linux-kernel - Re: [PATCH v2] nd_btt: Make BTT lanes preemptible

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <74efb332edd8012f772fa05ed385f8a9df70b57d.camel@intel.com>
Date:   Fri, 22 Sep 2023 22:22:58 +0000
From:   "Verma, Vishal L" <vishal.l.verma@...el.com>
To:     "tglozar@...hat.com" <tglozar@...hat.com>,
        "nvdimm@...ts.linux.dev" <nvdimm@...ts.linux.dev>
CC:     "Williams, Dan J" <dan.j.williams@...el.com>,
        "Jiang, Dave" <dave.jiang@...el.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "Weiny, Ira" <ira.weiny@...el.com>
Subject: Re: [PATCH v2] nd_btt: Make BTT lanes preemptible

On Wed, 2023-09-20 at 07:37 +0200, Tomas Glozar wrote:
> 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.
> 
> 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.
> 
> BUG example occurring when running ndctl tests on PREEMPT_RT kernel:
> 
> BUG: sleeping function called from invalid context at
> kernel/locking/spinlock_rt.c:48
> in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 4903, name:
> libndctl
> preempt_count: 1, expected: 0
> RCU nest depth: 0, expected: 0
> Preemption disabled at:
> [<ffffffffc1313db5>] nd_region_acquire_lane+0x15/0x90 [libnvdimm]
> Call Trace:
>  <TASK>
>  dump_stack_lvl+0x8e/0xb0
>  __might_resched+0x19b/0x250
>  rt_spin_lock+0x4c/0x100
>  ? btt_write_pg+0x2d7/0x500 [nd_btt]
>  btt_write_pg+0x2d7/0x500 [nd_btt]
>  ? local_clock_noinstr+0x9/0xc0
>  btt_submit_bio+0x16d/0x270 [nd_btt]
>  __submit_bio+0x48/0x80
>  __submit_bio_noacct+0x7e/0x1e0
>  submit_bio_wait+0x58/0xb0
>  __blkdev_direct_IO_simple+0x107/0x240
>  ? inode_set_ctime_current+0x51/0x110
>  ? __pfx_submit_bio_wait_endio+0x10/0x10
>  blkdev_write_iter+0x1d8/0x290
>  vfs_write+0x237/0x330
>  ...
>  </TASK>
> 
> Fixes: 5212e11fde4d ("nd_btt: atomic sector updates")
> Signed-off-by: Tomas Glozar <tglozar@...hat.com>

Thanks for the update to include the failure signature.
I read the commentary around migrate_disable(), and it makes sense to
use it as a replacement for disabling preemption here. I suppose at
some point there will be a move to remove migrate_disable(), and at
that point we'd have to rework this code to remove the implicit
lockless mode when operating under per-cpu assumptions. Until then,

Reviewed-by: Vishal Verma <vishal.l.verma@...el.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);
>