Commit Graph

1107 Commits

Author SHA1 Message Date
Kyle Evans
b29bf2f84e libbe(3)/be(8): Drop WARNS overrides, fix all fallout
Based on the idea that we shouldn't have all-new library and utility going
into base that need WARNS=1...

- Decent amount of constification
- Lots of parentheses
- Minor other nits
2018-07-25 15:14:35 +00:00
Kyle Evans
268af06d3e Normalize bectl(8)/libbe(3) Makefiles, remove Makefile copyright/license
Approved by:	hselaskey
2018-07-24 19:55:02 +00:00
Kyle Evans
70a11a8eea libbe(3): Add to cddl build, adjust src.libnames.mk as needed 2018-07-24 15:42:23 +00:00
Michael Tuexen
53e0911116 Improve TCP related tests for dtrace.
Ensure that the TCP connections are terminated gracefully as expected
by the test. Use appropriate numbers for sent/received packets.
In addition, enable tst.localtcpstate.ksh, which should pass, but
doesn't until https://reviews.freebsd.org/D16369 is committed.

Reviewed by:		markj@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16288
2018-07-22 10:50:59 +00:00
Michael Tuexen
be029a4979 Test that the dtrace UDP receive probe fires.
This test ensures that the fix committed in
https://svnweb.freebsd.org/changeset/base/336551
actually works.

Reviewed by:		dteske@, markj@, rrs@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16046
2018-07-20 15:37:29 +00:00
Michael Tuexen
10b803a40d Adjust comment to reality since r286171.
Sponsored by:		Netflix, Inc.
2018-07-15 20:42:47 +00:00
Michael Tuexen
e0f9b8233f Don't require a local sshd for the local TCP state dtrace test
This change is similar to the one done in r286171 for
tst.ipv4localtcp.ksh. This not only reduces the requirements on the
system used for testing but results also in a graceful teardown of
the TCP connection.

Reviewed by:		gnn@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16276
2018-07-15 20:41:16 +00:00
Michael Tuexen
dc9f20b3f3 Fix the UDP tests for dtrace.
The code imported from opensolaris was depending on ping supporting
UDP for sending probes. Since this is not supported by ping on FreeBSD
use a perl script instead.
The remote test requires the usage of ksh93, so state that in the
sheband.
Enable the local test, but keep the remote test disabled, since it
requires a remote machine on the LAN.

Reviewed by:		markj@, gnn@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16268
2018-07-15 20:34:22 +00:00
Michael Tuexen
60e4fb3a3f Return the intended return code.
This bug was spotted by markj@ in D16268 because I copied this code part
and used it there. So fix it.

Sponsored by:		Netflix, Inc.
2018-07-14 19:53:41 +00:00
Michael Tuexen
49d7124b18 Fix shebangs and execute bit of test scripts.
Since we don't have /usr/bin/ksh, use a generic way of specifying
ksh. Some of the tests only run with ksh93, so use this shell
for these tests. Two of the tests don't have the execute bit set,
so fix this, too.

Reviewed by:		markj@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16270
2018-07-14 19:49:14 +00:00
Michael Tuexen
d073fd606c Add support for TCP state names used by Solaris.
For compatibility, add the TCP state names used by Solaris
and given in the Dtrace Guide available at
https://docs.oracle.com/cd/E37838_01/html/E61035/glhgu.html#OSDTGglhmv

Reviewed by:		markj@
Sponsored by:		Netflix, Inc.
Differential Revision:	https://reviews.freebsd.org/D16269
2018-07-14 17:12:04 +00:00
Michael Tuexen
7575d3df5b The IP, TCP, and UDP provider report IP addresses as strings.
In some cases, the required information is not available and the
UDP provider reported an empty string in this case and the IP
and TCP provider reported a NULL pointer.

This patch changes the value provided in this case to the string
"<unknown>". This make the behaviour consistent and in-line with
the behaviour of Solaris.

Reviewed by:		markj@, dteske@, gnn@
Differential Revision:	https://reviews.freebsd.org/D15855
2018-06-18 18:35:29 +00:00
Eric van Gyzen
79d4ee83de Fix markup in zfs(8); no content change
Sponsored by:	Dell EMC
2018-06-15 15:28:31 +00:00
Mark Johnston
ecbde90073 Process CUs with a language attribute of DW_LANG_Mips_Assembler.
At the moment ctfconvert(1) does not do much with such CUs, but
that may not be true in the future, and we run ctfconvert on several
assembly files during the build.

