marius [Sat, 28 Jan 2012 23:26:50 +0000 (23:26 +0000)]
MFC: r225891
Re-reading the Schizo errata suggests that it's actually tolerable to
also use the streaming buffer of pre version 5/revision 2.3 hardware as
long as we stay away from context flushes (which iommu(4) so far doesn't
take advantage of). OpenSolaris does the same.
marius [Sat, 28 Jan 2012 23:25:28 +0000 (23:25 +0000)]
MFC: r225890
- Add protective parentheses to macros as far as possible.
- Move {r,w,}mb() to the top of this file where they live on most of the
other architectures.
marius [Sat, 28 Jan 2012 23:24:03 +0000 (23:24 +0000)]
MFC: r225889, r228222
In total store which we use for running the kernel and all of the userland
atomic operations behave as if they were followed by a CPU memory barrier
so there's no need to include ones in the acquire variants of atomic(9) and
it's sufficient to just use include compiler memory barriers to satisfy
the requirements of atomic(9). Removing the CPU memory barriers results in
a small performance improvement, specifically this is sufficient to
compensate the performance loss seen in the worldstone benchmark seen when
using SCHED_ULE instead of SCHED_4BSD.
This change is inspired by Linux even more radically doing the equivalent
thing some time ago.
Thanks go to Peter Jeremy for additional testing.
marius [Sat, 28 Jan 2012 23:15:02 +0000 (23:15 +0000)]
MFC: r225886
- Right-justify backslashes as suggested by style(9).
- Rename ATOMIC_INC_ULONG to ATOMIC_INC_LONG in order to be consistent with
the names of the other macros in this file an adjust accordingly.
marius [Sat, 28 Jan 2012 23:12:55 +0000 (23:12 +0000)]
MFC: r228022, r228026
For sparc64 also adjust the geometry of da(4) driven disks to not overflow
the 16-bit cylinders field of the VTOC8 disk label (at around 502GB). The
geometry chosen for disks above that limit allows to use disks up to 2TB,
which is the limit of the extended VTOC8 format. The geometry used for
disks smaller than the 16-bit cylinders limit stays the same as used by
cam_calc_geometry(9) for extended translation.
Thanks to Hans-Joerg Sirtl for providing hardware for testing this change.
rmacklem [Sat, 28 Jan 2012 01:45:19 +0000 (01:45 +0000)]
MFC: r230100
Tai Horgan reported via email that there were two places in
the new NFSv4 server where the code follows the wrong list.
Fortunately, for these fairly rare cases, the lc_stateid[]
lists are normally empty. This patch fixes the code to
follow the correct list.
qingli [Fri, 27 Jan 2012 02:13:27 +0000 (02:13 +0000)]
MFC 227460
A default route learned from the RAs could be deleted manually
after its installation. This removal may be accidental and can
prevent the default route from being installed in the future if
the associated default router has the best preference. The cause
is the lack of status update in the default router on the state
of its route installation in the kernel FIB. This patch fixes
the described problem.
dumbbell [Thu, 26 Jan 2012 19:46:13 +0000 (19:46 +0000)]
MFC r228259:
Support domain-search in dhclient(8)
The "domain-search" option (option 119) allows a DHCP server to publish
a list of implicit domain suffixes used during name lookup. This option
is described in RFC 3397.
For instance, if the domain-search option says:
".example.org .example.com"
and one wants to resolve "foobar", the resolver will try:
1. "foobar.example.org"
2. "foobar.example.com"
The file /etc/resolv.conf is updated with a "search" directive if the
DHCP server provides "domain-search".
A regression test suite is included in this patch under
tools/regression/sbin/dhclient.
MFC r229000:
Invalid Domain Search option isn't considered as a fatal error
In the original Domain Search option patch, an invalid option value
would cause the whole lease to be rejected. However, DHCP servers who
emit such an invalid value are more common than I thought. With this new
patch, just the option is rejected, not the entire lease.
glebius [Wed, 25 Jan 2012 13:47:55 +0000 (13:47 +0000)]
Merge r230127 from head/:
Restore functionality to pack several kernels into release. All
kernels specified by KERNCONF are built and packed into release.
The first one is packed into kernel.txz, all others to
kernel.CONFIG.txz.
rmacklem [Wed, 25 Jan 2012 01:45:19 +0000 (01:45 +0000)]
MFC: r229956
jwd@ reported via email that the "CacheSize" field reported by "nfsstat -e -s"
would go negative after using the "-z" option to zero out the stats.
