* [PATCH 0/4] add missing of_node_put
@ 2015-10-22 9:02 Julia Lawall
2015-10-22 9:02 ` [PATCH 1/4] of/overlay: " Julia Lawall
` (6 more replies)
0 siblings, 7 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 9:02 UTC (permalink / raw)
To: Rob Herring
Cc: kernel-janitors, Frank Rowand, Grant Likely, devicetree, linux-kernel
The various for_each device_node iterators performs an of_node_get on each
iteration, so a break out of the loop requires an of_node_put.
The complete semantic patch that fixes this problem is
(http://coccinelle.lip6.fr):
// <smpl>
@r@
local idexpression n;
expression e1,e2;
iterator name for_each_node_by_name, for_each_node_by_type,
for_each_compatible_node, for_each_matching_node,
for_each_matching_node_and_match, for_each_child_of_node,
for_each_available_child_of_node, for_each_node_with_property;
iterator i;
statement S;
expression list [n1] es;
@@
(
(
for_each_node_by_name(n,e1) S
|
for_each_node_by_type(n,e1) S
|
for_each_compatible_node(n,e1,e2) S
|
for_each_matching_node(n,e1) S
|
for_each_matching_node_and_match(n,e1,e2) S
|
for_each_child_of_node(e1,n) S
|
for_each_available_child_of_node(e1,n) S
|
for_each_node_with_property(n,e1) S
)
&
i(es,n,...) S
)
@@
local idexpression r.n;
iterator r.i;
expression e;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
return n;
|
+ of_node_put(n);
? return ...;
)
...
}
@@
local idexpression r.n;
iterator r.i;
expression e;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? break;
)
...
}
... when != n
@@
local idexpression r.n;
iterator r.i;
expression e;
identifier l;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? goto l;
)
...
}
...
l: ... when != n// </smpl>
---
drivers/of/irq.c | 7 +++++--
drivers/of/overlay.c | 5 ++++-
drivers/of/platform.c | 8 ++++++--
drivers/of/unittest.c | 8 ++++++--
4 files changed, 21 insertions(+), 7 deletions(-)
^ permalink raw reply [flat|nested] 23+ messages in thread
* [PATCH 1/4] of/overlay: add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
@ 2015-10-22 9:02 ` Julia Lawall
2015-10-22 9:02 ` [PATCH 2/4] of/platform: " Julia Lawall
` (5 subsequent siblings)
6 siblings, 0 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 9:02 UTC (permalink / raw)
To: Pantelis Antoniou
Cc: kernel-janitors, Rob Herring, Frank Rowand, Grant Likely,
devicetree, linux-kernel
for_each_child_of_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/of/overlay.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/of/overlay.c b/drivers/of/overlay.c
index 24e025f..54e5af9 100644
--- a/drivers/of/overlay.c
+++ b/drivers/of/overlay.c
@@ -149,6 +149,7 @@ static int of_overlay_apply_one(struct of_overlay *ov,
pr_err("%s: Failed to apply single node @%s/%s\n",
__func__, target->full_name,
child->name);
+ of_node_put(child);
return ret;
}
}
@@ -417,8 +418,10 @@ static int overlay_subtree_check(struct device_node *tree,
return 1;
for_each_child_of_node(tree, child) {
- if (overlay_subtree_check(child, dn))
+ if (overlay_subtree_check(child, dn)) {
+ of_node_put(child);
return 1;
+ }
}
return 0;
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 2/4] of/platform: add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
2015-10-22 9:02 ` [PATCH 1/4] of/overlay: " Julia Lawall
@ 2015-10-22 9:02 ` Julia Lawall
[not found] ` <1445504571-19838-1-git-send-email-Julia.Lawall-L2FTfq7BK8M@public.gmane.org>
` (4 subsequent siblings)
6 siblings, 0 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 9:02 UTC (permalink / raw)
To: Rob Herring
Cc: kernel-janitors, Frank Rowand, Grant Likely, devicetree, linux-kernel
for_each_child_of_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
local idexpression n;
expression root,e;
@@
for_each_child_of_node(root,n) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? break;
)
...
}
... when != n
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/of/platform.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index 7b33e03..64daf05 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -409,8 +409,10 @@ int of_platform_bus_probe(struct device_node *root,
if (!of_match_node(matches, child))
continue;
rc = of_platform_bus_create(child, matches, NULL, parent, false);
- if (rc)
+ if (rc) {
+ of_node_put(child);
break;
+ }
}
of_node_put(root);
@@ -451,8 +453,10 @@ int of_platform_populate(struct device_node *root,
for_each_child_of_node(root, child) {
rc = of_platform_bus_create(child, matches, lookup, parent, true);
- if (rc)
+ if (rc) {
+ of_node_put(child);
break;
+ }
}
of_node_set_flag(root, OF_POPULATED_BUS);
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 3/4] of/unittest: add missing of_node_put
[not found] ` <1445504571-19838-1-git-send-email-Julia.Lawall-L2FTfq7BK8M@public.gmane.org>
@ 2015-10-22 9:02 ` Julia Lawall
0 siblings, 0 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 9:02 UTC (permalink / raw)
To: Rob Herring
Cc: kernel-janitors-u79uwXL29TY76Z2rM5mHXA, Frank Rowand,
Grant Likely, devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA
for_each_child_of_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
Combine the puts into code at the end of the function, for conciseness.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/of/unittest.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
index 9f71770..e16ea57 100644
--- a/drivers/of/unittest.c
+++ b/drivers/of/unittest.c
@@ -205,16 +205,20 @@ static int __init of_unittest_check_node_linkage(struct device_node *np)
if (child->parent != np) {
pr_err("Child node %s links to wrong parent %s\n",
child->name, np->name);
- return -EINVAL;
+ rc = -EINVAL;
+ goto put_child;
}
rc = of_unittest_check_node_linkage(child);
if (rc < 0)
- return rc;
+ goto put_child;
count += rc;
}
return count + 1;
+put_child:
+ of_node_put(child);
+ return rc;
}
static void __init of_unittest_check_tree_linkage(void)
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 4/4] of/irq: add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
` (2 preceding siblings ...)
