am335x: 5.18.x: system stalling

[Yegor Yefremov]

On Wed, Jun 1, 2022 at 12:06 PM Ard Biesheuvel <ardb at kernel.org> wrote:
>
> On Wed, 1 Jun 2022 at 12:04, Yegor Yefremov <yegorslists at googlemail.com> wrote:
> >
> > On Wed, Jun 1, 2022 at 11:28 AM Ard Biesheuvel <ardb at kernel.org> wrote:
> > >
> > > On Wed, 1 Jun 2022 at 10:08, Ard Biesheuvel <ardb at kernel.org> wrote:
> > > >
> > > > On Wed, 1 Jun 2022 at 09:59, Arnd Bergmann <arnd at arndb.de> wrote:
> > > > >
> > > > > On Wed, Jun 1, 2022 at 9:36 AM Yegor Yefremov
> > > > > <yegorslists at googlemail.com> wrote:
> > > > > > On Tue, May 31, 2022 at 5:23 PM Arnd Bergmann <arnd at arndb.de> wrote:
> > > > > > > I've pushed a modified branch now, with that fix on the broken commit,
> > > > > > > and another change to make CONFIG_IRQSTACKS user-selectable rather
> > > > > > > than always enabled. That should tell us if the problem is in the SMP
> > > > > > > patching or in the irqstacks.
> > > > > > >
> > > > > > > Can you test the top of this branch with CONFIG_IRQSTACKS disabled,
> > > > > > > and (if that still stalls) retest the fixed commit f0191ea5c2e5 ("[PART 1]
> > > > > > > ARM: implement THREAD_INFO_IN_TASK for uniprocessor systems")?
> > > > > >
> > > > > > 1. the top of this branch with CONFIG_IRQSTACKS disabled stalls
> > > > > > 2. f0191ea5c2e5 with the same config - not
> > > > >
> > > > > Ok, perfect, that does narrow down the problem quite a bit: The final
> > > > > patch has seven changes, all of which can be done individually because
> > > > > in each case the simplified version in f0191ea5c2e5 is meant to run
> > > > > the exact same instructions as the version after the change, when running
> > > > > on a uniprocessor machine such as your am335x.
> > > > >
> > > > > You have already shown earlier that the get_current() and
> > > > > __my_cpu_offset() functions are not to blame here, as reverting
> > > > > only those does not change the behavior.
> > > > >
> > > > > This leaves the is_smp() check in set_current(), and the
> > > > > four macros in <asm/assembler.h>. I don't see anything obviously
> > > > > wrong with any of those five, but I would bet on the macros
> > > > > here. Can you try bisecting into this commit, maybe reverting
> > > > > the changes to set_current and get_current first, and then
> > > > > narrowing it down to (hopefully) a single macro that causes the
> > > > > problem?
> > > > >
> > > >
> > > > set_current() is never called by the primary CPU, which is why the
> > > > is_smp() check was removed from there in 57a420435edcb0b94 ("ARM: drop
> > > > pointless SMP check on secondary startup path").
> > > >
> > > > So that leaves only the four macros in asm/assembler.h, but I don't
> > > > see anything obviously wrong with those either.
> > >
> > > I pushed a patch on top of Arnd's branch at the link below that gets
> > > rid of the subsections, and uses normal branches (and code patching)
> > > to switch between the thread ID register and the LDR to retrieve the
> > > CPU offset and the current pointer. I have no explanation whether or
> > > why it could make a difference, but I think it's worth a try.
> >
> > The link to your repo is missing.
> >
>
> Oops, sorry :-)
>
> https://git.kernel.org/pub/scm/linux/kernel/git/ardb/linux.git/log/?h=am335x-stall-test

I have tested your branch and it stalls:

[   69.924298] rcu: INFO: rcu_sched self-detected stall on CPU
[   69.930986] rcu:     0-...!: (2600 ticks this GP)
idle=6f5/1/0x40000004 softirq=2257/2257 fqs=0
[   69.940551]  (t=2600 jiffies g=3413 q=11)
[   69.945187] rcu: rcu_sched kthread timer wakeup didn't happen for
2599 jiffies! g3413 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402
[   69.957111] rcu:     Possible timer handling issue on cpu=0
timer-softirq=1261
[   69.964668] rcu: rcu_sched kthread starved for 2600 jiffies! g3413
f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=0
[   69.975638] rcu:     Unless rcu_sched kthread gets sufficient CPU
time, OOM is now expected behavior.
[   69.985170] rcu: RCU grace-period kthread stack dump:
[   69.990708] task:rcu_sched       state:I stack:    0 pid:   10
ppid:     2 flags:0x00000000
[   70.000250] [<c0b683b4>] (__schedule) from [<c0b68cf8>] (schedule+0x54/0xe8)
[   70.008705] [<c0b68cf8>] (schedule) from [<c0b6f4fc>]
(schedule_timeout+0xa8/0x210)
[   70.017449] [<c0b6f4fc>] (schedule_timeout) from [<c01d8594>]
(rcu_gp_fqs_loop+0x118/0x6b4)
[   70.026875] [<c01d8594>] (rcu_gp_fqs_loop) from [<c01dc4c4>]
(rcu_gp_kthread+0x138/0x30c)
[   70.036074] [<c01dc4c4>] (rcu_gp_kthread) from [<c0164dd8>]
(kthread+0x13c/0x164)
[   70.044559] [<c0164dd8>] (kthread) from [<c0100150>]
(ret_from_fork+0x14/0x44)
[   70.052732] rcu: Stack dump where RCU GP kthread last ran:
[   70.058773] NMI backtrace for cpu 0
[   70.062840] CPU: 0 PID: 5 Comm: kworker/0:0 Not tainted 5.16.0-rc1 #1
[   70.070003] Hardware name: Generic AM33XX (Flattened Device Tree)
[   70.076698] Workqueue: events dbs_work_handler
[   70.082258] [<c01115f0>] (unwind_backtrace) from [<c010bfd4>]
(show_stack+0x10/0x14)
[   70.091113] [<c010bfd4>] (show_stack) from [<d00299f0>] (0xd00299f0)
[   70.099045] NMI backtrace for cpu 0
[   70.103188] CPU: 0 PID: 5 Comm: kworker/0:0 Not tainted 5.16.0-rc1 #1
[   70.110357] Hardware name: Generic AM33XX (Flattened Device Tree)
[   70.117027] Workqueue: events dbs_work_handler
[   70.122491] [<c01115f0>] (unwind_backtrace) from [<c010bfd4>]
(show_stack+0x10/0x14)
[   70.131254] [<c010bfd4>] (show_stack) from [<d00299f0>] (0xd00299f0)