In the Bay Area this week

I'm up in MPK this week. (Wed, Thur, and leaving Friday night.) If any fellow Solaris geeks up here are up for a pub outing, e-mail me.

doing my part

I just committed a change to move 7 previously closed drivers in Nevada to the open source tree under usr/src. This change involved nothing more than Makefile and copyright block editing, so it was pretty much a no-brainer. (Though the heavy lifting of the legal review had already been done.)

The drivers moved were: bscv, bscbus, i2bsc, gptwo_cpu, gptwocfg, todstarcat, and todm5819p_rmc.

Admittedly, none of these are likely to exist on your hardware, but it does help to have more bits open. Hopefully someday /usr/closed will either cease to exist or become its own consolidation separate from Nevada.

eri conversion update

Hmm... it looks like I never posted the webrev... well here it is, the webrev for the eri(7d) conversion to Nemo.

Now, the second bit of good news here is that the PSARC case for this as been submitted as PSARC 2007/243. Note that the case isn't published publicly at the time of writing, but it should be soon.

afe and dmfe cases approved

FYI, the afe and dmfe cases I had at PSARC (2007/229 and 2007/221 respectively) were approved. I've already put back the dmfe code. The afe code will be committed by Alan DuBoff. I've got pre-approval to do a follow-up putback to convert afe to GLDv3 afterwards.

Note that as a result of Crossbow, there are some changes coming in GLDv3, so it is still inappropriate to use GLDv3 for unbundled drivers. (The biggest of these changes is support for "polling", where the network stack can disable interrupts on the NIC and run a separate thread to poll the device for inbound packets. On extremely high traffic systems, this can have a big impact on overall system throughput by avoiding the extra context switches.)

afe PSARC case number

The PSARC fasttrack to integrate afe into Nevada was assigned case number PSARC 2007/229. Notably, this case was not submitted by me (I'm not even on the interest list!), and is being done as a result of the BSD license terms for afe. It will probably be reviewed at next week's PSARC meeting.

Death to IEN-116

Finally, over 20 years since the late Jon Postel said Death To IEN-116, we have finally removed it from OpenSolaris. Who says changes in Solaris take too long?

eri tests look good.. call for more testers

As predicted, the area of biggest risk in my conversion of eri to GLDv3 was in fact the kstat handling. However, I appear to have that all worked out now, and the binary is working flawlessly on my SunBlade 100. Even suspend/resume works fine. However, I've not yet integrated this code properly into a workspace to generate a webrev, but I will do so soon. (Probably tomorrow... I'd like to get my two other RTIs put back first.)

One of the biggest concerns about this effort was the added risk that doing this conversion might bring to the "stable" eri driver. So, I'm asking the community for help. If you want to help out with testing, especially if you have higher end systems or want to do some benchmark comparisons, please let me know.

(I don't have specific test suites to give out that this time... its of more value frankly to have people using their own tests right now, that way we get broader test coverage than perhaps we might with a single test suite.)

Please let me know. Thanks! (Oh yeah, if you have an eri you want to try with new GLDv3-based 802.3ad link aggregation features, I'd be game for that, too!)

(PS. An obvious consequence of this effort is that it will be easy to do the work to convert hme, gem, and qfe, which share a lot common heritage with the eri driver. So, maybe there is yet hope for those, as well.)

GLDv3 experiences

I've just finished (still testing!) my port of eri to GLDv3. Between that and eri, and looking at existing GLDv3 drivers (bge, rge, e1000g), I think I have gathered some operational experience that I hope we can use to improve Nemo. (So, anyone who says my time spent on converting eri was wasted is wrong... because if nothing else it gained some more operational experience with GLDv3.)

Executive summary of the takeaways I have gotten so far, that I think are worth noting:

• There is still a lot of code duplicated across even GLDv3 drivers (more below)
• Lock management is so much simplified
• GLDv3 kstats need "work"
• we really, really need Brussels... it can't come soon enough.
• some drivers can probably be changed internally to work even better with GLDv3 than a naive port

So here's the detailed stuff.

