Don't perform the acpi_DeviceIsPresent() check for PCI-PCI bridges. If
we are probing a PCI-PCI bridge it is because we found one by enumerating
the devices on a PCI bus, so the bridge is definitely present. A few
BIOSes report incorrect status (_STA) for some bridges that claimed they
were not present when in fact they were.
While here, move this check earlier for Host-PCI bridges so attach fails
before doing any work that needs to be torn down.
peter [Wed, 3 Jul 2013 07:03:19 +0000 (07:03 +0000)]
Replace the #define for "iconv" so it is for the function name instead of
a macro with parameters. Remove a __DECONST hack and add consts instead
for gnu libiconv API compatability. This makes it work with things like
devel/boost-libs that expects to use "iconv" as though it were a pointer.
A problem with the old NFS client where large writes to large files
would sometimes result in a corrupted file was reported via email.
This problem appears to have been caused by r251719 (reverting
r251719 fixed the problem). Although I have not been able to
reproduce this problem, I suspect it is caused by another thread
increasing np->n_size after the mtx_unlock(&np->n_mtx) but before
the vnode_pager_setsize() call. Since the np->n_mtx mutex serializes
updates to np->n_size, doing the vnode_pager_setsize() with the
mutex locked appears to avoid the problem.
Unfortunately, vnode_pager_setsize() where the new size is smaller,
cannot be called with a mutex held.
This patch returns the semantics to be close to pre-r251719 (actually
pre-r248567, r248581, r248567 for the new client) such that the call to
vnode_pager_setsize() is only delayed until after the mutex is
unlocked when np->n_size is shrinking. Since the file is growing
when being written, I believe this will fix the corruption.
A better solution might be to replace the mutex with a sleep lock,
but that is a non-trivial conversion, so this fix is hoped to be
sufficient in the meantime.
Reported by: David G. Lawrence (dg@dglawrence.com)
Tested by: David G. Lawrence (to be done soon)
Reviewed by: kib
MFC after: 1 week
Make it so that 'patch < FUBAR' and 'patch -i FUBAR' operate the same.
The former makes a copy of stdin, but was not accurately putting the
content of stdin into a temp file. This lead to the undercounting
the number of lines in hunks containing NUL characters when reading
from stdin. Thus resulting in "unexpected end of file in patch" errors.
- Allow ND6_IFF_AUTO_LINKLOCAL for IFT_BRIDGE. An interface with IFT_BRIDGE
is initialized with !ND6_IFF_AUTO_LINKLOCAL && !ND6_IFF_ACCEPT_RTADV
regardless of net.inet6.ip6.accept_rtadv and net.inet6.ip6.auto_linklocal.
To configure an autoconfigured link-local address (RFC 4862), the
following rc.conf(5) configuration can be used:
ifconfig_bridge0_ipv6="inet6 auto_linklocal"
- if_bridge(4) now removes IPv6 addresses on a member interface to be
added when the parent interface or one of the existing member
interfaces has an IPv6 address. if_bridge(4) merges each link-local
scope zone which the member interfaces form respectively, so it causes
address scope violation. Removal of the IPv6 addresses prevents it.
- if_lagg(4) now removes IPv6 addresses on a member interfaces
unconditionally.
- Set reasonable flags to non-IPv6-capable interfaces. [*]
Fix a panic when leaving MC group in a kernel with VIMAGE enabled.
in_leavegroup() is called from an asynchronous task, and
igmp_change_state() requires that curvnet is set by the caller.
When acquiring a lease, record the value of the BOOTP siaddr field
contained in the DHCP offer, and write it out to the lease file
as an unquoted value of the "next-server" keyword. The value is ignored
when the lease is read back by dhclient, however other applications
are free to parse it.
The intent behind this change is to allow easier interoperability
with automated installation systems e.g. Cobbler, Foreman, Razor;
FreeBSD installation kernels can automatically probe the network
to discover deployment servers. There are no plans to MFC this
change unless a backport is specifically requested.
The syntax of the "next-server <ip>" lease keyword is intended to be
identical to that used by the ISC DHCPD server in its configuration files.
The required defines are already present in dhclient but were unused before
this change. (Note: This is NOT the same as Option 66, tftp-server-name).
