]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
8 years agoRe-add AES-CBC ciphers to the default cipher list on the server.
des [Fri, 11 Mar 2016 00:23:10 +0000 (00:23 +0000)]
Re-add AES-CBC ciphers to the default cipher list on the server.

PR: 207679

8 years agoUpgrade to OpenSSH 7.2p2.
des [Fri, 11 Mar 2016 00:15:29 +0000 (00:15 +0000)]
Upgrade to OpenSSH 7.2p2.

8 years agoWhen -S is specified dump summary to stdout, not stderr, so it's
sobomax [Thu, 10 Mar 2016 23:19:35 +0000 (23:19 +0000)]
When -S is specified dump summary to stdout, not stderr, so it's
easier to capture and process it with external tools via pipe.

8 years agocxgbe(4): Improvements to the code that deals with the firmware's log.
np [Thu, 10 Mar 2016 23:17:26 +0000 (23:17 +0000)]
cxgbe(4): Improvements to the code that deals with the firmware's log.

- Query the location of the log very early during attach.  Refresh the
  location later after establishing contact with the firmware.
- Save the log's location as a flat address in devlog_params.
- Use a memory window instead of backdoor access to the EDC/MC to read
  the log.

8 years agoAdd -S option to print out summary after compression has been
sobomax [Thu, 10 Mar 2016 21:36:24 +0000 (21:36 +0000)]
Add -S option to print out summary after compression has been
completed.

MFC after:  2 weeks

8 years agocxgbe(4): Fix bug in r296603. The memory window needs to be
np [Thu, 10 Mar 2016 20:36:32 +0000 (20:36 +0000)]
cxgbe(4): Fix bug in r296603.  The memory window needs to be
repositioned if the start address isn't in the window already.  One
of the bounds check used the end address instead.

8 years agoRemove redundant files already tracked by tools/build/mk/OptionalObsoleteFiles.inc.
bdrewery [Thu, 10 Mar 2016 20:15:27 +0000 (20:15 +0000)]
Remove redundant files already tracked by tools/build/mk/OptionalObsoleteFiles.inc.

These files are installed, likely after r288230.  In
tools/build/mk/OptionalObsoleteFiles.inc they are bound
to the MK_BINUTILS option rather than unconditionally
deleted here.

Reported by: Kurt Lidl <lidl@pix.net>
MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division

8 years agoVendor import of OpenSSH 7.2p1.
des [Thu, 10 Mar 2016 20:10:25 +0000 (20:10 +0000)]
Vendor import of OpenSSH 7.2p1.

8 years agoFix mdoc markup.
brueffer [Thu, 10 Mar 2016 18:21:03 +0000 (18:21 +0000)]
Fix mdoc markup.

8 years agoRevert r292074 (by smh): Limit stripesize reported from nvd(4) to 4K
mav [Thu, 10 Mar 2016 17:13:10 +0000 (17:13 +0000)]
Revert r292074 (by smh): Limit stripesize reported from nvd(4) to 4K

I believe that this patch handled the problem from the wrong side.
Instead of making ZFS properly handle large stripe sizes, it made
unrelated driver to lie in reported parameters to workaround that.

Alternative solution for this problem from ZFS side was committed at
r296615.

Discussed with: smh

8 years agoMake ZFS ignore stripe sizes above SPA_MAXASHIFT (8KB).
mav [Thu, 10 Mar 2016 16:39:46 +0000 (16:39 +0000)]
Make ZFS ignore stripe sizes above SPA_MAXASHIFT (8KB).

If device has stripe size bigger then maximal sector size supported by
ZFS, there is nothing can be done to avoid read-modify-write cycles.
Taking that stripe size into account will only reduce space efficiency
and pointlessly bother user with warnings that can not be fixed.

Discussed with: smh

8 years agoAdd support for ddb(4).
br [Thu, 10 Mar 2016 15:51:43 +0000 (15:51 +0000)]
Add support for ddb(4).

Sponsored by: DARPA, AFRL
Sponsored by: HEIF5

8 years agoMake ZFS more picky to GEOM stripe sizes and offsets.
mav [Thu, 10 Mar 2016 14:18:14 +0000 (14:18 +0000)]
Make ZFS more picky to GEOM stripe sizes and offsets.

Use of misaligned or non-power-of-2 stripes is not really useful for ZFS,
since increased ashift won't help to avoid read-modify-write cycles, and
only reduce pool space efficiency and compression rates.

8 years agocxgb(4): Remove redundant part of an assertion.
np [Thu, 10 Mar 2016 14:17:24 +0000 (14:17 +0000)]
cxgb(4): Remove redundant part of an assertion.

