royger [Wed, 20 Aug 2014 14:57:20 +0000 (14:57 +0000)]
pci: make MSI(-X) enable and disable methods of the PCI bus
Make the functions pci_disable_msi, pci_enable_msi and pci_enable_msix
methods of the newbus PCI bus. This code should not include any
functional change.
kib [Wed, 20 Aug 2014 08:15:23 +0000 (08:15 +0000)]
Do not busy the UFS mount point inside VOP_RENAME(). The
kern_renameat() already starts write on the mp, which prevents
parallel unmount from proceed. Busying mp after vn_start_write()
deadlocks the unmount.
Reported and tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
kib [Wed, 20 Aug 2014 08:13:03 +0000 (08:13 +0000)]
Correct the test for condition to suspend UFS filesystem during
unmount. There is no need to suspend read-only filesystem, while we
need suspension on modificable mount point.
Reported by: rwatson
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
kib [Wed, 20 Aug 2014 08:02:38 +0000 (08:02 +0000)]
Add arch-specific macro SFBUF_PHYS_DMAP(), which should translate the
physical address of the page to direct map address, in case
SFBUF_OPTIONAL_DIRECT_MAP returns true. The case of PowerPC AIM
64bit, where the page physical address is identical to the direct map
address, is accidental.
Reviewed by: alc
Sponsored by: The FreeBSD Foundation
ian [Wed, 20 Aug 2014 00:33:37 +0000 (00:33 +0000)]
Don't stop other legs of a parallel build due to a failure in make_check.
The whole point is to see if there's any failure, which is handled by
building a newer version of make.
delphij [Wed, 20 Aug 2014 00:14:41 +0000 (00:14 +0000)]
Add zdb into rescue environment.
On amd64, this would increase the binary size by 1.1MiB and
make it possible to examine zpool status offline, useful for
recovery and diagnostic purposes.
Submitted by: sef
Obtained from: FreeNAS
MFC after: 2 weeks
bdrewery [Tue, 19 Aug 2014 21:04:31 +0000 (21:04 +0000)]
Avoid showing stale errors when nmount(2) fails.
Sometimes nmount(2) will fail without setting errmsg. The previous (ignored)
error would then be shown as the reason for the failed call if the next
nmount(2) also fails without [ENOENT,ENOTSUP].
An example is when there is a tmpfs mounted with -o size. vfs_filteropt() adds
'size' as an error in errmsg due to 'size' not being in tmpfs_updateopts. Then
tmpfs_mount returns [ENOTSUP] from nmount(2), which is then ignored. The next
call may race with an unmount causing an invalid [EINVAL] that then does log an
error, with the tmpfs errmsg.
The race itself is a separate issue to fix as it is expected to have an
[ENOENT] returned instead.
In this example the mount being shown is actually nullfs, not tmpfs that the
error is from.
mountd[740]: can't delete exports for /poudriere/data/.m/exp-head-commit-test-devel/04/.npkg: Invalid argument mount option <size> is unknown
It should only show:
mountd[740]: can't delete exports for /poudriere/data/.m/exp-head-commit-test-devel/04/.npkg: Invalid argument
bdrewery [Tue, 19 Aug 2014 15:30:56 +0000 (15:30 +0000)]
Use bsd.lib.mk here as all other csu Makefiles do.
This effectively reverts r124752.
There's no reason this should be different. It resulted in needing NO_PIE in
the original opt-out NO_PIE commit as this was not using the proper framework.
bdrewery [Tue, 19 Aug 2014 15:04:32 +0000 (15:04 +0000)]
Revert r267233 for now. PIE support needs to be reworked.
1. 50+% of NO_PIE use is fixed by adding -fPIC to INTERNALLIB and other
build-only utility libraries.
2. Another 40% is fixed by generating _pic.a variants of various libraries.
3. Some of the NO_PIE use is a bit absurd as it is disabling PIE (and ASLR)
where it never would work anyhow, such as csu or loader. This suggests
there may be better ways of adding support to the tree. Many of these
cases can be fixed such that -fPIE will work but there is really no
reason to have it in those cases.
4. Some of the uses are working around hacks done to some Makefiles that are
really building libraries but have been using bsd.prog.mk because the code
is cleaner. Had they been using bsd.lib.mk then NO_PIE would not have
been needed.
We likely do want to enable PIE by default (opt-out) for non-tree consumers
(such as ports). For in-tree though we probably want to only enable PIE
(opt-in) for common attack targets such as remote service daemons and setuid
utilities. This is also a great performance compromise since ASLR is expected
to reduce performance. As such it does not make sense to enable it in all
utilities such as ls(1) that have little benefit to having it enabled.
marcel [Mon, 18 Aug 2014 23:45:40 +0000 (23:45 +0000)]
For vendors like Juniper, extensibility for sockets is important. A
good example is socket options that aren't necessarily generic. To
this end, OSD is added to the socket structure and hooks are defined
for key operations on sockets. These are:
o soalloc() and sodealloc()
o Get and set socket options
o Socket related kevent filters.