[not found] ` <1445504571-19838-1-git-send-email-Julia.Lawall-L2FTfq7BK8M@public.gmane.org>
@ 2015-10-22 9:02 ` Julia Lawall
2015-10-22 13:04 ` [PATCH 0/4] " Rob Herring
` (2 subsequent siblings)
6 siblings, 0 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 9:02 UTC (permalink / raw)
To: Rob Herring
Cc: kernel-janitors, Frank Rowand, Grant Likely, devicetree, linux-kernel
for_each_matching_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
local idexpression n;
expression e;
identifier l;
@@
for_each_matching_node(n,...) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? goto l;
)
...
}
...
l: ... when != n
// </smpl>
Besides the issue found by the semantic patch, this code also stores the
device_node value in a list, which requires an of_node_get, and then cleans
up the list on exit from the function, which requires an of_node_put.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/of/irq.c | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/of/irq.c b/drivers/of/irq.c
index 171798d..902b89b 100644
--- a/drivers/of/irq.c
+++ b/drivers/of/irq.c
@@ -501,10 +501,12 @@ void __init of_irq_init(const struct of_device_id *matches)
* pointer, interrupt-parent device_node etc.
*/
desc = kzalloc(sizeof(*desc), GFP_KERNEL);
- if (WARN_ON(!desc))
+ if (WARN_ON(!desc)) {
+ of_node_put(np);
goto err;
+ }
- desc->dev = np;
+ desc->dev = of_node_get(np);
desc->interrupt_parent = of_irq_find_parent(np);
if (desc->interrupt_parent = np)
desc->interrupt_parent = NULL;
@@ -575,6 +577,7 @@ void __init of_irq_init(const struct of_device_id *matches)
err:
list_for_each_entry_safe(desc, temp_desc, &intc_desc_list, list) {
list_del(&desc->list);
+ of_node_put(desc->dev);
kfree(desc);
}
}
^ permalink raw reply related [flat|nested] 23+ messages in thread
* Re: [PATCH 0/4] add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
` (3 preceding siblings ...)
2015-10-22 9:02 ` [PATCH 4/4] of/irq: " Julia Lawall
@ 2015-10-22 13:04 ` Rob Herring
[not found] ` <CAL_JsqJGS+L=FZAwPs-wBSYiX8kQWqZyaLTGhadYVNT3=Qte_Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2015-10-22 15:08 ` Rob Herring
2015-11-18 22:04 ` Julia Lawall
6 siblings, 1 reply; 23+ messages in thread
From: Rob Herring @ 2015-10-22 13:04 UTC (permalink / raw)
To: Julia Lawall, Russell King - ARM Linux
Cc: kernel-janitors, Frank Rowand, Grant Likely, devicetree, linux-kernel
+Russell who raised issues with these iterators recently.
On Thu, Oct 22, 2015 at 4:02 AM, Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> The various for_each device_node iterators performs an of_node_get on each
> iteration, so a break out of the loop requires an of_node_put.
Thanks for this.
Are there any plans to use coccinelle as a checker that developers can
run to be warned of these problems?
Rob
> The complete semantic patch that fixes this problem is
> (http://coccinelle.lip6.fr):
>
> // <smpl>
> @r@
> local idexpression n;
> expression e1,e2;
> iterator name for_each_node_by_name, for_each_node_by_type,
> for_each_compatible_node, for_each_matching_node,
> for_each_matching_node_and_match, for_each_child_of_node,
> for_each_available_child_of_node, for_each_node_with_property;
> iterator i;
> statement S;
> expression list [n1] es;
> @@
>
> (
> (
> for_each_node_by_name(n,e1) S
> |
> for_each_node_by_type(n,e1) S
> |
> for_each_compatible_node(n,e1,e2) S
> |
> for_each_matching_node(n,e1) S
> |
> for_each_matching_node_and_match(n,e1,e2) S
> |
> for_each_child_of_node(e1,n) S
> |
> for_each_available_child_of_node(e1,n) S
> |
> for_each_node_with_property(n,e1) S
> )
> &
> i(es,n,...) S
> )
>
> @@
> local idexpression r.n;
> iterator r.i;
> expression e;
> expression list [r.n1] es;
> @@
>
> i(es,n,...) {
> ...
> (
> of_node_put(n);
> |
> e = n
> |
> return n;
> |
> + of_node_put(n);
> ? return ...;
> )
> ...
> }
>
> @@
> local idexpression r.n;
> iterator r.i;
> expression e;
> expression list [r.n1] es;
> @@
>
> i(es,n,...) {
> ...
> (
> of_node_put(n);
> |
> e = n
> |
> + of_node_put(n);
> ? break;
> )
> ...
> }
> ... when != n
>
> @@
> local idexpression r.n;
> iterator r.i;
> expression e;
> identifier l;
> expression list [r.n1] es;
> @@
>
> i(es,n,...) {
> ...
> (
> of_node_put(n);
> |
> e = n
> |
> + of_node_put(n);
> ? goto l;
> )
> ...
> }
> ...
> l: ... when != n// </smpl>
>
> ---
>
> drivers/of/irq.c | 7 +++++--
> drivers/of/overlay.c | 5 ++++-
> drivers/of/platform.c | 8 ++++++--
> drivers/of/unittest.c | 8 ++++++--
> 4 files changed, 21 insertions(+), 7 deletions(-)
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 0/4] add missing of_node_put
[not found] ` <CAL_JsqJGS+L=FZAwPs-wBSYiX8kQWqZyaLTGhadYVNT3=Qte_Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
@ 2015-10-22 13:19 ` Julia Lawall
0 siblings, 0 replies; 23+ messages in thread
From: Julia Lawall @ 2015-10-22 13:19 UTC (permalink / raw)
To: Rob Herring
Cc: Russell King - ARM Linux, kernel-janitors-u79uwXL29TY76Z2rM5mHXA,
Frank Rowand, Grant Likely, devicetree-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA
On Thu, 22 Oct 2015, Rob Herring wrote:
> +Russell who raised issues with these iterators recently.
>
> On Thu, Oct 22, 2015 at 4:02 AM, Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> > The various for_each device_node iterators performs an of_node_get on each
> > iteration, so a break out of the loop requires an of_node_put.
>
> Thanks for this.
>
> Are there any plans to use coccinelle as a checker that developers can
> run to be warned of these problems?
