Hi, On Mon, Jun 14, 2010 at 09:00:05AM +0200, olafbuddenha...@gmx.net wrote: > > or that modifying Mach for this purpose is misguided and should be > > avoided (for instance by embedding the initrd in a dedicated section > > of an elf module). > > Hm... I don't remember this variont actually being mentioned on IRC -- > but I think it indeed was brought up in the original discussion on the > list a couple of months ago.
It was discussed recently on IRC by Sergio and Samuel (don't remember when, exactly, perhaps it was after this thread was posted). Samuel's main objection to this idea is that it requires a new utility to pack images into ELF files. I agree, because not only does it complicate things short term with additional differences to how initrd is generated on Linux (I might be overstating this, I'm not familiar with the details), it also makes it impossible to generate images from systems which doesn't have this utility. > Also, having the initrd implementation in a userspace process will > require some changes in the bootstrap process... OTOH, these will be > necessary for userspace drivers anyways. I believe these two things are enough not to consider any such approach at this time, though we might want to reconsider once such changes are in place. (For instance, it might be possible to make Mach create a task without thread if it can't execute the image, then you could just read this task with the task store if you know its PID.) Regards, Fredrik
