Remove #ifdef INET6 throughout.

(dyoung)

Do not display expired or reclaimed vestigial TIME_WAIT entries.

(dyoung)

Use fls64() only for types strictly bigger than 4 bytes.

ok christos@. Will ask for a pull-up, using 64 bits ops even for 4-bytes
int's is plain overkill.

(jym)

Reduces the resources demanded by TCP sessions in TIME_WAIT-state using
methods called Vestigial Time-Wait (VTW) and Maximum Segment Lifetime
Truncation (MSLT).

MSLT and VTW were contributed by Coyote Point Systems, Inc.

Even after a TCP session enters the TIME_WAIT state, its corresponding
socket and protocol control blocks (PCBs) stick around until the TCP
Maximum Segment Lifetime (MSL) expires.  On a host whose workload
necessarily creates and closes down many TCP sockets, the sockets & PCBs
for TCP sessions in TIME_WAIT state amount to many megabytes of dead
weight in RAM.

Maximum Segment Lifetimes Truncation (MSLT) assigns each TCP session to
a class based on the nearness of the peer.  Corresponding to each class
is an MSL, and a session uses the MSL of its class.  The classes are
loopback (local host equals remote host), local (local host and remote
host are on the same link/subnet), and remote (local host and remote
host communicate via one or more gateways).  Classes corresponding to
nearer peers have lower MSLs by default: 2 seconds for loopback, 10
seconds for local, 60 seconds for remote.  Loopback and local sessions
expire more quickly when MSLT is used.

Vestigial Time-Wait (VTW) replaces a TIME_WAIT session's PCB/socket
dead weight with a compact representation of the session, called a
"vestigial PCB".  VTW data structures are designed to be very fast and
memory-efficient: for fast insertion and lookup of vestigial PCBs,
the PCBs are stored in a hash table that is designed to minimize the
number of cacheline visits per lookup/insertion.  The memory both
for vestigial PCBs and for elements of the PCB hashtable come from
fixed-size pools, and linked data structures exploit this to conserve
memory by representing references with a narrow index/offset from the
start of a pool instead of a pointer.  When space for new vestigial PCBs
runs out, VTW makes room by discarding old vestigial PCBs, oldest first.
VTW cooperates with MSLT.

It may help to think of VTW as a "FIN cache" by analogy to the SYN
cache.

A 2.8-GHz Pentium 4 running a test workload that creates TIME_WAIT
sessions as fast as it can is approximately 17% idle when VTW is active
versus 0% idle when VTW is inactive.  It has 103 megabytes more free RAM
when VTW is active (approximately 64k vestigial PCBs are created) than
when it is inactive.

(dyoung)

*_drain() routines may be called with locks held, so instead of doing
any work in *_drain(), set a drain-needed flag.  Do the work in the
fasttimo handler.

Contributed by Coyote Point Systems, Inc.

(dyoung)

Fix PR bin/22230.

(jruoho)

Link {daily, monthly, weekly}.conf.5 to {daily, monthly, weekly}.5 and move
things around a little. Part of fixing PR bin/22230.

(jruoho)

arp_drain() may be called with locks held, so instead of doing any work
in arp_drain(), set a drain-needed flag.  Do the work in the fasttimo
handler.

Contributed by Coyote Point Systems, Inc.

(dyoung)

Add a test case for PR bin/26453.

(jruoho)

Fix same copy-paste error here. (Why atf(7) does not warn about these?)

(jruoho)

Fix build failure reported by pgoyette@.

(jruoho)

Fixes for the advlock method. It can now sustain pkgsrc/devel/locktests
with glusterfs as backend

(manu)

Call advlock method if supplied

(manu)

Fix build (libperfuse is still not built by default, but time is coming)

(manu)

Clarify the comment on how to disable password authentication (i.e. the
combination of PasswordAuthentication=no and UsePam=yes still allow password
authentication). Fixes PR bin/32313 from Curt Sampson.

(jruoho)

Fix copy-paste error.

(jruoho)

Remove \*[q] -- not necessary, just use plain double quotes instead.

(wiz)

TAB/space cosmetics.

(tsutsui)

Identify AMD Family 11h. From PR bin/41188 by FUKAUMI Naoki.

