> Waiting for write barriers to clear is key to implementing fsync()
> efficiently and correctly.
Well, all you want is a feature to write a set of blocks and be
acknowledged of completion of the write before you send more data, but
OTOH you would not want to serialize fsync() operations, see my "groups"
that I told previously. That would probably involve tagging data blocks
in the long run. Not sure if the current tag command API of ATA can
already provide that, if so, all is there, and the barrier can be
implemented in the driver rather than the drive firmware.
-- Matthias Andree
"They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety." Benjamin Franklin - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to email@example.com More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/