ocfs2: update comments in masklog.h

 * only emit the appropriage printk() when the caller passes in a constant

 * mask, as is almost always the case.

 *

 * All this bitmask nonsense is hidden from the /proc interface so that Joel

 * doesn't have an aneurism.  Reading the file gives a straight forward

 * indication of which bits are on or off:

 * 	ENTRY off

 * 	EXIT off

 * All this bitmask nonsense is managed from the files under

 * /sys/fs/o2cb/logmask/.  Reading the files gives a straightforward

 * indication of which bits are allowed (allow) or denied (off/deny).

 * 	ENTRY deny

 * 	EXIT deny

 * 	TCP off

 * 	MSG off

 * 	SOCKET off

 * 	ERROR off

 * 	NOTICE on

 * 	ERROR allow

 * 	NOTICE allow

 *

 * Writing changes the state of a given bit and requires a strictly formatted

 * single write() call:

 *

 * 	write(fd, "ENTRY on", 8);

 * 	write(fd, "allow", 5);

 *

 * would turn the entry bit on.  "1" is also accepted in the place of "on", and

 * "off" and "0" behave as expected.

 * Echoing allow/deny/off string into the logmask files can flip the bits

 * on or off as expected; here is the bash script for example:

 *

 * Some trivial shell can flip all the bits on or off:

 *

 * log_mask="/proc/fs/ocfs2_nodemanager/log_mask"

 * cat $log_mask | (

 * 	while read bit status; do

 * 		# $1 is "on" or "off", say

 * 		echo "$bit $1" > $log_mask

 * log_mask="/sys/fs/o2cb/log_mask"

 * for node in ENTRY EXIT TCP MSG SOCKET ERROR NOTICE; do

 *	echo allow >"$log_mask"/"$node"

 * done

 * )

 *

 * The debugfs.ocfs2 tool can also flip the bits with the -l option:

 *

 * debugfs.ocfs2 -l TCP allow

 */

/* for task_struct */