[coreboot] write/verify bug

Paul Fox pgf at foxharp.boston.ma.us
Fri Jun 12 01:41:00 CEST 2009

ron wrote:
 > Boy, this discussion is getting really complicated. That's a sign that
 > it is time to STOP! and think it all over :-)

in my defense, i thought i had.  :-)

 > We're going to guess the order of operations based on arguments,
 > except when we ignore the argument order of operations, because we
 > think we're smarter than the user? And then we're going to reread the
 > file and ....
 > I'm an old unix buzzard. Our philosophy was, processes are cheap.
 > That's why things like ls | grep are deemed acceptable.

we also often create convenience options for ourselves, like
the 'z' option to tar.  but i take your point.

 > Tools do one thing (or one thing at a time) and do it well. flashrom
 > is getting radically over-engineered.
 > Please rethink this.
 > flashrom -e -- erase
 > flashrom -r -- read
 > flashrom -w -- write
 > flashrom -v -- verify
 > flashrom -any combination -- mistake
 > Otherwise, no matter how you set this up, you're going to violate
 > somebody's principle of least surprise. This thread already shows how.

i gues i didn't see that much controversy here.  the only real
surprise i've seen was mine, first at having my chip cleared
instead of programmed :-), and then when folks pointed out
flashrom's erase-by-default behavior (i.e., erase even without
-E).  but try as i might, i can't imagine anyone wanting to
_regularly_ perform the steps of erase, write, and verify in any
other order than that one.

if the consensus is that flashrom do only one thing at a time,
that's fine by me.  the current behavior is wrong, and needs to
be fixed, somehow.

 paul fox, pgf at foxharp.boston.ma.us (arlington, ma, where it's 55.4 degrees)

More information about the coreboot mailing list