2020-10-02 14:03:35

by Julia Lawall

[permalink] [raw]
Subject: kzfree script

Denis,

In the rule proposing kzfree_sensitive, I think it would be helpful to
also highlight the memset line.

julia


2020-10-02 14:12:00

by Denis Efremov (Oracle)

[permalink] [raw]
Subject: Re: kzfree script

Hi,

On 10/2/20 5:01 PM, Julia Lawall wrote:
> Denis,
>
> In the rule proposing kzfree_sensitive, I think it would be helpful to
> also highlight the memset line.

What do you mean? It's "highlighted" in context mode. Do you mean adding
position argument to memset call and showing this position in the warning
messages?

Thanks,
Denis

2020-10-02 14:16:39

by Julia Lawall

[permalink] [raw]
Subject: Re: kzfree script



On Fri, 2 Oct 2020, Denis Efremov wrote:

> Hi,
>
> On 10/2/20 5:01 PM, Julia Lawall wrote:
> > Denis,
> >
> > In the rule proposing kzfree_sensitive, I think it would be helpful to
> > also highlight the memset line.
>
> What do you mean? It's "highlighted" in context mode. Do you mean adding
> position argument to memset call and showing this position in the warning
> messages?

Yes, that seems to be what I mean. 0-day generated a message from the
script, and I had to hunt around for the reason why it was doing that. So
it would be nice to have the memset highlighted. It seems that the
non-patch 0-day messages are generated from the report mode.

thanks,
julia

2020-10-02 14:20:12

by Denis Efremov (Oracle)

[permalink] [raw]
Subject: Re: kzfree script

On 10/2/20 5:13 PM, Julia Lawall wrote:
>
>
> On Fri, 2 Oct 2020, Denis Efremov wrote:
>
>> Hi,
>>
>> On 10/2/20 5:01 PM, Julia Lawall wrote:
>>> Denis,
>>>
>>> In the rule proposing kzfree_sensitive, I think it would be helpful to
>>> also highlight the memset line.
>>
>> What do you mean? It's "highlighted" in context mode. Do you mean adding
>> position argument to memset call and showing this position in the warning
>> messages?
>
> Yes, that seems to be what I mean. 0-day generated a message from the
> script, and I had to hunt around for the reason why it was doing that. So
> it would be nice to have the memset highlighted. It seems that the
> non-patch 0-day messages are generated from the report mode.
>

Ok, I will send a patch for it.

Thanks,
Denis