1. code duplication
   
   The duplicated code falls into three major areas. ioctls (mostly ndd(1M) and loopback handling for SunVTS), kstats, and MII. For now I want to focus on the MII bit. It turns out that pretty much every Ethernet device on the planet talks to a transceiver (whether integrated into the same chip as the MAC controller or not) using MII/GMII. We have tons of logic surrounding MII and GMII replicated across each driver, and frequently the decisions made by one driver are different than those in another.
   
   There exists an old i386 driver called mii, which was an abortive attempt to create a common module/framework for MII and PHY handling. (Only used by the obsolete dnet driver at present.) I think this should be revived. Its been shown to work well for BSD Unix (at least NetBSD, but I'm pretty sure all of them), and it would really help simplify a lot of code. The eri driver, for example, probably has a couple thousand lines of MII related auto-negotiation logic in it.
   
   And of course, each of these negotiation frameworks takes a slightly different set of tunables and configuration parameters, exports different statistics, etc.
   


2. Lock management is so much simplified
   
   It's reallyeasy to write a GLDv3 driver that doesn't hold locks across GLDv3 routines. I suspect a lot of deadlocks/hangs/panics are going to be solved by moving drivers to GLDv3. (Of course, we've seen locking problems higher in the stack as a result... see recent deadlocks in dls, etc. But we only need to solve those once with GLDv3. Yay.)
   


3. The kstat framework for GLDv3 is just plain broken.
   
   There are several problems here.
   
   
   • All kstats for a media type are included, regardless of whether or not they make sense for a specific device. For example, the cap_rem_fault is not supported by most of the drivers yet, but yet, when the driver doesn't have support in mac_stat(), the statistic is included in kstat output as 0. However, pretty much any system with an 802.3u compliant MII does in fact support the rem_fault MII field. So in this case, just because the driver isn't exporting the stat, the framework is creating an outright lie. This is probably true of other stats as well. For example, if hardware isn't prepared to report runt_errors, then it doesn't make sense to claim that value as "zero".... because you might be flooding the device with bad packets, which just get dropped on the floor (perhaps getting accounted in some other, less granular "BadPackets" counter or somesuch.) Better to say nothing than to tell a lie, IMO.
     
   
   
   • kstat's are normally "snapshotted", so that you can take a snapshot of all stats in time at once. This is common with some hardware devices, too. Getting these stats may be expensive though. (For example reclaiming transmit buffers, so you can collect transmit status, etc. Acquiring locks. With some devices you might even have to do an expensive collection effort that would normally cv_wait for an interrupt.) Having to go through this several times (once for each stat collected) for a single snapshot is ... inefficient. It would be nice to add a mac_stat_update() entry point, which is separate from the mac_stat() entry point. (Even better, also add a mac_stat_done() to release any resources acquired by the first call.) The good news, I think, is that hopefully we aren't going to have to support DLPI DL_GET_STATISTICS_REQ, so it should be safe to cv_wait in mac_stat() related calls now (unlike with older GLDv2.) We aren't supporting the DLPI statistics calls, are we? Please say we aren't....
     
   
   
   • If the driver wants to export any additional driver-specific statistics, it has to do the whole kstat dance itself, in addition to the nemo mac_stat() entry point. Lets try to find a way for drivers to export/register additional driver specific kstats within the existing nemo framework, please?
     
   
   
   • Duplication. E.g. for bge, there is a "bge0" kstat, created by dls, as well as a "mac" kstat created by the mac module. Both of these will have some common counters, like ipackets64, brdcstxmt, etc. What's worse, one stat in particular, "unknowns" is counted by the dls framework in the "bge0" stat, but is not counted by the "mac" stat. This can lead to confusion. The duplication also makes worse the snapshot problem already mentioned, since it appears that most of the stats are generated just by calling the mac_stat() a second time for the same values already recorded in the "mac" kstat.
     
   
   
   • Inadequate list of kstats in the default set. I found several kstat which were missing. We got several of them getting fixed as a result of PSARC 2007/220, but I've since found a few others. E.g. Ethernet devices commonly can detect "jabber timeouts". These should be reported somehow. Also, stats about network related interrupts are really important, and aren't included by default. I consider this a significant shortcoming. I guess devices should register a KSTAT_TYPE_INTR kstat, but approximately none of them do today.
     
   
   
   • Stat cleanups in drivers. This is mostly a driver-specific problem, but look at the kstat output on bge and e1000g, and see what I'm talking about. There is a total lack of consistency here.
     
