Hello!
the syscalls operate in a process context and only modify the process context. So if you open a file in the context of process A you can only access the file from process A. Process B don't know anything about these open file.Damn, the kernel can not open any file for itself ?
Sure it can, with kernel_open :-)
Even putting the loop device in the core kernel (like /kern, /shm, etc...) would not work without this block device ? So, I'm stuck with my idea, no loopback device for MiNT.
Hehe, don't give up so fast. There exist more or less a block device abstraction. This is the buffer cache as used by the xfs. It's just missing an interface to block device drivers so you can add other things (at the moment only XHDI and BIOS is supported). So we need such an interface and we need to migrate the XHDI and BIOS driver to this interface (we should handle then the partition table self so we need routines to read partition data too). Then a loopback device driver that route access to a file can be easily added.
Regards, Frank -- ATARI FALCON 060 // MILAN 060 ----------------------------------------- http://www.cs.uni-magdeburg.de/~fnaumann/ e-Mail: fnaumann@freemint.de