Commit Graph

628 Commits

Author SHA1 Message Date
kib
42e09be9c2 If libthr.so is dlopened without RTLD_GLOBAL flag, the libthr symbols
do not participate in the global symbols namespace, but rtld locks are
still replaced and functions are interposed.  In particular,
__pthread_map_stacks_exec is resolved to the libc version.  If a
library is loaded later, which requires adjustment of the stack
protection mode, rtld calls into libc __pthread_map_stacks_exec due to
the symbols scope.  The libc version might recurse into binder and
recursively acquire rtld bind lock, causing the hang.

Make libc __pthread_map_stacks_exec() interposed, which synchronizes
rtld locks and version of the stack exec hook when libthr loaded,
regardless of the symbol scope control or symbol resolution order.

The __pthread_map_stacks_exec() symbol is removed from the private
version in libthr since libc symbol now operates correctly in presence
of libthr.

Reported and tested by:	markj
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2016-02-08 19:24:13 +00:00
br
aad03cf95c Add pthread MD part for RISC-V.
Reviewed by:	andrew
Sponsored by:	DARPA, AFRL
Sponsored by:	HEIF5
Differential Revision:	https://reviews.freebsd.org/D5063
2016-01-27 14:10:50 +00:00
vangyzen
e39bbe5207 libthr: const-ify two variables
Make the default umutex and urwlock initializers const,
because they can be, and as a microoptimization.

MFC after:	5 days
Sponsored by:	Dell Inc.
2016-01-13 22:34:55 +00:00
kib
27f78ada40 Typo. 2015-12-27 01:14:42 +00:00
jilles
507e66a915 libthr: Don't use both __sys_open() and __sys_openat(). 2015-12-20 16:33:56 +00:00
br
e10e8212ee Start support for the RISC-V 64-bit architecture developed by UC Berkeley.
RISC-V is a new ISA designed to support computer research and education, and
is now become a standard open architecture for industry implementations.

This is a minimal set of changes required to run 'make kernel-toolchain'
using external (GNU) toolchain.

The FreeBSD/RISC-V project home: https://wiki.freebsd.org/riscv.

Reviewed by:	andrew, bdrewery, emaste, imp
Sponsored by:	DARPA, AFRL
Sponsored by:	HEIF5
Differential Revision:	https://reviews.freebsd.org/D4445
2015-12-11 22:55:23 +00:00
bdrewery
bb97a3d43c Fix LDADD/DPADD that should be LIBADD.
Sponsored by:	EMC / Isilon Storage Division
2015-12-04 03:17:47 +00:00
bdrewery
e13d6f8b3f META MODE: Prefer INSTALL=tools/install.sh to lessen the need for xinstall.host.
This both avoids some dependencies on xinstall.host and allows
bootstrapping on older releases to work due to lack of at least 'install -l'
support.

Sponsored by:	EMC / Isilon Storage Division
2015-11-25 19:10:28 +00:00
ngie
dd1f618367 Refactor the test/ Makefiles after recent changes to bsd.test.mk (r289158) and
netbsd-tests.test.mk (r289151)

- Eliminate explicit OBJTOP/SRCTOP setting
- Convert all ad hoc NetBSD test integration over to netbsd-tests.test.mk
- Remove unnecessary TESTSDIR setting
- Use SRCTOP where possible for clarity

MFC after: 2 weeks
Sponsored by: EMC / Isilon Storage Divison
2015-10-12 08:16:03 +00:00
kib
b522c482ce Style. Use ANSI definition, wrap long lines, no initialization in
declaration for locals.

Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-09-08 08:48:53 +00:00
kib
36ed4c2481 In the pthread_once(), if the initializer has already run, then the
calling thread is supposed to see accesses issued by the initializer.
This means that the read of the once_control->state variable should
have an acquire semantic.  Use atomic_thread_fence_acq() when the
value read is ONCE_DONE, instead of straightforward atomic_load_acq(),
to only put a barrier when needed (*).

On the other hand, the updates of the once_control->state with the
intermediate progress state do not need to synchronize with other
state accesses, remove _acq suffix.

Reviewed by:	alc (previous version)
Suggested by:	alc (*)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-09-08 08:41:07 +00:00
kib
def7a72957 Pre-resolve symbols required for the deferred signal processing. This
avoids recursion into rtld when leaving libthr critical section for
the deferred signal delivery.

For the same reason, use syscall(2) instead of referencing
__sys_sigreturn(2).  Syscall() is already pre-resolved for fork()
interceptor.

Tested by:	Andre Meiser <ortadur@web.de>
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-08-10 17:02:42 +00:00
vangyzen
c03b18934d Disable SSE in libthr
Clang emits SSE instructions on amd64 in the common path of
pthread_mutex_unlock.  If the thread does not otherwise use SSE,
this usage incurs a context-switch of the FPU/SSE state, which
reduces the performance of multiple real-world applications by a
non-trivial amount (3-5% in one application).

