* [bug report] irqchip/mbigen: move to use bus_get_dev_root()
@ 2023-05-04 7:34 Dan Carpenter
2023-05-04 11:56 ` Kefeng Wang
0 siblings, 1 reply; 4+ messages in thread
From: Dan Carpenter @ 2023-05-04 7:34 UTC (permalink / raw)
To: gregkh; +Cc: Kefeng Wang, Hanjun Guo, Ma Jun, Ming Lei, kernel-janitors
Hello Greg Kroah-Hartman,
The patch fea087fc291b: "irqchip/mbigen: move to use
bus_get_dev_root()" from Mar 13, 2023, leads to the following Smatch
static checker warning:
drivers/irqchip/irq-mbigen.c:258 mbigen_of_create_domain()
error: potentially dereferencing uninitialized 'child'.
drivers/irqchip/irq-mbigen.c
235 static int mbigen_of_create_domain(struct platform_device *pdev,
236 struct mbigen_device *mgn_chip)
237 {
238 struct device *parent;
239 struct platform_device *child;
240 struct irq_domain *domain;
241 struct device_node *np;
242 u32 num_pins;
243
244 for_each_child_of_node(pdev->dev.of_node, np) {
245 if (!of_property_read_bool(np, "interrupt-controller"))
246 continue;
247
248 parent = bus_get_dev_root(&platform_bus_type);
249 if (parent) {
Smatch is concerned that "parent" can be NULL. Probably unlikely in
real life.
250 child = of_platform_device_create(np, NULL, parent);
251 put_device(parent);
252 if (!child) {
253 of_node_put(np);
254 return -ENOMEM;
255 }
256 }
257
--> 258 if (of_property_read_u32(child->dev.of_node, "num-pins",
259 &num_pins) < 0) {
260 dev_err(&pdev->dev, "No num-pins property\n");
261 of_node_put(np);
262 return -EINVAL;
263 }
264
265 domain = platform_msi_create_device_domain(&child->dev, num_pins,
266 mbigen_write_msg,
267 &mbigen_domain_ops,
268 mgn_chip);
269 if (!domain) {
270 of_node_put(np);
271 return -ENOMEM;
272 }
273 }
274
275 return 0;
276 }
regards,
dan carpenter
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [bug report] irqchip/mbigen: move to use bus_get_dev_root()
2023-05-04 7:34 [bug report] irqchip/mbigen: move to use bus_get_dev_root() Dan Carpenter
@ 2023-05-04 11:56 ` Kefeng Wang
2023-05-04 14:46 ` Dan Carpenter
0 siblings, 1 reply; 4+ messages in thread
From: Kefeng Wang @ 2023-05-04 11:56 UTC (permalink / raw)
To: Dan Carpenter, gregkh; +Cc: Hanjun Guo, Ma Jun, Ming Lei, kernel-janitors
On 2023/5/4 15:34, Dan Carpenter wrote:
> Hello Greg Kroah-Hartman,
>
> The patch fea087fc291b: "irqchip/mbigen: move to use
> bus_get_dev_root()" from Mar 13, 2023, leads to the following Smatch
> static checker warning:
>
> drivers/irqchip/irq-mbigen.c:258 mbigen_of_create_domain()
> error: potentially dereferencing uninitialized 'child'.
>
> drivers/irqchip/irq-mbigen.c
> 235 static int mbigen_of_create_domain(struct platform_device *pdev,
> 236 struct mbigen_device *mgn_chip)
> 237 {
> 238 struct device *parent;
> 239 struct platform_device *child;
> 240 struct irq_domain *domain;
> 241 struct device_node *np;
> 242 u32 num_pins;
> 243
> 244 for_each_child_of_node(pdev->dev.of_node, np) {
> 245 if (!of_property_read_bool(np, "interrupt-controller"))
> 246 continue;
> 247
> 248 parent = bus_get_dev_root(&platform_bus_type);
> 249 if (parent) {
>
> Smatch is concerned that "parent" can be NULL. Probably unlikely in
> real life.
How about move bus_get_dev_root() out of the loop.
