Received-SPF: pass (google.com: domain of linux-kernel+bounces-81908-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Message-ID: <94043c84-0b03-491a-9dd4-2a792d33bca0@arm.com>
Date: Mon, 26 Feb 2024 16:27:16 +0000
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH net-next v3 1/7] dma: compile-out DMA sync op calls when
 not used
Content-Language: en-GB
To: Alexander Lobakin <aleksander.lobakin@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>,
 Eric Dumazet <edumazet@google.com>, Jakub Kicinski <kuba@kernel.org>,
 Paolo Abeni <pabeni@redhat.com>, Christoph Hellwig <hch@lst.de>,
 Marek Szyprowski <m.szyprowski@samsung.com>, Joerg Roedel <joro@8bytes.org>,
 Will Deacon <will@kernel.org>,
 Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
 "Rafael J. Wysocki" <rafael@kernel.org>,
 Magnus Karlsson <magnus.karlsson@intel.com>,
 Maciej Fijalkowski <maciej.fijalkowski@intel.com>,
 Alexander Duyck <alexanderduyck@fb.com>, bpf@vger.kernel.org,
 netdev@vger.kernel.org, iommu@lists.linux.dev, linux-kernel@vger.kernel.org
References: <20240214162201.4168778-1-aleksander.lobakin@intel.com>
 <20240214162201.4168778-2-aleksander.lobakin@intel.com>
 <893ad3a4-ba24-43cf-8200-b8cd7742622d@arm.com>
 <6b003271-cd83-4091-89c6-bb37da62afef@intel.com>
From: Robin Murphy <robin.murphy@arm.com>
In-Reply-To: <6b003271-cd83-4091-89c6-bb37da62afef@intel.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

On 19/02/2024 12:53 pm, Alexander Lobakin wrote:
> From: Robin Murphy <robin.murphy@arm.com>
> Date: Wed, 14 Feb 2024 17:20:50 +0000
> 
>> On 2024-02-14 4:21 pm, Alexander Lobakin wrote:
> 
> [...]
> 
>>> -static inline void dma_sync_single_for_cpu(struct device *dev,
>>> dma_addr_t addr,
>>> -        size_t size, enum dma_data_direction dir)
>>> +static inline void __dma_sync_single_for_cpu(struct device *dev,
>>> +        dma_addr_t addr, size_t size, enum dma_data_direction dir)
>>
>> To me it would feel more logical to put all the wrappers inside the
>> #ifdef CONFIG_HAS_DMA and not touch these stubs at all (what does it
>> mean to skip an inline no-op?). Or in fact, if dma_skip_sync() is
>> constant false for !HAS_DMA, then we could also just make the external
>> function declarations unconditional and remove the stubs. Not a critical
>> matter though, and I defer to whatever Christoph thinks is most
>> maintainable.
> 
> It's done like that due to that I'm adding a runtime check in the second
> patch. I don't feel like touching this twice makes sense.

Huh? Why would anything need touching twice? All I'm saying is that it's 
pretty pointless to add any invocations of dma_skip_sync() in !HAS_DMA 
paths where we already know the whole API is stubbed out anyway. The 
only cases which are worth differentiating here are HAS_DMA + 
DMA_NEED_SYNC vs. HAS_DMA + !DMA_NEED_SYNC (with the subsequent runtime 
check then just subdividing the former).

> 
> [...]
> 
>>> @@ -348,18 +348,72 @@ static inline void dma_unmap_single_attrs(struct
>>> device *dev, dma_addr_t addr,
>>>        return dma_unmap_page_attrs(dev, addr, size, dir, attrs);
>>>    }
>>>    +static inline void __dma_sync_single_range_for_cpu(struct device *dev,
>>> +        dma_addr_t addr, unsigned long offset, size_t size,
>>> +        enum dma_data_direction dir)
>>> +{
>>> +    __dma_sync_single_for_cpu(dev, addr + offset, size, dir);
>>> +}
>>> +
>>> +static inline void __dma_sync_single_range_for_device(struct device
>>> *dev,
>>> +        dma_addr_t addr, unsigned long offset, size_t size,
>>> +        enum dma_data_direction dir)
>>> +{
>>> +    __dma_sync_single_for_device(dev, addr + offset, size, dir);
>>> +}
>>
>> There is no need to introduce these two.
> 
> I already replied to this in the previous thread. Some subsys may want
> to check for the shortcut earlier to avoid call ladders of their own
> functions. See patch 6 for example where I use this one.

Ugh, no. If the page pool code wants to be clever poking around and 
sidestepping parts of the documented API, it can flippin' well open-code 
a single addition to call __dma_sync_single_for_device() directly 
itself. I'm not at all keen on having to maintain "common" APIs for such 
niche trickery.

Thanks,
Robin.