Long-shot Windows XP question
Randy J. Ray
rjray at blackperl.com
Tue Apr 1 21:40:38 BST 2008
> This may not help you, but I thought the standard way to do this sort of
> thing was to have three partitions: one for Linux, one for Windows, and
> a common FAT32 slice for the other two sides to share & exchange files.
That's what I have. However, if you hibernate Windows (as opposed to a complete
shutting-down) it retains a ramdisk-skewed impression of what is on the VFAT
slice. Return to Linux and add/change a file, then return to Windows, and it
has no idea the new data is there. Make a change within Windows at this point,
and it will use the ramdisk/cache impression to obliterate the Linux-source
> Alternatively (maybe), would VMWare get you anywhere?
Not in this case... the main use of the XP side is still games like Civ IV or
Total War (neither of which I have much bandwidth for these days anyway). For
that, I have to run natively.
I finally found a thread on ubuntuforums.org, and found a service that someone
had written that hooks into the hibernate sequence and forces a complete flush
of the disk caches before hibernation. Now that I have that, the core problem
is gone. In its place, I now have a new problem: the first time I delete a file
from the VFAT slice within Linux, the fs becomes read-only (the delete
succeeds, though). If I umount and re-mount it, it's writable again. For
exactly one delete operation, then read-only. If I copy data *to* it, it keeps
I so fucking hate Windows...
(Ironically, this job will be giving me a Macbook Pro on or around Thursday,
kitted with VMWare Fusion and having access to a 1Tb Time Capsule. At that
point, I won't need to use my personal laptop for this. But I might as well
solve this problem as ignore it and hope I never need to worry about it in the
Randy J. Ray rjray at blackperl.com
Sunnyvale, CA http://www.rjray.org
More information about the london.pm