One aspect about hhook that appears to be not fully baked is the return
semantics (the return value from the hook is ignored in hhook_run_hooks()
at the time of commit). To support return values, the socket_hhook_data
structure contains a 'status' field to hold return values.
Submitted by: Anuranjan Shukla <anshukla@juniper.net>
Obtained from: Juniper Networks, Inc.
imp [Mon, 18 Aug 2014 21:04:44 +0000 (21:04 +0000)]
Create the native-xtools target. This target creates only the cross
building toolchain for the host computer. This toolchain produces
TARGET_ARCH and assumes the rest of the system contains libraries for
the target. It is intended to be used in a "qemu-user jail" where all
the binaries would otherwise be the target architecture's to build
ports. However, emulation of the compilers is too slow, so we build
native binaries for that. Rather than use the xdev produced binaries,
with all their weird links and paths, these binaries use the native
paths. They will not work unless installed into the qemu-user jail.
se [Mon, 18 Aug 2014 20:55:11 +0000 (20:55 +0000)]
Minor fixes to convert-keymap.pl (conversion of 8bit characters in the form
'x' with high bit set) and to KBDFILES.map (fix encodings and comment out a
few redundant keymap files).
alc [Mon, 18 Aug 2014 20:28:08 +0000 (20:28 +0000)]
There exists a possible sequence of page table page allocation failures
starting with a superpage demotion by pmap_enter() that could result in
a PV list lock being held when pmap_enter() is just about to return
KERN_RESOURCE_SHORTAGE. Consequently, the KASSERT that no PV list locks
are held needs to be replaced with a conditional unlock.
This adds initial support for PPC32 ELF PIC (Position Independent Code; the
-fPIC variety), thus rectifying a long-standing deficiency in the PowerPC
backend.
hselasky [Mon, 18 Aug 2014 14:30:43 +0000 (14:30 +0000)]
Use the "bSubslotSize" and "bSubFrameSize" fields to obtain the actual
sample size. According to the USB audio frame format specification
from USB.org, the value in the "bBitResolution" field can be less than
the actual sample size, depending on the actual hardware, and should
not be used for this computation.
se [Mon, 18 Aug 2014 09:40:19 +0000 (09:40 +0000)]
Import the tools used to convert the keymap files from SYSCONS (in locale
dependent encoding) to NEWCONS (Unicode).
The file "LANG.map" is used to convert INDEX.keymaps. It has 3 columns:
- the language ID as used in the source file
- the language ID to be used in the generated file (e.g. "iw" -> "he")
- the encoding of the menu texts for this language
The conversion result is written to STDOUT.
The file "KBDFILES.map" is used to batch convert keymap files. It's
columns are:
- the encoding used for the keymap sounce file
- the name of the source file
- the name of the generated file
The output files are created in the TEMP sub-directory of the vt keymap
directory, in order to preserve (possibly uncommitted) keymap files in
/usr/src/share/vt/keymaps.
The convert-keymap.pl script can be directly executed by passing the
source file name and the encoding on the command line. It writes to
STDOUT and generates hex Unicode codepoints by default. (This can be
changed to decimal in the script.)
While written for the one-time conversion of the SYSCONS keymaps into
the format required for NEWCONS, I think these tools may be useful for
easy conversion of possible further SYSCONS keymap files, that have not
been committed to the source tree.
imp [Mon, 18 Aug 2014 02:45:06 +0000 (02:45 +0000)]
/usr/libexec/ld.so.1 never was a thing on FreeBSD/arm. This was the
FreeBSD 3.x and 4.x run time linker. FreeBSD/arm's first release was
5.0. Retire this long-dead code.
imp [Mon, 18 Aug 2014 02:44:56 +0000 (02:44 +0000)]
Expand the elf brandelf infrastructure to give access to the whole ELF
header (Elf_Ehdr) to determine if a particular interpretor wants to
accept it or not. Use this mechanism to filter EABI arm on OABI arm
kernels, and vice versa. This method could also be used to implement
OABI on EABI arm kernels, if desired, or to allow a single mips kernel
to run o32, n32 and n64 binaries.
se [Sun, 17 Aug 2014 19:54:21 +0000 (19:54 +0000)]
Attempt at converting the SYSCONS keymaps to Unicode for use with NEWCONS.
I have spent many hours comparing source and destination formats, and hope
to have caught the most severe conversion errors.
Files were converted with a Perl script which I'll shortly commit to the
tools directory. This script is a much enhanced version of the one
provided by ray@ and is expected to support the full kbdmap(5) syntax.
