Hi kernel geeks,
Please forgive me if this questions was asked already in the past !!!
I have a question about shared memory mappings between kernel and
user process...(.like accessing device memory directly from user space
through device mmap interface)...I presume this mechanism use
remap_pfn_range() or similar....when such mappings are in place, (
whats the "U/S" flag in the underlying physical page indicating CPL of
the page) for such pages ? such page descriptors are stored in
normal GDT or considered as global pages ? Any special policies of
TLB flush of translations to such pages ?
With a shared memory between user process and kernel using such
mechanism, is it equivalent to zero-copy transfer between user space
and kernel space? ( I mean any extra overhead in memory access to
such pages)
I would appreciate your thoughts on this.
--
Best regards,
Peter Chacko
NetDiox computing systems,
Network storage & OS training and research.
Bangalore, India.
http://www.netdiox.com
080 2664 0708