Xref: utzoo comp.binaries.ibm.pc.d:293 comp.sys.ibm.pc:15792
Path: utzoo!attcan!uunet!husc6!psuvax1!psuhcx!matt
From: matt@psuhcx (Matt Cohen)
Newsgroups: comp.binaries.ibm.pc.d,comp.sys.ibm.pc
Subject: Flushot PLUS TRASHES CMOS RAM
Message-ID: <82@psuhcx.psu.edu>
Date: 22 May 88 06:10:57 GMT
Sender: matt@psuhcx.psu.edu
Reply-To: matt@psuhcx (Matt Cohen)
Organization: Penn State University Engineering Computer Lab
Lines: 97


	
I was testing FLUSHOT Plus to see if it was worth the
$10.00 fee the Author, Ross Greenberg is charging.
I knew that a large number of hours had been put into the program.
The documentation made it lk prettgood, even though it was a bit
rambly. 
I created a FLUSHOT.DAT file, and put in 37 lines
	of the form
	C=filename

	When I loaded Flushot, I got a message saying that
	there was no room for a table, and the machine hung.

	I had not read the documentation closely enough. It turns
	out that you have to put a 'dummy' checksum in each line
	like this:
		C=filename[12345]
	Where 12345 is the dummy number. When flushot is started,
	it checksums the file, and reports the new number, which
	you have to write down and type in FOR EACH FILE!
	I then rewrote the FLUSHOT.DAT file with only two programs,
	command.com and a.bat checksummed. Flushot checked them on
	startup, but did not perform as advertised when I ran A.BAT,
	changed it, and ran it again.
	Flushot claims that it checksums files  whenever they are
	loaded by MSDOS. I guess this does not apply to BATCH files.
 	I was going to test checksumming of .EXE files, but FLUSHOT
	trashed my CMOS ram.	
	
FLUSHOT PROTECTS CMOS RAM ?

	Finally, I added two more lines to FLUSHOT.DAT with dummy
	checksums. I restarted FLUSHOT and got the following
	message:
	
		CMOS RAM HAS BEEN CHANGED. Y TO CONTINUE, ANY OTHER
		KEY TO PROCEED
	
	Followed by a long garbled bunch of characters!.
        Naturally, when I rebooted I could not boot from the 
	Hard Disk, until I restored the setup information.

	My CMOS ram was trashed by FLUSHOT! I hoped that no damage
	had been performed to my FAT!.
	I then restored my ram with a CMOS-SAV progam which I wrote
	for such a purpose, and reloaded flushot.

	I then ran a program which zeroed out my CMOS ram using
	MS C outp() function, without a whimper from FLUSHOT.
	
	Note that I had no TSR'S present when this happened. I 
	have a Leading Edge AT clone (Made by Mitsubishi, same
	as SPERRY IT). I am running DOS 3.1.


	I considered the possibility of Ross Greenberg enforcing
	his $10.00 fee by putting counters into flushot (since I
	had to restart it each time I changed anything in the
	FLUSHOT.DAT file and did this a number of times)
	and put the idea aside. (That was a pretty virulent dissertation 
	in the manual about *worms*, maybe he thinks that people
	who don't buy his software are *worms*?!? :-)
 	What I think Ross will accomplish by these threats, rewards,
	challenges is ENCOURAGE scores of copycats to write viruses
	to beat flushot (which is buggy).


	My conclusion is that FLUSHOT Plus does not perform as
	advertised (in my case, anyway) and I would not use it
	or even trust it with my data. 
	The checksum protection is quite limited in number of
	files, and the
	method of entering the checksum is quite painful.

	The bugs in the program might be excusable if  
	the program was public domain or shareware in the
	sense that you pay for it only if you think it is 
	valuable (not if you use it, since technically, I
	owe Ross Greenberg $10 since I used it) .
	I think that it tries to do too much, and ends up doing 
	too little, even the wrong thing altogether. This shows
	poor design and testing practice.

	When I support a shareware program, I am not paying the
	author for his time, I am paying for a finished product.
	And a finished product, FLUSHOT PLUS is not !

  
	The above is my opinion, and no-one is liable for it but
	myself. I reserve the right to deny everything.

			Matt Cohen (matt@psuhcx)