There are three issues I have found so far:
1. of_node_put before a continue
2. missing of_node_put before a break
3. missing of_node get before some more permanent storage of the value
The semantic patches for the first two seem to be working pretty well, so
I could submit them for integration into the kernel. The third one, I
have only looked at in an ad hoc way so far.
julia
>
> Rob
>
> > The complete semantic patch that fixes this problem is
> > (http://coccinelle.lip6.fr):
> >
> > // <smpl>
> > @r@
> > local idexpression n;
> > expression e1,e2;
> > iterator name for_each_node_by_name, for_each_node_by_type,
> > for_each_compatible_node, for_each_matching_node,
> > for_each_matching_node_and_match, for_each_child_of_node,
> > for_each_available_child_of_node, for_each_node_with_property;
> > iterator i;
> > statement S;
> > expression list [n1] es;
> > @@
> >
> > (
> > (
> > for_each_node_by_name(n,e1) S
> > |
> > for_each_node_by_type(n,e1) S
> > |
> > for_each_compatible_node(n,e1,e2) S
> > |
> > for_each_matching_node(n,e1) S
> > |
> > for_each_matching_node_and_match(n,e1,e2) S
> > |
> > for_each_child_of_node(e1,n) S
> > |
> > for_each_available_child_of_node(e1,n) S
> > |
> > for_each_node_with_property(n,e1) S
> > )
> > &
> > i(es,n,...) S
> > )
> >
> > @@
> > local idexpression r.n;
> > iterator r.i;
> > expression e;
> > expression list [r.n1] es;
> > @@
> >
> > i(es,n,...) {
> > ...
> > (
> > of_node_put(n);
> > |
> > e = n
> > |
> > return n;
> > |
> > + of_node_put(n);
> > ? return ...;
> > )
> > ...
> > }
> >
> > @@
> > local idexpression r.n;
> > iterator r.i;
> > expression e;
> > expression list [r.n1] es;
> > @@
> >
> > i(es,n,...) {
> > ...
> > (
> > of_node_put(n);
> > |
> > e = n
> > |
> > + of_node_put(n);
> > ? break;
> > )
> > ...
> > }
> > ... when != n
> >
> > @@
> > local idexpression r.n;
> > iterator r.i;
> > expression e;
> > identifier l;
> > expression list [r.n1] es;
> > @@
> >
> > i(es,n,...) {
> > ...
> > (
> > of_node_put(n);
> > |
> > e = n
> > |
> > + of_node_put(n);
> > ? goto l;
> > )
> > ...
> > }
> > ...
> > l: ... when != n// </smpl>
> >
> > ---
> >
> > drivers/of/irq.c | 7 +++++--
> > drivers/of/overlay.c | 5 ++++-
> > drivers/of/platform.c | 8 ++++++--
> > drivers/of/unittest.c | 8 ++++++--
> > 4 files changed, 21 insertions(+), 7 deletions(-)
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 0/4] add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
` (4 preceding siblings ...)
2015-10-22 13:04 ` [PATCH 0/4] " Rob Herring
@ 2015-10-22 15:08 ` Rob Herring
2015-11-18 22:04 ` Julia Lawall
6 siblings, 0 replies; 23+ messages in thread
From: Rob Herring @ 2015-10-22 15:08 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, Frank Rowand, Grant Likely, devicetree, linux-kernel
On Thu, Oct 22, 2015 at 4:02 AM, Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> The various for_each device_node iterators performs an of_node_get on each
> iteration, so a break out of the loop requires an of_node_put.
> ---
>
> drivers/of/irq.c | 7 +++++--
> drivers/of/overlay.c | 5 ++++-
> drivers/of/platform.c | 8 ++++++--
> drivers/of/unittest.c | 8 ++++++--
> 4 files changed, 21 insertions(+), 7 deletions(-)
Series applied, thanks.
Rob
^ permalink raw reply [flat|nested] 23+ messages in thread
* [PATCH 0/4] add missing of_node_put
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
` (5 preceding siblings ...)
2015-10-22 15:08 ` Rob Herring
@ 2015-11-18 22:04 ` Julia Lawall
2015-11-18 22:04 ` [PATCH 1/4] mtd: brcmnand: improve memory management Julia Lawall
` (3 more replies)
6 siblings, 4 replies; 23+ messages in thread
From: Julia Lawall @ 2015-11-18 22:04 UTC (permalink / raw)
To: linux-arm-kernel
The various for_each device_node iterators performs an of_node_get on each
iteration, so a break out of the loop requires an of_node_put.
The complete semantic patch that fixes this problem is
(http://coccinelle.lip6.fr):
// <smpl>
@r@
local idexpression n;
expression e1,e2;
iterator name for_each_node_by_name, for_each_node_by_type,
for_each_compatible_node, for_each_matching_node,
for_each_matching_node_and_match, for_each_child_of_node,
for_each_available_child_of_node, for_each_node_with_property;
iterator i;
statement S;
expression list [n1] es;
@@
(
(
for_each_node_by_name(n,e1) S
|
for_each_node_by_type(n,e1) S
|
for_each_compatible_node(n,e1,e2) S
|
for_each_matching_node(n,e1) S
|
for_each_matching_node_and_match(n,e1,e2) S
|
for_each_child_of_node(e1,n) S
|
for_each_available_child_of_node(e1,n) S
|
for_each_node_with_property(n,e1) S
)
&
i(es,n,...) S
)
@@
local idexpression r.n;
iterator r.i;
expression e;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
return n;
|
+ of_node_put(n);
? return ...;
)
...
}
@@
local idexpression r.n;
iterator r.i;
expression e;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? break;
)
...
}
... when != n
@@
local idexpression r.n;
iterator r.i;
expression e;
identifier l;
expression list [r.n1] es;
@@
i(es,n,...) {
...
(
of_node_put(n);
|
e = n
|
+ of_node_put(n);
? goto l;
)
...
