The point is, I know my laptop generally operates at 60 degrees and this goes up to 80 degrees (never shows any higher) if the machine is working really hard. So what's the point, apart from having some eye-candy in the tray? I've disabled it.

[Feek]

Now that I've fixed this problem, I was thinking about the usefulness of this application. The point is, I know my laptop generally operates at 60 degrees and this goes up to 80 degrees (never shows any higher) if the machine is working really hard. So what's the point, apart from having some eye-candy in the tray? I've disabled it.

I admit that I haven't dealt with CPU temperature monitoring much before.
Later, I casually began to watch.
Your decision (to disable it) led me to ask whether it was important or not.

Searching the internet brought me some information:
"If you have a new computer or laptop, you don't need much to know the temperature of the processor, graphics card and other components.
However, with older computers, it may be quite appropriate to know how you're doing - dust, less functional thermal paste on the processor, maybe even problems with fans can lead to rising temperatures."

My laptop is over 10 years old, so nothing new.
If I understand correctly, temperature monitoring can alert me to a problem in time .

[cobaka]

@Feek & others

The temperature of the CPU "floats" on ambient.
Here where I live temperatures in winter are around 16 to 22C.
In summer we have days where the temperature will often tip +36C and on rare days hit +40 (or more).
In summer I watch the CPU temperature closely. The life of electronic equipment falls considerably for every +10C rise in temperature.

Cobaka

[amethyst]

Now that I've fixed this problem, I was thinking about the usefulness of this application. The point is, I know my laptop generally operates at 60 degrees and this goes up to 80 degrees (never shows any higher) if the machine is working really hard. So what's the point, apart from having some eye-candy in the tray? I've disabled it.

I admit that I haven't dealt with CPU temperature monitoring much before.
Later, I casually began to watch.
Your decision (to disable it) led me to ask whether it was important or not.

Searching the internet brought me some information:
"If you have a new computer or laptop, you don't need much to know the temperature of the processor, graphics card and other components.
However, with older computers, it may be quite appropriate to know how you're doing - dust, less functional thermal paste on the processor, maybe even problems with fans can lead to rising temperatures."

My laptop is over 10 years old, so nothing new.
If I understand correctly, temperature monitoring can alert me to a problem in time .

My laptop is 15 years old. As I say, the temp reading is generally 60 degrees during normal use and goes up to 80 when the machine really works hard. You can actually hear how hard the machine is working but the temp reading will stay at 80 degrees so that really does not say much apart from that your machine is now working harder. My laptop switches off when there's overload (this happens very seldomly). Now, if the utility was so sophisticated that it could tell me that I'm now in a danger zone and that the machine will switch off in say a minute at the current rate, then it may be useful but otherwise....

[wizard]

temperature monitoring can alert me to a problem in time

I always include a temp monitor, particularly for laptops that have their fan intake on the bottom. Users will place them mmmm on their laps, beds, carpet, furniture blocking the intake. Then you have problems. Years ago I was given an old Gateway laptop that the cpu had gotten so hot it was distorting the keyboard keys above it. More than once they have reminded me it was time to clean the heatsink and fan.

Thanks
wizard