Re: [PATCH net-next 1/7] dma: compile-out DMA sync op calls when not used

From: Robin Murphy <robin.murphy@arm.com>
To: Simon Horman <horms@kernel.org>,
	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
Subject: Re: [PATCH net-next 1/7] dma: compile-out DMA sync op calls when not used
Date: Wed, 31 Jan 2024 17:14:06 +0000	[thread overview]
Message-ID: <abdb46cc-b49e-4bc7-b703-678b079b120f@arm.com> (raw)
In-Reply-To: <20240131165258.GA401365@kernel.org>

On 31/01/2024 4:52 pm, Simon Horman wrote:
> On Fri, Jan 26, 2024 at 02:54:50PM +0100, Alexander Lobakin wrote:
>> Some platforms do have DMA, but DMA there is always direct and coherent.
>> Currently, even on such platforms DMA sync operations are compiled and
>> called.
>> Add a new hidden Kconfig symbol, DMA_NEED_SYNC, and set it only when
>> either sync operations are needed or there is DMA ops or swiotlb
>> enabled. Set dma_need_sync() and dma_skip_sync() (stub for now)
>> depending on this symbol state and don't call sync ops when
>> dma_skip_sync() is true.
>> The change allows for future optimizations of DMA sync calls depending
>> on compile-time or runtime conditions.
>>
>> Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
> 
> Hi Alexander,
> 
> This seems to cause x86_64 allmodconfig builds to fail:

Oh yeah, the sync_single_range definitions shouldn't need touching at 
all, since they're unconditional wrappers around regular sync_single 
invocations (which already may or may not do anything).

Thanks,
Robin.

> 
>   ../drivers/media/platform/ti/omap3isp/ispstat.c:82:35: error: ‘dma_sync_single_range_for_device’ undeclared (first use in this function); did you mean ‘dma_sync_sgtable_for_device’?
>      82 |                                   dma_sync_single_range_for_device);
>         |                                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>         |                                   dma_sync_sgtable_for_device
>   ../drivers/media/platform/ti/omap3isp/ispstat.c:82:35: note: each undeclared identifier is reported only once for each function it appears in
>   ../drivers/media/platform/ti/omap3isp/ispstat.c: In function ‘isp_stat_buf_sync_magic_for_cpu’:
>   ../drivers/media/platform/ti/omap3isp/ispstat.c:94:35: error: ‘dma_sync_single_range_for_cpu’ undeclared (first use in this function); did you mean ‘dma_sync_sgtable_for_cpu’?
>      94 |                                   dma_sync_single_range_for_cpu);
>         |                                   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>         |                                   dma_sync_sgtable_for_cpu