Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Branch: user/lstewart/…
Commits on Nov 20, 2008
  1. Fix a bug in kthread_exit() which led to potential deadlock in code that

    lstewart authored
    was relying on the documented behaviour in the kthread(9) man page. This brings
    the wakeup() related behaviour of kthread_exit in 7.x inline with 8.x.
    I've tested the fix with a simple scenario that was previously deadlocking. A
    waiting thread waits on a terminating thread by calling mtx_sleep() with the
    "chan" argument set to the *thread obtained by calling
    Without this patch, the waiting thread never gets woken because kthread_exit()
    calls wakeup() on the *proc. With this patch, the wait thread gets woken and
    all is good in the world again.
    For more details, refer to:
    Discussed with:	attilio@, julian@
  2. @delphij

    MFC: Document igb(4).

    delphij authored
    Approved by:	re (hrs)
Commits on Nov 19, 2008
  1. MFC: Add workaround for a back reference when no corresponding

    hrs authored
    parenthesized subexpression is defined.  For example, the
    following command line caused unexpected behavior like
    segmentation fault:
     % echo test | sed -e 's/test/\1/'
    Approved by:	re (kib)
    PR:		bin/126682
  2. @delphij

    MFC revisions 183653 183654 183714 (simon),

    delphij authored
    all recent em(4) and add igb(4).
    Approved by:	re (hrs)
Commits on Nov 18, 2008
  1. @delphij

    MFC r183718 and r185057:

    delphij authored
    Add some examples to demostrate gpart(8).
    Approved by:	re (kib)
  2. MFC 184934:

    ambrisko authored
    For now on every 10 cyclinder groups flush the buffer cache to free
    up space.  If the buffer cache fills up then the disk systems can
    grind to a halt.
    PR:		128832
    Approved by:	re (kensmith)
Commits on Nov 17, 2008
  1. MFC 182085, 184897, 184933, 184974, 184975

    ambrisko authored
    This addresses:
         1) New HW support from LSI (they want to get the HW support
            into a FreeBSD release before the HW comes out!).
         2) Bug fix in the ioctl path with zero length command.
         3) Adds functional support so that SCSI sense will work
            when requested from user-land.  This is a little strange,
            since they store a user-land pointer in the "data" space
            of the ioctl packet.  So copy the data out into another
            structure that is a union of the data and pointer (mfi_sense_ptr
            union).  This way we can cleanly use the pointer, as pointer on
            both 64bit and 32bit machines (i386/amd64).
    Submitted by:	LSI
    Approved by:	re (kib)
  2. @kostikbel

    MFC r184587:

    kostikbel authored
    Add the ffsll and flsll functions. These are ffs and fls operating
    on long long arguments.
    Approved by:	re (kensmith)
Commits on Nov 16, 2008
  1. MFC: r182876

    marius authored
    Adapt the locking of esp(4) to MPSAFE cam(4) (so that ncr53c9x_intr()
    actually is INTR_MPSAFE now) and fix a couple of bugs which lead to
    panics, amongst other improvements.
    Approved by:	re (kib)
  2. MFC: r180692

    marius authored
    o Fix style(9) bugs and similar nits.
    o Merge ncr53c9x.c from NetBSD:
      1.115: fix variable shadowing
      1.118: __inline -> inline
      1.121: fix empty if
    Approved by:	re (kib)
Commits on Nov 14, 2008
  1. Merge r183539 and r183558: Do match on class to avoid false positives,

    imp authored
    return BUS_GENERIC_PROBE for a generic match.
    Approved by:	re@
  2. MFC the BIND 9.4.2-P2 update

    dougb authored
    Approved by:	re (kib)
Commits on Nov 13, 2008
  1. @bsdjhb

    MFC: Adjust the license statement to more closely match a standard

    bsdjhb authored
    3-clause BSD license.
    Approved by:	re (kib)
  2. @DarkHelmet433

