Dear Tech-kernel people,
   I am working on a device driver for an ATM-25Mz Efficent card.  This
card requires a buffer where data can be read or writen to the card.  It
also requires that the area be contigous and be alligned to 256 byte
boundry.
   Is there any concerns or optimization issues I should be concerned
about relative to calling the kernel memory allocator?
   There does not seem to be a way to request the memory be allocated
on a 256 byte boundry but it may be allocated on a page boundry from
the manual page.  Is the size of the block tucked away just before the
block of memory like it is on some memory allocators?

   Suggestions, pointers welcome.
johnh...