It has been exercised in a university protocol testbed environment, with
Cobbler and an mfsBSD image containing pc-sysinstall (driven by Cobbler
Cheetah templates). The SYSLINUX memdisk driver is used to boot mfsBSD.
Currently this approach requires that a dedicated system profile has
been created for the node where FreeBSD is to be deployed. If this
is not present, the pc-sysinstall wrapper will be unable to obtain
a node configuration. There is code in progress to allow mfsBSD images
to obtain the required hints from the memdisk environment by parsing
the MBFT ACPI chunk. This is non-standard as it is not linked into
the platform's ACPI RSDT.
- Update newvers.sh to include svn revision in uname(1) if the
system has svnliteversion.
- If svnliteversion is not found, look for svnversion in /usr/bin
and /usr/local/bin, since svnlite can be installed as svn if
WITH_SVN is set.[1]
- Remove /bin from binary search paths.[1]
Import an implementation of the CAIA Delay-Gradient (CDG) congestion control
algorithm, which is based on the 2011 v0.1 patch release and described in the
paper "Revisiting TCP Congestion Control using Delay Gradients" by David Hayes
and Grenville Armitage. It is implemented as a kernel module compatible with the
modular congestion control framework.
CDG is a hybrid congestion control algorithm which reacts to both packet loss
and inferred queuing delay. It attempts to operate as a delay-based algorithm
where possible, but utilises heuristics to detect loss-based TCP cross traffic
and will compete effectively as required. CDG is therefore incrementally
deployable and suitable for use on shared networks.
In collaboration with: David Hayes <david.hayes at ieee.org> and
Grenville Armitage <garmitage at swin edu au>
MFC after: 4 days
Sponsored by: Cisco University Research Program and FreeBSD Foundation
andrew [Tue, 2 Jul 2013 08:04:41 +0000 (08:04 +0000)]
Work around an ARM EABI issue where clang would sometimes incorrectly align
the stack in a leaf function that uses TLS.
The issue is, when using TLS, the function is no longer a leaf as it calls
__aeabi_read_tp. With statically linked programs this is not an issue as
it doesn't make use of the stack, however with dynamically linked
applications we enter rtld which does use the stack and makes assumptions
about it's alignment.
This is only a temporary fix until a better patch can be made and submitted
upstream.
Revert the simplification of the i_gen calculation.
It is still a good idea to avoid zero values and for the case
of old filesystems there is probably no advantage in using
the complete 32 bits anyways.
Document the fact that an NFSv4 mount uses the host uuid to uniquely
identify the client to the server. As such, NFSv4 mounts
will break if host_enable="NO" is specified.
This is a content change.
Add a SIGINFO handler to devd. It will send useful statistics to syslog or
stderr as appropriate. Currently, the only statistic printed is the number of
events received.
sbin/devd/devd.cc
All output will now go to syslog(3) if devd is daemonized, or stderr
if it's running in the foreground.
sbin/devd/devd.8
Remove the "-D" flag. Filtering messages by priority now
happens in the usual syslog way. For performance reasons, a few
extra-verbose debugging statements are now conditional on the "-d" (do
not daemonize) flag.
etc/syslog.conf
etc/newsyslog.conf
Direct messages from devd(8) to /var/log/devd.log, but leave it
disabled by default
Reviewed by: eadler
Approved by: gibbs (co-mentor)
MFC after: never (removed a command-line option from devd)
A recent version of the oldnfs NFS client in head/current
will crash when doing a large write, since m_get2() would
return NULL. This patch fixes the problem, since nfsm_uiotombuf()
will allocate additional mbufs, as required.
Remove forced timeout of in-flight commands from mfi_timeout.
While this prevents commands getting stuck forever there is no way to guarantee
that data from the command hasn't been committed to the device.
In addition older mfi firmware has a bug that would cause the controller to
frequently stall IO for over our timeout value, which when combined with
a forced timeout often resulted in panics in UFS; which would otherwise be
avoided when the command eventually completed if left alone.
For reference this timeout issue is resolved in Dell FW package 21.2.1-0000.
Fixed FW package version for none Dell controller will likely vary.
Further simplify the i_gen calculation for older disks.
