NetBSD/src gcejoLlrescue list.ldd

   rescue/ldd: Add a couple new symbols for recent rtld changes.

   Should fix:

   #      link  rescue/rescue
   ...
   /tmp/build/2026.07.01.20.39.44-i386/tools/lib/gcc/i486--netbsdelf/14.3.0/../../../../i486--netbsdelf/bin/ld: /tmp/build/2026.07.01.20.39.44-i386/obj/usr.bin/ldd/elf32/libldd_elf32.a(load.o): in function `_rtld_load_object':
   load.c:(.text+0x23b): undefined reference to `_rtld_objgen'
   /tmp/build/2026.07.01.20.39.44-i386/tools/lib/gcc/i486--netbsdelf/14.3.0/../../../../i486--netbsdelf/bin/ld: load.c:(.text+0x256): undefined reference to `_rtld_objrelocpending'

   PR lib/59751: dlclose is not MT-safe depending on the libraries
   unloaded
VersionDeltaFile
1.8+2-1rescue/list.ldd
+2-11 files

NetBSD/src s9PmQw9doc 3RDPARTY

   libevent-2.1.13 out
VersionDeltaFile
1.2227+3-3doc/3RDPARTY
+3-31 files

NetBSD/src ESox8jIlibexec/ld.elf_so rtld.c load.c, usr.bin/ldd ldd.c

   ld.elf_so(1): Bump _rtld_objgen when changing, not reading, objlist.

   Prompted by:

   PR lib/59751: dlclose is not MT-safe depending on the libraries
   unloaded
VersionDeltaFile
1.226+5-4libexec/ld.elf_so/rtld.c
1.51+3-2libexec/ld.elf_so/load.c
1.30+3-2usr.bin/ldd/ldd.c
+11-83 files

NetBSD/src 1hvHvI5libexec/ld.elf_so rtld.c load.c, usr.bin/ldd ldd_elfxx.c

   ld.elf_so(1): Resolve several races in dlopen/dlclose.

   This is difficult because, although rtld generally has a single
   exclusive lock, i.e., generally runs single-threaded itself, it can't
   hold this lock while calling constructors/destructors (init/fini or
   ifunc) -- if it did, then, for example, lazy symbol binding that
   happens during the constructor/destructor would deadlock against
   itself.

   And whenever rtld drops the lock to call constructors/destructors,
   any objects it is working on, during dlopen or dlclose, might have
   been concurrently closed and invalidated by the time it gets the lock
   again.

   The key point is that anywhere we pass a sigset_t *mask parameter
   during dlopen or dlclose, we might release the rtld lock to sleep and
   then reacquire the lock.  And anywhere we might release and reacquire
   the lock, any objects we hold may be invalidated -- unless we hold some
   reference to prevent invalidation.  And any object we find in the list

    [129 lines not shown]
VersionDeltaFile
1.225+457-33libexec/ld.elf_so/rtld.c
1.50+84-15libexec/ld.elf_so/load.c
1.156+37-5libexec/ld.elf_so/rtld.h
1.28+16-10libexec/ld.elf_so/search.c
1.121+6-2libexec/ld.elf_so/reloc.c
1.9+3-3usr.bin/ldd/ldd_elfxx.c
+603-682 files not shown
+608-748 files

NetBSD/src nWoD803sys/dev/marvell mvsdio.c, sys/dev/sdmmc sdmmcvar.h

   sdmmc: remove unused cmd fields

   c_dmaseg and c_dmaoff are only written using memset(&cmd, 0) when
   constructing a command, so they always assume a constant value and can be
   removed.
VersionDeltaFile
1.9+4-5sys/dev/marvell/mvsdio.c
1.40+1-3sys/dev/sdmmc/sdmmcvar.h
+5-82 files

NetBSD/src qe0Bxomsys/arch/mips/mips cache.c

   Wire up L2 ops for the MIPS32/64 detection path.

   Previously it detected L2 cache but installed no ops (assuming coherency).
VersionDeltaFile
1.70+60-2sys/arch/mips/mips/cache.c
+60-21 files

NetBSD/src DD7qQd7sys/arch/mips/mips mips_machdep.c

   Remove the false CPU_MIPS_D_CACHE_COHERENT flag for XBurst.

