hdparm /dev/hdc showed up use_dma, but it didn't with the -i flag. Unfortunately, I found that out after using hdparm -d1 /dev/hdc so I don't know if it was on beforehand. I'd say it was, as with -i it spat out 'DMA modes: mdma0 mdma1 mdma2'.
Thanks for the suggestion. Alas, the problem persists.
[edit] There seems to be a memory leak.
Code:
Before copy
Mem: 904496k total, 268288k used, 636208k free, 39344k buffers
Swap: 128484k total, 0k used, 128484k free, 121588k cached
Copied 95Mb
Mem: 904496k total, 570116k used, 334380k free, 60496k buffers
Swap: 128484k total, 0k used, 128484k free, 328704k cached
Copied 78Mb
Mem: 904496k total, 761868k used, 142628k free, 60696k buffers
Swap: 128484k total, 0k used, 128484k free, 487904k cached
Copied 4.0Gb
Mem: 904496k total, 897872k used, 6624k free, 33744k buffers
Swap: 128484k total, 0k used, 128484k free, 610640k cached
The mouse movement returns to normal and some of the RAM is regained if I delete the file afterward, but that's not very helpful. :\
Code:
Deleted 4.0Gb
Mem: 904496k total, 616044k used, 288452k free, 34192k buffers
Swap: 128484k total, 0k used, 128484k free, 326064k cached
Hmm... it also seems the same thing is happenening for both hard disks. I guess I hadn't tried copying to the first one before.