This patch fixes that by not zeroing out the srvcache_size field
for "-z", since it is the size of the cache and not a counter
pluknet [Tue, 24 Jan 2012 10:28:19 +0000 (10:28 +0000)]
MFC r230256:
Fix the "lock &zrl->zr_mtx already initialized" assertion by initializing
the allocated memory before calling mtx_init(9) on mtx pointing to it.
Otherwize, random contents of uninitialized memory might occasionally
trigger the assertion.
Reported by: Pavel Polyakov <bsd kobyla org>
Reviewed by: pjd
jh [Mon, 23 Jan 2012 16:28:35 +0000 (16:28 +0000)]
MFC r229694:
r222004 changed sbuf_finish() to not clear the buffer error status. As a
consequence sbuf_len() will return -1 for buffers which had the error
status set prior to sbuf_finish() call. This causes a problem in
pfs_read() which purposely uses a fixed size sbuf to discard bytes which
are not needed to fulfill the read request.
Work around the problem by using the full buffer length when
sbuf_finish() indicates an overflow. An overflowed sbuf with fixed size
is always full.
pho [Sun, 22 Jan 2012 18:27:24 +0000 (18:27 +0000)]
MFC: r228360
Move cpu_set_upcall(newtd, td) up before the first call of
thread_free(newtd). This to avoid a possible page fault in
cpu_thread_clean() as seen on amd64 with syscall fuzzing.
rmacklem [Sun, 22 Jan 2012 05:16:31 +0000 (05:16 +0000)]
MFC: r229802
opt_inet6.h was missing from some files in the new NFS subsystem.
The effect of this was, for clients mounted via inet6 addresses,
that the DRC cache would never have a hit in the server. It also
broke NFSv4 callbacks when an inet6 address was the only one available
in the client. This patch fixes the above, plus deletes opt_inet6.h
from a couple of files it is not needed for.
alc [Sat, 21 Jan 2012 19:21:42 +0000 (19:21 +0000)]
MFC r228923, r228935, and r229007
Eliminate many of the unnecessary differences between the native and
paravirtualized pmap implementations for i386.
Fix a bug in the Xen pmap's implementation of
pmap_extract_and_hold(): If the page lock acquisition is retried,
then the underlying thread is not unpinned.
Wrap nearby lines that exceed 80 columns.
Merge r216333 and r216555 from the native pmap
When r207410 eliminated the acquisition and release of the page
queues lock from pmap_extract_and_hold(), it didn't take into
account that pmap_pte_quick() sometimes requires the page queues
lock to be held. This change reimplements pmap_extract_and_hold()
such that it no longer uses pmap_pte_quick(), and thus never
requires the page queues lock.
Merge r177525 from the native pmap
Prevent the overflow in the calculation of the next page
directory. The overflow causes the wraparound with consequent
corruption of the (almost) whole address space mapping.
Strictly speaking, r177525 is not required by the Xen pmap because
the hypervisor steals the uppermost region of the normal kernel
address space. I am nonetheless merging it in order to reduce the
number of unnecessary differences between the native and Xen pmap
implementations.
alc [Sat, 21 Jan 2012 18:11:12 +0000 (18:11 +0000)]
MFC r228746
The Xen pmap doesn't support superpages. So, there is no point in it
initializing structures, like the pv table, that are only used to
implement superpages. In fact, some of the unnecessary code in
pmap_init() was actually doing harm. It was preventing the kernel from
booting on virtual machines with more than 768 MB of memory.
rmh [Sat, 21 Jan 2012 17:22:50 +0000 (17:22 +0000)]
MFC r227827
Define __FreeBSD_kernel__ macro in sys/param.h.
__FreeBSD_kernel__ indicates that this system uses the kernel of FreeBSD,
which by definition is always true on FreeBSD. This macro is also defined
on other systems that use the kernel of FreeBSD, such as GNU/kFreeBSD.
It is tempting to use this macro in userland code when we want to enable
kernel-specific routines, and in fact it's fine to do this in code that
is part of FreeBSD itself. However, be aware that as presence of this
macro is still not widespread (e.g. older FreeBSD versions, 3rd party
compilers, etc), it is STRONGLY DISCOURAGED to check for this macro in
external applications without also checking for __FreeBSD__ as an
alternative.
alc [Sat, 21 Jan 2012 08:26:41 +0000 (08:26 +0000)]
MFC r228398
Avoid the possibility of integer overflow in the calculation of
VM_KMEM_SIZE_MAX. Specifically, if the user/kernel address space split
was changed such that the kernel address space was greater than or equal
to 2 GB, then overflow would occur.
alc [Sat, 21 Jan 2012 05:03:10 +0000 (05:03 +0000)]
MFC r226163, r228317, and r228324
Fix the handling of an empty kmem map by sysctl_kmem_map_free().