Instead of this change, I experimented with eagerly switching the
FPU state at context-switch time.  This did not help.  Most of the
cost seems to be in the read/write of memory--as kib@ stated--and
not in the #NM handling.  I tested on machines with and without
XSAVEOPT.

One counter-argument to this change is that most applications already
use SIMD, and the number of applications and amount of SIMD usage
are only increasing.  This is absolutely true.  I agree that--in
general and in principle--this change is in the wrong direction.
However, there are applications that do not use enough SSE to offset
the extra context-switch cost.  SSE does not provide a clear benefit
in the current libthr code with the current compiler, but it does
provide a clear loss in some cases.  Therefore, disabling SSE in
libthr is a non-loss for most, and a gain for some.

I refrained from disabling SSE in libc--as was suggested--because
I can't make the above argument for libc.  It provides a wide variety
of code; each case should be analyzed separately.

https://lists.freebsd.org/pipermail/freebsd-current/2015-March/055193.html

Suggestions from:	dim, jmg, rpaulo
Approved by:	kib (mentor)
MFC after:	2 weeks
Sponsored by:	Dell Inc.
2015-08-05 12:53:55 +00:00
pfg
3702cad714 cosmetic: whitespaces-tab before EOL
Obtained from:	cpi-llvm project
2015-07-08 16:35:24 +00:00
sjg
852129abd1 new depends 2015-06-16 23:37:19 +00:00
bapt
594e07bd1b Revert r284417 it is not necessary anymore 2015-06-15 19:28:07 +00:00
bapt
9fb85ece8b Enforce overwritting SHLIBDIR
Since METAMODE has been added, sys.mk loads bsd.mkopt.mk which ends load loading
bsd.own.mk which then defines SHLIBDIR before all the Makefile.inc everywhere.

This makes /lib being populated again.

Reported by:	many
2015-06-15 15:34:20 +00:00
kib
19fcbfd486 Fix typo in comment.
MFC after:	3 days
2015-06-14 19:19:46 +00:00
sjg
008d7c831f Add META_MODE support.
Off by default, build behaves normally.
WITH_META_MODE we get auto objdir creation, the ability to
start build from anywhere in the tree.

Still need to add real targets under targets/ to build packages.

Differential Revision:       D2796
Reviewed by: brooks imp
2015-06-13 19:20:56 +00:00
sjg
75a137820d dirdeps.mk now sets DEP_RELDIR 2015-06-08 23:35:17 +00:00
sjg
65145fa4c8 Merge sync of head 2015-05-27 01:19:58 +00:00
kib
40d3cfa056 Some third-party malloc(3) implementations use pthread_setspecific(3)
to handle per-thread information.  Since our pthread_setspecific()
implementation calls calloc(3) to allocate per-thread specific data
storage, things get complicated.

Switch the allocator to use bare mmap(2).  There is some loss of the
allocated page, since e.g. on amd64, PTHREAD_KEYS_MAX * sizeof(struct
pthread_specific_elem) is 3K (it actually spans whole page due to
padding), but I believe it is more acceptable than additional code for
specialized allocator().

The alternatives would either to make the specific data array be part of
the struct thread, or use internal bindings to call the libc malloc,
avoiding interposing.

Also do the style pass over the thr_spec.c, esp. simplify the
conditionals nesting by returning early when an error detected.
Remove trivial comments.

Found by:	yuri@rawbw.com
PR:	200138
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2015-05-15 08:40:17 +00:00
andrew
d2f646661f Disable the tests that use makecontext on arm64, it still needs to be
written.
2015-04-27 13:56:20 +00:00
ngie
e70955b9f3 Build/install libc, librt, libthr, and msun NetBSD test suites on all
architectures

MFC after: 1 week
2015-04-27 06:49:27 +00:00
pfg
90a59d1e9b _pthread_cleanup_push: fix allocator sizeof operand mismatch
Same fix appears to be in DragonFly's libthread_xu.

Found by:	Clang Static Analyzer
MFC after:	1 week
2015-04-22 16:51:21 +00:00
kib
2254748ed0 The lseek(2), mmap(2), truncate(2), ftruncate(2), pread(2), and
pwrite(2) syscalls are wrapped to provide compatibility with pre-7.x
kernels which required padding before the off_t parameter.  The
fcntl(2) contains compatibility code to handle kernels before the
struct flock was changed during the 8.x CURRENT development.  The
shims were reasonable to allow easier revert to the older kernel at
that time.

Now, two or three major releases later, shims do not serve any
purpose.  Such old kernels cannot handle current libc, so revert the
compatibility code.

Make padded syscalls support conditional under the COMPAT6 config
option.  For COMPAT32, the syscalls were under COMPAT6 already.

