Commit Graph

181467 Commits

Author SHA1 Message Date
Devin Teske
790062d26b Improve debugging for f_dialog_*_constrain() to print constrained rather
than unconstrained values.
2013-06-04 00:30:05 +00:00
Simon J. Gerraty
5a37b9e138 bmake does not like universe_${target}_prologue in the .ORDER
when there is an explicit dependency for it anyway.
Remove it so both bmake and fmake do the same thing.

Reviewed by:	OBrien
2013-06-03 23:51:46 +00:00
Hiren Panchasara
15e45369fe Fixing a typo.
Approved by:	sbruno (mentor)
MFC after:	3 days
2013-06-03 22:22:53 +00:00
Xin LI
21d1182ecf Fix a typo: XPORT_SPI should be tested against transport, nor protocol.
Submitted by:	Sascha Wildner <swildner dragonflybsd org>
Reviewed by:	mjacob
MFC after:	2 weeks
2013-06-03 21:52:19 +00:00
Ed Schouten
6e8727211d Add ATOMIC_*_LOCK_FREE constants.
These constants are part of the C standard. Both Clang and GCC seem to
export these constants under the name __GCC_ATOMIC_*_LOCK_FREE.
2013-06-03 20:52:20 +00:00
Pedro F. Giffuni
532ebe1313 ext2fs: space vs tab.
Obtained from:	Christoph Mallon
MFC after:	3 days
2013-06-03 20:33:05 +00:00
Steve Kargl
1a287d1ddf Change a comma to a semicolon.
Remove a blank line that crept into the declarations.

Fix a comment to show a sign on a NaN.
2013-06-03 20:09:22 +00:00
Pedro F. Giffuni
fc3ea958b2 ext2fs: Small cosmetic fixes.
Make a long macro readable and sort a header.

Obtained from:	Christoph Mallon
MFC after:	3 days
2013-06-03 20:02:45 +00:00
Steve Kargl
3ffff4bad5 ld80 and ld128 implementations of expm1l(). This code started life
as a fairly faithful implementation of the algorithm found in

PTP Tang, "Table-driven implementation of the Expm1 function
in IEEE floating-point arithmetic," ACM Trans. Math. Soft., 18,
211-222 (1992).

Over the last 18-24 months, the code has under gone significant
optimization and testing.

Reviewed by:	bde
Obtained from:	bde (most of the optimizations)
2013-06-03 19:51:32 +00:00
Adrian Chadd
904e385eba Fix the order of TX shutdown and reset.
* Grab the reset lock first, so any subsequent interrupt, TX, RX work
  will fail

* Then shut down interrupts

* Then wait for TX/RX to finish running

At this point no further work will be running, so it's safe to do the
reset path code.

PR:		kern/179232
2013-06-03 19:39:37 +00:00
Steve Kargl
42e4111cab Fix two comments that got lost in the disentanglement of the larger diff. 2013-06-03 19:29:03 +00:00
Adrian Chadd
cc7b47dd1b Fix receive on the AR9285 (Kite) with only one antenna connected.
The main problem here is that fast and driver RX diversity isn't actually
configured; I need to figure out why that is.  That said, this makes
the single-antenna connected AR9285 and AR2427 (AR9285 w/ no 11n) work
correctly.

PR:		kern/179269
2013-06-03 19:14:29 +00:00
Steve Kargl
8cc74771f2 ld80/s_expl.c:
* Use integral numerical constants, and let the compiler do the
  conversion to long double.

ld128/s_expl.c:

* Use integral numerical constants, and let the compiler do the
  conversion to long double.
* Use the ENTERI/RETURNI macros, which are no-ops on ld128.  This
  however makes the ld80 and ld128 identical.

Reviewed by:	bde (as part of larger diff)
2013-06-03 19:13:44 +00:00
Steve Kargl
35cbca6a7f Micro-optimization: move the unary mius operator to operate
on a literal constant.

Obtained from:	bde
2013-06-03 18:57:35 +00:00
Steve Kargl
a3f70b4ed8 Add a comment to note that bde supplied most, if not all,
of the optimizations.
2013-06-03 18:53:40 +00:00
Pedro F. Giffuni
4f69a09308 ext2fs: Update Block Group Descriptor struct.
Uncover some, previously reserved, fields that are used by Ext4.
These are currently unused but it is good to have them for future
reference.

Reviewed by:	bde
MFC after:	3 days
2013-06-03 18:52:14 +00:00
Steve Kargl
1783063f18 ld80/s_expl.c:
* In the special case x = -Inf or -NaN, use a micro-optimization
  to eliminate the need to access u.xbits.man.

* Fix an off-by-one for small arguments |x| < 0x1p-65.

ld128/s_expl.c:

* In the special case x = -Inf or -NaN, use a micro-optimization
  to eliminate the need to access u.xbits.manh and u.xbits.manl.

* Fix an off-by-one for small arguments |x| < 0x1p-114.

