Hi jpeps.jpeps wrote:I tried dog...same problem. I tried before going to bed and then in the morning. Must be a problem with the new kernel01micko wrote:Yeah that looks weird. Any output in the kernel logs? dmesg?jpeps wrote:I'm currious about how slacko uses cache. There seems to be a wide variation, even on a fresh boot. Here's two consecutive boots:Also, after subtracting out buff and cache, I'm already showing more memory used in slacko than lucid 525, and I've got a lot more loaded in lucid.. including devx, google-earth, more apps, etc.Code: Select all
Mem: 1014724K used, 19788K free, 0K shrd, 19496K buff, 908328K cached Mem: 222108K used, 812404K free, 0K shrd, 28940K buff, 132512K cached
Have you tried 3 headed dog? Is it the same?
Code: Select all
dog:Mem: 438108K used, 596404K free, 0K shrd, 47040K buff, 312900K cached dog:Mem: 1019076K used, 15436K free, 0K shrd, 39440K buff, 874324K cached
Could this be hardware specific?
I don't pretend to be any sort of expert on kernel compiling or memory usage. I did compile this kernel and it has been pretty solid for most users for a couple of months. The DOTconfig is where it usually is. It's slightly modified from BK's DOTconfig for k2.6.35.8 and I used Iguleder's script to compile in a full install of Slacko so there are no gccc mismatches, this would be an issue in 3HD.
So, it sure looks like a kernel bug but I think it's in the actual kernel source, not the compile.
I'm very reluctant to change kernels at this late stage, but if it has to be done I'll compile 2.6.39.4 in Slacko, PAE disabled, usb-storage compiled in, or, offer a second version with the later kernel, PAE enabled. Heaps of 3rd party stuff would need to be recompiled.