Message-ID: <59b703ac-9660-a0e5-09f7-c5b69d4f39cb@redhat.com>
Date:   Wed, 8 Mar 2023 16:18:55 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.8.0
Subject: Re: [PATCH] maple_tree: export symbol mas_preallocate()
Content-Language: en-US
To:     Christoph Hellwig <hch@lst.de>
Cc:     Vlastimil Babka <vbabka@suse.cz>,
        "Liam R. Howlett" <Liam.Howlett@Oracle.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org, maple-tree@lists.infradead.org
References: <20230302011035.4928-1-dakr@redhat.com>
 <20230302015242.xi3y53okc4skmonn@revolver>
 <9abef101-3a9b-0bdd-7139-ced7d5e28ebe@suse.cz>
 <da65d59e-f918-d7d5-644b-33cc51c2ba6c@redhat.com>
 <20230308144343.GA16259@lst.de>
From:   Danilo Krummrich <dakr@redhat.com>
Organization: RedHat
In-Reply-To: <20230308144343.GA16259@lst.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 3/8/23 15:43, Christoph Hellwig wrote:
> On Wed, Mar 08, 2023 at 03:34:29PM +0100, Danilo Krummrich wrote:
>> On 3/8/23 11:59, Vlastimil Babka wrote:
>>> On 3/2/23 02:52, Liam R. Howlett wrote:
>>>> Thanks for the patch.  This should indeed be exported.
>>>>
>>>> Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
>>>
>>> What uses it? Don't we export only things used by in-kernel modules?
>>
>> mas_preallocate() is part of the maple tree's 'Advanced API'. All other
>> functions of this API are exported already.
>>
>> More specifically, e.g. mas_store_prealloc() is exported which doesn't make
>> a lot of sense without mas_preallocate() being available too.
>>
>> I don't think it is used outside of mm yet, but as a generic tree
>> implementation maple tree is pretty likely to be used in code built as
>> module, e.g. drivers.
> 
> If anyone wants to use they can add the export as needed.
> 

Valid perception, though personally I wouldn't share it in this specific 
case.

Anyway, it's just that all other functions of the 'Advanced API' are 
exported already without them having other users than mm either, e.g. 
mas_store_prealloc() can't be used without mas_preallocate(), hence 
exporting one, but not the other doesn't really make sense.

Consistency wise I think we should either also export mas_preallocate() 
or don't export the other ones either.