From: Frank Rowand <[email protected]>
Fix boot warning 'Trying to vfree() nonexistent vm area'
from arch_timer_mem_of_init().
Refactored code attempts to read and iounmap using address frame
instead of address ioremap(frame->cntbase).
Fixes: c389d701dfb70 ("clocksource: arm_arch_timer: split MMIO timer probing.")
Signed-off-by: Frank Rowand <[email protected]>
---
WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:1514 iounmap+0x14/0x18
Trying to vfree() nonexistent vm area (ee821000)
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.12.0-rc1-dirty #1
Hardware name: Generic DT based system
[<c030f574>] (unwind_backtrace) from [<c030b8e0>] (show_stack+0x10/0x14)
[<c030b8e0>] (show_stack) from [<c05f5204>] (dump_stack+0x6c/0x8c)
[<c05f5204>] (dump_stack) from [<c031dcdc>] (__warn+0xd0/0xf8)
[<c031dcdc>] (__warn) from [<c031ddac>] (warn_slowpath_fmt+0x38/0x48)
[<c031ddac>] (warn_slowpath_fmt) from [<c0315e78>] (iounmap+0x14/0x18)
[<c0315e78>] (iounmap) from [<c0d36e98>] (arch_timer_mem_of_init+0x224/0x414)
[<c0d36e98>] (arch_timer_mem_of_init) from [<c0d36868>] (clocksource_probe+0x44/0xa8)
[<c0d36868>] (clocksource_probe) from [<c0d00ae8>] (start_kernel+0x228/0x3a0)
[<c0d00ae8>] (start_kernel) from [<0020807c>] (0x20807c)
drivers/clocksource/arm_arch_timer.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
index 4bed671e490e..8b5c30062d99 100644
--- a/drivers/clocksource/arm_arch_timer.c
+++ b/drivers/clocksource/arm_arch_timer.c
@@ -1209,9 +1209,9 @@ static int __init arch_timer_of_init(struct device_node *np)
return 0;
}
- rate = readl_relaxed(frame + CNTFRQ);
+ rate = readl_relaxed(base + CNTFRQ);
- iounmap(frame);
+ iounmap(base);
return rate;
}
--
Frank Rowand <[email protected]>
Hi Frank,
On 10 June 2017 at 08:26, <[email protected]> wrote:
> From: Frank Rowand <[email protected]>
>
> Fix boot warning 'Trying to vfree() nonexistent vm area'
> from arch_timer_mem_of_init().
>
> Refactored code attempts to read and iounmap using address frame
> instead of address ioremap(frame->cntbase).
>
> Fixes: c389d701dfb70 ("clocksource: arm_arch_timer: split MMIO timer probing.")
>
> Signed-off-by: Frank Rowand <[email protected]>
Reviewed-by: Fu Wei <[email protected]>
> ---
>
> WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:1514 iounmap+0x14/0x18
> Trying to vfree() nonexistent vm area (ee821000)
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.12.0-rc1-dirty #1
> Hardware name: Generic DT based system
> [<c030f574>] (unwind_backtrace) from [<c030b8e0>] (show_stack+0x10/0x14)
> [<c030b8e0>] (show_stack) from [<c05f5204>] (dump_stack+0x6c/0x8c)
> [<c05f5204>] (dump_stack) from [<c031dcdc>] (__warn+0xd0/0xf8)
> [<c031dcdc>] (__warn) from [<c031ddac>] (warn_slowpath_fmt+0x38/0x48)
> [<c031ddac>] (warn_slowpath_fmt) from [<c0315e78>] (iounmap+0x14/0x18)
> [<c0315e78>] (iounmap) from [<c0d36e98>] (arch_timer_mem_of_init+0x224/0x414)
> [<c0d36e98>] (arch_timer_mem_of_init) from [<c0d36868>] (clocksource_probe+0x44/0xa8)
> [<c0d36868>] (clocksource_probe) from [<c0d00ae8>] (start_kernel+0x228/0x3a0)
> [<c0d00ae8>] (start_kernel) from [<0020807c>] (0x20807c)
>
> drivers/clocksource/arm_arch_timer.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
> index 4bed671e490e..8b5c30062d99 100644
> --- a/drivers/clocksource/arm_arch_timer.c
> +++ b/drivers/clocksource/arm_arch_timer.c
> @@ -1209,9 +1209,9 @@ static int __init arch_timer_of_init(struct device_node *np)
> return 0;
> }
>
> - rate = readl_relaxed(frame + CNTFRQ);
> + rate = readl_relaxed(base + CNTFRQ);
>
> - iounmap(frame);
> + iounmap(base);
Great thanks for your patch, this is a bug. So sorry for this typo.
