[PATCH] net: caif: check return value of alloc_netdev

[PATCH] net: caif: check return value of alloc_netdev

[Rasmus Villemoes]

I don't know if dev can actually be NULL here, but the test should be
above alloc_netdev(), to avoid leaking the struct net_device in case
dev is actually NULL. And of course the return value from alloc_netdev
should be tested.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
---
Maybe the existing code was supposed to be "if (!ndev)", and dev
cannot be NULL, but then -ENODEV is a slightly odd return
value. Doing both tests seems to be the safe choice.

 drivers/net/caif/caif_spi.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/caif/caif_spi.c b/drivers/net/caif/caif_spi.c
index de3962014af7..4721948a92f6 100644
--- a/drivers/net/caif/caif_spi.c
+++ b/drivers/net/caif/caif_spi.c
@@ -730,11 +730,14 @@ int cfspi_spi_probe(struct platform_device *pdev)
 	int res;
 	dev = (struct cfspi_dev *)pdev->dev.platform_data;
 
-	ndev = alloc_netdev(sizeof(struct cfspi), "cfspi%d",
-			    NET_NAME_UNKNOWN, cfspi_setup);
 	if (!dev)
 		return -ENODEV;
 
+	ndev = alloc_netdev(sizeof(struct cfspi), "cfspi%d",
+			    NET_NAME_UNKNOWN, cfspi_setup);
+	if (!ndev)
+		return -ENOMEM;
+
 	cfspi = netdev_priv(ndev);
 	netif_stop_queue(ndev);
 	cfspi->ndev = ndev;
-- 
2.6.1

Re: [PATCH] net: caif: check return value of alloc_netdev

[David Miller]

From: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Date: Mon,  9 Nov 2015 13:19:10 +0100

> I don't know if dev can actually be NULL here, but the test should be
> above alloc_netdev(), to avoid leaking the struct net_device in case
> dev is actually NULL. And of course the return value from alloc_netdev
> should be tested.
> 
> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> ---
> Maybe the existing code was supposed to be "if (!ndev)", and dev
> cannot be NULL, but then -ENODEV is a slightly odd return
> value. Doing both tests seems to be the safe choice.

This is made painfully impossible to figure out because there seems
to be no in-tree code instantiating platform_device objects for
this caif_spi framework, nor are there even in-tree implementations
of the init_xfer et al. required methods.

So your fix is as good as any, applied, thanks.