   
   
4. We really need Brussels.
   
   From the above, you see the problems with kstats. There are similar problems with NDD. The amount of code scattered around different drivers trying to figure out NIC tuning is boggling. And most of it isn't what you'd call "sterling examples of quality". The eri driver was full of some really, really fragile code in this. (Deleting one tunable ... the instance ndd parameter... required updating no fewer than 4 different locations in the driver. And they weren't conveniently co-located.
   
   Interpretation of values, handling, all of it is terribly replicated across so many drivers. I can't wait to eradicate this crufty, horrid code, and replace it with something nice and sane from Brussels.



5. Some drivers can change internally to work even better with GLDv3.
   
   In eri, for example, I think we can be smart on the transmit side, so that, for example, when a group of mblks comes down, we don't kick the hardware and resync the descriptor rings until all the packets are queued for transmit. This would help amortize some per-packet expenses across multiple packets.
   
   Other drivers can benefit from multiaddress support. dmfe falls into that category.
   
   That said, my approach so far has been the naive conversion. I'd like to revisit a few of them to enhance them to take advantage of the superior design in GLDv3, but first I want to get them put back.
   

dmfe crossbow conversion

In case you ever wondered what it takes to convert a "simple" GLDv2 driver to Nemo, have a look at the webrev I posted earlier today.

I'm hoping that this work will get integrated soon. As an upshot, dmfe with this change "just works" with dladm show-dev.

report from the battery team

I'm now a member of the "battery team". I had a very productive con-call with the folks involved, and I think we are going to soon have a better common framework for battery APIs in the kernel so that SPARC systems can also take advantage of the gnome battery applet. Watch this space!

afe integration web rev posted

For the curious, I've posted a webrev containing the changes required to integrate afe into Nevada.

The driver includes changes from the stock AFE driver for Solaris, including some lint fixes, and changes to use the stock Solaris sys/miireg.h.

I'd love to make more changes to this driver, but at the moment I don't want to cause a test reset. Once the driver is integrated, I have a bunch more improvements coming... Nemo, multiple mac address support, VLAN support, link notification support (needed for NWAM), as well as code reduction by using some features that are now part of stock Solaris (like the common MII framework!)

Tadpole SPARCLE support putback

Core support for SPARCLE was just putback! I'm getting ready to post an initial tadpmu for public review soon, as well. This should make you SPARCLE/Sun Ultra 3 owners out there happy.

Not All Broadcom GigE's are Equal

Recently, I posted a blog entry where I described that "Not All GigE Are Equal", strongly advocating the use of Broadcom GigE devices when faced with a choice.

However, after spending time in the code, I've discovered that there is quite a range of differences amongst Broadcom gigE devices.

I had considered listing a full table of them, but it seems that would be a bit onerous. Take a look at usr/src/uts/common/io/bge/bge_chip2.c if you want to find out the gory details. But in the mean time, here are my recommendations:

If you have PCI or PCI-X: Choose a bcm5704 if you can. It has pretty much full feature support, but you need to pick a recent revision (newer than A0.) Look for pci ids of pci14e4,1646, pci14e4,16a8, or pci14e4,1649. These chips alls support PCI-X, multiple rings, full checksum offload, and multiple hardware tx and rx rings.

If you have PCIe: As far as I can tell, all of the PCIe chips that are Solaris supported lack support for multiple hardware tx/rx rings. This is really unfortunate, as it will have a negative impact on Crossbow benefits. But apart from that, it looks like the 5714 and 5714 series are your best bet. They both support jumbo frames, and they both have full checksum offload support. Look for pci ids of pci14e4,1668, pci14e4,1669, pci14e4,1678, or pci14e4,1679.

What this really says, is if you have to choose between a PCI-X card and a PCIe card, surprisingly, choose the PCI-X card (if you can get a 5704). Save your PCIe for framebuffers or HBAs. (Or, better, 10G cards like Neptune.)

blogger Atom bugs

As part of setting up the Tadpole project, I tried to use a feed direct from Blogger, but the OpenSolaris tonic infrastructure doesn't like it. Apparently the feed has some problems, which you can see by looking at the output from feedvalidator. Anyway, I was able to work around by using feedburner to convert the blogger Atom feed into a clean RSS feed. Maybe at some point some Blogger staff will look at this and see what the problem is.

hackergotchi... thanks Gman!

