]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
11 years agoUse a higher TTL (128) for DHCP packets. This matches the ISC DHCP client.
John Baldwin [Mon, 22 Apr 2013 15:02:32 +0000 (15:02 +0000)]
Use a higher TTL (128) for DHCP packets.  This matches the ISC DHCP client.

PR: bin/170279
MFC after: 1 week

11 years agoBuild uart_dev_lpc.c on arm only. This fixes pc98 build.
Yoshihiro Takahashi [Mon, 22 Apr 2013 13:02:41 +0000 (13:02 +0000)]
Build uart_dev_lpc.c on arm only.  This fixes pc98 build.

11 years agoKeep up with negative addrlen check removal in r249649.
Sergey Kandaurov [Mon, 22 Apr 2013 09:18:50 +0000 (09:18 +0000)]
Keep up with negative addrlen check removal in r249649.

11 years agoPanic if UMA_ZONE_PCPU is created at early stages of boot, when mp_ncpus
Gleb Smirnoff [Mon, 22 Apr 2013 09:02:23 +0000 (09:02 +0000)]
Panic if UMA_ZONE_PCPU is created at early stages of boot, when mp_ncpus
isn't yet initialized. Otherwise we will panic at first allocation later.

Sponsored by: Nginx, Inc.

11 years agoInitialize GIC_PMRR register on ARM GIC.
Damjan Marion [Mon, 22 Apr 2013 08:28:53 +0000 (08:28 +0000)]
Initialize GIC_PMRR register on ARM GIC.

Provided by: Thomas Skibo

11 years agoClarify to the user the amount of time required to load the menus for
Devin Teske [Mon, 22 Apr 2013 06:30:29 +0000 (06:30 +0000)]
Clarify to the user the amount of time required to load the menus for
each view in the details dialog.

11 years agoNew helper functions for common widgets.
Devin Teske [Mon, 22 Apr 2013 06:28:45 +0000 (06:28 +0000)]
New helper functions for common widgets.

11 years agoProper fix for copy/paste error (first attempt r249756).
Devin Teske [Mon, 22 Apr 2013 06:27:14 +0000 (06:27 +0000)]
Proper fix for copy/paste error (first attempt r249756).

11 years agoFix a copy/paste error.
Devin Teske [Mon, 22 Apr 2013 06:20:34 +0000 (06:20 +0000)]
Fix a copy/paste error.

11 years agoUpdate comment for accuracy.
Devin Teske [Mon, 22 Apr 2013 06:13:30 +0000 (06:13 +0000)]
Update comment for accuracy.

11 years agoStyle nit (to be consistent across project).
Devin Teske [Mon, 22 Apr 2013 06:12:22 +0000 (06:12 +0000)]
Style nit (to be consistent across project).

11 years agoFix two bugs when setting up a plip interface. First, dhcp status was not
Devin Teske [Mon, 22 Apr 2013 06:10:27 +0000 (06:10 +0000)]
Fix two bugs when setting up a plip interface. First, dhcp status was not
being properly cleared each iteration of the loop. Second, values weren't
properly quoted when sent to the editor function.

11 years agoConvert over the etherswitch framework to use VLAN IDs per port, rather
Adrian Chadd [Mon, 22 Apr 2013 05:52:18 +0000 (05:52 +0000)]
Convert over the etherswitch framework to use VLAN IDs per port, rather
than VLAN groups.

Some chips (eg this rtl8366rb) has a VLAN group per port - you first
define a set of VLANs in a vlan group, then you assign a VLAN group
to a port.

Other chips (eg the AR8xxx switch chips) have a VLAN ID array per
port - there's no group per se, just a list of vlans that can be
configured.

So for now, the switch API will use the latter and rely on drivers
doing the heavy lifting if one wishes to use the VLAN group method.
Maybe later on both can be supported.

PR: kern/177878
PR: kern/177873
Submitted by: Luiz Otavio O Souza <loos.br@gmail.com>
Reviewed by: ray

11 years agoUI improvements. First, implement --default-item whenever and wherever
Devin Teske [Mon, 22 Apr 2013 05:52:06 +0000 (05:52 +0000)]
UI improvements. First, implement --default-item whenever and wherever
possible to save keystrokes. Second, overhaul startup/rcdelete for much
improved performance. Last, but not least, kill-off useage of --clear
and implement --keep-tite in harmony to minimize jarring transitions.
Also, fix local variable names where necessary while we're here with
other minor comment-enhancements/typo-corrections.