Obtained from:	bde
2013-06-03 18:51:34 +00:00
Steve Kargl
31407861b8 ld80/s_expl.c:
* Update the evaluation of the polynomial.  This allows the removal
  of the now unused variables t23 and t45.

ld128/s_expl.c:

* Update the evaluation of the polynomial and the intermediate
  result t.  This update allows several numerical constants to be
  written as double rather than long double constants.   Update
  the constants as appropriate.

Obtained from:	bde
2013-06-03 18:40:00 +00:00
Steve Kargl
199b8e343d Rename a few P2, P3, ... coefficients to A2, A3, ... missed in
my previous commit.
2013-06-03 18:18:08 +00:00
Steve Kargl
f3049ab5f3 Update a comment to reflect that we are using an endpoint of
an interval instead of a midpoint.
2013-06-03 18:14:18 +00:00
Steve Kargl
ad36b00fcb Add a u suffix to the IEEEl2bits unions o_threshold and u_threshold,
and use macros to access the e component of the unions.  This allows
the portions of the code in ld80 to be identical to the ld128 code.

Obtained from:	bde
2013-06-03 18:07:04 +00:00
Steve Kargl
4aa8c9453f Introduce the macro LOG2_INTERVAL, which is log2(number of intervals).
Use the macroi as a micro-optimization to convert a subtraction and
division to a shift.

Obtained from:	bde
2013-06-03 17:51:08 +00:00
John Baldwin
24150d37d3 - Fix a couple of inverted panic messages for shared/exclusive mismatches
of a lock within a single thread.
- Fix handling of interlocks in WITNESS by properly requiring the interlock
  to be held exactly once if it is specified.
2013-06-03 17:41:11 +00:00
Steve Kargl
03e1315345 Whitespace. 2013-06-03 17:40:52 +00:00
Konstantin Belousov
9138579845 Assert that interrupts are enabled in the trap handlers on x86 before
calling generic code to deliver signals.

Discussed with:	bde
Tested by:	pho
MFC after:	1 week
2013-06-03 17:40:05 +00:00
John Baldwin
95d28652af - Handle the recursed/not recursed flags with RA_RLOCKED in rw_assert().
- Tweak a panic message.
2013-06-03 17:38:57 +00:00
Konstantin Belousov
d39116f5d5 Be more generous when donating the current thread time to the owner of
the vnode lock while iterating over the free vnode list.  Instead of
yielding, pause for 1 tick.  The change is reported to help in some
virtualized environments.

Submitted by:	Roger Pau Monn? <roger.pau@citrix.com>
Discussed with:	jilles
Tested by:	pho
MFC after:	2 weeks
2013-06-03 17:36:43 +00:00
Steve Kargl
bb23de67bb * Rename the polynomial coefficients from P2, P3, ... to A2, A3, ....
The names now coincide with the name used in PTP Tang's paper.

* Rename the variable from s to tbl to better reflect that
  this is a table, and to be consistent with the naming scheme
  in s_exp2l.c

Reviewed by:	bde (as part of larger diff)
2013-06-03 17:36:26 +00:00
Konstantin Belousov
be6ec55376 Remove irrelevant comments.
Discussed with:	alc
MFC after:	3 days
2013-06-03 17:30:40 +00:00
Navdeep Parhar
7bbe4d1f68 cxgbe(4): t4fw_cfg must be explicitly loaded if the driver is being
loaded via loader.conf.

Submitted by:	jwd@
MFC after:	3 days
2013-06-03 17:30:21 +00:00
Steve Kargl
b419a5506a * Style(9). Start non-Copyright fancy formatted comments with /**.
Reviewed by:	bde (as part of larger diff)
2013-06-03 17:24:46 +00:00
Steve Kargl
a1d69112c1 ld80/s_expl.c:
* Update Copyright years to include 2013.

ld128/s_expl.c:

* Correct and update Copyright years.  This code originated from
  the ld80 version, so it should reflect the same time period.

Reviewed by:	bde (as part of larger diff)
2013-06-03 17:21:43 +00:00
Ed Schouten
49111f0092 Add libiconv based versions of *c16*() and *c32*().
I initially thought wchar_t was locale independent, but this seems to be
only the case on Linux. This means that we cannot depend on the *wc*()
routines to implement *c16*() and *c32*(). Instead, use the Citrus
libiconv that is part of libc.

I'll see if there is anything I can do to make the existing functions
somewhat useful in case the system is built without libiconv in the
nearby future. If not, I'll simply remove the broken implementations.

Reviewed by:	jilles, gabor
2013-06-03 17:17:56 +00:00
Jason Evans
f8ca2db1f8 Update jemalloc to version 3.4.0. 2013-06-03 14:36:28 +00:00
Ed Maste
9bdb572a00 Switch to 2-clause license
Approved by:	luigi@
2013-06-03 13:11:48 +00:00
Andre Oppermann
33e0730e2d Specify a maximum TSO length limiting the segment chain to what the
Xen host side can handle after defragmentation.

This prevents the driver from throwing away too long TSO chains and
improves the performance on Amazon AWS instances with 10GigE virtual
interfaces to the normally expected throughput.

