Re: kern/59100: compat fixes from Nov 22 2024 break cvsup 1.6ZC static binary

["Onno van der Linden via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR kern/59100; it has been noted by GNATS.

From: Onno van der Linden <o.vd.linden%quicknet.nl@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/59100: compat fixes from Nov 22 2024 break cvsup 1.6ZC
 static binary
Date: Wed, 26 Feb 2025 21:23:46 +0100

 compat_netbsd32_16_init() in compat/netbsd32/netbsd32_compat_16.c 
 places netbsd32_machdep_md_16_init() inside an #ifdef construction
 that's skipped on amd64. Due to the skipping of netbsd32_machdep_md_16_init()
 netbsd32_sendsig_sigcontext_16_hook.hooked isn't set to a non-zero value
 and sigcontext_valid in netbsd32___sigaction_sigtramp() thus remains 0.
 
 Instead of the quick and dirty hack of setting  sigcontext_valid to 1 in
 netbsd32___sigaction_sigtramp() netbsd32_machdep_md_16_init() should be called
 for architectures (like amd64) that need it.
 
 >  So why isn't the 10.99.12 kernel ktrace not showing any NAMI lines for the kmod files ?
 
 That's also caused by the current version of netbsd32_compat_16.c
 The exec_sigcode_alloc() inside of #ifdef __HAVE_MD_NETBSD32_SIGCODE
 probably triggers some internal kmod loading which might or might not
 be necessary.