2023-07-05 00:22:22

by Yu Hao

Subject: [PATCH] net: wireless: cisco: Fix possible uninit bug

The struct cap_rid should be initialized by function readCapabilityRid.
However, there is not return value check. Iit is possible that
the function readCapabilityRid returns error code and cap_rid.softCap
is not initialized. But there is a read later for this field.

Signed-off-by: Yu Hao <[email protected]>
---
drivers/net/wireless/cisco/airo.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/cisco/airo.c
b/drivers/net/wireless/cisco/airo.c
index 7c4cc5f5e1eb..b3736d76a5d5 100644
--- a/drivers/net/wireless/cisco/airo.c
+++ b/drivers/net/wireless/cisco/airo.c
@@ -6950,8 +6950,11 @@ static int airo_get_range(struct net_device *dev,
CapabilityRid cap_rid; /* Card capability info */
int i;
int k;
+ int status;

- readCapabilityRid(local, &cap_rid, 1);
+ status = readCapabilityRid(local, &cap_rid, 1);
+ if (status != SUCCESS)
+ return ERROR;

dwrq->length = sizeof(struct iw_range);
memset(range, 0, sizeof(*range));
--
2.34.1

2023-07-06 08:03:41

by Paolo Abeni

Subject: Re: [PATCH] net: wireless: cisco: Fix possible uninit bug

On Tue, 2023-07-04 at 16:50 -0700, Yu Hao wrote:
> The struct cap_rid should be initialized by function readCapabilityRid.
> However, there is not return value check. Iit is possible that
> the function readCapabilityRid returns error code and cap_rid.softCap
> is not initialized. But there is a read later for this field.
>
> Signed-off-by: Yu Hao <[email protected]>
> ---
> drivers/net/wireless/cisco/airo.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/wireless/cisco/airo.c
> b/drivers/net/wireless/cisco/airo.c
> index 7c4cc5f5e1eb..b3736d76a5d5 100644
> --- a/drivers/net/wireless/cisco/airo.c
> +++ b/drivers/net/wireless/cisco/airo.c
> @@ -6950,8 +6950,11 @@ static int airo_get_range(struct net_device *dev,
> CapabilityRid cap_rid; /* Card capability info */
> int i;
> int k;
> + int status;
>
> - readCapabilityRid(local, &cap_rid, 1);
> + status = readCapabilityRid(local, &cap_rid, 1);
> + if (status != SUCCESS)
> + return ERROR;

This value is returned directly to the network stack, you must use a
standard error code instead (e.g. -EINVAL).

Also, please add a suitable 'Fixes' tag, thanks!

Paolo

2023-07-10 03:07:16

by Yu Hao

Subject: Re: [PATCH] net: wireless: cisco: Fix possible uninit bug

Sure.

I found that the related code is from Linux-2.6.12-rc2.
In this situation, the 'Fixes' tag should be 'Fixes: 1da177e4c3f4
("Linux-2.6.12-rc2")'?

Yu Hao

On Thu, Jul 6, 2023 at 12:54 AM Paolo Abeni <[email protected]> wrote:
>
> On Tue, 2023-07-04 at 16:50 -0700, Yu Hao wrote:
> > The struct cap_rid should be initialized by function readCapabilityRid.
> > However, there is not return value check. Iit is possible that
> > the function readCapabilityRid returns error code and cap_rid.softCap
> > is not initialized. But there is a read later for this field.
> >
> > Signed-off-by: Yu Hao <[email protected]>
> > ---
> > drivers/net/wireless/cisco/airo.c | 5 ++++-
> > 1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/wireless/cisco/airo.c
> > b/drivers/net/wireless/cisco/airo.c
> > index 7c4cc5f5e1eb..b3736d76a5d5 100644
> > --- a/drivers/net/wireless/cisco/airo.c
> > +++ b/drivers/net/wireless/cisco/airo.c
> > @@ -6950,8 +6950,11 @@ static int airo_get_range(struct net_device *dev,
> > CapabilityRid cap_rid; /* Card capability info */
> > int i;
> > int k;
> > + int status;
> >
> > - readCapabilityRid(local, &cap_rid, 1);
> > + status = readCapabilityRid(local, &cap_rid, 1);
> > + if (status != SUCCESS)
> > + return ERROR;
>
> This value is returned directly to the network stack, you must use a
> standard error code instead (e.g. -EINVAL).
>
> Also, please add a suitable 'Fixes' tag, thanks!
>
> Paolo
>

2023-07-10 16:36:58

by Paolo Abeni

Subject: Re: [PATCH] net: wireless: cisco: Fix possible uninit bug

On Sun, 2023-07-09 at 19:52 -0700, Yu Hao wrote:
> I found that the related code is from Linux-2.6.12-rc2.
> In this situation, the 'Fixes' tag should be 'Fixes: 1da177e4c3f4
> ("Linux-2.6.12-rc2")'?

Exactly!

Cheers,

Paolo