   There is no DMA coherency on that core.
VersionDeltaFile
1.310+6-4sys/arch/mips/mips/mips_machdep.c
+6-41 files

NetBSD/src iTRJuMzbin/sh jobs.c trap.c

   /bin/sh fix trap/jobid builtins when run from grandchild

   The trap and jobid builtins are designed to allow

        T=$(trap)

   (etc) to see (output) the traps (or jobs for jobid) of parent shell (the one
   running the assignment) to be seen from inside a subshell (here the command
   substitution, which it usually would be).

   But all that should be reported are the traps (jobs) of the parent of
   that subshell, not recursively up the chain.

   That is

        ( T=$(trap) )

   should usually be equivalent to T='' as inside a subshell environment
   all non-ignored traps are defined to be reset to their defaults

    [33 lines not shown]
VersionDeltaFile
1.127+42-9bin/sh/jobs.c
1.63+9-5bin/sh/trap.c
1.200+3-3bin/sh/eval.c
+54-173 files

NetBSD/src olzB1M0tests/bin/sh t_builtins.sh

   sh tests: builtins: add more content

   Many of the t_builtins sh tests were just placeholders for things to come.
   They've been that way for years...

   This adds some actual content to some of them - still very primitive in
   most cases, but better than was there before.   There are still too many
   empty tests (test the builtin exists, but that's it).

   Also, add a test for the trap builtin - that wasn't there before as traps
   are tested in other of the sh tests ... this new one is therefor not testing
   that traps work (or not in any way that's meaningful), rather it is testing
   that the various ways the trap command can be used all work.

   Currently two of the sub-tests will fail, the new trap test, and the
   (was previously empty) jobid test (testing a NetBSD extension builtin).

   Adding those failing tests was the inspiration for all of this.  The bug
   they expose is exotic, and never likely to be encountered, or do any harm

    [7 lines not shown]
VersionDeltaFile
1.9+235-37tests/bin/sh/t_builtins.sh
+235-371 files

NetBSD/src 71eZpxbtests/bin/sh t_arith.sh

   sh tests: arithmetic: fix test for support of pre ++/-- ops

   The C prefix and postfix unary ++ and -- operators are optional
   in POSIX sh (supported in NetBSD's sh).   To allow these tests to be
   used to test other shells, the tests generally attempt to verify that
   optional features are supported, before testing them.

   The prefix ++ and -- tests, to see if there was support in ${TEST_SH}
   were not guaranteed to work, as ++X might be preincrement X, or it
   might just be + + X (ie: applying the unary + operator twice).

   Fix that by testing whether ++X and + + X give the same result, if
   they do, then prefix ++ is not supported (also if ++X generates an error)

   Just the same for the prefix -- operator.

   The postfix operator tests are OK, X-- is either X-- or an error.

   NFCI for tests of the NetBSD sh.   (This change has been in my local
   set of sh tests for a long time now.)
VersionDeltaFile
1.11+11-3tests/bin/sh/t_arith.sh
+11-31 files

NetBSD/src nIb7Nhesys/net if_wg.c

   Avoid stupid macro name clash: m68k/include/reg.h defines R0 and R1
   as the register offsets for returning values from functions in the.
   These clash with local buffers R0 and R1 here.
   Rename them to cookie_R0 and cookie_R1 to make all m68k ports build again.
VersionDeltaFile
1.140+7-7sys/net/if_wg.c
+7-71 files

NetBSD/src XHggdlWetc Makefile, etc/unbound unbound.conf Makefile

   unbound: Install a default configuration.

   This way, setting unbound=YES is enough to get a working local
   recursive resolver listening on 127.0.0.1:53 and [::1]:53.

   PR misc/60339: unbound=YES in rc.conf should be enough to get a
   recursive resolver

   Additionally, this includes the workaround for:

   PR bin/60325: unbound emits warning: so-sndbuf 4194304 was not
   granted
VersionDeltaFile
1.1+49-0etc/unbound/unbound.conf
1.84+15-2share/man/man8/afterboot.8
1.3+14-1external/bsd/unbound/etc/rc.d/unbound
1.1+14-0etc/unbound/Makefile
1.195+12-1share/man/man5/rc.conf.5
1.477+2-2etc/Makefile
+106-62 files not shown
+110-88 files

NetBSD/src 2kji2lusys/arch/x86/x86 efi_machdep.c

   x86/efi_machdep.c: Nix extra blank line.

