skra [Tue, 10 Nov 2015 13:20:21 +0000 (13:20 +0000)]
Fix cp15 PAR definition and function. While here, add cp15 ATS1CPW
function which checks an address for privileged (PL1) write access.
The function is inlined so it does not bring any cost, but makes
function set for checking privileged access complete.
hselasky [Tue, 10 Nov 2015 12:20:22 +0000 (12:20 +0000)]
Add mlx5 and mlx5en driver(s) for ConnectX-4 and ConnectX-4LX cards
from Mellanox Technologies. The current driver supports ethernet
speeds up to and including 100 GBit/s. Infiniband support will be
done later.
The code added is not compiled by default, which will be done by a
separate commit.
mmel [Tue, 10 Nov 2015 11:45:41 +0000 (11:45 +0000)]
ARM: Improve robustness of locore_v6.S and fix errors.
- boot page table is not allocated in data section, so must be
cleared before use
- map only one section (1 MB) for SOCDEV mapping (*)
- DSB must be used for ensuring of finishing TLB operations
- Invalidate BTB when appropriate
ngie [Tue, 10 Nov 2015 10:59:40 +0000 (10:59 +0000)]
- Move the testing entries up for netbsd-tests/pjdfstest
- Add pjd to contrib/pjdfstest
- Add atf to the list; add jmmv
- Add tests
- Add share/mk/*.test.mk
hselasky [Tue, 10 Nov 2015 09:27:52 +0000 (09:27 +0000)]
Update the wsp driver to support newer touch pads, like found in
MacBookPro11,4 and MacBook12,1. This update adds support for the
force touch parameter.
bapt [Tue, 10 Nov 2015 08:11:27 +0000 (08:11 +0000)]
return "US-ASCII" instead of "POSIX" for "C" and "POSIX" locales
as it used to be in previous version of the locales. Returning
"POSIX" has too many fallouts.
np [Tue, 10 Nov 2015 02:29:19 +0000 (02:29 +0000)]
cxgbe/t4_tom: add a knob to the default configuration file to tune
the TOE for LAN operation. It is possible to set this to other values
(cluster for networks with little loss and really tight RTTs, and wan
for relatively large RTTs and/or lossy networks) depending on the
environment in which the TOE is being used.
None of this affects plain NIC operation in any way.
avos [Mon, 9 Nov 2015 23:46:13 +0000 (23:46 +0000)]
urtwn(4): refactor and fix TX path.
- Split urtwn_tx_start() into urtwn_tx_data() and urtwn_tx_start()
(the last will be used for beacon updates / raw xmit path).
- Remove unneeded code from _urtwn_getbuf().
- Use CCK11 for data frames in 11b mode.
- Send EAPOL frames at 1 Mbps.
- Reduce code duplication in urtwn_tx_data().
- Fix sequence numbering.
- Add IEEE80211_RADIOTAP_F_WEP flag for encrypted frames.
- Check URTWN_RUNNING flag under lock.
Tested with RTL8188EU, STA mode.
Reviewed by: kevlo
Approved by: adrian (mentor)
Differential Revision: https://reviews.freebsd.org/D4017
bdrewery [Mon, 9 Nov 2015 23:37:04 +0000 (23:37 +0000)]
FAST_DEPEND: Fix building of wrong source files in some cases.
Similar to the original reason for these dependency hints to be added,
in r124637, the missing-dependency file case can lead to building of the
wrong source.
A clear example of this is in gnu/lib/libstdc++ where the .PATH contains
both contrib/gcc and contrib/libstdc++/src.
contrib/gcc has a debug.c.
contrib/libstdc++/src has a debug.cc.
When building for the objects of debug.o, debug.So, and debug.po, it is
ambiguous for which src file to use due to the suffix transformation
rules, even though the proper one is listed first in .PATH.
This was normally avoided due to these dependency hints for the initial
build, and then mkdep would add an explicit 'debug.o: debug.cc'
dependency into the .depend file. WITH_FAST_DEPEND does not generate
the .depend file with these, but puts them into .depend.debug.o instead.
Rather than extending the exists() check to each object's .depend.*
file, just enable the hint when when using WITH_FAST_DEPEND. It fixes
the problem and seems to be safe enough to use since it is mapping SRCS
back to OBJS, rather than letting make make assumptions from OBJS to
SRCS.
A similar check mapping objects to headers is present in some mk files
but was not extended here for FAST_DEPEND since it has not yet been
found to be a problem.
bdrewery [Mon, 9 Nov 2015 23:36:57 +0000 (23:36 +0000)]
Fix 'make cleanworld' to respect TARGET/TARGET_ARCH for cross-build objtrees.
