mii related fallout

So there was some fallout from my mii push. Elxl devices had a nasty panic, which I fixed in time for build 119. Some older i82557 (iprb) devices had problems as well. The push for this fix went into build 120. To everyone affected, please accept my apologies. Build 120 should be stable for these devices.

Common MII/GMII layer integrated

Folks working with 802.3 (Ethernet) hardware (10/100/1000) can rejoice, as I've now integrated PSARC 2009/319, which provides a common MII and GMII layer for Ethernet device drivers. The first batch of drivers (iprb, dmfe, and afe) have been converted. This brings support for SunVTS netlbtest, dladm based link management, and even (for some devices) 802.3 flow control (aka Pause Frames) to these drivers.

I'll encourage folks working with other drivers to update their drivers to support the new framework. On average it cuts about 1500 lines from most drivers, and generally improves device functionality.

And yes, it works with Ethernet, even 1000Base-X should work (although only 1000Base-T has been tested.)

Any device that supports Clause 22 of 802.3 should work. Clause 45 (typically 10Gb) is not supported, however.

audiovia97 webrev posted

I've posted the webrev for audiovia97. This was covered by PSARC 2009/321. This is a driver for older Via 82C686 south bridges, used with 32-bit Via C3 and Pentium-III class processors.

The driver was written by 4Front (apparently building upon other Boomer work I've done). All I've done for this is simple packaging, cstyle fixes, and integration.

Let me know if you want early access to binaries! (You'll need Boomer or build 115 or later installed to use it.)

audiocmi integrated

I just pushed the audiocmi driver. Users with C-Media 8738, 8768, and 8338 devices can now enjoy Boomer using 16-bit stereo audio on these devices.

Note that while some of these devices can support multichannel surround, the Solaris driver for them does not support this kind of usage at present. If someone in the open source community would like to contribute support for this, please let me know.

proposal to change policy of SPARC deliverables

Historically, subsystems (drivers, etc.) were supposed to deliver on both SPARC and x86 platforms (and now amd64 as well) unless a really good argument was supplied. While I've long been a supporter of this philosophy, I think a time is coming to consider making a specific policy exception.

The problem I'm running into is with legacy PCI devices. Some of these legacy PCI devices can work SPARC. But devices that don't support 66 MHz frequently (but not always) have problems on SPARC workstations. The biggest problem is that many SPARC workstations have 32-bit 33 MHz slots that don't supply 3.3V. This affects quite a number of cards that require 3.3V power to operate. Some devices will operate with only 5V, but only marginally. I've seen a number of failures that I think can be traced to this problem, and now I don't recommend using PCI devices not specifically qualified for these platforms with SPARC systems. (Sun makes that same recommendation, btw.)

Furthermore, identifying suitable SPARC hardware can be a challenge. SPARC systems these days are slower than their x86 counterparts, yet for the most part remain more expensive, and there are few inexpensive options available for open source developers. And, SPARC desktop systems are effectively off the market -- Sun doesn't have any current SPARC desktop models for sale, and that's not likely to change anytime soon.

So what I'd like to propose is a policy change that recommends authors working on drivers for legacy PCI cards be given a blanket waiver for SPARC delivery. (Examples are recent NIC drivers such as "vr", "sfe", and "bfe", and audio hardware such as "audiopci", "audiocmi", and perhaps in the future drivers for Creative Audigy cards.)

I still believe that developers working with modern PCIe devices should deliver both SPARC and x86 binaries though. PCIe is available on last generation SPARC desktop (Ultra 25/45) hardware as well as current generation server products.

driver private headers

Some may have noticed in a few of my changes lately that I'm moving some header files around. This post explains the change (and encourages others to do the same).

Header files that are only useful to one subsystem (a kernel module, or driver, for example) should, IMO, only be located in the subsystem for which they are used. For example, do we really need to ship a header file in /usr/include/sys/ that has all the register definitions for the DEC tulip ethernet (dnet.h)? Wouldn't it be sufficient to have those definitions just where they belong, alongside the source for the dnet driver itself?

Locating the files for such subsystems in their own directory instead of a common directory (e.g. common/io/dnet/dnet.h instead of common/sys/dnet.h) has several positive ramifications:

1. It makes it readily apparent that the header file has no content needed outside of the kernel or subsystem. So changes can be more freely made (no userland dependencies, for example.)
   
2. The header file is not delivered to customers (except as part of the entire source code), shrinking the amount of disk space consumed on development workstations.
   