(jruoho)

Add a test for PR bin/41880.

(jruoho)

convert the pssTable mutex into a kmutex/cv.

(mrg)

A test for non-critical/low PR bin/42179.

(jruoho)

Verify that PR bin/42628 is no longer an issue.

(jruoho)

use C calling conventions via BEGIN_DECLS/END_DECLS

(agc)

Add a test for non-critical/low PR bin/43141.

(jruoho)

Some naive tests for the bswap(3) family.

(jruoho)

convert a lockless + spltty() combo to a IPL_VM mutex and cv.

hopefully this will avoid the beep-didn't-stop problem i had recently
that i was able to fix by calling wakeup() on pcppi's softc from ddb.

(mrg)

Xref bswap(3). Simplify STANDARDS.

(jruoho)

Constify, update some comments, use memset() to clear pgs.

(rmind)

Sync with head.

(jym)

Update few comments.

(rmind)

Mention POSIX semaphore improvements and PowerPC conversion to use PCU.

(rmind)

Update description of rmind-uvmplock branch.

(rmind)

PR/44928: Henning Petersen: Remove missing include dirs in Makefile.inc

(christos)

Clarify STANDARDS (i.e. cf. ualarm(3) is no longer in POSIX).

(jruoho)

Few basic tests for getitimer(2) and setitimer(2).

(jruoho)

Note also that the interaction between setitimer(2) and alarm(3) or sleep(3)
is unspecified.

(jruoho)

Remove NOTES as it is documented in itimerval(3). Remove #ifdefs from
SYNOPSIS (these are visible enough in the body of the text). Use a list for
the timers. Improve wording and markup.

(jruoho)

Fix "removing options MIPS3 cause DS5000 to loop during boot"
(i.e. no options MIPS3 kernel doesn't work on MIPS1) problem.

Reported and tested by Erik Bertelsen on port-pmax:
http://mail-index.NetBSD.org/port-pmax/2011/04/21/msg000099.html

(tsutsui)

Add support for new SUNIX PCI seriel and/or parallel adapters, with new
SUNIX vendor Id 0x1fd4 and product Id 0x1999. Constants were taken from
Linux open source driver distributed by SUNIX.

The following models should be supported now:
5008 1 parallel port
5016 16 serial ports, only 8 supported right now (PUC_MAX_PORTS limit)
5027 1 serial port
5037 2 serial ports
5056 4 serial ports
5066 8 serial ports, this model is the only one that was actually tested.
5069 1 serial port and 1 parallel port
5079 2 serial ports and 1 parallel port
5099 4 serial ports and 1 parallel port

(manu)

Don't conflict SUNIX new vendor Id with the previous one

(manu)

Add new PCI vendor Id for Sunix, and the corresponding device Id.

(manu)

Add back usr/libdata/debug/usr/tests/lib/libc/termios, needed for
debug builds.

(njoly)

Don't print a time_t with %lu

(gson)

Fix BCM5709 PHY detection for ethernet PHYs (the SerDes case being already
handled):
- export bge(4) and bnx(4) CHIP ID and PHY flags to brgphy(4). Move to
"unsigned int" rather than "int", and reuse the same softc members for
chipid and phyflags (behavior controlled by the sc_isbge/isbnx boolean).
- apply bug fix for revisions A and B, so that autonegotiation can
complete (from OpenBSD).

Bug reported by Rivo Nurges via private mail, patch tested and
confirmed working by him (with thanks!)

(jym)

Also add the Atffile to the sets list

(pgoyette)

Use the standard "PR foo/1234" notation such that the PR number is parsed
correctly for the atf(7) HTML reports.

(jruoho)

Make it self-contained (add missing stdbool.h include when used for libkvm)

(martin)

Get this right.  It was not the debugdata directory that was missing...

(pgoyette)

convert rb_mutex to a kmutex/cv.

(mrg)

Verify that nanosleep(2) can suspend for one nanosecond (or longer).

(jruoho)

Initial import of the in-kernel iSCSI initiator, kindly contributed by
Wasabi Systems under a 2-clause BSD license, to the othersrc tree.

The initiator itself can be loaded as an lkm, and works successfully
on 2.0 (the "burner" host), running against the NetBSD target on a 5.99 host.


burner# modload -v -s -p /usr/lkm/iscsi_post.sh /usr/lkm/iscsidrv.o
modload: reserving 36864 bytes of memory
Module loaded as ID 0
burner# iscsid
iSCSI Daemon loaded
burner# iscsictl add_send_target -a 172.16.135.133
Added Send Target 1
burner# iscsictl refresh_targets
OK
burner# iscsictl list_targets
    1: iqn.1994-04.org.netbsd.iscsi-target:target0
2: 172.16.135.133:3260,1
burner# iscsictl login -P 2
Created Session 2, Connection 1
burner# iscsictl list_sessions
Session 2: Target iqn.1994-04.org.netbsd.iscsi-target:target0
burner# newfs /dev/rsd0a
/dev/rsd0a: 100.0MB (204800 sectors) block size 8192, fragment size 1024
using 4 cylinder groups of 25.00MB, 3200 blks, 6144 inodes.
super-block backups (for fsck -b #) at:
    32,  51232, 102432, 153632,
burner# mount /dev/sd0a /mnt
burner# df
Filesystem  1K-blocks    Used    Avail Capacity  Mounted on
/dev/wd0a    4066094  186994  3675795    4%    /
kernfs              1        1        0  100%    /kern
/dev/sd0a      99247        1    94283    0%    /mnt
burner# dmesg | egrep '(scsibus|sd0)'
scsibus0 at bha2: 16 targets, 8 luns per target
scsibus0: waiting 2 seconds for devices to settle...
scsibus1 at iscsi0: 1 target, 16 luns per target
sd0 at scsibus1 target 0 lun 0: <NetBSD, NetBSD iSCSI, 0> disk fixed
sd0: fabricating a geometry
sd0: 100 MB, 100 cyl, 64 head, 32 sec, 512 bytes/sect x 204800 sectors
sd0: fabricating a geometry
sd0: fabricating a geometry
sd0: fabricating a geometry
burner#

on the target end:

Reading configuration from `/etc/iscsi/targets'
target0:rw:0.0.0.0/0
extent0:/tmp/iscsi-target0:0:104857600
DISK: 1 logical unit (204800 blocks, 512 bytes/block), type iscsi fs
DISK: LUN 0: 100 MB disk storage for "target0"
TARGET: iSCSI Qualified Name (IQN) is iqn.1994-04.org.netbsd.iscsi-target
> iSCSI Discovery login  successful from iqn.1994-04.org.netbsd:iscsi.burner.cupertino.alistaircrooks.com:0 on 172.16.135.137 disk -1, ISID 70368764559360, TSIH 1
< iSCSI Discovery logout successful from iqn.1994-04.org.netbsd:iscsi.burner.cupertino.alistaircrooks.com:0 on 172.16.135.137 disk -1, ISID 70368764559360, TSIH 1
> iSCSI Normal login  successful from iqn.1994-04.org.netbsd:iscsi.burner.cupertino.alistaircrooks.com:0 on 172.16.135.137 disk 0, ISID 70368764559360, TSIH 2


Status:

Vendor Tag: WASABI
Release Tags: iscsi-initiator-base

(agc)

Make sure to enable the vector unit in the PSL when returning to userland.

(matt)

Fix broken stack.  The r1 is stack pointer.  Not cpu_info pointer.

(kiyohara)

Don't call pcu_do_op in pcu_load to save current context, use pcu_cpu_op
instead (since it will deal with ci->ci_pcu_curlwp[id] being NULL).

(matt)

Fix cut/paste error.

(matt)

Add tests/lib/libc/termios subdirectory to accomodate jruoho's new test
t_tcsetpgrp

(pgoyette)

Limit sys/pcu.h inclusion within _KERNEL scope, plus _KMEMUSER.

(rmind)

Move powerpc to use pcu to manage FPU/AltiVec/SPE.

(matt)

A little PMAP_OEA64 pmap support.

(matt)

Add nandfcm device glue

(matt)

Add LBC FCM defintions (for NAND).

(matt)

mpc85xx uses MBRs

(matt)

A little more constification

(matt)

use a unique wchan identifier for node_queue_cv than the old default.

(mrg)

re-apply the reverted previous, after fixing the missed wakeup:

convert eq_mutex to a kmutex/cv.
convert ltsleep() used for plain timeout into kpause().

(mrg)

rever the previous for now; it hangs reconstruction.

(mrg)

Extend PCU:
- Add pcu_ops_t::pcu_state_release() operation for PCU_RELEASE case.
- Add pcu_switchpoint() to perform release operation on context switch.
- Sprinkle const, misc.  Also, sync MIPS with changes.

Per discussions with matt@.

(rmind)

Add an IPI for xcalls.

(matt)

- Rather use pmap_pte_set() when modifying a PTE.

- Call pmap_pte_flush() when issuing PTE modifications. While it is a NOOP
for native x86, it is not for Xen. It will flush all operations that are
possibly waiting in the queue, like xpq_queue_pte_update(). Do it for
each level, this function is only called at boot time and is not
performance critical.

While here:
- No need to cast early_zerop to void with memset().
- Move common variables out of the #ifdef's.
- KNF

(jym)

Test also ttyname_r(3).

(jruoho)

Verify PR standards/44921.

(jruoho)

Naive tests for clearerr(3).

Converted from the "PCCA" test suite by Stathis Kamperis.

(jruoho)

Reset ZS chip properly in zscninit() so that zscngetc() works
even before zsc is attached during configure.
(i.e. make userconf(4) work on serial console)

(tsutsui)

fix some bugs and simplify locking
from rmind

(ahoka)

Remove trailing tabs, add RCS IDs.

(rmind)

Add the new 'termios' subdirectory.

(jruoho)

Add a test for PR lib/41673 (another trivial errno "bug").

(jruoho)

convert eq_mutex to a kmutex.
convert ltsleep() used for plain timeout into kpause().

(mrg)

Improve further so that this might actually fail.

(jruoho)

Actually, as this is nanosleep(2), test 1-10 nano second sleep times.

(jruoho)

Simplify the previous. (Apparently the sleep time is not that accurate.)

(jruoho)

Few basic tests for nanosleep(2).

(jruoho)

Two elementary tests for gettimeofday(2).

(jruoho)

convert rf_debug_mem_mutex to a kmutex, and fix RF_DEBUG_MEM option.

(mrg)

PR lib/42587 was fixed; remove xfail.

(jruoho)

convert rf_tracing_mutex to a kmutex.

(mrg)

convert the (unused) rf_sparet_wait_mutex to a real mutex/cv.

(mrg)

Start eliminating "super-user" from the documentation.

(christos)

More lim_free() fallout

(pgoyette)

nice should always return EPERM, not EACCES

(christos)

if donice fails, don't keep going with the next process.

(christos)

Catch up with lim_privatise() change (missed in previous commit).

(rmind)

- Remove FORK_SHARELIMIT and PL_SHAREMOD, simplify lim_privatise().
- Use kmem(9) for struct plimit::pl_corename.

(rmind)

convert mcpair to kmutex.

(mrg)

Merge duplicate code fragments into a new lim_setcorename() routine.

(rmind)

Rename limfree() to lim_free(), misc clean up.  No functional change.

(rmind)

sysctl_proc_corename: improve comments, clean up, move a check for
KAUTH_REQ_PROCESS_CORENAME_SET earlier, do not bother to strcmp().

(rmind)

Save a few global registers during set/longjmp - at least the application
registers and the only currently used system register (%g7).
Sparc now passes the setjmp tests (and should be able to build perl again).

(martin)

Fix previous: the way we calculate, we need to test the size of sigjmp_buf

(martin)

Add another compile time assert: catch overflows of jmp_buf

(martin)

Save/restore some global registers (sigsetjmp compatible). Does not realy
make a difference currently, but is a hard to spot blackhole if something
ever starts using the application registers.

(martin)

Rearange jmp_buf content again: storing the sigsetjmp() save_mask midway
inside the environment is not a good idea, we may overwrite it when
fetching other details of the sigjmp environment.
Instead move it to the very end again.
Do not bother to save volatile (caller saved) global registers.
Fixes PR port-sparc64/44918.

(martin)

Small typo fix.

(jym)

correct comment

(plunky)

add a define for pcb_sp

(christos)

Share the amd64 and i386 files.
XXX: Is there a better way to get the cpu family?

(christos)

ufs_mknod: change vnode type to VNON before it gets unlocked.  Closes a small
    window where the vnode could have type VCHR but op vector ffs_vnodeop_p.

(hannken)

Save/restore errno in error path in dupgood. Check that n>0 before
accessing banner[n-1].

Approved a while ago by pooka@

(alnsn)

Add new files.

(alnsn)

Add Iomega Storcenter dsrtc device (DS1337).

(phx)

Detect QNAP board last, because a Realtek chip at pci device 15 is used
on several boards.

(phx)

Few fundamental consistency checks for kill(2).

(jruoho)

- convert rf_printf_mutex to a kmutex
- convert rf_rad_lock and the per-raid "cv" to per-raid kmutex/and real cv
- use rf_mutex_init() in places, and move it with the similar definitions

(mrg)

PR/42320 doesn't seem to be fixed, mark the failure as expected.

(alnsn)

Add files required for the new testcase in t_awk.sh.

(alnsn)

Add simplistic atabus(4) rescan support.

(jakllsch)

Move the connection code of xbdback(4) and xvif(4) backends in separate
functions. The frontend watch function is easier to read, and mixing
switch() with goto's error paths is rather error-prone.

While here, sprinkle some aprint_*.

Tested under amd64 dom0 with i386 PAE and amd64 domUs.

(jym)

Small comment improvement.

(rmind)

Silence xenbus_read_target() in ENOENT case (== entry is missing from
Xenstore). The error case does not bring much here; assume that the value
is 0.

Print the error code when writing the ``target'' value fails.

(jym)

Apply DRY: xpmap_{mtop,ptom}() can reuse xpmap_{mtop,ptom}_masked() for
the frame number lookup.

No functional change.

(jym)

When we stretch a short frame to 60 bytes in rge_send() do not return 60,
but the original frame size.

(phx)

simplify.  Don't limit data way size to a page.

(matt)

whitespace cleanup.

(matt)

minor cleanup (remove redundant static, fix whitespace).

(matt)

Remove redundant opt_multiprocessor.h include

(matt)

Add/move some COP0_SYNC.  increment tlb miss counter on mips3 and mips64r2
#ifdef MIPS3 / #endif more nops

(matt)

Add whitespace

(matt)

KNF cleanup.

(matt)

whitespace cleanup

(matt)

fix mips32/mips64 program on mips32/mips64 r2 bug.

(matt)

constification.

(matt)

remove redundant opt_multiprocessor.h include

(matt)

constification

(matt)

Use sizeof(*pte) instead sizeof(type)

(matt)

cleanup cpu.h includes

(matt)

Deal with uarea mapped in either kseg0 or xkphys.

(matt)

whitespace cleanup

(matt)

add cop2 instructions

(matt)

ras atomicvec is no more.

(matt)

Since the RAS lock stubs are the default, don't bother assigning them just
make the default.

(matt)

define<space> -> define<tab>

(matt)

minor cleanups.

(matt)

Use M_ZERO with malloc instead of doing a explicit memset

(matt)

minor constification, format cleanups

(matt)

Use mips64 for m[tf]c2, not arch=xlr

(matt)

cpu_number() return an u_int, use %u for format

(matt)

Provide OHCI companions devices to EHCI.

(matt)

Add some P10xx SVRs

(matt)

Use the wrtee inlines for booke.

(matt)

Mention PowerPC BookE support and addition of PCU.

(rmind)

Remove (off by default) option to build Postfix with Cyrus SASL support.
As "libsasl" integration can now be considered stable it is no longer
necessary to suppor this hack.

(tron)

Add test cases for PR bin/6764: 'trap ... 0' is supposed to execute the
command when the shell exits but it does not work in ksh when the shells
exits "implicitly" (without an explicit exit/return statement).

These new tests cover both sh and ksh.  The ksh part of this does not
strictly belong to this directory, but I think it'd be nice to extend
all the tests in here to cover both interpreters whenever that makes
sense, much like we do with the file system tests.

(jmmv)

New sentence, new line. Use more markup.

(wiz)

Re-add -q description that got lost when -x was added.
Sort descriptions.
Bump date.

(wiz)

Sort sections, SEE ALSO, and fix xref.

(wiz)

missing files

(mbalmer)