This simplifies the logic to always try removing the objdir if it exists
and to fallback on a 'cleandir' if no objdir exists. The reasoning for
this is to avoid rm -rf src/* (r126024)
bapt [Mon, 9 Nov 2015 22:06:22 +0000 (22:06 +0000)]
locales: Enforce US-ASCII encoding (limited to 7-bit)
The US-ASCII format was getting treated identically to POSIX. It is
supposed to throw an ILSEQ errno if a value of 0x80 or greater is
encountered, so let's bring back the "ASCII" handling.
While here, change nl_codeset to return US-ASCII only when the encoding
really is "US-ASCII". Before "C" and "POSIX" encoding returned this
string, so now they return "POSIX".
mav [Mon, 9 Nov 2015 18:33:36 +0000 (18:33 +0000)]
Introduce portal group options in ctl.conf.
While CTL has concept of port options, used at least for iSCSI ports now,
before this change it was impossible to set them manually. There still
no user-configurable port options now, but I am planning to change that.
cem [Mon, 9 Nov 2015 16:50:42 +0000 (16:50 +0000)]
linuxkpi/sysfs.h: Cast arg2 through intptr_t to avoid GCC warning
The code compiles fine under Clang, but GCC on PPC is less permissive about
integer and pointer sizes. (An intmax_t is clearly *large enough* to hold a
pointer value.)
adrian [Mon, 9 Nov 2015 15:59:42 +0000 (15:59 +0000)]
ath(4): begin fleshing out a "reset type" extension to force cold/warn resets.
Right now the only way to force a cold reset is:
* The HAL itself detects it's needed, or
* The sysctl, setting all resets to be cold.
Trouble is, cold resets take quite a bit longer than warm resets.
However, there are situations where a cold reset would be nice.
Specifically, after a stuck beacon, BB/MAC hang, stuck calibration results,
etc.
The vendor HAL has a separate method to set the reset reason (which is
how HAL_RESET_BBPANIC gets set) which informs the HAL during the reset path
why it occured. This is almost but not quite the same; I may eventually
unify both approaches in the future.
This commit just extends HAL_RESET_TYPE to include both status (eg BBPANIC)
and type (eg do COLD.) None of the HAL code uses it yet though; that'll
come later.
It also is a big no-op in each HAL - I need to go teach each of the HALs
about cold/warm reset through this path.
royger [Mon, 9 Nov 2015 12:22:44 +0000 (12:22 +0000)]
xen-blkfront: add support for unmapped IO
Using unmapped IO is really beneficial when running inside of a VM,
since it avoids IPIs to other vCPUs in order to invalidate the
mappings.
This patch adds unmapped IO support to blkfront. The following tests
results have been obtained when running on a Xen host without HAP:
PVHVM
3165.84 real 6354.17 user 4483.32 sys
PVHVM with unmapped IO
2099.46 real 4624.52 user 2967.38 sys
This is because when running using shadow page tables TLB flushes and
range invalidations are much more expensive, so using unmapped IO
provides a very important performance boost.
The implementation of bus_dmamap_load_ma_triv currently calls
_bus_dmamap_load_phys on each page that is part of the passed in buffer.
Since each page is treated as an individual buffer, the resulting behaviour
is different from the behaviour of _bus_dmamap_load_buffer. This breaks
certain drivers, like Xen blkfront.
If an unmapped buffer of size 4096 that starts at offset 13 into the first
page is passed to the current _bus_dmamap_load_ma implementation (so the ma
array contains two pages), the result is that two segments are created, one
with a size of 4083 and the other with size 13 (because two independant
calls to _bus_dmamap_load_phys are performed, one for each physical page).
If the same is done with a mapped buffer and calling _bus_dmamap_load_buffer
the result is that only one segment is created, with a size of 4096.
This patch relegates the usage of bus_dmamap_load_ma_triv in x86 bounce
buffer code to drivers requesting BUS_DMA_KEEP_PG_OFFSET and implements
_bus_dmamap_load_ma so that it's behaviour is the same as the mapped version
(_bus_dmamap_load_buffer). This patch only modifies the x86 bounce buffer
code, other arches are left untouched.
melifaro [Mon, 9 Nov 2015 10:35:33 +0000 (10:35 +0000)]
Decompose arp_ifinit() into arp_add_ifa_lle() and arp_announce_ifaddr().
Rename arp_ifinit2() into arp_announce_ifaddr().