11 years agoUse new f_getvar for clarity and general code follow-up to r249746.
Devin Teske [Mon, 22 Apr 2013 05:32:27 +0000 (05:32 +0000)]
Use new f_getvar for clarity and general code follow-up to r249746.

11 years agoetherswitchcfg(8) crashes when you don't set vlangroup members.
Adrian Chadd [Mon, 22 Apr 2013 05:30:20 +0000 (05:30 +0000)]
etherswitchcfg(8) crashes when you don't set vlangroup members.
Fix this to require an argument.

PR: kern/177872
Submitted by: Luiz Otavio O Souza <loos.br@gmail.com>

11 years agoInitialise this to zero before using it to configure the vlangroup
Adrian Chadd [Mon, 22 Apr 2013 05:27:13 +0000 (05:27 +0000)]
Initialise this to zero before using it to configure the vlangroup
information.

PR: kern/177871
Submitted by: Luiz Otavio O Souza <loos.br@gmail.com>

11 years agoAdd new flags `-d' (sets debug=1) and `-D file' (sets debugFile) and
Devin Teske [Mon, 22 Apr 2013 05:02:34 +0000 (05:02 +0000)]
Add new flags `-d' (sets debug=1) and `-D file' (sets debugFile) and
improve debugging initialization. Also fixup USAGE statements while we're
here. Also, change initialization of main program to _not_ change working
directory, allowing the debugFile to be relative without confusion.

11 years agoSimplify vm_radix_{add,dec}lev().
Alan Cox [Mon, 22 Apr 2013 01:26:13 +0000 (01:26 +0000)]
Simplify vm_radix_{add,dec}lev().

Sponsored by: EMC / Isilon Storage Division

11 years agoAdd missing static keyword, to make the code build with WARNS=6.
Ed Schouten [Sun, 21 Apr 2013 22:47:24 +0000 (22:47 +0000)]
Add missing static keyword, to make the code build with WARNS=6.

11 years agoUnbreak the build of newfs_nandfs.
Ed Schouten [Sun, 21 Apr 2013 22:36:14 +0000 (22:36 +0000)]
Unbreak the build of newfs_nandfs.

Mark global variables static. While there, make some small style(9)
cleanups and remove a variable that is unused.

11 years agoPlug static llentry leak (ipv4 & ipv6 were affected).
Oleg Bulyzhin [Sun, 21 Apr 2013 21:28:38 +0000 (21:28 +0000)]
Plug static llentry leak (ipv4 & ipv6 were affected).

PR: kern/172985
MFC after: 1 month

11 years agoRemove code duplication.
Mikolaj Golub [Sun, 21 Apr 2013 20:51:53 +0000 (20:51 +0000)]
Remove code duplication.

11 years agoDocument a few expansions for the $PS1 and $PS2 environmental variables.
Joel Dahl [Sun, 21 Apr 2013 19:55:38 +0000 (19:55 +0000)]
Document a few expansions for the $PS1 and $PS2 environmental variables.

PR: 173410
Submitted by: Derek Wood <ddwood@outlook.com>
Reviewed by: jilles

11 years agodup(2): Remove incorrect sentence about getdtablesize().
Jilles Tjoelker [Sun, 21 Apr 2013 19:42:04 +0000 (19:42 +0000)]
dup(2): Remove incorrect sentence about getdtablesize().

There are no getdtablesize() bounds on the file descriptor to be duplicated;
it only has to be open. If the RLIMIT_NOFILE rlimit was decreased after
opening the file descriptor, it may be greater than or equal to
getdtablesize() but still valid.

MFC after: 1 week

11 years agoFix typos in my last commit.
Warner Losh [Sun, 21 Apr 2013 17:23:27 +0000 (17:23 +0000)]
Fix typos in my last commit.

Submitted by: Niclas Zeising and Benjamin Kaduk

11 years agoNote merge to head for sendmail 8.14.7.
Gregory Neil Shapiro [Sun, 21 Apr 2013 17:13:58 +0000 (17:13 +0000)]
Note merge to head for sendmail 8.14.7.

