Message-ID: <20190215114617.lfag4sarqzbtca6c@shell.armlinux.org.uk>
Date:   Fri, 15 Feb 2019 11:46:17 +0000
From:   Russell King - ARM Linux admin <linux@...linux.org.uk>
To:     Andrew Lunn <andrew@...n.ch>,
        Vivien Didelot <vivien.didelot@...il.com>
Cc:     netdev@...r.kernel.org
Subject: [BUG] 4.20: mv88e6xxx: WARNING: possible circular locking dependency
 detected

Hi Andrew, Vivien,

I decided to try adding support for the DSA switch interrupt on
SolidRun's Clearfog platform, but I notice having done so I get:

WARNING: possible circular locking dependency detected
4.20.0+ #297 Not tainted
------------------------------------------------------
systemd-udevd/157 is trying to acquire lock:
ecc4a080 (&chip->reg_lock){+.+.}, at: __setup_irq+0x640/0x704

but task is already holding lock:
edf9c940 (&desc->request_mutex){+.+.}, at: __setup_irq+0xa0/0x704

which lock already depends on the new lock.


the existing dependency chain (in reverse order) is:

-> #1 (&desc->request_mutex){+.+.}:
       mutex_lock_nested+0x1c/0x24
       __setup_irq+0xa0/0x704
       request_threaded_irq+0xd0/0x150
       mv88e6xxx_probe+0x41c/0x694 [mv88e6xxx]
       mdio_probe+0x2c/0x54
       really_probe+0x200/0x2c4
       driver_probe_device+0x5c/0x174
       __driver_attach+0xd8/0xdc
       bus_for_each_dev+0x58/0x7c
       bus_add_driver+0xe4/0x1f0
       driver_register+0x7c/0x110
       mdio_driver_register+0x24/0x58
       do_one_initcall+0x74/0x2e8
       do_init_module+0x60/0x1d0
       load_module+0x1968/0x1ff4
       sys_finit_module+0x8c/0x98
       ret_fast_syscall+0x0/0x28
       0xbee82ae8

-> #0 (&chip->reg_lock){+.+.}:
       __mutex_lock+0x50/0x8b8
       mutex_lock_nested+0x1c/0x24
       __setup_irq+0x640/0x704
       request_threaded_irq+0xd0/0x150
       mv88e6xxx_g2_irq_setup+0xcc/0x1b4 [mv88e6xxx]
       mv88e6xxx_probe+0x44c/0x694 [mv88e6xxx]
       mdio_probe+0x2c/0x54
       really_probe+0x200/0x2c4
       driver_probe_device+0x5c/0x174
       __driver_attach+0xd8/0xdc
       bus_for_each_dev+0x58/0x7c
       bus_add_driver+0xe4/0x1f0
       driver_register+0x7c/0x110
       mdio_driver_register+0x24/0x58
       do_one_initcall+0x74/0x2e8
       do_init_module+0x60/0x1d0
       load_module+0x1968/0x1ff4
       sys_finit_module+0x8c/0x98
       ret_fast_syscall+0x0/0x28
       0xbee82ae8

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&desc->request_mutex);
                               lock(&chip->reg_lock);
                               lock(&desc->request_mutex);
  lock(&chip->reg_lock);

