Austin Group issue 62 [1] dropped the async-signal-safe requirement
for fork and provided a async-signal-safe _Fork replacement that
does not run the atfork handlers. It will be included in the next
POSIX standard.
It allow to close a long standing issue to make fork AS-safe (BZ#4737).
As indicated on the bug, besides the internal lock for the atfork
handlers itself; there is no guarantee that the handlers itself will
not introduce more AS-safe issues.
The idea is synchronize fork with the required internal locks to allow
children in multithread processes to use mostly of standard function
(even though POSIX states only AS-safe function should be used). On
signal handles, _Fork should be used intead and only AS-safe functions
should be used.
For testing, the new tst-_Fork only check basic usage. I also added
a new tst-mallocfork3 which uses the same strategy to check for
deadlock of tst-mallocfork2 but using threads instead of subprocesses
(and it does deadlock if it replaces _Fork with fork).
[1] https://austingroupbugs.net/view.php?id=62
The valgrind/helgrind test suite needs a way to make stack dealloction
more prompt, and this feature seems to be generally useful.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
AMD define different flags for IRPB, IBRS, and STIPBP [1], so new
x86_64_cpu are added and IBRS_IBPB is only tested for Intel.
The SSDB is also defined and implemented different on AMD [2],
and also a new AMD_SSDB flag is added. It should map to the
cpuinfo 'ssdb' on recent AMD cpus.
It fixes tst-cpu-features-cpuinfo and tst-cpu-features-cpuinfo-static
on recent AMD cpus.
Checked on x86_64-linux-gnu on AMD Ryzen 9 5900X.
[1] https://developer.amd.com/wp-content/resources/Architecture_Guidelines_Update_Indirect_Branch_Control.pdf
[2] https://bugzilla.kernel.org/show_bug.cgi?id=199889
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
* NEWS: Don't imply the default will always be 32-bit.
* manual/creature.texi (Feature Test Macros):
Say that _TIME_BITS and _FILE_OFFSET_BITS defaults
may change in future releases.
A new build flag, _TIME_BITS, enables the usage of the newer 64-bit
time symbols for legacy ABI (where 32-bit time_t is default). The 64
bit time support is only enabled if LFS (_FILE_OFFSET_BITS=64) is
also used.
Different than LFS support, the y2038 symbols are added only for the
required ABIs (armhf, csky, hppa, i386, m68k, microblaze, mips32,
mips64-n32, nios2, powerpc32, sparc32, s390-32, and sh). The ABIs with
64-bit time support are unchanged, both for symbol and types
redirection.
On Linux the full 64-bit time support requires a minimum of kernel
version v5.1. Otherwise, the 32-bit fallbacks are used and might
results in error with overflow return code (EOVERFLOW).
The i686-gnu does not yet support 64-bit time.
This patch exports following rediretions to support 64-bit time:
* libc:
adjtime
adjtimex
clock_adjtime
clock_getres
clock_gettime
clock_nanosleep
clock_settime
cnd_timedwait
ctime
ctime_r
difftime
fstat
fstatat
futimens
futimes
futimesat
getitimer
getrusage
gettimeofday
gmtime
gmtime_r
localtime
localtime_r
lstat_time
lutimes
mktime
msgctl
mtx_timedlock
nanosleep
nanosleep
ntp_gettime
ntp_gettimex
ppoll
pselec
pselect
pthread_clockjoin_np
pthread_cond_clockwait
pthread_cond_timedwait
pthread_mutex_clocklock
pthread_mutex_timedlock
pthread_rwlock_clockrdlock
pthread_rwlock_clockwrlock
pthread_rwlock_timedrdlock
pthread_rwlock_timedwrlock
pthread_timedjoin_np
recvmmsg
sched_rr_get_interval
select
sem_clockwait
semctl
semtimedop
sem_timedwait
setitimer
settimeofday
shmctl
sigtimedwait
stat
thrd_sleep
time
timegm
timerfd_gettime
timerfd_settime
timespec_get
utime
utimensat
utimes
utimes
wait3
wait4
* librt:
aio_suspend
mq_timedreceive
mq_timedsend
timer_gettime
timer_settime
* libanl:
gai_suspend
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Commit 68ab82f566 added support for the scv
syscall ABI on powerpc. Since then systems that have kernel and processor
support started using scv. However adding the proper support for a new syscall
ABI requires changes to several other projects (e.g. qemu, valgrind, strace,
kernel), which are gradually receiving support.