11 years agoMinor changes to force commit these files so new freebsd*.cf files are
Gregory Neil Shapiro [Sun, 21 Apr 2013 17:11:45 +0000 (17:11 +0000)]
Minor changes to force commit these files so new freebsd*.cf files are
built to use the new sendmail-8.14.6/cf tree.

MFC after: 4 days

11 years agoEmbed revision id in the library.
Mikolaj Golub [Sun, 21 Apr 2013 17:10:10 +0000 (17:10 +0000)]
Embed revision id in the library.

MFC after: 29 days

11 years agoUpdate for sendmail 8.14.7
Gregory Neil Shapiro [Sun, 21 Apr 2013 17:09:51 +0000 (17:09 +0000)]
Update for sendmail 8.14.7

MFC after: 4 days

11 years agoMerge sendmail 8.14.7 to HEAD
Gregory Neil Shapiro [Sun, 21 Apr 2013 17:08:44 +0000 (17:08 +0000)]
Merge sendmail 8.14.7 to HEAD

MFC after: 4 days

11 years agoImport sendmail 8.14.7
Gregory Neil Shapiro [Sun, 21 Apr 2013 16:35:04 +0000 (16:35 +0000)]
Import sendmail 8.14.7

11 years agoFix manual page.
Hans Petter Selasky [Sun, 21 Apr 2013 16:09:35 +0000 (16:09 +0000)]
Fix manual page.

PR: usb/177896
Submitted by: Leif Velcro <velcroleaf@rocketmail.com>

11 years agoAdd OHCI controller ID.
Hans Petter Selasky [Sun, 21 Apr 2013 16:02:50 +0000 (16:02 +0000)]
Add OHCI controller ID.

MFC after: 2 weeks
Submitted by: Dmitry Luhtionov <dmitryluhtionov@gmail.com>

11 years agoMerge the 386 and amd64 versions of the fenv.h, to make cc -m32
Konstantin Belousov [Sun, 21 Apr 2013 13:31:55 +0000 (13:31 +0000)]
Merge the 386 and amd64 versions of the fenv.h, to make cc -m32
compilations which use fenv.h work.

Reviewed by: tjil
Sponsored by: The FreeBSD Foundation

11 years agoRemove redundant definitions of _ALIGN and _ALIGNBYTES.
Tijl Coosemans [Sun, 21 Apr 2013 11:12:44 +0000 (11:12 +0000)]
Remove redundant definitions of _ALIGN and _ALIGNBYTES.

11 years agoAdd example.
Joel Dahl [Sun, 21 Apr 2013 10:30:19 +0000 (10:30 +0000)]
Add example.

PR: 177025
Submitted by: Fernando <fernando.apesteguia@gmail.com>
Reviewed by: theraven

11 years agoMove EXAMPLES descriptions to before the actual command.
Joel Dahl [Sun, 21 Apr 2013 10:08:33 +0000 (10:08 +0000)]
Move EXAMPLES descriptions to before the actual command.

PR: 177870
Submitted by: Bjorn Heidotting <b.heidotting@yahoo.com>

11 years agoSince we didn't break the loop, we should set i to -1 to start from the
Andrey V. Elsukov [Sun, 21 Apr 2013 09:10:35 +0000 (09:10 +0000)]
Since we didn't break the loop, we should set i to -1 to start from the
beginning.

Submitted by: Steven Hartland
MFC after: 1 week

11 years agoFix an off by one calculation in wiipic_dispatch().
Rui Paulo [Sun, 21 Apr 2013 08:35:38 +0000 (08:35 +0000)]
Fix an off by one calculation in wiipic_dispatch().

11 years agoRemove cross-references to nonexistent CPU_SET(3) manpage.
Joel Dahl [Sun, 21 Apr 2013 06:46:41 +0000 (06:46 +0000)]
Remove cross-references to nonexistent CPU_SET(3) manpage.
Also fix cpu_getaffinity(2) document title.

PR: 176317
Submitted by: brucec

11 years agoWhen doing BAW tracking, don't dereference a NULL pointer if the BAW
Adrian Chadd [Sun, 21 Apr 2013 00:41:15 +0000 (00:41 +0000)]
When doing BAW tracking, don't dereference a NULL pointer if the BAW
slot is actually NULL.

11 years agoAdd more details about updating across the clang divide.
Warner Losh [Sat, 20 Apr 2013 23:33:51 +0000 (23:33 +0000)]
Add more details about updating across the clang divide.
Add hints about having a root shell around on live updates.
Couple of additional nits.

11 years agoThere's some races (likely in the BAR handling, sigh) which is causing
Adrian Chadd [Sat, 20 Apr 2013 22:46:31 +0000 (22:46 +0000)]
There's some races (likely in the BAR handling, sigh) which is causing
the pause/resume code to not be called completely symmetrically.

I'll chase down the root cause of that soon; this at least works around
the bug and tells me when it happens.

11 years agoAdd missing Pp. Also remove some minor whitespace.
Joel Dahl [Sat, 20 Apr 2013 22:26:33 +0000 (22:26 +0000)]
Add missing Pp. Also remove some minor whitespace.

PR: 177995
Submitted by: olgeni

11 years agomdoc: end function context properly.
Joel Dahl [Sat, 20 Apr 2013 22:16:49 +0000 (22:16 +0000)]
mdoc: end function context properly.

11 years agoWrong cast.
Mikolaj Golub [Sat, 20 Apr 2013 15:37:33 +0000 (15:37 +0000)]
Wrong cast.

MFC after: 1 month

11 years agoEnable libcompiler-rt on MIPS.
Ed Schouten [Sat, 20 Apr 2013 14:44:28 +0000 (14:44 +0000)]
Enable libcompiler-rt on MIPS.

Originally we disabled libcompiler-rt on MIPS and SPARC64, because of an
issue where __clzdi2 and __ctzdi2 would cause endless recursion. This
bug has been fixed in r230021 already, but for some reason we only
switched to libcompiler-rt on SPARC64 -- not MIPS.

This means we can finally use <stdatomic.h> on all our architectures.

11 years agoFix compilation.
Kenneth D. Merry [Sat, 20 Apr 2013 14:33:55 +0000 (14:33 +0000)]
Fix compilation.

Pointy hat to: ken

11 years agobin/df: Fix unitialized use in prtstat
Ulrich Spörlein [Sat, 20 Apr 2013 12:41:05 +0000 (12:41 +0000)]
bin/df: Fix unitialized use in prtstat

While here:
- use NULL in the context of pointers
- use memset instead of bzero throughout the file
- free memory to appease clang static analyzer

Found by: Coverity Scan (the UNINIT one)

11 years agoMake DELAY(9) reference pause(9).
Edward Tomasz Napierala [Sat, 20 Apr 2013 12:36:53 +0000 (12:36 +0000)]
Make DELAY(9) reference pause(9).

11 years agoMerge of vendor of 249690, tzdata2013c
Edwin Groothuis [Sat, 20 Apr 2013 11:42:10 +0000 (11:42 +0000)]
Merge of vendor of 249690, tzdata2013c

- antarctica: AusAQ and ATAQ have been removed.
- Antarctica/Macquarie has been moved to australasia file and AU.
- Asia/Hebron, Palestine updated for 2013.
- Paraguay stays with DST for the whole year.

11 years agoVendor import of tzdata2013c
Edwin Groothuis [Sat, 20 Apr 2013 11:30:14 +0000 (11:30 +0000)]
Vendor import of tzdata2013c

- antarctica: AusAQ and ATAQ have been removed.
- Antarctica/Macquarie has been moved to australasia file and AU.
- Asia/Hebron, Palestine updated for 2013.
- Paraguay stays with DST for the whole year.

Obtained from: ftp://ftp.iana.org/tz/releases/

11 years agoBump date.
Mikolaj Golub [Sat, 20 Apr 2013 08:25:14 +0000 (08:25 +0000)]
Bump date.

11 years agoSync gcore(1) with the recent changes in kernel code aimed at adding
Mikolaj Golub [Sat, 20 Apr 2013 08:23:19 +0000 (08:23 +0000)]
Sync gcore(1) with the recent changes in kernel code aimed at adding
procstat notes to a process core file.

Suggested by: jhb
MFC after: 1 month

11 years agoMake use of newly added libprocstat(3) ability to extract procstat
Mikolaj Golub [Sat, 20 Apr 2013 08:22:09 +0000 (08:22 +0000)]
Make use of newly added libprocstat(3) ability to extract procstat
info from a process core file.

So now one can run procstat(1) on a process core e.g. to get a list of
files opened by a process when it crashed:

root@lisa:/ # procstat -f /root/vi.core
  PID COMM               FD T V FLAGS     REF  OFFSET PRO NAME
  658 vi               text v r r--------   -       - -   /usr/bin/vi
  658 vi               ctty v c rw-------   -       - -   /dev/pts/0
  658 vi                cwd v d r--------   -       - -   /root
  658 vi               root v d r--------   -       - -   /
  658 vi                  0 v c rw-------  11    3208 -   /dev/pts/0
  658 vi                  1 v c rw-------  11    3208 -   /dev/pts/0
  658 vi                  2 v c rw-------  11    3208 -   /dev/pts/0
  658 vi                  3 v r r----n-l-   1       0 -   /tmp/vi.0AYKz3Lps7
  658 vi                  4 v r rw-------   1       0 -   /var/tmp/vi.recover/vi.GaGYsz
  658 vi                  5 v r rw-------   1       0 -   -

PR: kern/173723
Suggested by: jhb
MFC after: 1 month

11 years agoUse procstat_getkstack(3) for retrieving process kernel stacks
Mikolaj Golub [Sat, 20 Apr 2013 08:19:06 +0000 (08:19 +0000)]
Use procstat_getkstack(3) for retrieving process kernel stacks
instead of direct sysctl calls.

MFC after: 1 month

11 years agoAdd procstat_getkstack function to dump kernel stacks of a process.
Mikolaj Golub [Sat, 20 Apr 2013 08:17:20 +0000 (08:17 +0000)]
Add procstat_getkstack function to dump kernel stacks of a process.

MFC after: 1 month

11 years agoUse libprocstat(3) to retrieve ELF auxiliary vector.
Mikolaj Golub [Sat, 20 Apr 2013 08:15:43 +0000 (08:15 +0000)]
Use libprocstat(3) to retrieve ELF auxiliary vector.

MFC after: 1 month

11 years agoThis should have been committed in r249666.
Mikolaj Golub [Sat, 20 Apr 2013 08:13:35 +0000 (08:13 +0000)]
This should have been committed in r249666.

MFC after: 1 month

11 years agoAdd procstat_getauxv function to retrieve a process auxiliary vector.
Mikolaj Golub [Sat, 20 Apr 2013 08:10:47 +0000 (08:10 +0000)]
Add procstat_getauxv function to retrieve a process auxiliary vector.

MFC after: 1 month

11 years agoUse libprocstat(3) to retrieve process command line arguments and
Mikolaj Golub [Sat, 20 Apr 2013 08:08:29 +0000 (08:08 +0000)]
Use libprocstat(3) to retrieve process command line arguments and
environment variables.

MFC after: 1 month

11 years agoExtend libprocstat with functions to retrieve process command line
Mikolaj Golub [Sat, 20 Apr 2013 08:07:04 +0000 (08:07 +0000)]
Extend libprocstat with functions to retrieve process command line
arguments and environment variables.

Suggested by: stas
Reviewed by: jhb and stas (initial version)
MFC after: 1 month

11 years agoUse libprocstat(3) when retrieving binary information for a process.
Mikolaj Golub [Sat, 20 Apr 2013 08:05:04 +0000 (08:05 +0000)]
Use libprocstat(3) when retrieving binary information for a process.

MFC after: 1 month

11 years agoAdd procstat_getosrel function to retrieve a process osrel info.
Mikolaj Golub [Sat, 20 Apr 2013 08:03:56 +0000 (08:03 +0000)]
Add procstat_getosrel function to retrieve a process osrel info.

MFC after: 1 month

11 years agoAdd procstat_getpathname function to retrieve a process executable.
Mikolaj Golub [Sat, 20 Apr 2013 08:02:43 +0000 (08:02 +0000)]
Add procstat_getpathname function to retrieve a process executable.

MFC after: 1 month

11 years agoUse procstat_getrlimit(3) for retrieving rlimit information instead of
Mikolaj Golub [Sat, 20 Apr 2013 08:01:00 +0000 (08:01 +0000)]
Use procstat_getrlimit(3) for retrieving rlimit information instead of
direct sysctl calls.

MFC after: 1 month

11 years agoAdd procstat_getrlimit function to retrieve a process resource limits info.
Mikolaj Golub [Sat, 20 Apr 2013 07:59:44 +0000 (07:59 +0000)]
Add procstat_getrlimit function to retrieve a process resource limits info.