diff --git a/drivers/irqchip/irq-mbigen.c b/drivers/irqchip/irq-mbigen.c
index eada5e0e3eb9..954a55ed1f2b 100644
--- a/drivers/irqchip/irq-mbigen.c
+++ b/drivers/irqchip/irq-mbigen.c
@@ -240,26 +240,25 @@ static int mbigen_of_create_domain(struct
platform_device *pdev,
struct irq_domain *domain;
struct device_node *np;
u32 num_pins;
+ int ret = -ENODEV;
+
+ parent = bus_get_dev_root(&platform_bus_type);
+ if (!parent)
+ return ret;
for_each_child_of_node(pdev->dev.of_node, np) {
if (!of_property_read_bool(np, "interrupt-controller"))
continue;
- parent = bus_get_dev_root(&platform_bus_type);
- if (parent) {
- child = of_platform_device_create(np, NULL, parent);
- put_device(parent);
- if (!child) {
- of_node_put(np);
- return -ENOMEM;
- }
- }
+ child = of_platform_device_create(np, NULL, parent);
+ if (!child)
+ goto out_put;
if (of_property_read_u32(child->dev.of_node, "num-pins",
&num_pins) < 0) {
dev_err(&pdev->dev, "No num-pins property\n");
- of_node_put(np);
- return -EINVAL;
+ ret = -EINVAL;
+ goto out_put;
}
domain = platform_msi_create_device_domain(&child->dev, num_pins,
@@ -267,12 +266,17 @@ static int mbigen_of_create_domain(struct
platform_device *pdev,
&mbigen_domain_ops,
mgn_chip);
if (!domain) {
- of_node_put(np);
- return -ENOMEM;
+ ret = -ENOMEM;
+ goto out_put;
}
}
return 0;
+
+out_put:
+ of_node_put(np);
+ put_device(parent);
+ return ret;
}
>
> regards,
> dan carpenter
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [bug report] irqchip/mbigen: move to use bus_get_dev_root()
2023-05-04 11:56 ` Kefeng Wang
@ 2023-05-04 14:46 ` Dan Carpenter
2023-05-05 8:55 ` Kefeng Wang
0 siblings, 1 reply; 4+ messages in thread
From: Dan Carpenter @ 2023-05-04 14:46 UTC (permalink / raw)
To: Kefeng Wang; +Cc: gregkh, Hanjun Guo, Ma Jun, Ming Lei, kernel-janitors
On Thu, May 04, 2023 at 07:56:07PM +0800, Kefeng Wang wrote:
>
>
> On 2023/5/4 15:34, Dan Carpenter wrote:
> > Hello Greg Kroah-Hartman,
> >
> > The patch fea087fc291b: "irqchip/mbigen: move to use
> > bus_get_dev_root()" from Mar 13, 2023, leads to the following Smatch
> > static checker warning:
> >
> > drivers/irqchip/irq-mbigen.c:258 mbigen_of_create_domain()
> > error: potentially dereferencing uninitialized 'child'.
> >
> > drivers/irqchip/irq-mbigen.c
> > 235 static int mbigen_of_create_domain(struct platform_device *pdev,
> > 236 struct mbigen_device *mgn_chip)
> > 237 {
> > 238 struct device *parent;
> > 239 struct platform_device *child;
> > 240 struct irq_domain *domain;
> > 241 struct device_node *np;
> > 242 u32 num_pins;
> > 243
> > 244 for_each_child_of_node(pdev->dev.of_node, np) {
> > 245 if (!of_property_read_bool(np, "interrupt-controller"))
> > 246 continue;
> > 247
> > 248 parent = bus_get_dev_root(&platform_bus_type);
> > 249 if (parent) {
> >
> > Smatch is concerned that "parent" can be NULL. Probably unlikely in
> > real life.
>
> How about move bus_get_dev_root() out of the loop.
>
Sounds good to me, but I don't know this code at all.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [bug report] irqchip/mbigen: move to use bus_get_dev_root()
2023-05-04 14:46 ` Dan Carpenter
@ 2023-05-05 8:55 ` Kefeng Wang
0 siblings, 0 replies; 4+ messages in thread
From: Kefeng Wang @ 2023-05-05 8:55 UTC (permalink / raw)
To: Dan Carpenter; +Cc: gregkh, Ming Lei, kernel-janitors, Hanjun Guo
On 2023/5/4 22:46, Dan Carpenter wrote:
> On Thu, May 04, 2023 at 07:56:07PM +0800, Kefeng Wang wrote:
>>
>>
>> On 2023/5/4 15:34, Dan Carpenter wrote:
>>> Hello Greg Kroah-Hartman,
>>>
>>> The patch fea087fc291b: "irqchip/mbigen: move to use
>>> bus_get_dev_root()" from Mar 13, 2023, leads to the following Smatch
>>> static checker warning:
>>>
>>> drivers/irqchip/irq-mbigen.c:258 mbigen_of_create_domain()
>>> error: potentially dereferencing uninitialized 'child'.
>>>
>>> drivers/irqchip/irq-mbigen.c
>>> 235 static int mbigen_of_create_domain(struct platform_device *pdev,
>>> 236 struct mbigen_device *mgn_chip)
>>> 237 {
>>> 238 struct device *parent;
>>> 239 struct platform_device *child;
>>> 240 struct irq_domain *domain;
>>> 241 struct device_node *np;
>>> 242 u32 num_pins;
>>> 243
>>> 244 for_each_child_of_node(pdev->dev.of_node, np) {
>>> 245 if (!of_property_read_bool(np, "interrupt-controller"))
>>> 246 continue;
>>> 247
>>> 248 parent = bus_get_dev_root(&platform_bus_type);
>>> 249 if (parent) {
>>>
>>> Smatch is concerned that "parent" can be NULL. Probably unlikely in
>>> real life.
>>
>> How about move bus_get_dev_root() out of the loop.
>>
>
> Sounds good to me, but I don't know this code at all.
Post a patch[1], let's wait for review, and could you help to test it,
man thanks.
[1]
https://lore.kernel.org/all/20230505090654.12793-1-wangkefeng.wang@huawei.com/
>
> regards,
> dan carpenter
>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2023-05-05 8:55 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-05-04 7:34 [bug report] irqchip/mbigen: move to use bus_get_dev_root() Dan Carpenter
2023-05-04 11:56 ` Kefeng Wang
2023-05-04 14:46 ` Dan Carpenter
2023-05-05 8:55 ` Kefeng Wang
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.