PR: 207858
Submitted by: David Binderman

8 years agoMFV r296609: 6370 ZFS send fails to transmit some holes
mav [Thu, 10 Mar 2016 09:01:19 +0000 (09:01 +0000)]
MFV r296609: 6370 ZFS send fails to transmit some holes

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Chris Williamson <chris.williamson@delphix.com>
Reviewed by: Stefan Ring <stefanrin@gmail.com>
Reviewed by: Steven Burgess <sburgess@datto.com>
Reviewed by: Arne Jansen <sensille@gmx.net>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Paul Dagnelie <pcd@delphix.com>

In certain circumstances, "zfs send -i" (incremental send) can produce a
stream which will result in incorrect sparse file contents on the
target.

The problem manifests as regions of the received file that should be
sparse (and read a zero-filled) actually contain data from a file that
was deleted (and which happened to share this file's object ID).

Note: this can happen only with filesystems (not zvols, because they do
not free (and thus can not reuse) object IDs).

Note: This can happen only if, since the incremental source (FromSnap),
a file was deleted and then another file was created, and the new file
is sparse (i.e. has areas that were never written to and should be
implicitly zero-filled).

We suspect that this was introduced by 4370 (applies only if hole_birth
feature is enabled), and made worse by 5243 (applies if hole_birth
feature is disabled, and we never send any holes).

The bug is caused by the hole birth feature. When an object is deleted
and replaced, all the holes in the object have birth time zero. However,
zfs send cannot tell that the holes are new since the file was replaced,
so it doesn't send them in an incremental. As a result, you can end up
with invalid data when you receive incremental send streams. As a
short-term fix, we can always send holes with birth time 0 (unless it's
a zvol or a dataset where we can guarantee that no objects have been
reused).

Closes #37

openzfs/openzfs@adef853162e83f7cdf6b2d9af9756d434a9c743b

8 years agoDon't assume that bio_cmd is a bitfield.
imp [Thu, 10 Mar 2016 06:25:47 +0000 (06:25 +0000)]
Don't assume that bio_cmd is a bitfield.

Differential Revision: https://reviews.freebsd.org/D5591

8 years agoDon't assume that bio_cmd is a bit mask.
imp [Thu, 10 Mar 2016 06:25:39 +0000 (06:25 +0000)]
Don't assume that bio_cmd is a bit mask.

Differential Revision: https://reviews.freebsd.org/D5592

8 years agoDon't assume that bio_cmd is bit mask.
imp [Thu, 10 Mar 2016 06:25:31 +0000 (06:25 +0000)]
Don't assume that bio_cmd is bit mask.

Differential Revision: https://reviews.freebsd.org/D5593

8 years agoMove to new value for XPT_GET_SIM_KNOB to avoid clash with XPT_ATA_IO.
imp [Thu, 10 Mar 2016 06:25:05 +0000 (06:25 +0000)]
Move to new value for XPT_GET_SIM_KNOB to avoid clash with XPT_ATA_IO.

8 years agocxgbe(4): Add general purpose routines that offer safe access to the
np [Thu, 10 Mar 2016 06:15:31 +0000 (06:15 +0000)]
cxgbe(4): Add general purpose routines that offer safe access to the
chip's memory windows.  Convert existing users of these windows to the
new routines.

8 years agoFix bug in VNIC causing phony number of available TX descriptors
zbb [Thu, 10 Mar 2016 05:45:24 +0000 (05:45 +0000)]
Fix bug in VNIC causing phony number of available TX descriptors

TSO packets will signal segments TX completion in the separate CQ
descriptors. Each CQ descriptor for HW TSO will point to the same
SQ entry.
Do not invoke nicvf_put_sq_desc() for secondary segments to avoid
free_cnt corruption and eventually integer overflow that will result
in the negative free_cnt value and hence impossibility of further
transmission.

Reviewed by:   wma
Obtained from: Semihalf
Sponsored by:  Cavium
Differential Revision: https://reviews.freebsd.org/D5535

8 years agoFix "received NULL mbuf" bug in VNIC
zbb [Thu, 10 Mar 2016 05:23:46 +0000 (05:23 +0000)]
Fix "received NULL mbuf" bug in VNIC

Do not modify NIC_QSET_CQ_0_7_HEAD manually, especially
in non-atomic context.
It doesn't seem to be necessary to recreate CQ head after
interrupt clearing too.

Reviewed by:   wma
Obtained from: Semihalf
Sponsored by:  Cavium
Differential Revision: https://reviews.freebsd.org/D5533

8 years agocxgbe(4): Allow the addr/len pair that is being validated in
np [Thu, 10 Mar 2016 02:43:10 +0000 (02:43 +0000)]
cxgbe(4): Allow the addr/len pair that is being validated in
validate_mem_range to span multiple memory types.  Update
validate_mt_off_len to use validate_mem_range.

8 years agohyperv/hn: Make the # of TX rings configurable.
sephe [Thu, 10 Mar 2016 02:37:47 +0000 (02:37 +0000)]
hyperv/hn: Make the # of TX rings configurable.

Rename the tunables to avoid confusion.

MFC after: 1 week
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5578

8 years agohyperv/hn: Factor out hn_channel_attach
sephe [Thu, 10 Mar 2016 02:28:01 +0000 (02:28 +0000)]
hyperv/hn: Factor out hn_channel_attach

MFC after: 1 week
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5577

8 years agohyperv/hn: Move if_initname to an earlier place
sephe [Thu, 10 Mar 2016 02:13:42 +0000 (02:13 +0000)]
hyperv/hn: Move if_initname to an earlier place

So that functions shared w/ attach path could use if_printf().

While I'm here, remove unnecessary if_dunit and if_dname assignment.

MFC after: 1 week
Sponsored by: Microsoft OSTC
Differential Revision: https://reviews.freebsd.org/D5576

8 years agoDon't assume that bio_cmd is a bitfield.
imp [Thu, 10 Mar 2016 00:36:52 +0000 (00:36 +0000)]
Don't assume that bio_cmd is a bitfield.

Differential revision: https://reviews.freebsd.org/D5590

8 years agoDon't assume bio_cmd is a bit field.
imp [Thu, 10 Mar 2016 00:36:45 +0000 (00:36 +0000)]
Don't assume bio_cmd is a bit field.

Differential Revision: https://reviews.freebsd.org/D5594

8 years agoAdd raw RX-50 support. These are 400k single sided disks with 80
imp [Thu, 10 Mar 2016 00:36:38 +0000 (00:36 +0000)]
Add raw RX-50 support. These are 400k single sided disks with 80
tracks and 10 sectors per track. More exotic RX-50 types not
supported, nor is there support for de-interleaving the first two
tracks where the physical sectors are 0 1 2 3 4 5 6 7 8 9, but they
should be interpreted as 0 5 1 6 2 7 3 8 4 9. This is purely to read
the media with dd. The FAT that's on these disks won't work with
msdosfs anyway.

8 years agoStop assuming that bio_cmd is a bit field.
imp [Thu, 10 Mar 2016 00:33:06 +0000 (00:33 +0000)]
Stop assuming that bio_cmd is a bit field.

Differential Revision: https://reviews.freebsd.org/D5587

8 years agoActually send a asconf chunk, not only queue one.
tuexen [Thu, 10 Mar 2016 00:27:10 +0000 (00:27 +0000)]
Actually send a asconf chunk, not only queue one.

MFC after: 3 days

8 years agoDIRDEPS_BUILD: Connect MK_TESTS.
bdrewery [Wed, 9 Mar 2016 22:46:01 +0000 (22:46 +0000)]
DIRDEPS_BUILD: Connect MK_TESTS.

Sponsored by: EMC / Isilon Storage Division

8 years agoFix and connect setjmp test.
bdrewery [Wed, 9 Mar 2016 22:45:04 +0000 (22:45 +0000)]
Fix and connect setjmp test.

Sponsored by: EMC / Isilon Storage Division

8 years agoThese group names may be used as a cookie, so replace any non-fs-safe characters.
bdrewery [Wed, 9 Mar 2016 22:45:00 +0000 (22:45 +0000)]
These group names may be used as a cookie, so replace any non-fs-safe characters.

One example is in cddl/usr.sbin/dtrace/tests/common/aggs.  It could be
fixed but other uses of this would break, especially in the
DIRDEPS_BUILD which uses the group names for stage cookies.

MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division

8 years agoDIRDEPS_BUILD: Let PROGS bootstrapping work.
bdrewery [Wed, 9 Mar 2016 22:44:48 +0000 (22:44 +0000)]
DIRDEPS_BUILD: Let PROGS bootstrapping work.

- Support (DP|LIB)ADD_${PROG}.
- Support SRCS[._]${PROG}.
- Don't bootstrap DIRDEPS while recursing.

Sponsored by: EMC / Isilon Storage Division

8 years agoFix code so that buf_ring allocation for Tx Queues and their mutexes
davidcs [Wed, 9 Mar 2016 21:12:26 +0000 (21:12 +0000)]
Fix code so that buf_ring allocation for Tx Queues and their mutexes
is done during during bxe_attach() and freed during bxe_detach()

MFC after: 5 days

8 years agosh: Add test for 'set -o nolog'.
jilles [Wed, 9 Mar 2016 21:05:21 +0000 (21:05 +0000)]
sh: Add test for 'set -o nolog'.

The option does not do anything so check that the output of 'set +o' is
different.

8 years agosh: Avoid out-of-bounds access in setoptionbyindex() for 'set -o nolog'.
jilles [Wed, 9 Mar 2016 21:00:57 +0000 (21:00 +0000)]
sh: Avoid out-of-bounds access in setoptionbyindex() for 'set -o nolog'.

Reported by: hrs

8 years agoRemove these broken filemon tests.
bdrewery [Wed, 9 Mar 2016 20:15:03 +0000 (20:15 +0000)]
Remove these broken filemon tests.

They were not very useful in their current state.  It only ran a fork bomb,
confirmed headers/footers matched, hard-coded the number of expected entries
(rather than ensuring each entry is present when expected), and was missing a
sizeof_long.c file from r251368 which makes its intent for testing 32-bit
binaries unclear.

More extensive tests should be written with ATF now.

8 years agoFILEMON_SET_FD: Disallow changing the fd.
bdrewery [Wed, 9 Mar 2016 19:50:35 +0000 (19:50 +0000)]
FILEMON_SET_FD: Disallow changing the fd.

MFC after: 1 week
Suggested by: mjg
Sponsored by: EMC / Isilon Storage Division

8 years agoFor the MD_ROOT option don't inject /dev/md0 as root dev when ROOTDEVNAME
sobomax [Wed, 9 Mar 2016 19:36:25 +0000 (19:36 +0000)]
For the MD_ROOT option don't inject /dev/md0 as root dev when ROOTDEVNAME
is defined explicitly. It's kinda pointless and results in extra step in
boot sequence which is not really needed, i.e.:

md0: Embedded image 1331200 bytes at 0x8038b7b4
Trying to mount root from ufs:/dev/md0 []...
Mounting from ufs:/dev/md0 failed with error 22.
Trying to mount root from ufs:md0.uzip []...
warning: no time-of-day clock registered, system time will not be set accurately
start_init: trying /sbin/init

8 years agoRegen.
jhb [Wed, 9 Mar 2016 19:06:46 +0000 (19:06 +0000)]
Regen.

8 years agoSimplify AIO initialization now that it is standard.
jhb [Wed, 9 Mar 2016 19:05:11 +0000 (19:05 +0000)]
Simplify AIO initialization now that it is standard.

- Mark AIO system calls as STD and remove the helpers to dynamically
  register them.
- Use COMPAT6 for the old system calls with the older sigevent instead of
  an 'o' prefix.
- Simplify the POSIX configuration to note that AIO is always available.
- Handle AIO in the default VOP_PATHCONF instead of special casing it in
  the pathconf() system call.  fpathconf() is still hackish.
- Remove freebsd32_aio_cancel() as it just called the native one directly.

Reviewed by: kib
Sponsored by: Chelsio Communications
Differential Revision: https://reviews.freebsd.org/D5589

8 years agoUse ptrace(2) LWP events to track threads reliably in truss.
jhb [Wed, 9 Mar 2016 18:45:41 +0000 (18:45 +0000)]
Use ptrace(2) LWP events to track threads reliably in truss.

- truss can now log the system call invoked by a thread during a
  voluntary process exit.  No return value is logged, but the value passed
  to exit() is included in the trace output.  Arguments passed to thread
  exit system calls such as thr_exit() are not logged as voluntary thread
  exits cannot be distinguished from involuntary thread exits during a
  system call.
- New events are now reported for thread births and exits similar to the
  recently added events for new child processes when following forks.

Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D5561

8 years agoFix reporting of the CloudABI ABI in kdump.
jhb [Wed, 9 Mar 2016 18:38:30 +0000 (18:38 +0000)]
Fix reporting of the CloudABI ABI in kdump.

- Advertise the word size for CloudABI ABIs via the SV_LP64 flag.  All of
  the other ABIs include either SV_ILP32 or SV_LP64.
- Fix kdump to not assume a 32-bit ABI if the ABI flags field is non-zero
  but SV_LP64 isn't set.  Instead, only assume a 32-bit ABI if SV_ILP32 is
  set and fallback to the unknown value of "00" if neither SV_LP64 nor
  SV_ILP32 is set.

Reviewed by: kib, ed
Differential Revision: https://reviews.freebsd.org/D5560

8 years agoSecond argument of the mips_timer_init_params() is boolean, so
sobomax [Wed, 9 Mar 2016 18:38:03 +0000 (18:38 +0000)]
Second argument of the mips_timer_init_params() is boolean, so
pass in "1" for true consistently.

8 years agoMissed addition to r296563 to fix newer tools to work with older kernel.
mav [Wed, 9 Mar 2016 16:05:13 +0000 (16:05 +0000)]
Missed addition to r296563 to fix newer tools to work with older kernel.

8 years agoSet buffer to empty string to prevent duplicated output in some cases.
ae [Wed, 9 Mar 2016 14:47:05 +0000 (14:47 +0000)]
Set buffer to empty string to prevent duplicated output in some cases.

PR: 193888

8 years agoFix spelling of MAXNAMLEN.
trasz [Wed, 9 Mar 2016 13:45:03 +0000 (13:45 +0000)]
Fix spelling of MAXNAMLEN.

MFC after: 1 month
Sponsored by: The FreeBSD Foundation

8 years agoImprove U-Boot API detection
sgalabov [Wed, 9 Mar 2016 11:45:48 +0000 (11:45 +0000)]
Improve U-Boot API detection

Until now, ubldr has been trying to locate the U-Boot API using a hint
address (U-Boot's current stack pointer), aligning it to 1MiB and going
over a 3MiB (or 1MiB in case of MIPS) memory region searching for a
valid API signature.