}
...
l: ... when != n// </smpl>
---
drivers/iio/adc/qcom-spmi-vadc.c | 4 +++-
drivers/mtd/nand/brcmnand/brcmnand.c | 14 ++++++++++----
drivers/mtd/nand/sunxi_nand.c | 4 +++-
drivers/power/reset/at91-reset.c | 1 +
4 files changed, 17 insertions(+), 6 deletions(-)
^ permalink raw reply [flat|nested] 23+ messages in thread
* [PATCH 1/4] mtd: brcmnand: improve memory management
2015-11-18 22:04 ` Julia Lawall
@ 2015-11-18 22:04 ` Julia Lawall
2015-11-18 22:44 ` Brian Norris
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
` (2 subsequent siblings)
3 siblings, 1 reply; 23+ messages in thread
From: Julia Lawall @ 2015-11-18 22:04 UTC (permalink / raw)
To: Brian Norris
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper
This patch addresses several related memory management issues in the probe
function:
1. for_each_available_child_of_node performs an of_node_get on each
iteration, so a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_available_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
2. The devm_kzalloc'd data is not used if brcmnand_init_cs fails. Free it
immediately, using devm_kfree in this case, instead of waiting for the
remove function.
3. If the continue is not taken, then host is added to a list, that has a
lifetime beyond the end of the for_each_available_child_of_node loop body.
Thus, of_node_get is needed on child, which is referenced by host. A
corresponding of_node_put is needed in the remove function.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
One could consider whether the of_node_get should be on host->of_node,
which looks more similar to the thing that is stored in the list. I used
child, to be more similar to the of_node_put in the same function.
drivers/mtd/nand/brcmnand/brcmnand.c | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
index 2a437c7..b0cb55d 100644
--- a/drivers/mtd/nand/brcmnand/brcmnand.c
+++ b/drivers/mtd/nand/brcmnand/brcmnand.c
@@ -2237,16 +2237,20 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
struct brcmnand_host *host;
host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
- if (!host)
+ if (!host) {
+ of_node_put(child);
return -ENOMEM;
+ }
host->pdev = pdev;
host->ctrl = ctrl;
host->of_node = child;
ret = brcmnand_init_cs(host);
- if (ret)
+ if (ret) {
+ devm_kfree(dev, host);
continue; /* Try all chip-selects */
-
+ }
+ of_node_get(child);
list_add_tail(&host->node, &ctrl->host_list);
}
}
@@ -2264,8 +2268,10 @@ int brcmnand_remove(struct platform_device *pdev)
struct brcmnand_controller *ctrl = dev_get_drvdata(&pdev->dev);
struct brcmnand_host *host;
- list_for_each_entry(host, &ctrl->host_list, node)
+ list_for_each_entry(host, &ctrl->host_list, node) {
+ of_node_put(host->of_node);
nand_release(&host->mtd);
+ }
dev_set_drvdata(&pdev->dev, NULL);
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put
2015-11-18 22:04 ` Julia Lawall
2015-11-18 22:04 ` [PATCH 1/4] mtd: brcmnand: improve memory management Julia Lawall
@ 2015-11-18 22:04 ` Julia Lawall
2015-11-23 7:26 ` Chen-Yu Tsai
` (2 more replies)
2015-11-18 22:04 ` [PATCH 3/4] iio: adc: spmi-vadc: " Julia Lawall
2015-11-18 22:04 ` [PATCH 4/4] power/reset: at91-reset: " Julia Lawall
3 siblings, 3 replies; 23+ messages in thread
From: Julia Lawall @ 2015-11-18 22:04 UTC (permalink / raw)
To: linux-arm-kernel
for_each_child_of_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/mtd/nand/sunxi_nand.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/mtd/nand/sunxi_nand.c b/drivers/mtd/nand/sunxi_nand.c
index 2ed52e4..1bbcc0c 100644
--- a/drivers/mtd/nand/sunxi_nand.c
+++ b/drivers/mtd/nand/sunxi_nand.c
@@ -1391,8 +1391,10 @@ static int sunxi_nand_chips_init(struct device *dev, struct sunxi_nfc *nfc)
for_each_child_of_node(np, nand_np) {
ret = sunxi_nand_chip_init(dev, nfc, nand_np);
- if (ret)
+ if (ret) {
+ of_node_put(nand_np);
return ret;
+ }
}
return 0;
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 3/4] iio: adc: spmi-vadc: add missing of_node_put
2015-11-18 22:04 ` Julia Lawall
2015-11-18 22:04 ` [PATCH 1/4] mtd: brcmnand: improve memory management Julia Lawall
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
@ 2015-11-18 22:04 ` Julia Lawall
2015-11-21 18:25 ` Jonathan Cameron
2015-11-18 22:04 ` [PATCH 4/4] power/reset: at91-reset: " Julia Lawall
3 siblings, 1 reply; 23+ messages in thread
From: Julia Lawall @ 2015-11-18 22:04 UTC (permalink / raw)
To: Jonathan Cameron
Cc: kernel-janitors, Hartmut Knaack, Lars-Peter Clausen,
Peter Meerwald, linux-iio, linux-kernel,
Russell King - ARM Linux, Thomas Petazzoni, Andrew Lunn,
Bjorn Helgaas, Jason Cooper
for_each_available_child_of_node performs an of_node_get on each iteration,
so a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_available_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/iio/adc/qcom-spmi-vadc.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/iio/adc/qcom-spmi-vadc.c b/drivers/iio/adc/qcom-spmi-vadc.c
index 0c4618b..c2babe5 100644
--- a/drivers/iio/adc/qcom-spmi-vadc.c
+++ b/drivers/iio/adc/qcom-spmi-vadc.c
@@ -839,8 +839,10 @@ static int vadc_get_dt_data(struct vadc_priv *vadc, struct device_node *node)
for_each_available_child_of_node(node, child) {
ret = vadc_get_dt_channel_data(vadc->dev, &prop, child);
- if (ret)
+ if (ret) {
+ of_node_put(child);
return ret;
+ }
vadc->chan_props[index] = prop;
^ permalink raw reply related [flat|nested] 23+ messages in thread
* [PATCH 4/4] power/reset: at91-reset: add missing of_node_put
2015-11-18 22:04 ` Julia Lawall
` (2 preceding siblings ...)