   No functional change intended.
VersionDeltaFile
1.8+2-3sys/arch/x86/x86/efi_machdep.c
+2-31 files

NetBSD/src K93C2bJsys/arch/x86/x86 efi_machdep.c

   x86/efi: Bypass kASAN memcpy when copying the efi runtime.

   Fixes crash at boot with `options KASAN' enabled on EFI systems.

   We could try to find some way to exclude the EFI runtime mappings
   from kASAN, e.g. by hacking kasan_md_unsupported.  But there's a lot
   of detailed exclusions to pass through to that (every entry in the
   EFI memory map), and this is the only memcpy in the whole kernel that
   all those details are likely to be applicable.

   PR kern/60391: kernel panic in efi_init() with "options KASAN"
VersionDeltaFile
1.7+9-3sys/arch/x86/x86/efi_machdep.c
+9-31 files

NetBSD/src OZVCckcsys/arch/ofppc/conf EFIKA

   Add kernel config for EFIKA 5200B. The board is supported now.
VersionDeltaFile
1.1+271-0sys/arch/ofppc/conf/EFIKA
+271-01 files

NetBSD/src 9QpNxuBsys/arch/powerpc/mpc5200 mpc5200_ac97.c

   Add debugging to MPC5200B AC97 driver.

   An attempt to hunt down rare cold boot race.
VersionDeltaFile
1.2+26-7sys/arch/powerpc/mpc5200/mpc5200_ac97.c
+26-71 files

NetBSD/src Ui0b2ucsys/arch/powerpc/conf files.ofw, sys/arch/powerpc/oea ofw_subr.S

   Introduce EFIKA_OFW_WORKAROUNDS - patch OFW trampoline for EFIKA.

   Long story short, is that SmartFirmware (EFIKA's OpenFirmware) misbehaves
   on every OF call from the kernel by zeroing SDR1 and messing with BATs.

   In the olden days, these problems were worked around with FIRMWORKSBUGS
   option. However, that option does not work currently on EFIKA. Presumaly,
   something has changed between 2012 and now, which causes FIRMWORKSBUGS
   to hang on SmartFirmware very early (before NetBSD copyright banner is
   printed).

   So we get this kludge, which only does the bare minimum to let OFW
   call succeed, and the kernel continue after the call.
VersionDeltaFile
1.21+48-1sys/arch/powerpc/oea/ofw_subr.S
1.7+9-1sys/arch/powerpc/conf/files.ofw
+57-22 files

NetBSD/src eage6aZlib/libc/citrus/modules citrus_viqr.c

   citrus/VIQR: unsigned chlen for simplicity; sprinkle assertions.

   PR 59019: various iconv issues
VersionDeltaFile
1.9+14-3lib/libc/citrus/modules/citrus_viqr.c
+14-31 files

NetBSD/src 0n1zEmmlib/libc/citrus/modules citrus_utf7.c

   citrus/UTF7: unsigned chlen for simplicity; sprinkle assertions.

   PR 59019: various iconv issues
VersionDeltaFile
1.9+26-6lib/libc/citrus/modules/citrus_utf7.c
+26-61 files

NetBSD/src 10gCbbolib/libc/citrus/modules citrus_hz.c

   citrus/HZ: Be more careful about encoding buffer.

   1. Use unsigned char for chlen, since it's never negative and only
      needs to be big enough for the ch array.

   2. Expand the ch array by space for two escape sequences, each of
      which is two characters.

      I don't think it is currently possible with the i18n data we ship
      for it to emit two escape sequences in a row -- there is a code
      path to do this, but I don't think it can be reached without
      setting escape sequences for variable 1 and extending this in some
      way.  Nevertheless, let's err on the side of safety.

   3. Sprinkle assertions everywhere.

   PR 59019: various iconv issues
VersionDeltaFile
1.7+22-6lib/libc/citrus/modules/citrus_hz.c
+22-61 files

NetBSD/src D74v71Slib/i18n_module Makefile.inc, lib/libc/citrus/modules citrus_viqr.c citrus_hz.c

   i18n_modules: Build with _DIAGNOSTIC.