This change introduces an alternative way of doing this, namely the
following:
- both U-Boot's bootelf and go commands actually pass argc and argv to
  the entry point (e.g., ubldr's start function, but they should also
  be passed over to main() transparently)
- so, instead of trying to go and look for a valid API signature, we
  look at the parameters passed to main()
- if there's an option '-a' with argument, which is a valid hexadecimal
  unsigned  long number (x), we try to verify whether we have a valid
  API signature at address x. If so - we use it. If not - we fallback
  to the original way of locating the API signature.

The U-Boot change, which causes the API structure address to be
exported as an environment variable, was committed to mainline U-Boot
as commit 22aa61f707574dd569296f521fcfc46a05f51c48

Reviewed by: andrew, adrian
Approved by: adrian (mentor)
Sponsored by: Smartcom - Bulgaria AD
Differential Revision: https://reviews.freebsd.org/D5492

8 years agoAdd new IOCTL compat shims for ABI breakage caused by r296510:
mav [Wed, 9 Mar 2016 11:16:15 +0000 (11:16 +0000)]
Add new IOCTL compat shims for ABI breakage caused by r296510:
MFV r296505: 6531 Provide mechanism to artificially limit disk performance

8 years agoAdd support for IPPROTO_IPV6 socket layer for getsockopt/setsockopt calls.
ae [Wed, 9 Mar 2016 09:12:40 +0000 (09:12 +0000)]
Add support for IPPROTO_IPV6 socket layer for getsockopt/setsockopt calls.
Also add mapping for several options from RFC 3493 and 3542.

Reviewed by: dchagin
Tested by: Joe Love <joe at getsomwhere dot net>
MFC after: 2 weeks

8 years agoFollow-up r295827: Don't enable meta stats when recursing PROGS.
bdrewery [Wed, 9 Mar 2016 03:22:23 +0000 (03:22 +0000)]
Follow-up r295827: Don't enable meta stats when recursing PROGS.

Sponsored by: EMC / Isilon Storage Division

8 years agoDIRDEPS_BUILD+PROGS: Fix staging not respecting (BINDIR|PROGNAME)[._]${PROG}.
bdrewery [Wed, 9 Mar 2016 03:22:20 +0000 (03:22 +0000)]
DIRDEPS_BUILD+PROGS: Fix staging not respecting (BINDIR|PROGNAME)[._]${PROG}.

Observed in tests/sys/kern.

Sponsored by: EMC / Isilon Storage Division

8 years agoRemove things set already by bsd.progs.mk.
bdrewery [Wed, 9 Mar 2016 03:22:17 +0000 (03:22 +0000)]
Remove things set already by bsd.progs.mk.

MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division

8 years agoPROGS: Track child meta files.
bdrewery [Wed, 9 Mar 2016 03:22:08 +0000 (03:22 +0000)]
PROGS: Track child meta files.

This will allow Makefile.depend to properly capture all dependencies.
It is not 100% optimal but works.  Other options would be to use *.meta
here which would include too much or to keep a Makefile.depend per PROG
and include it from the main Makefile.depend which would not be
straight forward.

Sponsored by: EMC / Isilon Storage Division

8 years agocxgbe(4): Rename regwin_lock to reg_lock. It is used to protect access
np [Tue, 8 Mar 2016 22:23:30 +0000 (22:23 +0000)]
cxgbe(4): Rename regwin_lock to reg_lock.  It is used to protect access
to indirect registers only.

8 years agoDIRDEPS_BUILD: Update clang dependencies after r296417.
bdrewery [Tue, 8 Mar 2016 21:26:44 +0000 (21:26 +0000)]
DIRDEPS_BUILD: Update clang dependencies after r296417.

Sponsored by: EMC / Isilon Storage Division

8 years agoDon't ever create object directories here with MK_AUTO_OBJ.
bdrewery [Tue, 8 Mar 2016 21:26:25 +0000 (21:26 +0000)]
Don't ever create object directories here with MK_AUTO_OBJ.

Sponsored by: EMC / Isilon Storage Division

8 years agodrm/i915: Update to match Linux 3.8.13
dumbbell [Tue, 8 Mar 2016 20:33:02 +0000 (20:33 +0000)]
drm/i915: Update to match Linux 3.8.13

This update brings initial support for Haswell GPUs.

Tested by: Many users of FreeBSD, PC-BSD and HardenedBSD
Relnotes: yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D5554

8 years agoBetter english.
dchagin [Tue, 8 Mar 2016 19:40:01 +0000 (19:40 +0000)]
Better english.

Submitted by: Kevin P. Neal
MFC after: 1 week

8 years agocxgbe(4): Reshuffle and rototill t4_hw.c, solely to reduce diffs with
np [Tue, 8 Mar 2016 19:34:56 +0000 (19:34 +0000)]
cxgbe(4): Reshuffle and rototill t4_hw.c, solely to reduce diffs with
the internal shared code.