X-MFC with:	r334883
2018-06-11 16:33:36 +00:00
Mark Johnston
c5fda9bac0 Don't process DWARF generated from non-C/C++ code.
ctfconvert(1) is not designed to handle DWARF generated from such code,
and will generally fail in non-obvious ways.  Use an explicit check to
help catch such potential failures.

Reported by:	Johannes Lundberg <johalun0@gmail.com>
MFC after:	2 weeks
2018-06-09 15:10:49 +00:00
Sean Eric Fagan
69724399c4 This originated from ZFS On Linux, as
d4a72f2386

During scans (scrubs or resilvers), it sorts the blocks in each transaction
group by block offset; the result can be a significant improvement. (On my
test system just now, which I put some effort to introduce fragmentation into
the pool since I set it up yesterday, a scrub went from 1h2m to 33.5m with the
changes.) I've seen similar rations on production systems.

Approved by:	Alexander Motin
Obtained from:	ZFS On Linux
Relnotes:	Yes (improved scrub performance, with tunables)
Differential Revision:	https://reviews.freebsd.org/D15562
2018-06-08 17:38:28 +00:00
Sean Bruno
8d7181d1e0 Unbreak dtrace runtime for udp after svn r334719 SO_REUSEPORT commit.
Submitted by:	Johannes Lundberg <johalun0@gmail.com>
Sponsored by:	Limeligght Networks
2018-06-07 15:27:07 +00:00
Sean Bruno
1a43cff92a Load balance sockets with new SO_REUSEPORT_LB option.
This patch adds a new socket option, SO_REUSEPORT_LB, which allow multiple
programs or threads to bind to the same port and incoming connections will be
load balanced using a hash function.

Most of the code was copied from a similar patch for DragonflyBSD.

However, in DragonflyBSD, load balancing is a global on/off setting and can not
be set per socket. This patch allows for simultaneous use of both the current
SO_REUSEPORT and the new SO_REUSEPORT_LB options on the same system.

Required changes to structures:
Globally change so_options from 16 to 32 bit value to allow for more options.
Add hashtable in pcbinfo to hold all SO_REUSEPORT_LB sockets.

Limitations:
As DragonflyBSD, a load balance group is limited to 256 pcbs (256 programs or
threads sharing the same socket).

This is a substantially different contribution as compared to its original
incarnation at svn r332894 and reverted at svn r332967.  Thanks to rwatson@
for the substantive feedback that is included in this commit.

Submitted by:	Johannes Lundberg <johalun0@gmail.com>
Obtained from:	DragonflyBSD
Relnotes:	Yes
Sponsored by:	Limelight Networks
Differential Revision:	https://reviews.freebsd.org/D11003
2018-06-06 15:45:57 +00:00
Devin Teske
e52d921647 dwatch(1): Update manual to reference actual release
MFC after:	3 days
X-MFC-to:	stable/11
Sponsored by:	Smule, Inc.
2018-06-03 23:49:02 +00:00
Devin Teske
7d958cc9aa dwatch(1): Fix "-t test" for post-processing profiles
Profiles that perform post-processing of the DTrace output were
dropping the "-t test" option on the floor. Fix handling of this
option for said profiles.

X-MFC-to:	stable/11
X-MFC-with:	r334261-334262
Sponsored by:	Smule, Inc.
2018-05-29 22:36:37 +00:00
Devin Teske
1b114e50c9 dwatch(1): Eliminate ANSI dimming in developer mode
"Developer mode" (passing of "-dev" options), which enables debugging
features on compilation error, used to dim lines unrelated to error.

That proved distracting and feedback from testers also confirmed that
simply highlighting the line the compiler complains about is enough.