3. No exceptions entry is required to suppress the file from delivery - so the file is referenced in fewer places. (Fewer places to change if the file needs to move, be removed, etc.)
4. The header file is easier for humans to locate with the source.
5. The header file is faster for the compiler to locate -- it will be in the first directory searched - slightly faster build times. (If enough subsystems do this, it might start to make a noticeable improvement.)
   
6. The common include directory becomes slightly more manageable with each such change. (Big directories with hundreds of files are awkward.)
   
7. The file can't be used/abused by other subsystems (at least not without making such abuse obvious), because it really is in a private directory.

The upshot of all this is that a bunch of driver-private header files have been slowly moving out of uts/common/sys/ and into better locations. Hopefully other driver developers will consider doing the same.

driver.conf files considered evil

Just a quick note on driver.conf files that some device drivers deliver.

I believe that driver.conf files, and the tunables that are usually put in them, are for the most part, a byproduct of inadequate architecture The average driver should use no tunables in driver.conf properties.

As a result, I've started going out of my way to remove driver.conf files from device drivers for which I'm responsible. I think that even the "advanced" tunables (such as the interrupt rate used for audio drivers) falls outside the scope of what users should normally tune, and that delivering a configuration file is actively harmful.

One of Solaris' great strengths historically has been the approach to "self-tuning", so that lots of configuration is not necessary. I'd like to see that continue. The next time you think about adding a driver.conf property, consider carefully if there might not be a better solution. (Either self tuning, providing a default that works for everyone or nearly everyone, or using another system like Brussels to provide configuration overrides.)

(PCI devices in particular are very painful to configure with driver.conf... see the pci man page to find out why!)

audiocmi webrev posted

I've posted a webrev for the audiocmi driver. This driver supports CMI 8738 and similar parts. This is for Boomer, and I hope to get it integrated into build 117. If you want to test a binary, drop me an e-mail!

spwr opensource and GLDv3?

If you'd like the spwr driver to be open sourced and GLDv3 compliant (with VLAN support, link notification, dladm and ndd support, and the rest of the goodies), and have a few spare cards that you can give me, please feel free to contact me. I'm willing to do the work as a side project.

I'm also interested in obtaining 100Base-X fiber cards, and fiber cables (VF45 format), especially if someone also has access to the 3M VOL-N100VF+TX card (which is theoretically supported by "afe", but which I'm not entirely convinced works properly because I've never actually had a card on hand that I could test!) I need cards and cable... at the moment I have neither.

Note that this is not Sun commissioned work, but an interesting side project that caught my interest. If nobody replies with hardware, then its unlikely that I'll do any of the work.

Going to CommunityOne After All

Well, it appears that I'll be at CommunityOne (Monday only) after all! This is going to be a fairly big event -- the big event for OpenSolaris this year, I think. I'm looking forward to the chance to meet a bunch of folks that I didn't get the chance to meet previously. Hope to see you there!

hme for x86 RTI submitted

I've just submitted the RTI for hme. If the RTI advocate approves it in time, it will be in build 115, otherwise in build 116. This will allow you to use your old PCI qfe boards with OpenSolaris on x86 systems.

It also represents a significant simplifications of the code. Thanks to the folks who've helped with testing!

audio1575 driver for x86

This driver, with surround sound (5.1) support is pushed into Build 115. Stay tuned.

iprb suspend/resume and quiesce support in b115

I've just pushed an updated iprb with suspend/resume and quiesce support. Its still in the closed tree, but even so this improvement should help out folks who are stuck with one of these on their system board. Enjoy.

on the evils of auto-bounce/discard mailing lists

Some of you have already seen this rant from me. But I think its important enough to bring to the greater attention of the community.

Some mailing lists in OpenSolaris are configured to automatically bounce or discard messages sent from a mailing list that is not subscribed to that list. This is, IMO, a fairly toxic configuration.

At first glance, the idea seems good -- if your list only accepts member submissions, then you'll not have to deal with all the spam, and you don't have to moderate. The list can basically run from that point completely unadministered. Sounds good, doesn't it.

The problem is that this configuration is toxic to many discussions and to some users. As an example, I tend to get involved in many cross discipline conversations -- partly as a result of my membership on ARC.

And yet, my replies, often to important discussions about cases that might be interesting to certain communities, are often bounced back, because I simply am not subscribed to those lists. I don't want to subscribe to every list -- and I shouldn't have to in order to be an effective ARC member. A lot of times I just give up -- so communities are missing out on relevant conversation because of these configurations. This is a serious impediment to collaboration.

