Re: [PATCH] seq_file: Allow private data to be supplied on seq_open

[Rob Jones <rob.jones@codethink.co.uk>]

Hi Steve,

On 07/08/14 14:32, Steven Whitehouse wrote:
> Hi,
>
> On 07/08/14 13:58, Rob Jones wrote:
> [snip]
>>
>> On a related subject, Having looked at a few uses of seq_file, I must
>> say that some users seem to make assumptions about the internal
>> workings of the module. Dangerous behaviour as only some behaviours are
>> documented.
>>
>> e.g. The behaviour that "struct seq_file" pointer is stored in
>> file->private_data is documented and can therefore be relied upon but
>> the fact that the output buffer and its size are only defined at the
>> first output (and can therefore be pre-defined and pre-allocated by
>> user code) is not documented and could therefore change without warning.
>>
>> This second behaviour is assumed in, for example, module fs/gfs2/glock.c
>> which could, therefore, stop working properly without warning if the
>> internal behaviour was changed.
>>
> While it is undocumented, it is I understand, how this feature was
> intended to be used, so I think that it is safe to do this in the GFS2
> case. Here is a ref to the thread which explains how it landed up like
> that:
> https://www.redhat.com/archives/cluster-devel/2012-June/msg00000.html

No criticism was intended of that particular piece of code, It has been
there for a couple of years and is, presumably, still working :-)

It was just a general point about things needing to be written down. A
behaviour such as you were relying on can be a very positive thing but
it would be of much greater use if it was written down in the file docs.

I completely missed seq_file_private() because I was looking at the
docs more than the code. If it had been written down in the docs it
would have saved me quite a bit of time, similarly, if the buffer
allocation behaviour was documented, changes to seq_file.c would not be
made that could break your code.

God knows, I'm not a fan of unnecessary documentation but where it's
useful I'm all for it.

>
> Steve.

-- 
Rob Jones
Codethink Ltd
mailto:rob.jones@codethink.co.uk
tel:+44 161 236 5575