This patch introduces the use of devm_kzalloc and does away
with kfree in the probe function.
Signed-off-by: Vaishali Thakkar <[email protected]>
---
Some comments on this patch and driver:
- Here, after this change there will be a kfree on the failure of
the regmap_read too. I am not sure if driver is currently missing
it or there is something else happening here. Also, I think there
should be soc_device_unregister under the regmap_read.
- Moreover, get_maintainer.pl is not pointing me towards any particular
mailing list or maintainers. So, I am sending this patch to all
commiters as there are only three commits for this driver.
---
drivers/soc/versatile/soc-realview.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/soc/versatile/soc-realview.c b/drivers/soc/versatile/soc-realview.c
index e642c45..0889710 100644
--- a/drivers/soc/versatile/soc-realview.c
+++ b/drivers/soc/versatile/soc-realview.c
@@ -103,7 +103,8 @@ static int realview_soc_probe(struct platform_device *pdev)
if (IS_ERR(syscon_regmap))
return PTR_ERR(syscon_regmap);
- soc_dev_attr = kzalloc(sizeof(*soc_dev_attr), GFP_KERNEL);
+ soc_dev_attr = devm_kzalloc(&pdev->dev, sizeof(*soc_dev_attr),
+ GFP_KERNEL);
if (!soc_dev_attr)
return -ENOMEM;
@@ -115,10 +116,9 @@ static int realview_soc_probe(struct platform_device *pdev)
soc_dev_attr->machine = "RealView";
soc_dev_attr->family = "Versatile";
soc_dev = soc_device_register(soc_dev_attr);
- if (IS_ERR(soc_dev)) {
- kfree(soc_dev_attr);
+ if (IS_ERR(soc_dev))
return -ENODEV;
- }
+
ret = regmap_read(syscon_regmap, REALVIEW_SYS_ID_OFFSET,
&realview_coreid);
if (ret)
--
1.9.1
On Mon, Jul 27, 2015 at 5:23 AM, Vaishali Thakkar
<[email protected]> wrote:
> This patch introduces the use of devm_kzalloc and does away
> with kfree in the probe function.
>
> Signed-off-by: Vaishali Thakkar <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
Please get this through ARM SoC or tell me if I should handle it.
> Some comments on this patch and driver:
> - Here, after this change there will be a kfree on the failure of
> the regmap_read too. I am not sure if driver is currently missing
> it or there is something else happening here. Also, I think there
> should be soc_device_unregister under the regmap_read.
Likely. There are sloppy errors everywhere.
> - Moreover, get_maintainer.pl is not pointing me towards any particular
> mailing list or maintainers. So, I am sending this patch to all
> commiters as there are only three commits for this driver.
This is my code, also my stupid bugs :/
Yours,
Linus Walleij
On Tue, Jul 28, 2015 at 4:17 PM, Vaishali Thakkar
<[email protected]> wrote:
> On 28 Jul 2015 17:35, "Linus Walleij" <[email protected]> wrote:
>> Reviewed-by: Linus Walleij <[email protected]>
>> Please get this through ARM SoC or tell me if I should handle it.
>
> So, do you want me to send this in ARM SoC mailing list?
[email protected] is preferred, if you want
the ARM SoC maintainers to apply the patch directly not
that it has my Review tag, send it to [email protected]
and ask them to apply it.
>> Likely. There are sloppy errors everywhere.
>
> Can I go for handling them?? I am not sure how much important this driver
> is.
Yes, and it's kind of important-ish since people look at
reference designs when doing other board support and the
realviews are also supported in QEMU making them kind
of popular.
Yours,
Linus Walleij
On Wed, Jul 29, 2015 at 2:00 PM, Linus Walleij <[email protected]> wrote:
> On Tue, Jul 28, 2015 at 4:17 PM, Vaishali Thakkar
> <[email protected]> wrote:
>> On 28 Jul 2015 17:35, "Linus Walleij" <[email protected]> wrote:
>
>>> Reviewed-by: Linus Walleij <[email protected]>
>>> Please get this through ARM SoC or tell me if I should handle it.
>>
>> So, do you want me to send this in ARM SoC mailing list?
>
> [email protected] is preferred, if you want
> the ARM SoC maintainers to apply the patch directly not
> that it has my Review tag, send it to [email protected]
> and ask them to apply it.
Ok. I will send this to [email protected]
with your Review tag.
>>> Likely. There are sloppy errors everywhere.
>>
>> Can I go for handling them?? I am not sure how much important this driver
>> is.
>
> Yes, and it's kind of important-ish since people look at
> reference designs when doing other board support and the
> realviews are also supported in QEMU making them kind
> of popular.
I see. Then I'll go for it.
Thank You.
> Yours,
> Linus Walleij
--
Vaishali