* Regresion with dsa_port_disable
@ 2019-08-24 22:53 Andrew Lunn
2019-08-24 23:12 ` Vivien Didelot
0 siblings, 1 reply; 5+ messages in thread
From: Andrew Lunn @ 2019-08-24 22:53 UTC (permalink / raw
To: Vivien Didelot; +Cc: netdev
Hi Vivien
I just booted a ZII devel C and got a new warning splat.
WARNING: CPU: 0 PID: 925 at kernel/irq/manage.c:1708 __free_irq+0xc8/0x2c4
Trying to free already-free IRQ 0
Modules linked in:
CPU: 0 PID: 925 Comm: kworker/0:2 Not tainted 5.3.0-rc5-01151-g7ff758fcdf65 #231
Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
Workqueue: events deferred_probe_work_func
Backtrace:
[<8010d9e4>] (dump_backtrace) from [<8010dd9c>] (show_stack+0x20/0x24)
r7:8016edf8 r6:00000009 r5:00000000 r4:9ec67944
[<8010dd7c>] (show_stack) from [<8083b03c>] (dump_stack+0x24/0x28)
[<8083b018>] (dump_stack) from [<8011c108>] (__warn.part.3+0xcc/0xf8)
[<8011c03c>] (__warn.part.3) from [<8011c1ac>] (warn_slowpath_fmt+0x78/0x94)
r6:000006ac r5:80a8cbf0 r4:80d07088
[<8011c138>] (warn_slowpath_fmt) from [<8016edf8>] (__free_irq+0xc8/0x2c4)
r3:00000000 r2:80a8cca8
r7:9f486668 r6:9ee25268 r5:9f486600 r4:9ee25268
[<8016ed30>] (__free_irq) from [<8016f07c>] (free_irq+0x38/0x74)
r10:9eeb3600 r9:9e412040 r8:00000009 r7:9ee26040 r6:9ee2404c r5:9ee242c8
r4:9ee25268 r3:00000c00
[<8016f044>] (free_irq) from [<805a244c>] (mv88e6390x_serdes_irq_free+0x68/0x98)
r5:9ee242c8 r4:9ee24040
[<805a23e4>] (mv88e6390x_serdes_irq_free) from [<8059bc94>] (mv88e6xxx_port_disable+0x58/0x98)
r7:9ee26040 r6:00000009 r5:9ee2404c r4:9ee24040
[<8059bc3c>] (mv88e6xxx_port_disable) from [<80806f70>] (dsa_port_disable+0x44/0x50)
r7:9ee26040 r6:9ee26d74 r5:00000009 r4:9ee26040
[<80806f2c>] (dsa_port_disable) from [<80805df0>] (dsa_register_switch+0x964/0xab8)
r5:9efe194c r4:9ee26d38
[<8080548c>] (dsa_register_switch) from [<8059b734>] (mv88e6xxx_probe+0x730/0x778)
r10:80943e64 r9:9fbf77d0 r8:00000000 r7:80d07088 r6:9e410040 r5:00000000
r4:9e40e800
[<8059b004>] (mv88e6xxx_probe) from [<80582da8>] (mdio_probe+0x40/0x64)
r10:00000012 r9:80d5eccc r8:00000000 r7:00000000 r6:8141f358 r5:9e40e800
r4:80d5eccc
[<80582d68>] (mdio_probe) from [<80518858>] (really_probe+0x100/0x2d8)
r5:9e40e800 r4:8141f354
The previous code was careful to balance mv88e6352_serdes_irq_setup()
with mv88e6390x_serdes_irq_free(). I _think_ your change broke this
balance, and we now try to free an interrupt which was never
allocated.
Andrew
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regresion with dsa_port_disable
2019-08-24 22:53 Regresion with dsa_port_disable Andrew Lunn
@ 2019-08-24 23:12 ` Vivien Didelot
2019-08-24 23:16 ` Andrew Lunn
0 siblings, 1 reply; 5+ messages in thread
From: Vivien Didelot @ 2019-08-24 23:12 UTC (permalink / raw
To: Andrew Lunn; +Cc: netdev
Hi Andrew,
On Sun, 25 Aug 2019 00:53:06 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> I just booted a ZII devel C and got a new warning splat.