Obtained from: Chelsio Communications

8 years agoPut a commit message from r296502 about Linux alarm() system call
dchagin [Tue, 8 Mar 2016 19:20:57 +0000 (19:20 +0000)]
Put a commit message from r296502 about Linux alarm() system call
behaviour to the source.

Suggested by: emaste@

MFC after: 1 week

8 years agoLoad linux64 module for amd64 if Linux abi enabled.
dchagin [Tue, 8 Mar 2016 19:08:55 +0000 (19:08 +0000)]
Load linux64 module for amd64 if Linux abi enabled.

Reviewed by: emaste@
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D5567

8 years agoMFV r296540: 4448 zfs diff misprints unicode characters
mav [Tue, 8 Mar 2016 18:53:00 +0000 (18:53 +0000)]
MFV r296540: 4448 zfs diff misprints unicode characters

Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com>
Reviewed by: Toomas Soome <tsoome@me.com>
Approved by: Matthew Ahrens <mahrens@delphix.com>
Author: Joshua M. Clulow <jmc@joyent.com>

illumos/illumos-gate@b211eb9181f99c20acbf4c528f94cb44b4ca8c31

8 years agoMFV r296538: 6544 incorrect comment in libzfs.h about offline status
mav [Tue, 8 Mar 2016 18:48:20 +0000 (18:48 +0000)]
MFV r296538: 6544 incorrect comment in libzfs.h about offline status

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Gerhard Roethlin <git@the-color-black.net>

illumos/illumos-gate@cb605c4d8ab24b5a900b8b4ca85db65c22d05fad

8 years agoMFV r296536: 6551 cmd/zpool: cleanup gcc warnings
mav [Tue, 8 Mar 2016 18:39:39 +0000 (18:39 +0000)]
MFV r296536: 6551 cmd/zpool: cleanup gcc warnings

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Andy Stormont <astormont@racktopsystems.com>
Approved by: Robert Mustacchi <rm@joyent.com>

illumos/illumos-gate@b327cd3f3b4dab4f29e7140159b1e01ed2ceef2a

8 years agoMFV r296534: 6550 cmd/zfs: cleanup gcc warnings
mav [Tue, 8 Mar 2016 18:35:53 +0000 (18:35 +0000)]
MFV r296534: 6550 cmd/zfs: cleanup gcc warnings

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Andy Stormont <astormont@racktopsystems.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Igor Kozhukhov <ikozhukhov@gmail.com>

illumos/illumos-gate@c16bcc4577f389573eff411c7b7e040294078c3b

8 years agoMFV r296532: 6637 replacing "dontclose" with "should_close"
mav [Tue, 8 Mar 2016 18:32:31 +0000 (18:32 +0000)]
MFV r296532: 6637 replacing "dontclose" with "should_close"

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: David Schwartz <dschwartz783@gmail.com>

illumos/illumos-gate@d189620258b3c9b0e2f7e2104840be2eee7c68e5

8 years agoMFV r296529:
mav [Tue, 8 Mar 2016 18:28:24 +0000 (18:28 +0000)]
MFV r296529:
6672 arc_reclaim_thread() should use gethrtime() instead of ddi_get_lbolt()
6673 want a macro to convert seconds to nanoseconds and vice-versa

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
Reviewed by: Robert Mustacchi <rm@joyent.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Eli Rosenthal <eli.rosenthal@delphix.com>

illumos/illumos-gate@a8f6344fa0921599e1f4511e41b5f9a25c38c0f9

8 years agoMFV r296527: 6659 nvlist_free(NULL) is a no-op
mav [Tue, 8 Mar 2016 18:11:38 +0000 (18:11 +0000)]
MFV r296527:  6659 nvlist_free(NULL) is a no-op

Reviewed by: Toomas Soome <tsoome@me.com>
Reviewed by: Marcel Telka <marcel@telka.sk>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Josef 'Jeff' Sipek <jeffpc@josefsipek.net>

illumos/illumos-gate@aab83bb83be7342f6cfccaed8d5fe0b2f404855d

8 years agoRecord command exit status in the typescript file when running simple commands.
bdrewery [Tue, 8 Mar 2016 18:05:23 +0000 (18:05 +0000)]
Record command exit status in the typescript file when running simple commands.

Also capitalize 'command:'.

Relnotes: yes
MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Division

8 years agoJust exit in the child if execve(2) fails.
bdrewery [Tue, 8 Mar 2016 18:05:20 +0000 (18:05 +0000)]
Just exit in the child if execve(2) fails.

No functional change.

This is mostly addressing a false-positive from the clang static
analyzer due to it thinking that done() was being called with freed
memory, however the kill(0, SIGTERM) made the done() never reached.
It doesn't make sense to the show the footer from the child anyhow, nor
does it make sense to kill the process group here since the execve(2)
failed in the child.  This code was leftover from many years of refactoring.

