Re: [PATCH] per-task delay accounting taskstats interface: controlexit data through cpumasks]

From: Shailabh Nagar
Date: Thu Jul 06 2006 - 18:26:12 EST


Andrew Morton wrote:
> Thomas Graf <tgraf@xxxxxxx> wrote:
>
>>* Shailabh Nagar <nagar@xxxxxxxxxxxxxx> 2006-07-06 07:37
>>
>>>@@ -37,9 +45,26 @@ static struct nla_policy taskstats_cmd_g
>>> __read_mostly = {
>>> [TASKSTATS_CMD_ATTR_PID] = { .type = NLA_U32 },
>>> [TASKSTATS_CMD_ATTR_TGID] = { .type = NLA_U32 },
>>>+ [TASKSTATS_CMD_ATTR_REGISTER_CPUMASK] = { .type = NLA_STRING },
>>>+ [TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK] = { .type = NLA_STRING },};
>>
>>>+ na = info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK];
>>>+ if (nla_len(na) > TASKSTATS_CPUMASK_MAXLEN)
>>>+ return -E2BIG;
>>>+ rc = cpulist_parse((char *)nla_data(na), mask);
>>
>>This isn't safe, the data in the attribute is not guaranteed to be
>>NUL terminated. Still it's probably me to blame for not making
>>this more obvious in the API.
>>
>
>
> Thanks, that was an unpleasant bug.
>
>
>>I've attached a patch below extending the API to make it easier
>>for interfaces using NUL termianted strings,
>
>
> In the interests of keeping this work decoupled from netlink enhancements
> I'd propose the below.

The patch looks good. I was also thinking of simply modifying the input
to have the null termination and modify later when netlink provides
generic support.


> Is it bad to modify the data at nla_data()?
>
>
> --- a/kernel/taskstats.c~per-task-delay-accounting-taskstats-interface-control-exit-data-through-cpumasks-fix
> +++ a/kernel/taskstats.c
> @@ -299,6 +299,23 @@ cleanup:
> return 0;
> }
>
> +static int parse(struct nlattr *na, cpumask_t *mask)
> +{
> + char *data;
> + int len;
> +
> + if (na == NULL)
> + return 1;
> + len = nla_len(na);
> + if (len > TASKSTATS_CPUMASK_MAXLEN)
> + return -E2BIG;
> + if (len < 1)
> + return -EINVAL;
> + data = nla_data(na);
> + data[len - 1] = '\0';
> + return cpulist_parse(data, *mask);
> +}
> +
> static int taskstats_user_cmd(struct sk_buff *skb, struct genl_info *info)
> {
> int rc = 0;
> @@ -309,27 +326,17 @@ static int taskstats_user_cmd(struct sk_
> struct nlattr *na;
> cpumask_t mask;
>
> - if (info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK]) {
> - na = info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK];
> - if (nla_len(na) > TASKSTATS_CPUMASK_MAXLEN)
> - return -E2BIG;
> - rc = cpulist_parse((char *)nla_data(na), mask);
> - if (rc)
> - return rc;
> - rc = add_del_listener(info->snd_pid, &mask, REGISTER);
> + rc = parse(info->attrs[TASKSTATS_CMD_ATTR_REGISTER_CPUMASK], &mask);
> + if (rc < 0)
> return rc;
> - }
> + if (rc == 0)
> + return add_del_listener(info->snd_pid, &mask, REGISTER);
>
> - if (info->attrs[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK]) {
> - na = info->attrs[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK];
> - if (nla_len(na) > TASKSTATS_CPUMASK_MAXLEN)
> - return -E2BIG;
> - rc = cpulist_parse((char *)nla_data(na), mask);
> - if (rc)
> - return rc;
> - rc = add_del_listener(info->snd_pid, &mask, DEREGISTER);
> + rc = parse(info->attrs[TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK], &mask);
> + if (rc < 0)
> return rc;
> - }
> + if (rc == 0)
> + return add_del_listener(info->snd_pid, &mask, DEREGISTER);
>
> /*
> * Size includes space for nested attributes
> _
>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/