MFC after: 1 month

11 years agoUse procstat_getumask(3) for retrieving umaks information instead of
Mikolaj Golub [Sat, 20 Apr 2013 07:58:20 +0000 (07:58 +0000)]
Use procstat_getumask(3) for retrieving umaks information instead of
direct sysctl.

MFC after: 1 month

11 years agoAdd procstat_getumask function to retrieve a process umask.
Mikolaj Golub [Sat, 20 Apr 2013 07:57:08 +0000 (07:57 +0000)]
Add procstat_getumask function to retrieve a process umask.

MFC after: 1 month

11 years agoUse procstat_getgroups(3) for retrieving groups information instead of
Mikolaj Golub [Sat, 20 Apr 2013 07:55:31 +0000 (07:55 +0000)]
Use procstat_getgroups(3) for retrieving groups information instead of
direct sysctl.

MFC after: 1 month

11 years agoAdd procstat_getgroups function to retrieve process groups.
Mikolaj Golub [Sat, 20 Apr 2013 07:54:07 +0000 (07:54 +0000)]
Add procstat_getgroups function to retrieve process groups.

MFC after: 1 month

11 years agoUse more generic procstat_getvmmap(3) for retrieving VM layout of a process.
Mikolaj Golub [Sat, 20 Apr 2013 07:52:23 +0000 (07:52 +0000)]
Use more generic procstat_getvmmap(3) for retrieving VM layout of a process.

MFC after: 1 month

11 years agoUse procstat_getprocs(3) for retrieving thread information instead of
Mikolaj Golub [Sat, 20 Apr 2013 07:50:59 +0000 (07:50 +0000)]
Use procstat_getprocs(3) for retrieving thread information instead of
direct sysctl calls.

MFC after: 1 month

11 years agoAdd procstat_getvmmap function to get VM layout of a process.
Mikolaj Golub [Sat, 20 Apr 2013 07:49:35 +0000 (07:49 +0000)]
Add procstat_getvmmap function to get VM layout of a process.

MFC after: 1 month

11 years agoMake libprocstat(3) extract procstat notes from a process core file.
Mikolaj Golub [Sat, 20 Apr 2013 07:47:26 +0000 (07:47 +0000)]
Make libprocstat(3) extract procstat notes from a process core file.

PR: kern/173723
Suggested by: jhb
Glanced by: kib
MFC after: 1 month

11 years agoAdd note about fagility of the clang upgrade process.
Warner Losh [Sat, 20 Apr 2013 01:12:23 +0000 (01:12 +0000)]
Add note about fagility of the clang upgrade process.

11 years agoExpose CAM_BOOT_DELAY as a kernel conf item now.
Sean Bruno [Sat, 20 Apr 2013 00:33:37 +0000 (00:33 +0000)]
Expose CAM_BOOT_DELAY as a kernel conf item now.

This allows users who boot without loader to adjust their environments
around slightly buggy or slow hardware.

PR: kern/161809
Submitted by: rozhuk.im@gmail.com
MFC after: 2 weeks

11 years agoMerge ACPICA 20130418.
Jung-uk Kim [Fri, 19 Apr 2013 23:49:34 +0000 (23:49 +0000)]
Merge ACPICA 20130418.

11 years agoInitialise the chainmask fields regardless of whether 11n support
Adrian Chadd [Fri, 19 Apr 2013 21:49:11 +0000 (21:49 +0000)]
Initialise the chainmask fields regardless of whether 11n support
is compiled in or not.

This fixes issues with people running -HEAD but who build modules
without doing a "make buildkernel KERNCONF=XXX", thus picking up
opt_*.h.  The resulting module wouldn't have 11n enabled and the
chainmask configuration would just be plain wrong.

11 years agomostly whitespace changes:
Luigi Rizzo [Fri, 19 Apr 2013 21:08:21 +0000 (21:08 +0000)]
mostly whitespace changes:
- remove vestiges of the old memory allocator
- clean up some comments

11 years agoUpdate chio(1) and ch(4) to support reporting element designators.
Kenneth D. Merry [Fri, 19 Apr 2013 20:03:51 +0000 (20:03 +0000)]
Update chio(1) and ch(4) to support reporting element designators.

This allows mapping a tape drive in a changer (as reported by
'chio status') to a sa(4) driver instance by comparing the
serial numbers.