>
> WARNING: CPU: 0 PID: 925 at kernel/irq/manage.c:1708 __free_irq+0xc8/0x2c4
> Trying to free already-free IRQ 0
> Modules linked in:
> CPU: 0 PID: 925 Comm: kworker/0:2 Not tainted 5.3.0-rc5-01151-g7ff758fcdf65 #231
> Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
> Workqueue: events deferred_probe_work_func
> Backtrace:
> [<8010d9e4>] (dump_backtrace) from [<8010dd9c>] (show_stack+0x20/0x24)
> r7:8016edf8 r6:00000009 r5:00000000 r4:9ec67944
> [<8010dd7c>] (show_stack) from [<8083b03c>] (dump_stack+0x24/0x28)
> [<8083b018>] (dump_stack) from [<8011c108>] (__warn.part.3+0xcc/0xf8)
> [<8011c03c>] (__warn.part.3) from [<8011c1ac>] (warn_slowpath_fmt+0x78/0x94)
> r6:000006ac r5:80a8cbf0 r4:80d07088
> [<8011c138>] (warn_slowpath_fmt) from [<8016edf8>] (__free_irq+0xc8/0x2c4)
> r3:00000000 r2:80a8cca8
> r7:9f486668 r6:9ee25268 r5:9f486600 r4:9ee25268
> [<8016ed30>] (__free_irq) from [<8016f07c>] (free_irq+0x38/0x74)
> r10:9eeb3600 r9:9e412040 r8:00000009 r7:9ee26040 r6:9ee2404c r5:9ee242c8
> r4:9ee25268 r3:00000c00
> [<8016f044>] (free_irq) from [<805a244c>] (mv88e6390x_serdes_irq_free+0x68/0x98)
> r5:9ee242c8 r4:9ee24040
> [<805a23e4>] (mv88e6390x_serdes_irq_free) from [<8059bc94>] (mv88e6xxx_port_disable+0x58/0x98)
> r7:9ee26040 r6:00000009 r5:9ee2404c r4:9ee24040
> [<8059bc3c>] (mv88e6xxx_port_disable) from [<80806f70>] (dsa_port_disable+0x44/0x50)
> r7:9ee26040 r6:9ee26d74 r5:00000009 r4:9ee26040
> [<80806f2c>] (dsa_port_disable) from [<80805df0>] (dsa_register_switch+0x964/0xab8)
> r5:9efe194c r4:9ee26d38
> [<8080548c>] (dsa_register_switch) from [<8059b734>] (mv88e6xxx_probe+0x730/0x778)
> r10:80943e64 r9:9fbf77d0 r8:00000000 r7:80d07088 r6:9e410040 r5:00000000
> r4:9e40e800
> [<8059b004>] (mv88e6xxx_probe) from [<80582da8>] (mdio_probe+0x40/0x64)
> r10:00000012 r9:80d5eccc r8:00000000 r7:00000000 r6:8141f358 r5:9e40e800
> r4:80d5eccc
> [<80582d68>] (mdio_probe) from [<80518858>] (really_probe+0x100/0x2d8)
> r5:9e40e800 r4:8141f354
>
> The previous code was careful to balance mv88e6352_serdes_irq_setup()
> with mv88e6390x_serdes_irq_free(). I _think_ your change broke this
> balance, and we now try to free an interrupt which was never
> allocated.
What do you mean by "balance mv88e6352_serdes_irq_setup() with
mv88e6390x_serdes_irq_free()"?
Vivien
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regresion with dsa_port_disable
2019-08-24 23:12 ` Vivien Didelot
@ 2019-08-24 23:16 ` Andrew Lunn
2019-08-25 0:53 ` Vivien Didelot
0 siblings, 1 reply; 5+ messages in thread
From: Andrew Lunn @ 2019-08-24 23:16 UTC (permalink / raw
To: Vivien Didelot; +Cc: netdev
On Sat, Aug 24, 2019 at 07:12:20PM -0400, Vivien Didelot wrote:
> Hi Andrew,
>
> On Sun, 25 Aug 2019 00:53:06 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> > I just booted a ZII devel C and got a new warning splat.
