> >- another time, suspension worked but it took a long time in the > >"waiting phase" between writing processes to disk and writing the > > I have two hypotheses for that: > > 1.) IIRC the kernel (pmdisk) said something about 50'000+ blocks or > pages being written to disk (usually it's just 10'000+). If this > number is about pages (4kb, right?), this would be about 200MB, which
Right. > nears the swap limit (I'm not sure if this run was with the 2nd swap > already prioritized higher). So.. what's the symptom of swap > collisions? The suspend should just abort. > 2.) is pmdisk ~high memory clean? I have 512MB of RAM here, IIRC I > had to enable some kernel option for this (but CONFIG_HIGHMEM is not > set; dunno remember which it was, would have to look deeper). No idea really. If it wasn't it would always fail, I suspect. Freeing memory is done by using the regular swap interface, that should take care of most of the memory contents anyway. Michael