The designators can be ASCII (which is printed out directly), binary
(which is printed in hex format) or UTF-8, which is printed in either
native UTF-8 format if the terminal can support it, or in %XX notation
for non-ASCII characters.  Thanks to Hiroki Sato <hrs@> for the
explaining UTF-8 printing and example UTF-8 printing code.

chio.h: Modify the changer_element_status structure to add new
fields and definitions from the SMC3r16 spec.

Rename the original CHIOGSTATUS ioctl to OCHIOGTATUS and
define a new CHIOGSTATUS ioctl.

Clean up some tab/space issues.

chio.c:  For the 'status' subcommand, print the designator field
if it is supplied by a device.

scsi_ch.h: Add new flags for DVCID and CURDATA to the READ
ELEMENT STATUS command structure.

Add a read_element_status_device_id structure
for the data fields in the new standard. Add new
unions, dt_or_obsolete and voltage_devid, to hold
and address data from either SCSI-2 or newer devices.

scsi_ch.c: Implement support for fetching device IDs with READ
ELEMENT STATUS data.

Add new arguments to scsi_read_element_status() to
allow the user to request the DVCID and CURDATA bits.
This isn't compiled into libcam (it's only an internal
kernel interface), so we don't need any special
handling for the API change.

If the user issues the new CHIOGSTATUS ioctl, copy all of
the available element status data out.  If he issues the
OCHIOGSTATUS ioctl, we don't copy the new fields in the
structure.

Fix a bug in chopen() that would result in the peripheral
never getting unheld if chgetparams() failed.

Sponsored by: Spectra Logic
Submitted by: Po-Li Soong
MFC After: 1 week

11 years agoAdd the Clang specific -Wmissing-variable-declarations to WARNS=6.
Ed Schouten [Fri, 19 Apr 2013 19:45:00 +0000 (19:45 +0000)]
Add the Clang specific -Wmissing-variable-declarations to WARNS=6.

This compiler flag enforces that that people either mark variables
static or use an external declarations for the variable, similar to how
-Wmissing-prototypes works for functions.

Due to the fact that Yacc/Lex generate code that cannot trivially be
changed to not warn because of this (lots of yy* variables), add a
NO_WMISSING_VARIABLE_DECLARATIONS that can be used to turn off this
specific compiler warning.

Announced on: toolchain@

11 years agoFix -Wmissing-variable-declarations compiler warnings.
Ed Schouten [Fri, 19 Apr 2013 19:38:39 +0000 (19:38 +0000)]
Fix -Wmissing-variable-declarations compiler warnings.

References:
https://www.illumos.org/issues/3700

11 years agoImplement a very basic multi-PHY aware switch device.
Adrian Chadd [Fri, 19 Apr 2013 17:50:38 +0000 (17:50 +0000)]
Implement a very basic multi-PHY aware switch device.

This is intended to be used as a stop-gap for switch devices
which expose multiple ethernet PHYs but we don't have a driver
for - here, etherswitchcfg and the general switch configuration
API can be used to interface to said PHYs.

Submitted by: Luiz Otavio O Souza <loos.br@gmail.com>

11 years agoInclude PID in the error message which is printed when the maxproc limit
Jaakko Heinonen [Fri, 19 Apr 2013 15:19:29 +0000 (15:19 +0000)]
Include PID in the error message which is printed when the maxproc limit
is exceeded. Improve formatting of the message while here.

PR: kern/60550
Submitted by: Lowell Gilbert, bde

11 years agoDon't compare unsigned socklen_t against < 0.
Gleb Smirnoff [Fri, 19 Apr 2013 13:40:13 +0000 (13:40 +0000)]
Don't compare unsigned socklen_t against < 0.

Reviewed by: jhb

11 years agosem: Restart the POSIX sem_* calls after signals with SA_RESTART set.
Jilles Tjoelker [Fri, 19 Apr 2013 10:16:00 +0000 (10:16 +0000)]
sem: Restart the POSIX sem_* calls after signals with SA_RESTART set.

Programs often do not expect an [EINTR] return from sem_wait() and POSIX
only allows it if the signal was installed without SA_RESTART. The timeout
in sem_timedwait() is absolute so it can be restarted normally.

The umtx call can be invoked with a relative timeout and in that case
[ERESTART] must be changed to [EINTR]. However, libc does not do this.

The old POSIX semaphore implementation did this correctly (before r249566),
unlike the new umtx one.

It may be desirable to avoid [EINTR] completely, which matches the pthread
functions and is explicitly permitted by POSIX. However, the kernel must
return [EINTR] at least for signals with SA_RESTART clear, otherwise pthread
cancellation will not abort a semaphore wait. In this commit, only restore
the 8.x behaviour which is also permitted by POSIX.

Discussed with: jhb
MFC after: 1 week

11 years agoAdd a debug statement to log the currently chosen chainmask configuration.
Adrian Chadd [Fri, 19 Apr 2013 08:06:45 +0000 (08:06 +0000)]
Add a debug statement to log the currently chosen chainmask configuration.

11 years ago.. don't know how this snuck into this commit. Sorry.
Adrian Chadd [Fri, 19 Apr 2013 08:01:34 +0000 (08:01 +0000)]
.. don't know how this snuck into this commit. Sorry.

Fix compile build before anyone notices.

11 years agoPrint out the chainmask configuration.
Adrian Chadd [Fri, 19 Apr 2013 07:56:22 +0000 (07:56 +0000)]
Print out the chainmask configuration.

11 years agoUse uint32_t for fields that are fetched via ath_hal_getcapability().
Adrian Chadd [Fri, 19 Apr 2013 06:59:10 +0000 (06:59 +0000)]
Use uint32_t for fields that are fetched via ath_hal_getcapability().

11 years agoImport ACPICA 20130418.
Jung-uk Kim [Fri, 19 Apr 2013 05:49:53 +0000 (05:49 +0000)]
Import ACPICA 20130418.

11 years agoFix the uart(4) module build. Without uart_dev_lpc the module cannot be loaded.
Justin Hibbits [Fri, 19 Apr 2013 05:46:16 +0000 (05:46 +0000)]
Fix the uart(4) module build.  Without uart_dev_lpc the module cannot be loaded.

11 years agoAn even more refined version of r249484, until we can come up with
Tim Kientzle [Fri, 19 Apr 2013 05:13:57 +0000 (05:13 +0000)]
An even more refined version of r249484, until we can come up with
a good fix for the -print-file-name=include breakage.

As suggested by Andrey Chernov.

11 years agoFix the symlink creation from r249484 so that repeated
Tim Kientzle [Fri, 19 Apr 2013 05:01:14 +0000 (05:01 +0000)]
Fix the symlink creation from r249484 so that repeated
installs work correctly.

Suggested by Tijl Coosemans.

11 years agoAttempt to mitigate poor initialization of arc4 by one-shot
Andrey A. Chernov [Fri, 19 Apr 2013 00:30:52 +0000 (00:30 +0000)]
Attempt to mitigate poor initialization of arc4 by one-shot
reinitialization from yarrow right after good entropy is harvested.

Approved by:    secteam (delphij)
MFC after:      1 week

11 years agoWhen an NFS unmount occurs, once vflush() writes the last dirty
Rick Macklem [Thu, 18 Apr 2013 23:20:16 +0000 (23:20 +0000)]
When an NFS unmount occurs, once vflush() writes the last dirty
buffer for the last vnode on the mount back to the server, it
returns. At that point, the code continues with the unmount,
including freeing up the nfs specific part of the mount structure.
It is possible that an nfsiod thread will try to check for an
empty I/O queue in the nfs specific part of the mount structure
after it has been free'd by the unmount. This patch avoids this problem by
setting the iodmount entries for the mount back to NULL while holding the
mutex in the unmount and checking the appropriate entry is non-NULL after
acquiring the mutex in the nfsiod thread.

Reported and tested by: pho
Reviewed by: kib
MFC after: 2 weeks

11 years agocxgbe(4): Refuse to install T5 firmwares on a T4 card (and vice versa).
Navdeep Parhar [Thu, 18 Apr 2013 22:54:41 +0000 (22:54 +0000)]
cxgbe(4): Refuse to install T5 firmwares on a T4 card (and vice versa).

MFC after: 1 week

11 years agoRecover missing arp_ifinit() call.
Oleg Bulyzhin [Thu, 18 Apr 2013 20:13:33 +0000 (20:13 +0000)]
Recover missing arp_ifinit() call.

MFC after: 2 weeks