> >
> > WARNING: CPU: 0 PID: 925 at kernel/irq/manage.c:1708 __free_irq+0xc8/0x2c4
> > Trying to free already-free IRQ 0
> > Modules linked in:
> > CPU: 0 PID: 925 Comm: kworker/0:2 Not tainted 5.3.0-rc5-01151-g7ff758fcdf65 #231
> > Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
> > Workqueue: events deferred_probe_work_func
> > Backtrace:
> > [<8010d9e4>] (dump_backtrace) from [<8010dd9c>] (show_stack+0x20/0x24)
> > r7:8016edf8 r6:00000009 r5:00000000 r4:9ec67944
> > [<8010dd7c>] (show_stack) from [<8083b03c>] (dump_stack+0x24/0x28)
> > [<8083b018>] (dump_stack) from [<8011c108>] (__warn.part.3+0xcc/0xf8)
> > [<8011c03c>] (__warn.part.3) from [<8011c1ac>] (warn_slowpath_fmt+0x78/0x94)
> > r6:000006ac r5:80a8cbf0 r4:80d07088
> > [<8011c138>] (warn_slowpath_fmt) from [<8016edf8>] (__free_irq+0xc8/0x2c4)
> > r3:00000000 r2:80a8cca8
> > r7:9f486668 r6:9ee25268 r5:9f486600 r4:9ee25268
> > [<8016ed30>] (__free_irq) from [<8016f07c>] (free_irq+0x38/0x74)
> > r10:9eeb3600 r9:9e412040 r8:00000009 r7:9ee26040 r6:9ee2404c r5:9ee242c8
> > r4:9ee25268 r3:00000c00
> > [<8016f044>] (free_irq) from [<805a244c>] (mv88e6390x_serdes_irq_free+0x68/0x98)
> > r5:9ee242c8 r4:9ee24040
> > [<805a23e4>] (mv88e6390x_serdes_irq_free) from [<8059bc94>] (mv88e6xxx_port_disable+0x58/0x98)
> > r7:9ee26040 r6:00000009 r5:9ee2404c r4:9ee24040
> > [<8059bc3c>] (mv88e6xxx_port_disable) from [<80806f70>] (dsa_port_disable+0x44/0x50)
> > r7:9ee26040 r6:9ee26d74 r5:00000009 r4:9ee26040
> > [<80806f2c>] (dsa_port_disable) from [<80805df0>] (dsa_register_switch+0x964/0xab8)
> > r5:9efe194c r4:9ee26d38
> > [<8080548c>] (dsa_register_switch) from [<8059b734>] (mv88e6xxx_probe+0x730/0x778)
> > r10:80943e64 r9:9fbf77d0 r8:00000000 r7:80d07088 r6:9e410040 r5:00000000
> > r4:9e40e800
> > [<8059b004>] (mv88e6xxx_probe) from [<80582da8>] (mdio_probe+0x40/0x64)
> > r10:00000012 r9:80d5eccc r8:00000000 r7:00000000 r6:8141f358 r5:9e40e800
> > r4:80d5eccc
> > [<80582d68>] (mdio_probe) from [<80518858>] (really_probe+0x100/0x2d8)
> > r5:9e40e800 r4:8141f354
> >
> > The previous code was careful to balance mv88e6352_serdes_irq_setup()
> > with mv88e6390x_serdes_irq_free(). I _think_ your change broke this
> > balance, and we now try to free an interrupt which was never
> > allocated.
>
> What do you mean by "balance mv88e6352_serdes_irq_setup() with
> mv88e6390x_serdes_irq_free()"?
Hi Vivien
It never called mv88e6390x_serdes_irq_free() unless
mv88e6352_serdes_irq_setup() had been called first.
mv88e6390x_serdes_irq_free() makes the assumption there actually is an
interrupt to free. I suspect your changes now call
mv88e6390x_serdes_irq_free() unconditionally.
Andrew
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regresion with dsa_port_disable
2019-08-24 23:16 ` Andrew Lunn
@ 2019-08-25 0:53 ` Vivien Didelot
2019-08-25 8:35 ` Marek Behun
0 siblings, 1 reply; 5+ messages in thread
From: Vivien Didelot @ 2019-08-25 0:53 UTC (permalink / raw
To: Andrew Lunn; +Cc: netdev
Hi Andrew,
On Sun, 25 Aug 2019 01:16:53 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> On Sat, Aug 24, 2019 at 07:12:20PM -0400, Vivien Didelot wrote:
> > Hi Andrew,
> >
> > On Sun, 25 Aug 2019 00:53:06 +0200, Andrew Lunn <andrew@lunn.ch> wrote:
> > > I just booted a ZII devel C and got a new warning splat.
