31-May-2016 14:06 GMT.
[News] OpenBSD end pegasos support, (Dale's Story)ANN.lu
Posted on 26-Mar-2004 06:10 GMT by Dale Rahn125 comments
View flat
View list
OpenBSD/pegasos gone. From what originally sounded like promising Open Source support, the relationship between Genesi and Dale Rahn has turned quite sour. OpenBSD/pegasos gone.

From what originally sounded like promising Open Source support, the relationship between Genesi and Dale Rahn has turned quite sour.

I was initially contacted in Oct 2002 regarding porting OpenBSD to the Pegasos I system. After some months, they contracted me to port and support OpenBSD on the Pegasos I for a small amount of money and 5 boards for the use of myself and other OpenBSD developers. Due to the production/hardware problems in the Pegasos I boards, I ended up receiving one board. After the port was up and running with snapshots and source changes available, it became clear that those systems were not to ship in quantity, and the effort was stopped.

I was then contacted early in October 2003 asking me to update the port for the 3.4 release. Because the funding for the DARPA Grant which had been paying me was running out, I updated the code, made several improvements and had it ready for a simultaneous release with the official OpenBSD release.

I was hired on Oct 27 2003 as a non-benefits employee. Genesi wanted me to port and support OpenBSD on the Pegasos II system. When I was hired, I tried to impress on Bill Buck (who hired me over the phone) that documentation was required to write the software and support a commercial offering based on the Pegasos boards. I was reassured that documentation would not be a problem.

Genesi was in talks with ShopIP regarding shipping Pegasos based firewall boxes running ShopIP's crunchbox software. They were chasing the high revenue opportunity. Because of my status in OpenBSD and the fact that Genesi was (to be) paying me a salary I worked on finishing the Pegasos port and getting it into the OpenBSD tree.

Things appeared to be going well, however a minor delay was announced with our Dec 1 paychecks. They were to be delayed until about Dec 10 due to 'stock market issues'. I was flown out to New York City to help with the presentation of the 'Guardian', the Pegasos Crunchbox at the InfoSec conference. When at the conference some unusual negotiation was occurring between ShopIP and Genesi, there were some questions raised as to who the 'IP' of ShopIP belonged as apparently the developer had not been paid for nearly 2 years. This apparently came out about the time that Genesi and ShopIP were negotiating how the guardian was to be configured and the proceeds distributed.

As the conference ended, the delayed payday arrived. Several of the other 'employees' of Genesi were quite anxious to receive their checks and a plan was made to FedEx the checks from the conference. Since I was at the conference with Paul, I was written a check (for Oct and Nov) and handed it.

Later I find out that the other checks which were to be FedEx'ed were never sent. The check which I was handed (for $10,000) turned out to be dated 12/11/01 (two years previous), the bank refused to honor the check and did not even attempt to cash it. (Much later I find out that the account did not have the money present, even if the check had gone thru the bank).

After living thru a very uncomfortable Christmas, my smallest and most miserable in my life because I had not seen any money from Genesi, I finally blew up shortly after the new year and told them I was ready to walk. Part of me was still hoping to get the 15,000 they owed me at that point, or I would have walked then. This resulted in them paying me for one month of work ('we have no money left') to keep me around.

The Pegasos II which had just shipped and was still sitting unopened on the floor might have also influenced that money. So, I was mollified and the OpenBSD port was started. Unfortunately, because of non-standard PCI probing methods in the new northbridge, porting the software took longer than expected.

After some workarounds from hints from thrice forwarded emails, I was able to configure the system and get it running. It would have been quite straightforward to have written this properly if they had provided the documentation as they had planned. I pointed this out and stated that to write the gigabit ethernet which was to be used in the Guardian product, it would need that documentation otherwise a performant correct driver could not be written. bPlan and the Linux developer Sven (who by that time was unemployed due to Thendic-France closing) had the documentation. I, however, as a Genesi employee was not allowed to obtain the documentation.

Finally, (to shut me up I think), I was sent the linux driver for the onboard gigabit. I had previously told them that having a linux driver as reference would be nice but could not replace documentation, as most linux drivers do not lend themselves to being comprehended in such a way that the hardware is understandable and that a driver for OpenBSD could be written. That and the fact I found they were not configuring the gig-E interrupt correctly, but were tying it to the generic IRQ9 so that it gets serviced occasionally, was quite amusing.

I continued attempting to improve the port for some time, meanwhile working on OpenBSD/cats, however little progress was made on Pegasos.

It was announced that a new Pegasos II board run would occur and hints appeared on the developer chat areas that new firmware enhancements were to appear. Wanting OpenBSD to be able to support the new features (reset!) of this firmware, I attempted to obtain a copy. After about a week of begging on the chat sessions, I was able to locate someone who had a copy of the early firmware and obtain a copy. The normal source, the developer and the other bPlan representatives ignored all requests.

When I tested this new firmware on my machine, I panicked. OpenBSD would not boot. It would load the kernel and hang. I IMMEDIATELY contacted the Genesi and bPlan people claiming that this firmware had problems and that these problems needed to be resolved before the hardware was released. I offered my time to help locate the change which caused the problem, whether it was an error in the new firmware or modified assumptions that both the firmware and the OpenBSD kernel make. I received exactly one response stating some things had changed, however the reply was not specific enough to use to start debugging the issue. All other attempts at contact were ignored. About 1.5 weeks later the new boards were released and according to information by one of the users, released with firmware made the day I contacted them (which was newer that what I tested). No further communication was received after _many_ attempts were made.

After over a week of this, the announcement that (again?) there would be no payments made (was the end of the month again) and that the health benefits which they had been promising for the last 3 months would be delayed at least another month, I resigned (the date, March 1st). It was quite clear that Bill Buck who was supposed to be CEO of Genesi and 'in charge' had no power over bPlan. bPlan didn't care at all about OpenBSD or the overall quality of the board in the embedded market. I received a simple reply that was an acceptance of my resignation, and a statement that "We will make every effort to pay you for the services you have provided over the past several months".

At the point I left Genesi, I did not threaten or put any conditions on OpenBSD/pegasos. I continued asking Genesi/bPlan for fixed firmware or comments on what changed. No response ever came.

Finally I sent them a note stating that it was too late for any changes to go into the OpenBSD release (which is at the end of the 3.5 release cycle) and that we would not be able to release OpenBSD/pegasos based on the support provided and that it would be removed from the tree after 3.5. This finally provoked a response: Since "you were not able to work on Genesi related projects after February 4th" any "consulting fees" would not paid past that date. In addition, I was to be billed for the two boards which were sent to Theo de Raadt.

This is how I was treated in the employ of Genesi, I hope that no others will be hurt by this company's empty promises or as a result of using their shoddy products.

So here I am after having worked for Genesi for 4.5 months, paid for almost _5_weeks_, and searching for a job. Sigh.

Back to Top