Re: linux-master: WARNING: suspicious RCU usage in mem_allocator_disconnect

[Grygorii Strashko]

On 03/12/2019 12:28, Grygorii Strashko wrote:
> Hi All,
>
> While placing intf down I'm getting below splat with debug options enabled.
> Not sure how to fix it, so will be appreciated for any help.\

And it seems introduced by commit:

commit c3f812cea0d7006469d1cf33a4a9f0a12bb4b3a3
Author: Jonathan Lemon <jonathan.lemon@xxxxxxxxx>
Date:   Thu Nov 14 14:13:00 2019 -0800

    page_pool: do not release pool until inflight == 0.


> =========================================================
> [Â 333.933896]
> [Â 333.935511] =============================
> [Â 333.939552] WARNING: suspicious RCU usage
> [Â 333.943724] 5.4.0-08849-ga6eb3c7b339b-dirty #40 Not tainted
> [Â 333.949335] -----------------------------
> [Â 333.953445] ./include/linux/rcupdate.h:273 Illegal context switch in RCU read-side critical section!
> [Â 333.962698]
> [Â 333.962698] other info that might help us debug this:
> [Â 333.962698]
> [Â 333.970752]
> [Â 333.970752] rcu_scheduler_active = 2, debug_locks = 1
> [Â 333.977391] 2 locks held by ifconfig/1007:
> [Â 333.981520]Â #0: c10b18ec (rtnl_mutex){+.+.}, at: devinet_ioctl+0xc4/0x850
> [Â 333.988534]Â #1: c103e838 (rcu_read_lock){....}, at: rhashtable_walk_start_check+0x0/0x3dc
> [Â 333.996939]
> [Â 333.996939] stack backtrace:
> [Â 334.001334] CPU: 0 PID: 1007 Comm: ifconfig Not tainted 5.4.0-08849-ga6eb3c7b339b-dirty #40
> [Â 334.009733] Hardware name: Generic DRA72X (Flattened Device Tree)
> [Â 334.015878] [<c0113330>] (unwind_backtrace) from [<c010d23c>] (show_stack+0x10/0x14)
> [Â 334.023675] [<c010d23c>] (show_stack) from [<c09f9e08>] (dump_stack+0xe4/0x11c)
> [Â 334.031038] [<c09f9e08>] (dump_stack) from [<c016e4a4>] (___might_sleep+0x1e8/0x2bc)
> [Â 334.038834] [<c016e4a4>] (___might_sleep) from [<c0a17bd0>] (__mutex_lock+0x38/0xa18)
> [Â 334.046716] [<c0a17bd0>] (__mutex_lock) from [<c0a185cc>] (mutex_lock_nested+0x1c/0x24)
> [Â 334.054774] [<c0a185cc>] (mutex_lock_nested) from [<c0858208>] (mem_allocator_disconnect+0xf8/0x288)
> [Â 334.063966] [<c0858208>] (mem_allocator_disconnect) from [<c085df50>] (page_pool_release+0x230/0x3b4)
> [Â 334.073242] [<c085df50>] (page_pool_release) from [<c085e12c>] (page_pool_destroy+0x58/0x11c)
> [Â 334.081822] [<c085e12c>] (page_pool_destroy) from [<c0771554>] (cpsw_destroy_xdp_rxqs+0x88/0xa0)
> [Â 334.090663] [<c0771554>] (cpsw_destroy_xdp_rxqs) from [<c0774638>] (cpsw_ndo_stop+0x100/0x10c)
> [Â 334.099331] [<c0774638>] (cpsw_ndo_stop) from [<c0814fdc>] (__dev_close_many+0xac/0x130)
> [Â 334.107475] [<c0814fdc>] (__dev_close_many) from [<c0824068>] (__dev_change_flags+0xc8/0x1f0)
> [Â 334.116053] [<c0824068>] (__dev_change_flags) from [<c08241a8>] (dev_change_flags+0x18/0x48)
> [Â 334.124545] [<c08241a8>] (dev_change_flags) from [<c08efc3c>] (devinet_ioctl+0x6c0/0x850)
> [Â 334.132775] [<c08efc3c>] (devinet_ioctl) from [<c08f2d98>] (inet_ioctl+0x1f8/0x3b4)
> [Â 334.140483] [<c08f2d98>] (inet_ioctl) from [<c07f4594>] (sock_ioctl+0x398/0x5f4)
> [Â 334.147929] [<c07f4594>] (sock_ioctl) from [<c03279b4>] (do_vfs_ioctl+0x9c/0xa08)
> [Â 334.155461] [<c03279b4>] (do_vfs_ioctl) from [<c0328384>] (ksys_ioctl+0x64/0x74)
> [Â 334.162905] [<c0328384>] (ksys_ioctl) from [<c01011ac>] (__sys_trace_return+0x0/0x14)
> [Â 334.170781] Exception stack(0xed517fa8 to 0xed517ff0)
> [Â 334.175870] 7fa0:ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 0007b4ec bee79d84 00000003 00008914 bee79a80 0007b4ec
> [Â 334.184099] 7fc0: 0007b4ec bee79d84 bee79d84 00000036 bee79c4c bee79c4c bee79a80 00000003
> [Â 334.192325] 7fe0: 0009d1ec bee79a14 0003214b b6e94f7c
> [Â 334.197604] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:938
