-----BEGIN PGP SIGNED MESSAGE-----
On 11/20/07, Andrew Greenberg <[EMAIL PROTECTED]> wrote:
> Huh. So, I want to look at that. I really thought we solved this
> problem before flight, so I think we should all sit down and make sure
> the post-processing isn't missing something. I swear the ADEOS patch
> fixed all this junk.
The ADEOS patch ensured that we didn't lose any packets while reading
them off the CAN bus from the microcontrollers. However, in that
interrupt handler, we sampled the value of 'jiffies', which was
incremented by the timer interrupt. We could have pulled the timer
interrupt into the real-time ADEOS domain, but didn't--so the timer
interrupt didn't get to execute any more than anything else when the IDE
driver had interrupts disabled for long periods.
That means that all messages received in those intervals where
interrupts were disabled(-ish, cf. ADEOS) have the same, stale,
timestamp; then the timer handler gets to run again, fast-forwards
jiffies to catch up, and the next message received has a correct
This kind of crap is why I don't want to fly a 2.4 kernel again.
> Anyway: happy processing! See you all at the next meeting.
Which I believe is a week from tomorrow. :-)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
-----END PGP SIGNATURE-----
psas-software mailing list