Meanwhile, having a way to disable scv on glibc at build time can be useful for
distros that may encounter conflicts with projects that still do not support the
scv ABI, buying time until proper support is added.
This commit adds a --disable-scv option that disables scv support and uses sc
for all syscalls, like before commit 68ab82f566.
Reviewed-by: Raphael M Zinsly <rzinsly@linux.ibm.com>
Now that thread cancellation state is not accessed concurrently anymore,
it is possible to move it out the 'cancelhandling'.
The code is also simplified: CANCELLATION_P is replaced with a
internal pthread_testcancel call and the CANCELSTATE_BIT{MASK} is
removed.
With this behavior pthread_setcancelstate does not require to act on
cancellation if cancel type is asynchronous (is already handled either
by pthread_setcanceltype or by the signal handler).
Checked on x86_64-linux-gnu and aarch64-linux-gnu.
ISO C2X has made some changes to the handling of feature test macros
related to features from the floating-point TSes, and to exactly what
such features are present in what headers, that require corresponding
changes in glibc.
* For the few features that were controlled by
__STDC_WANT_IEC_60559_BFP_EXT__ (and the corresponding DFP macro) in
C2X, there is now instead a new feature test macro
__STDC_WANT_IEC_60559_EXT__ covering both binary and decimal FP.
This controls CR_DECIMAL_DIG in <float.h> (provided by GCC; I
implemented support for the new feature test macro for GCC 11) and
the totalorder and payload functions in <math.h>. C2X no longer
says anything about __STDC_WANT_IEC_60559_BFP_EXT__ (so it's
appropriate for that macro to continue to enable exactly the
features from TS 18661-1).
* The SNAN macros for each floating-point type have moved to <float.h>
(and been renamed in the process). Thus, the copies in <math.h>
should only be defined for __STDC_WANT_IEC_60559_BFP_EXT__, not for
C2X.
* The fmaxmag and fminmag functions have been removed (replaced by new
functions for the new min/max operations in IEEE 754-2019). Thus
those should also only be declared for
__STDC_WANT_IEC_60559_BFP_EXT__.
* The _FloatN / _FloatNx handling for the last two points in glibc is
trickier, since __STDC_WANT_IEC_60559_TYPES_EXT__ is still in C2X
(the integration of TS 18661-3 as an Annex, that is, which hasn't
yet been merged into the C standard git repository but has been
accepted by WG14), so C2X with that macro should not declare some
things that are declared for older standards with that macro. The
approach taken here is to provide the declarations (when
__STDC_WANT_IEC_60559_TYPES_EXT__ is enabled) only when (defined
__USE_GNU || !__GLIBC_USE (ISOC2X)), so if C2X features are enabled
then those declarations (that are only in TS 18661-3 and not in C2X)
will only be provided if _GNU_SOURCE is defined as well. Thus
_GNU_SOURCE remains a superset of the TS features as well as of C2X.
Some other somewhat related changes in C2X are not addressed here.
There's an open proposal not to include the fmin and fmax functions
for the _FloatN / _FloatNx types, given the new min/max operations,
which could be handled like the previous point if adopted. And the
fromfp functions have been changed to return a result in floating type
rather than intmax_t / uintmax_t; my inclination there is to treat
that like that change of totalorder type (new symbol versions etc. for
the ABI change; old versions become compat symbols and are no longer
supported as an API).
Tested for x86_64 and x86.
This patch optimizes the performance of memcpy/memmove for A64FX [1]
which implements ARMv8-A SVE and has L1 64KB cache per core and L2 8MB
cache per NUMA node.
The performance optimization makes use of Scalable Vector Register
with several techniques such as loop unrolling, memory access
alignment, cache zero fill, and software pipelining.
SVE assembler code for memcpy/memmove is implemented as Vector Length
Agnostic code so theoretically it can be run on any SOC which supports
ARMv8-A SVE standard.
We confirmed that all testcases have been passed by running 'make
check' and 'make xcheck' not only on A64FX but also on ThunderX2.
And also we confirmed that the SVE 512 bit vector register performance
is roughly 4 times better than Advanced SIMD 128 bit register and 8
times better than scalar 64 bit register by running 'make bench'.
[1] https://github.com/fujitsu/A64FX
Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com>
Reviewed-by: Szabolcs Nagy <Szabolcs.Nagy@arm.com>
This patch is a test helper script to change Vector Length for child
process. This script can be used as test-wrapper for 'make check'.
