In the Linux kernel, the following vulnerability has been resolved:
block: don't allow the same type rq_qos add more than once
In our test of iocost, we encountered some list add/del corruptions of innerwalk list in ioctimer_fn.
The reason can be described as follows:
cpu 0 cpu 1 iocqoswrite iocqoswrite
ioc = qtoioc(queue); if (!ioc) { ioc = kzalloc(); ioc = qtoioc(queue); if (!ioc) { ioc = kzalloc(); ... rqqosadd(q, rqos); } ... rqqosadd(q, rqos); ... }
When the io.cost.qos file is written by two cpus concurrently, rqqos may be added to one disk twice. In that case, there will be two iocs enabled and running on one disk. They own different iocgs on their active list. In the ioctimerfn function, because of the iocgs from two iocs have the same root iocg, the root iocg's walklist may be overwritten by each other and this leads to list add/del corruptions in building or destroying the inner_walk list.
And so far, the blk-rq-qos framework works in case that one instance for one type rq_qos per queue by default. This patch make this explicit and also fix the crash above.