Few Tests - abort - WARNING: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected
<7> [111.707392] i915 0000:00:02.0: [drm:intel_modeset_setup_hw_state [i915]] [PLANE:42:cursor A] min_cdclk 0 kHz
<7> [111.707631] i915 0000:00:02.0: [drm:intel_bw_crtc_update [i915]] pipe A data rate 0 num active planes 0
<4> [111.707915]
<4> [111.707920] =====================================================
<4> [111.707923] WARNING: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected
<4> [111.707925] 6.13.0-rc1-CI_DRM_15771-g26715037f44d+ #1 Not tainted
<4> [111.707928] -----------------------------------------------------
<4> [111.707930] kworker/u32:19/2627 [HC0[0]:SC0[2]:HE1:SE0] is trying to acquire:
<4> [111.707933] ffffffff836af018 (napi_hash_lock){+.+.}-{3:3}, at: napi_hash_add+0x30/0xf0
<4> [111.707943]
and this task is already holding:
<4> [111.707945] ffff88811c510da0 (&tp->lock){+.-.}-{3:3}, at: tg3_chip_reset+0x17b/0xa20 [tg3]
<4> [111.707956] which would create a new lock dependency:
<4> [111.707958] (&tp->lock){+.-.}-{3:3} -> (napi_hash_lock){+.+.}-{3:3}
<4> [111.707964]
but this new dependency connects a SOFTIRQ-irq-safe lock:
<4> [111.707966] (&tp->lock){+.-.}-{3:3}
<4> [111.707968]
... which became SOFTIRQ-irq-safe at:
<4> [111.707972] lock_acquire+0xc9/0x300
<4> [111.707977] _raw_spin_lock+0x2f/0x60
<7> [111.707869] i915 0000:00:02.0: [drm:intel_modeset_setup_hw_state [i915]] [PLANE:48:primary B] min_cdclk 0 kHz
<4> [111.707982] tg3_timer+0x37/0xdb0 [tg3]
<4> [111.707987] call_timer_fn+0xab/0x2a0
<4> [111.707989] __run_timers+0x266/0x350
<4> [111.707991] run_timer_softirq+0x8a/0x100
<4> [111.707992] handle_softirqs+0xd7/0x4d0
<4> [111.707994] __irq_exit_rcu+0x13f/0x160
<4> [111.707996] irq_exit_rcu+0xe/0x20
<4> [111.707998] sysvec_apic_timer_interrupt+0xa0/0xc0
<4> [111.708001] asm_sysvec_apic_timer_interrupt+0x1b/0x20
<4> [111.708005] cpuidle_enter_state+0x130/0x890
<4> [111.708007] cpuidle_enter+0x2e/0x50
<4> [111.708011] call_cpuidle+0x22/0x60
<4> [111.708013] do_idle+0x1ff/0x260
<4> [111.708015] cpu_startup_entry+0x29/0x30
<4> [111.708018] start_secondary+0x12e/0x160
<4> [111.708021] common_startup_64+0x13e/0x141
<4> [111.708024]
to a SOFTIRQ-irq-unsafe lock:
<4> [111.708025] (napi_hash_lock){+.+.}-{3:3}
<4> [111.708027]
... which became SOFTIRQ-irq-unsafe at:
<4> [111.708028] ...
<4> [111.708028] lock_acquire+0xc9/0x300
<4> [111.708031] _raw_spin_lock+0x2f/0x60
<4> [111.708034] napi_hash_add+0x30/0xf0
<4> [111.708036] napi_enable+0x104/0x190
<4> [111.708038] tg3_napi_enable+0x60/0x120 [tg3]
<4> [111.708043] tg3_start+0x60a/0xea0 [tg3]
<4> [111.708048] tg3_open+0x130/0x2d0 [tg3]
<4> [111.708053] __dev_open+0x10e/0x1e0
<4> [111.708055] __dev_change_flags+0x1fd/0x290
<4> [111.708057] dev_change_flags+0x27/0x80
<4> [111.708060] do_setlink.isra.0+0x420/0x1320
<4> [111.708063] rtnl_setlink+0x262/0x4b0
<4> [111.708065] rtnetlink_rcv_msg+0x450/0x6e0
<4> [111.708067] netlink_rcv_skb+0x5d/0x110
<4> [111.708070] rtnetlink_rcv+0x15/0x30
<4> [111.708071] netlink_unicast+0x1b2/0x2a0
<4> [111.708073] netlink_sendmsg+0x233/0x490
<4> [111.708075] __sys_sendto+0x242/0x250
<4> [111.708078] __x64_sys_sendto+0x24/0x40
<4> [111.708080] x64_sys_call+0x2540/0x2650
<4> [111.708083] do_syscall_64+0x91/0x180
<4> [111.708086] entry_SYSCALL_64_after_hwframe+0x76/0x7e
<4> [111.708088]
other info that might help us debug this:
<4> [111.708089] Possible interrupt unsafe locking scenario: