Marcel Moolenaar [Wed, 25 Jun 2014 15:22:14 +0000 (15:22 +0000)]
* Handle ++x as well as x++ while converting.
* Add special case handling where normal conversion would not work
(some APIs have special names)
* Fix conversion for function calls involving ifnet
Submitted by: Sreekanth Rupavatharam <rupavath@juniper.net>
Obtained from: Juniper Networks, Inc.
xen/virtio: fix balloon drivers to not mark pages as WIRED
Prevent the Xen and VirtIO balloon drivers from marking pages as
wired. This prevents them from increasing the system wired page count,
which can lead to mlock failing because of hitting the limit in
vm.max_wired.
In the Xen case make sure pages are zeroed before giving them back to
the hypervisor, or else we might be leaking data. Also remove the
balloon_{append/retrieve} and link pages directly into the
ballooned_pages queue using the plinks.q field in the page struct.
dev/virtio/balloon/virtio_balloon.c:
- Don't allocate pages with VM_ALLOC_WIRED.
dev/xen/balloon/balloon.c:
- Don't allocate pages with VM_ALLOC_WIRED.
- Make sure pages are zeroed before giving them back to the
hypervisor.
- Remove the balloon_entry struct and the balloon_{append/retrieve}
functions and use the page plinks.q entry to link the pages
directly into the ballooned_pages queue.
Daichi GOTO [Wed, 25 Jun 2014 05:39:30 +0000 (05:39 +0000)]
Fixed an IIC timing issue between the glxiic master and a slave of
peripheral devices. When transmitting (rx) from slave to master,
sometimes nAKC delays. As a result, some slaves fails their
transmission.
Davide Italiano [Wed, 25 Jun 2014 03:54:02 +0000 (03:54 +0000)]
Continue the crusade towards a dev_clone()-free kernel, removing its
usage from dtrace. The dtrace code already uses cdevpriv(9) since FreeBSD
8, so this change should be quite harmless.
Reviewed by: markj
Approved by: markj
MFC after: never
Alan Cox [Wed, 25 Jun 2014 03:30:03 +0000 (03:30 +0000)]
Now that vm_map_insert() sets MAP_ENTRY_GROWS_{DOWN,UP} on the stack entries
that it creates (r267645), we can place the check that blocks map entry
coalescing on stack entries in vm_map_simplify_entry() where it properly
belongs.
Warner Losh [Tue, 24 Jun 2014 22:15:27 +0000 (22:15 +0000)]
Make sure that the sub-makes for unwind.h start from the CURDIR
(/usr/src) tree rather than the OBJDIR (/usr/obj) tree. This fixes
broken incremental builds with the canonical MAKESYSPATH workaround
of .../share/mk. This is a gross kludge.
Garrett Wollman [Tue, 24 Jun 2014 20:23:18 +0000 (20:23 +0000)]
Catch up with many years of changes:
o Document PF_LOCAL as being an alias for PF_UNIX
o Document POSIX standardization of this interface using AF_*
constants rather than PF_* constants, and note the three particular
families which POSIX standardizes.
o Note anticipated POSIX standardization of SOCK_CLOEXEC.
o Delete from listing protocol families that FreeBSD doesn't support
(in some cases, like PF_PUP, has never supported).
o Add to listing some current protocol families that have been
introduced in the last decade or so.
o Document the correspondence of PF_* and AF_* constants.
We should probably change the documentation to make the AF_* constants
primary, but this commit does not do so.
Put the aesni_cipher_setup() and aesni_cipher_process() functions into
the file which is compiled with SSE disabled. The functions set up
the FPU context for kernel, and compiler optimizations which could
lead to use of XMM registers before the fpu_kern_enter(9) is called or
after fpu_kern_leave(9), panic the machine.
Discussed with: jmg
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Neel Natu [Tue, 24 Jun 2014 02:02:51 +0000 (02:02 +0000)]
Provide APIs to directly get 'lowmem' and 'highmem' size directly.
Previously the sizes were inferred indirectly based on the size of the mappings
at 0 and 4GB respectively. This works fine as long as size of the allocation is
identical to the size of the mapping in the guest's address space. However, if
the mapping is disjoint then this assumption falls apart (e.g., due to the
legacy BIOS hole between 640KB and 1MB).
Add FPU_KERN_KTHR flag to fpu_kern_enter(9), which avoids saving FPU
context into memory for the kernel threads which called
fpu_kern_thread(9). This allows the fpu_kern_enter() callers to not
check for is_fpu_kern_thread() to get the optimization.
Apply the flag to padlock(4) and aesni(4). In aesni_cipher_process(),
do not leak FPU context state on error.
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Use correct names for the flags. MAP_ENTRY_GROWS_* have the same
numerical values as MAP_STACK_GROWS_*, but the former is for entries'
eflags, while the later for the cow argument of vm_map_insert().
Alexander Kabaev [Mon, 23 Jun 2014 03:45:39 +0000 (03:45 +0000)]
Restore the check for non-NULL dmatag in sndbuf_free.
The sound drivers that use own buffer management can use sndbuf_setup
and not do any busdma allocation, so the driver will end up with the
managed buffer but no valid dma map and tag for it. Avoid calling
bus_dmamem_free in such cases.
Mark Johnston [Mon, 23 Jun 2014 02:00:14 +0000 (02:00 +0000)]
Fix some bugs when fetching probe arguments in i386. Firstly ensure that
the 4 byte-aligned dtrace_invop_callsite can be found and that it
immediately follows the call to dtrace_invop(). Secondly, fix some pointer
arithmetic to account for differences between struct i386_frame and illumos'
struct frame. Finally, ensure that dtrace_getarg() isn't inlined. It works
by following a fixed number of frame pointers to the probe site, so inlining
breaks it.
Mateusz Guzik [Mon, 23 Jun 2014 01:28:18 +0000 (01:28 +0000)]
Tidy up fd-related functions called by do_execve
o assert in each one that fdp is not shared
o remove unnecessary NULL checks - all userspace processes have fdtables
and kernel processes cannot execve
o remove comments about the danger of fd_ofiles getting reallocated - fdtable
is not shared and fd_ofiles could be only reallocated if new fd was about to be
added, but if that was possible the code would already be buggy as setugidsafety
work could be undone
Mark Johnston [Mon, 23 Jun 2014 01:10:56 +0000 (01:10 +0000)]
Fix a couple of bugs on amd64 when fetching probe arguments beyond the
first five for probes entered through a UD fault (i.e. FBT probes).
Specifically, handle the fact that dtrace_invop_callsite must be
16 byte-aligned and thus may not immediately follow the call to
dtrace_invop() in dtrace_invop_start(). Also fetch register arguments and
the stack pointer through a struct trapframe instead of a struct reg.
Change suggestion how to set MAKESYSPATH as broken incremental build
workaround. Magic ".../share/mk" (search directories up to /)
does not work for f.e. /usr/src/gnu/lib/libgcc because the path
inside is starting from /usr/obj hierarchy and ends up in
/usr/share/mk, not in the /usr/src/share/mk where src.opts.mk is.
IMHO proper fixing of incremental build is needed urgently.
Mateusz Guzik [Sun, 22 Jun 2014 21:37:27 +0000 (21:37 +0000)]
Don't take filedesc lock in fdunshare().
We can read refcnt safely and only care if it is equal to 1.
If it could suddenly change from 1 to something bigger the code would be
buggy even in the previous form and transitions from > 1 to 1 are equally racy
and harmless (we copy even though there is no need).
Gavin Atkinson [Sun, 22 Jun 2014 16:35:03 +0000 (16:35 +0000)]
With the migration from GNATS to Bugzilla, we no longer need a tool to
graph GNATS PR statistics. It would be nice to have a similar tool for
the new bug database, but that would likely be a from-scratch rewrite.
John-Mark Gurney [Sun, 22 Jun 2014 10:00:33 +0000 (10:00 +0000)]
convert to using pidfile... This prevents multiple wpa_supplicants
running at the same time causing problems w/ wifi not working..
the patch will be submitted upstream... The next step if someone wants
to push it upstream is to break os_unix.c up so that all these other
utilities don't need libutil..