Usage examples:
~/build$ make check subdirs=string \
test-wrapper='~/glibc/sysdeps/unix/sysv/linux/aarch64/vltest.py 16'
~/build$ ~/glibc/sysdeps/unix/sysv/linux/aarch64/vltest.py 16 \
make test t=string/test-memcpy
~/build$ ~/glibc/sysdeps/unix/sysv/linux/aarch64/vltest.py 32 \
./debugglibc.sh string/test-memmove
~/build$ ~/glibc/sysdeps/unix/sysv/linux/aarch64/vltest.py 64 \
./testrun.sh string/test-memset
Previously, the source file nptl/cancellation.c was compiled multiple
times, for libc, libpthread, librt. This commit switches to a single
implementation, with new __pthread_enable_asynccancel@@GLIBC_PRIVATE,
__pthread_disable_asynccancel@@GLIBC_PRIVATE exports.
The almost-unused CANCEL_ASYNC and CANCEL_RESET macros are replaced
by LIBC_CANCEL_ASYNC and LIBC_CANCEL_ASYNC macros. They call the
__pthread_* functions unconditionally now. The macros are still
needed because shared code uses them; Hurd has different definitions.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cover key corner cases (e.g., whether errno is set) that are well
settled in glibc, fix some examples to avoid integer overflow, and
update some other dated examples (code needed for K&R C, e.g.).
* manual/charset.texi (Non-reentrant String Conversion):
* manual/filesys.texi (Symbolic Links):
* manual/memory.texi (Allocating Cleared Space):
* manual/socket.texi (Host Names):
* manual/string.texi (Concatenating Strings):
* manual/users.texi (Setting Groups):
Use reallocarray instead of realloc, to avoid integer overflow issues.
* manual/filesys.texi (Scanning Directory Content):
* manual/memory.texi (The GNU Allocator, Hooks for Malloc):
* manual/tunables.texi:
Use code font for 'malloc' instead of roman font.
(Symbolic Links): Don't assume readlink return value fits in 'int'.
* manual/memory.texi (Memory Allocation and C, Basic Allocation)
(Malloc Examples, Alloca Example):
* manual/stdio.texi (Formatted Output Functions):
* manual/string.texi (Concatenating Strings, Collation Functions):
Omit pointer casts that are needed only in ancient K&R C.
* manual/memory.texi (Basic Allocation):
Say that malloc sets errno on failure.
Say "convert" rather than "cast", since casts are no longer needed.
* manual/memory.texi (Basic Allocation):
* manual/string.texi (Concatenating Strings):
In examples, use C99 declarations after statements for brevity.
* manual/memory.texi (Malloc Examples): Add portability notes for
malloc (0), errno setting, and PTRDIFF_MAX.
(Changing Block Size): Say that realloc (p, 0) acts like
(p ? (free (p), NULL) : malloc (0)).
Add xreallocarray example, since other examples can use it.
Add portability notes for realloc (0, 0), realloc (p, 0),
PTRDIFF_MAX, and improve notes for reallocating to the same size.
(Allocating Cleared Space): Reword now-confusing discussion
about replacement, and xref "Replacing malloc".
* manual/stdio.texi (Formatted Output Functions):
Don't assume message size fits in 'int'.
* manual/string.texi (Concatenating Strings):
Fix undefined behavior involving arithmetic on a freed pointer.
Finally remove all mpa related files, headers, declarations, probes, unused
tables and update makefiles.
Reviewed-By: Paul Zimmermann <Paul.Zimmermann@inria.fr>
This code adds new flag - '--allow-time-setting' to cross-test-ssh.sh
script to indicate if it is allowed to alter the date on the system
on which tests are executed. This change is supposed to be used with
test systems, which use virtual machines for testing.
The GLIBC_TEST_ALLOW_TIME_SETTING env variable is exported to the
remote environment on which the eligible test is run and brings no
functional change when it is not.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
The TUNABLE_SET interface took a primitive C type argument, which
resulted in inconsistent type conversions internally due to incorrect
dereferencing of types, especialy on 32-bit architectures. This
change simplifies the TUNABLE setting logic along with the interfaces.
Now all numeric tunable values are stored as signed numbers in
tunable_num_t, which is intmax_t. All calls to set tunables cast the
input value to its primitive type and then to tunable_num_t for
storage. This relies on gcc-specific (although I suspect other
compilers woul also do the same) unsigned to signed integer conversion
semantics, i.e. the bit pattern is conserved. The reverse conversion
is guaranteed by the standard.