Sponsored by:	Smule, Inc.
2018-05-27 22:32:45 +00:00
Devin Teske
e08e9e9990 dwatch(1): Guard against error when given -t "*..."
dwatch allows you to customnize the predicate (condition) for when
information is displayed. The DTrace syntax for this is:

	probe[, ...] /predicate/ { [actions] }

But if predicate is something like "*args[1]!=NULL" to test that
the first pointer in an array is non-NULL, the syntax produced is:

	probe[, ...] /*arg1!=NULL/ { [actions] }

The issue being that "/*" is the beginning of a comment and thus
the following error is emitted:

	dtrace: failed to compile script /dev/stdin:
		line 535: /* encountered inside a comment

This patch adds whitespace around the argument given to -t,
producing:

	probe[, ...] / *arg1!=NULL / { [actions] }

Sponsored by:	Smule, Inc.
2018-05-27 22:27:47 +00:00
Matt Macy
1f52c1db90 ctf dwarf: don't report "no dwarf entry" as if it were an error 2018-05-19 18:50:58 +00:00
Matt Macy
d0ba1baed3 ctfconvert: silence useless enum has too many values warning 2018-05-19 06:31:17 +00:00
Devin Teske
1764b74809 dwatch(1): Refactor sendrecv profile
The profile for send(2)/recv(2) observation has been refactored to
eliminate alloca() in favor of translations available in HEAD.

Sponsored by:	Smule, Inc.
2018-05-12 06:23:30 +00:00
Devin Teske
f134539c02 dwatch(1): Expose process for ip/tcp/udp
Knowing the value of execname during these probes is of some value even
if it is commonly the interrupt kernel thread (intr[12]) -- quite often
it is not, but that depends on the probe.

Sponsored by:	Smule, Inc.
2018-05-12 06:18:15 +00:00
Devin Teske
a061d97027 dwatch(1): Allow `-E code' to override profile EVENT_DETAILS
This allows quick changes to the formatted output of a profile.

Sponsored by:	Smule, Inc.
2018-05-12 06:01:41 +00:00
Devin Teske
6a21b3cdf4 dwatch(1): Export ARGV to profiles loaded via load_profile()
A module that wishes to post-process the output needs to know which
arguments were passed in order to re-execute a child in a pipe-chain.
Further, the expansion of ARGV needs to be such that items are escaped
properly.

Sponsored by:	Smule, Inc.
2018-05-12 05:49:31 +00:00
Devin Teske
4a73674e32 dwatch(1): Simplify info message test
The info() function already tests $QUIET

Sponsored by:	Smule, Inc.
2018-05-12 05:43:47 +00:00
Devin Teske
5f2202d104 dwatch(1): Separate default values so `-[BK] num' don't affect usage
If you were to pass an invalid option after `-B num' or `-K num' you
would see that the usage statement would show the value you passed
instead of the actual default.

Moving the default values to separate variables that are unaffected
by the options-parsing allows the usage statement to correctly show
the hard-coded default values if no flags are used.

Sponsored by:	Smule, Inc.
2018-05-12 05:41:28 +00:00
Devin Teske
4a84c26cfc dwatch(1): Bugfix, usage displayed with `-1Q'
A return statement should have been an exit in list_profiles().
If the user passed `-Q' to list profiles and asked for one-line
per profile (`-1'), list_profiles() would not exit as should.

Sponsored by:	Smule, Inc.
2018-05-12 05:36:47 +00:00
Sean Bruno
86784a0a22 Cleanup sundry clang warnings for code that is not upstream in illumos.
https://github.com/illumos/illumos-gate/edit/master/usr/src/lib/libzfs/common/libzfs_sendrecv.c

Patch our version of it to quiesce warnings until someone decides to sync
up our code:

libzfs_sendrecv.c:2555:30: warning: format specifies type 'unsigned long'
  but the argument has type 'uint64_t' (aka 'unsigned long long') [-Wformat]
                                sprintf(guidname, "%lu", thisguid);
                                                   ~~~   ^~~~~~~~
                                                   %llu
libzfs_sendrecv.c:2612:29: warning: format specifies type 'unsigned long'
  but the argument has type 'uint64_t' (aka 'unsigned long long') [-Wformat]
                        sprintf(guidname, "%lu", parent_fromsnap_guid);
                                           ~~~   ^~~~~~~~~~~~~~~~~~~~
                                           %llu
libzfs_sendrecv.c:2645:29: warning: format specifies type 'unsigned long'
  but the argument has type 'uint64_t' (aka 'unsigned long long') [-Wformat]
                        sprintf(guidname, "%lu", parent_fromsnap_guid);
                                           ~~~   ^~~~~~~~~~~~~~~~~~~~
                                           %llu

Reviewed by:	allanjude
Differential Revision:	https://reviews.freebsd.org/D15325
2018-05-06 16:22:02 +00:00
Eitan Adler
4822188974 zpool(8): correct list of default properties in 'list'.
The default provides output in the following form:
```
NAME       SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP  DEDUP
HEALTH  ALTROOT
```

this corrects the man page.

Also submitted upstream as
https://github.com/openzfs/openzfs/pull/632/files (with slightly
different changes needed)
2018-04-28 01:14:16 +00:00
Sean Bruno
7875017ca9 Revert r332894 at the request of the submitter.
Submitted by:	Johannes Lundberg <johalun0_gmail.com>
Sponsored by:	Limelight Networks
2018-04-24 19:55:12 +00:00
Sean Bruno
7b7796eea5 Load balance sockets with new SO_REUSEPORT_LB option
This patch adds a new socket option, SO_REUSEPORT_LB, which allow multiple
programs or threads to bind to the same port and incoming connections will be
load balanced using a hash function.

Most of the code was copied from a similar patch for DragonflyBSD.

However, in DragonflyBSD, load balancing is a global on/off setting and can not
be set per socket. This patch allows for simultaneous use of both the current
SO_REUSEPORT and the new SO_REUSEPORT_LB options on the same system.

Required changes to structures
Globally change so_options from 16 to 32 bit value to allow for more options.
Add hashtable in pcbinfo to hold all SO_REUSEPORT_LB sockets.

Limitations
As DragonflyBSD, a load balance group is limited to 256 pcbs
(256 programs or threads sharing the same socket).

Submitted by:	Johannes Lundberg <johanlun0@gmail.com>
Sponsored by:	Limelight Networks
Differential Revision:	https://reviews.freebsd.org/D11003
2018-04-23 19:51:00 +00:00
Devin Teske
4e16c227ac dwatch(1): Remove the line used to demonstrate `-dev' option
In recently added sendrecv profile, there was a line purposefully
added to introduce a compilation error in which `-dev' is used to
debug the entry. Removing the entry.

Sponsored by:	Smule, Inc.
2018-04-22 02:40:21 +00:00
Devin Teske
f3cf700b57 dwatch(1): Add profile for send(2)/recv(2) syscalls
Sponsored by:	Smule, Inc.
2018-04-22 02:36:06 +00:00
Devin Teske
bcce9a2b33 dwatch(1): Add `-dev' option to aid debugging of profiles
The options `-d' (debug), `-e' (exit after compile), and `-v' (verbose)
when combined in any order (though best remembered as `-dev') will run
the conflated script through dtrace(1), test for error conditions, and
show the line that dtrace(1) failed at (with context).

If no errors are found, the output is the same as `-e[v]'.

When writing a new profile for dwatch(1), you can quickly test to
make sure it compiles by running `dwatch -devX profile_name' where
profiles live in /usr/libexec/dwatch or /usr/local/libexec/dwatch
(the latter being where profiles installed via ports should go).

Sponsored by:	Smule, Inc.
2018-04-22 02:20:17 +00:00
Alexander Motin
e4d6c7fc17 MFV man pages update from r329502: 7614 zfs device evacuation/removal.
MFC after:	3 days
2018-04-17 02:33:54 +00:00
Andriy Gapon
81f187e576 allow ZFS pool to have temporary name for duration of current import
The change adds -t <name> option to zpool create and -t option to zpool
import in its form with an old name and a new name.  This allows to
import (or create) a pool under a name that's different from its real,
permanent name without affecting that name.  This is useful when working
with VM images or images of other physical systems if they happen to
have a ZFS pool with the same name as the host system.

The changes come from ZoL with some small tweaks.
The porting has been done by julian.

The change is being submitted to OpenZFS:
https://github.com/openzfs/openzfs/pull/600

Submitted by:	julian
Reviewed by:	smh
MFC after:	2 weeks
Sponsored by:	Panzura (porting)
Differential Revision: https://reviews.freebsd.org/D14972
2018-04-12 10:37:26 +00:00
Alexander Motin
849a7ce2d5 MFV r331712:
9280 Assertion failure while running removal_with_ganging test with 4K devices

illumos/illumos-gate@243952c7ee

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Approved by: Garrett D'Amore <garrett@damore.org>
Author: Matt Ahrens <Matt.Ahrens@delphix.com>
2018-03-28 23:17:29 +00:00
Alexander Motin
5c4561f332 MFV r331706:
9235 rename zpool_rewind_policy_t to zpool_load_policy_t

illumos/illumos-gate@5dafeea3eb

We want to be able to pass various settings during import/open of a pool,
which are not only related to rewind. Instead of adding a new policy and
duplicate a bunch of code, we should just rename rewind_policy to a more
generic term like load_policy.

