The the __freq_reg_info() never returns NULL and the callers don't check
for NULL. This initialization to set "reg_rule = NULL;" is just there
to make GCC happy but it's not required in current GCCs.
The problem is that Smatch sees the initialization and concludes that
this function can return NULL so it complains that the callers are not
checking for it.
Smatch used to be able to parse this correctly but we recently changed
the code from:
- for (bw = MHZ_TO_KHZ(20); bw >= min_bw; bw = bw / 2) {
+ for (bw = MHZ_TO_KHZ(bws[i]); bw >= min_bw; bw = MHZ_TO_KHZ(bws[i--])) {
Originally Smatch used to understand that this code always iterates
through the loop once, but the change from "MHZ_TO_KHZ(20)" to
"MHZ_TO_KHZ(bws[i])" is too complicated for Smatch.
Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
---
Feel free to throw this patch into the bit bucket if you want... I
don't feel strongly about it either way. Just a piece of functional
art.
net/wireless/reg.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index 6043a9d33d61..3dab859641e1 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -1616,8 +1616,8 @@ static const struct ieee80211_reg_rule *
__freq_reg_info(struct wiphy *wiphy, u32 center_freq, u32 min_bw)
{
const struct ieee80211_regdomain *regd = reg_get_regdomain(wiphy);
- const struct ieee80211_reg_rule *reg_rule = NULL;
const u32 bws[] = {0, 1, 2, 4, 5, 8, 10, 16, 20};
+ const struct ieee80211_reg_rule *reg_rule;
int i = ARRAY_SIZE(bws) - 1;
u32 bw;
--
2.28.0