Eliminate zeroing ifa_rtrequest: it was used for calling arp_rtrequest()
which was responsible for handling route cloning requests. It became
obsolete since r186119 (L2/L3 split).
ngie [Mon, 9 Nov 2015 09:39:07 +0000 (09:39 +0000)]
Follow up to r290259 dealing with makefs(8)
- Don't use contractions (don't -> do not)
- Change "throw away" to "discard" when describing the -o keep-bad-images option
- Revert author e-mail split I brought over from NetBSD, effectively reverting
the change bapt made in r267668
has been left undisturbed. The target `make sh-tests` has been moved over
from tools/regression/lib/libc/gen/Makefile to
lib/libc/tests/gen/Makefile and made into a PHONY target
case2.0 and case3.0 test input generation isn't being done automatically.
This needs additional discussion.
ngie [Mon, 9 Nov 2015 02:29:38 +0000 (02:29 +0000)]
Reduce the Makefile snippet complexity a bit
- Set BINDIR to TESTSDIR globally (and subsequently, remove all
`${FILESGROUP}DIR` setting because BINDIR is set to `TESTSDIR`)
- Set MAN to "" globally, instead of per-PROG
marius [Mon, 9 Nov 2015 00:19:04 +0000 (00:19 +0000)]
- Although it doesn't make a whole lot of sense to enable RX and TX
before their initial configuration is done, it turns out that r281337
has the inverse effect on some older chips. Moreover, as with newer
chips before, two chips seemingly identical according to their MAC
revisions may behave differently in this regard, with most working
but a few not, making changes extremely hard to test.
Closer inspection of the corresponding Linux code suggests that RX
and TX should only be enabled after their initial configuration with
RTL8168G and later chips, i. e. RTL8106E{,US}, RTL8107E, as well as
RTL8168{EP,G,GU,H}, so limit the new code path to these. [1]
- Distinguish between RTL8168H and RTL8107E, with the latter being the
10/100-Mbit/s-only variant of the former.
- For MAC variants that can only do Fast Ethernet at a maximum, ensure
that we don't advertise Gigabit Ethernet speed.
- In re_stop(), do the inverse of re_init_locked() and enable RXDV
gate on RTL8168G and later chips again, matching what Linux does.
andreast [Sun, 8 Nov 2015 21:06:51 +0000 (21:06 +0000)]
Improve r290373, do a runtime check rather than a compile time switch. I
learned that the Power8 and the PS3 have a mix of OFW and FDT. Both have AIM
defined. But currently they are not affected. They have no I2C devices under
OFW.
This version was tested on a Quad G5 and build tested for armv6*.
skra [Sun, 8 Nov 2015 18:48:35 +0000 (18:48 +0000)]
Fix pagelist bus_dmamap_t map handling. Memory for pagelist is allocated
by bus_dmamem_alloc() which creates associated bus_dmamap_t map for us.
When this memory is freed by bus_dmamem_free(), the map is freed as well.
Thus there is no need to free it explicitly by bus_dmamap_destroy(),
which leads to double freeing.
trasz [Sun, 8 Nov 2015 17:33:48 +0000 (17:33 +0000)]
Userspace part of reroot support. This makes it possible to change
the root filesystem without full reboot, using "reboot -r". This can
be used to to eg. boot from a temporary md_image preloaded by loader(8),
setup an iSCSI session, and continue booting from rootfs mounted over
iSCSI.
Reviewed by: kib@, bapt@
MFC after: 1 month
Relnotes: yes
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D3693
tijl [Sun, 8 Nov 2015 14:26:50 +0000 (14:26 +0000)]
Since r289279 bufinit() uses mp_ncpus, but some architectures set this
variable during mp_start() which is too late. Move this to mp_setmaxid()
where other architectures set it and move x86 assertions to MI code.
hselasky [Sun, 8 Nov 2015 09:37:26 +0000 (09:37 +0000)]
Avoid using the bounce buffer when the source or destination buffer is
32-bits aligned. Merge the two bounce buffers into a single one. Some
rough tests showed that the DWC OTG throughput on RPI2 increased by
10% after this patch.
ngie [Sun, 8 Nov 2015 08:27:51 +0000 (08:27 +0000)]
printfloat_test and scanfloat_test need symbols from msun; these are automatically
provided on amd64, but not i386. Add libm to DPADD/LDADD to unbreak the i386
tinderbox
gonzo [Sun, 8 Nov 2015 04:15:52 +0000 (04:15 +0000)]
Fix some of WITNESS complaints and bootup lock by removing msg_avail
condvar/mutex. They're basically no-op because error is not propagated
up the call chain. We still report message failures in VC service callback
gonzo [Sun, 8 Nov 2015 03:53:31 +0000 (03:53 +0000)]
Fix locking for VCHI driver by matching sleepable/non-sleepable APIs:
- Emulate Linux mutex API using sx(9) locks with only exclusive operations
instead of mutex(9), in Linux mutexes are sleepable.
