Dmitry Chagin [Sun, 25 Jun 2017 15:21:51 +0000 (15:21 +0000)]
PFS_DELEN is the sum of the permanent part of the struct dirent and
fixed size for the name buffer PFS_NAMELEN.
As r318736 was commited (ino64 project) the size of the permanent part
of the struct dirent was changed, so calulate PFS_DELEN properly.
Andrew Turner [Sun, 25 Jun 2017 13:22:49 +0000 (13:22 +0000)]
Stop calling cpu_dcache_wb_range from PTE_SYNC.
We set the shareability attributes in TCR_EL1 on boot. These tell the
hardware the pagetables are in cached memory so there is no need to flush
the entries from the cache to memory.
This has about 4.2% improvement in system time and 2.7% improvement in
user time for a buildkernel -j48 on a ThunderX.
Keep the old code for now to allow for further comparisons.
Gleb Smirnoff [Sun, 25 Jun 2017 01:41:07 +0000 (01:41 +0000)]
Provide sbsetopt() that handles socket buffer related socket options.
It distinguishes between data flow sockets and listening sockets, and
in case of the latter doesn't change resource limits, since listening
sockets don't hold any buffers, they only carry values to be inherited
by their children.
Rick Macklem [Sat, 24 Jun 2017 20:01:21 +0000 (20:01 +0000)]
Add two new compound RPCs to the NFSv4.1/pNFS client.
When the NFSv4.1 client is doing pNFS, it needs to get an Open and
a Layout for every file it will be doing I/O on. The current code
does two separate RPCs to get these. This patch adds two new compounds
that do the both the Open and LayoutGet in the same RPC, reducing the
RPC count.
It also factors out the code that sets up and parses the LayoutGet operation
into separate functions, so that the code doesn't get duplicated for
these new RPCs.
This patch is fairly large, but should only affect the NFSv4.1 client
when the "pnfs" option is specified.
Ed Maste [Sat, 24 Jun 2017 18:37:51 +0000 (18:37 +0000)]
Clean up stale dependencies after r320278
Our current approach to dependency tracking cannot cope with switching
generated asm syscall stubs into C wrappers. Perpetuate the hack in
Makefile.inc1 to paper over the problem until we can take a holistic
approach to fixing dependency problems.
Alan Cox [Sat, 24 Jun 2017 17:10:33 +0000 (17:10 +0000)]
Increase the pageout cluster size to 32 pages.
Decouple the pageout cluster size from the size of the hash table entry
used by the swap pager for mapping (object, pindex) to a block on the
swap device(s), and keep the size of a hash table entry at its current
size.
Guard, requested by the MAP_GUARD mmap(2) flag, prevents the reuse of
the allocated address space, but does not allow instantiation of the
pages in the range. It is useful for more explicit support for usual
two-stage reserve then commit allocators, since it prevents accidental
instantiation of the mapping, e.g. by mprotect(2).
Use guards to reimplement stack grow code. Explicitely track stack
grow area with the guard, including the stack guard page. On stack
grow, trivial shift of the guard map entry and stack map entry limits
makes the stack expansion. Move the code to detect stack grow and
call vm_map_growstack(), from vm_fault() into vm_map_lookup().
As result, it is impossible to get random mapping to occur in the
stack grow area, or to overlap the stack guard page.
Do not try to unmark MAP_ENTRY_IN_TRANSITION marked by other thread.
The issue is catched by "vm_map_wire: alien wire" KASSERT at the end
of the vm_map_wire(). We currently check for MAP_ENTRY_WIRE_SKIPPED
flag before ensuring that the wiring_thread is curthread. For HOLESOK
wiring, this means that we might see WIRE_SKIPPED entry from different
wiring.
The fix it by only checking WIRE_SKIPPED if the entry is put
IN_TRANSITION by us. Also fixed a typo in the comment explaining the
situation.
Reported and tested by: pho
Reviewed by: alc
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Emmanuel Vadot [Sat, 24 Jun 2017 16:41:26 +0000 (16:41 +0000)]
Allwinner: Add support for H2 Plus SoC
H2+ SoC is a stripped down version of H3 without gigabit ethernet and 4K HDMI.
Also add sun8i-h2-plus-orangepi-zero.dts to the build as we run on this board.