2015-11-18 22:04 ` [PATCH 3/4] iio: adc: spmi-vadc: " Julia Lawall
@ 2015-11-18 22:04 ` Julia Lawall
2015-12-05 0:42 ` Sebastian Reichel
3 siblings, 1 reply; 23+ messages in thread
From: Julia Lawall @ 2015-11-18 22:04 UTC (permalink / raw)
To: Sebastian Reichel
Cc: kernel-janitors, Dmitry Eremin-Solenikov, David Woodhouse,
linux-pm, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper
for_each_matching_node performs an of_node_get on each iteration, so
a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression e,e1;
local idexpression np;
@@
for_each_matching_node(np, e1) {
... when != of_node_put(np)
when != e = np
(
return np;
|
+ of_node_put(np);
? return ...;
)
...
}
// </smpl>
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
drivers/power/reset/at91-reset.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/power/reset/at91-reset.c b/drivers/power/reset/at91-reset.c
index 3f6b5dd..1b5d450 100644
--- a/drivers/power/reset/at91-reset.c
+++ b/drivers/power/reset/at91-reset.c
@@ -198,6 +198,7 @@ static int __init at91_reset_probe(struct platform_device *pdev)
at91_ramc_base[idx] = of_iomap(np, 0);
if (!at91_ramc_base[idx]) {
dev_err(&pdev->dev, "Could not map ram controller address\n");
+ of_node_put(np);
return -ENODEV;
}
idx++;
^ permalink raw reply related [flat|nested] 23+ messages in thread
* Re: [PATCH 1/4] mtd: brcmnand: improve memory management
2015-11-18 22:04 ` [PATCH 1/4] mtd: brcmnand: improve memory management Julia Lawall
@ 2015-11-18 22:44 ` Brian Norris
2015-11-19 6:13 ` Julia Lawall
0 siblings, 1 reply; 23+ messages in thread
From: Brian Norris @ 2015-11-18 22:44 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper,
Kamal Dasu
On Wed, Nov 18, 2015 at 11:04:11PM +0100, Julia Lawall wrote:
> This patch addresses several related memory management issues in the probe
> function:
>
> 1. for_each_available_child_of_node performs an of_node_get on each
> iteration, so a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_available_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
Good catch again
> 2. The devm_kzalloc'd data is not used if brcmnand_init_cs fails. Free it
> immediately, using devm_kfree in this case, instead of waiting for the
> remove function.
Same
> 3. If the continue is not taken, then host is added to a list, that has a
> lifetime beyond the end of the for_each_available_child_of_node loop body.
> Thus, of_node_get is needed on child, which is referenced by host. A
> corresponding of_node_put is needed in the remove function.
This one's a bit silly. We really shouldn't be keeping the reference in
'host' at all. Also, as of commit 215a02fd3087 ("mtd: grab a reference to
the MTD of_node before registering it"), the MTD core will actually be
refcounting the node for us, too, so this isn't really necessary.
I have a patch to remove brcmnand_host::of_node (appended below), which
should make this step obsolete, and be more obvious that no extra
of_node_get()'ing is required.
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
>
> ---
>
> One could consider whether the of_node_get should be on host->of_node,
> which looks more similar to the thing that is stored in the list. I used
> child, to be more similar to the of_node_put in the same function.
>
> drivers/mtd/nand/brcmnand/brcmnand.c | 14 ++++++++++----
> 1 file changed, 10 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
> index 2a437c7..b0cb55d 100644
> --- a/drivers/mtd/nand/brcmnand/brcmnand.c
> +++ b/drivers/mtd/nand/brcmnand/brcmnand.c
> @@ -2237,16 +2237,20 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
> struct brcmnand_host *host;
>
> host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
> - if (!host)
> + if (!host) {
> + of_node_put(child);
> return -ENOMEM;
In code reading, I noticed that we don't actually cleanup for prior
iterations of brcmnand_init_cs() here. i.e., if we're exiting here, we
should be doing nand_release() on all previously-registered chips.
> + }
> host->pdev = pdev;
> host->ctrl = ctrl;
> host->of_node = child;
>
> ret = brcmnand_init_cs(host);
> - if (ret)
> + if (ret) {
> + devm_kfree(dev, host);
> continue; /* Try all chip-selects */
> -
> + }
> + of_node_get(child);
> list_add_tail(&host->node, &ctrl->host_list);
> }
> }
> @@ -2264,8 +2268,10 @@ int brcmnand_remove(struct platform_device *pdev)
> struct brcmnand_controller *ctrl = dev_get_drvdata(&pdev->dev);
> struct brcmnand_host *host;
>
> - list_for_each_entry(host, &ctrl->host_list, node)
> + list_for_each_entry(host, &ctrl->host_list, node) {
> + of_node_put(host->of_node);
> nand_release(&host->mtd);
> + }
>
> dev_set_drvdata(&pdev->dev, NULL);
>
Patch to kill off some of this:
---8<---
From 6c51a9ef1325e7b06a7623c1fbca1adf6eeb8253 Mon Sep 17 00:00:00 2001
From: Brian Norris <computersforpeace@gmail.com>
Date: Wed, 18 Nov 2015 14:33:24 -0800
Subject: [PATCH] mtd: brcmnand: drop brcmnand_host::of_node field
We don't actually need to stash a copy of this device_node indefinitely;
we only need it in brcmnand_init_cs().
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Cc: <bcm-kernel-feedback-list@broadcom.com>
Cc: Kamal Dasu <kdasu.kdev@gmail.com>
---
drivers/mtd/nand/brcmnand/brcmnand.c | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
index c395b4a75fb1..351438a62aaa 100644
--- a/drivers/mtd/nand/brcmnand/brcmnand.c
+++ b/drivers/mtd/nand/brcmnand/brcmnand.c
@@ -176,7 +176,6 @@ struct brcmnand_cfg {
struct brcmnand_host {
struct list_head node;
- struct device_node *of_node;
struct nand_chip chip;
struct mtd_info mtd;
@@ -1896,10 +1895,9 @@ static int brcmnand_setup_dev(struct brcmnand_host *host)
return 0;
}
-static int brcmnand_init_cs(struct brcmnand_host *host)
+static int brcmnand_init_cs(struct brcmnand_host *host, struct device_node *dn)
{
struct brcmnand_controller *ctrl = host->ctrl;
- struct device_node *dn = host->of_node;
struct platform_device *pdev = host->pdev;
struct mtd_info *mtd;
struct nand_chip *chip;
@@ -2231,9 +2229,8 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
return -ENOMEM;
host->pdev = pdev;
host->ctrl = ctrl;
- host->of_node = child;
- ret = brcmnand_init_cs(host);
+ ret = brcmnand_init_cs(host, child);
if (ret)
continue; /* Try all chip-selects */
--
2.6.0.rc2.230.g3dd15c0
^ permalink raw reply related [flat|nested] 23+ messages in thread
* Re: [PATCH 1/4] mtd: brcmnand: improve memory management
2015-11-18 22:44 ` Brian Norris
@ 2015-11-19 6:13 ` Julia Lawall
2015-11-19 19:26 ` Brian Norris
0 siblings, 1 reply; 23+ messages in thread
From: Julia Lawall @ 2015-11-19 6:13 UTC (permalink / raw)
To: Brian Norris
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper,
Kamal Dasu
On Wed, 18 Nov 2015, Brian Norris wrote:
> On Wed, Nov 18, 2015 at 11:04:11PM +0100, Julia Lawall wrote:
> > This patch addresses several related memory management issues in the probe
> > function:
> >
> > 1. for_each_available_child_of_node performs an of_node_get on each
> > iteration, so a break out of the loop requires an of_node_put.
