Hi all,
After merging the cifs tree, today's linux-next build (powerpc
ppc64_defconfig) produced these warnings:
fs/cifs/sess.c: In function 'cifs_select_sectype':
fs/cifs/sess.c:527:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
switch (requested) {
^~~~~~
fs/cifs/sess.c:540:2: note: here
case CIFS_NEGFLAVOR_LANMAN:
^~~~
fs/cifs/connect.c: In function 'cifs_parse_mount_options':
fs/cifs/connect.c:1825:8: warning: this statement may fall through [-Wimplicit-fallthrough=]
if (!(tmp_end < end && tmp_end[1] == delim)) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
fs/cifs/connect.c:1832:3: note: here
case Opt_pass:
^~~~
Exposed by my use of -Wimplicit-fallthrough
I am not sure why these only turned up now (as opposed to earlier today).
--
Cheers,
Stephen Rothwell
Both of those cases are intentional fallthroughs and there are
existing comments in the code noting the reasons for them to
fallthrough
(also can see the reasoning for these in the commits which introduced
them from Sachin c369c9a4a7c82) and dde2356c84662)
On Sun, Nov 25, 2018 at 6:14 PM Stephen Rothwell <[email protected]> wrote:
>
> Hi all,
>
> After merging the cifs tree, today's linux-next build (powerpc
> ppc64_defconfig) produced these warnings:
>
> fs/cifs/sess.c: In function 'cifs_select_sectype':
> fs/cifs/sess.c:527:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
> switch (requested) {
> ^~~~~~
> fs/cifs/sess.c:540:2: note: here
> case CIFS_NEGFLAVOR_LANMAN:
> ^~~~
> fs/cifs/connect.c: In function 'cifs_parse_mount_options':
> fs/cifs/connect.c:1825:8: warning: this statement may fall through [-Wimplicit-fallthrough=]
> if (!(tmp_end < end && tmp_end[1] == delim)) {
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> fs/cifs/connect.c:1832:3: note: here
> case Opt_pass:
> ^~~~
>
> Exposed by my use of -Wimplicit-fallthrough
>
> I am not sure why these only turned up now (as opposed to earlier today).
>
> --
> Cheers,
> Stephen Rothwell
--
Thanks,
Steve
Hi Steve,
On Sun, 25 Nov 2018 18:31:40 -0600 Steve French <[email protected]> wrote:
>
> Both of those cases are intentional fallthroughs and there are
> existing comments in the code noting the reasons for them to
> fallthrough
>
> (also can see the reasoning for these in the commits which introduced
> them from Sachin c369c9a4a7c82) and dde2356c84662)
I am not questioning that :-)
The gcc warning can be turned off by adding a /* fall through */
comment at the point the fall through happens. Kees and others are
working on the several hundred other places that need annotating.
This one just popped up.
--
Cheers,
Stephen Rothwell
On Sun, Nov 25, 2018 at 4:52 PM Stephen Rothwell <[email protected]> wrote:
>
> Hi Steve,
>
> On Sun, 25 Nov 2018 18:31:40 -0600 Steve French <[email protected]> wrote:
> >
> > Both of those cases are intentional fallthroughs and there are
> > existing comments in the code noting the reasons for them to
> > fallthrough
> >
> > (also can see the reasoning for these in the commits which introduced
> > them from Sachin c369c9a4a7c82) and dde2356c84662)
>
> I am not questioning that :-)
>
> The gcc warning can be turned off by adding a /* fall through */
> comment at the point the fall through happens. Kees and others are
> working on the several hundred other places that need annotating.
Right. The goal is to avoid adding any _new_ cases of this. :)
> This one just popped up.
It's already working! :) Thanks Stephen!
-Kees
--
Kees Cook
Hi Kees,
On Sun, 25 Nov 2018 21:48:10 -0800 Kees Cook <[email protected]> wrote:
>
> On Sun, Nov 25, 2018 at 4:52 PM Stephen Rothwell <[email protected]> wrote:
> >
> > On Sun, 25 Nov 2018 18:31:40 -0600 Steve French <[email protected]> wrote:
> > >
> > > Both of those cases are intentional fallthroughs and there are
> > > existing comments in the code noting the reasons for them to
> > > fallthrough
> > >
> > > (also can see the reasoning for these in the commits which introduced
> > > them from Sachin c369c9a4a7c82) and dde2356c84662)
> >
> > I am not questioning that :-)
> >
> > The gcc warning can be turned off by adding a /* fall through */
> > comment at the point the fall through happens. Kees and others are
> > working on the several hundred other places that need annotating.
>
> Right. The goal is to avoid adding any _new_ cases of this. :)
>
> > This one just popped up.
>
> It's already working! :) Thanks Stephen!
Actually this one was the product of how I filter out warnings ... it
was an old one, but the line numbers changed. I have now started
ignoring these if just the line numbers change.
--
Cheers,
Stephen Rothwell