openssh bugs - Dec 2019 - [Bug 3102] New: Compilation error in sshbuf-misc.c on IRIX 6.5 with MIPSPro 7.4

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

            Bug ID: 3102
           Summary: Compilation error in sshbuf-misc.c on IRIX 6.5 with
                    MIPSPro 7.4
           Product: Portable OpenSSH
           Version: 8.1p1
          Hardware: MIPS64
                OS: IRIX
            Status: NEW
          Severity: minor
          Priority: P5
         Component: Miscellaneous
          Assignee: unassigned-bugs at mindrot.org
          Reporter: ole.weidner at protonmail.ch

cc-1515 c99: ERROR File = sshbuf-misc.c, Line = 225
  A value of type "int" cannot be assigned to an entity of type
"void
*".

        if ((p = memmem(sshbuf_ptr(b) + start_offset,


This can be easily fixed by changing line 215 from

        void *p;

to 

        size_t p;

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

Darren Tucker <dtucker at dtucker.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dtucker at dtucker.net

--- Comment #1 from Darren Tucker <dtucker at dtucker.net> ---
(In reply to Ole Weidner from comment #0)
> cc-1515 c99: ERROR File = sshbuf-misc.c, Line = 225
>   A value of type "int" cannot be assigned to an entity of type
> "void *".
> 
>         if ((p = memmem(sshbuf_ptr(b) + start_offset,
> 
> This can be easily fixed by changing line 215 from void *p to 
> to size_t p
That's not correct, though.  Although it's not standardized, the
consensus seems to be memmem returns a void *
(http://man7.org/linux/man-pages/man3/memmem.3.html) and that's
certainly what OpenSSH expects (https://man.openbsd.org/memmem).

Does IRIX define its own memmem to return an integer type (ie is
HAVE_MEMMEM in config.h) or is this a result of a missing prototype? 
(It's defined in openbsd-compat.h which is pulled in via includes.h,
but it's possible there might be some reason it doesn't work).

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

--- Comment #2 from Ole Weidner <ole.weidner at protonmail.ch> ---
HAVE_MEMMEM is not defined in config.h and I couldn't find any traces
of it in the IRIX system includes.

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

--- Comment #3 from Darren Tucker <dtucker at dtucker.net> ---
Created attachment 3342
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3342&action=edit
check if memmem is declared and use for compat prototype

(In reply to Ole Weidner from comment #2)
> HAVE_MEMMEM is not defined in config.h and I couldn't find any
> traces of it in the IRIX system includes.
If you don't have HAVE_MEMMEM in config.h then I'm not sure what's
going on.  You could check config.log and look for the memmem test
output.

I did find one way to reproduce a similar error: where configure is
picking up a memmem from somewhere (maybe libc, or maybe one of the
other dependencies, I reproduced with gcc and a -lmemmem in LIBS) but
that will result in HAVE_MEMMEM in config.h.

The compat code only defines the prototype #ifndef HAVE_MEMMEM, thus in
this case nothing defines it so it defaults to int and you get the
mismatch.

The attached patch fixes what I found  (note that you will need to run
autoreconfig from autoconf 2.69 to rebuild configure before running
it).  It's worth trying for your case too.

The other potential gotcha: if you have a different config.h elsewhere
in your search path you might not be using the one you think you are.

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

Darren Tucker <dtucker at dtucker.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|                            |3079


Referenced Bugs:

https://bugzilla.mindrot.org/show_bug.cgi?id=3079
[Bug 3079] Tracking bug for 8.2 release
-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

--- Comment #4 from Darren Tucker <dtucker at dtucker.net> ---
I've committed that diff, I would be interested to know whether or not
it makes a difference in this case.  You can try as snapshot
(https://www.mindrot.org/openssh_snap/) or current git tree.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

--- Comment #5 from Ole Weidner <ole.weidner at protonmail.ch> ---
The diff indeed makes a difference and openssh compiles without any
issue now The following warning pops up during linkage:

ld32: WARNING 84 : openbsd-compat/libopenbsd-compat.a is not used for
resolving any symbol.

The executable works fine.

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org
         Resolution|---                         |FIXED
             Status|NEW                         |RESOLVED

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

https://bugzilla.mindrot.org/show_bug.cgi?id=3102

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #6 from Damien Miller <djm at mindrot.org> ---
closing resolved bugs as of 8.6p1 release

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.