Skip to contentSkip to conversation

pstore: Tidy up an error check

PATCH In review Dan Carpenter 2 participants 3 messages Dec 2, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 pstore: Tidy up an error check

The crypto_alloc_comp() function never returns NULL, it returns error pointers on error.

---
 fs/pstore/platform.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/pstore/platform.c +1 -1
index 36714df37d5d..b7a2a2a31dee 100644
@@ -315,7 +315,7 @@ static void allocate_buf_for_compression(void)
 	}
 
 	ctx = crypto_alloc_comp(zbackend->name, 0, 0);
-	if (IS_ERR_OR_NULL(ctx)) {
+	if (IS_ERR(ctx)) {
 		kfree(buf);
 		pr_err("crypto_alloc_comp('%s') failed: %ld\n", zbackend->name,
 		       PTR_ERR(ctx));
Signature · 1 line
2.29.2
#1 lore raw
Quoted 4 lines · Dan Carpenter

The crypto_alloc_comp() function never returns NULL, it returns error pointers on error.

I replied to an identical patch yesterday, actually:
https://lore.kernel.org/lkml/202012011215.B9BF24A6D@keescook/

Using IS_ERR_OR_NULL() is more robust, and this isn't fast path, so I'd prefer to keep it that way.

-Kees

Quoted 20 lines
---
 fs/pstore/platform.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/pstore/platform.c +1 -1
index 36714df37d5d..b7a2a2a31dee 100644
@@ -315,7 +315,7 @@ static void allocate_buf_for_compression(void)
 	}
 
 	ctx = crypto_alloc_comp(zbackend->name, 0, 0);
-	if (IS_ERR_OR_NULL(ctx)) {
+	if (IS_ERR(ctx)) {
 		kfree(buf);
 		pr_err("crypto_alloc_comp('%s') failed: %ld\n", zbackend->name,
 		       PTR_ERR(ctx));
Signature · 1 line
2.29.2
Signature · 1 line
Kees Cook
#2 lore raw
Quoted 4 lines · Kees Cook
Quoted 3 lines · Dan Carpenter

The crypto_alloc_comp() function never returns NULL, it returns error pointers on error.

Quoted 8 lines
Quoted 1 line

I replied to an identical patch yesterday, actually:
https://lore.kernel.org/lkml/202012011215.B9BF24A6D@keescook/

Using IS_ERR_OR_NULL() is more robust, and this isn't fast path, so I'd prefer to keep it that way.

The NULL return doesn't make any sense though because crypto_alloc_comp() isn't optional... When a function returns both error pointers and NULLs then the NULL is special kind of success.

	p = get_feature();

If "p" is an error pointer that means an error happened. If "p" is NULL that means the feature is disabled in the .config or whatever. We can't return a valid pointer because the feature doesn't exist but it's also not an error so it doesn't return an error pointer. The code should not print a warning, maybe an info level printk at most. Then the driver should continue operating with the feature turned off.

Two of the callers for crypto_alloc_comp() check for error pointers and NULL and three only check for error pointers. It's inconsistent.

regards,
dan carpenter