[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <65036a57ea900_35db10294ec@iweiny-mobl.notmuch>
Date: Thu, 14 Sep 2023 13:17:27 -0700
From: Ira Weiny <ira.weiny@...el.com>
To: Tomáš Glozar <tglozar@...il.com>,
<nvdimm@...ts.linux.dev>
CC: <dan.j.williams@...el.com>, <vishal.l.verma@...el.com>,
<dave.jiang@...el.com>, <ira.weiny@...el.com>,
<linux-kernel@...r.kernel.org>, Tomas Glozar <tglozar@...hat.com>
Subject: Re: [PATCH] nd_btt: Make BTT lanes preemptible
Tomáš Glozar wrote:
> From: Tomas Glozar <tglozar@...hat.com>
>
> nd_region_acquire_lane uses get_cpu, which disables preemption. This is
> an issue on PREEMPT_RT kernels, since btt_write_pg and also
> nd_region_acquire_lane itself take a spin lock, resulting in BUG:
> sleeping function called from invalid context.
Is the bug in 1 of 2 places?
1) When btt_write_pg()->lock_map() (when the number of lanes is < number
of cpus) and the lane is acquired is called?
*or*
2) When nd_region_acquire_lane() internally trys to take it's lock?
A copy/paste of the BUG observed would have been more clear I think.
Regardless I *think* this is ok but I'm worried I don't fully understand
what the problem is.
Ira
>
> Fix the issue by replacing get_cpu with smp_process_id and
> migrate_disable when needed. This makes BTT operations preemptible, thus
> permitting the use of spin_lock.
>
> Fixes: 5212e11fde4d ("nd_btt: atomic sector updates")
> Signed-off-by: Tomas Glozar <tglozar@...hat.com>
> ---
> drivers/nvdimm/region_devs.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/nvdimm/region_devs.c b/drivers/nvdimm/region_devs.c
> index 0a81f87f6f6c..e2f1fb99707f 100644
> --- a/drivers/nvdimm/region_devs.c
> +++ b/drivers/nvdimm/region_devs.c
> @@ -939,7 +939,8 @@ unsigned int nd_region_acquire_lane(struct nd_region *nd_region)
> {
> unsigned int cpu, lane;
>
> - cpu = get_cpu();
> + migrate_disable();
> + cpu = smp_processor_id();
> if (nd_region->num_lanes < nr_cpu_ids) {
> struct nd_percpu_lane *ndl_lock, *ndl_count;
>
> @@ -958,16 +959,15 @@ EXPORT_SYMBOL(nd_region_acquire_lane);
> void nd_region_release_lane(struct nd_region *nd_region, unsigned int lane)
> {
> if (nd_region->num_lanes < nr_cpu_ids) {
> - unsigned int cpu = get_cpu();
> + unsigned int cpu = smp_processor_id();
> struct nd_percpu_lane *ndl_lock, *ndl_count;
>
> ndl_count = per_cpu_ptr(nd_region->lane, cpu);
> ndl_lock = per_cpu_ptr(nd_region->lane, lane);
> if (--ndl_count->count == 0)
> spin_unlock(&ndl_lock->lock);
> - put_cpu();
> }
> - put_cpu();
> + migrate_enable();
> }
> EXPORT_SYMBOL(nd_region_release_lane);
>
> --
> 2.39.3
>
Powered by blists - more mailing lists