2010-08-11 10:12:06

by Julia Lawall

[permalink] [raw]
Subject: [PATCH 5/5] drivers/serial: Return -ENOMEM on memory allocation failure

From: Julia Lawall <[email protected]>

In this code, 0 is returned on memory allocation failure, even though other
failures return -ENOMEM or other similar values.

A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)

// <smpl>
@@
expression ret;
expression x,e1,e2,e3;
@@

ret = 0
... when != ret = e1
*x = \(kmalloc\|kcalloc\|kzalloc\)(...)
... when != ret = e2
if (x == NULL) { ... when != ret = e3
return ret;
}
// </smpl>

Signed-off-by: Julia Lawall <[email protected]>

---
I believe this code also leaks earlier instances of port, which are only
referenced by card_ptr, which is freed in the error handling code at the
end of the function. A lot of operations are done on port on each
iteration, however, so I'm not sure whether it is good enough to just free
them. Perhaps there is some way to call ioc3uart_remove?

drivers/serial/ioc3_serial.c | 1 +
1 file changed, 1 insertion(+)

diff --git a/drivers/serial/ioc3_serial.c b/drivers/serial/ioc3_serial.c
index 93de907..800c546 100644
--- a/drivers/serial/ioc3_serial.c
+++ b/drivers/serial/ioc3_serial.c
@@ -2044,6 +2044,7 @@ ioc3uart_probe(struct ioc3_submodule *is, struct ioc3_driver_data *idd)
if (!port) {
printk(KERN_WARNING
"IOC3 serial memory not available for port\n");
+ ret = -ENOMEM;
goto out4;
}
spin_lock_init(&port->ip_lock);


2010-08-11 15:59:55

by Patrick Gefre

[permalink] [raw]
Subject: Re: [PATCH 5/5] drivers/serial: Return -ENOMEM on memory allocation failure

Julia Lawall wrote:
> From: Julia Lawall <[email protected]>
>
> In this code, 0 is returned on memory allocation failure, even though other
> failures return -ENOMEM or other similar values.
>
> A simplified version of the semantic match that finds this problem is as
> follows: (http://coccinelle.lip6.fr/)
>
> // <smpl>
> @@
> expression ret;
> expression x,e1,e2,e3;
> @@
>
> ret = 0
> ... when != ret = e1
> *x = \(kmalloc\|kcalloc\|kzalloc\)(...)
> ... when != ret = e2
> if (x == NULL) { ... when != ret = e3
> return ret;
> }
> // </smpl>
>
> Signed-off-by: Julia Lawall <[email protected]>
>

Signed-off-by: Pat Gefre <[email protected]>



> ---
> I believe this code also leaks earlier instances of port, which are only
> referenced by card_ptr, which is freed in the error handling code at the
> end of the function. A lot of operations are done on port on each
> iteration, however, so I'm not sure whether it is good enough to just free
> them. Perhaps there is some way to call ioc3uart_remove?
>

Yes you are right, there should be something like this for out4:

out4:
for (phys_port = 0; phys_port < PORTS_PER_CARD; phys_port++) {
port = card_ptr->ic_port[phys_port].icp_port;
if (port) {
pci_free_consistent(port->ip_idd->pdev,
TOTAL_RING_BUF_SIZE,
(void *)port->ip_cpu_ringbuf,
port->ip_dma_ringbuf);
kfree(port);
}
}
kfree(card_ptr);
return ret;




> drivers/serial/ioc3_serial.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/serial/ioc3_serial.c b/drivers/serial/ioc3_serial.c
> index 93de907..800c546 100644
> --- a/drivers/serial/ioc3_serial.c
> +++ b/drivers/serial/ioc3_serial.c
> @@ -2044,6 +2044,7 @@ ioc3uart_probe(struct ioc3_submodule *is, struct ioc3_driver_data *idd)
> if (!port) {
> printk(KERN_WARNING
> "IOC3 serial memory not available for port\n");
> + ret = -ENOMEM;
> goto out4;
> }
> spin_lock_init(&port->ip_lock);

2010-08-11 16:23:32

by Julia Lawall

[permalink] [raw]
Subject: Re: [PATCH 5/5] drivers/serial: Return -ENOMEM on memory allocation failure

> > I believe this code also leaks earlier instances of port, which are only
> > referenced by card_ptr, which is freed in the error handling code at the
> > end of the function. A lot of operations are done on port on each
> > iteration, however, so I'm not sure whether it is good enough to just free
> > them. Perhaps there is some way to call ioc3uart_remove?
> >
>
> Yes you are right, there should be something like this for out4:
>
> out4:
> for (phys_port = 0; phys_port < PORTS_PER_CARD; phys_port++) {
> port = card_ptr->ic_port[phys_port].icp_port;
> if (port) {
> pci_free_consistent(port->ip_idd->pdev,
> TOTAL_RING_BUF_SIZE,
> (void *)port->ip_cpu_ringbuf,
> port->ip_dma_ringbuf);
> kfree(port);
> }
> }
> kfree(card_ptr);
> return ret;

Actually, pci_alloc_consistent is only called when phys_port is 0. In the
subsequent cases, the ip_dma_ringbuf field is just initialized to the
previous value. So it could be:

out4:
for (phys_port = 0; phys_port < PORTS_PER_CARD; phys_port++) {
port = card_ptr->ic_port[phys_port].icp_port;
if (port) {
if (phys_port == 0)
pci_free_consistent(port->ip_idd->pdev,
TOTAL_RING_BUF_SIZE,
(void *)port->ip_cpu_ringbuf,
port->ip_dma_ringbuf);
kfree(port);
}
}
kfree(card_ptr);
return ret;

julia