- Emulate Linux rwlock_t using rwlock(9) instead of sx(9). rwlock_t
in Linux are spin locks
markj [Sun, 8 Nov 2015 01:38:56 +0000 (01:38 +0000)]
- Consistently use PROC_ASSERT_HELD() to verify that a process' hold count
is non-zero.
- Include the process address in the PROC_ASSERT_HELD() and
PROC_ASSERT_NOT_HELD() assertion messages so that the corresponding
process can be found easily when debugging.
markj [Sun, 8 Nov 2015 01:36:18 +0000 (01:36 +0000)]
Ensure that deactivated pages that are not expected to be reused are
reclaimed in FIFO order by the pagedaemon. Previously we would enqueue
such pages at the head of the inactive queue, yielding a LIFO reclaim order.
bdrewery [Sun, 8 Nov 2015 00:50:18 +0000 (00:50 +0000)]
Add built-in ccache build support via WITH_CCACHE_BUILD option.
ccache is mostly beneficial for frequent builds where -DNO_CLEAN is not
used to achieve a safe pseudo-incremental build. This is explained in
more detail upstream [1] [2]. It incurs about a 20%-28% hit to populate the
cache, but with a full cache saves 30-50% in build times. When combined with
the WITH_FAST_DEPEND feature it saves up to 65% since ccache does cache the
resulting dependency file, which it does not do when using mkdep(1)/'CC
-E'. Stats are provided at the end of this message.
This removes the need to modify /etc/make.conf with the CC:= and CXX:=
lines which conflicted with external compiler support [3] (causing the
bootstrap compiler to not be built which lead to obscure failures [4]),
incorrectly invoked ccache in various stages, required CCACHE_CPP2 to avoid
Clang errors with parenthesis, and did not work with META_MODE.
The option name was picked to match the existing option in ports. This
feature is available for both in-src and out-of-src builds that use
/usr/share/mk.
Linking, assembly compiles, and pre-processing avoid using ccache since it is
only overhead. ccache does nothing special in these modes, although there is
no harm in calling it for them.
CCACHE_COMPILERCHECK is set to 'content' when using the in-tree bootstrap
compiler to hash the content of the compiler binary to determine if it
should be a cache miss. For external compilers the 'mtime' option is used
as it is more efficient and likely to be correct. Future work may optimize the
'content' check using the same checks as whether a bootstrap compiler is needed
to be built.
The CCACHE_CPP2 pessimization is currently default in our devel/ccache
port due to Clang requiring it. Clang's -Wparentheses-equality,
-Wtautological-compare, and -Wself-assign warnings do not mix well with
compiling already-pre-processed code that may have expanded macros that
trigger the warnings. GCC has so far not had this issue so it is allowed to
disable the CCACHE_CPP2 default in our port.
Sharing a cache between multiple checkouts, or systems, is explained in
the ccache manual. Sharing a cache over NFS would likely not be worth
it, but syncing cache directories between systems may be useful for an
organization. There is also a memcached backend available [5]. Due to using
an object directory outside of the source directory though you will need to
ensure that both are in the same prefix and all users use the same layout. A
possible working layout is as follows:
Source: /some/prefix/src1
Source: /some/prefix/src2
Source: /some/prefix/src3
Objdir: /some/prefix/obj
Environment: CCACHE_BASEDIR='${SRCTOP:H}' MAKEOBJDIRPREFIX='${SRCTOP:H}/obj'
This will use src*/../obj as the MAKEOBJDIRPREFIX and tells ccache to replace
all absolute paths to be relative. Using something like this is required due
to -I and -o flags containing both SRC and OBJDIR absolute paths that ccache
adds into its hash for the object without CCACHE_BASEDIR.
distcc can be hooked into by setting CCACHE_PREFIX=/usr/local/bin/distcc.
I have not personally tested this and assume it will not mix well with
using the bootstrap compiler.
The cache from buildworld can be reused in a subdir by first running
'make buildenv' (from r290424).
Note that the cache is currently different depending on whether -j is
used or not due to ccache enabling -fdiagnostics-color automatically if
stderr is a TTY, which bmake only does if not using -j.
The system I used for testing was:
WITNESS
Build options: -j20 WITH_LLDB=yes WITH_DEBUG_FILES=yes WITH_CCACHE_BUILD=yes
DISK: ZFS 3-way mirror with very slow disks using SSD l2arc/log.