For instance, we'd like to set spa->spa_import_flags from the nvlist,
rather from a flags parameter passed to spa_import as in some cases we want
those flags not only for the import case, but also for the open case. One
such flag could be ZFS_IMPORT_MISSING_LOG (as used in zdb) which would
allow zfs to open a pool when logs are missing.

Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Pavel Zakharov <pavel.zakharov@delphix.com>
2018-03-28 22:29:06 +00:00
Alexander Motin
0b0c76bc58 MFV r331695, 331700: 9166 zfs storage pool checkpoint
illumos/illumos-gate@8671400134

The idea of Storage Pool Checkpoint (aka zpool checkpoint) deals with
exactly that.  It can be thought of as a “pool-wide snapshot” (or a
variation of extreme rewind that doesn’t corrupt your data).  It remembers
the entire state of the pool at the point that it was taken and the user
can revert back to it later or discard it.  Its generic use case is an
administrator that is about to perform a set of destructive actions to ZFS
as part of a critical procedure.  She takes a checkpoint of the pool before
performing the actions, then rewinds back to it if one of them fails or puts
the pool into an unexpected state.  Otherwise, she discards it.  With the
assumption that no one else is making modifications to ZFS, she basically
wraps all these actions into a “high-level transaction”.

Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Serapheim Dimitropoulos <serapheim.dimitro@delphix.com>
2018-03-28 22:01:27 +00:00
Alexander Motin
4bada7a0a2 Partial MFV r329753:
8809 libzpool should leverage work done in libfakekernel

illumos/illumos-gate@f06dce2c1f

Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Gordon Ross <gordon.w.ross@gmail.com>
Approved by: Richard Lowe <richlowe@richlowe.net>
Author: Andrew Stormont <astormont@racktopsystems.com>

We do not have libfakekernel, but need to reduce code divergence.
2018-03-28 20:41:15 +00:00
Conrad Meyer
2f68cdb944 ctfconvert: Fix minor memory leaks in STABS parser
In an error case, free leaked objects.  Does anything use STABS anymore?
Probably not.

Reported by:	Coverity
Sponsored by:	Dell EMC Isilon
2018-03-27 22:49:06 +00:00
Conrad Meyer
52f72944b8 ctfconvert/ctfmerge: Fix a memory leak enumerating DWARF files
Reported by:	Coverity
Sponsored by:	Dell EMC Isilon
2018-03-26 23:20:37 +00:00
Conrad Meyer
f5147e312f libctf: Don't construct pointers to out of bounds array offsets
Just attempting to do the pointer arithmetic is undefined behavior.

No functional change intended.

Reported by:	Coverity
Sponsored by:	Dell EMC Isilon
2018-03-26 22:02:36 +00:00
Conrad Meyer
e796cc77c5 libctf: Appease Coverity overrun warnings
Rather than zeroing and reading into the a smaller union member the full
union size, just zero and read directly into the union.

No functional change intended.

Reported by:	Coverity
Sponsored by:	Dell EMC Isilon
2018-03-26 21:57:44 +00:00
Alexander Motin
e76e77a972 MFV r331407: 9213 zfs: sytem typo
illumos/illumos-gate@edc8ef7d92

Reviewed by: C Fraire <cfraire@me.com>
Reviewed by: Andy Fiddaman <omnios@citrus-it.co.uk>
Approved by: Joshua M. Clulow <josh@sysmgr.org>
Author: Toomas Soome <tsoome@me.com>
2018-03-23 02:30:29 +00:00
Alexander Motin
b8436536c9 MFV r331400: 8484 Implement aggregate sum and use for arc counters
In pursuit of improving performance on multi-core systems, we should
implements fanned out counters and use them to improve the performance of
some of the arc statistics. These stats are updated extremely frequently,
and can consume a significant amount of CPU time.

Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: Paul Dagnelie <pcd@delphix.com>
2018-03-23 02:15:05 +00:00