> >
> > A simplified version of the semantic patch that fixes this problem is as
> > follows (http://coccinelle.lip6.fr):
> >
> > // <smpl>
> > @@
> > expression root,e;
> > local idexpression child;
> > @@
> >
> > for_each_available_child_of_node(root, child) {
> > ... when != of_node_put(child)
> > when != e = child
> > (
> > return child;
> > |
> > + of_node_put(child);
> > ? return ...;
> > )
> > ...
> > }
> > // </smpl>
>
> Good catch again
>
> > 2. The devm_kzalloc'd data is not used if brcmnand_init_cs fails. Free it
> > immediately, using devm_kfree in this case, instead of waiting for the
> > remove function.
>
> Same
>
> > 3. If the continue is not taken, then host is added to a list, that has a
> > lifetime beyond the end of the for_each_available_child_of_node loop body.
> > Thus, of_node_get is needed on child, which is referenced by host. A
> > corresponding of_node_put is needed in the remove function.
>
> This one's a bit silly. We really shouldn't be keeping the reference in
> 'host' at all. Also, as of commit 215a02fd3087 ("mtd: grab a reference to
> the MTD of_node before registering it"), the MTD core will actually be
> refcounting the node for us, too, so this isn't really necessary.
>
> I have a patch to remove brcmnand_host::of_node (appended below), which
> should make this step obsolete, and be more obvious that no extra
> of_node_get()'ing is required.
OK. Should I resend my patch without this change?
julia
> > Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
> >
> > ---
> >
> > One could consider whether the of_node_get should be on host->of_node,
> > which looks more similar to the thing that is stored in the list. I used
> > child, to be more similar to the of_node_put in the same function.
> >
> > drivers/mtd/nand/brcmnand/brcmnand.c | 14 ++++++++++----
> > 1 file changed, 10 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
> > index 2a437c7..b0cb55d 100644
> > --- a/drivers/mtd/nand/brcmnand/brcmnand.c
> > +++ b/drivers/mtd/nand/brcmnand/brcmnand.c
> > @@ -2237,16 +2237,20 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
> > struct brcmnand_host *host;
> >
> > host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
> > - if (!host)
> > + if (!host) {
> > + of_node_put(child);
> > return -ENOMEM;
>
> In code reading, I noticed that we don't actually cleanup for prior
> iterations of brcmnand_init_cs() here. i.e., if we're exiting here, we
> should be doing nand_release() on all previously-registered chips.
>
> > + }
> > host->pdev = pdev;
> > host->ctrl = ctrl;
> > host->of_node = child;
> >
> > ret = brcmnand_init_cs(host);
> > - if (ret)
> > + if (ret) {
> > + devm_kfree(dev, host);
> > continue; /* Try all chip-selects */
> > -
> > + }
> > + of_node_get(child);
> > list_add_tail(&host->node, &ctrl->host_list);
> > }
> > }
> > @@ -2264,8 +2268,10 @@ int brcmnand_remove(struct platform_device *pdev)
> > struct brcmnand_controller *ctrl = dev_get_drvdata(&pdev->dev);
> > struct brcmnand_host *host;
> >
> > - list_for_each_entry(host, &ctrl->host_list, node)
> > + list_for_each_entry(host, &ctrl->host_list, node) {
> > + of_node_put(host->of_node);
> > nand_release(&host->mtd);
> > + }
> >
> > dev_set_drvdata(&pdev->dev, NULL);
> >
>
> Patch to kill off some of this:
>
> ---8<---
> From 6c51a9ef1325e7b06a7623c1fbca1adf6eeb8253 Mon Sep 17 00:00:00 2001
> From: Brian Norris <computersforpeace@gmail.com>
> Date: Wed, 18 Nov 2015 14:33:24 -0800
> Subject: [PATCH] mtd: brcmnand: drop brcmnand_host::of_node field
>
> We don't actually need to stash a copy of this device_node indefinitely;
> we only need it in brcmnand_init_cs().
