linux-kernel - Re: [PATCH V2] dmaengine: sprd: Support two-stage dma interrupt

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] [day] [month] [year] [list]

Message-ID: <18c6c091-6eb6-cc56-68d6-2b9811da9618@linux.alibaba.com>
Date:   Fri, 21 Oct 2022 09:35:17 +0800
From:   Baolin Wang <baolin.wang@...ux.alibaba.com>
To:     Vinod Koul <vkoul@...nel.org>
Cc:     Cixi Geng <gengcixi@...il.com>, orsonzhai@...il.com,
        zhang.lyra@...il.com, dmaengine@...r.kernel.org,
        linux-kernel@...r.kernel.org, Cixi Geng <cixi.geng1@...soc.com>
Subject: Re: [PATCH V2] dmaengine: sprd: Support two-stage dma interrupt

On 10/20/2022 7:22 PM, Vinod Koul wrote:
> On 20-10-22, 16:15, Baolin Wang wrote:
> 
>>>> why is sprd_dma_int_type part of driver interface. sprd_dma_int_type is
>>>> used only by this driver and should be moved into the driver..
>>
>> Now we can not move this into dma driver, since we have some drivers in the
>> mainline will set the DMA interrupt type, such as spi-sprd.c, sprd_serial.c
>> and sprd-pcm-compress.c.
> 
> That may not sound right.. Why should peripheral set the DMA
> interrupt..?

That's because SPRD DMA controller supplies several different interrupt 
types for different scenarios of users, for example:

@SPRD_DMA_FRAG_INT: fragment done interrupt when one fragment request is 
done.
@SPRD_DMA_BLK_INT: block done interrupt when one block request is done.
@SPRD_DMA_BLK_FRAG_INT: block and fragment interrupt when one fragment 
or one block request is done.
@SPRD_DMA_TRANS_INT: tansaction done interrupt when one transaction 
request is done.
......

Some users may want use Linklist interrupt, and others may just want 
tansaction interrupt. So exposing these interrupt types for users to 
decide seems more suitable.