Gman (Glynn) made a hackergotchi from a photo I sent him, which is used on planet.opensolaris.org. His gimp-fu is great. Thanks Gman!

Tadpole project live

The Tadpole project is now live!

First Tadpole code review posted

The first review for Tadpole platform support is online now. Please let me know your thoughts, after reading it. There will be more good stuff coming soon, I hope. (Also, if you have a Tadpole platform other than a SPARCLE or UltraBook IIi, and are willing to test, please let me know!)

Thanks!

Who's Who?

I just received two e-mails (identical to each other) stating this:

Dear DAmore Garrett,

The Heritage Registry of Who's Who is recognizing you for possible inclusion in the upcoming 2007-08 Edition. Please go to http://theheritageregistryofwhoswho.com and click on the invitation button.

Thank You,

Chris Jespersen

I'm not entirely convinced this is a worthwhile thing... but I'm willing to play along until they ask me for money. Anyone else out there received these before?

Tadpole project proposed

FYI, I recently proposed a new project to track improvements to support for Tadpole platforms in OpenSolaris. It looks like it got the seconds needed, so I'm just waiting for the infrastructure to be created.

first putback!

I just made my first putback to ON (6487387 pcic driver contains obsolete & private Tadpole code that should be removed).

While this is nothing earth shattering, hopefully I'll be making a lot more commits soon.

Inland Empire Solaris Users?

I've been wondering how many other OpenSolaris users there are out there in the Inland Empire. I recently met one close to me, which surprised me quite a bit. I figured I was the only one within at least a 30 mile radius.

If there are others of you out there, please drop me a line. I'd like to inquire as to whether it makes sense to consider starting a User's Group for the area. Possibly we could join up with any other User Groups for Southern California.

For the record I live in southwest Riverside county, not far from Temecula and Murrieta. (For those of you not familiar with the west coast, the Inland Empire refers to a large region of southern California that is separated from the coastal areas of Orange and Los Angeles counties by a range of coastal mountains. I often have joked that I'm about 65 miles from any natural technology center, but now I'm not so sure. And I think a lot of people who commute to places like San Diego and LA live out here.)

ancient history (IEN-116 must die!)

Funny note. When I came back to Sun (two weeks ago), I discovered that an ancient PSARC case (2002/356) for the removal of the Trivial Name Server (in.tnamed) had never been completed. So for 5-odd years since we've continued to ship this long-since-obsolete protocol. I'm going to go ahead and drive forward with the actual removal... at the time I did it as a case study in how much process was involved with even a simple EOF. Lets see how long this one takes. (For the record, the IEN-116 protocol was obsolete as far back as 1986, when J. Postel first requested vendors ditch it.)

afe and mxfe pending updates

Those of you using afe (and also mxfe) will be pleased to note that the time is fast approaching when afe will hopefully be integrated into Solaris Nevada. There is a PSARC fasttrack scheduled for it next week if I understand correctly. (I don't have the case number yet.)

There are a few ramifications of this. One of the most immediate is that I'm going to be winding down support for versions of Solaris earlier than 10. In fact, I no longer have any personal installations running anything less than S10u3, and most everything is running Nevada.

The other reason for me to do this is so that I can immediately start taking advantage of some features that are present in Solaris 10 and Nevada. For example, I want to add support for DLPI link notification, and ultimately (in Nevada) port to GLDv3.

The GLDv3 has some compelling features, and as a result afe and mxfe will gain support for features like vlans, jumbo frames, and interrupt blanking. And, they'll also benefit from the increased performance gains afforded by the GLDv3 framework.

It isn't clear to me that I'll be supporting GLDv3 for Solaris 10 (the interfaces are not yet public), but at least in Nevada I will. And even for S10, I'll probably be using new GLDv2 features that are not available to older releases. (Like the DLPI link notification.)

Before I do this, I will be spinning one last significant bug fix release for afe and mxfe, which addresses several significant bugs found by Sun's QA group. (Including the fact that afe has not functioned properly with multicast since it was first written!)

Watch the web page for more details.

