Sed,
thanks for your report! I'll to reproduce the problem when I return from my vacation in ten days.
roux@ecoledoc.lip6.fr (Cedric ROUX) writes:
[...]
When i run it under fiasco, it runs a little, then it blocks, so i press ctrl+c, that nornmally should end up the program and give me back the shell, but not, it just continues a little more, and stops again, and so on until the end of the program. Then i see a lot of "Aiee: scheduling in interrupt 00422ba (activ 10000)" with sometimes a 0 in place of the 10000.
This looks like a scheduling problem in Fiasco: An interrupt thread is still active while the Linux scheduling code is run. This should not happen because interrupt threads have a higher priority and never block. Interesting.
[...]
Third, but it was before patch3, i ran through my hd, making ls, cat, rm and so on, and at one time, after a ls, and before the result being displayed, i receive this message "page fault in signal handler". Maybe patch3 fixes that, huh ?
Maybe yes, but I'm not sure.
And last, i normally have shift+page-up to see last lines, but it does not works under fiasco (maybe kernel 2.0.21 doesn't handle it anyway). Little problem, but problem.
This doesn't work because hardware scrolling is disabled in L4Linux. We did this because L4Linux shares the console's video memory with Fiasco/L4 and potentially other L4 tasks.
I soon gonna be able to debug with a second machine, so my next reports will be better.
That's cool!
Michael