[External] Re: [PATCH 1/5] maple_tree: Move the check forward to avoid static check warning.

[Liam R. Howlett]

* Peng Zhang <zhangpeng.00 at bytedance.com> [231116 04:18]:
> Hello Liam,
> 
> 在 2023/11/13 23:09, Dan Carpenter 写道:
> > On Mon, Nov 13, 2023 at 10:44:02AM +0800, Peng Zhang wrote:
> > > 
> > > 
> > > 在 2023/11/10 22:49, Liam R. Howlett 写道:
> > > > * Peng Zhang <zhangpeng.00 at bytedance.com> [231109 07:43]:
> > > > > Put the check for gap before its reference to avoid Smatch static check
> > > > > warnings. This is not a bug, it's just a validation program.
> > > > > 
> > > > > Reported-by: Dan Carpenter <dan.carpenter at linaro.org>
> > > > > Closes: http://lists.infradead.org/pipermail/maple-tree/2023-November/003046.html
> > > > > Signed-off-by: Peng Zhang <zhangpeng.00 at bytedance.com>
> > > > > ---
> > > > >    lib/maple_tree.c | 6 +++++-
> > > > >    1 file changed, 5 insertions(+), 1 deletion(-)
> > > > > 
> > > > > diff --git a/lib/maple_tree.c b/lib/maple_tree.c
> > > > > index aaba453b0d30..1dfbcb74787a 100644
> > > > > --- a/lib/maple_tree.c
> > > > > +++ b/lib/maple_tree.c
> > > > > @@ -7249,6 +7249,11 @@ static void mas_validate_gaps(struct ma_state *mas)
> > > > >    counted:
> > > > >    	if (mt == maple_arange_64) {
> > > > > +		if (!gaps) {
> > > > > +			MT_BUG_ON(mas->tree, 1);
> > > > > +			return;
> > > > 
> > > > Since MT_BUG_ON() will BUG_ON(), is there a reason to return?
> > > > 
> > > > The only reason I've used braces to BUG_ON() is to include extra error
> > > > text, but you haven't provided any here.  I think you could do this (or
> > > > MT_BUG_ON() variant, if you prefer):
> > > > 
> > > > MAS_BUG_ON(mas, gaps == NULL);
> > > Because MT_BUG_ON() in the maple tree may not invoke BUG_ON() and continue
> > > execution, this code is written to prevent the static checker from issuing
> > > further warnings. (Dan Carpenter informed me of this a long time ago.)
> > 
> > Yeah.  Smatch will see this as a test for NULL followed by an untest
> > dereference.  We could do the return as you suggest, or we could
> > 
> > 1) ignore the smatch warning
> > 
> > or
> > 
> > 2) make smatch ignore MT_BUG_ON() like it does with other ASSERT()
> >     statements.  (This is fine because we're hopefully not putting
> >     side effects into the MAS_BUG_ON() call like:
> > 
> > 	MAS_BUG_ON(mas, p = kmalloc());
> Liam, What do you think is the best way to handle this?
> I'm planning to submit the next version to LKML.

Sorry, I was away for the majority of the week.  I think we should make
smatch ignore MT_BUG_ON() and MAS_BUG_ON() so it is fixed for all future
patches.

Thanks,
Liam