But from the oops it looks either like potentially like memory corruption, or corruption of the cred. The oops reports
invalid opcode: 0000 [#1] SMP
however the piece of code triggering this is used all the time, so the more likely scenario is that the cred does not have the correct security labeling which would trigger an oops. The oops would not normally be an invalid opcode, but its possible the handling within the oops is triggering the invalid opcode.
The line of code triggering this oops is
BUG_ON(!ctx || !ctx->label);
which going with the bad cred hypothesis means the actual problem is else where. Tracing down the conditions that cause the bad cred may be difficult.
With this being reported as fine in 14.04.5 with all 4.4.x kernels, and without further input to help trace down the conditions that cause this, I am inclined to close this bug as fixed.
In testing I have not been able to reproduce.
But from the oops it looks either like potentially like memory corruption, or corruption of the cred. The oops reports
invalid opcode: 0000 [#1] SMP
however the piece of code triggering this is used all the time, so the more likely scenario is that the cred does not have the correct security labeling which would trigger an oops. The oops would not normally be an invalid opcode, but its possible the handling within the oops is triggering the invalid opcode.
The line of code triggering this oops is
BUG_ON(!ctx || !ctx->label);
which going with the bad cred hypothesis means the actual problem is else where. Tracing down the conditions that cause the bad cred may be difficult.
With this being reported as fine in 14.04.5 with all 4.4.x kernels, and without further input to help trace down the conditions that cause this, I am inclined to close this bug as fixed.