2010-07-30 11:08:22

by Kulikov Vasiliy

Subject: [PATCH 3/9] staging: otus: check kmalloc() return value

kmalloc() may fail, if so return error from zfwUsbSubmitControl().

Signed-off-by: Kulikov Vasiliy <[email protected]>
---
drivers/staging/otus/wrap_usb.c | 5 +++++
1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/drivers/staging/otus/wrap_usb.c b/drivers/staging/otus/wrap_usb.c
index 93459ca..9f04047 100644
--- a/drivers/staging/otus/wrap_usb.c
+++ b/drivers/staging/otus/wrap_usb.c
@@ -104,6 +104,11 @@ u32_t zfwUsbSubmitControl(zdev_t *dev, u8_t req, u16_t value, u16_t index,

if (size > 0) {
buf = kmalloc(size, GFP_KERNEL);
+ if (buf == NULL) {
+ pr_err("zfwUsbSubmitControl() failed, "
+ "kmalloc() returned NULL\n");
+ return 1;
+ }
memcpy(buf, (u8_t *)data, size);
} else
buf = NULL;
--
1.7.0.4

2010-07-30 12:31:23

by walter harms

Subject: Re: [PATCH 3/9] staging: otus: check kmalloc() return value

Kulikov Vasiliy schrieb:
> kmalloc() may fail, if so return error from zfwUsbSubmitControl().
>
> Signed-off-by: Kulikov Vasiliy <[email protected]>
> ---
> drivers/staging/otus/wrap_usb.c | 5 +++++
> 1 files changed, 5 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/staging/otus/wrap_usb.c b/drivers/staging/otus/wrap_usb.c
> index 93459ca..9f04047 100644
> --- a/drivers/staging/otus/wrap_usb.c
> +++ b/drivers/staging/otus/wrap_usb.c
> @@ -104,6 +104,11 @@ u32_t zfwUsbSubmitControl(zdev_t *dev, u8_t req, u16_t value, u16_t index,
>
> if (size > 0) {
> buf = kmalloc(size, GFP_KERNEL);
> + if (buf == NULL) {
> + pr_err("zfwUsbSubmitControl() failed, "
> + "kmalloc() returned NULL\n");
> + return 1;
> + }
> memcpy(buf, (u8_t *)data, size);
> } else
> buf = NULL;


We had a memdup() somewhere had'nt we ?

re,
wh

2010-07-30 12:49:31

by Jiri Slaby

Subject: Re: [PATCH 3/9] staging: otus: check kmalloc() return value

On 07/30/2010 02:31 PM, walter harms wrote:
>
>
> Kulikov Vasiliy schrieb:
>> kmalloc() may fail, if so return error from zfwUsbSubmitControl().
>>
>> Signed-off-by: Kulikov Vasiliy <[email protected]>
>> ---
>> drivers/staging/otus/wrap_usb.c | 5 +++++
>> 1 files changed, 5 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/staging/otus/wrap_usb.c b/drivers/staging/otus/wrap_usb.c
>> index 93459ca..9f04047 100644
>> --- a/drivers/staging/otus/wrap_usb.c
>> +++ b/drivers/staging/otus/wrap_usb.c
>> @@ -104,6 +104,11 @@ u32_t zfwUsbSubmitControl(zdev_t *dev, u8_t req, u16_t value, u16_t index,
>>
>> if (size > 0) {
>> buf = kmalloc(size, GFP_KERNEL);
>> + if (buf == NULL) {
>> + pr_err("zfwUsbSubmitControl() failed, "
>> + "kmalloc() returned NULL\n");
>> + return 1;
>> + }
>> memcpy(buf, (u8_t *)data, size);
>> } else
>> buf = NULL;
>
>
> We had a memdup() somewhere had'nt we ?

Yes, but it is not what he is changing. (Patches welcome.)

The patch is OK, except try to avoid function names in printks next
time. Function names are constantly changing, but developers tend not to
change printk strings. (%s + __func__)

regards,
--
js

2010-07-30 14:07:33

by Dan Carpenter

Subject: Re: [PATCH 3/9] staging: otus: check kmalloc() return value

On Fri, Jul 30, 2010 at 03:08:00PM +0400, Kulikov Vasiliy wrote:
> if (size > 0) {
> buf = kmalloc(size, GFP_KERNEL);
> + if (buf == NULL) {
> + pr_err("zfwUsbSubmitControl() failed, "
> + "kmalloc() returned NULL\n");

This isn't a big deal, but the pr_err() isn't needed. kmalloc() already
prints a message unless __GFP_NOWARN is used.

regards,
dan carpenter