>
> Signed-off-by: Brian Norris <computersforpeace@gmail.com>
> Cc: <bcm-kernel-feedback-list@broadcom.com>
> Cc: Kamal Dasu <kdasu.kdev@gmail.com>
> ---
> drivers/mtd/nand/brcmnand/brcmnand.c | 7 ++-----
> 1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
> index c395b4a75fb1..351438a62aaa 100644
> --- a/drivers/mtd/nand/brcmnand/brcmnand.c
> +++ b/drivers/mtd/nand/brcmnand/brcmnand.c
> @@ -176,7 +176,6 @@ struct brcmnand_cfg {
>
> struct brcmnand_host {
> struct list_head node;
> - struct device_node *of_node;
>
> struct nand_chip chip;
> struct mtd_info mtd;
> @@ -1896,10 +1895,9 @@ static int brcmnand_setup_dev(struct brcmnand_host *host)
> return 0;
> }
>
> -static int brcmnand_init_cs(struct brcmnand_host *host)
> +static int brcmnand_init_cs(struct brcmnand_host *host, struct device_node *dn)
> {
> struct brcmnand_controller *ctrl = host->ctrl;
> - struct device_node *dn = host->of_node;
> struct platform_device *pdev = host->pdev;
> struct mtd_info *mtd;
> struct nand_chip *chip;
> @@ -2231,9 +2229,8 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
> return -ENOMEM;
> host->pdev = pdev;
> host->ctrl = ctrl;
> - host->of_node = child;
>
> - ret = brcmnand_init_cs(host);
> + ret = brcmnand_init_cs(host, child);
> if (ret)
> continue; /* Try all chip-selects */
>
> --
> 2.6.0.rc2.230.g3dd15c0
>
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 1/4] mtd: brcmnand: improve memory management
2015-11-19 6:13 ` Julia Lawall
@ 2015-11-19 19:26 ` Brian Norris
2015-11-19 21:32 ` [PATCH 1/4 v2] " Julia Lawall
0 siblings, 1 reply; 23+ messages in thread
From: Brian Norris @ 2015-11-19 19:26 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper,
Kamal Dasu
On Thu, Nov 19, 2015 at 07:13:45AM +0100, Julia Lawall wrote:
> On Wed, 18 Nov 2015, Brian Norris wrote:
> > On Wed, Nov 18, 2015 at 11:04:11PM +0100, Julia Lawall wrote:
> > > 3. If the continue is not taken, then host is added to a list, that has a
> > > lifetime beyond the end of the for_each_available_child_of_node loop body.
> > > Thus, of_node_get is needed on child, which is referenced by host. A
> > > corresponding of_node_put is needed in the remove function.
> >
> > This one's a bit silly. We really shouldn't be keeping the reference in
> > 'host' at all. Also, as of commit 215a02fd3087 ("mtd: grab a reference to
> > the MTD of_node before registering it"), the MTD core will actually be
> > refcounting the node for us, too, so this isn't really necessary.
> >
> > I have a patch to remove brcmnand_host::of_node (appended below), which
> > should make this step obsolete, and be more obvious that no extra
> > of_node_get()'ing is required.
>
> OK. Should I resend my patch without this change?
Sure, that'd be good. Then I could merge/rebase mine on top.
Thanks,
Brian
^ permalink raw reply [flat|nested] 23+ messages in thread
* [PATCH 1/4 v2] mtd: brcmnand: improve memory management
2015-11-19 19:26 ` Brian Norris
@ 2015-11-19 21:32 ` Julia Lawall
2015-12-01 1:56 ` Brian Norris
0 siblings, 1 reply; 23+ messages in thread
From: Julia Lawall @ 2015-11-19 21:32 UTC (permalink / raw)
To: Brian Norris
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper,
Kamal Dasu
This patch addresses two related memory management issues in the probe
function:
1. for_each_available_child_of_node performs an of_node_get on each
iteration, so a break out of the loop requires an of_node_put.
A simplified version of the semantic patch that fixes this problem is as
follows (http://coccinelle.lip6.fr):
// <smpl>
@@
expression root,e;
local idexpression child;
@@
for_each_available_child_of_node(root, child) {
... when != of_node_put(child)
when != e = child
(
return child;
|
+ of_node_put(child);
? return ...;
)
...
}
// </smpl>
2. The devm_kzalloc'd data is not used if brcmnand_init_cs fails. Free it
immediately, using devm_kfree in this case, instead of waiting for the
remove function.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
---
v2: Drop of_node_get addition, as another patch will get rid of the need
for it completely.
drivers/mtd/nand/brcmnand/brcmnand.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c
index 2a437c7..551d474 100644
--- a/drivers/mtd/nand/brcmnand/brcmnand.c
+++ b/drivers/mtd/nand/brcmnand/brcmnand.c
@@ -2237,15 +2237,19 @@ int brcmnand_probe(struct platform_device *pdev, struct brcmnand_soc *soc)
struct brcmnand_host *host;
host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
- if (!host)
+ if (!host) {
+ of_node_put(child);
return -ENOMEM;
+ }
host->pdev = pdev;
host->ctrl = ctrl;
host->of_node = child;
ret = brcmnand_init_cs(host);
- if (ret)
+ if (ret) {
+ devm_kfree(dev, host);
continue; /* Try all chip-selects */
+ }
list_add_tail(&host->node, &ctrl->host_list);
}
^ permalink raw reply related [flat|nested] 23+ messages in thread
* Re: [PATCH 3/4] iio: adc: spmi-vadc: add missing of_node_put
2015-11-18 22:04 ` [PATCH 3/4] iio: adc: spmi-vadc: " Julia Lawall
@ 2015-11-21 18:25 ` Jonathan Cameron
0 siblings, 0 replies; 23+ messages in thread
From: Jonathan Cameron @ 2015-11-21 18:25 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, Hartmut Knaack, Lars-Peter Clausen,
Peter Meerwald, linux-iio, linux-kernel,
Russell King - ARM Linux, Thomas Petazzoni, Andrew Lunn,
Bjorn Helgaas, Jason Cooper
On 18/11/15 22:04, Julia Lawall wrote:
> for_each_available_child_of_node performs an of_node_get on each iteration,
> so a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_available_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Thanks. Applied to the fixes-togreg branch of iio.git and marked for stable.