The arc was fully populated with src tree files and ccache objects.
RAM: 76GiB
CPU: Intel(R) Xeon(R) CPU L5520 @2.27GHz
2 package(s) x 4 core(s) x 2 SMT threads = hw.ncpu=16
The WITH_FAST_DEPEND feature was used for comparison here as well to show
the dramatic time savings with a full cache.
buildworld:
x buildworld-before
+ buildworld-ccache-empty
* buildworld-ccache-full
% buildworld-ccache-full-fastdep
# buildworld-fastdep
+-------------------------------------------------------------------------------+
|% * # +|
|% * # +|
|% * # xxx +|
| |A |
| A|
| A |
|A |
| A |
+-------------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 3 3744.13 3794.31 3752.25 3763.5633 26.935139
+ 3 4519 4525.04 4520.73 4521.59 3.1104823
Difference at 95.0% confidence
758.027 +/- 43.4565
20.1412% +/- 1.15466%
(Student's t, pooled s = 19.1726)
* 3 1823.08 1827.2 1825.62 1825.3 2.0785572
Difference at 95.0% confidence
-1938.26 +/- 43.298
-51.5007% +/- 1.15045%
(Student's t, pooled s = 19.1026)
% 3 1266.96 1279.37 1270.47 1272.2667 6.3971113
Difference at 95.0% confidence
-2491.3 +/- 44.3704
-66.1952% +/- 1.17895%
(Student's t, pooled s = 19.5758)
# 3 3153.34 3155.16 3154.2 3154.2333 0.91045776
Difference at 95.0% confidence
-609.33 +/- 43.1943
-16.1902% +/- 1.1477%
(Student's t, pooled s = 19.0569)
buildkernel:
x buildkernel-before
+ buildkernel-ccache-empty
* buildkernel-ccache-empty-fastdep
% buildkernel-ccache-full
# buildkernel-ccache-full-fastdep
@ buildkernel-fastdep
+-------------------------------------------------------------------------------+
|# @ % * |
|# @ % * x + |
|# @ % * xx ++|
| MA |
| MA|
| A |
| A |
|A |
| A |
+-------------------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 3 571.57 573.94 571.79 572.43333 1.3094401
+ 3 727.97 731.91 728.06 729.31333 2.2492295
Difference at 95.0% confidence
156.88 +/- 4.17129
27.4058% +/- 0.728695%
(Student's t, pooled s = 1.84034)
* 3 527.1 528.29 528.08 527.82333 0.63516402
Difference at 95.0% confidence
-44.61 +/- 2.33254
-7.79305% +/- 0.407478%
(Student's t, pooled s = 1.02909)
% 3 400.4 401.05 400.62 400.69 0.3306055
Difference at 95.0% confidence
-171.743 +/- 2.16453
-30.0023% +/- 0.378128%
(Student's t, pooled s = 0.954969)
# 3 201.94 203.34 202.28 202.52 0.73020545
Difference at 95.0% confidence
-369.913 +/- 2.40293
-64.6212% +/- 0.419774%
(Student's t, pooled s = 1.06015)
@ 3 369.12 370.57 369.3 369.66333 0.79033748
Difference at 95.0% confidence
-202.77 +/- 2.45131
-35.4225% +/- 0.428227%
(Student's t, pooled s = 1.0815)
bdrewery [Sun, 8 Nov 2015 00:50:06 +0000 (00:50 +0000)]
Add a hack to workaround ZSH as BUILDENV_SHELL breaking CPUTYPE.
ZSH considers CPUTYPE a magic variable that will be the output of 'uname -m'
even if already set in environment when starting up.
The CPUTYPE?= check in Makefile.inc1 and supporting overriding CPUTYPE
manually in the buildenv shell make automatic workarounds too tricky
here. ZSH should really respect variables set in the environment before
trashing them.
r289932 accidentally broke the rule skip calculation. The address family
argument to PF_ANEQ() is now important, and because it was set to 0 the macro
always evaluated to false.
This resulted in incorrect skip values, which in turn broke the rule
evaluations.
imp [Sat, 7 Nov 2015 22:52:06 +0000 (22:52 +0000)]
Implement the phy-mode property for ate and macb. If it is set to
"rmii", use rmii mode for the MAC, otherwise use MII mode. The code is
somewhat duplicated between these drivers for this.
Also, add AT91RM9200 compatibility strings to the ate driver. In the
future, there's a good chance that ate will lose the MACB support and
only attach to the AT91RM9200 EMAC device since the macb works now
that RMII support has been added to it.
projects / people / liuw / freebsd.git / log