[DRE-maint] Bug#918383: ruby-ferret 0.11.8.6-2: FTBFS, alignment problem

[Steve McIntyre]

Source: ruby-ferret
Version: 0.11.8.6-2
Severity: important
User: debian-arm at lists.debian.org
Usertags: alignment

Hi!

I've been doing a full rebuild of the Debian archive, building all
source packages targeting armel and armhf using arm64 hardware. We are
planning in future to move all of our 32-bit armel/armhf builds to
using arm64 machines, so this rebuild is to identify packages that
might have problems with this configuration.

A feature of the arm64 kernel is that it does *not* support fixing up
code with broken alignment, so code that might have built and run OK
on our older armel/armhf build machines due to kernel fixups will now
fail.

When building your package, I've found a bus error (aka alignment
fault) on both armel and armhf. The full logs are online at

  https://www.einval.com/debian/arm/rebuild-logs/armel/FAIL/ruby-ferret_0.11.8.6-2_armel.log
  https://www.einval.com/debian/arm/rebuild-logs/armhf/FAIL/ruby-ferret_0.11.8.6-2_armhf.log

for reference

I've done a quick bit of debugging to find the source of the
bug. Here's a gdb stacktrace and variable printout to demonstrate the
problem.

(sid-armhf)steve at mjolnir:~/debian/build/ruby-ferret/ruby-ferret-0.11.8.6$ file core
core: ELF 32-bit LSB core file, ARM, version 1 (SYSV), SVR4-style, from 'ruby2.5 debian/ruby-tests.rb', real uid: 1000, effective uid: 1000, real gid: 1000, effective gid: 1000, execfn: '/usr/bin/ruby2.5', platform: 'v8l'
(sid-armhf)steve at mjolnir:~/debian/build/ruby-ferret/ruby-ferret-0.11.8.6$ gdb /usr/bin/ruby2.5 core
...
Reading symbols from /usr/bin/ruby2.5...(no debugging symbols found)...done.
[New LWP 15636]
[New LWP 15639]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/arm-linux-gnueabihf/libthread_db.so.1".
Core was generated by `ruby2.5 debian/ruby-tests.rb'.
Program terminated with signal SIGBUS, Bus error.
#0  0xf7975e8e in frt_p_new (mp=mp at entry=0x10c7a98, doc_num=doc_num at entry=0, pos=pos at entry=0) at index.c:5180
5180        return occ;
[Current thread is 1 (Thread 0xf7b5c210 (LWP 15636))]
(gdb) bt
#0  0xf7975e8e in frt_p_new (mp=mp at entry=0x10c7a98, doc_num=doc_num at entry=0, pos=pos at entry=0) at index.c:5180
#1  0xf7975f1c in dw_add_posting (mp=mp at entry=0x10c7a98, curr_plists=curr_plists at entry=0x1121438, fld_plists=fld_plists at entry=0x11a6eb8, doc_num=doc_num at entry=0, 
    text=text at entry=0x11c0c68 "20040601", len=len at entry=8, pos=pos at entry=0) at index.c:5523
#2  0xf797628c in frt_dw_invert_field (dw=dw at entry=0xd22bc0, fld_inv=fld_inv at entry=0xf740075a, df=df at entry=0xf6ece8) at index.c:5587
#3  0xf79763f0 in frt_dw_add_doc (dw=dw at entry=0xd22bc0, doc=doc at entry=0xc80b18) at index.c:5662
#4  0xf79764da in frt_iw_add_doc (iw=iw at entry=0xffc218, doc=doc at entry=0xc80b18) at index.c:6337
#5  0xf7987ff0 in frb_iw_add_doc (self=18250200, rdoc=<optimized out>) at r_index.c:1530
#6  0xf7ec4dfa in ?? () from /usr/lib/arm-linux-gnueabihf/libruby-2.5.so.2.5
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb) p mp 
$1 = (FrtMemoryPool *) 0x10c7a98
(gdb) p p
$2 = (FrtPosting *) 0xf7402e7e
(gdb) p *p
$3 = {freq = 0, doc_num = 0, first_occ = 0x0, next = 0x0}
(gdb) p &p->doc_num 
$4 = (int *) 0xf7402e82
(gdb) p &p->first_occ 
$5 = (FrtOccurence **) 0xf7402e86

Ignore the line number 5180 - it looks like gdb is getting the exact
line number wrong. The fault is in the function p_new():

Posting *p_new(MemoryPool *mp, int doc_num, int pos)
{
    Posting *p = MP_ALLOC(mp, Posting);
    p->doc_num = doc_num;
    p->first_occ = occ_new(mp, pos);
    p->freq = 1;
    p->next = NULL;
    return p;
}

MP_ALLOC() has returned a misaligned pointer, and the code following
here therefore is trying to fill things in misaligned too. p->doc_num
is an int, but is not aligned appropriately...

-- System Information:
Debian Release: 9.6
  APT prefers stable-updates
  APT policy: (500, 'stable-updates'), (500, 'stable-debug'), (500, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-8-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8), LANGUAGE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)