Hello Guido,
* Guido Witmond <guido@...231...> [2015-08-03 21:57:20 +0200]:
Last weekend I compiled Foc_x86_64 and tested it with the onboard ethernet cards of my server box.
It worked! It worked just as good as the Intel e1000 driver. (Meaning, after a short while, all throughput ceased. :-( But that is eiter Foc or Lightttpd)
That is great news, indeed. Regarding lighttpd there is at least one problem we know of (see issue #987 [1]). In general it seems to work fine, however. We do not stress test lighty regularly though but do use tools like ``ab'' [2] occasionally (I am afraid a cannot remember the parameters we used). So, as a starting point, it might be a good idea to narrow down the workload which triggers the faulting behaviour.
In reply to your former E-Mail asking about debugging lighty, the messages should already be printed to the serial console or rather to the configured LOG service. You might need to configure it first to be more verbose by editing its config file which is certainly described in its documentation. There are also the usual DEBUG defines that can be enabled at compile time (you would have to add them to ``repos/ports/src/app/lighttpd/target.inc''). IIRC lightly uses one event loop to handle all events (bascially it will use select()). Instrumenting this loop is probably the fastest way to determine if lighty still “lives” even when it appears to be not reacting.
Regards Josef
[1] https://github.com/genodelabs/genode/issues/987 [2] http://httpd.apache.org/docs/2.4/programs/ab.html