> > >
> > > WARNING: CPU: 0 PID: 925 at kernel/irq/manage.c:1708 __free_irq+0xc8/0x2c4
> > > Trying to free already-free IRQ 0
> > > Modules linked in:
> > > CPU: 0 PID: 925 Comm: kworker/0:2 Not tainted 5.3.0-rc5-01151-g7ff758fcdf65 #231
> > > Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
> > > Workqueue: events deferred_probe_work_func
> > > Backtrace:
> > > [<8010d9e4>] (dump_backtrace) from [<8010dd9c>] (show_stack+0x20/0x24)
> > > r7:8016edf8 r6:00000009 r5:00000000 r4:9ec67944
> > > [<8010dd7c>] (show_stack) from [<8083b03c>] (dump_stack+0x24/0x28)
> > > [<8083b018>] (dump_stack) from [<8011c108>] (__warn.part.3+0xcc/0xf8)
> > > [<8011c03c>] (__warn.part.3) from [<8011c1ac>] (warn_slowpath_fmt+0x78/0x94)
> > > r6:000006ac r5:80a8cbf0 r4:80d07088
> > > [<8011c138>] (warn_slowpath_fmt) from [<8016edf8>] (__free_irq+0xc8/0x2c4)
> > > r3:00000000 r2:80a8cca8
> > > r7:9f486668 r6:9ee25268 r5:9f486600 r4:9ee25268
> > > [<8016ed30>] (__free_irq) from [<8016f07c>] (free_irq+0x38/0x74)
> > > r10:9eeb3600 r9:9e412040 r8:00000009 r7:9ee26040 r6:9ee2404c r5:9ee242c8
> > > r4:9ee25268 r3:00000c00
> > > [<8016f044>] (free_irq) from [<805a244c>] (mv88e6390x_serdes_irq_free+0x68/0x98)
> > > r5:9ee242c8 r4:9ee24040
> > > [<805a23e4>] (mv88e6390x_serdes_irq_free) from [<8059bc94>] (mv88e6xxx_port_disable+0x58/0x98)
> > > r7:9ee26040 r6:00000009 r5:9ee2404c r4:9ee24040
> > > [<8059bc3c>] (mv88e6xxx_port_disable) from [<80806f70>] (dsa_port_disable+0x44/0x50)
> > > r7:9ee26040 r6:9ee26d74 r5:00000009 r4:9ee26040
> > > [<80806f2c>] (dsa_port_disable) from [<80805df0>] (dsa_register_switch+0x964/0xab8)
> > > r5:9efe194c r4:9ee26d38
> > > [<8080548c>] (dsa_register_switch) from [<8059b734>] (mv88e6xxx_probe+0x730/0x778)
> > > r10:80943e64 r9:9fbf77d0 r8:00000000 r7:80d07088 r6:9e410040 r5:00000000
> > > r4:9e40e800
> > > [<8059b004>] (mv88e6xxx_probe) from [<80582da8>] (mdio_probe+0x40/0x64)
> > > r10:00000012 r9:80d5eccc r8:00000000 r7:00000000 r6:8141f358 r5:9e40e800
> > > r4:80d5eccc
> > > [<80582d68>] (mdio_probe) from [<80518858>] (really_probe+0x100/0x2d8)
> > > r5:9e40e800 r4:8141f354
> > >
> > > The previous code was careful to balance mv88e6352_serdes_irq_setup()
> > > with mv88e6390x_serdes_irq_free(). I _think_ your change broke this
> > > balance, and we now try to free an interrupt which was never
> > > allocated.
> >
> > What do you mean by "balance mv88e6352_serdes_irq_setup() with
> > mv88e6390x_serdes_irq_free()"?
>
> Hi Vivien
>
> It never called mv88e6390x_serdes_irq_free() unless
> mv88e6352_serdes_irq_setup() had been called first.
> mv88e6390x_serdes_irq_free() makes the assumption there actually is an
> interrupt to free. I suspect your changes now call
> mv88e6390x_serdes_irq_free() unconditionally.
OK I think you meant info->ops->serdes_irq_free and
info->ops->serdes_irq_setup, otherwise it's confusing.
I think I know what's going on, I'll look into it soon.
Thanks!
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regresion with dsa_port_disable
2019-08-25 0:53 ` Vivien Didelot
@ 2019-08-25 8:35 ` Marek Behun
0 siblings, 0 replies; 5+ messages in thread
From: Marek Behun @ 2019-08-25 8:35 UTC (permalink / raw
To: Vivien Didelot; +Cc: Andrew Lunn, netdev
On Sat, 24 Aug 2019 20:53:49 -0400
Vivien Didelot <vivien.didelot@gmail.com> wrote:
> OK I think you meant info->ops->serdes_irq_free and
> info->ops->serdes_irq_setup, otherwise it's confusing.
>
> I think I know what's going on, I'll look into it soon.
Either dsa_port_setup is calling dsa_port_disable for
DSA_PORT_TYPE_UNUSED and this causes that, or, as line
Workqueue: events deferred_probe_work_func
suggests, probe is deferred, dsa_tree_setup calls
dsa_tree_teardown_switches, and thus dsa_port_disable is called.
We should add a check into dsa_port_disable if that port was successfuly
enabled, or something like that.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2019-08-25 8:35 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-08-24 22:53 Regresion with dsa_port_disable Andrew Lunn
2019-08-24 23:12 ` Vivien Didelot
2019-08-24 23:16 ` Andrew Lunn
2019-08-25 0:53 ` Vivien Didelot
2019-08-25 8:35 ` Marek Behun
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.