On Tue, 2024-10-15 at 14:42 +0200, Greg Kroah-Hartman wrote:
> 5.10-stable review patch. If anyone has any objections, please let
> me know.
This patch is buggy and must not be cherry-picked without also having:
a09c17240bd ("io_uring/sqpoll: retain test for whether the CPU is
valid")
7f44beadcc1 ("io_uring/sqpoll: do not put cpumask on stack")
Best regards,
Felix Moessbauer
>
> ------------------
>
> From: Felix Moessbauer <felix.moessbauer@xxxxxxxxxxx>
>
> The submit queue polling threads are userland threads that just never
> exit to the userland. When creating the thread with
> IORING_SETUP_SQ_AFF,
> the affinity of the poller thread is set to the cpu specified in
> sq_thread_cpu. However, this CPU can be outside of the cpuset defined
> by the cgroup cpuset controller. This violates the rules defined by
> the
> cpuset controller and is a potential issue for realtime applications.
>
> In b7ed6d8ffd6 we fixed the default affinity of the poller thread, in
> case no explicit pinning is required by inheriting the one of the
> creating task. In case of explicit pinning, the check is more
> complicated, as also a cpu outside of the parent cpumask is allowed.
> We implemented this by using cpuset_cpus_allowed (that has support
> for
> cgroup cpusets) and testing if the requested cpu is in the set.
>
> Fixes: 37d1e2e3642e ("io_uring: move SQPOLL thread io-wq forked
> worker")
> Cc: stable@xxxxxxxxxxxxxxx # 6.1+
> Signed-off-by: Felix Moessbauer <felix.moessbauer@xxxxxxxxxxx>
> Link:
> https://lore.kernel.org/r/20240909150036.55921-1-felix.moessbauer@xxxxxxxxxxx
> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> ---
> io_uring/io_uring.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> --- a/io_uring/io_uring.c
> +++ b/io_uring/io_uring.c
> @@ -56,6 +56,7 @@
> #include <linux/mm.h>
> #include <linux/mman.h>
> #include <linux/percpu.h>
> +#include <linux/cpuset.h>
> #include <linux/slab.h>
> #include <linux/blkdev.h>
> #include <linux/bvec.h>
> @@ -8571,10 +8572,12 @@ static int io_sq_offload_create(struct i
> return 0;
>
> if (p->flags & IORING_SETUP_SQ_AFF) {
> + struct cpumask allowed_mask;
> int cpu = p->sq_thread_cpu;
>
> ret = -EINVAL;
> - if (cpu >= nr_cpu_ids || !cpu_online(cpu))
> + cpuset_cpus_allowed(current, &allowed_mask);
> + if (!cpumask_test_cpu(cpu, &allowed_mask))
> goto err_sqpoll;
> sqd->sq_cpu = cpu;
> } else {
>
>
--
Siemens AG, Technology
Linux Expert Center