   Fix some bitrot that it turned up here.

   PR 59019: various iconv issues
VersionDeltaFile
1.8+19-6lib/libc/citrus/modules/citrus_viqr.c
1.6+6-4lib/libc/citrus/modules/citrus_hz.c
1.17+6-4lib/libc/citrus/modules/citrus_iconv_std.c
1.15+2-1lib/i18n_module/Makefile.inc
+33-154 files

NetBSD/src GLq16Yytests/lib/libc/locale t_iconv.c

   iconv(3): Fix printing bytes on test failure.

   Don't sign-extend to a 32-bit quantity causing the left to be padded
   with f's; just print the 8-bit quantity.

   PR 59019: various iconv issues
VersionDeltaFile
1.8+5-5tests/lib/libc/locale/t_iconv.c
+5-51 files

NetBSD/src exLfYVQtests/lib/libc/locale t_iconv.c

   iconv(3): Expand ISO-2022-CN tests a little.

   This one is broken for a partial input case, not yet sure why.

   PR 59019: various iconv issues
VersionDeltaFile
1.7+19-4tests/lib/libc/locale/t_iconv.c
+19-41 files

NetBSD/src tLcn45Jtests/lib/libc/locale t_iconv.c

   iconv(3): Add a test case for ISO-2022-CN.

   From Mark Johnston of FreeBSD.

   PR 59019: various iconv issues
VersionDeltaFile
1.6+7-2tests/lib/libc/locale/t_iconv.c
+7-21 files

NetBSD/src iXtkor3lib/libc/citrus/modules citrus_zw.c, tests/lib/libc/locale t_iconv.c

   iconv(3): Fix buffer overrun in ZW encoding.

   Based on a report by Nick Wellnhofer.

   PR 59019: various iconv issues
VersionDeltaFile
1.7+5-2lib/libc/citrus/modules/citrus_zw.c
1.5+3-3tests/lib/libc/locale/t_iconv.c
+8-52 files

NetBSD/src KdxNOXslib/libc/citrus/modules citrus_viqr.c, tests/lib/libc/locale t_iconv.c

   iconv(3): Fix buffer overrun in VIQR encoding.

   Based on a report by Nick Wellnhofer.

   PR 59019: various iconv issues
VersionDeltaFile
1.7+7-3lib/libc/citrus/modules/citrus_viqr.c
1.4+3-3tests/lib/libc/locale/t_iconv.c
+10-62 files

NetBSD/src XWZH6ktlib/libc/citrus/modules citrus_utf7.c, tests/lib/libc/locale t_iconv.c

   iconv(3): Fix buffer overrun in UTF-7 encoding.

   Based on a report by Nick Wellnhofer.

   PR 59019: various iconv issues
VersionDeltaFile
1.8+16-6lib/libc/citrus/modules/citrus_utf7.c
1.3+5-5tests/lib/libc/locale/t_iconv.c
+21-112 files

NetBSD/src Pv4xBtflib/libc/citrus/modules citrus_hz.c, tests/lib/libc/locale t_iconv.c

   iconv(3): Fix null pointer dereference with HZ8 encoding.

   Based on a report by Nick Wellnhofer.

   PR 59019: various iconv issues
VersionDeltaFile
1.5+4-2lib/libc/citrus/modules/citrus_hz.c
1.2+2-4tests/lib/libc/locale/t_iconv.c
+6-62 files

NetBSD/src zZNqQJ5distrib/sets/lists/debug mi, distrib/sets/lists/tests mi

   iconv(3): Draft some automatic tests.

   Based on a report by Nick Wellnhofer.

   PR 59019: various iconv issues
VersionDeltaFile
1.1+354-0tests/lib/libc/locale/t_iconv.c
1.1423+2-1distrib/sets/lists/tests/mi
1.19+2-1tests/lib/libc/locale/Makefile
1.515+2-1distrib/sets/lists/debug/mi
+360-34 files

NetBSD/src UzWRGXksys/arch/ofppc/include intr.h

   Increase the number of possible interrupt sources.

   Needed for MPC5200B, harmless for others.
VersionDeltaFile
1.17+2-2sys/arch/ofppc/include/intr.h
+2-21 files