Hello Michal, I completely agree. As I mentioned previously, I would like to avoid touching the VFS and this seems to be the only solution. My original question was whether there is some support in the VFS for this already. The rest of the communication here is me trying to justify the way I asked.
I started working on implementing that "hacky solution" yesterday and I'm getting close to having a layer between reads and writes and their implementation corresponding to the type of the given file descriptor. Thank you for your insights, Jan 2015-03-17 20:42 GMT+01:00 Michal Koutný <[email protected]>: > Hi Jan. > > 2015-03-17 19:41 GMT+01:00 Jan Mareš <[email protected]>: > > If those structures are not all supported by the VFS I will > > have to distinguish between the different types judging by the received > > HANDLE. The only way I can imagine implementing that is a hack similar to > > the one described in the previous message. > I think that it's not necessary to have all these objects supported by > VFS. It's a viewpoint affected by the Unix principle: "Everything is a > file (descriptor)". I don't know Windows internals but it seems to me > that there must be a dispatcher that passes handles to appropriate > subsystem (filesystem, scheduler, networking, whatever), which > provides uniform access ("everything is a handle") to various types of > resources. > > I don't see it as a hack (although I don't have precise idea of the > implementation) to implement this dispatcher in userspace in HelenOS. > For instance as a part of the POSIX compatibility layer (though there > might be better places). > > Michal > > _______________________________________________ > HelenOS-devel mailing list > [email protected] > http://lists.modry.cz/listinfo/helenos-devel >
_______________________________________________ HelenOS-devel mailing list [email protected] http://lists.modry.cz/listinfo/helenos-devel
