[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7a5e5b07-2870-d47e-c364-b09d004cf873@gpxsee.org>
Date: Wed, 21 Sep 2022 17:35:14 +0200
From: Martin Tůma <tumic@...see.org>
To: Lizhi Hou <lizhi.hou@....com>
Cc: dmaengine@...r.kernel.org, linux-i2c@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-media@...r.kernel.org,
martin.tuma@...iteqautomotive.com, mchehab@...nel.org,
michal.simek@...inx.com, vkoul@...nel.org
Subject: Re: [PATCH v2 2/3] Added Xilinx XDMA IP core driver
On 21. 09. 22 17:15, Lizhi Hou wrote:
>
> On 9/21/22 02:25, Martin Tůma wrote:
>> > Currently, the V3 patch series does not support register user logic
>> > interrupt yet.
>>
>> This is a showstopper for almost every XDMA based PCIe card. As the
>> driver "consumes" the whole register space (including the user IRQs
>> enable/disable registers), there is AFAIK no way how to enable the
>> user IRQs when this driver is loaded.
>>
>> > Could you convert your driver to use this?
>>
>> Not without the user IRQs.
>
> I provided the patch link for user logic IRQ support in previous reply.
> You may pull it and patch it on top of the V3 patch series.
>
>
> Lizhi
>
>>
>> M.
I'm sorry, I have missed that. With that modification it should be
(theoretically - haven't tested it yet) possible to use our v4l2 driver
with this XDMA driver instead of the one I have provided in my patches.
I will now try to rewrite our driver and test it with your XDMA driver
to be sure it works. It will however definitely need a "V4" of your XDMA
driver with that patch on github before it can be included into the
kernel and be usable for all XDMA-based PCIe cards.
M.
Powered by blists - more mailing lists