Import postfix 3.11.2 (previous was 3.10.1)
Changes in 3.11.2
Bugfix (defect introduced: Postfix 3.11): the proxymap(8) daemon
dereferenced an uninitialized pointer after a request protocol
error. This daemon is not exposed to local or remote users. Found
by Claude Opus 4.6.
Bugfix (defect introduced: 20260309) a change, to set the service_name
default value to "amnesiac", violated a test that parameter names
in postconf output must match 1:1 with parameter names in the
postlink script.
Changes in 3.11.1
Bugfix (defect introduced: 20260219): alias_maps errors when
default_database_type was not set in main.cf. Fix by Michael Tokarev.
[97 lines not shown]
Pull up following revision(s) (requested by mrg in ticket #282):
libexec/httpd/bozohttpd.c: revision 1.153
libexec/httpd/bozohttpd.c: revision 1.154
libexec/httpd/CHANGES: revision 1.58
libexec/httpd/tilde-luzah-bozo.c: revision 1.17
libexec/httpd/bozohttpd.8: revision 1.102
Fix double free of hr_file (later handled by bozo_clean_request())
Also hr_file_free should be used for free(3) (it fixes "//~user/" case
in which hr_file points to the last slash in the prefix).
Fix minor memory leak
bump version for last 2 memory fixes.
make.1 indent description of '-' impact in jobs-mode
Shift the '.El' so that the paragraph describing how jobs-mode handles
lines starting with '-' is clearly part of the description of '-'.
Pull up following revision(s) (requested by riastradh in ticket #1266):
sys/kern/vfs_subr.c: revision 1.503
Correct reversed sense of IMNT_MPSAFE test in VFS_MOUNT(9).
Pull up following revision(s) (requested by riastradh in ticket #281):
sys/kern/vfs_subr.c: revision 1.503
Correct reversed sense of IMNT_MPSAFE test in VFS_MOUNT(9).
Additionaly pull up the following for ticket #262:
sys/arch/x86/x86/via_padlock.c 1.38
to fix the i386 build:
Update to match recent cryptodev update
<opencrypto/cryptodev.h> recently changed a bunch of fields from
signed to unsigned. To avoid gcc stupidly complaining about
comparisons of different signedness ints (I mean, how can
for (i = 0; i < unsigned_thing; i++)
ever see i become negative?) change i from int to unsigned int.
Also pull up for ticket #1262:
sys/arch/x86/x86/via_padlock.c 1.38
to fix the i386 build:
Update to match recent cryptodev update
<opencrypto/cryptodev.h> recently changed a bunch of fields from
signed to unsigned. To avoid gcc stupidly complaining about
comparisons of different signedness ints (I mean, how can
for (i = 0; i < unsigned_thing; i++)
ever see i become negative?) change i from int to unsigned int.
Apply patch, requested by mrg in ticket #280:
distrib/sets/lists/base/shl.mi
distrib/sets/lists/comp/mi
distrib/sets/lists/debug/shl.mi
Fix set lists after pullups for bind, xz, OpenSSH and OpenSSL.
Fix dumpfs to allow NAME=wedge-label on command line
This is a follow-on from the PR bin/59957 changes, and is being
included with that PR so it is easier to locate all the tools that
have been changed this way, so they can all be updated again once
the new code is moved to libutil where it belongs, and out of src/sbin/fsck.
When handling of NAME=whatever was added to dumpfs, it was done
in a way where it could only work when dumpfs was used as
dumpfs mount-point
and /etc/fsck contained
NAME=whatever mount-point [rest of fstab entry]
That covers the vast majority of actual uses, but a simple
dumpfs NAME=whatever
[5 lines not shown]
Pull up following revision(s) (requested by skrll in ticket #279):
common/lib/libc/atomic/atomic_cas_16_cas.c: revision 1.4
common/lib/libc/atomic/atomic_cas_16_cas.c: revision 1.5
common/lib/libc/atomic/atomic_cas_8_cas.c: revision 1.4
common/lib/libc/atomic/atomic_cas_32_cas.c: revision 1.3
common/lib/libc/atomic/atomic_cas_8_cas.c: revision 1.5
common/lib/libc/atomic/atomic_cas_8_cas.c: revision 1.6
tests/lib/libc/atomic/t___sync_compare_and_swap.c: revision 1.4
tests/lib/libc/atomic/t___sync_compare_and_swap.c: revision 1.5
common/lib/libc/atomic/atomic_init_testset.c: revision 1.22
common/lib/libc/atomic/atomic_init_testset.c: revision 1.23
common/lib/libc/atomic/atomic_cas_64_cas.c: revision 1.4
common/lib/libc/atomic/atomic_cas_by_cas32.c: revision 1.5
PR/56839 GCC emits wrong codes for compare_and_swap_1 bultins on armv5 (el & eb)
There is mismatch in signedness of the GCC builtin __sync_* function arguments
and the _atomic_* functions so we cannot directly alias them. Instead write
[13 lines not shown]
Pull up following revision(s) (requested by roy in ticket #278):
sys/net/if.c: revision 1.537
sys/net/if.h: revision 1.310
if: replace link state change queue with state transition
if_link_queue is now a normal bitmask rather than holding a queue.
It holds three bits to mirror the link state - UNKNOWN, DOWN and UP.
There are also some bits to indicate that the link state has been scheduled
for change and if it has been locked for changes (ie the interface being
destroyed).
The logic is simple - transitioning to DOWN will remove UNKNOWN and UP,
transitioning to UNKNOWN will remove UP (no driver should do this).
This means that even in the event of transitions happening faster than
the kernel can spit them out, the correct state of the link will be
[7 lines not shown]
Pull up following revision(s) (requested by mrg in ticket #277):
sys/fs/cd9660/cd9660_rrip.c: revision 1.19
cd9660: make sure that NM records are at least 5 bytes long.
avoids an integer underflow when this length has 5 subtracted from it
for a later path.
Reported by Adam Crosser, Praetorian