But it goes beyond that -- I'm also known by at least four different e-mail address -- one personal address, one @opensolaris.org, one first.last@sun.com, and one nickname@sun.com. People know me by all of those addresses -- so I'll be CC'd on conversations using any one of those addresses. When I reply, unless I am careful to remember to use the address I've subscribed to the list as, it will bounce on certain lists. Now its been pointed out that I can fix this by subscribing all of my addresses to the mailing lists I'm member of -- but who wants to subscribe to every list they're on 4 times? This is a serious impediment to collaboration.

The other situation is when someone has some new bit of information that they would like to bring to attention to a group of individuals, or ask a question, without having to be a member of the group. If I think I've found a bug in the TCP stack, should I have to be a member of networking-discuss@ in order to ask the group about it, or post the information? I suspect many such newbies hit the auto-bounce barrier for some of these groups, and just give up. The threshold for participation is simply too great. This is a serious impediment to collaboration.

So, all of these issues seem like they are negatively impacting collaboration. What is the solution?

Easy: moderate your lists properly. For heavily trafficked lists, it might take a few minutes a day to do this, but configure the lists to hold posts from non-members for moderation. If you identify a couple of volunteers to share the list password with, you can spread the chore, so that it is not too onerous for any one individual.

Those of you list owners with auto-discard/bounce set, please consider changing to a regular moderated list format. As attractive as the idea of a configuration where you don't have to do any work is, such configurations are actually hurting the group.

I'm done ranting about this for now. Thank you.

giving up on iprb

Due to snafus with the various legal departments involved, we seem to have hit another roadblock getting iprb open sourced. (Why is it that whenever lawyers get involved, everything seems to take five times longer than it otherwise would? Do corporate lawyers bill by the hour just like their normal "free-agent" counterparts?)

Rather than continue to wait for this to get resolved, I've decided to move forward and get some of the important fixes into iprb even though it is still closed source. Most notably, my changes allow iprb to support suspend-to-ram, and fast reboot.

I've submitted the RTI for these changes, so hopefully this will be in either build 115 or (more likely) build 116. As far as opening the actual source up -- even though there is nothing in the source that is not already made public by Intel -- I'm not holding my breath.

audio1575 driver for x86

I'm getting ready to submit an RTI to add M1575 (Acer/Uli) support to OpenSolaris. As part of the work, I've taken the M1575 driver from OpenSolaris (which was used for the SPARC Ultra 25, and 45 workstations) and added quiesce() and multichannel surround support. I'm told it works nicely with 5.1 channel surround. If you have this chip (e.g. an ATI SB200 motherboard), let me know and I'll see if I can get a binary to you. (You'll need to have Boomer RC3 or build 115 installed, though.)

7zip saves the day

I recently received a document that was LHA archived. I was struggling to find a suitable decompressor, since Gnome archive-manager complained that the format was unsupported.

Nicely, though, 7zip had no trouble uncompressing the file. And, 7zip is stock on all recent builds of OpenSolaris. Very cool. :-)

hme/qfe x86 and SPARC binaries

hme (and by extension qfe) cards can be used on x86 systems if you download the test driver that I've supplied. Just download, extract, and follow the directions in the README file.

I'm also interested in verification that the driver performs properly on SPARC; binaries are in the same archive. The code was significantly simplified, and I need to know it works properly.

Finally, I'm seriously in need of code reviewers. I can't integrate the changes without a proper review. If you can volunteer to help, check out the webrev that I have posted.

Thanks!

(Note: on x86 qfe cards will show up as "hme" devices... this is normal and expected. The "qfe" name is an artifact only found on SPARC systems for historical reasons... the actual qfe card consists of 4 hme devices behind a PCI bridge.)

Via Rhine driver integrated

Joost Mulders last night integrated his Via Rhine "vr" driver. While I don't personally have this hardware, I know it is very common. I did code review his driver, which I found to represent very tight and well-written code. Congratulations Joost!

Saurabh Misra is working on a home-grown "bfe" (Broadcom Fast Ethernet) driver as well, which also looks very good and should be integrating soon.

ALI 5451 southbridge AC'97 audio integrated

I just integrated support for the ALi Southbridge AC'97 controller found on certain x86 motherboards The necessary driver (audiots from the SPARC world, actually) is available on x86 platforms starting with build 115. For the moment this driver only supports stereo, but if someone has a multichannel configuration that they want to have supported, they should contact me and I'll work with them to get the testing necessary to enable this done.