The flag is not implemented, all FreeBSD architectures correctly
handle locks on normal cacheable mappings. On the other hand, the
flag was specified by some software, so it is kept in the header as
nop. Removal from the man page should discourage its use.
Emmanuel Vadot [Sat, 24 Jun 2017 09:33:25 +0000 (09:33 +0000)]
loader.efi: Disable smbios for arm
The smbios code does a lot of unaligned access, since we don't really
care about smbios info on ARM (not all board expose information and those
who does don't expose useful ones) disable smbios for this arch (at least
for now).
Ed Maste [Fri, 23 Jun 2017 18:58:28 +0000 (18:58 +0000)]
makefs: add copies of NetBSD makefs msdos source files
We do not treat makefs as contrib code. Import copies of makefs msdos
files from NetBSD so that we can track our changes to these files.
These are copied from NetBSD, with only a change to use __FBSDID and
$FreeBSD$ instead of __KERNEL_RCSID and $NetBSD$. A copy of the
original $NetBSD$ tag remains in each source file.
These two files were missed in r320212. Also remove a stray blank line
added in msdosfs_vfsops.c.
Submitted by: Siva Mahadevan
Sponsored by: The FreeBSD Foundation
Ed Maste [Fri, 23 Jun 2017 18:41:49 +0000 (18:41 +0000)]
Allow Clang's integrated assembler to assemble boot0
dim@ compared clang IAS-built and GNU as-built boot0 and found them
equivalent. IAS encoded one instruction using two bytes where GNU as
used three, and another instruction using three bytes where GNU as used
two. The net result is equivalent and tested, so there is no need to
force IAS off for boot0.
Bryan Drewery [Fri, 23 Jun 2017 18:26:33 +0000 (18:26 +0000)]
packages: Allow stageworld/stagekernel to run with make jobs.
The -B was originally added in projects/release-pkg r289381 as a copy
of what 'make world' did at the time. The -B was removed from
the 'installworld' call in 'world' in r303844 though. The staging
of files is safe to run in parallel.
Warner Losh [Fri, 23 Jun 2017 18:06:20 +0000 (18:06 +0000)]
Forward compatibility for ino64.
Add forward compatibility so that new binaries can run on old
kernels. If the new system call from ino64 isn't available on your
system, then the old one will be used and the results translated. The
stat and statfs families of functions are fully emulated. While not
required by policy, in this case it is helpful to our users to provide
this compatibility. In this case, it allows rollback of the kernel
after installing a new userland should a problem be discovered. It
also prevents foot-shooting if a user does an install before rebooting
with the new kernel. Finally, it allows the use case where one needs
to run new binaries on an old kernel as part of an upgrade process.
The getdirentries family uses tricks that may not work on remote
filesystems. Specifically, it uses a buffer 1/4 the size requested to
get the data from he old syscall.
The code carefully uses direct syscalls for old system calls to avoid
referencing freebsd11_* symbols, which contaminate ld-elf.so.1's
export table due to its use of stat functions, which causes errno to
be incorrect in client programs due to the wrong *stat* function being
resolved in some cases.
This code should removed sometime after 12 is branched.
Tested on: 12-current binaries on a 10.3-beta kernel run and return
consistent results. 12-current kernel and userland with
packages from before ino64 was committed also work.
Ed Maste [Fri, 23 Jun 2017 15:57:58 +0000 (15:57 +0000)]
enable --build-id for the kernel link
A Build-ID is an identifier generated at link time to uniquely identify
ELF binaries. It allows efficient confirmation that an executable or
shared library and a corresponding standalone debuginfo file match.
(Otherwise, a checksum of the debuginfo file must be calculated when
opening it in a debugger.)
The FreeBSD base system includes GNU bfd ld 2.17.50 as the linker for
architectures other than arm64. Build-ID support was added to bfd ld
shortly after that version, so was not previously available to us.
We can now start making use of Build-ID as we migrate to using lld or
bfd ld from ports, conditionally enabled based on the LINKER_TYPE and
LINKER_VERSION make variables added in r320244 and subsequent commits.
Reviewed by: dim
MFC after: 3 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D11314
Michael Tuexen [Fri, 23 Jun 2017 09:27:31 +0000 (09:27 +0000)]
Use a longer buffer for messages in ERROR chunks.
This allows them to be sent in a non truncated way and addresses a warning
given by newver versions of gcc.
Thanks to Anselm Jonas Scholl for reporting it and providing a patch.
Ed Maste [Fri, 23 Jun 2017 03:55:00 +0000 (03:55 +0000)]
change GNU ld LINKER_TYPE from binutils to bfd
GNU binutils includes two linkers: ld.bfd and ld.gold. For clarity use
LINKER_TYPE=bfd to refer to ld.bfd, the original binutils linker that
identifies itself as "GNU ld".
Cy Schubert [Fri, 23 Jun 2017 01:05:49 +0000 (01:05 +0000)]
Replace the leap-seconds file in r320242 from USNO -
ftp://tycho.usno.navy.mil/pub/ntp/leap-seconds.3701462400 - with a
leap-seconds file from NIST at ftp://time.nist.gov/pub/. The USNO
version of the file changes the last documented leap second update
time whereas the NIST version does not. The expiration of the USNO
version of the file is also one month short.
Requested by: ian@
Obtained from: ftp://time.nist.gov/pub/leap-seconds.3676924800
MFC after: 3 days
Glen Barber [Fri, 23 Jun 2017 00:08:36 +0000 (00:08 +0000)]
In release/release.sh:
- Rename chroot_arm_armv6_build_release() to chroot_arm_build_release()
and make it hardware agnostic (such as armv6 -vs- armv7 -vs- arm64).
- Evaluate EMBEDDED_TARGET differently so release/tools/arm.subr can
be used for arm/armv6 and arm64/aarch64.
- Update comments and copyright.
In release/tools/arm.subr:
- In arm_create_disk(), change the default alignment from 63 to 512k,
fixing a boot issue on arm64 and EFI. [1]
- Update comments and copyright.
Add a RPI3 configuration file, pieces obtained from Crochet.
Obtained from: Crochet [1]
MFC after: 5 days
X-MFC-Note: maybe
Sponsored by: The FreeBSD Foundation
As per https://datacenter.iers.org/eop/-/somos/5Rgv/latest/16:
INTERNATIONAL EARTH ROTATION AND REFERENCE SYSTEMS SERVICE (IERS)
SERVICE INTERNATIONAL DE LA ROTATION TERRESTRE ET DES SYSTEMES DE REFERENCE
SERVICE DE LA ROTATION TERRESTRE
OBSERVATOIRE DE PARIS
61, Av. de l'Observatoire 75014 PARIS (France)
Tel. : 33 (0) 1 40 51 23 35
FAX : 33 (0) 1 40 51 22 91
Internet : services.iers@obspm.fr
Paris, 9 January 2017
Bulletin C 53
To authorities responsible
for the measurement and
distribution of time
INFORMATION ON UTC - TAI
NO leap second will be introduced at the end of June 2017.
The difference between Coordinated Universal Time UTC and the
International Atomic Time TAI is :
from 2017 January 1, 0h UTC, until further notice : UTC-TAI = -37 s
Leap seconds can be introduced in UTC at the end of the months of December
or June, depending on the evolution of UT1-TAI. Bulletin C is mailed every
six months, either to announce a time step in UTC, or to confirm that there
will be no time step at the next possible date.
Christian BIZOUARD
Director
Earth Orientation Center of IERS
Observatoire de Paris, France
Obtained from: ftp://tycho.usno.navy.mil/pub/ntp/leap-seconds.3701462400
MFC after: 3 days
Ed Schouten [Thu, 22 Jun 2017 18:39:52 +0000 (18:39 +0000)]
Use __ISO_C_VISIBLE, as opposed to testing __STDC_VERSION__.
FreeBSD's C library uses __STDC_VERSION__ to determine whether the
compiler provides language features specific to a certain version of the
C standard. __ISO_C_VISIBLE is used to specify which library features
need to be exposed.
max_align_t currently uses __STDC_VERSION__, even though it should be
using __ISO_C_VISIBLE to remain consistent with the rest of the headers
in include/.
