small recursive function stack overflow crashes sbcl. Windows 8/ 64bit sbcl

[Rolf Madsen]

The following piece of code:

(defun req-sqrt(v)
    (sqrt (req-sqrt v)))
(req-sqrt 1.0)
(req-sqrt 1.0)

reliable crashes both sbcl 1.1.8 and 1.1.12 on my setup: windows 8 64 bit with 64 bit sbcl.

actually replacing sqrt with sin or + 1 has same behaviour, possibly any function will do.

On 1.1.12 one might have to try twice.

*features*
(:SWANK :QUICKLISP :ASDF3 :ASDF2 :ASDF :OS-WINDOWS :NON-BASE-CHARS-EXIST-P
 :ASDF-UNICODE :ALIEN-CALLBACKS :ANSI-CL :ASH-RIGHT-VOPS
 :C-STACK-IS-CONTROL-STACK :COMMON-LISP :COMPARE-AND-SWAP-VOPS
 :COMPLEX-FLOAT-VOPS :CYCLE-COUNTER :FLOAT-EQL-VOPS :GENCGC
 :IEEE-FLOATING-POINT :INLINE-CONSTANTS :LINKAGE-TABLE :LITTLE-ENDIAN
 :MEMORY-BARRIER-VOPS :MULTIPLY-HIGH-VOPS :OS-PROVIDES-DLOPEN
 :OS-PROVIDES-PUTWC :PACKAGE-LOCAL-NICKNAMES :RAW-INSTANCE-INIT-VOPS :SB-DOC
 :SB-DYNAMIC-CORE :SB-EVAL :SB-FUTEX :SB-LDB :SB-PACKAGE-LOCKS :SB-QSHOW
 :SB-SAFEPOINT :SB-SAFEPOINT-STRICTLY :SB-SIMD-PACK :SB-SOURCE-LOCATIONS
 :SB-TEST :SB-THREAD :SB-THRUPTION :SB-UNICODE :SB-WTIMER :SBCL
 :STACK-ALLOCATABLE-CLOSURES :STACK-ALLOCATABLE-FIXED-OBJECTS
 :STACK-ALLOCATABLE-LISTS :STACK-ALLOCATABLE-VECTORS
 :STACK-GROWS-DOWNWARD-NOT-UPWARD :UNWIND-TO-FRAME-AND-CALL-VOP :WIN32 :X86-64)

[Jan Moringen (scymtym)]

> reliable crashes [...]

Can you describe the failure mode in more detail?

> On 1.1.12 one might have to try twice.

What happens in the first try? The expected "Binding stack exhausted. PROCEED WITH CAUTION."?

Running SBCL 1.1.12 under wine (on x86) produces the following:

$ wine sbcl
This is SBCL 1.1.12, an implementation of ANSI Common Lisp.
[...]
* (defun req-sqrt(v)
    (sqrt (req-sqrt v)))
REQ-SQRT
* (req-sqrt 1.0)
err:seh:setup_exception_record stack overflow 1792 bytes in thread 0009 eip 7ef633e8 esp 00540c30 stack 0x540000-0x541000-0x740000

Is this equivalent to what you experience?

[Rolf Madsen (rolfrm)]

first of all I should probably note that no optimizations are enabled. im running a new install and just enter what i wrote above. It is installed next to an older 1.1.8 installation, but i dont think they interfer in any way. I have simply installed the .msi from sbcl.org.

I have tested more versions now: 1.1.17 x86 and x86-64. Still same problem.

somehow if i run through slime it crashes the first time always and through cmd.exe it always crashes the second time.

>Can you describe the failure mode in more detail?

sbcl.exe exits. Windows pops up a messagebox saying
"sbcl.exe stopped working..."

and then shows some more details. Tried to translate the names from danish so they might be a bit inaccurate.

Problem signature:
  name of problem event: APPCRASH
  program name: sbcl.exe
  Programversion: 0.0.0.0
  Timestamp for program: 524e570c
  Error module name: StackHash_2264
  Error module version: 0.0.0.0
  Timestamp for Error module: 00000000
  Exception code: c0000005
  Undtagelsesforskydning(dont know translation "exception shift"?): PCH_89
  OS-version: 6.3.9600.2.0.0.768.101
  Country standard-id: 1030
  More information 1: 2264
  More information 2: 2264db07e74365624c50317d7b856ae9
  More information 3: 875f
  More information 4: 875fa2ef9d2bdca96466e8af55d1ae6e

>What happens in the first try?
When i run through sbcl.exe i get sb-kernel::control-stack-exhausted and then "...this is possible due to heavily nested or infinetely recursive function cals, or a tail call that SBCL cannot or has not optimized away"

>Is this equivalent to what you experience?
it is not.

** cmd.exe dump (i replaced "req-sqrt" with "r" and used + 1 instead of sqrt) **
C:\Program Files (x86)\Steel Bank Common Lisp\1.1.17>sbcl.exe
This is SBCL 1.1.17, an implementation of ANSI Common Lisp.
More information about SBCL is available at <http://www.sbcl.org/>.

SBCL is free software, provided as is, with absolutely no warranty.
It is mostly in the public domain; some portions are provided under
BSD-style licenses. See the CREDITS and COPYING files in the
distribution for more information.

WARNING: the Windows port is fragile, particularly for multithreaded
code. Unfortunately, the development team currently lacks the time
and resources this platform demands.
* (defun r (v) (+ 1 (r v))) (r 1)

R
*
debugger invoked on a SB-KERNEL::CONTROL-STACK-EXHAUSTED in thread
#<THREAD "main thread" RUNNING {240F3CF1}>:
  Control stack exhausted (no more space for function call frames).
This is probably due to heavily nested or infinitely recursive function
calls, or a tail call that SBCL cannot or has not optimized away.

PROCEED WITH CAUTION.

Type HELP for debugger help, or (SB-EXT:EXIT) to exit from SBCL.

restarts (invokable by number or by possibly-abbreviated name):
  0: [ABORT] Exit debugger, returning to top level.
0

(R 1)
0] 0

* (r 1)

C:\Program Files (x86)\Steel Bank Common Lisp\1.1.17>

** **
sbcl crashes between the last (r 1) and C:\Program FIles (x86)...

[Douglas Katzman (dougk)]

Not sure what's being asked. Of course if overflows the stack since it's infinitely recursive.

[Luís Oliveira (luismbo)]

Should be fixed in https://sourceforge.net/p/sbcl/sbcl/ci/15870aab53d9440006d379a1b97b0f00b651f170/