| Message ID | 5149EFFA.1050408@codeaurora.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 03/20/2013 06:20 PM, Stephen Boyd wrote: > On 03/20/13 10:09, Gregory CLEMENT wrote: >> On 03/13/2013 07:17 PM, Stephen Boyd wrote: >>> Separate the armada 370xp local timers from the local timer API. >>> This will allow us to remove ARM local timer support in the near >>> future and makes this driver multi-architecture friendly. >> At first view the code looks good, but when I applied your patch set on >> linux-next, build it and run it on a Armada XP based board (AX3 with 2 cores), >> it crashed: > [...] >> Internal error: Oops - undefined instruction: 0 [#1] SMP ARM >> Modules linked in: >> CPU: 0 Not tainted (3.9.0-rc3-next-20130319-00010-g728b448 #153) >> PC is at 0xe92d45f0 >> LR is at armada_370_xp_timer_interrupt+0x3c/0x4c >> pc : [<e92d45f0>] lr : [<c023c2bc>] psr: 600001d3 >> sp : c0635eb8 ip : 00000000 fp : c063c3f0 >> r10: 000003ff r9 : 00000000 r8 : 00000010 >> r7 : c22b3c40 r6 : ef007c00 r5 : c0640fcc r4 : c0053e30 >> r3 : e92d45f0 r2 : fffffffe r1 : c22b3c40 r0 : c0053e30 >> Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel >> Control: 10c53c7d Table: 0000406a DAC: 00000015 >> Process swapper/0 (pid: 0, stack limit = 0xc0634238) >> Stack: (0xc0635eb8 to 0xc0636000) >> 5ea0: ef004c80 c0063224 >> 5ec0: 00000010 00000010 00000000 c0660ac0 c0635f18 c005fcb8 c0632b90 c000ed94 >> 5ee0: c0313c60 60000153 00000001 c00085a8 c0313c54 c0313c60 60000153 ffffffff >> 5f00: c0635f4c 00000000 562f5842 c06360c0 00000000 c000db60 0000001a ffff8ad0 >> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >> 5f60: 00000021 00000000 00000003 00000004 0000006e c065fcc0 c067924c c063ceb8 >> 5f80: c063cc84 c006d8c0 00000005 c065fcc0 c067924c c0421764 c22ad780 c063c42c >> 5fa0: 562f5842 c063cca8 c06605c0 c04379c0 c22ad780 00000000 562f5842 00000000 >> 5fc0: 00000000 c0417754 ffffffff ffffffff c04172dc 00000000 00000000 c04379c0 >> 5fe0: 10c53c7d c063c414 c04379bc c063febc 0000406a 00008074 00000000 00000000 >> [<c023c2bc>] (armada_370_xp_timer_interrupt+0x3c/0x4c) from [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) >> [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) from [<c005fcb8>] (generic_handle_irq+0x20/0x30) >> [<c005fcb8>] (generic_handle_irq+0x20/0x30) from [<c000ed94>] (handle_IRQ+0x38/0x90) >> [<c000ed94>] (handle_IRQ+0x38/0x90) from [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) >> [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) from [<c000db60>] (__irq_svc+0x40/0x50) >> Exception stack(0xc0635f18 to 0xc0635f60) >> 5f00: 0000001a ffff8ad0 >> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >> [<c000db60>] (__irq_svc+0x40/0x50) from [<c0313c60>] (calibrate_delay+0x378/0x528) >> [<c0313c60>] (calibrate_delay+0x378/0x528) from [<c0417754>] (start_kernel+0x250/0x2dc) >> [<c0417754>] (start_kernel+0x250/0x2dc) from [<00008074>] (0x8074) >> Code: 1fe7deb7 cd5772dd fff5692e ed55f79e (7ed5a5f7) >> >> >> >> I am trying to figure out what happened. > > Argh. Stupid