    MFC: 184547,184548: add END() macros for Assembler code in libc. This

    DarkHelmet433 authored
    causes the symbol table to include the size of the function.
    Approved by:	re (kib)
  3. @rpaulo

    MFC r176428:

    rpaulo authored
    	Add AS lookup functionality. On each hop we query a whois server to
    	find the corresponding AS for that IP (-a switch).
    	We can also choose a different whois server with the -A switch. The
    	default is
    	Obtained from:	       NetBSD
    Approved by:	       re
  4. @kostikbel

    MFC r184501:

    kostikbel authored
    The code in linux_proc_exit() contains a race when multiple linux based
    processes exits at the same time. The linux_emuldata structure is freed
    but p->p_emuldata is left as a dangling pointer to the just freed
    The check for W_EXIT in the loop scanning the child processes isn't safe
    since the state of the child process can change right afterwards. Lock
    the process and check the W_EXIT before delivering signal.
    Approved by:	re (kensmith)
  5. MFC r184253:

    yongari authored
      Use auto-negotiation for manual media type selection. This fixes
      establishment of 10/100Mbps link on Atheros AR8121(L1E).
    Approved by:	re (kib)
  6. @DarkHelmet433

    MFC 184828,184829: Fix si_addr in 32 bit signals.

    DarkHelmet433 authored
    Approved by:	re (kib)
Commits on Nov 12, 2008
  1. (MFC r184484)

    joerg authored
    When running a "chio return" operation using a physical source unit
    rather than a voltag name, do not set the CESR_VOLTAGS flags in the
    CHIOGSTATUS command requesting the current status.  As voltags are an
    optional feature that must be handled as "reserved" by media changers
    not implementing the feature, always setting CESR_VOLTAGS resulted in
    the command being aborted with an `Invalid field in CDB', and
    consequently the "chio return" failed, for media changers that do not
    support voltags.
    Approved by:	re (kib)
  2. @delphij

    MFC revision 1801453: Don't leak DMA map if not freed.

    delphij authored
    Submitted by:	kevlo
    Approved by:	re (kensmith)
Commits on Nov 11, 2008
  1. @rwatson

    Merge r184304 from head to stable/7:

    rwatson authored
    In both dropwithreset paths in tcp_input.c, drop the tcbinfo lock
    sooner to decomplicate locking and eliminate the need for a rather
    chatty comment about why we have to handle the global lock in a
    special way for the benefit of ipfw and pf cred rules.
    Approved by:	re (kib)
Commits on Nov 10, 2008
  1. @delphij

    Merge the following bce(4) changes:

    delphij authored
    r176448 (davidch)
     - Added loose RX MTU functionality to allow frames larger than 1500
       bytes to be accepted even though the interface MTU is set to 1500.
     - Implemented new TCP header splitting/jumbo frame support which uses
       two chains for receive traffic rather than the original single
       receive chain.
     - Added additional debug support code.
    r178132 (davidch)
     - Fixed a problem with the send chain consumer index which would cause
       TX traffic to sit in the send chain until a received packet kick
       started the interrupt handler.  This would cause extremely slow
       performance when used with NFS over UDP.
     - Removed untested polling code.
     - Updated copyright year in the file header.
     - Removed inadvertent ^M's created by DOS text editor.
    r178853 (scottl)
    The BCE chips appear to have an undocumented requirement that RX frames
    be aligned on an 8 byte boundary.  Prior to rev 1.36 (now r176448) this
    wasn't a problem because mbuf clusters tend be naturally aligned.  The
    switch to using split buffers with the first buffer being the embedded
    data area of the mbuf has broken this assumption, at least on i386,
    causing a complete failure of RX functionality.  Fix this for now by
    using a full cluster for the first RX buffer.  A more sophisticated
    approach could be done with the old buffer scheme to realign the m_data
    pointer with m_adj(), but I'm also not clear on performance benefits of
    this old scheme or the performance implications of adding an m_adj()
    call to every allocation.
    r179436 (jhb)
    Trim an extra semi-colon.
    r179695 (davidch)
     - Fixed kern/123696 by increasing firmware timeout value from 100 to
     - Fixed a problem on i386 architecture when using split header/jumbo
       frame firmware caused by hardware alignment requirements.
     - Added #define BCE_USE_SPLIT_HEADER to allow the feature to be
       enabled/disabled.  Enabled by default.
    PR:             kern/123696
    r179771 (davidch)
     - Added support for BCM5709 and BCM5716 controllers.
    r182293 (davidch)
     - Updated support for 5716.
     - Added some additional code for debug builds.
     - Fixed a problem printing physical memory on 64bit system during
     - Modified some of the context memory and mailbox register names to
       more clearly distinguish their use.
     - Added memory barriers for Intel CPUs when accessing host memory data
       structures which are written by hardware.
    Approved by:	re (kib)
  2. @rwatson