Having a zero here is not really a problem and this is more
similar to what is done in newfs_random().
peter [Mon, 1 Jul 2013 08:06:26 +0000 (08:06 +0000)]
__weak_alias() doesn't exist on FreeBSD. Use __weak_reference();
Expose iconv functions as weak symbols as well as their internal
remapped #define names. This is necessary for autoconf compatability -
on Linux it appears that #include <iconv.h> isn't a link time
prerequisite for their version that's built into glibc.
Initialize the pthread rwlock. Note that upstream has three
separate locks. The file-local static lock appears intentional.
I'm using this as a ports-compatible compile-time substitute for
converters/libiconv on one of my personal machines.
Don't assume that UFS on-disk format of a directory is the same as
defined by <sys/dirent.h>
Always start parsing at DIRBLKSIZ aligned offset, skip first entries if
uio_offset is not DIRBLKSIZ aligned. Return EINVAL if buffer is too
small for single entry.
Preallocate buffer for cookies. Cookies will be replaced with d_off
field in struct dirent at later point.
Skip entries with zero inode number.
Stop mangling dirent in ufs_extattr_iterate_directory().
Reviewed by: kib
Sponsored by: Google Summer Of Code 2011
In UFS, i_gen is a random generated value and there is not way for
it to be negative. Actually, the value of i_gen is just used to
match bit patterns and it is of not consequence if the values are
signed or not.
Following other filesystems, set it to unsigned and use it as such,
Fix issues with zeroing and fetching the counters, on x86 and ppc64.
Issues were noted by Bruce Evans and are present on all architectures.
On i386, a counter fetch should use atomic read of 64bit value,
otherwise carry from the increment on other CPU could be lost for the
given fetch, making error of 2^32. If 64bit read (cmpxchg8b) is not
available on the machine, it cannot be SMP and it is enough to disable
preemption around read to avoid the split read.
On x86 the counter increment is not atomic on purpose, which makes it
possible for the store of the incremented result to override just
zeroed per-cpu slot. The effect would be a counter going off by
arbitrary value after zeroing. Perform the counter zeroing on the
same processor which does the increments, making the operations
mutually exclusive. On i386, same as for the fetching, if the
cmpxchg8b is not available, machine is not SMP and we disable
preemption for zeroing.
PowerPC64 is treated the same as amd64.
For other architectures, the changes made to allow the compilation to
succeed, without fixing the issues with zeroing or fetching. It
should be possible to handle them by using the 64bit loads and stores
atomic WRT preemption (assuming the architectures also converted from
using critical sections to proper asm). If architecture does not
provide the facility, using global (spin) mutex would be non-optimal
but working solution.
Noted by: bde
Sponsored by: The FreeBSD Foundation
ed [Sun, 30 Jun 2013 10:38:20 +0000 (10:38 +0000)]
Make atomic_fetch_add() and atomic_fetch_sub() work for pointers with GCC 4.2.
According to the standard, atomic_fetch_*() has to behave identical to
regular arithmetic. This means that for pointer types, we have to apply
the stride when doing addition/subtraction.
The GCC documentation seems to imply this is done for __sync_*() as
well. Unfortunately, both tests and Googling seems to reveal this is not
really the case. Fix this by performing the multiplication with the
stride manually.
ed [Sun, 30 Jun 2013 08:59:33 +0000 (08:59 +0000)]
Convert this piece of code to use C11 atomics.
As mentioned before, we should at least aim to have one piece of code in
both user space and kernel space that uses C11 atomics, to get some
coverage. This piece of code can be migrated trivially, so it's a good
candidate.
ed [Sun, 30 Jun 2013 08:54:41 +0000 (08:54 +0000)]
Make various fixes to <stdatomic.h>.
- According to the standard, memory_order is a type. Use a typedef.
- atomic_*_fence() and atomic_flag_*() are described by the standard as
functions. Use inline functions to implement them.
- Only expose the atomic_*_explicit() functions in kernel space. We
should not use the short-hand functions, as they will always use
memory_order_seq_cst.
yongari [Sun, 30 Jun 2013 05:12:18 +0000 (05:12 +0000)]
Fix triggering false watchdog timeout when controller is in PAUSE
state. Previously it used to check if controller has sent a
PAUSE frame to the remote peer.
Reported by: David Imhoff via Brad Smith <brad@OpenBSD.org>
Submitted by: davidch (initial version)
Reviewed by: davidch, David Imhoff <dimhoff_devel@xs4all.nl>