MFC after: 1 week
Sponsored by: EMC / Isilon Storage Division

8 years agoFilemon: Attach from the child to avoid racing with the parent attach.
bdrewery [Tue, 8 Mar 2016 18:05:02 +0000 (18:05 +0000)]
Filemon: Attach from the child to avoid racing with the parent attach.

This is the same as how the bmake filemon usage works.

This also fixes failed attach not properly flushing the TTY.

MFC after: 1 week
Relnotes: yes
Sponsored by: EMC / Isilon Storage Division

8 years agoMFV r296522: 6541 Pool feature-flag check defeated if "verify" is included
mav [Tue, 8 Mar 2016 17:58:02 +0000 (17:58 +0000)]
MFV r296522: 6541 Pool feature-flag check defeated if "verify" is included
in the dedup property value

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Laager <rlaager@wiktel.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: ilovezfs <ilovezfs@icloud.com>

illumos/illumos-gate@971640e6aa954c91b0706543741aa4570299f4d7

8 years agoMFV r296520: 6562 Refquota on receive doesn't account for overage
mav [Tue, 8 Mar 2016 17:53:42 +0000 (17:53 +0000)]
MFV r296520:  6562 Refquota on receive doesn't account for overage

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Yuri Pankov <yuri.pankov@nexenta.com>
Reviewed by: Toomas Soome <tsoome@me.com>
Approved by: Gordon Ross <gwr@nexenta.com>
Author: Dan McDonald <danmcd@omniti.com>

illumos/illumos-gate@5f7a8e6d750cb070a3347f045201c6206caee6aa

8 years agoMFV r296518: 5027 zfs large block support (add copyright)
mav [Tue, 8 Mar 2016 17:51:09 +0000 (17:51 +0000)]
MFV r296518:  5027 zfs large block support (add copyright)

Author: Matthew Ahrens <matt@mahrens.org>

illumos/illumos-gate@c3d26abc9ee97b4f60233556aadeb57e0bd30bb9

8 years agoboot1.efi: use += to append to LDFLAGS
emaste [Tue, 8 Mar 2016 17:45:56 +0000 (17:45 +0000)]
boot1.efi: use += to append to LDFLAGS

This is for consistency with loader.efi's Makefile and simplifies some
out-of-tree experimentation.

8 years agoMFV r296515: 6536 zfs send: want a way to disable setting of
mav [Tue, 8 Mar 2016 17:43:21 +0000 (17:43 +0000)]
MFV r296515:  6536 zfs send: want a way to disable setting of
DRR_FLAG_FREERECORDS

Reviewed by: Anil Vijarnia <avijarnia@racktopsystems.com>
Reviewed by: Kim Shrier <kshrier@racktopsystems.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Andrew Stormont <astormont@racktopsystems.com>

illumos/illumos-gate@880094b6062aebeec8eda6a8651757611c83b13e

8 years agoMFV r296513: 6450 scrub/resilver unnecessarily traverses snapshots created
mav [Tue, 8 Mar 2016 17:34:58 +0000 (17:34 +0000)]
MFV r296513: 6450 scrub/resilver unnecessarily traverses snapshots created
after the scrub started

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Matthew Ahrens <mahrens@delphix.com>

illumos/illumos-gate@38d61036746e2273cc18f6698392e1e29f87d1bf

8 years agoMFV r296511: 6537 Panic on zpool scrub with DEBUG kernel
mav [Tue, 8 Mar 2016 17:32:24 +0000 (17:32 +0000)]
MFV r296511: 6537 Panic on zpool scrub with DEBUG kernel

Reviewed by: Steve Gonczi <gonczi@comcast.net>
Reviewed by: Dan McDonald <danmcd@omniti.com>
Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Matthew Ahrens <mahrens@delphix.com>
Author: Gary Mills <gary_mills@fastmail.fm>

illumos/illumos-gate@8c04a1fa3f7d569d48fe9b5342d0bd4c533179b9

8 years agoMFV r296505: 6531 Provide mechanism to artificially limit disk performance
mav [Tue, 8 Mar 2016 17:27:13 +0000 (17:27 +0000)]
MFV r296505: 6531 Provide mechanism to artificially limit disk performance

Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Dan McDonald <danmcd@omniti.com>
Author: Prakash Surya <prakash.surya@delphix.com>

illumos/illumos-gate@97e81309571898df9fdd94aab1216dfcf23e060b

8 years agoDelete empty sys/modules/aio/ directory
ngie [Tue, 8 Mar 2016 16:47:23 +0000 (16:47 +0000)]
Delete empty sys/modules/aio/ directory

This was missed in r296277

X-MFC with: r296277
Sponsored by: EMC / Isilon Storage Division

8 years agoFollow-up r296324: Fix STATICOBJS dependency guesses conditions.
bdrewery [Tue, 8 Mar 2016 16:12:55 +0000 (16:12 +0000)]
Follow-up r296324: Fix STATICOBJS dependency guesses conditions.

