I made the code pretty much 64-bit safe ... if the comment is there,
it's probably because I haven't tested it. :)
--Matt
P.S. I'll be back in the office on Monday, and I
can start integrating your code into the patch ASAP.
On Thu, 10 Feb 2000, Brian Hall wrote:
|>OK, I can now dump header and pages to the swapfile. DUMP_LEVEL = 0,1,2,3,4
|>and compression are tested and "working" (defined as not crashing or causing
an
|>oops; at least the header is verified correct; and something is being written
|>out to the swap device besides just the header).
|>
|>I see a note in __dump_compress that it needs fixing for >32bit systems, what
|>needs to be done there ? I don't see an obvious need to change the return
value
|>type, it'll be much less than needs a 64 bit field, and the rest of the fxn is
|>only relative to the current page being compressed. There is a similar note
|>about fixing for >32bit systems in __dump_execute, but I think I covered that
by
|>switching to 64 bit variables.
|>
|>
|>P.S. My testing has gone a lot faster since I started using the SysRq keys to
|>sync and remount the disk R/O to avoid the fsck on reboot!
|>
|>--
|>Brian Hall <brianw.hall@xxxxxxxxxx>
|>http://www.bigfoot.com/~brihall
|>Linux Consultant
|>
|