Enjoy.

kaBOOM!

Boomer has integrated! Yay... this was a long road, and a lot of code -- one of the biggest projects I've worked on (and the hg push was far and away the single biggest push or putback I've ever done.)

Anyway, we were the first integration into build 115 of ON. See the flag day message for more details. Enjoy!

-- Garrett

preliminary audio driver for cmi8738

I've got a preliminary driver working for the CMedia 8738, 8768, and 8338 series of devices. This driver works with Boomer (you'll need RC3 installed, or -- hopefully -- Nevada build 115 when it comes out, and it seems to be working except for some issues with the record gain being a bit faint. (Hopefully I'll resolve those soon.)

The driver lacks support for advanced features like surround sound or jack retasking, mostly because I can't easily test them -- I have only an ancient 8738 -033 (stereo only) part available to me. (Also, it seems that CMedia has had some rather unusual changes with each revision of the chip, making proper support of advanced features such as SPDIF or surround sound a bit troublesome.)

So, that said, I'd like to hear from folks who have this part and can use a driver for it. If you want to test, drop me a line. Thanks.

Going to KCA 2009

It looks like I'll be in Brisbane, Australia for the Kernel Conference Australia 2009. The primary mission is to talk about Boomer -- the exact details aren't locked in yet. But I've been asked to chat about GLDv3 and driver porting in general as well.

Anyway, for those of you down under, I look forward to the chance to meet you. This will be my first trip south of the equator, and I'm looking forward to it. Hopefully my wife will be able to join me there as well!

boomer rc3, audio driver for ali5451

I've posted Boomer RC3 yesterday. It will be the last separate binary release of Boomer Phase I -- we're on track for integration into Nevada B115 -- C-Team Commitment review is on Monday.

That said, we've found one problem, which is of a P3 nature, which affects the ability of certain applications (audacity is the only one found so far) to record -- the record process doesn't start properly when triggered. I've got a fix for this, and can offer a binary to anyone who needs it for RC3. We've still not yet figured out whether this problem will be fixed in Boomer Phase I prior to integration, or just after integration -- but in either case it will be build 115 as well.

The other thing is that I've got a binary driver built for the audio controller found on the ALi 5451 southbridge. This is a somewhat older part, but folks might have them on certain motherboards. If you've got one, and you want a binary driver, let me know -- I'll be happy to supply one -- but you'll need to run Boomer RC3 or an OpenSolaris build that has Boomer integrated (i.e. build 115). We'll be handling this one as an RFE after Boomer Phase I integration -- probably also in Build 115. (Note that Dev at 4Front has already tested this binary for me, so I know its good to go -- it is just a recompilation on x86 of the audiots SPARC driver that is in Boomer.)

Anyway, keep your fingers crossed for our integration into build 115.

rtls open sourced!

I just pushed

6822752 rtls should be open source

It will be in build 113. There are some possible projects contributors could do involving that code (such as conversion to Brussels). Contact me if you're interested.

I've Arrived

Today something cool happened. Casper Dik asked me to review some changes he had to fix the Cardbus support in the pci nexus code.

What's so cool about that? Well, its the first time I've ever reviewed anything from Casper.

Casper, from back in the mid-90's, was kind of a folk hero of mine (and not just me, I think) back when I was doing system administration in my prior life at Qualcomm. I'd never have believed some 15 years ago that Casper would be asking me for code review.

Cool. :-)

Another Boomer Release Candidate Posted

I've posted an updated release candidate 1b for Boomer today, and you can download it from the files area.

The only remaining known P1-P3 issues in this release are a problem that affects users of TX (Trusted Extensions) and lack of BFU support. Obviously we're feeling pretty good about this release, and we're on track for integration in b115.

As usual issues should be reported using the Bugster category development/audio-ng/software. Alternatively, you can send mail to opensound-discuss@sun.com.

Boomer Release Candidate 1a Available

I've posted a release candidate for Boomer today, and you can download it from the files area. This release candidate includes USB support, and fixes for some critical bugs, so if you installed the beta, please upgrade.

There are still some known issues, but we believe we're quickly closing to being ready to integrate in Build 115.

Bugs should be reported using the Bugster category development/audio-ng/software. Alternatively, you can send mail to opensound-discuss@sun.com.

Enjoy!