Re: Test failure on ARM

[Jerry J. <log...@gm...>]

On Thu, May 24, 2012 at 5:24 PM, Jerry James <log...@gm...> wrote:
> Maybe these problems have been fixed since the 2.49 release.  I tried
> building with mercurial head, only to have a different build failure
> because host_cpu_c_abi is set to "armel" instead of "arm" in makemake,
> resulting in all of the ARM-specific code being skipped.  That's all
> the time I have to mess with it today.  I'll try to patch that up
> tomorrow.

I don't know what the right fix for this is.  Just to get the build
going again, I took the dumb but simple approach:

--- src/makemake.in.orig	2012-05-24 09:42:10.000000000 -0600
+++ src/makemake.in	2012-05-24 17:28:54.160000792 -0600
@@ -1130,6 +1130,10 @@
     * ) cpu=${host_cpu}
   esac
 fi
+# Kludge to get the arm build going again
+if test "$cpu" = armel; then
+  cpu=arm
+fi

 test "${verbose}" = true -o "${verbose}" = yes && \
   cat <<EOF >&2

I kicked off another build with mercurial tip this morning.  It just failed:

boot/lisp.run -B . -M boot/lispinit.mem -norc -q -i i18n/preload.lisp
-i syscalls/preload.lisp -i regexp/preload.lisp -x (saveinitmem
"base/lispinit.mem")
;; Loading file i18n/preload.lisp ...
;; Loaded file i18n/preload.lisp
;; Loading file syscalls/preload.lisp ...
;; Loaded file syscalls/preload.lisp
;; Loading file regexp/preload.lisp ...
;; Loaded file regexp/preload.lisp
*** - handle_fault error2 ! address = 0x36bcd0f4 not in
[0x1abf0000,0x1ad3bf28) !
SIGSEGV cannot be cured. Fault address = 0x36bcd0f4.
GC count: 0
Space collected by GC: 0 0
Run time: 0 420000
Real time: 1 410038
GC time: 0 0
Permanently allocated: 90656 bytes.
Currently in use: 1971168 bytes.
Free space: 390112 bytes.
./clisp-link: line 97: 22968 Segmentation fault      (core dumped) "$@"


I put a breakpoint at spvw_fault.d:220, and ran it again:
(gdb) bt
#0  handle_fault (verbose=0, address=915259636) at ../src/spvw_fault.d:220
#1  sigsegv_handler (fault_address=0x368dc0f4, serious=0)
    at ../src/spvw_sigsegv.d:74
#2  0xb6f33c54 in ?? () from /lib/libsigsegv.so.2
#3  <signal handler called>
#4  gc_morris2 (page=<optimized out>) at ../src/spvw_garcol.d:404
#5  gar_col_normal () at ../src/spvw_garcol.d:2327
#6  0x00030a34 in do_gar_col () at ../src/spvw_garcol.d:3055
#7  0x000d4eec in with_gc_statistics (fun=fun@entry=0x30a18 <do_gar_col>)
    at ../src/predtype.d:3123
#8  0x00026d18 in gar_col (level=level@entry=1) at ../src/spvw_garcol.d:3065
#9  0x00032454 in savemem (stream=<optimized out>,
    executable=executable@entry=0) at ../src/spvw_memfile.d:434
#10 0x0005eea4 in C_savemem () at ../src/pathname.d:8851
#11 0x00038564 in interpret_bytecode_ (closure=0x19cf84a1, codeptr=0x19cf80c8,
    byteptr=0x368dc0f4 <Address 0x368dc0f4 out of bounds>)
    at ../src/eval.d:7655
#12 0x00040ab0 in eval_closure (closure=<optimized out>) at ../src/eval.d:3917
#13 eval1 (form=<optimized out>, form@entry=0x42100) at ../src/eval.d:3091
#14 0x00042100 in eval (form=0x42100) at ../src/eval.d:2966
#15 0x000c82b8 in C_read_eval_print () at ../src/debug.d:409
#16 0x00033850 in funcall_subr (fun=0x1c17e6, args_on_stack=<optimized out>)
    at ../src/eval.d:5221
#17 0x0003d438 in interpret_bytecode_ (closure=0x19d3bf55, codeptr=0x19cf68fc,
    byteptr=0x368dc0f4 <Address 0x368dc0f4 out of bounds>)
    at ../src/eval.d:6793
#18 0x00044384 in funcall_closure (closure=<optimized out>,
    args_on_stack=<optimized out>) at ../src/eval.d:5624
#19 0x0004ad54 in C_driver () at ../src/control.d:1999
#20 0x0003d328 in interpret_bytecode_ (closure=0x19cf6afd, codeptr=0x19cf689c,
    byteptr=0x368dc0f4 <Address 0x368dc0f4 out of bounds>)
    at ../src/eval.d:6799
#21 0x00044384 in funcall_closure (closure=<optimized out>,
    args_on_stack=<optimized out>) at ../src/eval.d:5624
#22 0x00032078 in main_actions (p=0x1d7ce0) at ../src/spvw.d:3606
#23 0x00020378 in main (argc=15, argv=0xbefff8b4) at ../src/spvw.d:3877
(gdb) info locals
heap = 0x1d74fc
pa_address = 915259392
obj = 0x368dc0f4
heapnr = 0
(gdb) print *heap
$1 = {pages = {start = 431894528, end = 433331964, gcpriv = {
      firstmarked = 0x19be3000, l = 431894528, d = 431894528,
      next = 0x19be3000}}, heap_limit = 433332224,
  heap_gen0_start = 431894528, heap_gen0_end = 433254184,
  heap_gen1_start = 433254400, physpages = 0x1dbad0}

Any ideas?  I'll kick off another build with generational GC turned
off, but these builds take several hours, so I probably won't see the
results until Monday morning.
-- 
Jerry James
http://www.jamezone.org/