Not All GigE Are Equal

As a consequence of work I've been doing lately since I joined Sun, I've learned some things that folks who care a great deal about performance might like to know.

The most important of these is that not all gigabit cards are created equal. And even among those that are, some of them get preferential treatment at Sun.

One surprise: the gigE device that gets the most preferential treatment is not a Sun branded NIC. In fact, its a device that you can readily find at your local computer retailer.

I speak of bge.

The bge (Broadcom) NIC has some very, very sophisticated logic on it, that Crossbow is going to be able to take advantage of to get you some very nice performance acceleration, plus some greatly added support for QoS and stack virtualization. If you're thinking about a NIC, my first choice would be a Broadcom NIC.

The Cassini (Sun Gigaswift) has many of the same features, but costs more, and is harder to find. And, the Cassini isn't supported by some of the crossbow features -- yet. This issue will of course get resolved, but for the immediate now, your best bet is a broadcom NIC, especially if you want to run Solaris.

The other commodity NICs (RealTek 8169, Intel Pro/G, etc.) are certainly nice enough, but the features on these nics are an incremental update over similar 100 Mbit hardware, and don't hold a candle to the separate hardware rings, advanced classification engines, and similar features present in the broadcom and cassini hardware. (Noteably, these features will be more important with 10G NICs, and devices like Neptune -- Sun's 10G offering, will be featuring them prominently.) And finally, now that 10G and stack virtualization need these features, Solaris is going to start taking advantage of them. Some of this is already in Nevada, and more is on the way soon.

I wouldn't be surprised if other high-end NIC developers (Intel? Marvell?) start offering these features in future updates, although I expect some players (such as RealTek) will continue to focus on much simpler (and hence cheaper) devices.

Congratulations new OGB

Its official!

The OpenSolaris Constitution has been ratified. Yay!

Congratulations to the new OGB as well. I'm generally very pleased with the election results, despite not getting elected myself.

A few interesting tid bits:

1) Rich Teer is the only non-Sun OGB member. (And apparently he has done some work for Sun.)
2) The entire OGB seems to be made up of engineers.
3) Neither female member was elected.
4) There are two ARC members sitting. (James Carlson and Alan Coopersmith.)
5) There seems to be good geographical representation... i.e. MPK (and SFBay in general) don't seem to be overweighted.
6) Several members have sat on other FOSS boards (at least Glynn and Alan)

I would have liked to see someone with more marketing and program management elected.

In future years, I'd like to see the process for Core Contributor grants revised. I think only folks who are active in the community should have this role ... I think a lot of people got grants just because they committed code as part of their day jobs at Sun. Also, there should be a limit to how many Core Contributors a given community can elect... the large number of user-group contributors could have had an adverse effect on results. I would also like to see 1) term limits, and 2) limits on the number of members working for (affiliated with) any one employer. (Not more than 3.) But those are ideas for the new OGB to think about.

Again, congratulations to the new OGB, and a BIG thank-you to everyone who voted (regardless of whether you voted for me or not.)

Just Voted

I just cast my ballot for the OGB and ratification of the constitution.

I will not tell you who I voted for, but I will say two things.

First off, the decision was hard. There are some excellent candidates running. I'm pretty confident that we are going to have a great OGB, made of up reasonable individuals who are passionate about OpenSolaris. (Yes, I did vote for myself, but I also voted for quite a few other people... there are 7 seats, after all.)

Second, I did vote to ratify the Draft Consitution. I hope you did, too.

The window of time to cast your ballot is quickly drawing to a close. Polls close on Monday, so be sure to cast your ballot before then. Unlike some others, I waited a bit, primarily because I wanted to hear what some of the other candidates had to say. So even if you have not voted yet, please do so today. Even if you abstain from voting on the candidates, at least make a statement on the ratification of the constitution. I believe it is important to get at least a 51% turnout. I believe there are around 260 eligible voters, so far only 85 ballots have been cast.

So, as Glynn says, think of the kittens and VOTE!

first day at Sun

Apparently, I am not the only one who started here at Sun yesterday. Auspicious? Wait and see...

Meantime, I spent yesterday meeting the group, and getting to know what I'll be working on. There's some exciting stuff going on, and while I can't really talk about it now, the good news is that eventually the good stuff will make into OpenSolaris.