Remove WITHOUT_SYSCALL_COMPAT build option, which only purpose was to
(partially) disable the removed shims.

Reviewed by:	jhb, imp (previous versions)
Discussed with:	peter
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-04-18 21:50:13 +00:00
kib
9a774084c8 Make wait6(2), waitid(3) and ppoll(2) cancellation points. The
waitid() function is required to be cancellable by the standard.  The
wait6() and ppoll() follow the other syscalls in their groups.

Reviewed by:	jhb, jilles (previous versions)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-04-18 21:35:41 +00:00
andrew
f98580c864 Add pthread_md.h for arm64.
Differential Revision:	https://reviews.freebsd.org/D2137
Reviewed by:	kib
Sponsored by:	The FreeBSD Foundation
2015-03-30 19:10:09 +00:00
kib
6531ee3ae5 Make kevent(2) a cancellation point.
Note that to cancel blocked kevent(2) call, changelist must be empty,
since we cannot cancel a call which already made changes to the
process state.  And in reverse, call which only makes changes to the
kqueue state, without waiting for an event, is not cancellable.  This
makes a natural usage model to migrate kqueue loop to support
cancellation, where existing single kevent(2) call must be split into
two: first uncancellable update of kqueue, then cancellable wait for
events.

Note that this is ABI-incompatible change, but it is believed that
there is no cancel-safe code that relies on kevent(2) not being a
cancellation point.  Option to preserve the ABI would be to keep
kevent(2) as is, but add new call with flags to specify cancellation
behaviour, which only value seems to add complications.

Suggested and reviewed by:	jilles
Sponsored by:	The FreeBSD Foundation
MFC after:	2 weeks
2015-03-29 19:14:41 +00:00
andrew
0c72282747 Start to import support for the AArch64 architecture from ARM. This change
only adds support for kernel-toolchain, however it is expected further
changes to add kernel and userland support will be committed as they are
reviewed.

As our copy of binutils is too old the devel/aarch64-binutils port needs
to be installed to pull in a linker.

To build either TARGET needs to be set to arm64, or TARGET_ARCH set to
aarch64. The latter is set so uname -p will return aarch64 as existing
third party software expects this.

Differential Revision:	https://reviews.freebsd.org/D2005
Relnotes:	Yes
Sponsored by:	The FreeBSD Foundation
2015-03-19 13:53:47 +00:00
jkim
17c127beed Fix a typo in comment and explain the reason. 2015-03-09 20:26:42 +00:00
kib
0e077650ee Propagate errors from _thr_umutex_unlock2 through mutex_unlock_common.
Errors from _thr_umutex_unlock2 should "never happen" in normal
circumstances.  If they do, however, return them to the application
so it can fail early and loudly.  Hiding the errors will only delay
the inevitable failure, making it harder to find and diagnose.

Submitted by:	Eric van Gyzen <eric_van_gyzen@dell.com>
Obtained from:	Dell Inc.
PR:	198914
MFC after:	1 week
2015-02-25 16:18:26 +00:00
kib
1cd0dffdca Properly interpose libc spinlocks, was missed in r276630. In
particular, stdio locking was affected.

Reported and tested by:	"Matthew D. Fuller" <fullermd@over-yonder.net>
Sponsored by:	The FreeBSD Foundation
MFC after:	3 days
2015-02-14 11:47:40 +00:00
kib
036bae5626 Update libthr(3) man page to reflect the work done to support dlopen.
Noted and reviewed by:	bdrewery
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-02-12 17:16:54 +00:00
kib
98a3318e37 Fully initialize allocated memory for the new barrier. The
b_destroying member was left uninitialized, which caused spurious
EBUSY.

PR:	197365
Noted by:	Florent Guiliani <fguiliani@verisign.com>
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-02-06 12:18:38 +00:00
andrew
4721e5b10a Merge all the copies of _tcb_ctor and _tcb_dtor.
The amd64, i386, and sparc64 versions were identical, with the one
difference where the former two used inline asm instead of _tcb_get. I
have compared the function before and after replacing the asm with _tcb_get
and found the object files to be identical.

The arm, mips, and powerpc versions were almost identical. The only
difference was the powerpc version used an alignment of 1 where arm and
mips used 16. As this is an increase in alignment is will be safe.

Along with this arm, mips, and powerpc all passed, when initial was true,
the value returned from _tcb_get as the first argument to
_rtld_allocate_tls. This would then return this pointer back to the caller.
We can remove these extra calls by checking if initial is set and setting
the thread control block directly. As this is what the sparc64 code does
we can use it directly.

As after these observations all the architectures can now have identical
code we can merge them into a common file.

Differential Revision:	https://reviews.freebsd.org/D1556
Reviewed by:	kib
Sponsored by:	The FreeBSD Foundation
2015-01-21 16:41:05 +00:00
kib
6ef4d0dc99 Fix bug in r276630. Do not allow pthread_sigmask() to block SIGCANCEL.
Reported and tested by:	royger
Sponsored by:	The FreeBSD Foundation
MFC after:	3 days
2015-01-21 16:13:37 +00:00
kib
aa476ee143 Reduce the size of the interposing table and amount of
cancellation-handling code in the libthr.  Translate some syscalls
into their more generic counterpart, and remove translated syscalls
from the table.

List of the affected syscalls:
creat, open -> openat
raise -> thr_kill
sleep, usleep -> nanosleep
pause -> sigsuspend
wait, wait3, waitpid -> wait4

Suggested and reviewed by:	jilles (previous version)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-01-11 22:16:31 +00:00
jhibbits
5de1b857a0 Avoid use of register variables. Clang 3.5 treats this as undefined behavior,
and bad things happen.

MFC after:	1 week
2015-01-06 03:50:43 +00:00
kib
6b2710fe56 Avoid calling internal libc function through PLT or accessing data
though GOT, by staticizing and hiding.  Add setter for
__error_selector to hide it as well.

Suggested and reviewed by:	jilles
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-01-05 01:06:54 +00:00
kib
8d1dfb4106 Fix known issues which blow up the process after dlopen("libthr.so")
(or loading a dso linked to libthr.so into process which was not
linked against threading library).

- Remove libthr interposers of the libc functions, including
  __error(). Instead, functions calls are indirected through the
  interposing table, similar to how pthread stubs in libc are already
  done.  Libc by default points either to syscall trampolines or to
  existing libc implementations.  On libthr load, libthr rewrites the
  pointers to the cancellable implementations already in libthr.  The
  interposition table is separate from pthreads stubs indirection
  table to not pull pthreads stubs into static binaries.

- Postpone the malloc(3) internal mutexes initialization until libthr
  is loaded.  This avoids recursion between calloc(3) and static
  pthread_mutex_t initialization.

- Reinstall signal handlers with wrapper on libthr load.  The
  _rtld_is_dlopened(3) is used to avoid useless calls to sigaction(2)
  when libthr is statically referenced from the main binary.

In the process, fix openat(2), swapcontext(2) and setcontext(2)
interposing.  The libc symbols were exported at different versions
than libthr interposers.  Export both libc and libthr versions from
libc now, with default set to the higher version from libthr.

Remove unused and disconnected swapcontext(3) userspace implementation
from libc/gen.

No objections from:	deischen
Tested by:	pho, antoine (exp-run) (previous versions)
Sponsored by:	The FreeBSD Foundation
MFC after:	1 week
2015-01-03 18:38:46 +00:00
emaste
fda27c9937 Revert r274772: it is not valid on MIPS
Reported by:	sbruno
2014-11-25 03:50:31 +00:00
emaste
c7e313326d Use canonical __PIC__ flag
It is automatically set when -fPIC is passed to the compiler.

Reviewed by:	dim, kib
Sponsored by:	The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D1179
2014-11-21 02:05:48 +00:00
sjg
b137080f19 Merge from head@274682 2014-11-19 01:07:58 +00:00
ngie
7e7662c988 Add reachover Makefiles for contrib/netbsd-tests/lib/libpthread as
lib/libthr/tests

A variant of this code has been tested on amd64/i386 for some time by
EMC/Isilon on 10-STABLE/11-CURRENT. It builds on other architectures, but the
code will remain off until it's proven it works on virtual hardware or real
hardware on other architectures

Original work by: pho
Sponsored by: EMC / Isilon Storage Division
2014-11-16 06:35:20 +00:00
pluknet
f0e5f3d683 Fix description of mutex acquisition.
Reviewed by:	kib
X-MFC with:	r272070
Sponsored by:	Nginx, Inc.
2014-09-26 04:33:27 +00:00
kib
b1bed1b450 Expand the libthr(3) manpage to document knobs accepted by libthr.so
and explain some internal working of the library, neccessary to
understand the knobs effects.

Reviewed by:	bjk, pluknet
Sponsored by:	The FreeBSD Foundation
MFC after:	3 weeks
2014-09-24 12:41:39 +00:00
kib
a14de80022 Switch the defaults to not split the RLIMIT_STACK-sized initial thread
stack into the stacks of the created threads.  Add knob
LIBPTHREAD_SPLITSTACK_MAIN to restore the older behaviour.

Sponsored by:	The FreeBSD Foundation
MFC after:	3 weeks
2014-09-24 12:39:12 +00:00
rpaulo
d9d51ad995 Fix typo in a comment. 2014-09-02 18:21:19 +00:00
sjg
d7cd1d425c Merge head from 7/28 2014-08-19 06:50:54 +00:00