The naming convention used is:
<2-letter country code>.<variant>.kbd
Only if there are multiple layouts for different languages:
<2-letter country code>-<2-letter language code>.<variant>.kbd
In nearly all cases, the keyboards are country specific, only. Currently
there is only one case where the language was added ("ch-fr.kbd" for
the Swiss-French keyboard layout).
I choose to write Unicode character codes as hex numbers. While this
increases the diff to the SYSCONS keymap files for the trivial cases
(conversion from ISO8859-1), it really helps to verify the more complex
cases against a Unicode table (which is indexed by hex numbers).
This commit does not cover all files that have been converted, since I
need to sort out which ones to use, if there were several with different
source encodings to choose from.
Review and test of the keymap files is highly desirable before 10.1 is
released. I'd also appreciate educated opinions regarding the optimum
variant (to be made available as the default for each language).
Since there are no NEWCONS keymaps in 10-STABLE, I plan to MFC after
the minimum allowed delay of 3 days, to allow at least a few weeks to
test and improve what will be in the next release.
se [Sun, 17 Aug 2014 11:59:23 +0000 (11:59 +0000)]
Add a few missing entries and fix entries that are obviously wrong.
The use of the old ISO language code "iw" for Hebrew was inconsistent
and it is replaced by the new language code "he" (which was already
used for the keyboard menu entry, but not for the menu heading or the
default font).
These changes are in preparation of the conversion of this file and
the keymap definitions to Unicode for use with NEWCONS.
trasz [Sun, 17 Aug 2014 09:44:42 +0000 (09:44 +0000)]
Bring in the new automounter, similar to what's provided in most other
UNIX systems, eg. MacOS X and Solaris. It uses Sun-compatible map format,
has proper kernel support, and LDAP integration.
There are still a few outstanding problems; they will be fixed shortly.
ian [Sun, 17 Aug 2014 02:56:58 +0000 (02:56 +0000)]
When the initarm_* routines were renamed to platform_* and moved to their
own header file, the lovely block of comments explaining what the generic
init code expects of the soc implementations got lost, restore it.
ian [Sun, 17 Aug 2014 02:53:36 +0000 (02:53 +0000)]
Rename the old initarm_* functions to the new platform_* names. Also
move the registration of the static device map table into the function
intended to do devmap init stuff.
markj [Sat, 16 Aug 2014 21:42:55 +0000 (21:42 +0000)]
Factor out the common code for function boundary tracing instead of
duplicating the entire implementation for both x86 and powerpc. This makes
it easier to add support for other architectures and has no functional
impact.
melifaro [Sat, 16 Aug 2014 19:13:52 +0000 (19:13 +0000)]
Add support for reading i2c SFP/SFP+ data from NIC driver and
presenting most interesting fields via ifconfig -v.
This version supports Intel ixgbe driver only.
Tested on: Cisco,Intel,Mellanox,ModuleTech,Molex transceivers
MFC after: 2 weeks
luigi [Sat, 16 Aug 2014 15:00:01 +0000 (15:00 +0000)]
Update to the current version of netmap.
Mostly bugfixes or features developed in the past 6 months,
so this is a 10.1 candidate.
Basically no user API changes (some bugfixes in sys/net/netmap_user.h).
In detail:
1. netmap support for virtio-net, including in netmap mode.
Under bhyve and with a netmap backend [2] we reach over 1Mpps
with standard APIs (e.g. libpcap), and 5-8 Mpps in netmap mode.
2. (kernel) add support for multiple memory allocators, so we can
better partition physical and virtual interfaces giving access
to separate users. The most visible effect is one additional
argument to the various kernel functions to compute buffer
addresses. All netmap-supported drivers are affected, but changes
are mechanical and trivial
3. (kernel) simplify the prototype for *txsync() and *rxsync()
driver methods. All netmap drivers affected, changes mostly mechanical.
4. add support for netmap-monitor ports. Think of it as a mirroring
port on a physical switch: a netmap monitor port replicates traffic
present on the main port. Restrictions apply. Drive carefully.
5. if_lem.c: support for various paravirtualization features,
experimental and disabled by default.
Most of these are described in our ANCS'13 paper [1].
Paravirtualized support in netmap mode is new, and beats the
numbers in the paper by a large factor (under qemu-kvm,
we measured gues-host throughput up to 10-12 Mpps).
A lot of refactoring and additional documentation in the files
in sys/dev/netmap, but apart from #2 and #3 above, almost nothing
of this stuff is visible to other kernel parts.
Example programs in tools/tools/netmap have been updated with bugfixes
and to support more of the existing features.
This is meant to go into 10.1 so we plan an MFC before the Aug.22 deadline.