Reviewed by: dim
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D11303
Andriy Gapon [Thu, 22 Jun 2017 17:10:34 +0000 (17:10 +0000)]
MFV r319950: 5220 L2ARC does not support devices that do not provide 512B access
FreeBSD note: the actual change has been in FreeBSD since r297848. This
commit accounts for integration of that change with subsequent changes,
especially r320156 (MFV of r318946) and r314274.
https://www.illumos.org/issues/5220
There are disk devices that have logical sector size larger than 512B, for
example 4KB. That is, their physical sector size is larger than 512B and they
do not provide emulation for 512B sector sizes. For such devices both a data
offset and a data size must be properly aligned. L2ARC should arrange that
because it uses physical I/O.
zio_vdev_io_start() performs a necessary transformation if io_size is not
aligned to vdev_ashift, but that is done only for logical I/O. Something
similar should be done in L2ARC code.
* a temporary write buffer should be allocated if the original buffer is
not going to be compressed and its size is not aligned
* size of a temporary compression buffer should be ashift aligned
* for the reads, if a size of a target buffer is not sufficiently large and
it is not aligned then a temporary read buffer should be allocated
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Saso Kiselkov <saso.kiselkov@nexenta.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Author: Andriy Gapon <avg@FreeBSD.org>
https://www.illumos.org/issues/8056
The send size estimate for a zvol can be too low, if the size of the record
headers (dmu_replay_record_t's) is a significant portion of the size.
This is typically the case when the data is highly compressible, especially
with embedded blocks.
The problem is that dmu_adjust_send_estimate_for_indirects() assumes that
blocks are the size of the "recordsize" property (128KB).
However, for zvols, the blocks are the size of the "volblocksize" property
(8KB). Therefore, we estimate that there will be 16x less record headers than
there really will be.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Paul Dagnelie <pcd@delphix.com>
Andriy Gapon [Thu, 22 Jun 2017 16:52:22 +0000 (16:52 +0000)]
MFV r318947: 7578 Fix/improve some aspects of ZIL writing.
FreeBSD note: this commit removes small differences between what mav
committed to FreeBSD in r308782 and what ended up committed to illumos
after addressing all review comments.
https://www.illumos.org/issues/7578
After some ZIL changes 6 years ago zil_slog_limit got partially broken
due to zl_itx_list_sz not updated when async itx'es upgraded to sync.
Actually because of other changes about that time zl_itx_list_sz is not
really required to implement the functionality, so this patch removes
some unneeded broken code and variables.
Original idea of zil_slog_limit was to reduce chance of SLOG abuse by
single heavy logger, that increased latency for other (more latency critical)
loggers, by pushing heavy log out into the main pool instead of SLOG. Beside
huge latency increase for heavy writers, this implementation caused double
write of all data, since the log records were explicitly prepared for SLOG.
Since we now have I/O scheduler, I've found it can be much more efficient
to reduce priority of heavy logger SLOG writes from ZIO_PRIORITY_SYNC_WRITE
to ZIO_PRIORITY_ASYNC_WRITE, while still leave them on SLOG.
Existing ZIL implementation had problem with space efficiency when it
has to write large chunks of data into log blocks of limited size. In some
cases efficiency stopped to almost as low as 50%. In case of ZIL stored on
spinning rust, that also reduced log write speed in half, since head had to
uselessly fly over allocated but not written areas. This change improves
the situation by offloading problematic operations from z*_log_write() to
zil_lwb_commit(), which knows real situation of log blocks allocation and
can split large requests into pieces much more efficiently. Also as side
effect it removes one of two data copy operations done by ZIL code WR_COPIED
case.
While there, untangle and unify code of z*_log_write() functions.
Also zfs_log_write() alike to zvol_log_write() can now handle writes crossing
block boundary, that may also improve efficiency if ZPL is made to do that.
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Andriy Gapon <avg@FreeBSD.org>
Reviewed by: Steven Hartland <steven.hartland@multiplay.co.uk>
Reviewed by: Brad Lewis <brad.lewis@delphix.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Author: Alexander Motin <mav@FreeBSD.org>
Ed Maste [Thu, 22 Jun 2017 15:09:42 +0000 (15:09 +0000)]
retire arm64 kernel module linker workaround
Relocatable linking in aarch64 ld from binutils 2.25.1 does not work.
The linker corrupts the references to the external symbols which are
defined by other object in the linking set and should therefore lose
the GOT entry.
