On 14.3.2012 16:27, Maurizio Lombardi wrote: > On Tue, Mar 13, 2012 at 10:57 AM, Jiri Svoboda <[email protected]> > wrote: >> Let's first look at the overall architecture, instead. In the end you >> probably want to have some >> 'time server' which provides current time to applications upon request. >> The time server probably combines multiple sources of information / time >> sources, >> depending on what's available (e.g. CMOS, timer interrupt, tick register). > > This is probably the best solution, a completely generic time server and a > bunch of separate hw specific drivers. > > When an application executes the time() function a request will be sent via > IPC > to the time server and the latter will reply with the current time, if > I got it right.
On the other hand, introducing several layers of time services and drivers may seem unnecessarily complicated. In my view, the driver would be device specific, but it would talk a generic time protocol, pretty much in the same way as block device drivers are device specific, but they speak the unifying block device protocol. >>> <jermar> not so sure rtc needs to be a char device >> That would probably make no sense. > > :) Can't help but notice that everytime Jakub says "I'm not so sure about it" > what he really means is "I'm *damn* sure about it". I am not so sure about that either :-) Jakub _______________________________________________ HelenOS-devel mailing list [email protected] http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