 *** DEADLOCK ***

2 locks held by systemd-udevd/157:
 #0: ee040868 (&dev->mutex){....}, at: __driver_attach+0x70/0xdc
------------[ cut here ]------------
 #1: edf9c940 (&desc->request_mutex){+.+.}, at: __setup_irq+0xa0/0x704
WARNING: CPU: 0 PID: 152 at kernel/locking/lockdep.c:355
stack backtrace:
downgrading a read lock
Modules linked in:
CPU: 1 PID: 157 Comm: systemd-udevd Not tainted 4.20.0+ #297
 marvell_cesa(+) mv88e6xxx(+) dsa_core devlink xhci_plat_hcd(+) xhci_hcd armada_Hardware name: Marvell Armada 380/385 (Device Tree)
[<c0019638>] (unwind_backtrace) from [<c0014888>] (show_stack+0x10/0x14)
[<c0014888>] (show_stack) from [<c07f55c0>] (dump_stack+0x9c/0xd4)
[<c07f55c0>] (dump_stack) from [<c0088afc>] (print_circular_bug+0x284/0x2d8)
[<c0088afc>] (print_circular_bug) from [<c0086b5c>] (__lock_acquire+0x15d4/0x19b[<c0086b5c>] (__lock_acquire) from [<c0087828>] (lock_acquire+0xc4/0x1dc)
[<c0087828>] (lock_acquire) from [<c080fe68>] (__mutex_lock+0x50/0x8b8)
[<c080fe68>] (__mutex_lock) from [<c0810758>] (mutex_lock_nested+0x1c/0x24)
[<c0810758>] (mutex_lock_nested) from [<c009e060>] (__setup_irq+0x640/0x704)
[<c009e060>] (__setup_irq) from [<c009e2e0>] (request_threaded_irq+0xd0/0x150)
[<c009e2e0>] (request_threaded_irq) from [<bf0dc970>] (mv88e6xxx_g2_irq_setup+0x[<bf0dc970>] (mv88e6xxx_g2_irq_setup [mv88e6xxx]) from [<bf0d5a90>] (mv88e6xxx_p[<bf0d5a90>] (mv88e6xxx_probe [mv88e6xxx]) from [<c050d420>] (mdio_probe+0x2c/0x[<c050d420>] (mdio_probe) from [<c0496eac>] (really_probe+0x200/0x2c4)
[<c0496eac>] (really_probe) from [<c0497140>] (driver_probe_device+0x5c/0x174)
[<c0497140>] (driver_probe_device) from [<c0497330>] (__driver_attach+0xd8/0xdc)[<c0497330>] (__driver_attach) from [<c0495494>] (bus_for_each_dev+0x58/0x7c)
[<c0495494>] (bus_for_each_dev) from [<c04963d4>] (bus_add_driver+0xe4/0x1f0)
[<c04963d4>] (bus_add_driver) from [<c0498038>] (driver_register+0x7c/0x110)
[<c0498038>] (driver_register) from [<c050d338>] (mdio_driver_register+0x24/0x58[<c050d338>] (mdio_driver_register) from [<c000afdc>] (do_one_initcall+0x74/0x2e[<c000afdc>] (do_one_initcall) from [<c00d4994>] (do_init_module+0x60/0x1d0)
[<c00d4994>] (do_init_module) from [<c00d39e0>] (load_module+0x1968/0x1ff4)
[<c00d39e0>] (load_module) from [<c00d4248>] (sys_finit_module+0x8c/0x98)
[<c00d4248>] (sys_finit_module) from [<c0009000>] (ret_fast_syscall+0x0/0x28)
Exception stack(0xed42bfa8 to 0xed42bff0)
bfa0:                   00020000 00000000 0000000b b6e814b5 00000000 010b31e0
bfc0: 00020000 00000000 00000000 0000017b 010b1b30 00020000 00000000 010b31e0
bfe0: bee82af8 bee82ae8 b6e7b2ac b6ddad70
CPU: 0 PID: 152 Comm: systemd-udevd Not tainted 4.20.0+ #297
Hardware name: Marvell Armada 380/385 (Device Tree)
[<c0019638>] (unwind_backtrace) from [<c0014888>] (show_stack+0x10/0x14)
[<c0014888>] (show_stack) from [<c07f55c0>] (dump_stack+0x9c/0xd4)
[<c07f55c0>] (dump_stack) from [<c00312bc>] (__warn+0xf8/0x124)
[<c00312bc>] (__warn) from [<c00313b0>] (warn_slowpath_fmt+0x38/0x48)
[<c00313b0>] (warn_slowpath_fmt) from [<c0087518>] (lock_downgrade+0x14c/0x1b8)
[<c0087518>] (lock_downgrade) from [<c0081650>] (downgrade_write+0x14/0xd4)
[<c0081650>] (downgrade_write) from [<c0189d10>] (__do_munmap+0x2b8/0x31c)
[<c0189d10>] (__do_munmap) from [<c0189dd4>] (__vm_munmap+0x60/0xa0)
[<c0189dd4>] (__vm_munmap) from [<c0009000>] (ret_fast_syscall+0x0/0x28)
Exception stack(0xed443fa8 to 0xed443ff0)
3fa0:                   010a8240 00001000 b665f000 00001000 00000000 00000000
3fc0: 010a8240 00001000 00000000 0000005b 00000000 00000007 b6ee5f10 00000000
3fe0: fbad2418 bee7f124 b6d7c7b4 b6ddafac
irq event stamp: 83666
hardirqs last  enabled at (83665): [<c001d6b0>] do_page_fault+0x190/0x360
hardirqs last disabled at (83666): [<c080e474>] __schedule+0xbc/0x9c4
softirqs last  enabled at (82980): [<c000a484>] __do_softirq+0x344/0x540
softirqs last disabled at (82971): [<c00386e0>] irq_exit+0x124/0x144
---[ end trace c91466d44e5e3485 ]---

This is caused by the locking order inversion in mv88e6xxx_probe:

        mutex_lock(&chip->reg_lock);
        if (chip->irq > 0)
                err = mv88e6xxx_g1_irq_setup(chip);
        else
                err = mv88e6xxx_irq_poll_setup(chip);
        mutex_unlock(&chip->reg_lock);

Here, we take chip->reg_lock, and then call into mv88e6xxx_g1_irq_setup()
which then calls request_threaded_irq(), taking the request_mutex.
However, when we request the g2 interrupt, we call request_threaded_irq()
again, which takes the request_mutex, which then goes on to call
chip_bus_lock().  This comes through to mv88e6xxx_g1_irq_bus_lock,
which then tries to grab chip->reg_lock.

It looks to me like the mutex_lock()/unlock() for reg_lock should be
moved inside mv88e6xxx_g1_irq_free_common() and
mv88e6xxx_g1_irq_setup_common(), which will avoid holding it while
calling request_threaded_irq() or setting up the delayed work.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up

Powered by blists - more mailing lists