2019-03-29 09:03:39

by Xiaoming Ni

[permalink] [raw]
Subject: [PATCH] clk:Fix divide by 0 error in divider_ro_round_rate_parent

In the function divider_recalc_rate The judgment of the return value of
_get_div indicates that the return value of _get_div may be 0.
In order to avoid the divide-by-zero error, add check the return value
of _get_div in the divider_ro_round_rate_parent

Signed-off-by: nixiaoming <[email protected]>
---
drivers/clk/clk-divider.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/clk/clk-divider.c b/drivers/clk/clk-divider.c
index e5a1726..0854e3e 100644
--- a/drivers/clk/clk-divider.c
+++ b/drivers/clk/clk-divider.c
@@ -347,6 +347,8 @@ long divider_ro_round_rate_parent(struct clk_hw *hw, struct clk_hw *parent,
int div;

div = _get_div(table, val, flags, width);
+ if (!div) /* avoid divide-by-zero */
+ return -EINVAL;

/* Even a read-only clock can propagate a rate change */
if (clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT) {
--
1.8.5.6



2019-03-29 22:42:56

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH] clk:Fix divide by 0 error in divider_ro_round_rate_parent

Quoting nixiaoming (2019-03-29 02:05:24)
> In the function divider_recalc_rate The judgment of the return value of

Please write divider_recalc_rate() with parenthesis to show it's a
function.

> _get_div indicates that the return value of _get_div may be 0.

__get_div()

> In order to avoid the divide-by-zero error, add check the return value
> of _get_div in the divider_ro_round_rate_parent
>
> Signed-off-by: nixiaoming <[email protected]>

Is this your name? nixiaoming? Or is it written some other way?

> ---
> drivers/clk/clk-divider.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/clk/clk-divider.c b/drivers/clk/clk-divider.c
> index e5a1726..0854e3e 100644
> --- a/drivers/clk/clk-divider.c
> +++ b/drivers/clk/clk-divider.c
> @@ -347,6 +347,8 @@ long divider_ro_round_rate_parent(struct clk_hw *hw, struct clk_hw *parent,
> int div;
>
> div = _get_div(table, val, flags, width);
> + if (!div) /* avoid divide-by-zero */
> + return -EINVAL;

How does _get_div() return 0? What is the value of 'flags' here when
this goes wrong?

>
> /* Even a read-only clock can propagate a rate change */
> if (clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT) {
> --
> 1.8.5.6
>

Wow that's a 5 year old version of git!


2019-03-30 02:04:53

by Xiaoming Ni

[permalink] [raw]
Subject: RE: [PATCH] clk:Fix divide by 0 error in divider_ro_round_rate_parent

On 3/30/2019 6:42 AM Stephen Boyd wrote:
>Quoting nixiaoming (2019-03-29 02:05:24)
>> In the function divider_recalc_rate The judgment of the return value of
>
>Please write divider_recalc_rate() with parenthesis to show it's a
>function.
>
>> _get_div indicates that the return value of _get_div may be 0.
>
>__get_div()
Thank you for your guidance, I will correct it later in the patch.

>> In order to avoid the divide-by-zero error, add check the return value
>> of _get_div in the divider_ro_round_rate_parent
>>
>> Signed-off-by: nixiaoming <[email protected]>
>
>Is this your name? nixiaoming? Or is it written some other way?

Chinese name 倪小明
>> ---
>> drivers/clk/clk-divider.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/clk/clk-divider.c b/drivers/clk/clk-divider.c
>> index e5a1726..0854e3e 100644
>> --- a/drivers/clk/clk-divider.c
>> +++ b/drivers/clk/clk-divider.c
>> @@ -347,6 +347,8 @@ long divider_ro_round_rate_parent(struct clk_hw *hw, struct clk_hw *parent,
>> int div;
>>
>> div = _get_div(table, val, flags, width);
>> + if (!div) /* avoid divide-by-zero */
>> + return -EINVAL;
>
>How does _get_div() return 0? What is the value of 'flags' here when
>this goes wrong?

divider_ro_round_rate_parent() and divider_recalc_rate() are functions
of the EXPORT_SYMBOL_GPL attribute

If _get_div() can return 0 in the argument of divider_recalc_rate()
Then should be able to return 0 in divider_ro_round_rate_parent()

>
>>
>> /* Even a read-only clock can propagate a rate change */
>> if (clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT) {
>> --
>> 1.8.5.6
>>
>
>Wow that's a 5 year old version of git!
>
>