mlx4_core 0000:07:00.0: swiotlb buffer is full and OOM observed during stress test on reset_controller

[Sagi Grimberg]

> Hi experts
>
> I reproduced this issue during stress test on reset_controller, could you help check it, thanks.
>
> Reproduce steps on initiator side:
> num=0
> while [ 1 ]
> do
> 	echo "-------------------------------$num"
> 	echo 1 >/sys/block/nvme0n1/device/reset_controller || exit 1
> 	((num++))
> done
>
> Here is the full log:
> http://pastebin.com/mek9fb0b

I'm using CX5-LX device and have not seen any issues with it.

Would it be possible to retest with kmemleak?