FreshBSD

    Eliminate (most) double decrements of remaining space in the pretty
    print buffer. Use consistant idiom to increment pointer to data
    being consumed, instead of hiding some increments.

    Assemble outgoing packets using iovec to combine the (optional)
    ethernet header, the ip header, the udp header and the packet.
    Rather than manually concatenating the headers and then using iovec
    to bundle in the packet. Eliminates some memcpy()s and convoluted
    data copying.

    make the word-matching of the autoinstaller config file allow skipping
    the trailing question mark, such as in "hostname = foo"
    
    noted by, discussed with, and ok rpe@

    Fix typo; that teaches me to steal other people's diffs!

    Make obj->pages a simple array instead of an array of bus_dma_segment_t's.
    Simplifies things a bit and reduces the diffs with Linux a bit too.

    Better bus error diagnostics (this really only affects IP28). Also, do what
    the comments say and really only ignore speculative accesses causing bus
    errors when coming from the kernel.

    Revert change to sha256: ramdisks don't build with this and the obvious
    conf/files change isn't enough due to size limits

    Revert the previous commit; this is not the right approach.

    Use an u_int32_t instead of an unsigned long to store the magic number
    we're about to send.  Should a bug on big-endian LP64 archs reported by
    James MacMahon (jwm at operand dot ca).  ok millert@

    update a few RC4 references to ChaCha. thanks sthen

    Use a correct pexp and unbreak stop/reload.  The old and wrong pexp
    in /var/run/rc.d/identd has to be manually removed.
    Reported by Adam Jeanguenat (avj at voyager dot 6v6 dot org).
    ok dcoppa@ lteo@

    revert to return EPERM for sem_init pshared until it really works
    reported by zhuk

    remove unused include dating from the pre-termios code and accidentally
    reintroduced with the 4.4lite merge 19 years ago

    do not include empty (and already removed) files; ok syl@

    remove empty files; ok syl@

    Add Yota LU156 to the device list.
    ok brad@

    typo fix from gavin atkinson, freebsd -r259189;

    for automated installs, use the hostname from the dhcp lease by default
    
    ok rpe@

    md5 smells bad. switch to sha256. ok deraadt matthew mikeb

    Stop the madness!  Prevent GCC from inlining these unsafe functions:
    sprintf, vsprintf, stpcpy, strcat, strcpy.  We're hitting the linker
    again, therefore the warning will show up now.

    Clean up the ioctl handler to be in line with most of the other
    Ethernet drivers.
    
    ok sasano@

    Stray memcpy() using src size.

    dead assignment;  from david hill, ok claudio

    Drop the f0, f1, f2 gettytab capabilities that were used to poke
    magic numbers into sgttyb.  The "modern" replacement for f# is the
    set of i#, o#, c#, l# to poke magic numbers into termios.
    ok miod@

    Disable the USE_OLD_TTY hack that re-encoded termios Bnnn speeds
    as sgtty Bnnn speeds.  This means ospeed has to grow from short to
    int to hold all possible values.  Bump major version.
    
    While there, also fix a bug in _nc_baudrate().
    
    ok nicm@, millert@, deraadt@

    Add infrastructure to create un-biglocked task queues.  Stolen from blambert@
    who is slacking to much.
    
    ok dlg@

    Oops. Missed file.
    
    Rework [priv_]write_file() and associated struct imsg_write_file
    to use imsg_composev() rather than allocating memory, copying data
    into it, and then passing it to imsg_compose() to copy data again
    into imsg buffer.
    
    ok pelikan@

    Rework [priv_]write_file() and associated struct imsg_write_file
    to use imsg_composev() rather than allocating memory, copying data
    into it, and then passing it to imsg_compose() to copy data again
    into imsg buffer.
    
    ok pelikan@

    Remove the printcap fc, fs, xc, xs capabilities from lpd that were used
    to configure a tty by poking magic numbers into sgtty.  If needed, the
    ms capability, which uses symbolic stty modes, can be used instead.
    Better description of ms from FreeBSD.
    ok beck@, deraadt@, jmc@ (man page)

    Unset fuse_mnt in fusefs_unmount after vflushing and freeing fbufs in use.
    
    This way, if fuse unmounts a FS without FBT_DESTROY, the fuse_mnt pointer
    is already invalidated.  Also, on weird unmount situations with vnodes in
    use, vflush() them before doing fuse_device_cleanup().
    
    Tested with ntfs-3g, ok syl@.

    Remove an attempt to free a non-heap object.
    
    from David Hill thanks.
    ok stsp@.

    Make the response lookup of a question match on entire words, where
    anything not being a whitespace is considered part of a word.
    
    deraadt@ likes it, rpe@ has approved of the idea

    Bring i915_gem_mmap_ioctl a bit closer to the Linux code.

    At resume, do not spin flushing characters in from the chip.  There shouldn't
    be any characters.  If the chip is not actually there for some reason, we'd
    be spinning so early in the resume sequence, we'd probably go mad trying to
    find the reason..
    ok kettenis
    request: people with serial ports on their laptops, try to run a "cu" over
    a suspend/resume cycle, and see if you see "input noise"

    Remove MD intagp(4) code that is unused now that inteldrm(4) manages the GTT
    all by itself.

    use new com.c bus-independent suspend/resume code, rather than trying to
    craft broken local code here
    ok kettenis

    initialize staterr, no functional change;  from david hill

    At suspend/resume time cope with timeouts, DTR on resume in a bus
    bus-independent fashion.
    ok kettenis

    sync

    Refactor processing of classless static routes option (121). Add
    more data validation, reduce magic number count, shorter parameter
    name, reduce mem*() usage, etc.

    use a separate task queue for sensors;
    with input from and ok kettenis, deraadt

    Don't keep removed files (B_INVAL bufs) in the buffer cache. Free buffers
    once brelse() is called and the buffer is B_INVAL.
    ok jsing@ krw@

    Fix crash on urndis(4) detach
    
    
    When detaching my nokia c2-01 dohooks in if_detach would panic.
    
    This was due to a failure to setup a pipe in urndis_ctrl_init()
    which results in a TIMEOUT on receive with side-effects later on
    during detach due to assumptions regarding the existence of interface
    hooks during dohooks on if_detach tear-down.
    
    That resulted on dereferencing a NULL function pointer which triggered
    a panic.
    
    
    Discussed with mpi@, tested and okay giovanni@.

    distingush between sa_msgid not set and 0; otherwise we start
    dropping messages if we usually are the initiator and the peer
    initiates rekeying first.  ok mikeb@

    Fix issues with dirent returning junk due to removal of bzero in version 1.14.
    Ensure d_name string is nul terminated and the correct length in d_namlen.
    noticed by pelikan@
    ok syl@ guenther@

    jmc has high expectations.

    replace workq_task_add with a task. stash teh struct in the ata xfer thing
    so its now reliable.

    remove unused defines

    Add Authors for the public domain ed25519/nacl code.
    see also http://nacl.cr.yp.to/features.html
            All of the NaCl software is in the public domain.
    and http://ed25519.cr.yp.to/software.html
            The Ed25519 software is in the public domain.

    forgot to say: aobj grow/shrink tweaks are mostly from Pedro Martelletto,
    like the rest of tmpfs