Eliminate the possibility of 32-bit arithmetic overflow in the
calculation of vm_kmem_size that may occur if the system
administrator has specified a vm.vm_kmem_size tunable value that
exceeds the hard cap.
lstewart [Sat, 21 Jan 2012 03:59:31 +0000 (03:59 +0000)]
MFC r229898:
Consumers of bpfdetach() expect it to remove all bpf_if structs from the
bpf_iflist list which reference the specified ifnet. The existing implementation
only removes the first matching bpf_if found in the list, effectively leaking
list entries if an ifnet has been bpfattach()ed multiple times with different
DLTs.
Fix the leak by performing the detach logic in a loop, stopping when all bpf_if
structs referencing the specified ifnet have been detached and removed from the
bpf_iflist list.
Whilst here, also:
- Remove the unnecessary "bp->bif_ifp == NULL" check, as a bpf_if should never
exist in the list with a NULL ifnet pointer.
- Except when INVARIANTS is in the kernel config, silently ignore the case where
no bpf_if referencing the specified ifnet is found, as it is harmless and does
not require user attention.
gnn [Wed, 18 Jan 2012 21:54:34 +0000 (21:54 +0000)]
MFC: 229965
Fix for PR 138526.
Add the ability for /dev/null and /dev/zero to accept
being set into non blocking mode via fcntl(). This
brings the code into compliance with IEEE Std 1003.1-2001
as referenced in another PR, 94729.
truckman [Wed, 18 Jan 2012 21:49:37 +0000 (21:49 +0000)]
MFC: r229984
Pass the arguments to mtx_init() in the correct order. There should be
no change to the binary because the value of MTX_DEF is zero and there
is a visible function prototype.
gavin [Tue, 17 Jan 2012 13:52:04 +0000 (13:52 +0000)]
Merge r229085 from head:
Default to not performing the early-boot memory tests when we detect we
are booting inside a VM. There are three reasons to disable this:
o It causes the VM host to believe that all the tested pages or RAM are
in use. This in turn may force the host to page out pages of RAM
belonging to other VMs, or otherwise cause problems with fair resource
sharing on the VM cluster.
o It adds significant time to the boot process (around 1 second/Gig in
testing)
o It is unnecessary - the host should have already verified that the
memory is functional etc.
Note that this simply changes the default when in a VM - it can still be
overridden using the hw.memtest.tests tunable.
pho [Tue, 17 Jan 2012 11:04:58 +0000 (11:04 +0000)]
MFC: r228218, r228219, 228220, 228221
Rename copyin_timeout32 to umtx_copyin_timeout32 and move parameter
check here. Include check for negative seconds value.
Add umtx_copyin_timeout() and move parameter checks here.
Add declaration of umtx_copyin_timeout()
Use umtx_copyin_timeout() to copy and check timeout parameter.
kib [Tue, 17 Jan 2012 06:57:55 +0000 (06:57 +0000)]
MFC r229934:
Change the type of the paging_in_progress refcounter from u_short to
u_int. With the auto-sized buffer cache on the modern machines, UFS
metadata can generate more the 65535 pages belonging to the buffers
undergoing i/o, overflowing the counter.
To keep the layout of other fields of the struct vm_object intact
on stable/9, put enlarged paging_in_progress at the end of the structure,
and put a placeholder in the place of old pip counter.
kib [Tue, 17 Jan 2012 06:40:41 +0000 (06:40 +0000)]
MFC r229435:
Add special loader environment variables 'comconsole_port' and
'comconsole_pcidev'.
Change allows to use ISA serial ports other than COM1 for the
loader/kernel console without loader recompilation, and to configure
console on the PCI-attached UARTs.
pluknet [Mon, 16 Jan 2012 14:55:42 +0000 (14:55 +0000)]
MFC r228985,229881:
Unset the environment variables associated with individual menu items
before invoking the kernel.
Get rid of a spurious warning on the console when booting the kernel
from the interactive loader(8) prompt and beastie_disable="YES" is set
in loader.conf(5).
jhb [Fri, 13 Jan 2012 20:35:12 +0000 (20:35 +0000)]
MFC 229665,229672,229700:
Remove the assertion from tcp_input() that rcv_nxt is always greater
than or equal to rcv_adv and fix tcp_twstart() to handle this case by
assuming the last window was zero rather than a negative value.
