Jetson TK1 Pro X locks up with gk20a_fifo_handle_mmu_fault

Hello, world.

I’m trying to do some development with a TK1 Pro. I’m running X on it and trying to do my development from there, with a bit of help from distcc for cross-compilation. Several times a day, the X server locks up, leaving a console message like this:

[ 6628.569156] gk20a gk20a.0: gk20a_fifo_handle_mmu_fault: mmu fault on engine 0
, engine subid 1 (hub), client 4 (fe), addr 0x00000000:0x00000000, type 0 (pde),
info 0x000004c0,inst_ptr 0xf3972000
[ 6628.569156]
[ 6628.587855] gk20a gk20a.0: gk20a_fifo_set_ctx_mmu_error: channel 1 generated
a mmu fault
[ 6628.595955] gk20a gk20a.0: gk20a_set_error_notifier: error notifier set to 31
[ 6628.595955]
[ 6628.614308] gk20a gk20a.0: gk20a_fifo_handle_dropped_mmu_fault: dropped mmu f
ault (0x00000000)
[ 6628.623026] gk20a gk20a.0: fifo_error_isr: channel reset initated from fifo_e

I’m running vibrante linux, which wants to identify itself as vibrante-vcm30t124-linux.

Any ideas what I can do here? Is this a software thing (I see a few mentions of something similar) or is it time to replace my development machine?

–Harold Z.