OGB interview

Simon Phipps (aka webmink) interviewed me yesterday as part of the series of interviews he is giving to the OGB candidates. This is an excellent way to get some idea of each of the candidates; if you're a Core Contributor, I recommend checking his blog out.

OGB Postition Statement

As you probably know, I'm running for a seat on the OpenSolaris Governing Board (OGB). I've answered a number of questions already, but I really like how Keith Wesolows has set up a position statement, so much so that I'm going to do the same. Much of my layout and presentation follows his, but rest assured that there are real meaty differences here. :-)

Before I do that, after reviewing a lot of the material that has been submitted by nominees, I do think there is a really good group of candidates who are as passionate and level-headed as I am running... its likely that I'll be happy with the board that is elected regardless of whether I'm on it or not. Now, on to the details... I'll start with my positions on things, along with a short bio at the end.

• The Constitution

YES.

The document isn't perfect, but it is an excellent start, and we can address issues with it as we move along. Without it, the OGB cannot exist, and the project will not be able to exist as a legitimate entity independent of Sun.

• Community Structure

We can do better.

The biggest issues I have with the way communities and projects are laid out now, is that it seems very ad-hoc. However, I'm not sure that we should throw the baby out with the bath water. I think there is a need for communities to exist without specific code bases, and many projects may have efforts which span otherwise logical code boundaries (e.g.. work in both the ON consolidation and the Gnome consolidation.)

Simply providing better guidance on the creation of groups, the conditions under which a group should be terminated, and clarifying how projects are created and terminated can be done.

I do believe that every Project should have a Group that "owns" it, and that some of the way that we have created Groups (using DTrace as an example) should really have been projects under some umbrella Project.

• Change control, SCM, and ARCs

Open it up!

Historically "significant" changes to Solaris have had to go through ARCs. (Architectural Review Committees.) As painful as the process can often seem to be, I think that the end result has been the creation of the finest operating system on the planet. So I don't want to muck with it too much.

However, I believe that if OpenSolaris is going to really be Open, then the review process for it needs to be open, and that these ARCs need to include representation from stakeholders in the community. And I think there can be some minor tweaking to help the community in dealing with this process. I would consider this the first step that OGB should be pushing Sun for, simply because it does not require any change in the technology or tools to make it happen right now.

Ultimately, the community needs to have the ability to access the code directly, and there needs to be some powers that are delegated to the community, such as RTI advocacy. More on that later.

• Licensing

No simple answer.

This seems to be a hot-bed of controversy. Let me state some things unequivocally. First off, I do not believe that changing licensing is at this time likely to significantly improve mindshare or marketshare of OpenSolaris. Second, I do believe there is a real risk of a license-driven fork if the "core" of OpenSolaris is dual-licensed. Third, for various reasons, I believe that wholesale relicensing of OpenSolaris to GPL (either v2 or the draft v3 that I've seen) is bad idea that will ultimately severely harm a number of commercial stakeholders (among them Sun, but not just Sun.)

However, I'm not a license zealot, and I believe that there may be cases where dual licensing to enable certain components of OpenSolaris to be used in foreign projects may ultimately be in the best interests of the project. If the goal of a project is to have as many people using a certain technology, then it is also in the best interest to eliminate the barriers to use. For example, the project may decide that Linux adoption of DTrace is something that has concrete benefits. In this case, those portions could be dual licensed, on a case by case basis.

One important stumbling block, however, is the concern about enhancements to the code from an external project under a license other than CDDL. Ultimately I would prefer to see such contributions able to come back to Open Solaris under CDDL. I'm not sure how best to achieve this, though; and if the concern arises, it may be a good idea to get some legal assistance with the matter.

• On Commercial Use and Intellectual Property

I make my living writing software.

Furthermore, I've worked at Sun in the past, expect to start working their again in about two weeks, and have most recently worked at a significant Sun licensee and technical partner.

I would like to see Open Solaris continue to encourage commercial endeavors based on it. We should continue to have a friendly relationship with Sun and other companies (such as my soon-to-be-previous employer) and encourage their participation.

I also believe that there is a case for software patents, but that the patent process is badly busted and needs repair. You won't find me at a protest against patents or against companies that make money selling software, but you also won't find any content I don't have a legal right to have on any of my systems. But you will find me contributing time, energy, and actual code to Open Source projects I believe in. Right now Open Solaris is the first and foremost of those.

