Skip to contentSkip to conversation

cxl: Add guest-specific code

bug report Applied Dan Carpenter 3 participants 4 messages Jul 14, 2016 – Jul 20, 2016 Char and misc drivers

lore
Share & tools
lore
#0 lore raw

Hello Christophe Lombard,

The patch 14baf4d9c739: "cxl: Add guest-specific code" from Mar 4, 2016, leads to the following static checker warning:

	drivers/misc/cxl/guest.c:1115 cxl_guest_init_adapter()
	error: we previously assumed 'adapter->guest' could be null (see line 1114)

drivers/misc/cxl/guest.c
1105 struct cxl *cxl_guest_init_adapter(struct device_node *np, struct platform_device *pdev) 1106 {

  1107          struct cxl *adapter;
  1108          bool free = true;
  1109          int rc;
  1110  
  1111          if (!(adapter = cxl_alloc_adapter()))
  1112                  return ERR_PTR(-ENOMEM);
  1113  
  1114          if (!(adapter->guest = kzalloc(sizeof(struct cxl_guest), GFP_KERNEL))) {
  1115                  free_adapter(adapter);
                        ^^^^^^^^^^^^^^^^^^^^^

We can't call free_adapter() if adapter->guest is NULL.

  1116                  return ERR_PTR(-ENOMEM);
  1117          }
	drivers/misc/cxl/guest.c:919 afu_properties_look_ok()
	warn: unsigned 'afu->crs_len' is never less than zero.

drivers/misc/cxl/guest.c
907 static int afu_properties_look_ok(struct cxl_afu *afu) 908 {
909 if (afu->pp_irqs < 0) {
910 dev_err(&afu->dev, "Unexpected per-process minimum interrupt value\n");
911 return -EINVAL;
912 }
913
914 if (afu->max_procs_virtualised < 1) {
915 dev_err(&afu->dev, "Unexpected max number of processes virtualised value\n");
916 return -EINVAL;
917 }
918

   919          if (afu->crs_len < 0) {
                    ^^^^^^^^^^^^^^^^

Remove this test. Unsigned is never less than zero.

   920                  dev_err(&afu->dev, "Unexpected configuration record size value\n");
   921                  return -EINVAL;
   922          }
   923  
   924          return 0;
   925  }

regards,
dan carpenter

#1 lore raw

PATCH 1/1 cxl: fix potential NULL dereference in free_adapter()

If kzalloc() fails when allocating adapter->guest in cxl_guest_init_adapter(), we call free_adapter() before erroring out. free_adapter() in turn attempts to dereference adapter->guest, which in this case is NULL.

In free_adapter(), skip the adapter->guest cleanup if adapter->guest is NULL.

Fixes 14baf4d9c739 ("cxl: Add guest-specific code")

---

Build tested only.

---
 drivers/misc/cxl/guest.c | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
drivers/misc/cxl/guest.c +9 -7
index ee7148e..9aa58a7 100644
@@ -1055,16 +1055,18 @@ static void free_adapter(struct cxl *adapter)
 	struct irq_avail *cur;
 	int i;
 
-	if (adapter->guest->irq_avail) {
-		for (i = 0; i < adapter->guest->irq_nranges; i++) {
-			cur = &adapter->guest->irq_avail[i];
-			kfree(cur->bitmap);
+	if (adapter->guest) {
+		if (adapter->guest->irq_avail) {
+			for (i = 0; i < adapter->guest->irq_nranges; i++) {
+				cur = &adapter->guest->irq_avail[i];
+				kfree(cur->bitmap);
+			}
+			kfree(adapter->guest->irq_avail);
 		}
-		kfree(adapter->guest->irq_avail);
+		kfree(adapter->guest->status);
+		kfree(adapter->guest);
 	}
-	kfree(adapter->guest->status);
 	cxl_remove_adapter_nr(adapter);
-	kfree(adapter->guest);
 	kfree(adapter);
 }
 
Signature · 2 lines
Andrew Donnellan              OzLabs, ADL Canberra
[email protected]  IBM Australia Limited
#2 lore raw

Re: cxl: fix potential NULL dereference in free_adapter()

Quoted 11 lines · Andrew Donnellan

If kzalloc() fails when allocating adapter->guest in cxl_guest_init_adapter(), we call free_adapter() before erroring out. free_adapter() in turn attempts to dereference adapter->guest, which in this case is NULL.

In free_adapter(), skip the adapter->guest cleanup if adapter->guest is NULL.

Fixes 14baf4d9c739 ("cxl: Add guest-specific code")

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/8fbaa51d43ef2c6a72849ec340

cheers

#3 lore raw
In reply to Dan Carpenter
Quoted 3 lines · Dan Carpenter
   919          if (afu->crs_len < 0) {
                    ^^^^^^^^^^^^^^^^

Remove this test. Unsigned is never less than zero.

Is there another lower bound that we should be checking against here?

Signature · 2 lines
Andrew Donnellan              OzLabs, ADL Canberra
[email protected]  IBM Australia Limited