It happened in my last v24 patchset: https://lkml.org/lkml/2017/4/14/363
Hope this fix can be merged ASAP.
>
> return rate;
> }
> --
> Frank Rowand <[email protected]>
>
--
Best regards,
Fu Wei
Software Engineer
Red Hat
On Fri, Jun 09 2017 at 5:26:32 pm BST, [email protected] wrote:
> From: Frank Rowand <[email protected]>
>
> Fix boot warning 'Trying to vfree() nonexistent vm area'
> from arch_timer_mem_of_init().
>
> Refactored code attempts to read and iounmap using address frame
> instead of address ioremap(frame->cntbase).
>
> Fixes: c389d701dfb70 ("clocksource: arm_arch_timer: split MMIO timer probing.")
>
> Signed-off-by: Frank Rowand <[email protected]>
> ---
>
> WARNING: CPU: 0 PID: 0 at mm/vmalloc.c:1514 iounmap+0x14/0x18
> Trying to vfree() nonexistent vm area (ee821000)
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.12.0-rc1-dirty #1
> Hardware name: Generic DT based system
> [<c030f574>] (unwind_backtrace) from [<c030b8e0>] (show_stack+0x10/0x14)
> [<c030b8e0>] (show_stack) from [<c05f5204>] (dump_stack+0x6c/0x8c)
> [<c05f5204>] (dump_stack) from [<c031dcdc>] (__warn+0xd0/0xf8)
> [<c031dcdc>] (__warn) from [<c031ddac>] (warn_slowpath_fmt+0x38/0x48)
> [<c031ddac>] (warn_slowpath_fmt) from [<c0315e78>] (iounmap+0x14/0x18)
> [<c0315e78>] (iounmap) from [<c0d36e98>] (arch_timer_mem_of_init+0x224/0x414)
> [<c0d36e98>] (arch_timer_mem_of_init) from [<c0d36868>] (clocksource_probe+0x44/0xa8)
> [<c0d36868>] (clocksource_probe) from [<c0d00ae8>] (start_kernel+0x228/0x3a0)
> [<c0d00ae8>] (start_kernel) from [<0020807c>] (0x20807c)
>
> drivers/clocksource/arm_arch_timer.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
> index 4bed671e490e..8b5c30062d99 100644
> --- a/drivers/clocksource/arm_arch_timer.c
> +++ b/drivers/clocksource/arm_arch_timer.c
> @@ -1209,9 +1209,9 @@ static int __init arch_timer_of_init(struct device_node *np)
> return 0;
> }
>
> - rate = readl_relaxed(frame + CNTFRQ);
> + rate = readl_relaxed(base + CNTFRQ);
>
> - iounmap(frame);
> + iounmap(base);
>
> return rate;
> }
Ah, nice find.
Acked-by: Marc Zyngier <[email protected]>
Thomas, Daniel: Can you queue this for the next -rc please?
Thanks,
M.
--
Jazz is not dead, it just smell funny.
On 10/06/2017 11:49, Marc Zyngier wrote:
> On Fri, Jun 09 2017 at 5:26:32 pm BST, [email protected] wrote:
>> From: Frank Rowand <[email protected]>
>>
>> Fix boot warning 'Trying to vfree() nonexistent vm area'
>> from arch_timer_mem_of_init().
>>
>> Refactored code attempts to read and iounmap using address frame
>> instead of address ioremap(frame->cntbase).
>>
>> Fixes: c389d701dfb70 ("clocksource: arm_arch_timer: split MMIO timer probing.")
>>
>> Signed-off-by: Frank Rowand <[email protected]>
>> ---
[ ... ]
> Ah, nice find.
>
> Acked-by: Marc Zyngier <[email protected]>
>
> Thomas, Daniel: Can you queue this for the next -rc please?
Yep, applied to my tree.
Thanks.
-- Daniel
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog