Justin Hibbits [Fri, 21 Oct 2016 02:16:11 +0000 (02:16 +0000)]
Remove a hack requiring dtsec0 to always be enabled for mdio.
Instead replace it with a different hack, that turns fman into a simplebus
subclass, and maps its children within its address space.
Since all PHY communication is done through dtsec0's mdio space, the FDT
contains a reference to the dtsec0 mdio handle in all nodes that need it.
Instead of using Freescale's implementation for MII access, use our own (copied
loosely from the eTSEC driver, and could possibly be merged eventually). This
lets us access the registers directly rather than needing a full dtsec interface
just to access the registers.
Future directions will include turning fman into more of a simplebus, and not
mapping the region and playing games. This will require changes to the dtsec
driver to make it a child of fman, and possibly other drivers as well.
Rick Macklem [Thu, 20 Oct 2016 23:53:16 +0000 (23:53 +0000)]
A problem w.r.t. interoperation between the FreeBSD NFSv4.1 server with
delegations enabled and the Linux NFSv4.1 client was reported in
reviews.freebsd.org/D7891.
I believe that the FreeBSD server behaviour conforms to the RFC and that
the Linux client has a bug. Therefore, I do not think the proposed patch
is appropriate. When nfsrv_writedelegifpos is non-zero, the FreeBSD
server will issue a write delegation for a read open if possible.
The Linux client then erroneously assumes that the credentials used for
the read open can write the file.
This patch reverses the default value for nfsrv_writedelegifpos to 0 so
that the default behaviour is Linux compatible and adds a sysctl that can
be used to set nfsrv_writedelegifpos.
This change should only affect users that are mounting a FreeBSD server
with delegations enabled (they are not enabled by default) with a Linux
NFSv4.1 client mount.
Mark Johnston [Thu, 20 Oct 2016 23:02:30 +0000 (23:02 +0000)]
gmirror: Release pending regular requests when synchronization stops.
Normally gmirror allows colliding requests to proceed whenever a
synchronization request completes and advances to the next offset. However
if an I/O request collides with one of the final g_mirror_syncreqs, nothing
releases it once synchronization completes, resulting in an apparent I/O
hang. The same problem can occur if synchronization is aborted by an
I/O error. Therefore, be sure to requeue pending requests when
mirror synchronization is stopped for any reason.
While here, remove some dead code from g_mirror_regular_release().
Fix building of llvm's unwind if gcc has been also built
when building gcc an unwind.h header is generate in the cc_tool directory
which is included in the CFLAGS before the path where the llvm's unwind.h file
lives
Reviewed by: emaste
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D7357
Mateusz Guzik [Thu, 20 Oct 2016 19:55:50 +0000 (19:55 +0000)]
cache: fix up a corner case in r307650
If no negative entry is found on the last list, the ncp pointer will be
left uninitialized and a non-null value will make the function assume an
entry was found.
Add make rules to build LLVM IR from C/C++ sources.
As a foundation for future work with LLVM's Intermediate Representation (IR),
add new suffix rules that can be used to build .llo (text) or .bco (bitcode)
files from C or C++ sources. This compilation step uses the same CFLAGS, etc.,
as are used for building .o files, with the exception of optimization flags.
Many of the things we would like to do with IR (e.g., instrumentation) work
better with unoptimized code, so our approach is to build .c->.bco without
optimization and then apply the optimization in post-analysis,
post-instrumentation linking.
The overall result of these changes is:
* one can "make foo.llo" or "make foo.bco" wherever "make foo.o" was supported
* new make variables IR_CFLAGS and IR_CXXFLAGS are available to inspect the
flags that are used by Clang to generate the IR
These new rules are added unconditionally to our non-POSIX suffix rule set,
since we cannot inspect COMPILER_TYPE in sys.mk. Future changes that depend
on these rules (e.g., building IR versions of binaries from bsd.prog.mk)
should use COMPILER_TYPE to determine when we can expect IR rules to succeed.
Andrew Gallatin [Thu, 20 Oct 2016 13:48:29 +0000 (13:48 +0000)]
Clear mbuf hashtype on loopback when RSS is enabled.
The hashtype on an outgoing mbuf reflects the correct hash on the
transmit side of the connection. If this hash persists on loopback,
the receiving RSS/PCBGROUP code will use it to look up the pcbgroup
for the transmit side, which will often not match the pcbgroup for the
receive side of the connection. This leads to TCP connections
hanging, and dropping the SYN/ACK packet. This is essentially
the same as having a hardware network card generate mbufs with an
incorrect RSS hash.
There are a number of places which can set the hash on transmit,
so the simplest fix is to simply clear the hash at loopback time.
Clearing the hash allows a new, correct hash to be calculated in
software on the receive side.
Reviewed by: jtl
Discussed with: adrian
Sponsored by: Netflix
Wojciech Macek [Thu, 20 Oct 2016 11:23:59 +0000 (11:23 +0000)]
Support for MSI-X on Annapurna Alpine
This patch adds support for MSI-X interrupts
on Annapurna Alpine platform. MSI-X on Alpine
work similarly to GICv2m, i.e. some range of
SPI interrupts is reserved in GIC and individual
SPIs can be triggered by MSI-X messages.
This SPI range is defined in FDT.
Andrew Turner [Thu, 20 Oct 2016 09:22:10 +0000 (09:22 +0000)]
Add support for the fpu_kern(9) KPI on arm64. It hooks into the existing
VFP code to store the old context, with lazy loading of the new context
when needed.
FPU_KERN_NOCTX is missing as this is unused in the crypto code this has
been tested with, and I am unsure on the requirements of the UEFI
Runtime Services.
Reviewed by: kib
Obtained from: ABT Systeems Ltd
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D8276
Ed Maste [Wed, 19 Oct 2016 20:56:21 +0000 (20:56 +0000)]
Put each SUBDIR on a separate line for ease of maintenance
Additional patches to this file are in progress, and having each SUBDIR
entry on a separate line makes it easier to change the order in which
the patches are reviewed, tested, and applied.
Ed Maste [Wed, 19 Oct 2016 19:57:36 +0000 (19:57 +0000)]
Force MK_GDB to no if MK_BINUTILS == no
This is currently encoded in conditional blocks in gnu/lib/Makefile and
gnu/usr.bin/Makefile. Set it via src.opts.mk to make GDB's dependency
on binutils more clear.
Partial workaround for Intel PCI adapters reading past the end of the
host-programmed DMA regions. This change seemingly fixes the
descriptor fetches, but the packet memory accesses are left
problematic.
Reviewed by: emaste, erj, sbruno
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D8282
Pietro Cerutti [Wed, 19 Oct 2016 14:51:25 +0000 (14:51 +0000)]
Chase a cornercase in printenv and sync its behaviour with builtin's
The cornercase is when printenv is passed a parameter in the form VAR=val,
where VAR=val exists in the environment. In this case, printenv would print a
spurious newline and returns 0.
Warner Losh [Wed, 19 Oct 2016 13:45:51 +0000 (13:45 +0000)]
Document the unique requirements for upgrading more
prominently. Unlike in the past (dating back to the 4.x branch point),
you cannot upgrade from any point on the past couple of stable
branches to -current. Due to a bug in clang that existed before
r286035 in stable/9 or r286033 in stable/10, we cannot compile llvm
that's in 11.x or -current. Unfortunately, these revisions are after
9.3R and 10.3R on their respective branches. stable/11 and 11.0R are
not affected.
This also affects the upgrade path to 11 (including 11.0R) from
stable/9 and stable/10 (which would otherwise work, were it not for
this bug).
Add FFS pager, which uses buffer cache read operation to validate pages.
See the comments for more detailed description of the algorithm.
The pager is used unconditionally when the block size of the
underlying device is larger than the machine page size, since local
vnode pager cannot handle the configuration [1]. Otherwise, the
vfs.ffs.use_buf_pager sysctl allows to switch to the local pager.
Measurements demonstrated no regression in the ever-important
buildworld benchmark, and small (~5%) throughput improvements in the
special microbenchmark configuration for dbench over swap-backed
md(4).
Code can be generalized and reused for other filesystems which use
buffer cache.
Adrian Chadd [Wed, 19 Oct 2016 05:03:46 +0000 (05:03 +0000)]
[net80211] Initial full-offload scan support.
