[PATCH 1/2] powerpc: Fix a memory leak in error handling paths

If we exit the for_each_of_cpu_node loop early, the reference on the
current node must be decremented, otherwise there is a leak.

Fixes: a94fe366340a ("powerpc: use for_each_of_cpu_node iterator")
Signed-off-by: Christophe JAILLET <redacted>
---
Strangely, the commit above added the needed of_node_put in one place but
missed 2 other places!
This is strange, so maybe I misunderstand something. Review carefully
---
 arch/powerpc/platforms/powermac/feature.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/powerpc/platforms/powermac/feature.c b/arch/powerpc/platforms/powermac/feature.c
index 5c77b9a24c0e..e612222f7d2e 100644
--- a/arch/powerpc/platforms/powermac/feature.c
+++ b/arch/powerpc/platforms/powermac/feature.c

@@ -1060,6 +1060,7 @@ core99_reset_cpu(struct device_node *node, long param, long value)
 			continue;
 		if (param == *num) {
 			reset_io = *rst;
+			of_node_put(np);
 			break;
 		}
 	}

@@ -1506,6 +1507,7 @@ static long g5_reset_cpu(struct device_node *node, long param, long value)
 			continue;
 		if (param == *num) {
 			reset_io = *rst;
+			of_node_put(np);
 			break;
 		}
 	}

-- 
2.27.0

'arch/powerpc/platforms/powermac/feature.c' triggers many checkpatch.pl
warnings.
The code looks old and not very active, so fixing them all does not make
so much sense and will hide some 'git blame' information.

So only apply a few fixes that save a few lines of code.

Signed-off-by: Christophe JAILLET <redacted>
---
 arch/powerpc/platforms/powermac/feature.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/platforms/powermac/feature.c b/arch/powerpc/platforms/powermac/feature.c
index e612222f7d2e..155e8586692e 100644
--- a/arch/powerpc/platforms/powermac/feature.c
+++ b/arch/powerpc/platforms/powermac/feature.c

@@ -83,8 +83,7 @@ struct macio_chip *macio_find(struct device_node *child, int type)
 }
 EXPORT_SYMBOL_GPL(macio_find);
 
-static const char *macio_names[] =
-{
+static const char *macio_names[] = {
 	"Unknown",
 	"Grand Central",
 	"OHare",

@@ -119,8 +118,7 @@ struct feature_table_entry {
 	feature_call	function;
 };
 
-struct pmac_mb_def
-{
+struct pmac_mb_def {
 	const char*			model_string;
 	const char*			model_name;
 	int				model_id;

@@ -301,11 +299,10 @@ static long ohare_sleep_state(struct device_node *node, long param, long value)
 
 	if ((pmac_mb.board_flags & PMAC_MB_CAN_SLEEP) == 0)
 		return -EPERM;
-	if (value == 1) {
+	if (value == 1)
 		MACIO_BIC(OHARE_FCR, OH_IOBUS_ENABLE);
-	} else if (value == 0) {
+	else if (value == 0)
 		MACIO_BIS(OHARE_FCR, OH_IOBUS_ENABLE);
-	}
 
 	return 0;
 }

@@ -2992,7 +2989,6 @@ void pmac_register_agp_pm(struct pci_dev *bridge,
 	if (bridge != pmac_agp_bridge)
 		return;
 	pmac_agp_suspend = pmac_agp_resume = NULL;
-	return;
 }
 EXPORT_SYMBOL(pmac_register_agp_pm);
 

-- 
2.27.0

On Fri, Apr 23, 2021 at 9:40 AM Christophe JAILLET
[off-list ref] wrote:

If we exit the for_each_of_cpu_node loop early, the reference on the
current node must be decremented, otherwise there is a leak.

Fixes: a94fe366340a ("powerpc: use for_each_of_cpu_node iterator")
Signed-off-by: Christophe JAILLET <redacted>
---
Strangely, the commit above added the needed of_node_put in one place but
missed 2 other places!

Well, maintained it in one place and forgot to add in the other two.

This is strange, so maybe I misunderstand something. Review carefully
---
 arch/powerpc/platforms/powermac/feature.c | 2 ++
 1 file changed, 2 insertions(+)

Reviewed-by: Rob Herring <robh@kernel.org>

I don't really think patch 2 is worthwhile but that's up to the
powerpc maintainers.

Rob