[Qemu-devel] Crashing in tcp_close

[Brian Candler]

I have some reproducible-ish segfaults in qemu 2.7.0 (built from source) 
running under ubuntu 16.04, on a quad-core i7 Mac Mini Server.

I can reproduce these problems on a different Mac Mini, and I also 
replaced the RAM on mine, so I'm sure it's not hardware related.

It's somewhat painful to reproduce (taking about 30 minutes each 
attempt, and using a lot of network bandwidth).

This is using packer (packer.io) to create a VM and then using ansible 
to do a whole load of package installation and provisioning inside that 
VM.  packer starts qemu with a user-mode network interface.

If I part-build the VM, I can continue the build by restarting it under 
gdb and qemu directly at the command line, and get a backtrace. Here's 
the first one:

Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault.
0x00007ffff6a1bb5b in _int_free (av=0x7ffff6d5fb20 <main_arena>, 
p=<optimised out>, have_lock=0) at malloc.c:4006
4006    malloc.c: No such file or directory.
(gdb) bt
#0  0x00007ffff6a1bb5b in _int_free (av=0x7ffff6d5fb20 <main_arena>, 
p=<optimised out>, have_lock=0)
    at malloc.c:4006
#1  0x00007ffff6a1fabc in __GI___libc_free (mem=<optimised out>) at 
malloc.c:2969
#2  0x00005555559a6c0f in tcp_close (address@hidden) at 
slirp/tcp_subr.c:334
#3  0x00005555559a6c8f in tcp_drop (address@hidden, 
err=<optimised out>) at slirp/tcp_subr.c:298
#4  0x00005555559a816b in tcp_timers (timer=<optimised out>, 
tp=0x555556621ed0) at slirp/tcp_timer.c:179
#5  tcp_slowtimo (address@hidden) at slirp/tcp_timer.c:89
#6  0x00005555559a0be8 in slirp_pollfds_poll (pollfds=0x555556531f20, 
address@hidden)
    at slirp/slirp.c:576
#7  0x00005555559d4b0c in main_loop_wait (nonblocking=<optimised out>) 
at main-loop.c:508
#8  0x000055555573fea1 in main_loop () at vl.c:1908
#9  main (argc=<optimised out>, argv=<optimised out>, envp=<optimised 
out>) at vl.c:4604
(gdb)

So:

* Is this of interest?

* If so, what additional gdb output would you like me to provide?

* If developers want to reproduce this, let me know and I can probably 
send the VM qcow2 file and/or packer source privately off-list [I need 
to check permission for that]

Thanks,

Brian Candler.