    Merge r184298 from head to stable/7:

    rwatson authored
      Remove endearing but syntactically unnecessary "return;" statements
      directly before the final closeing brackets of some TCP functions.
    Approved by:	re (gnn)
  3. @jasone


    jasone authored
    	Revert to preferring mmap(2) over sbrk(2) when mapping memory, due
    	to potential extreme contention in the kernel for multi-threaded
    	applications on SMP systems.
    Approved by:	re (kib)
  4. MFC:

    davidxu authored
    	Add POSIX clock id CLOCK_THREAD_CPUTIME_ID support.
    Approved by:	re (kib)
Commits on Nov 7, 2008
  1. MFC: r184753

    blackend authored
    Update ports number and size of Ports Collection.
    Approved by:  re (kib)
  2. MFC: r183014

    bz authored
         (which was missed with r184739: r183013,r183017,r183032,r183034)
      Original commit message from HEAD:
    	Julian: "oops commit the version that compiles"
      Unbreak RELENG_7 including the missing sys/proc.h header file.
    Approved by:	re (kib)
  3. @kostikbel

    This is direct commit to stable/7 to accomodate KPI difference between

    kostikbel authored
    HEAD and the branch. Provide struct thread * argument to VOP_GETATTR.
    Approved by:	re (rwatson)
Commits on Nov 6, 2008
  1. @DarkHelmet433

    MFC: r184492 - add offset / fileid / fsid to KERN_PROC_VMMAP sysctl

    DarkHelmet433 authored
    for the benefit of valgrind.
    Approved by:	re (kib)
  2. MFC a rewrite of rt_check(). also revert the addition of

    julian authored
    rt_check_fib() which we discovered is not needed.
    fixes some hangs people have seen
    Approved by:	re (ken)
  3. Oops. Correct a mis-merge. This piece of r183860 didn't make it

    kensmith authored
    through while doing the merge.
    PR:		misc/128610
    Submitted by:	veron <dot> x503 <at> hotmail <dot> com
    Approved by:	re (blackend, kib)
Commits on Nov 5, 2008
  1. @dgerzo

    MFC rev. 183041:

    dgerzo authored
    - remove superfluous word
    PR:		docs/127401
    Submitted by:	Mick Charles Beaver <>
    Approved by:	re@ (blackend)
  2. @kostikbel

    MFC r182698 (by kan):

    kostikbel authored
    Make sure internal rtld malloc routines are not called from unlocked
    contexts as rtld's malloc is not thread safe and is only supposed to be
    called with exclusive bind lock already held.
    Approved by:	re (kensmith), kan
  3. @kostikbel

    MFC r183737 (by kan):

    kostikbel authored
    Allow strong symbols to override weak ones for lookups done through
    dlsym with RTLD_NEXT/RTLD_SELF handles.
    Allow symbols from to be located this way too.
    Approved by:	re (kensmith), kan
Something went wrong with that request. Please try again.