[PATCH net 0/2] net: ethoc: Error path and transmit fixes

[PATCH net 0/2] net: ethoc: Error path and transmit fixes

[Florian Fainelli]

Hi all,

This patch series contains two patches for the ethoc driver while testing on a
TS-7300 board where ethoc is provided by an on-board FPGA.

First patch was cooked after chasing crashes with invalid resources passed to
the driver.

Second patch was cooked after seeing that an interface configured with IP
192.168.2.2 was sending ARP packets for 192.168.0.0, no wonder why it could not
work.

I don't have access to any other platform using an ethoc interface so
it could be good to some testing on Xtensa for instance.

Florian Fainelli (2):
  net: ethoc: Fix early error paths
  net: ethoc: Correctly pad short packets

 drivers/net/ethernet/ethoc.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

-- 
2.7.4

[PATCH net 1/2] net: ethoc: Fix early error paths

[Florian Fainelli]

In case any operation fails before we can successfully go the point
where we would register a MDIO bus, we would be going to an error label
which involves unregistering then freeing this yet to be created MDIO
bus. Update all error paths to go to label free which is the only one
valid until either the clock is enabled, or the MDIO bus is allocated
and registered. This fixes kernel panics such as this one:

Fixes: a1702857724f ("net: Add support for the OpenCores 10/100 Mbps Ethernet MAC.")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/ethernet/ethoc.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/ethoc.c b/drivers/net/ethernet/ethoc.c
index 4edb98c3c6c7..06ae14a8e946 100644
--- a/drivers/net/ethernet/ethoc.c
+++ b/drivers/net/ethernet/ethoc.c
@@ -1086,7 +1086,7 @@ static int ethoc_probe(struct platform_device *pdev)
 	if (!priv->iobase) {
 		dev_err(&pdev->dev, "cannot remap I/O memory space\n");
 		ret = -ENXIO;
-		goto error;
+		goto free;
 	}
 
 	if (netdev->mem_end) {
@@ -1095,7 +1095,7 @@ static int ethoc_probe(struct platform_device *pdev)
 		if (!priv->membase) {
 			dev_err(&pdev->dev, "cannot remap memory space\n");
 			ret = -ENXIO;
-			goto error;
+			goto free;
 		}
 	} else {
 		/* Allocate buffer memory */
@@ -1106,7 +1106,7 @@ static int ethoc_probe(struct platform_device *pdev)
 			dev_err(&pdev->dev, "cannot allocate %dB buffer\n",
 				buffer_size);
 			ret = -ENOMEM;
-			goto error;
+			goto free;
 		}
 		netdev->mem_end = netdev->mem_start + buffer_size;
 		priv->dma_alloc = buffer_size;
@@ -1120,7 +1120,7 @@ static int ethoc_probe(struct platform_device *pdev)
 		128, (netdev->mem_end - netdev->mem_start + 1) / ETHOC_BUFSIZ);
 	if (num_bd < 4) {
 		ret = -ENODEV;
-		goto error;
+		goto free;
 	}
 	priv->num_bd = num_bd;
 	/* num_tx must be a power of two */
@@ -1133,7 +1133,7 @@ static int ethoc_probe(struct platform_device *pdev)
 	priv->vma = devm_kzalloc(&pdev->dev, num_bd*sizeof(void *), GFP_KERNEL);
 	if (!priv->vma) {
 		ret = -ENOMEM;
-		goto error;
+		goto free;
 	}
 
 	/* Allow the platform setup code to pass in a MAC address. */
-- 
2.7.4

Re: [PATCH net 1/2] net: ethoc: Fix early error paths

[Andrew Lunn]

On Mon, Jul 11, 2016 at 10:04:10AM -0700, Florian Fainelli wrote:
> In case any operation fails before we can successfully go the point
> where we would register a MDIO bus, we would be going to an error label
> which involves unregistering then freeing this yet to be created MDIO
> bus. Update all error paths to go to label free which is the only one
> valid until either the clock is enabled, or the MDIO bus is allocated
> and registered. This fixes kernel panics such as this one:
> 
> Fixes: a1702857724f ("net: Add support for the OpenCores 10/100 Mbps Ethernet MAC.")

Hi Florian

Did you intend to paste a panic here?

    Andrew

Re: [PATCH net 1/2] net: ethoc: Fix early error paths

[Florian Fainelli]

Hi Andrew,

On 07/11/2016 10:11 AM, Andrew Lunn wrote:
> On Mon, Jul 11, 2016 at 10:04:10AM -0700, Florian Fainelli wrote:
>> In case any operation fails before we can successfully go the point
>> where we would register a MDIO bus, we would be going to an error label
>> which involves unregistering then freeing this yet to be created MDIO
>> bus. Update all error paths to go to label free which is the only one
>> valid until either the clock is enabled, or the MDIO bus is allocated
>> and registered. This fixes kernel panics such as this one:
>>
>> Fixes: a1702857724f ("net: Add support for the OpenCores 10/100 Mbps Ethernet MAC.")
> 
> Hi Florian
> 
> Did you intend to paste a panic here?

I did, and then realized this was not so helpful since the NULL pointer
de-reference was fairly obvious, the driver is also not widely used
(AFAICT), I will respin the commit message to remove the last sentence,
thanks!
-- 
Florian

[PATCH net 2/2] net: ethoc: Correctly pad short packets

[Florian Fainelli]

Even though the hardware can be doing zero padding, we want the SKB to
be going out on the wire with the appropriate size. This fixes packet
truncations observed with e.g: ARP packets.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
---
 drivers/net/ethernet/ethoc.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/ethernet/ethoc.c b/drivers/net/ethernet/ethoc.c
index 06ae14a8e946..ca678d46c322 100644
--- a/drivers/net/ethernet/ethoc.c
+++ b/drivers/net/ethernet/ethoc.c
@@ -860,6 +860,11 @@ static netdev_tx_t ethoc_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	unsigned int entry;
 	void *dest;
 
+	if (skb_put_padto(skb, ETHOC_ZLEN)) {
+		dev->stats.tx_errors++;
+		goto out;
+	}
+
 	if (unlikely(skb->len > ETHOC_BUFSIZ)) {
 		dev->stats.tx_errors++;
 		goto out;
-- 
2.7.4