All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
To: Anuradha Ranasinghe <anuradha@tengriaero.com>
Cc: linux-media <linux-media@vger.kernel.org>,
	Lakshitha Dayasena <lakshitha@tengriaero.com>,
	Krishan Nilanga <krishan@tengriaero.com>,
	linux-pci@vger.kernel.org, l.stach@pengutronix.de,
	Richard.Zhu@freescale.com
Subject: Re: TW686x Linux Main Line Driver Issue
Date: Thu, 20 Apr 2017 11:12:47 -0300	[thread overview]
Message-ID: <CAAEAJfC0MdO2Uy8P0OajRHEc3seUiwLv0qqxLzM3b9eFFfuk8g@mail.gmail.com> (raw)
In-Reply-To: <CAOxqCCT6MOCLG+HHsuOU0zoq1zxRRJNFn0DYz9tOj-ez7+BNRA@mail.gmail.com>

On 20 April 2017 at 07:10, Anuradha Ranasinghe <anuradha@tengriaero.com> wrote:
> Dear All,
>
> This issue is associated to the Linux Mainline Kernel 4.1.15.2 (branch2)
> tw686x upstream driver and IMX6Q platform.
>
> We have an analog camera capture board (a custom one) based around tw6865.
> We are interfacing it with Nitrogen6_Max board (IMX6Q) . We use the
> aforementioned kernel with the boundary devices latest patches to the tw686x
> driver (having 3 DMA buffers) and system running on Ubuntu 16 Xenial Mate
> version.
> https://github.com/boundarydevices/linux-imx6/commits/7fcd22da6d731b36e5ab856551c41301fca9881f
>
> The driver initialization, device and composite signal detection work well
> as intended. But when the streaming started, frame rate becomes lower and
> after few frames, the whole system freezes. To get the camera to work to
> this level, we had to do :
>

What dma-mode are you using? Have you tried other dma-modes?

How many frames do you manage to obtain? I believe you should
debug this further and provide more information.

> 1. Disable PCI interrupts from the kernel (from menuconfig and pci=nomsi
> kernel command)

(CCing PCI people) Lucas, Richard: any idea about why is this parameter needed?

> 2. Set Coherent_Pool to 128M in boot args to get the memory allocation for
> driver. Without this driver does not enumerate.
>

Hm.. interesting.

> I can confirm that there is no issue in our hardware. I strictly followed
> the free scale data sheet recommendations. So I have few questions needing
> your answers :
>
>> Have you guys tried this driver for tw6865 or a related chip on same root
>> fs ? If not can you kindly mention the operating condition you had ?

FWIW, I have tested tw6869 on x86_64. It works well here, but only
using dma-mode=memcpy, as documented here:

http://lxr.free-electrons.com/source/drivers/media/pci/tw686x/tw686x-core.c#L19

>> Is attached patch required for higher kernel versions (4.1.15) to support
>> DMA accessing ?

I have no idea about that patch. The patch has nothing to do with tw686x,
but it's i.MX6 platform-specific, you should ask the author.

>> Is there any additional settings (cma allocation, memory mapping) required
>> for this newer kernel ?

That depends on your platform and usage.

>> I use the pipeline :
> gst-launch-1.0 v4l2src device=/dev/video0 ! video/x-raw,width=720,height=480
> ! autovideosink
> for an unknown reason, imxv4l2videosrc does not work at all for this pcie
> camera.
>

You need to ask imxv4l2videosrc's authors.

-- 
Ezequiel García, VanguardiaSur
www.vanguardiasur.com.ar

WARNING: multiple messages have this Message-ID (diff)
From: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
To: Anuradha Ranasinghe <anuradha@tengriaero.com>
Cc: linux-media <linux-media@vger.kernel.org>,
	Lakshitha Dayasena <lakshitha@tengriaero.com>,
	Krishan Nilanga <krishan@tengriaero.com>,
	linux-pci@vger.kernel.org, l.stach@pengutronix.de,
	Richard.Zhu@freescale.com
Subject: Re: TW686x Linux Main Line Driver Issue
Date: Thu, 20 Apr 2017 11:12:47 -0300	[thread overview]
Message-ID: <CAAEAJfC0MdO2Uy8P0OajRHEc3seUiwLv0qqxLzM3b9eFFfuk8g@mail.gmail.com> (raw)
In-Reply-To: <CAOxqCCT6MOCLG+HHsuOU0zoq1zxRRJNFn0DYz9tOj-ez7+BNRA@mail.gmail.com>