Mateusz Guzik [Sun, 22 Jun 2014 01:31:55 +0000 (01:31 +0000)]
fd: replace fd_nfiles with fd_lastfile where appropriate
fd_lastfile is guaranteed to be the biggest open fd, so when the intent
is to iterate over active fds or lookup one, there is no point in looking
beyond that limit.
Mark Johnston [Sat, 21 Jun 2014 19:29:40 +0000 (19:29 +0000)]
Allow creation of SDT probes from a module in which no providers are
defined. This ensures that the sdt:zfs:: probes appear despite the fact
the sdt provider is defined in the kernel rather than in zfs.ko.
Reported by: hiren
Tested by: hiren
MFC after: 2 weeks
Dimitry Andric [Sat, 21 Jun 2014 18:22:23 +0000 (18:22 +0000)]
Pull in r211435 from upstream llvm trunk (by Benjamin Kramer):
Legalizer: Add support for splitting insert_subvectors.
We handle this by spilling the whole thing to the stack and doing the
insertion as a store.
PR19492. This happens in real code because the vectorizer creates
v2i128 when AVX is enabled.
This fixes a "fatal error: error in backend: Do not know how to split
the result of this operator!" message encountered during compilation of
the net-p2p/libtorrent-rasterbar port.
Reported by: Evgeniy <iron@mail.ua>
MFC after: 3 days
Pedro F. Giffuni [Sat, 21 Jun 2014 01:43:56 +0000 (01:43 +0000)]
Revert r267675:
The code doesn't really benefit of using reallocf() in this case.
Also, the realloc() results being assigned temporary variable which
makes blind replacement with reallocf() mostly useless.
Glen Barber [Fri, 20 Jun 2014 21:53:50 +0000 (21:53 +0000)]
Fix a bug in bsdgrep(1) where patterns are not correctly
detected.
Certain criteria must be met for this bug to show up:
* the -w flag is specified, and
* neither -o or --color are specified, and
* the pattern is part of another word in the line, and
* the other word that contains the pattern occurs first
PR: 181973
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Navdeep Parhar [Fri, 20 Jun 2014 20:28:46 +0000 (20:28 +0000)]
Consider the total number of descriptors available (and not just those
that are ready to be reclaimed) when deciding whether to resume tx after
a stall.
John Baldwin [Fri, 20 Jun 2014 19:54:23 +0000 (19:54 +0000)]
Cap the percent CPU of individual threads at 100% to fix some of the
more obvious imprecision in the previous top changes.
Specifically, top uses a delta of clock_gettime() calls right after
invoking the kern.proc sysctl to fetch the process/thread list to
compute the time delta between the fetches. However, the kern.proc
sysctl handler does not run in constant time. It can spin on locks,
be preempted by an interrupt handler, etc. As a result, the time
between the gathering of stats for individual processes or threads
between subsequent kern.proc handlers can vary. If a "slow" kern.proc
run is followed by a "fast" kern.proc run, then the threads/processes
at the start of the "slow" run will have a longer time delta than the
threads/processes at the end. If the clock_gettime() time delta is
not itself skewed by preemption, then the delta may be too short for
a given thread/process resulting in a higher percent CPU than actual.
However, there is no good way to calculate the exact amount of overage,
nor to know which threads to subtract the overage from. Instead, just
punt and fix the definitely-wrong case of an individual thread having
more than 100% CPU.
Test LD_LIBRARY_PATH_FDS by linking a binary that requires a shared
library that isn't in any of the usual search paths. Ensure this fails
when we don't supply LD_LIBRARY_PATH_FDS or we pass invalid information
in it. Ensure it works when we pass the correct directory in various
places in the variable.
Add the LD_LIBRARY_PATH_FDS environmental variable.
This variable allows the loading of shared libraries via directory descriptors
rather than via library paths. If LD_LIBRARY_PATH_FDS=3:4:12, the directories
represented by file descriptors 3, 4 and 12 will searched for shared libraries
before the normal path-based mechanisms are used. This allows us to execute
unprivileged binaries from within a Capsicum sandbox even if they require
shared libraries.
Michael Tuexen [Fri, 20 Jun 2014 13:26:49 +0000 (13:26 +0000)]
Honor jails for unbound SCTP sockets when selecting source addresses,
reporting IP-addresses to the peer during the handshake, adding
addresses to the host, reporting the addresses via the sysctl
interface (used by netstat, for example) and reporting the
addresses to the application via socket options.
