[L2Ork-dev] Segfault using paste from clipboard on Linux

[Ivica Ico Bukvic]

It's by far the most SNAFU'd implementation of a for loop I've ever 
seen. So much so, I am not even sure what it exactly does:

    -->     for (g = x->gl_list; g2 = g->g_next; g = g2)
                     ;

I am assuming it is looking for the last element in the gl_list. If so, 
I have no idea how this has worked so far, since I do not see any kind 
of a check whether g->g_next is not NULL... This is also the only place 
in the entire function (canvas_f) that uses t_gobj *g2.

Running uner that assumption, I refactored it as:

g = x->gl_list;
if (g) // probably unnecessary but hey let's be pedantic
     while (g->g_next)
         g = g->g_next;

Once this is implemented, the segfault now happens on line 2282 
(immediately below) on the first pd_checkobject(&g->g_pd) which also 
segfaults with an invalid read of size 8...


On 6/18/2020 4:13 PM, Jonathan Wilkes wrote:
> On Thu, Jun 18, 2020 at 3:32 PM Ivica Bukvic <ico at vt.edu> wrote:
>> Valgrind is a bit more descriptive. It does not happen every time but it does happen nonetheless. Looks like Windows is a lot less forgiving. I was pasting the content of a simple patch multiple times (close the window, new window, paste again):
>>
>> #N canvas 487 261 450 300 10;
>> #X floatatom 145 63 5 0 0 0 - - -, f 5;
>> #X obj 241 123 print;
>> #X floatatom 126 134 5 0 0 0 - - -, f 5;
>> #X connect 0 0 1 0;
>>
>> ==12982== Invalid read of size 8
>> ==12982==    at 0x436EF3: canvas_f (g_canvas.c:2273)
> Something's off for me because 2273 isn't inside canvas_f().
>
> Can you tell me what's at that line number in whatever is your current branch?
>
> -Jonathan
>
>> ==12982==    by 0x49B977: pd_typedmess (m_class.c:779)
>> ==12982==    by 0x49B7DD: pd_typedmess (m_class.c:883)
>> ==12982==    by 0x4A4D80: binbuf_eval (m_binbuf.c:937)
>> ==12982==    by 0x4AC56B: socketreceiver_read (s_inter.c:615)
>> ==12982==    by 0x4AB554: sys_domicrosleep.constprop.3 (s_inter.c:226)
>> ==12982==    by 0x4AE35A: sys_pollgui (s_inter.c:1155)
>> ==12982==    by 0x4A84B9: m_pollingscheduler (m_sched.c:542)
>> ==12982==    by 0x4A84B9: m_mainloop (m_sched.c:622)
>> ==12982==    by 0x4AAEF9: sys_main (s_main.c:440)
>> ==12982==    by 0x5ED882F: (below main) (libc-start.c:291)
>> ==12982==  Address 0x8 is not stack'd, malloc'd or (recently) free'd
>> ==12982==
>> ==12982==
>> ==12982== Process terminating with default action of signal 11 (SIGSEGV)
>> ==12982==  Access not within mapped region at address 0x8
>> ==12982==    at 0x436EF3: canvas_f (g_canvas.c:2273)
>> ==12982==    by 0x49B977: pd_typedmess (m_class.c:779)
>> ==12982==    by 0x49B7DD: pd_typedmess (m_class.c:883)
>> ==12982==    by 0x4A4D80: binbuf_eval (m_binbuf.c:937)
>> ==12982==    by 0x4AC56B: socketreceiver_read (s_inter.c:615)
>> ==12982==    by 0x4AB554: sys_domicrosleep.constprop.3 (s_inter.c:226)
>> ==12982==    by 0x4AE35A: sys_pollgui (s_inter.c:1155)
>> ==12982==    by 0x4A84B9: m_pollingscheduler (m_sched.c:542)
>> ==12982==    by 0x4A84B9: m_mainloop (m_sched.c:622)
>> ==12982==    by 0x4AAEF9: sys_main (s_main.c:440)
>> ==12982==    by 0x5ED882F: (below main) (libc-start.c:291)
>> ==12982==  If you believe this happened as a result of a stack
>> ==12982==  overflow in your program's main thread (unlikely but
>> ==12982==  possible), you can try to increase the size of the
>> ==12982==  main thread stack using the --main-stacksize= flag.
>> ==12982==  The main thread stack size used in this run was 8388608.
>> ==12982==
>> ==12982== HEAP SUMMARY:
>> ==12982==     in use at exit: 355,586 bytes in 3,885 blocks
>> ==12982==   total heap usage: 7,748 allocs, 3,863 frees, 2,870,428 bytes allocated
>> ==12982==
>> ==12982== LEAK SUMMARY:
>> ==12982==    definitely lost: 416 bytes in 9 blocks
>> ==12982==    indirectly lost: 85 bytes in 10 blocks
>> ==12982==      possibly lost: 43,438 bytes in 1,323 blocks
>> ==12982==    still reachable: 311,647 bytes in 2,543 blocks
>> ==12982==         suppressed: 0 bytes in 0 blocks
>> ==12982== Rerun with --leak-check=full to see details of leaked memory
>> ==12982==
>> ==12982== For counts of detected and suppressed errors, rerun with: -v
>> ==12982== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
>>
>> --
>> Ivica Ico Bukvic, D.M.A.
>> Director, Creativity + Innovation
>> Institute for Creativity, Arts, and Technology
>>
>> Virginia Tech
>> Creative Technologies in Music
>> School of Performing Arts – 0141
>> Blacksburg, VA 24061
>> (540) 231-6139
>> ico at vt.edu
>>
>> ci.icat.vt.edu
>> www.icat.vt.edu
>> www.performingarts.vt.edu
>> l2ork.icat.vt.edu
>> ico.bukvic.net

-- 
Ivica Ico Bukvic, D.M.A.
Director, Creativity + Innovation
Institute for Creativity, Arts, and Technology

Virginia Tech
Creative Technologies in Music
School of Performing Arts – 0141
Blacksburg, VA 24061
(540) 231-6139
ico at vt.edu

www.icat.vt.edu
www.performingarts.vt.edu
l2ork.icat.vt.edu
ico.bukvic.net