A lot of this code has been contributed by my colleagues at UNIPI,
including Giuseppe Lettieri, Vincenzo Maffione, Stefano Garzarella.
peter [Sat, 16 Aug 2014 14:56:11 +0000 (14:56 +0000)]
Use mount protocol version 3 by default for showmount and umount.
mount_nfs effectively uses mount protocol v3 by default already.
v1 mount protocol is being removed along with nfsv2 by a high profile NFS
appliance vendor and our legacy v1 mount protocol usage causes rpc errors.
royger [Sat, 16 Aug 2014 10:47:24 +0000 (10:47 +0000)]
net: move interface removal notification up in if_detach_internal
This is needed to prevent having interfaces with ifp->if_addr == NULL
on bridge interfaces. Moving the notification event handlers up makes
sure the interfaces are removed before doing any more cleanup.
jilles [Fri, 15 Aug 2014 22:36:41 +0000 (22:36 +0000)]
sh: Mask off shift distance (<< and >>) in arithmetic.
In C, shift distances equal to or larger than the number of bits in the
operand result in undefined behaviour. As part of eliminating undefined
behaviour in arithmetic, mask off the distance like Java and JavaScript
specify and C on x86 usually does.
Assumption: conversion from unsigned to signed retains the two's complement
bits.
Assumption: uintmax_t has no padding bits.
ngie [Fri, 15 Aug 2014 21:22:49 +0000 (21:22 +0000)]
Fix atmconfig compilation when MK_ATM == yes and MK_BSNMP == no
Makefile.inc1:
Always compile gensnmptree with bootstrap-tools when MK_BSNMP != no
instead of depending on a potentially stale tool installed on the build host
sbin/atm/atmconfig/Makefile:
- Always remove oid.h to avoid cluttering up the build/src tree.
- Consolidate all of the RESCUE/MK_BSNMP != no logic under one
conditional to improve readability
- Remove unnecessary ${.OBJDIR} prefixing for oid.h and use ${.TARGET} instead
of spelling out oid.h
- Add a missing DPADD for ${LIBCRYPTO} when compiled MK_BSNMP == yes and
MK_OPENSSL == yes and not compiling for /rescue/rescue
sbin/atm/atmconfig/main.c:
Change #ifndef RESCUE to #ifdef WITH_BSNMP in main.c to make it
clear that we're compiling bsnmp support into atmconfig
imp [Fri, 15 Aug 2014 16:08:52 +0000 (16:08 +0000)]
Implement the FDT static pinctl/pinmux spec for Atmel. This will
configure the mux and config registers for PIO devices based on what
we find in the FDT. I developed it per the spec that had been
committed to Linux in the January 2014 time frame and haven't
updated. In short, bundles of pins are activated in specific ways for
specific configurations, and we implement all of that.
What's not included is a MI device infrastructure, any dynamic
run-time changing of these pins, etc. Also not included are hooks into
all the drivers to enable the latter (static at boot no driver changes
are needed). These larger questions will need to be answered once we
have more drivers like this for more platforms, or somebody has a heck
of a lot of time to research a bunch of platforms, the Linux solution
(which is good, but has its warts), etc.
kib [Fri, 15 Aug 2014 07:30:14 +0000 (07:30 +0000)]
Implement 'fast path' for the vm page fault handler. Or, it could be
called a scalable path. When several preconditions hold, the vm
object lock for the object containing the faulted page is taken in
read mode, instead of write, which allows parallel faults processing
in the region.
Namely, the fast path is taken when the faulted page already exists
and does not need copy on write, is already fully valid, and not busy.
For technical reasons, fast path is avoided when the fault is the
first write on the vnode object, or when the fault is for wiring or
debugger read or write.
On the fast path, pmap_enter(9) is passed the PMAP_ENTER_NOSLEEP flag,
since object lock is kept. Pmap might fail to create the entry, in
which case the fallback to slow path is performed.
Reviewed by: alc
Tested by: pho (previous version)
Hardware provided and hosted by: The FreeBSD Foundation and
Sentex Data Communications
Sponsored by: The FreeBSD Foundation
MFC after: 2 week
asomers [Thu, 14 Aug 2014 22:33:56 +0000 (22:33 +0000)]
Convert devd's client socket to type SOCK_SEQPACKET.
This change consists of two merges from projects/zfsd/head along with the
addition of an ATF test case for the new functionality.
sbin/devd/tests/Makefile
sbin/devd/tests/client_test.c
Add ATF test cases for reading events from both devd socket types.
r266519:
sbin/devd/devd.8
sbin/devd/devd.cc
Create a new socket, of type SOCK_SEQPACKET, for communicating with
clients. SOCK_SEQPACKET sockets preserve record boundaries,
simplying code in the client. The old SOCK_STREAM socket is retained
for backwards-compatibility with existing clients.
r269993:
sbin/devd/devd.8
Fix grammar bug.
CR: https://reviews.freebsd.org/rS266519
MFC after: 5 days
Sponsored by: Spectra Logic