• The Role of OGB

Strong non-technical leadership.

I believe OGB has a limited number of tasks to accomplish. First, I think OGB can help provide non-technical leadership for the Open Solaris umbrella. By that, I mean that OGB can help direct what markets the project should be pursuing, how we image ourselves, arrangement of constitutional matters (such as the election itself), etc. Second, OGB can help act as a liason between the community (and projects and groups within it), and other 3rd parties, including Sun. Third, this first OGB has, IMO, an important goal of getting some of the infrastructure and political issues worked out (perhaps by creation of a separate group) to finish making the project independent of Sun. Finally, it can act as the final arbiter in cases where an issue cannot be resolved at a lower level (such as at a project or group level.)

I do not believe that OGB should be providing any technical direction whatsoever, although some of its members will undoubtedly do so as a result of their other roles. Indeed, apart from the fairly large set of tasks the initial OGB has to perform to address organizational and administrative shortcomings, I do not expect the OGB to be very involved in the day to day running of the project.

• The role of Sun

A contributor, perhaps the most important contributor.

(Disclaimer: I will shortly be working for Sun -- indirectly, but they will be paying my salary nonetheless. I am running as an individual however, because I'm passionate about Solaris. I just also happen to be going to work for Sun also because I'm passionate about Solaris. Well, that, and the nice salary...)

Sun is a major contributor, and the copyright holder for nearly all of the source we work with. However, by creation of the Open Solaris project, Sun has made it clear that it wants the project to stand on its own.

There are various technical, administrative, and political barriers still standing in the way. None of them are insurmountable. I think OGB should be actively trying to drive solutions to these problems (perhaps not the actual solutions themselves, but make sure that the right people are working on them, and track the progress.)

• Technical Matters

Yes, I'm an engineer.

I think I've already stated that OGB shouldn't be involved in the technical direction of Open Solaris. However, a few items that have been asked already.

Binary compatibility. Much of the appeal of Solaris has been the fact that it has been stable from release to release. I would be hard pressed to find a case for breaking this. Stability levels and standards like the DDI have been a key value proposition for Solaris in the past, and will continue to be so if we don't break it.

Quality first. The various processes that Sun has had in place for Solaris for over a decade now have done a lot to ensure that Solaris is the high quality product it is. Change for its own sake is bad. Peer review is good. Oversight review by a strong (genius!) technical leadership (ARC) is even better. I don't want to change this culture. I do want to figure out how to bring non-Sun parties to the table though, and involve the whole community. Its going to be tough for some people who are used to "commit first, fix it later, and hope that someone else will document it" policies in certain other open source projects. But we'll all be better for it.

Now, all that being said, there are probably questions I forgot to answer, or simply haven't thought of. You can ask me, or the whole group of candidates at board-candidates-2007@opensolaris.org. I'll post any of my replies here.

About Me (Bio)

I'm a 35 year old software engineer from southern California, where I currently reside. (I telecommute.) I have a BS in Computer Science from San Diego State University (class of '95). My areas of expertise are kernel, device drivers, embedded systems, networking, and security. I've hacked on Linux, NetBSD, and Solaris, as well as some proprietary kernels (including the Sun Ray firmware). Far and away I prefer Solaris -- I could go on and on about its benefits... but you already know them, right?

I've worked as a UNIX systems administrator at Qualcomm, as a software engineer (systems software and device drivers) at Sun, and as a kernel, networking, and thin-client engineer at Tadpole Computer (now General Dynamics.) I'm expecting to start a new job as a Contractor at Sun working on Solaris starting March 19, 2007. (My whole resume is online if you really want to know the details.)

I also have a loving wife, and am blessed with two girls and a boy (ages 5, 7, and 6). We also have two cats.

When I'm not in front of a computer, or spending time at home with the family, I enjoy white water kayaking (including ocean surfing in playboats), sailing, skiing, swimming and reading.

New blog

This is my new blog, replacing the short-lived Celestial Bodies, Amphibians, and Reptiles blog I set up while an employee at General Dynamics (formerly Tadpole Computer.)

There will be more here in short bit, particularly an update on my position for the OpenSolaris Governing Board (OGB).