This is a very simple addition to the net80211 scan support.
It doesn't implement a replacement scan interface - it just disables
the pieces that we should disable to make this lifecycle a bit
more managable.
There's more work to come before full scan offload support is available
but it should be good enough for driver work.
* add a flag to say "full offload"
* don't do probe requests when scanning full-offload - firmware can do that
* don't do powersave transitions and buffering - firmware can do that
tested:
* iwm(4) - STA mode
* ath10k port (local, not in freebsd-head yet)
Policy for FreeBSD/arm64 kernel config is the same one as for x86
architectures: provide GENERIC kernel bootable on as many systems
as possible. Since there is no SMP support for RPi 3 yet new kernel
config was introduced: GENERIC-UP, which is effectively GENERIC with
SMP option disabled
Alan Somers [Tue, 18 Oct 2016 23:20:49 +0000 (23:20 +0000)]
Fix C++ includability of crypto headers with static array sizes
C99 allows array function parameters to use the static keyword for their
sizes. This tells the compiler that the parameter will have at least the
specified size, and calling code will fail to compile if that guarantee is
not met. However, this syntax is not legal in C++.
This commit reverts r300824, which worked around the problem for
sys/sys/md5.h only, and introduces a new macro: min_size(). min_size(x) can
be used in headers as a static array size, but will still compile in C++
mode.
1. Use taskqueue_create() instead of taskqueue_create_fast() for both
fastpath and slowpath taskqueues.
2. Service all transmits in taskqueue threads.
3. additional stats counters for keeping track of
- bd availability
- tx buf ring not emptied in the fp task queue.
These are drained via timeout taskqueue.
- tx attempts during link down.
Luigi Rizzo [Tue, 18 Oct 2016 16:18:25 +0000 (16:18 +0000)]
remove stale and unused code from various files
fix build on 32 bit platforms
simplify logic in netmap_virt.h
The commands (in net/netmap.h) to configure communication with the
hypervisor may be revised soon.
At the moment they are unused so this will not be a change of API.
Fix cpu assignment by assuring stride is non-zero, assert that all tasks
have a valid taskqueue.
ref: https://github.com/NextBSD/NextBSD/commit/db398176234fe3ce9f8e8b671f56000f8276feba
Andrew Turner [Tue, 18 Oct 2016 13:39:55 +0000 (13:39 +0000)]
Add PCB_FP_USERMASK so we can mask off floating point flags that should
not be sent to userspace, for example the future flag to tell when we are
using floating point in the kernel.
Obtained from: ABT Systems Ltd
Sponsored by: The FreeBSD Foundation
Ruslan Bukin [Tue, 18 Oct 2016 12:58:17 +0000 (12:58 +0000)]
fstyp tests comes with pre-created EL filesystems,
but fstyp cannot detect EL filesystem on EB machine,
so exclude test files from distribution and skip the
test.
Julien Charbon [Tue, 18 Oct 2016 07:16:49 +0000 (07:16 +0000)]
Fix a double-free when an inp transitions to INP_TIMEWAIT state
after having been dropped.
This fixes enforces in_pcbdrop() logic in tcp_input():
"in_pcbdrop() is used by TCP to mark an inpcb as unused and avoid future packet
delivery or event notification when a socket remains open but TCP has closed."
Warner Losh [Tue, 18 Oct 2016 05:43:12 +0000 (05:43 +0000)]
Add a new flag to mkimg (-a num) to specify the active partition for
those partitioning schemes that have this concept. Implement it as an
override for mbr's setting 0x80 in the flags for the first partition
when we have boot code.
Marcel Moolenaar [Tue, 18 Oct 2016 01:55:07 +0000 (01:55 +0000)]
o Provide a private definition for UUIDs (mkimg_uuid_t) because
UUIDs are not portable.
o Move mkimg_uuid() to a new file and merge both gpt_uuid_enc()
and vhd_uuid_enc() into a single mkimg_uuid_enc() that lives
in the same file.
o Move the OS-specific implementation of generating a UUID to
osdep_uuidgen() and provide the implementations for FreeBSD,
macOS and Linux.
o Expect the partitioning scheme headers to be found by having
a search to the directory in which the headers live. This
avoids conflicts on non-FreeBSD machines.