Fix EtherIP. TOS field must be initialized when the inner protocol is
PF_LINK, and multicast/broadcast flag should always be dropped because
the outer protocol uses unicast even when the inner address is not for
unicast. It had been broken since r236951 when gif_output() started to
use IFQ_HANDOFF().
Correct assertion. The shadowing object cannot be tmpfs vm object,
and tmpfs object cannot shadow. In other words, tmpfs vm object is
always at the bottom of the shadow chain.
Reported and tested by: bdrewery
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
marius [Thu, 24 Jul 2014 10:14:51 +0000 (10:14 +0000)]
Intel desktop Haswell CPUs may report benign corrected parity errors (see
HSD131 erratum in [1]) at a considerable rate. So filter these (default),
unless logging is enabled. Unfortunately, there really is no better way to
reasonably implement suppressing these errors than to just skipping them
in mca_log(). Given that they are reported for bank 0, they'd need to be
masked in MSR_MC0_CTL. However, P6 family processors require that register
to be set to either all 0s or all 1s, disabling way more than the one error
in question when using all 0s there. Alternatively, it could be masked for
the corresponding CMCI, but that still wouldn't keep the periodic scanner
from detecting these spurious errors. Apart from that, register contents of
MSR_MC0_CTL{,2} don't seem to be publicly documented, neither in the Intel
Architectures Developer's Manual nor in the Haswell datasheets.
Note that while HSD131 actually is only about C0-stepping as of revision
014 of the Intel desktop 4th generation processor family specification
update, these corrected errors also have been observed with D0-stepping
aka "Haswell Refresh".
marius [Thu, 24 Jul 2014 10:12:22 +0000 (10:12 +0000)]
Copying pages via temporary mappings in the !DMAP case of pmap_copy_pages()
involves updating the corresponding page tables followed by accesses to the
pages in question. This sequence is subject to the situation exactly described
in the "AMD64 Architecture Programmer's Manual Volume 2: System Programming"
rev. 3.23, "7.3.1 Special Coherency Considerations" [1, p. 171 f.]. Therefore,
issuing the INVLPG right after modifying the PTE bits is crucial (see also
r269050).
For the amd64 PMAP code, the order of instructions was already correct. The
above fact still is worth documenting, though.
marius [Thu, 24 Jul 2014 10:08:02 +0000 (10:08 +0000)]
- Copying and zeroing pages via temporary mappings involves updating the
corresponding page tables followed by accesses to the pages in question.
This sequence is subject to the situation exactly described in the "AMD64
Architecture Programmer's Manual Volume 2: System Programming" rev. 3.23,
"7.3.1 Special Coherency Considerations" [1, p. 171 f.]. Therefore, issuing
the INVLPG right after modifying the PTE bits is crucial.
For pmap_copy_page(), this has been broken in r124956 and later on carried
over to pmap_copy_pages() derived from the former, while all other places
in the i386 PMAP code use the correct order of instructions in this regard.
Fixing the latter breakage solves the problem of data corruption seen with
unmapped I/O enabled when running at least bare metal on AMD R-268D APUs.
However, this might also fix similar corruption reported for virtualized
environments.
- In pmap_copy_pages(), correctly set the cache bits on the source page being
copied. This change is thought to be a NOP for the real world, though. [2]
Add compat shims for XDEV and XDEV_ARCH so we don't break all the
users of them. Also, add a note to updating. We'll keep these shims at
least until the 12 branch.
The faulting instruction needs to be restarted when the exception handler
is done handling the fault. bhyve now does this correctly by setting
'vmexit[vcpu].inst_length' to zero so the %rip is not advanced.
A minor complication is that the fault injection APIs are used by instruction
emulation code that is shared by vmm.ko and bhyve. Thus the argument that
refers to 'struct vm *' in kernel or 'struct vmctx *' in userspace needs to
be loosely typed as a 'void *'.
