H1, Change from V1 -> V2: Fix a potential null pointer dereference as reported by kbuild test robot <fengguang...@intel.com>.
Inorder to report the kzalloc failure, I have just used a pr_debug statement.
If it looks ugly, should I change the function's return value to int and give
out an -ENOMEM ?
From 6615c7a995a6bb4dcf4c1e9ac59bca823ec821a5 Mon Sep 17 00:00:00 2001
From: Aurabindo J <m...@aurabindo.in>
Date: Wed, 21 May 2014 11:18:10 +0530
Subject: [PATCH] mfd: abx500-core: Fix compiler warning larger stack frame
On systems with CONFIG_FRAME_WARN=1024, compiler warns the allocation of
an object of struct device on stack. Make the allocation dynamically to
fix the warning.
Signed-off-by: Aurabindo J <m...@aurabindo.in>
---
drivers/mfd/abx500-core.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/drivers/mfd/abx500-core.c b/drivers/mfd/abx500-core.c
index f3a15aa..5e5098d 100644
--- a/drivers/mfd/abx500-core.c
+++ b/drivers/mfd/abx500-core.c
@@ -154,16 +154,21 @@ EXPORT_SYMBOL(abx500_startup_irq_enabled);
void abx500_dump_all_banks(void)
{
struct abx500_ops *ops;
- struct device dummy_child = {NULL};
+ struct device *dummy_child;
struct abx500_device_entry *dev_entry;
- list_for_each_entry(dev_entry, &abx500_list, list) {
- dummy_child.parent = dev_entry->dev;
- ops = &dev_entry->ops;
+ dummy_child = kzalloc(sizeof(struct device), GFP_KERNEL);
+ if (dummy_child != NULL) {
+ list_for_each_entry(dev_entry, &abx500_list, list) {
+ dummy_child->parent = dev_entry->dev;
+ ops = &dev_entry->ops;
- if ((ops != NULL) && (ops->dump_all_banks != NULL))
- ops->dump_all_banks(&dummy_child);
- }
+ if ((ops != NULL) && (ops->dump_all_banks != NULL))
+ ops->dump_all_banks(dummy_child);
+ }
+ kfree(dummy_child);
+ } else
+ pr_debug("kzalloc failure in abx500_dump_all_banks\n");
}
EXPORT_SYMBOL(abx500_dump_all_banks);
--
1.9.1
signature.asc
Description: OpenPGP digital signature
