Re: bin/60150: named(8) crashes at startup on NetBSD/i386 11.0_RC2

["RVP via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR bin/60150; it has been noted by GNATS.

From: RVP <rvp%SDF.ORG@localhost>
To: Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>
Cc: gnats-bugs%netbsd.org@localhost
Subject: Re: bin/60150: named(8) crashes at startup on NetBSD/i386 11.0_RC2
Date: Sun, 5 Apr 2026 10:51:36 +0000 (UTC)

 On Sat, 4 Apr 2026, Izumi Tsutsui wrote:
 
 > On NetBSD/i386, sizeof(max_align_t) is 12 bytes so
 > sizeof(union size_info) is also 12 byets.
 >
 
 Yes, and `alignof(max_align_t)' is 4 which'll also stomp on the flag bits
 encoded in the lowest 3 bits of the pointer address.
 
 > The returned addess of mallocx() calculated by `ptr = &si[1];`
 > is not 8 byte aligned.
 >
 
 Yes, it _must_ be at least 8-byte aligned, but on NetBSD/i386, as we've seen,
 it isn't. Don't know what the reasoning behind the `max_align_t __alignment;'
 there is. The usual way, if the pointer immediately following, should also be
 aligned, is to do `alignas(max_align_t) ...' but this requires a C11 compiler.
 
 In any case, someone should definitely ask the BIND people.
 
 > Actually the following ugly patch fixes the assertion of _cds_lfht_add()
 > in liburcu:
 >
 > ---
 > Index: dist/lib/isc/jemalloc_shim.h
 > ===================================================================
 > RCS file: /cvsroot/src/external/mpl/bind/dist/lib/isc/jemalloc_shim.h,v
 > retrieving revision 1.4
 > diff -u -p -d -r1.4 jemalloc_shim.h
 > --- dist/lib/isc/jemalloc_shim.h	26 Jan 2025 16:25:37 -0000	1.4
 > +++ dist/lib/isc/jemalloc_shim.h	3 Apr 2026 20:13:47 -0000
 > @@ -30,9 +30,17 @@ const char *malloc_conf = NULL;
 >
 > #include <stdlib.h>
 >
 > +#ifndef ALIGNMENT
 > +#define ALIGNMENT 8U
 > +#endif
 > +#ifndef roundup2
 > +#define roundup2(x,m)   ((((x) - 1) | ((m) - 1)) + 1)
 > +#endif
 > +
 > typedef union {
 > 	size_t size;
 > 	max_align_t __alignment;
 > +	uint8_t __roundup[roundup2(sizeof(max_align_t), ALIGNMENT)];
 > } size_info;
 >
 > static inline void *
 >
 
 I've fixed it slightly differently:
 
 ```
 --- src/external/mpl/bind/dist/lib/isc/jemalloc_shim.h.orig	2025-01-26 16:25:37.000000000 +0000
 +++ src/external/mpl/bind/dist/lib/isc/jemalloc_shim.h	2026-04-05 10:29:53.567824881 +0000
 @@ -32,7 +32,16 @@
 
   typedef union {
   	size_t size;
 -	max_align_t __alignment;
 +	/*
 +	 * should be `alignas(max_align_t) char __alignment;', but this,
 +	 * using NetBSD's stddef.h, yields a 4-byte alignment (16 if you
 +	 * use GCC's own stddef.h).
 +	 *
 +	 * Remove uncertainty by explicitly padding to 16 bytes, ie. malloc(3)
 +	 * alignment. (Actually, the pointer following this header need only
 +	 * be 8-byte aligned.)
 +	 */
 +	char __alignment[16];
   } size_info;
 
   static inline void *
 ```
 
 HTH,
 
 -RVP