1. Add CPUID_INDEX_14_ECX_0 for CPUID leaf 0x14 to detect PTWRITE feature
in EBX of CPUID leaf 0x14 with ECX == 0.
2. Add PTWRITE detection to CPU feature tests.
3. Add 2 static CPU feature tests.
The struct tag is actually entry (not ENTRY). The data member has
type void *, and it can point to binary data. Only the key member is
required to be a null-terminated string.
Reviewed-by: Arjun Shankar <arjun@redhat.com>
Add _SC_MINSIGSTKSZ for the minimum signal stack size derived from
AT_MINSIGSTKSZ, which is the minimum number of bytes of free stack
space required in order to gurantee successful, non-nested handling
of a single signal whose handler is an empty function, and _SC_SIGSTKSZ
which is the suggested minimum number of bytes of stack space required
for a signal stack.
If AT_MINSIGSTKSZ isn't available, sysconf (_SC_MINSIGSTKSZ) returns
MINSIGSTKSZ. On Linux/x86 with XSAVE, the signal frame used by kernel
is composed of the following areas and laid out as:
------------------------------
| alignment padding |
------------------------------
| xsave buffer |
------------------------------
| fsave header (32-bit only) |
------------------------------
| siginfo + ucontext |
------------------------------
Compute AT_MINSIGSTKSZ value as size of xsave buffer + size of fsave
header (32-bit only) + size of siginfo and ucontext + alignment padding.
If _SC_SIGSTKSZ_SOURCE or _GNU_SOURCE are defined, MINSIGSTKSZ and SIGSTKSZ
are redefined as
/* Default stack size for a signal handler: sysconf (SC_SIGSTKSZ). */
# undef SIGSTKSZ
# define SIGSTKSZ sysconf (_SC_SIGSTKSZ)
/* Minimum stack size for a signal handler: SIGSTKSZ. */
# undef MINSIGSTKSZ
# define MINSIGSTKSZ SIGSTKSZ
Compilation will fail if the source assumes constant MINSIGSTKSZ or
SIGSTKSZ.
The reason for not simply increasing the kernel's MINSIGSTKSZ #define
(apart from the fact that it is rarely used, due to glibc's shadowing
definitions) was that userspace binaries will have baked in the old
value of the constant and may be making assumptions about it.
For example, the type (char [MINSIGSTKSZ]) changes if this #define
changes. This could be a problem if an newly built library tries to
memcpy() or dump such an object defined by and old binary.
Bounds-checking and the stack sizes passed to things like sigaltstack()
and makecontext() could similarly go wrong.
Most packages have been tested with their latest releases, except for
Python, whose latest version is 3.9.1.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
In <sys/platform/x86.h>, define CPU features as enum instead of using
the C preprocessor magic to make it easier to wrap this functionality
in other languages. Move the C preprocessor magic to internal header
for better GCC codegen when more than one features are checked in a
single expression as in x86-64 dl-hwcaps-subdirs.c.
1. Rename COMMON_CPUID_INDEX_XXX to CPUID_INDEX_XXX.
2. Move CPUID_INDEX_MAX to sysdeps/x86/include/cpu-features.h.
3. Remove struct cpu_features and __x86_get_cpu_features from
<sys/platform/x86.h>.
4. Add __x86_get_cpuid_feature_leaf to <sys/platform/x86.h> and put it
in libc.
5. Make __get_cpu_features() private to glibc.
6. Replace __x86_get_cpu_features(N) with __get_cpu_features().
7. Add _dl_x86_get_cpu_features to GLIBC_PRIVATE.
8. Use a single enum index for each CPU feature detection.
9. Pass the CPUID feature leaf to __x86_get_cpuid_feature_leaf.
10. Return zero struct cpuid_feature for the older glibc binary with a
smaller CPUID_INDEX_MAX [BZ #27104].
11. Inside glibc, use the C preprocessor magic so that cpu_features data
can be loaded just once leading to more compact code for glibc.
256 bits are used for each CPUID leaf. Some leaves only contain a few
features. We can add exceptions to such leaves. But it will increase
code sizes and it is harder to provide backward/forward compatibilities
when new features are added to such leaves in the future.
When new leaves are added, _rtld_global_ro offsets will change which
leads to race condition during in-place updates. We may avoid in-place
updates by
1. Rename the old glibc.
2. Install the new glibc.
3. Remove the old glibc.
NB: A function, __x86_get_cpuid_feature_leaf , is used to avoid the copy
relocation issue with IFUNC resolver as shown in IFUNC resolver tests.
I used these shell commands:
../glibc/scripts/update-copyrights $PWD/../gnulib/build-aux/update-copyright
(cd ../glibc && git commit -am"[this commit message]")
and then ignored the output, which consisted lines saying "FOO: warning:
copyright statement not found" for each of 6694 files FOO.
I then removed trailing white space from benchtests/bench-pthread-locks.c
and iconvdata/tst-iconv-big5-hkscs-to-2ucs4.c, to work around this
diagnostic from Savannah:
remote: *** pre-commit check failed ...
remote: *** error: lines with trailing whitespace found
remote: error: hook declined to update refs/heads/master
Introduce a new _FORTIFY_SOURCE level of 3 to enable additional
fortifications that may have a noticeable performance impact, allowing
more fortification coverage at the cost of some performance.
With llvm 9.0 or later, this will replace the use of
__builtin_object_size with __builtin_dynamic_object_size.
__builtin_dynamic_object_size
-----------------------------
__builtin_dynamic_object_size is an LLVM builtin that is similar to
__builtin_object_size. In addition to what __builtin_object_size
does, i.e. replace the builtin call with a constant object size,
__builtin_dynamic_object_size will replace the call site with an
expression that evaluates to the object size, thus expanding its
applicability. In practice, __builtin_dynamic_object_size evaluates
these expressions through malloc/calloc calls that it can associate
with the object being evaluated.
A simple motivating example is below; -D_FORTIFY_SOURCE=2 would miss
this and emit memcpy, but -D_FORTIFY_SOURCE=3 with the help of
__builtin_dynamic_object_size is able to emit __memcpy_chk with the
allocation size expression passed into the function:
void *copy_obj (const void *src, size_t alloc, size_t copysize)
{
void *obj = malloc (alloc);
memcpy (obj, src, copysize);
return obj;
}
Limitations
-----------
If the object was allocated elsewhere that the compiler cannot see, or
if it was allocated in the function with a function that the compiler
does not recognize as an allocator then __builtin_dynamic_object_size
also returns -1.
Further, the expression used to compute object size may be non-trivial
and may potentially incur a noticeable performance impact. These
fortifications are hence enabled at a new _FORTIFY_SOURCE level to
allow developers to make a choice on the tradeoff according to their
environment.
In the next release of POSIX, free must preserve errno
<https://www.austingroupbugs.net/view.php?id=385>.
Modify __libc_free to save and restore errno, so that
any internal munmap etc. syscalls do not disturb the caller's errno.
Add a test malloc/tst-free-errno.c (almost all by Bruno Haible),
and document that free preserves errno.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Add Intel Linear Address Masking (LAM) support to <sys/platform/x86.h>.
HAS_CPU_FEATURE (LAM) can be used to detect if LAM is enabled in CPU.
LAM modifies the checking that is applied to 64-bit linear addresses,
allowing software to use of the untranslated address bits for metadata.
Add a new glibc tunable: mem.tagging. This is a decimal constant in
the range 0-255 but used as a bit-field.
Bit 0 enables use of tagged memory in the malloc family of functions.
Bit 1 enables precise faulting of tag failure on platforms where this
can be controlled.
Other bits are currently unused, but if set will cause memory tag
checking for the current process to be enabled in the kernel.
This patch adds the configuration machinery to allow memory tagging to be
enabled from the command line via the configure option --enable-memory-tagging.
The current default is off, though in time we may change that once the API
is more stable.
GCC 6.5 fails to correctly build ldconfig with recent ld.so.cache
commits, e.g.:
785969a047
elf: Implement a string table for ldconfig, with tail merging
If glibc is build with gcc 6.5.0:
__builtin_add_overflow is used in
<glibc>/elf/stringtable.c:stringtable_finalize()
which leads to ldconfig failing with "String table is too large".
This is also recognizable in following tests:
FAIL: elf/tst-glibc-hwcaps-cache
FAIL: elf/tst-glibc-hwcaps-prepend-cache
FAIL: elf/tst-ldconfig-X
FAIL: elf/tst-ldconfig-bad-aux-cache
FAIL: elf/tst-ldconfig-ld_so_conf-update
FAIL: elf/tst-stringtable
See gcc "Bug 98269 - gcc 6.5.0 __builtin_add_overflow() with small
uint32_t values incorrectly detects overflow"
(https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98269)
The align the GNU extension with the others one that accept specify
which clock to wait for (such as pthread_mutex_clocklock).