> [Â 334.206157] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1007, name: ifconfig
> [Â 334.214274] 2 locks held by ifconfig/1007:
> [Â 334.218401]Â #0: c10b18ec (rtnl_mutex){+.+.}, at: devinet_ioctl+0xc4/0x850
> [Â 334.225407]Â #1: c103e838 (rcu_read_lock){....}, at: rhashtable_walk_start_check+0x0/0x3dc
> [Â 334.233813] CPU: 0 PID: 1007 Comm: ifconfig Not tainted 5.4.0-08849-ga6eb3c7b339b-dirty #40
> [Â 334.242212] Hardware name: Generic DRA72X (Flattened Device Tree)
> [Â 334.248351] [<c0113330>] (unwind_backtrace) from [<c010d23c>] (show_stack+0x10/0x14)
> [Â 334.256147] [<c010d23c>] (show_stack) from [<c09f9e08>] (dump_stack+0xe4/0x11c)
> [Â 334.263506] [<c09f9e08>] (dump_stack) from [<c016e464>] (___might_sleep+0x1a8/0x2bc)
> [Â 334.271300] [<c016e464>] (___might_sleep) from [<c0a17bd0>] (__mutex_lock+0x38/0xa18)
> [Â 334.279181] [<c0a17bd0>] (__mutex_lock) from [<c0a185cc>] (mutex_lock_nested+0x1c/0x24)
> [Â 334.287238] [<c0a185cc>] (mutex_lock_nested) from [<c0858208>] (mem_allocator_disconnect+0xf8/0x288)
> [Â 334.296427] [<c0858208>] (mem_allocator_disconnect) from [<c085df50>] (page_pool_release+0x230/0x3b4)
> [Â 334.305703] [<c085df50>] (page_pool_release) from [<c085e12c>] (page_pool_destroy+0x58/0x11c)
> [Â 334.314281] [<c085e12c>] (page_pool_destroy) from [<c0771554>] (cpsw_destroy_xdp_rxqs+0x88/0xa0)
> [Â 334.323122] [<c0771554>] (cpsw_destroy_xdp_rxqs) from [<c0774638>] (cpsw_ndo_stop+0x100/0x10c)
> [Â 334.331788] [<c0774638>] (cpsw_ndo_stop) from [<c0814fdc>] (__dev_close_many+0xac/0x130)
> [Â 334.339931] [<c0814fdc>] (__dev_close_many) from [<c0824068>] (__dev_change_flags+0xc8/0x1f0)
> [Â 334.348510] [<c0824068>] (__dev_change_flags) from [<c08241a8>] (dev_change_flags+0x18/0x48)
> [Â 334.357000] [<c08241a8>] (dev_change_flags) from [<c08efc3c>] (devinet_ioctl+0x6c0/0x850)
> [Â 334.365228] [<c08efc3c>] (devinet_ioctl) from [<c08f2d98>] (inet_ioctl+0x1f8/0x3b4)
> [Â 334.372935] [<c08f2d98>] (inet_ioctl) from [<c07f4594>] (sock_ioctl+0x398/0x5f4)
> [Â 334.380380] [<c07f4594>] (sock_ioctl) from [<c03279b4>] (do_vfs_ioctl+0x9c/0xa08)
> [Â 334.387911] [<c03279b4>] (do_vfs_ioctl) from [<c0328384>] (ksys_ioctl+0x64/0x74)
> [Â 334.395355] [<c0328384>] (ksys_ioctl) from [<c01011ac>] (__sys_trace_return+0x0/0x14)
> [Â 334.403231] Exception stack(0xed517fa8 to 0xed517ff0)
> [Â 334.408319] 7fa0:ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 0007b4ec bee79d84 00000003 00008914 bee79a80 0007b4ec
> [Â 334.416548] 7fc0: 0007b4ec bee79d84 bee79d84 00000036 bee79c4c bee79c4c bee79a80 00000003
> [Â 334.424774] 7fe0: 0009d1ec bee79a14 0003214b b6e94f7c
>
>
> Enabled debug options:
> =================================================
> +CONFIG_LOCKUP_DETECTOR=y
> +CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC=y
> +CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC_VALUE=1
> +CONFIG_DETECT_HUNG_TASK=y
> +CONFIG_DEFAULT_HUNG_TASK_TIMEOUT=300
> +CONFIG_BOOTPARAM_HUNG_TASK_PANIC=y
> +CONFIG_BOOTPARAM_HUNG_TASK_PANIC_VALUE=1
> +CONFIG_PANIC_ON_OOPS=y
> +CONFIG_PANIC_ON_OOPS_VALUE=1
> +
> +CONFIG_DEBUG_RT_MUTEXES=y
> +CONFIG_DEBUG_PI_LIST=y
> +CONFIG_DEBUG_SPINLOCK=y
> +CONFIG_DEBUG_MUTEXES=y
> +CONFIG_DEBUG_WW_MUTEX_SLOWPATH=y
> +CONFIG_DEBUG_LOCK_ALLOC=y
> +CONFIG_PROVE_LOCKING=y
> +CONFIG_LOCKDEP=y
> +CONFIG_DEBUG_LOCKDEP=y
> +CONFIG_DEBUG_ATOMIC_SLEEP=y
> +CONFIG_DEBUG_LOCKING_API_SELFTESTS=n
> +CONFIG_STACKTRACE=y
> +CONFIG_DEBUG_BUGVERBOSE=y
> +CONFIG_DEBUG_LIST=y
> +CONFIG_DEBUG_SG=y
> +CONFIG_DEBUG_NOTIFIERS=y
> +
> +CONFIG_SPARSE_RCU_POINTER=y
> +CONFIG_RCU_CPU_STALL_TIMEOUT=60
> +CONFIG_RCU_CPU_STALL_INFO=y
> +CONFIG_RCU_TRACE=y
> +CONFIG_PROVE_RCU=y
> +CONFIG_PROVE_RCU_REPEATEDLY=y
> +
> +CONFIG_DMA_API_DEBUG=y

--
Best regards,
grygorii