Firmware is available from http://www.steventoth.net/linux/hvr22xx/firmwares/4019072

Comments welcome.

I want to take this opportunity to thank everyone involved in the SAA7164 project. If you’ve donated, tested, written to me, complained, given me words of encouragement or simply reported bugs… thank you!

What’s next for the SAA7164 project? Once the merge is complete we’ll rebuild saa7164-stable and saa7164-dev trees so they reflect the very latest source form LinuxTV.org, meaning you’ll have the very latest patches for your all of your other TV tuner cards along with support for the SAA7164 based products. We’ll continue to maintain the SAA7164 driver here at www.kernellabs.com so keep coming back for the latest news and updates.

Hmm. And after that? What about the Analog Encoder support? Stay tuned, keep reading the KernelLabs blog, the wheels of industry are beginning to turn. We hope to make an important announcement shortly.

Update: The tree was merged into linuxtv.org today, Sep 17th 2009.

Update: Patches for the issue are available in the saa7164-dev tree.

I’ve just begun testing these patches in a private tree and initial indications are very good. Feel free to try these patches if you’re experiencing bad behavior from youe HVR2200/50 driver.

As always, comments and feedback are welcome.

Update: These patches are now merged into the stable tree.

I spent six hours over the weekend logged in remotely to an Australian users system trying to track down a problem that was seemingly preventing the driver from responding correctly after being loaded. Requests into the firmware were being ignored for an apparently unknown reason. dmesg was showing massive numbers of I2C errors and general timeout warnings.

Well, after installing many different debug builds, tracing the driver manually, adding extra levels of debug, forcing certain things to happen …. nothing would work. Yes, nothing. A real head scratcher.

Yet, an almost identical board worked with tzap correctly in our development systems.

As a hunch I happened to notice that MythTV was set to automatically start. I disabled this, rebooted and tzap, dvbtraffic and everything else was working perfectly. At first glance it looks like this users MythTV configuration is breaking the driver, I need more investigation on this. I’ll do this and report back.

Why this post then? I’d also like to open up this post for comments by people who are using MythTV specifically with the HVR2200. Please describe either your successful or unsuccessful experiences. ‘Works for me’ is fine as a comment if you’re not in a writing mood.

As always, your feedback is appreciated.

The change history in the saa7164-stable repository shows that the tree hasn’t needed (or had) any digital TV related bug fixes in 2 months.

The last major piece of development work took place on May 19th in saa7164-buffering, in which I took a completely different approach to managing DMA buffering in an attempt to rule out some issues that could occur on busy systems with excessively high latency.

Over the last two months, on and off, I’ve been successfully running -stable in a second MythTV system, mirroring my production Myth schedule. I’m declaring victory. Personally speaking, if my production MythTV backend had a PCIe slot, I would of been using -stable in my production environment since June.

So, it’s no surprise to me that yesterdays set of test, each lasting 30 minutes, using a single HVR-2250 on -stable, that consisted of running a mixture of oddball/unusual LiveTV, background recording, overlapping background recordings, using every combination across both TV tuners in an attempt to find breakage, all tests completed without error. No failures.

I think saa7164-stable is ready for merge into the mainline kernel.

Yet, I still have reports from some people that the driver hard locks the system. I want to encourage anyone who has tested or is running -stable to speak up on this thread (again). Your opinions count.

Tell us, do you think -stable is ready for merge into the mainline kernel? If not, describe your specific issues, configuration and platform details. Let’s see if the community can help figure out why it’s completely reliable for most people and not for others.

As always, your feedback is always appreciated.

I’ll also be on IRC for most of the day (on and off) so if you have HVR-2250 or HVR-2200 hangs issues then feel free to contact me on irc.freenode.net #linuxtv.

Hmm, that would cause awkward things to happen and probably lead to a full system hang.

Time to add some jiffy monitoring to the deferred queue handing and gather some statistics on good and badly performing SAA7164 based systems.

Maybe the lockup is as simple as this.

Today, the only supported features of the HVR-2200 and the HVR-2250 products are the digital TV inputs. So, if you have access to DVB-T signals in Europe or Australia, or you’re lucky enough to be able to receive ATSC or Digital Cable (QAM Annex-B) in North America, Mexico or Canada then one of these two boards would be useful to you. You’re in luck. The HVR-22×0 series products are dual tuners so you can use both digital TV tuners independently and concurrently.

Today, if you need analog TV support (US cable comes to mind), remote control support, VBI / Closed Captions then you’re out of luck. The driver has no support for any of these features.

Generally speaking, the current driver is fairly stable for a large number of people but a handful are experiencing a full system lockup. I’m covering that separately on the blog so I’ll skip the detail here. Importantly, I have no plans to develop new driver features (analog etc) until we get digital TV support reliable and stable across all platforms, including 32 and 64bit. Why? It’s a question of overall quality and currently it’s not high enough for the Kernel Labs seal of approval.

Work is progressing on the driver. In the coming days / weeks I’ll isolate then fix the hang and prepare the driver for another round of testing, followed by requesting merge into the master repositories at linuxtv.org. Once this is complete then we’ll review the other missing features and outline a roadmap for implementation.

The long-term goal is to have all of the missing features working in the driver, where possible.

I also want to mention that I’m running the saa7164-dev tree on my 64bit ubuntu 8.10 (intrepid), 4GB 4 core box without any issues. I can reliably use MythTV to record many overlapping and long running TV shows, I can reliably watch live TV for over 24 hours without any issues.

The one hang I have witnessed was using the highly experiment saa7164-buffer tree, which has had zero test time – so this doesn’t surprise me.

Personally this is very frustrating. I’m still trying to figure this out.