casting again. Can you try this? Our emails must have crossed, your fix is also fine :) > > ---8<---- > > diff --git a/drivers/clocksource/time-armada-370-xp.c b/drivers/clocksource/time-armada-370-xp.c > index ee2e50c5..a9bf37a 100644 > --- a/drivers/clocksource/time-armada-370-xp.c > +++ b/drivers/clocksource/time-armada-370-xp.c > @@ -150,7 +150,7 @@ static irqreturn_t armada_370_xp_timer_interrupt(int irq, void *dev_id) > /* > * ACK timer interrupt and call event handler. > */ > - struct clock_event_device *evt = *(struct clock_event_device **)dev_id; > + struct clock_event_device *evt = dev_id; > > writel(TIMER0_CLR_MASK, local_base + LCL_TIMER_EVENTS_STATUS); > evt->event_handler(evt); >
On 03/20/2013 06:26 PM, Gregory CLEMENT wrote: > On 03/20/2013 06:20 PM, Stephen Boyd wrote: >> On 03/20/13 10:09, Gregory CLEMENT wrote: >>> On 03/13/2013 07:17 PM, Stephen Boyd wrote: >>>> Separate the armada 370xp local timers from the local timer API. >>>> This will allow us to remove ARM local timer support in the near >>>> future and makes this driver multi-architecture friendly. >>> At first view the code looks good, but when I applied your patch set on >>> linux-next, build it and run it on a Armada XP based board (AX3 with 2 cores), >>> it crashed: >> [...] >>> Internal error: Oops - undefined instruction: 0 [#1] SMP ARM >>> Modules linked in: >>> CPU: 0 Not tainted (3.9.0-rc3-next-20130319-00010-g728b448 #153) >>> PC is at 0xe92d45f0 >>> LR is at armada_370_xp_timer_interrupt+0x3c/0x4c >>> pc : [<e92d45f0>] lr : [<c023c2bc>] psr: 600001d3 >>> sp : c0635eb8 ip : 00000000 fp : c063c3f0 >>> r10: 000003ff r9 : 00000000 r8 : 00000010 >>> r7 : c22b3c40 r6 : ef007c00 r5 : c0640fcc r4 : c0053e30 >>> r3 : e92d45f0 r2 : fffffffe r1 : c22b3c40 r0 : c0053e30 >>> Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel >>> Control: 10c53c7d Table: 0000406a DAC: 00000015 >>> Process swapper/0 (pid: 0, stack limit = 0xc0634238) >>> Stack: (0xc0635eb8 to 0xc0636000) >>> 5ea0: ef004c80 c0063224 >>> 5ec0: 00000010 00000010 00000000 c0660ac0 c0635f18 c005fcb8 c0632b90 c000ed94 >>> 5ee0: c0313c60 60000153 00000001 c00085a8 c0313c54 c0313c60 60000153 ffffffff >>> 5f00: c0635f4c 00000000 562f5842 c06360c0 00000000 c000db60 0000001a ffff8ad0 >>> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >>> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >>> 5f60: 00000021 00000000 00000003 00000004 0000006e c065fcc0 c067924c c063ceb8 >>> 5f80: c063cc84 c006d8c0 00000005 c065fcc0 c067924c c0421764 c22ad780 c063c42c >>> 5fa0: 562f5842 c063cca8 c06605c0 c04379c0 c22ad780 00000000 562f5842 00000000 >>> 5fc0: 00000000 c0417754 ffffffff ffffffff c04172dc 00000000 00000000 c04379c0 >>> 5fe0: 10c53c7d c063c414 c04379bc c063febc 0000406a 00008074 00000000 00000000 >>> [<c023c2bc>] (armada_370_xp_timer_interrupt+0x3c/0x4c) from [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) >>> [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) from [<c005fcb8>] (generic_handle_irq+0x20/0x30) >>> [<c005fcb8>] (generic_handle_irq+0x20/0x30) from [<c000ed94>] (handle_IRQ+0x38/0x90) >>> [<c000ed94>] (handle_IRQ+0x38/0x90) from [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) >>> [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) from [<c000db60>] (__irq_svc+0x40/0x50) >>> Exception stack(0xc0635f18 to 0xc0635f60) >>> 5f00: 0000001a ffff8ad0 >>> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >>> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >>> [<c000db60>] (__irq_svc+0x40/0x50) from [<c0313c60>] (calibrate_delay+0x378/0x528) >>> [<c0313c60>] (calibrate_delay+0x378/0x528) from [<c0417754>] (start_kernel+0x250/0x2dc) >>> [<c0417754>] (start_kernel+0x250/0x2dc) from [<00008074>] (0x8074) >>> Code: 1fe7deb7 cd5772dd fff5692e ed55f79e (7ed5a5f7) >>> >>> >>> >>> I am trying to figure out what happened. >> >> Argh. Stupid casting again. Can you try this? > > Our emails must have crossed, your fix is also fine :) > I have also tested the patch set on a Armada 370 based board (Armada 370 DB), and it works well. So with this last chunk you can add my: Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com> and also my tested-by: Gregory CLEMENT <gregory.clement@free-electrons.com> > >> >> ---8<---- >> >> diff --git a/drivers/clocksource/time-armada-370-xp.c b/drivers/clocksource/time-armada-370-xp.c >> index ee2e50c5..a9bf37a 100644 >> --- a/drivers/clocksource/time-armada-370-xp.c >> +++ b/drivers/clocksource/time-armada-370-xp.c >> @@ -150,7 +150,7 @@ static irqreturn_t armada_370_xp_timer_interrupt(int irq, void *dev_id) >> /* >> * ACK timer interrupt and call event handler. >> */ >> - struct clock_event_device *evt = *(struct clock_event_device **)dev_id; >> + struct clock_event_device *evt = dev_id; >> >> writel(TIMER0_CLR_MASK, local_base + LCL_TIMER_EVENTS_STATUS); >> evt->event_handler(evt); >> > >
On 03/20/13 10:44, Gregory CLEMENT wrote: > On 03/20/2013 06:26 PM, Gregory CLEMENT wrote: >> On 03/20/2013 06:20 PM, Stephen Boyd wrote: >>> On 03/20/13 10:09, Gregory CLEMENT wrote: >>>> On 03/13/2013 07:17 PM, Stephen Boyd wrote: >>>>> Separate the armada 370xp local timers from the local timer API. >>>>> This will allow us to remove ARM local timer support in the near >>>>> future and makes this driver multi-architecture friendly. >>>> At first view the code looks good, but when I applied your patch set on >>>> linux-next, build it and run it on a Armada XP based board (AX3 with 2 cores), >>>> it crashed: >>> [...] >>>> Internal error: Oops - undefined instruction: 0 [#1] SMP ARM >>>> Modules linked in: >>>> CPU: 0 Not tainted (3.9.0-rc3-next-20130319-00010-g728b448 #153) >>>> PC is at 0xe92d45f0 >>>> LR is at armada_370_xp_timer_interrupt+0x3c/0x4c >>>> pc : [<e92d45f0>] lr : [<c023c2bc>] psr: 600001d3 >>>> sp : c0635eb8 ip : 00000000 fp : c063c3f0 >>>> r10: 000003ff r9 : 00000000 r8 : 00000010 >>>> r7 : c22b3c40 r6 : ef007c00 r5 : c0640fcc r4 : c0053e30 >>>> r3 : e92d45f0 r2 : fffffffe r1 : c22b3c40 r0 : c0053e30 >>>> Flags: nZCv IRQs off FIQs off Mode SVC_32 ISA ARM Segment kernel >>>> Control: 10c53c7d