Check on x86_64-linux-gnu.
Reviewed-by: Lukasz Majewski <lukma@denx.de>
The SXID_* tunable properties only influence processes that are
AT_SECURE, so make that a bit more explicit in the documentation and
comment.
Revisiting the code after a few years I managed to confuse myself, so
I imagine there could be others who may have incorrectly assumed like
I did that the SXID_ERASE tunables are not inherited by children of
non-AT_SECURE processes.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
I couldn't pinpoint which standard has added it, but no other POSIX
system supports it and/or no longer provide it. The 'struct vtimes'
also has a lot of drawbacks due its limited internal type size.
I couldn't also see find any project that actually uses this symbol,
either in some dignostic way (such as sanitizer). So I think it should
be safer to just move to compat symbol, instead of deprecated. The
idea it to avoid new ports to export such broken interface (riscv32
for instance).
Checked on x86_64-linux-gnu and i686-linux-gnu.
In the Spanish language, the digraph "ll" has not been considered a
separate letter since 1994:
https://www.rae.es/consultas/exclusion-de-ch-y-ll-del-abecedario
Since January 1998 (commit 49891c1062),
glibc's locale data no longer specifies "ch" and "ll" as separate
collation elements. So, it's better to not use "ll" in an example.
Also, the Czech "ch" is a better example as it collates in a more
surprising place.
Current systems do not have BSD terminals, so the fallback code in
posix_openpt/getpt does not do anything. Also remove the file system
check for /dev/pts. Current systems always have a devpts file system
mounted there if /dev/ptmx exists.
grantpt is now essentially a no-op. It only verifies that the
argument is a ptmx-descriptor. Therefore, this change indirectly
addresses bug 24941.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Some tunable values and their minimum/maximum values must be determinted
at run-time. Add TUNABLE_SET_WITH_BOUNDS and TUNABLE_SET_WITH_BOUNDS_FULL
to update tunable value together with minimum and maximum values.
__tunable_set_val is updated to set tunable value as well as min/max
values.
The __x86_shared_non_temporal_threshold determines when memcpy on x86
uses non_temporal stores to avoid pushing other data out of the last
level cache.
This patch proposes to revert the calculation change made by H.J. Lu's
patch of June 2, 2017.
H.J. Lu's patch selected a threshold suitable for a single thread
getting maximum performance. It was tuned using the single threaded
large memcpy micro benchmark on an 8 core processor. The last change
changes the threshold from using 3/4 of one thread's share of the
cache to using 3/4 of the entire cache of a multi-threaded system
before switching to non-temporal stores. Multi-threaded systems with
more than a few threads are server-class and typically have many
active threads. If one thread consumes 3/4 of the available cache for
all threads, it will cause other active threads to have data removed
from the cache. Two examples show the range of the effect. John
McCalpin's widely parallel Stream benchmark, which runs in parallel
and fetches data sequentially, saw a 20% slowdown with this patch on
an internal system test of 128 threads. This regression was discovered
when comparing OL8 performance to OL7. An example that compares
normal stores to non-temporal stores may be found at
https://vgatherps.github.io/2018-09-02-nontemporal/. A simple test
shows performance loss of 400 to 500% due to a failure to use
nontemporal stores. These performance losses are most likely to occur
when the system load is heaviest and good performance is critical.
The tunable x86_non_temporal_threshold can be used to override the
default for the knowledgable user who really wants maximum cache
allocation to a single thread in a multi-threaded system.
The manual entry for the tunable has been expanded to provide
more information about its purpose.
modified: sysdeps/x86/cacheinfo.c
modified: manual/tunables.texi
Add Intel Key Locker:
https://software.intel.com/content/www/us/en/develop/download/intel-key-locker-specification.html
support to <sys/platform/x86.h>. Intel Key Locker has
1. KL: AES Key Locker instructions.
2. WIDE_KL: AES wide Key Locker instructions.
3. AESKLE: AES Key Locker instructions are enabled by OS.
Applications should use
if (CPU_FEATURE_USABLE (KL))
and
if (CPU_FEATURE_USABLE (WIDE_KL))
to check if AES Key Locker instructions and AES wide Key Locker
instructions are usable.
