Skip to content
Commits on Feb 7, 2016
  1. <cpu/limits.h>: Adjust visibility of various *_{MIN,MAX} constants.

    Sascha Wildner committed
    This commit just affects POSIX environments (to reduce namespace
    pollution), not our default one where all of this is visible. No
    dports breakage was observed in my tests, as is expected because
    this change is...
    Taken-from: FreeBSD
  2. <cpu/limits.h>: Minor whitespace fixes.

    Sascha Wildner committed
  3. <grp.h>: Include <machine/stdint.h> instead of the whole of <sys/type…

    Sascha Wildner committed
    It is only needed for __uint32_t.
    After this commit, namespace pollution for <grp.h> is considerably lower.
  4. <grp.h>: Minor whitespace fixes.

    Sascha Wildner committed
  5. tools/tcp_stream: Kill netperf children proactively upon SIGINT

    Sepherosa Ziehau committed
    TCP_MAERTS test traps the first SIGINT, so use SIGKILL to terminate
    them, if we are about to die.
  6. tools/tcp_stream: Add dual side test support

    Sepherosa Ziehau committed
  7. drm/linux: Sync linux/math64.h with FreeBSD

    François Tigeot committed
    Removing 32-bit inline functions.
  8. tools/tcp_stream: Allow performing test to multiple hosts

    Sepherosa Ziehau committed
  9. @ivadasz

    Remove FreeBSD OS/ABI from raid driver blobs.

    ivadasz committed
    This fixes kgdb issues when the kernel was linked with
    The problem was caused by's behaviour of setting the OS/ABI in
    the output binary to FreeBSD when any of the input objects is tagged
    with the FreeBSD OS/ABI.
    The hpt27xx, hptmv and hptrr raid-drivers are using binary blobs, with
    the FreeBSD OS/ABI flag, so we should just set the OS/ABI of these
    objects to none after uudecode-ing them.
  10. tools/tcp_stream: Move sockbuf string to the main stack

    Sepherosa Ziehau committed
Commits on Feb 6, 2016
  1. linux/kref.h: Add kref_put_mutex()

    François Tigeot committed
  2. @markuspf

    etc/rc.d: Update rtsold

    markuspf committed
    * Remove precmd, it did not work because autocmd does not exist
    * Add postcmd, taken from FreeBSD head
Commits on Feb 5, 2016
  1. <libgen.h>: Clean up whitespace and some #if 0'd code.

    Sascha Wildner committed
  2. Don't remove manpages which we have to keep.

    Sascha Wildner committed
    They were not part of the library privatization (which can only
    affect section 3 manual pages). Those tools are consumers of
    privatized libs but of course they must keep their documentation.
  3. Expand all wildcards from the lib privatization.

    Sascha Wildner committed
    We want to see exactly what we remove, not just approximately.
    'make upgrade' changes are a pain to review when everything is
    muddled with patterns. And it gets worse over time because
    without a not yet upgraded system one has no choice but to
    restore a world from the time and replaying the 'make upgrade'
    to see what was going on.
    Also, if we ever need to bring back one file which is covered
    by a pattern later on (and we've had such cases), no one will
    know what the rest of the wildcard was covering aside from
    that file.
  4. Add many missing '.gz'.

    Sascha Wildner committed
Commits on Feb 3, 2016
  1. <time.h>: CLOCKS_PER_SEC must be of type clock_t.

    Sascha Wildner committed
    Note that its value violates the standards requirement that it be
    1000000. See the STANDARDS section of clock(3)'s manual page for
    further information.
  2. <time.h>: struct timespec's definition has to be visible for C11.

    Sascha Wildner committed
    See B.26 in the C11 standard.
    Note that this is not in the N1548 committee draft.
Commits on Feb 2, 2016
  1. <time.h>: C11 actually specifies that TIME_UTC be "greater than 0".

    Sascha Wildner committed
    In-discussion-with: xenu
Commits on Feb 1, 2016
  1. make upgrade: Add an obsolete time zone.

    Sascha Wildner committed
  2. @jrmarino

    bump __DragonFly_version for library privatizaton

    jrmarino committed
    Better late than never.   Version is now 400502
Commits on Jan 30, 2016
  1. Sync zoneinfo database with tzdata2016a from…

    Sascha Wildner committed
    * America/Cayman will not observe daylight saving this year after all.
        Revert our guess that it would.  (Thanks to Matt Johnson.)
    * Asia/Chita switches from +0800 to +0900 on 2016-03-27 at 02:00.
        (Thanks to Alexander Krivenyshev.)
    * Asia/Tehran now has DST predictions for the year 2038 and later,
        to be March 21 00:00 to September 21 00:00.  This is likely better
        than predicting no DST, albeit off by a day every now and then.
    * America/Metlakatla switched from PST all year to AKST/AKDT on
        2015-11-01 at 02:00.  (Thanks to Steffen Thorsen.)
    * America/Santa_Isabel has been removed, and replaced with a
        backward compatibility link to America/Tijuana.  Its contents were
        apparently based on a misreading of Mexican legislation.
    * Asia/Karachi's two transition times in 2002 were off by a minute.
        (Thanks to Matt Johnson.)
    * The Theory and asia files now mention the 2015 book "The Global
        Transformation of Time, 1870-1950", and cite a couple of reviews.
    * The America/Chicago entry now documents the informal use of US
        central time in Fort Pierre, South Dakota.  (Thanks to Rick
        McDermid, Matt Johnson, and Steve Jones.)
  2. @xenu

    libc - implement C11 function timespec_get()

    xenu committed with Sascha Wildner
  3. Revert "boot/acpi: Implement hint.acpi.0.rsdp kenv."

    Sascha Wildner committed
    This reverts commit 0a259e6.
Commits on Jan 29, 2016
  1. boot/acpi: Use an ACPICA constant instead of a magic number.

    Sascha Wildner committed
    While here, reword a comment to not refer to any specific version of
    the ACPI specification.
  2. boot/acpi: Implement hint.acpi.0.rsdp kenv.

    Sascha Wildner committed
    ACPICA can make use of it.
  3. @zrj-rimwis

    kernel/isofs: Add all iso9660 number conversion routines.

    zrj-rimwis committed with zrj
    Mainly to remove 7.2.2 and 7.3.2 routines from lib/libstand.
    While there, fix return type for 7.1.2 to be signed as per ISO9660.
    This commit breaks sysutils/hal dport again,
    it should have been patched in the first place.
    Taken-from: FreeBSD
  4. @zrj-rimwis

    kernel/isofs: Remove UNALIGNED_ACCESS

    zrj-rimwis committed with zrj
    Don't make assumptions against clever compilers.
Something went wrong with that request. Please try again.