Re: [PATCH v2 05/12] rcutorture: Add tests for SRCU up/down reader primitives

[Joel Fernandes]

On 3/31/2025 5:03 PM, Paul E. McKenney wrote:
> This commit adds a new rcutorture.n_up_down kernel boot parameter
> that specifies the number of outstanding SRCU up/down readers, which
> begin in kthread context and end in an hrtimer handler.  There is a new
> kthread ("rcu_torture_updown") that scans an per-reader array looking
> for elements whose readers have ended.  This kthread sleeps between one
> and two milliseconds between consecutive scans.
> 
> [ paulmck: Apply kernel test robot feedback. ]
> [ paulmck: Apply Z qiang feedback. ]
> 
> Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>

For completeness, posting our discussion for the archives, an issue exists in
this patch causing the following errors on an ARM64 machine with 288 CPUs:

When running SRCU-P test, we intermittently see:

[ 9500.806108] ??? Writer stall state RTWS_SYNC(21) g18446744073709551218 f0x0
->state 0x2 cpu 4
[ 9515.833356] ??? Writer stall state RTWS_SYNC(21) g18446744073709551218 f0x0
->state 0x2 cpu 4

It bisected to just this patch.

thanks,

 - Joel