There are two different versions of the ARM ABI depending on the
TARGET_ARCH. As these are sligntly different a package built for
one may not work on another. We need to detect which one we are on
by parsing the .ARM.attributes section.
This will only work on the ARM EABI as this section is part of the
ABI definition. As armv6 only supports the ARM EABI this is not a
problem for the oabi.
Older versions of libelf in FreeBSD fail to read the
.ARM.attributes section needed. As armv6 is unsupported on these
versions we can assume we are running on arm.
Submitted by: andrew
Approved by: re (delphij)
Obtained from: pkgng git
Clean up the Kerberos build by turning libheimipcc and libheimipcs into
private shared libraries, instead of hacked-together archives of PIC
objects. This makes it possible to build a static libkrb5 that works.
Fix the length calculation for the final block of a sendfile(2)
transmission which could be tricked into rounding up to the nearest
page size, leaking up to a page of kernel memory. [13:11]
In IPv6 and NetATM, stop SIOCSIFADDR, SIOCSIFBRDADDR, SIOCSIFDSTADDR
and SIOCSIFNETMASK at the socket layer rather than pass them on to the
link layer without validation or credential checks. [SA-13:12]
Prevent cross-mount hardlinks between different nullfs mounts of the
same underlying filesystem. [SA-13:13]
Peter Grehan [Tue, 10 Sep 2013 03:48:18 +0000 (03:48 +0000)]
Go way past 11 and bump bhyve's max vCPUs to 16.
This should be sufficient for 10.0 and will do
until forthcoming work to avoid limitations
in this area is complete.
Thanks to Bela Lubkin at tidalscale for the
headsup on the apic/cpu id/io apic ASL parameters
that are actually hex values and broke when
written as decimal when 11 vCPUs were configured.
Emit the correct standard library dependency line for C++ programs. In
the CLANG_IS_CC case, the default is now libc++. Only use libstdc++ if
!CLANG_IS_CC or it was explicitly requested in CXXFLAGS.
John Baldwin [Mon, 9 Sep 2013 18:11:59 +0000 (18:11 +0000)]
Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a mapping use
an address in the first 2GB of the process's address space. This flag should
have the same semantics as the same flag on Linux.
To facilitate this, add a new parameter to vm_map_find() that specifies an
optional maximum virtual address. While here, fix several callers of
vm_map_find() to use a VMFS_* constant for the findspace argument instead of
TRUE and FALSE.
Make the primary name of the OF console device /dev/ofwcons, and only
alias it to the contents of the output property if it is defined. This
avoids a panic when booting machines (QEMU) where the output-device
property is not defined.
Since output-device is free-form and potentially conflicts with other
entries in /dev, I also am not sure we should be doing the aliasing at
all, but this at least makes things work again.
These three files appeared in 6.0p1, which was imported into the vendor
branch but never merged to head. They were inadvertantly left out when
6.1p1 was merged to head. It didn't make any difference at the time,
because they were unused, but one of them is required for DNS-based host
key verification.
Add POWER CPUs to the kernel's knowledge. This does not imply we currently
actually run on any machines with POWER CPUs but avoids closing that door
unnecessarily.
Add hook called when every new processor is brought online -- including the
BSP -- so that platform modules have a chance to add the new CPU to any
internal bookkeeping.
Use the canonical bits for wired, etc. in the PTE. This is important for
interactions with certain kinds of hypervisors that look into the PTEs
more closely than they should.
Xin LI [Mon, 9 Sep 2013 05:01:18 +0000 (05:01 +0000)]
In r243868, the error message buffer errmsg have been changed from
an on-stack array to a pointer and therefore sizeof(errmsg) would
become 4 or 8 bytes depending on the architecture.
Fix this by using ERRMSGL in place of sizeof().
Submitted by: J David <j.david.lists@gmail.com>
MFC after: 3 days
Approved by: re (kib)
Alan Cox [Sun, 8 Sep 2013 21:30:53 +0000 (21:30 +0000)]
Prior to r254304, we only began scanning the active page queue when the
amount of free memory was close to the point at which we would begin
reclaiming pages. Now, we continuously scan the active page queue,
regardless of the amount of free memory. Consequently, we are continuously
calling pmap_ts_referenced() on active pages.