Reported by: antoine
Pointyhat to: bdrewery
Sponsored by: EMC / Isilon Storage Division

8 years agoDoes not leak fp. While here remove bogus cast of fp->f_data.
dchagin [Tue, 8 Mar 2016 15:55:43 +0000 (15:55 +0000)]
Does not leak fp. While here remove bogus cast of fp->f_data.

MFC after: 1 week

8 years agoLinux accept() system call return EOPNOTSUPP errno instead of EINVAL
dchagin [Tue, 8 Mar 2016 15:15:34 +0000 (15:15 +0000)]
Linux accept() system call return EOPNOTSUPP errno instead of EINVAL
for UDP sockets.

MFC after: 1 week

8 years agoAccording to POSIX and Linux implementation the alarm() system call
dchagin [Tue, 8 Mar 2016 15:12:49 +0000 (15:12 +0000)]
According to POSIX and Linux implementation the alarm() system call
is always successfull.
So, ignore any errors and return 0 as a Linux do.

XXX. Unlike POSIX, Linux in case when the invalid seconds value specified
always return 0, so in that case Linux does not return proper remining time.

MFC after: 1 week

8 years agoLink the newly created process to the corresponding parent as
dchagin [Tue, 8 Mar 2016 15:08:22 +0000 (15:08 +0000)]
Link the newly created process to the corresponding parent as
if CLONE_PARENT is set, then the parent of the new process will
be the same as that of the calling process.

MFC after: 1 week

8 years agoMention resolvconf(8) in resolv.conf(5).
trasz [Tue, 8 Mar 2016 11:04:08 +0000 (11:04 +0000)]
Mention resolvconf(8) in resolv.conf(5).

MFC after: 1 month
Sponsored by: The FreeBSD Foundation

8 years agocxgbe(4): Minor updates to the shared routines that deal with firmware images.
np [Tue, 8 Mar 2016 10:07:40 +0000 (10:07 +0000)]
cxgbe(4): Minor updates to the shared routines that deal with firmware images.

8 years agocxgbe(4): Fix t4_tp_get_rdma_stats.
np [Tue, 8 Mar 2016 09:40:45 +0000 (09:40 +0000)]
cxgbe(4): Fix t4_tp_get_rdma_stats.

8 years agocxgbe(4): Many new functions in the shared code, unused at this time.
np [Tue, 8 Mar 2016 09:34:56 +0000 (09:34 +0000)]
cxgbe(4): Many new functions in the shared code, unused at this time.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Use t4_link_down_rc_str in shared code to decode the reason
np [Tue, 8 Mar 2016 08:59:34 +0000 (08:59 +0000)]
cxgbe(4): Use t4_link_down_rc_str in shared code to decode the reason
the link is down, instead of doing it in OS specific code.

8 years agoAdd MIPS_INTRNG to sys/conf/options.mips
sgalabov [Tue, 8 Mar 2016 08:57:53 +0000 (08:57 +0000)]
Add MIPS_INTRNG to sys/conf/options.mips

This was somehow missed in the commit of
https://reviews.freebsd.org/D5182
although it was in the original diff submitted for review.

Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D5568

8 years agocxgbe(4): Updates to shared routines that get/set various parameters via
np [Tue, 8 Mar 2016 08:39:53 +0000 (08:39 +0000)]
cxgbe(4): Updates to shared routines that get/set various parameters via
the firmware.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Remove __devinit and SPEED_<foo> as part of catch up with
np [Tue, 8 Mar 2016 08:13:37 +0000 (08:13 +0000)]
cxgbe(4): Remove __devinit and SPEED_<foo> as part of catch up with
internal shared code.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Updates to the shared routines that deal with the serial EEPROM,
np [Tue, 8 Mar 2016 07:48:55 +0000 (07:48 +0000)]
cxgbe(4): Updates to the shared routines that deal with the serial EEPROM,
flash, and VPD.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Updates to mailbox routines in the shared code.
np [Tue, 8 Mar 2016 06:27:47 +0000 (06:27 +0000)]
cxgbe(4): Updates to mailbox routines in the shared code.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Update the interrupt handlers for hardware errors.
np [Tue, 8 Mar 2016 02:44:32 +0000 (02:44 +0000)]
cxgbe(4): Update the interrupt handlers for hardware errors.

Obtained from: Chelsio Communications

8 years agocxgbe(4): Overhaul the shared code that deals with the chip's TP block,
np [Tue, 8 Mar 2016 02:04:05 +0000 (02:04 +0000)]
cxgbe(4): Overhaul the shared code that deals with the chip's TP block,
which is responsible for filtering and RSS.

Add the ability to use filters that match on PF/VF (aka "VNIC id") while
here.  This is mutually exclusive with filtering on outer VLAN tag with
Q-in-Q.

Sponsored by: Chelsio Communications