Table: 0000406a DAC: 00000015 >>>> Process swapper/0 (pid: 0, stack limit = 0xc0634238) >>>> Stack: (0xc0635eb8 to 0xc0636000) >>>> 5ea0: ef004c80 c0063224 >>>> 5ec0: 00000010 00000010 00000000 c0660ac0 c0635f18 c005fcb8 c0632b90 c000ed94 >>>> 5ee0: c0313c60 60000153 00000001 c00085a8 c0313c54 c0313c60 60000153 ffffffff >>>> 5f00: c0635f4c 00000000 562f5842 c06360c0 00000000 c000db60 0000001a ffff8ad0 >>>> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >>>> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >>>> 5f60: 00000021 00000000 00000003 00000004 0000006e c065fcc0 c067924c c063ceb8 >>>> 5f80: c063cc84 c006d8c0 00000005 c065fcc0 c067924c c0421764 c22ad780 c063c42c >>>> 5fa0: 562f5842 c063cca8 c06605c0 c04379c0 c22ad780 00000000 562f5842 00000000 >>>> 5fc0: 00000000 c0417754 ffffffff ffffffff c04172dc 00000000 00000000 c04379c0 >>>> 5fe0: 10c53c7d c063c414 c04379bc c063febc 0000406a 00008074 00000000 00000000 >>>> [<c023c2bc>] (armada_370_xp_timer_interrupt+0x3c/0x4c) from [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) >>>> [<c0063224>] (handle_percpu_devid_irq+0x64/0x80) from [<c005fcb8>] (generic_handle_irq+0x20/0x30) >>>> [<c005fcb8>] (generic_handle_irq+0x20/0x30) from [<c000ed94>] (handle_IRQ+0x38/0x90) >>>> [<c000ed94>] (handle_IRQ+0x38/0x90) from [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) >>>> [<c00085a8>] (armada_370_xp_handle_irq+0xa4/0xb0) from [<c000db60>] (__irq_svc+0x40/0x50) >>>> Exception stack(0xc0635f18 to 0xc0635f60) >>>> 5f00: 0000001a ffff8ad0 >>>> 5f20: ffff8ad0 c06360c0 00000000 00000000 c04379c0 c22ad780 00000000 562f5842 >>>> 5f40: c06360c0 00000000 60000153 c0635f60 c0313c54 c0313c60 60000153 ffffffff >>>> [<c000db60>] (__irq_svc+0x40/0x50) from [<c0313c60>] (calibrate_delay+0x378/0x528) >>>> [<c0313c60>] (calibrate_delay+0x378/0x528) from [<c0417754>] (start_kernel+0x250/0x2dc) >>>> [<c0417754>] (start_kernel+0x250/0x2dc) from [<00008074>] (0x8074) >>>> Code: 1fe7deb7 cd5772dd fff5692e ed55f79e (7ed5a5f7) >>>> >>>> >>>> >>>> I am trying to figure out what happened. >>> Argh. Stupid casting again. Can you try this? >> Our emails must have crossed, your fix is also fine :) >> > I have also tested the patch set on a Armada 370 based board (Armada 370 DB), > and it works well. > So with this last chunk you can add my: > Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com> > and also my > tested-by: Gregory CLEMENT <gregory.clement@free-electrons.com> > Great. Thanks for testing.
diff --git a/drivers/clocksource/time-armada-370-xp.c b/drivers/clocksource/time-armada-370-xp.c index ee2e50c5..a9bf37a 100644 --- a/drivers/clocksource/time-armada-370-xp.c +++ b/drivers/clocksource/time-armada-370-xp.c @@ -150,7 +150,7 @@ static irqreturn_t armada_370_xp_timer_interrupt(int irq, void *dev_id) /* * ACK timer interrupt and call event handler. */ - struct clock_event_device *evt = *(struct clock_event_device **)dev_id; + struct clock_event_device *evt = dev_id; writel(TIMER0_CLR_MASK, local_base + LCL_TIMER_EVENTS_STATUS); evt->event_handler(evt);