Prior to this change, pmap_ts_referenced() would always demote superpage
mappings in order to obtain finer-grained reference information. This made
sense because we were coming under memory pressure and would soon have to
begin reclaiming pages. Now, however, with continuous scanning of the
active page queue, these demotions are taking a toll on performance. For
example, on one of my test machines, the running time for the HPCC Random
Access benchmark (also known as GUPS) has increased by 54%. To address this
problem, I have replaced the demotion with a heuristic for periodically
clearing the reference flag on superpage mappings.
Drain for the xbusy state for two places which potentially do
pmap_remove_all(). Not doing the drain allows the pmap_enter() to
proceed in parallel, making the pmap_remove_all() effects void.
The race results in an invalidated page mapped wired by usermode.
Reported and tested by: pho
Reviewed by: alc
Sponsored by: The FreeBSD Foundation
Approved by: re (glebius)
Add a stock libmap32.conf for amd64. The first two lines have no effect
except to document the hardcoded standard library search path for 32-bit
binaries. The third line performs the equivalent substitution for the
private library directory.
Ironically, these entries rely on functionality which is only available
in the COMPAT_32BIT version of rtld-elf.
Create a private library directory (LIBPRIVATEDIR) for libraries which
we don't want to expose but which can't or shouldn't be static.
To mark a library as private, define PRIVATELIB in its Makefile. It
will be installed in LIBPRIVATEDIR, which is normally /usr/lib/private
(or /usr/lib32/private for 32-bit libraries on 64-bit platforms).
To indicate that a program or library depends on a private library,
define USEPRIVATELIB in its Makefile. The correct version of
LIBPRIVATEDIR will be added to its run-time library search path.
Mark Murray [Sat, 7 Sep 2013 22:07:36 +0000 (22:07 +0000)]
Fix the build; Certain linkable symbols need to always be present.
Pass the pointy hat please.
Also unblock the software (Yarrow) generator for now. This will be
reverted; Yarrow needs to block until secure, not this behaviour
of serving as soon as asked.
Folks with specific requiremnts will be able to (can!) unblock this
device with any write, and are encouraged to do so in /etc/rc.d/*
scripting. ("Any" in this case could be "echo '' > /dev/random" as
root).
Jayachandran C. [Sat, 7 Sep 2013 18:26:16 +0000 (18:26 +0000)]
Netlogic XLP network driver update
Changes are to
- update board and network interface detection logic
- fix reading onboard CPLD in little-endian config
- print NAE frequency conrrectly for Bx chips
- update XAUI config to disable Rx/Tx until interface is up
Submitted by: Venkatesh J V <venkatesh.vivekanandan@broadcom.com>
Jayachandran C. [Sat, 7 Sep 2013 16:31:30 +0000 (16:31 +0000)]
Use a better version of memcpy/bcopy for mips kernel.
Use a variant of mips libc memcpy for kernel. This implementation uses
64-bit operations when compiled for 64-bit, and is significantly faster
in that case.
Alexander Motin [Sat, 7 Sep 2013 15:16:30 +0000 (15:16 +0000)]
Micro-optimize cpu_search(), allowing compiler to use more efficient inline
ffsl() implementation, when it is available, instead of homegrown iteration.
On dual-E5645 amd64 system (2x6x2 cores) under heavy I/O load that reduces
time spent inside cpu_search() from 19% to 13%, while IOPS increased by 5%.
Mark Murray [Sat, 7 Sep 2013 14:15:13 +0000 (14:15 +0000)]
Bring in some behind-the-scenes development, mainly By Arthur Mesh,
the rest by me.
o Namespace cleanup; the Yarrow name is now restricted to where it
really applies; this is in anticipation of being augmented or
replaced by Fortuna in the future. Fortuna is mentioned, but behind
#if logic, and is ignorable for now.
o The harvest queue is pulled out into its own modules.
o Entropy harvesting is emproved, both by being made more conservative,
and by separating (a bit!) the sources. Available entropy crumbs are
marginally improved.
o Selection of sources is made clearer. With recent revelations,
this will receive more work in the weeks and months to come.
