linux-kernel - Re: [PATCH v2 6/9] debugfs: Implement debugfs_create

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 26 Mar 2021 15:58:37 +0100
From:   Rasmus Villemoes <linux@...musvillemoes.dk>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     Greg KH <greg@...ah.com>, mingo@...nel.org, mgorman@...e.de,
        juri.lelli@...hat.com, vincent.guittot@...aro.org,
        dietmar.eggemann@....com, rostedt@...dmis.org, bsegall@...gle.com,
        bristot@...hat.com, joshdon@...gle.com, valentin.schneider@....com,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 6/9] debugfs: Implement debugfs_create_str()

On 26/03/2021 15.22, Peter Zijlstra wrote:
> On Fri, Mar 26, 2021 at 01:53:59PM +0100, Rasmus Villemoes wrote:
>> On 26/03/2021 12.38, Peter Zijlstra wrote:
> 
>>> +
>>> +again:
>>> +	rcu_read_lock();
>>> +	str = rcu_dereference(*(char **)file->private_data);
>>> +	len = strlen(str) + 1;
>>> +
>>> +	if (!copy || copy_len < len) {
>>> +		rcu_read_unlock();
>>> +		kfree(copy);
>>> +		copy = kmalloc(len + 1, GFP_KERNEL);
>>> +		if (!copy) {
>>> +			debugfs_file_put(dentry);
>>> +			return -ENOMEM;
>>> +		}
>>> +		copy_len = len;
>>> +		goto again;
>>> +	}
>>> +
>>> +	strncpy(copy, str, len);
>>> +	copy[len] = '\n';
>>> +	copy[len+1] = '\0';
>>> +	rcu_read_unlock();
>>
>> As noted (accidentally off-list), this is broken. I think you want this
>> on top
>>
>> - len = strlen(str) + 1;
>> + len = strlen(str);
> 
>   kmalloc(len + 2, ...);

No, because nul-terminating the stuff you pass to
simple_read_from_buffer is pointless cargo-culting. Yeah, read_file_bool
does it, but that's just bogus.

>> - strncpy(copy, str, len);
>> + memcpy(copy, str, len);
>>   copy[len] = '\n';
>> - copy[len+1] = '\0';
> 
> I'll go with strscpy() I tihnk, something like:
> 
> 	len = strscpy(copy, str, len);
> 	if (len < 0)
> 		return len;

To what end? The only way that could possibly return -EFOO is if the
nul-terminator in str vanished between the strlen() and here, and in
that case you have bigger problems.

> 	copy[len] = '\n';
> 	copy[len + 1] = '\0';
> 
>>> +EXPORT_SYMBOL_GPL(debugfs_read_file_str);
>>
>> Why?
> 
> Copy-pasta from debugfs_*_bool(). This thing seems to export everything
> and I figured I'd go along with that.

I thought the convention was not to export anything until the kernel
itself had a (modular) user. But I can't find that stated under
Documentation/ anywhere - but it does say "Every function that is
exported to loadable modules using
``EXPORT_SYMBOL`` or ``EXPORT_SYMBOL_GPL`` should have a kernel-doc
comment.". Anyway, the *_bool stuff doesn't seem to be something to be
copy-pasted.

Rasmus