This issue was reported by Bernd Walter.
Restore the ABI of the cpuctl(4) ioctl request CPUCTL_CPUID, use
separate argument structure with added level_type field for
CPUID_CPUID_COUNT request.
Reviewed by: attilio (previous version)
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Assert that the new entry is inserted into the right location in the
map entries list, and that it does not overlap with the previous and
next entries.
Reviewed by: alc
Sponsored by: The FreeBSD Foundation
MFC after: 1 week
Bryan Venteicher [Fri, 20 Jun 2014 02:49:03 +0000 (02:49 +0000)]
Handle multiple calls to rxq_eof for single packet completion
This requires the VMware vmxnet3 device to flip the start of packet
descriptor's generation before the rest of the packet's descriptors
have been loaded into the Rx ring. I've never observed this behavior,
and it seems to make the most sense not to do it this way. But it is
not a lot of work for the driver to handle this situation just in case.
Alexander Kabaev [Fri, 20 Jun 2014 01:45:03 +0000 (01:45 +0000)]
Set target->sbp field to valid value when sbp device is created.
The sbp_cam_detach_target can be called from sbp_post_explore function
on the first target that is not really attached and it was written with
the corresponding safety check in place to tolerate that. Unfortunately
the recent locking cleanup did add a locking assertion that tries to
dereference the target->sbp pointer unconditionally, which causes less
than desirable outcome. Since the assertion is useful, just initialize
the target sbp pointer once when sbp device is being initialized instead
of when the target is being attached. This makes assertion work in all
cases and fixes the crash on boot.
Attilio Rao [Thu, 19 Jun 2014 21:54:41 +0000 (21:54 +0000)]
Following comments in r242565 add the possibility to specify ecx when
performing cpuid calls.
Add also a new way to specify the level type to cpucontrol(8) as
reported in the manpage.
Alan Cox [Thu, 19 Jun 2014 21:05:07 +0000 (21:05 +0000)]
Eliminate a pointless call to vm_map_clip_start() from vm_map_growstack().
For this call to do anything at all we would have to have two overlapping
map entries.
John Baldwin [Thu, 19 Jun 2014 18:35:14 +0000 (18:35 +0000)]
Trust the state of a power resource that get from a working _STA method
instead of trying to cache it.
Previously, we only trusted the state if we did not have a cached state.
However, once a state was cached, the _STA method was always ignored.
Specifically, once a power resource had been turned on once (e.g.
during resume), the driver assumed it was always on even if _STA said it
was off and never turned it back on. This prevented the power resource
from being turned back on if a laptop was resumed twice, for example.
To fix, just remove the cached state entirely and always use the results
of _STA. The loops already skip any resources where _STA fails.
Alan Cox [Thu, 19 Jun 2014 16:26:16 +0000 (16:26 +0000)]
When MAP_STACK_GROWS_{DOWN,UP} are passed to vm_map_insert() set the
corresponding flag(s) in the new map entry. Previously, the caller was
responsible for setting them after vm_map_insert() returned.
Pass MAP_STACK_GROWS_DOWN to vm_map_insert() from vm_map_growstack() when
extending the stack in the downward direction.
Together these changes slightly simplify the caller's task when creating a
downward growing stack. In particular, the caller no longer needs to clip
the previous entry, because the new stack entry can't possibly coalesce
with the previous entry.
Alexander Motin [Thu, 19 Jun 2014 13:19:35 +0000 (13:19 +0000)]
Execute task management request directly in ctl_queue() context.
From one side it allows to remove CTL_FLAG_TASK_PENDING flag, handling of
which significantly complicates fine-grained locking. From the other side
it reduces task management requests latency even below then that flag could.
As downside, it denies task management code to sleep, but that is not needed
any way now.
Alexander Motin [Thu, 19 Jun 2014 12:43:41 +0000 (12:43 +0000)]
Add some more CTL_FLAG_ABORT check points.
This should allow to abort commands doing mostly disk I/O, such as VERIFY
or WRITE SAME. Before this change CTL_FLAG_ABORT was only checked around
data moves, which for these commands may not happen for a very long time.