Submitted by:	cperciva (earlier version)
Reviewed by:	cperciva
Tested by:	cperciva
MFC after:	1 week
2013-06-03 13:00:33 +00:00
Andre Oppermann
3c914c547e Allow drivers to specify a maximum TSO length in bytes if they are
limited in the amount of data they can handle at once.

Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to
change the limit.

The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything
less wouldn't be very useful anymore.  The upper limit is still at
IP_MAXPACKET (65536 bytes).  Raising it requires further auditing of
the IPv4/v6 code path's as the length field in the IP header would
overflow leading to confusion in firewalls and others packet handler on
the real size of the packet.

The placement into "struct ifnet" is a bit hackish but the best place
that was found.  When the stack/driver boundary is updated it should
be handled in a better way.

Submitted by:	cperciva (earlier version)
Reviewed by:	cperciva
Tested by:	cperciva
MFC after:	1 week (using spare struct members to preserve ABI)
2013-06-03 12:55:13 +00:00
Ed Maste
acbbd07aca Switch to 2-clause license
Approved by:	bms@
2013-06-03 12:43:09 +00:00
David Schultz
0921e73008 Bump __FreeBSD_version for the addition of the following functions to libm:
cacos, cacosf, cacosh, cacoshf,
  casin, casinf, casinh, casinhf,
  catan, catanf, catanh, catanhf,
  logl, log2l, log10l, log1pl
I am hoping kargl@ will commit expl and expm1l soon, in which case this
bump will cover those, too.

Requested by:	danfe
2013-06-03 09:33:51 +00:00
David Schultz
8ffb5b838e Add more tests for log functions. A few are commented out because the
long double versions don't pass yet.  (They are rather nit-picky cases,
so there's ongoing discussion with Bruce about whether it is worth the
performance cost.)
2013-06-03 09:15:15 +00:00
David Schultz
25a4d6bfda Add logl, log2l, log10l, and log1pl.
Submitted by:	bde
2013-06-03 09:14:31 +00:00
Konstantin Belousov
1e65d73c74 Do not map the shared page COW. If the process wired its address
space, fork(2) would cause shadowing of the physical object and
copying of the shared page into private copy, effectively preventing
updates for the exported timehands structure and stopping the clock.

Specify the maximum allowed permissions for the page to be read and
execute, preventing write from the user mode.

Reported and tested by:	<huanghwh@yahoo.com>
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2013-06-03 04:32:53 +00:00
Konstantin Belousov
91ddaeb725 Since the cause of the problems with the __fillcontextx() was
identified, unify the code of check_deferred_signal() for all
architectures, making the variant under #ifdef x86 common.

Tested by:	marius (sparc64)
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2013-06-03 04:22:42 +00:00
Konstantin Belousov
07d46f9c18 MFamd64: when printing the trap information, show the %esp value.
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2013-06-03 04:19:21 +00:00
Konstantin Belousov
92fab43f7f When auto-sizing the buffer cache, limit the amount of physical memory
used as the estimation of size, to 32GB.  This provides around 100K of
buffer headers and corresponding KVA for buffer map at the peak.
Sizing the cache larger is not useful, also resulting in the wasting
and exhausting of KVA for large machines.

Reported and tested by:	bdrewery
Sponsored by:	The FreeBSD Foundation
2013-06-03 04:16:48 +00:00
Konstantin Belousov
aed1e745d0 The vm_page lock is not needed around the call to vm_page_insert().
Submitted by:	alc
MFC after:	1 week
2013-06-03 04:11:42 +00:00
Alan Cox
b417181250 Require that the page lock is held, instead of the object lock, when
clearing the page's PGA_REFERENCED flag.  Since we are typically
manipulating the page's act_count field when we are clearing its
PGA_REFERENCED flag, the page lock is already held everywhere that we clear
the PGA_REFERENCED flag.  So, in fact, this revision only changes some
comments and an assertion.  Nonetheless, it will enable later changes to
object locking in the pageout code.

Introduce vm_page_assert_locked(), which completely hides the implementation
details of the page lock from the caller, and use it in
vm_page_aflag_clear().  (The existing vm_page_lock_assert() could not be
used in vm_page_aflag_clear().)  Over the coming weeks, I expect that we'll
either eliminate or replace the various uses of vm_page_lock_assert() with
vm_page_assert_locked().

Reviewed by:	attilio
Sponsored by:	EMC / Isilon Storage Division
2013-06-03 01:22:54 +00:00
Brian Somers
6a3193a773 Move ppp.8.m4 back to ppp.8 and de-ifdef it.
Requested by: joel

MFC after:	1 week
2013-06-02 23:56:38 +00:00
Devin Teske
dd5cc066af Like r250701, introduce another handy function for truncating variables to a
specific byte-length. Works like vsnprintf(3).
2013-06-02 23:25:27 +00:00
Devin Teske
f8e121ed61 Fix a typo. In this case, the typo caused the f_noyes() invocation to
display a global instead of the passed-argument; however since the global
always has the same value as the passed argument, it made no difference in
the realtime operation.
2013-06-02 23:23:29 +00:00