Jonathan
>
> ---
> drivers/iio/adc/qcom-spmi-vadc.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iio/adc/qcom-spmi-vadc.c b/drivers/iio/adc/qcom-spmi-vadc.c
> index 0c4618b..c2babe5 100644
> --- a/drivers/iio/adc/qcom-spmi-vadc.c
> +++ b/drivers/iio/adc/qcom-spmi-vadc.c
> @@ -839,8 +839,10 @@ static int vadc_get_dt_data(struct vadc_priv *vadc, struct device_node *node)
>
> for_each_available_child_of_node(node, child) {
> ret = vadc_get_dt_channel_data(vadc->dev, &prop, child);
> - if (ret)
> + if (ret) {
> + of_node_put(child);
> return ret;
> + }
>
> vadc->chan_props[index] = prop;
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-iio" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
@ 2015-11-23 7:26 ` Chen-Yu Tsai
2015-11-23 13:59 ` Boris Brezillon
2015-12-01 1:56 ` Brian Norris
2 siblings, 0 replies; 23+ messages in thread
From: Chen-Yu Tsai @ 2015-11-23 7:26 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Nov 19, 2015 at 6:04 AM, Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> for_each_child_of_node performs an of_node_get on each iteration, so
> a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Chen-Yu Tsai <wens@csie.org>
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
2015-11-23 7:26 ` Chen-Yu Tsai
@ 2015-11-23 13:59 ` Boris Brezillon
2015-12-01 1:56 ` Brian Norris
2 siblings, 0 replies; 23+ messages in thread
From: Boris Brezillon @ 2015-11-23 13:59 UTC (permalink / raw)
To: linux-arm-kernel
On Wed, 18 Nov 2015 23:04:12 +0100
Julia Lawall <Julia.Lawall@lip6.fr> wrote:
> for_each_child_of_node performs an of_node_get on each iteration, so
> a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Thanks,
Boris
--
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 1/4 v2] mtd: brcmnand: improve memory management
2015-11-19 21:32 ` [PATCH 1/4 v2] " Julia Lawall
@ 2015-12-01 1:56 ` Brian Norris
0 siblings, 0 replies; 23+ messages in thread
From: Brian Norris @ 2015-12-01 1:56 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, David Woodhouse, linux-mtd,
bcm-kernel-feedback-list, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper,
Kamal Dasu
On Thu, Nov 19, 2015 at 10:32:15PM +0100, Julia Lawall wrote:
> This patch addresses two related memory management issues in the probe
> function:
>
> 1. for_each_available_child_of_node performs an of_node_get on each
> iteration, so a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_available_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> 2. The devm_kzalloc'd data is not used if brcmnand_init_cs fails. Free it
> immediately, using devm_kfree in this case, instead of waiting for the
> remove function.
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
>
> ---
>
> v2: Drop of_node_get addition, as another patch will get rid of the need
> for it completely.
Pushed to l2-mtd.git
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
2015-11-23 7:26 ` Chen-Yu Tsai
2015-11-23 13:59 ` Boris Brezillon
@ 2015-12-01 1:56 ` Brian Norris
2 siblings, 0 replies; 23+ messages in thread
From: Brian Norris @ 2015-12-01 1:56 UTC (permalink / raw)
To: linux-arm-kernel
On Wed, Nov 18, 2015 at 11:04:12PM +0100, Julia Lawall wrote:
> for_each_child_of_node performs an of_node_get on each iteration, so
> a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression root,e;
> local idexpression child;
> @@
>
> for_each_child_of_node(root, child) {
> ... when != of_node_put(child)
> when != e = child
> (
> return child;
> |
> + of_node_put(child);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
>
Pushed to l2-mtd.git
^ permalink raw reply [flat|nested] 23+ messages in thread
* Re: [PATCH 4/4] power/reset: at91-reset: add missing of_node_put
2015-11-18 22:04 ` [PATCH 4/4] power/reset: at91-reset: " Julia Lawall
@ 2015-12-05 0:42 ` Sebastian Reichel
0 siblings, 0 replies; 23+ messages in thread
From: Sebastian Reichel @ 2015-12-05 0:42 UTC (permalink / raw)
To: Julia Lawall
Cc: kernel-janitors, Dmitry Eremin-Solenikov, David Woodhouse,
linux-pm, linux-kernel, Russell King - ARM Linux,
Thomas Petazzoni, Andrew Lunn, Bjorn Helgaas, Jason Cooper
[-- Attachment #1: Type: text/plain, Size: 685 bytes --]
Hi,
On Wed, Nov 18, 2015 at 11:04:14PM +0100, Julia Lawall wrote:
> for_each_matching_node performs an of_node_get on each iteration, so
> a break out of the loop requires an of_node_put.
>
> A simplified version of the semantic patch that fixes this problem is as
> follows (http://coccinelle.lip6.fr):
>
> // <smpl>
> @@
> expression e,e1;
> local idexpression np;
> @@
>
> for_each_matching_node(np, e1) {
> ... when != of_node_put(np)
> when != e = np
> (
> return np;
> |
> + of_node_put(np);
> ? return ...;
> )
> ...
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Thanks, queued.
-- Sebastian
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
^ permalink raw reply [flat|nested] 23+ messages in thread
end of thread, other threads:[~2015-12-05 0:42 UTC | newest]
Thread overview: 23+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-10-22 9:02 [PATCH 0/4] add missing of_node_put Julia Lawall
2015-10-22 9:02 ` [PATCH 1/4] of/overlay: " Julia Lawall
2015-10-22 9:02 ` [PATCH 2/4] of/platform: " Julia Lawall
[not found] ` <1445504571-19838-1-git-send-email-Julia.Lawall-L2FTfq7BK8M@public.gmane.org>
2015-10-22 9:02 ` [PATCH 3/4] of/unittest: " Julia Lawall
2015-10-22 9:02 ` [PATCH 4/4] of/irq: " Julia Lawall
2015-10-22 13:04 ` [PATCH 0/4] " Rob Herring
[not found] ` <CAL_JsqJGS+L=FZAwPs-wBSYiX8kQWqZyaLTGhadYVNT3=Qte_Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2015-10-22 13:19 ` Julia Lawall
2015-10-22 15:08 ` Rob Herring
2015-11-18 22:04 ` Julia Lawall
2015-11-18 22:04 ` [PATCH 1/4] mtd: brcmnand: improve memory management Julia Lawall
2015-11-18 22:44 ` Brian Norris
2015-11-19 6:13 ` Julia Lawall
2015-11-19 19:26 ` Brian Norris
2015-11-19 21:32 ` [PATCH 1/4 v2] " Julia Lawall
2015-12-01 1:56 ` Brian Norris
2015-11-18 22:04 ` [PATCH 2/4] mtd: nand: sunxi: add missing of_node_put Julia Lawall
2015-11-23 7:26 ` Chen-Yu Tsai
2015-11-23 13:59 ` Boris Brezillon
2015-12-01 1:56 ` Brian Norris
2015-11-18 22:04 ` [PATCH 3/4] iio: adc: spmi-vadc: " Julia Lawall
2015-11-21 18:25 ` Jonathan Cameron
2015-11-18 22:04 ` [PATCH 4/4] power/reset: at91-reset: " Julia Lawall
2015-12-05 0:42 ` Sebastian Reichel
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).