The code in tcp_input() already safely handled this case. It can happen
due to delayed ACKs along with a remote sender that sends data beyond
the window we previously advertised. If we have room in our socket buffer
for the extra data beyond the advertised window, we will accept it.
However, if the ACK for that segment is delayed, then we will not
effectively fixup rcv_adv to account for that extra data until the
next segment arrives and forces out an ACK. When that next segment
arrives, rcv_nxt will be beyond rcv_adv.
jhb [Fri, 13 Jan 2012 20:28:11 +0000 (20:28 +0000)]
MFC 228961:
Further relax the strictness of enforcing allocations to only come from
decoded ranges. Pass any request for a specific range that fails because
it is not in a decoded range for an ACPI Host-PCI bridge up to the parent
to see if it can still be allocated. This is based on the assumption that
many BIOSes are inconsistent/broken and that settings programmed into BARs
or resources assigned to other built-in components are more trustworthy than
the list of decoded resource ranges in _CRS. This effectively limits the
decoded ranges to only being used for "wildcard" ranges when allocating
fresh resources for a BAR, etc. At some point I would like to only be
this permissive during an early scan of firmware-assigned resources during
boot and to be strict about all later allocations, but that isn't viable
currently.
jhb [Fri, 13 Jan 2012 20:25:38 +0000 (20:25 +0000)]
MFC 228960:
Cap the priority calculated from the current thread's running tick count
at SCHED_PRI_RANGE to prevent overflows in the priority value. This can
happen due to irregularities with clock interrupts under certain
virtualization environments.
jhb [Fri, 13 Jan 2012 20:15:27 +0000 (20:15 +0000)]
MFC 229429:
Some small fixes to CPU accounting for threads:
- Only initialize the per-cpu switchticks and switchtime in sched_throw()
for the very first context switch on APs during boot. This avoids a
small gap between the middle of thread_exit() and sched_throw() where
time is not accounted to any thread.
- In thread_exit(), update the timestamp bookkeeping to track the changes
to mi_switch() introduced by td_rux so that the code once again matches
the comment claiming it is mimicing mi_switch(). Specifically, only
update the per-thread stats directly and depend on ruxagg() to update
p_rux rather than adjusting p_rux directly. While here, move the
timestamp bookkeeping as late in the function as possible.
jhb [Fri, 13 Jan 2012 19:54:00 +0000 (19:54 +0000)]
MFC 229427:
Fix a few bugs in the SRAT parsing code:
- Actually increment ndomain when building our list of known domains
so that we can properly renumber them to be 0-based and dense.
- If the number of domains exceeds the configured maximum (VM_NDOMAIN),
bail out of processing the SRAT and disable NUMA rather than hitting an
obscure panic later.
- Don't bother parsing the SRAT at all if VM_NDOMAIN is set to 1 to
disable NUMA (the default).
jhb [Fri, 13 Jan 2012 19:50:52 +0000 (19:50 +0000)]
MFC 229390,229420,229479:
Fix some races in the multicast code by removing places where we would
drop the IF_ADDR_LOCK while walking an interface's multicast address list:
- Use TAILQ_FOREACH() instead of TAILQ_FOREACH_SAFE() for some loops that
do not modify the queues they iterate over.
- When cancelling multicast timers on an interface, don't release the
reference on a group in the leaving state while iterating over the loop.
Instead, use the same approach used in igmp_ifdetach() and mld_ifdetach()
of placing the groups to free on a pending release list and then releasing
the references after dropping the IF_ADDR_LOCK.
- Use the mli_relinmhead list normally used to defer calls to
in6m_release_locked() to defer calls to mld_v1_transmit_report() until
after the IF_ADDR_LOCK is dropped.
jhb [Fri, 13 Jan 2012 19:20:09 +0000 (19:20 +0000)]
MFC 229414,229476,229477:
Various fixes to the SIOC[DG]LIFADDR ioctl handlers:
- Grab a reference on any matching interface address (ifa) before dropping
the IF_ADDR_LOCK() and release the reference after using it to prevent a
potential use-after-free.
- Fix the IPv4 ioctl handlers in in_lifaddr_ioctl() to work with IPv4
interface addresses rather than IPv6.
- Add missing interface address list locking in the IPv4 handlers.
jhb [Fri, 13 Jan 2012 19:10:53 +0000 (19:10 +0000)]
MFC 229400:
Improve the cscope target's handling of MD directories. Automatically
include the MACHINE_CPUARCH directory if it differents from MACHINE when
building an index for a single machine. Also, include the 'x86' directory
when building an index for i386, pc98, or amd64.