On 26/06/14 16:35, Ian Jackson wrote:
> Antti Kantee writes ("Re: pthreads and rump kernel application stack"):
>> Minor changes to MiniOS were necessary. The stack allocation is a bit
>> of a hack currently; MiniOS requires a constant-size and appropriately
>> aligned stack. Before merge to master, Ian, any comments on minor
>> further MiniOS divergence?
>
> AFAICT there are three things here:
> * Ability to specify the stack
> * Scheduler hook
> * Something to do with timeouts
> I don't quite understand the timeout changes or why minios's existing
> timeout machinery wasn't sufficient.
Yea, pretty much. The timeout changes are there mostly because a) we
need to know if the sleep timed out or if there was a wakeup b) pthread
condvars use absolute timestamps instead of relative ones.
> In more general terms it would be better to have a formal set of
> patches to minios to be sent to the minios maintainers and xen-devel.
> But most of what you have done doesn't seem controversial in spirit,
> so it probably won't make sorting out the minios fork much harder.
Since sorting out is required in any case, it's probably easier to do
that once we have everything working. I'm not sure the current changes
would apply against upstream.
> And getting threads going is great and I really don't want to
> discourage you from pressing on...
Pressed onward (i.e. merged the branch to master).
thanks,
antti
------------------------------------------------------------------------------
Open source business process management suite built on Java and Eclipse
Turn processes into business applications with Bonita BPM Community Edition
Quickly connect people, data, and systems into organized workflows
Winner of BOSSIE, CODIE, OW2 and Gartner awards
http://p.sf.net/sfu/Bonitasoft
_______________________________________________
rumpkernel-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/rumpkernel-users