Submitted by: Arthur Mesh (partly) <arthurmesh@gmail.com>
Davide Italiano [Sat, 7 Sep 2013 13:45:44 +0000 (13:45 +0000)]
- Use make_dev_credf(MAKEDEV_REF) instead of the race-prone make_dev()+
dev_ref() in the clone handlers that still use it.
- Don't set SI_CHEAPCLONE flag, it's not used anywhere neither in devfs
(for anything real)
Revert parts of r245132 and r245175. We don't need to write to the
IMAN register to clear the pending interrupt status bits. This patch
tries to solve problems seen on the MacBook Air, as reported by
Johannes Lundberg <johannes@brilliantservice.co.jp>
Peter Grehan [Sat, 7 Sep 2013 03:33:36 +0000 (03:33 +0000)]
Mask off the vector from the MSI-x data word.
Some o/s's set the trigger-mode level bit which
results in an invalid vector and pass-thru interrupts
not being delivered.
Remove unnecessary mediaClose (FTP operations are done with either ftp(1)
or fetch(1), neither of which are stateful, compared to how sysinstall(8)
did FTP operations, maintaining an open session until mediaClose).
Our code does not consider yet the case of hash collisions. This
is a rather annoying situation where two or more files that
happen to have the same hash value will not appear accessible.
The situation is not difficult to work-around but given that things
will just work without enabling htree we will save possible
embarrassments for the next release.
Rick Macklem [Fri, 6 Sep 2013 23:14:31 +0000 (23:14 +0000)]
Intermittent crashes in the NLM (rpc.lockd) code during system
shutdown was reporetd via email. The crashes occurred because the
client side NLM would attempt to use its socket after it had been
destroyed. Looking at the code, it would soclose() once the reference
count on the socket handling structure went to 0. Unfortunately,
nlm_host_get_rpc() will simply allocate a new socket handling structure
when none exists and use the now soclose()d socket. Since there doesn't
seem to be a safe way to determine when the socket is no longer needed,
this patch modifies the code so that it never soclose()es the socket.
Since there is only one socket ever created, this does not introduce a
leak when the rpc.lockd is stopped/restarted. The patch also disables
unloading of the nfslockd module, since it is not safe to do so (and
has never been safe to do so, from what I can see).
sys/amd64/include/cpufunc.h:
sys/amd64/amd64/pmap.c:
Move invltlb_globpcid() into cpufunc.h so that it can be
used by the Xen HVM version of tlb shootdown IPI handlers.
sys/x86/xen/xen_intr.c:
sys/xen/xen_intr.h:
Rename xen_intr_bind_ipi() to xen_intr_alloc_and_bind_ipi(),
and remove the ipi vector parameter. This api allocates
an event channel port that can be used for ipi services,
but knows nothing of the actual ipi for which that port
will be used. Removing the unused argument and cleaning
up the comments surrounding its declaration helps clarify
its actual role.
sys/amd64/amd64/mp_machdep.c:
sys/amd64/include/cpu.h:
sys/i386/i386/mp_machdep.c:
sys/i386/include/cpu.h:
Implement a generic framework for amd64 and i386 that allows
the implementation of certain CPU management functions to
be selected at runtime. Currently this is only used for
the ipi send function, which we optimize for Xen when running
on a Xen hypervisor, but can easily be expanded to support
more operations.
sys/amd64/include/pcpu.h:
sys/i386/include/pcpu.h:
sys/i386/include/smp.h:
Remove NR_VIRQS and NR_IPIS from FreeBSD headers. NR_VIRQS
is defined already for us in the xen interface files.
NR_IPIS is only needed in one file per Xen platform and is
easily inferred by the IPI vector table that is defined in
those files.
sys/i386/xen/mp_machdep.c:
Restructure to more closely match the HVM implementation by
performing table driven IPI setup.
Davide Italiano [Fri, 6 Sep 2013 21:02:43 +0000 (21:02 +0000)]
Retire netisr.netisr_direct and netisr.netisr_direct_force sysctls.
These were used to control/export dispatch policy but they're not anymore.
This commit cannot be MFC'ed to 9 because old netstat(9) binary relies
on such sysctl to work. On the other hand, there's no real reason to
keep'em around in 10.