Re: "insn does not satisfy its constraints" compiling perl

[John Paul Adrian Glaubitz <glaubitz%physik.fu-berlin.de@localhost>]

Hi John,

On Sat, 2026-02-28 at 20:33 +0000, John Klos wrote:
> My NetBSD 11_RC1 VAXstation VLC is working valiantly to compile the latest 
> perl from pkgsrc. Two days and 16 hours later:
> 
> gcc -c -DPERL_CORE -O2 -pthread -I/usr/include -fwrapv 
> -fno-strict-aliasing -pipe -I/usr/local/include -std=c99 -O2 -pthread 
> -I/usr/include -Wall -Werror=pointer-arith -Werror=vla -Wextra 
> -Wno-long-long -Wno-declaration-after-statement -Wc++-compat 
> -Wwrite-strings -Wno-use-after-free -DPIC -fPIC toke.c
> toke.c: In function 'Perl_valid_identifier_pve':
> toke.c:13972:1: error: insn does not satisfy its constraints:
> 13972 | }
>        | ^
> (insn 406 405 74 40 (parallel [
>              (set (reg:SI 0 %r0 [112])
>                  (zero_extract:SI (mem/u:SI (reg:SI 0 %r0) [2  S4 A32])
>                      (const_int 8 [0x8])
>                      (const_int 16 [0x10])))
>              (clobber (reg:CC 16 %psl))
>          ]) "toke.c":13945:12 470 {*extzv_aligned_2}
>       (expr_list:REG_UNUSED (reg:CC 16 %psl)
>          (nil)))
> during RTL pass: cprop_hardreg
> toke.c:13972:1: internal compiler error: in extract_constrain_insn, at 
> recog.cc:2698
> no stack trace because unwind library not available
> Please submit a full bug report, with preprocessed source (by using 
> -freport-bug).
> See <http://www.NetBSD.org/support/send-pr.html> for instructions.
> *** Error code 1

What version of GCC is this?

Could you create the preprocessed source together with the source (toke.c)
by passing -save-temps to the GCC command line?

gcc -c -DPERL_CORE -O2 -pthread -I/usr/include -fwrapv \
    -fno-strict-aliasing -pipe -I/usr/local/include -std=c99 -O2 -pthread \
    -I/usr/include -Wall -Werror=pointer-arith -Werror=vla -Wextra \ 
    -Wno-long-long -Wno-declaration-after-statement -Wc++-compat \
    -Wwrite-strings -Wno-use-after-free -DPIC -fPIC toke.c -save-temps

If it's a recent version of GCC, it should be reported to the GCC bug tracker:

https://gcc.gnu.org/bugzilla/

> I can't say I've seen this one before. For whatever reason, it compiled 
> fine with "-O0".

That's not unusual. It's usually one particular optimization that triggers
an internal compiler error. If it doesn't build with -O1 but with -O0, you
can try adding the following optimizations for O1 manually until it fails:

https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html


Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer
`. `'   Physicist
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913