The problem is fixed in later versions of GNU ld and does not exist in
the in-tree lld linker that we now use by default for arm64, so the
workaround can be removed.
Reviewed by: kib
MFC after: 3 weeks
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D11302
Cy Schubert [Thu, 22 Jun 2017 12:46:48 +0000 (12:46 +0000)]
In poolnodcommand(): TTL (-T) is only valid when adding a node to a
pool (ippool -a) not when removing a node from a pool (ippool -r).
Flag -T as an error in ippool -r.
Bryan Drewery [Thu, 22 Jun 2017 05:34:41 +0000 (05:34 +0000)]
Rework logic for skipping .depend/.meta file read/stat/writes.
- Rename _SKIP_READ_DEPEND to _SKIP_DEPEND since it also avoids writing.
- This now uses .NOMETA to avoid reading any .meta files related to
DEPENDOBJS. Objects not in OBJS/DEPENDOBJS may still have their .meta
files read in if they are in the dependency graph.
- This also avoids statting .meta and .depend files in the META_MODE +
-DNO_FILEMON case.
Xin LI [Thu, 22 Jun 2017 05:10:16 +0000 (05:10 +0000)]
Fix use-after-free introduced in r300388.
In r300388, endnetconfig() was called on nc_handle which would release
the associated netconfig structure, which means tmpnconf->nc_netid
would be a use-after-free.
Solve this by doing endnetconfig() in return paths instead.
Reported by: jemalloc via kevlo
Reviewed by: cem, ngie (earlier version)
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D11288
Ed Maste [Thu, 22 Jun 2017 02:46:36 +0000 (02:46 +0000)]
makefs: add copies of NetBSD makefs msdos source files
We do not treat makefs as contrib code. Import copies of makefs msdos
files from NetBSD so that we can track our changes to these files.
These are copied from NetBSD, with only a change to use __FBSDID and
$FreeBSD$ instead of __KERNEL_RCSID and $NetBSD$. A copy of the
original $NetBSD$ tag remains in each source file.
Submitted by: Siva Mahadevan
Sponsored by: The FreeBSD Foundation
Pedro F. Giffuni [Thu, 22 Jun 2017 02:43:32 +0000 (02:43 +0000)]
ext2fs: add dir_nlink feature support.
ext4 on linux has always supported more than 32000 directories through
the dir_nlink feature, but FreeBSD was unable to catch up on this feature.
As part of the 64 bit inode changes nlink_t has been extended and this
feature is now possible.
Conrad Meyer [Thu, 22 Jun 2017 02:19:39 +0000 (02:19 +0000)]
join(1): Fix field ordering for -v output
Per POSIX, join(1) (in modes other than -o) is a concatenation of selected
character fields. The joined field is first, followed by fields in the
order they occurred in the input files.
Our join(1) utility previously handled this correctly for lines with a match
in the other file. But it failed to order output fields correctly for
unmatched lines, printed in -a and -v modes.
A simple test case is:
$ touch a
$ echo "2 1" > b
$ join -v2 -2 2 a b
1 2
Rick Macklem [Thu, 22 Jun 2017 00:17:15 +0000 (00:17 +0000)]
Ensure that the credentials field of the NFSv4 client open structure is
initialized.
bdrewery@ has reported panics "newnfs_copycred: negative nfsc_ngroups".
The only way I can see that this occurs is that the credentials field of
the open structure gets used before being filled in.
I am not sure quite how this happens, but for the file create case, the
code is serialized via the vnode lock on the directory. If, somehow, a
link to the same file gets created just after file creation, this might
occur.
This patch ensures that the credentials field is initialized to a reasonable
set of credentials before the structure is linked into any list, so I
this should ensure it is initialized before use.
I am committing the patch now, since bdrewery@ notes that the panics
are intermittent and it may be months before he knows if the patch fixes
his problem.
Bryan Drewery [Wed, 21 Jun 2017 23:01:18 +0000 (23:01 +0000)]
Follow-up r308602: Don't add missing headers to .depend.tables.h.
This also avoids an error from egrep when a header is missing. This can happen
with something like WITHOUT_BLUETOOTH set when searching for
$include_dir/netgraph/bluetooth/include/ng_btsocket.h. The warning was
not an error (from set -e) due to being on the left side of a pipe. Now the
all_headers list is only filled with existing headers.