Fake out the deprecation of XDEV and XDEV_ARCH by
continuing to use the variables in the configuration
file, but switch XDEV= and XDEV_ARCH= to TARGET= and
TARGET_ARCH= appropriately.
cxgbe(4): Keep track of the clusters that have to be freed by the
custom free routine (rxb_free) in the driver. Fail MOD_UNLOAD with
EBUSY if any such cluster has been handed up to the kernel but hasn't
been freed yet. This prevents a panic later when the cluster finally
needs to be freed but rxb_free is gone from the kernel.
Deprecate the use of XDEV and XDEV_ARCH and just use TARGET/TARGET_ARCH
for the xdev build target, which is awesome and totally works.
Reapply svn R268377 with correct name of libsupc++ here as this does
resolve one dependancy race when building the xdev target.
the xdev target builds for amd64, i386, mips, mips64 and armv6 with this commit,
must be built as root, must be built from /usr/src, must not have a /usr/obj and
places the xdev tools in /usr/$TARGET_ARCH-freebsd
the xdev target still leaves some assorted files strewn about your /usr/src when
this is done and needs to be investigated further.
Fix a bug where include file `networking/services.subr' was not being
properly clustered.
Change `fixedsize' to `fixedsize = true' to satisfy newer versions of dot(1)
generating error "Warning: <stdin>: syntax error in line 27 near ','".
Remove bounding cluster surrounding various include-groupings, allowing
groups of include files to float freely to the greatest consumer (reducing
the density of rendered diagrams).
While here, change the bgcolor of include clusters from white to X11 color
`thistle' (a light purple).
Also, add the word "Subroutines" after include cluster labels.
Do not terminate the digraph definition with a semi-colon [;] after the
ending brace to satisfy newer versions of dot(1) generating error "Warning:
<stdin>: syntax error in line 940 near ';'".
-fix a condition so that fparseln() doesn't report spurious empty lines
eg after 2 comment lines, or on EOF after a single comment line
-no escape character means no escaped characters
modify the previous fix so that no pointless realloc()s are done in
the case of multiple empty continuation lines, and comment the code
to make the logics obvious
fparseln is now part of libc in NetBSD so this changes the previous
revision numbering.
Add example script `add_some_packages.sh', demonstrating how to install
a list of packages using the bsdconfig(8) API, providing dialog(8) based
user experience (versus plain console output were you to use pkg(8)
directly to install the same list of packages).
Remove example script `browse_packages_ftp.sh', made obsolete because
the digests.txz and packagesite.txz databases for pkg(8) are not
available via FTP (HTTP only to pkg.freebsd.org SRV hosts).
Update example script `browse_packages_http.sh', made to work with new
pkg(8) demonstrating how to generate a local package repository.
Fix a bug in `bsdconfig packages' where packages were listed twice. This
fix requires pkg(8) version 1.2.7_4 or higher.
NB: It is the introduction of pkg(8) 1.2.7_4 wherein I am also able to
drastically reduce the generation time of package dependencies prior to
the dialog display (by utilizing the new `-I' flag to pkg-rquery(8)).
While here, fix a positional argument nit for f_index_initialize() of
`packages/index.subr' include (the one and only argument is positional
argument 1 to state the by-ref handle, indicating the variable to set
in the caller's namespace; the nit I'm fixing here is that we were
querying positional argument 2 for this information incorrectly; caused
by a missing hunk back around SVN r257795).
Fix a bug in sysrc(8) described by PR bin/187458 "sysrc(8) silently and
unexpectedly bootstraps pkg". This was caused by an explicit entry in
`/usr/share/bsdconfig/common.subr' (used by sysrc(8)) that called pkg(8)
to populate the $PKG_ABI global (called in a way that allows implicit
bootstrap of pkg(8)). The solution to which was to find every place in
bsdconfig(8) that requires the $PKG_ABI global and add a layer of
protection by way of introducing the new API call f_musthavepkg_init()
(provided by new include, `/usr/share/bsdconfig/packages/musthavepkg.subr'
intented to mirror `/usr/share/bsdconfig/mustberoot.subr'). When the
$PKG_ABI global is required, you can use `f_musthavepkg_init' to cause
graceful premature termination in the event that pkg(8) has not yet been
bootstrapped, and thus cannot be used to populate $PKG_ABI.
NB: If running interactively ($nonInteractive is NULL or unset), the
f_musthavepkg_init() API call will attempt to bootstrap pkg(8), but only
if the user chooses "Yes" to a Yes/No dialog confirming it is ok to
bootstrap.
While here, simplify an if-conditional in `media/http.subr' include to
use sh(1) inline assignment-with-break.
Also, fix a bug in `media/http.subr' and `media/httpproxy.subr' wherein
the error messages for various HTTP failures were not finding their way
to the console (needed to pass stdout to pass-thru descriptor).
While here, remove the executable bit from `packages/categories.subr',
`packages/index.subr', and `packages/packages.subr' includes.
Fix issues where pkg(8) complains about PACKAGESITE being defined.
Previously, we would set $PACKAGESITE and export it. Now we only set
$PACKAGESITE for invocations of "pkg update" -- getting rid of all the
spurious warnings about PACKAGESITE being deprecated (it's still used
in the case of "pkg update" for simplicity versus having to configure
a config-file).
Remove the leading argument from invocations of f_index_initialize() in
`packages/packages.subr' include. The leading argument no longer means
what it used to, pre SVN r257995).
/usr/bin/sendbug and the sendbug(1) man page were part of send-pr and
are no longer installed since r267486. Add them to ObsoleteFiles.inc1,
which should have been done in r267486.
PR: 192010
Submitted by: Vick Khera
Pointyhat to: gavin
MFC after: 3 days
For "a"-mode files and rewind/fseek + fwrite combination return meaningful
value now, like Apple does, but avoid their __sflush physical write
performance degradation as much as possible.
ian [Tue, 22 Jul 2014 20:32:24 +0000 (20:32 +0000)]
Fix EABI unwind function declarations to match the function definitions.
The mismatch was noticed by clang, when using it to compile the in-tree
gcc and libstdc++.
In r264504, we prevented doing I/O for more than MAXPHYS by making
the assumption that consumers would respect bio_completed and/or
bio_resid to detect short reads. This assumption proved false and
file corruption was the result.
Create as many bios as we need to satisfy the original request.
Check the cached chunk every time we need to do I/O to increase the
hit rate.
Obtained from: junipre Networks, Inc.
MFC after: 1 week
bsd.lib.mk does not add OBJS etc to CLEANFILES so does not automatically
get them flagged as .NOPATH. This hurts people who don't use obj dirs.
Since its clean target seprate rm's for things, use NOPATH_FILES as list
to collect things that need .NOPATH.
bsd.obj.mk will add CLEANFILES to NOPATH_FILES and do the deed if needed.
peter [Mon, 21 Jul 2014 22:10:29 +0000 (22:10 +0000)]
Merge serf-1.3.6
+ Revert r2319 from serf 1.3.5: this change was making serf call handle_response
+ multiple times in case of an error response, leading to unexpected behavior.
dim [Mon, 21 Jul 2014 21:26:10 +0000 (21:26 +0000)]
Run mtree for BSD.tests.dist during make xdev-install, if the tests are
enabled (which they are in the default configuration). Otherwise, it
will fail because ${XDDESTDIR}/usr/include/atf-c does not exist.
Merge change from upstream linux kernel submitted by OpenBSD:
drm/radeon: fix-up some float to fixed conversion thinkos
Spotted by Brad Smith when porting to OpenBSD. Noticed-by: Brad Smith <brad@comstyle.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
ref: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=1cd73ff7
Remove #ifdef DUMBBELL_WIP in favor of upstream fix.
This change is a bit ugly, but so is the coupling between the i915
driver and syscons. It isn't worth developing a more elegant solution
only to support the legacy syscons console.
The hcreate(3) implementation and related functions we inherited
from NetBSD used to free() the key value, something that is not
supported by the standard implementation.
This would cause a segmentation fault when attempting to run
the examples from the opengroup and linux manpages. NetBSD
has added non-standard calls to provide the previous
behaviour but hdestroy is not very commonly used so at this
time it seems excessive to bring those to FreeBSD.
Bump the __FreeBSD_version as this is an ABI change.
In "procstat -v" (VM), spell out 'FL' to 'FLAG' since there are two extra
columns available anyway. Also left align as we tend to do for flags
fields, although you can't see that currently as the string fully fills
that available columns.
des [Sun, 20 Jul 2014 14:49:24 +0000 (14:49 +0000)]
Older versions of flex (like the one used in FreeBSD <= 9) expect the
output file name to follow the -o flag immediately, with no intervening
whitespace.
adrian [Sun, 20 Jul 2014 07:45:48 +0000 (07:45 +0000)]
Teach ixgbe(4) about rss_gethashconfig().
If RSS is enabled, ixgbe(4) will query the RSS API for the types of hashes
which should be used. It'll then only enable hashes that are exposed via
the RSS layer.
This way it won't try to do things like enable UDP hashing if RSS explicitly
states that it isn't supported in lookups.
adrian [Sun, 20 Jul 2014 07:43:41 +0000 (07:43 +0000)]
Disable the ixgbe(4) UDP 4-tuple hashing for the time being.
A mix of fragmented and non-fragmented UDP in a single stream will end up
being hashed differently, resulting in out-of-order behaviour in the receive
path.
adrian [Sun, 20 Jul 2014 07:36:59 +0000 (07:36 +0000)]
Implement rss_gethashconfig() - return the currently supported hash methods
by the stack.
Right now the stack isn't really setup for RSS with 4-tuple UDP hashing
for either IPv4 and IPv6.
The specifics:
* The UDP init path udp_init() and udplite_init() specify the hash as
2-tuple, so the PCBGROUPS code only tries a 2-tuple check;
* The PCBGROUPS and RSS code doesn't know about the UDP hash types
just yet, so they're never treated as valid hashes.
* For correctness, 4-tuple can't be enabled in the general case because
UDP datagrams can be more fragmented than IP datagrams may be.
Strictly speaking, TCP datagrams may also be fragmented and this could
cause issues with PCBGROUPS/RSS until the IP defragment path grows some
code to re-calculate the RSS hash.
I'll follow this commit up with awareness of the UDP 4-tuple for those
who wish to configure it, but for now it'll stay disabled.
No drivers (yet) know to use this function when RSS is enabled.
ian [Sat, 19 Jul 2014 22:13:27 +0000 (22:13 +0000)]
Add dl_unwind_find_exidx() for ARM EABI, required for C++ exception
handling. For statically linked apps this uses the __exidx_start/end
symbols set up by the linker. For dynamically linked apps it finds the
shared object that contains the given address and returns the location and
size of the exidx section in that shared object.
The dl_unwind_find_exidx() name is used by other BSD projects and Android,
and is mentioned in clang 3.5 comments as "the BSD interface" for finding
exidx data. GCC (in libgcc_s) expects the exact same API and functionality
to be provided by a function named __gnu_Unwind_Find_exidx(), so we provide
that with an alias ("strong reference").
A nested exception condition arises when a second exception is triggered while
delivering the first exception. Most nested exceptions can be handled serially
but some are converted into a double fault. If an exception is generated during
delivery of a double fault then the virtual machine shuts down as a result of
a triple fault.
vm_exit_intinfo() is used to record that a VM-exit happened while an event was
being delivered through the IDT. If an exception is triggered while handling
the VM-exit it will be treated like a nested exception.
vm_entry_intinfo() is used by processor-specific code to get the event to be
injected into the guest on the next VM-entry. This function is responsible for
deciding the disposition of nested exceptions.
des [Sat, 19 Jul 2014 20:55:13 +0000 (20:55 +0000)]
Check if the specified group is the user's primary group before
iterating over the (possibly empty) list of members. Otherwise, we
get a false negative when the target group has no members listed in
/etc/group. This went mostly unnoticed because root is explicitly
listed as a member of wheel, so the bug is never triggered in the most
common use case, which is su(8).