On 20 April 2017 at 07:10, Anuradha Ranasinghe <anuradha@tengriaero.com> wr=
ote:
> Dear All,
>
> This issue is associated to the Linux Mainline Kernel 4.1.15.2 (branch2)
> tw686x upstream driver and IMX6Q platform.
>
> We have an analog camera capture board (a custom one) based around tw6865=
.
> We are interfacing it with Nitrogen6_Max board (IMX6Q) . We use the
> aforementioned kernel with the boundary devices latest patches to the tw6=
86x
> driver (having 3 DMA buffers) and system running on Ubuntu 16 Xenial Mate
> version.
> https://github.com/boundarydevices/linux-imx6/commits/7fcd22da6d731b36e5a=
b856551c41301fca9881f
>
> The driver initialization, device and composite signal detection work wel=
l
> as intended. But when the streaming started, frame rate becomes lower and
> after few frames, the whole system freezes. To get the camera to work to
> this level, we had to do :
>

What dma-mode are you using? Have you tried other dma-modes?

How many frames do you manage to obtain? I believe you should
debug this further and provide more information.

> 1. Disable PCI interrupts from the kernel (from menuconfig and pci=3Dnoms=
i
> kernel command)

(CCing PCI people) Lucas, Richard: any idea about why is this parameter nee=
ded?

> 2. Set Coherent_Pool to 128M in boot args to get the memory allocation fo=
r
> driver. Without this driver does not enumerate.
>

Hm.. interesting.

> I can confirm that there is no issue in our hardware. I strictly followed
> the free scale data sheet recommendations. So I have few questions needin=
g
> your answers :
>
>> Have you guys tried this driver for tw6865 or a related chip on same roo=
t
>> fs ? If not can you kindly mention the operating condition you had ?

FWIW, I have tested tw6869 on x86_64. It works well here, but only
using dma-mode=3Dmemcpy, as documented here:

http://lxr.free-electrons.com/source/drivers/media/pci/tw686x/tw686x-core.c=
#L19

>> Is attached patch required for higher kernel versions (4.1.15) to suppor=
t
>> DMA accessing ?

I have no idea about that patch. The patch has nothing to do with tw686x,
but it's i.MX6 platform-specific, you should ask the author.

>> Is there any additional settings (cma allocation, memory mapping) requir=
ed
>> for this newer kernel ?

That depends on your platform and usage.

>> I use the pipeline :
> gst-launch-1.0 v4l2src device=3D/dev/video0 ! video/x-raw,width=3D720,hei=
ght=3D480
> ! autovideosink
> for an unknown reason, imxv4l2videosrc does not work at all for this pcie
> camera.
>

You need to ask imxv4l2videosrc's authors.

--=20
Ezequiel Garc=C3=ADa, VanguardiaSur
www.vanguardiasur.com.ar

       reply	other threads:[~2017-04-20 14:12 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CAOxqCCT6MOCLG+HHsuOU0zoq1zxRRJNFn0DYz9tOj-ez7+BNRA@mail.gmail.com>
2017-04-20 14:12 ` Ezequiel Garcia [this message]
2017-04-20 14:12   ` TW686x Linux Main Line Driver Issue Ezequiel Garcia
2017-04-24 10:38   ` Lucas Stach
     [not found]     ` <CAKXQXwLuG1A37NTPrE0abPWhMDGd=10Ud+xNa-4+k+8qMhD8tA@mail.gmail.com>
     [not found]       ` <CAKXQXwKeA6YCaepmjJJBf+Nc3bOO9aEnGmnHfnb2aDX3f6YXzw@mail.gmail.com>
2017-04-25 14:15         ` Ezequiel Garcia
2017-04-25 14:15           ` Ezequiel Garcia

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAAEAJfC0MdO2Uy8P0OajRHEc3seUiwLv0qqxLzM3b9eFFfuk8g@mail.gmail.com \
    --to=ezequiel@vanguardiasur.com.ar \
    --cc=Richard.Zhu@freescale.com \
    --cc=anuradha@tengriaero.com \
    --cc=krishan@tengriaero.com \
    --cc=l.stach@pengutronix.de \
    --cc=lakshitha@tengriaero.com \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.