TEXT	runtime·racecallbackthunk(SB), NOSPLIT|NOFRAME, $0-0
	// Handle command raceGetProcCmd (0) here.
	// First, code below assumes that we are on curg, while raceGetProcCmd
	// can be executed on g0. Second, it is called frequently, so will
	// benefit from this fast path.
	CMPQ	RARG0, $0
	JNE	rest
	get_tls(RARG0)
	MOVQ	g(RARG0), RARG0
	MOVQ	g_m(RARG0), RARG0
	MOVQ	m_p(RARG0), RARG0
	MOVQ	p_raceprocctx(RARG0), RARG0

-		  && __string2_1bptr_p (src) && n <= 8			      \
-		  ? __mempcpy_small (dest, __mempcpy_args (src), n)	      \
-		  : __mempcpy (dest, src, n)))
-#   endif
-/* In glibc we use this function frequently but for namespace reasons
-   we have to use the name `__mempcpy'.  */
-#   define mempcpy(dest, src, n) __mempcpy (dest, src, n)
-#  endif
-
-#  if !__GNUC_PREREQ (3, 0) || defined _FORCE_INLINES

TEXT	runtime·racecallbackthunk(SB), NOSPLIT|NOFRAME, $0
	// Handle command raceGetProcCmd (0) here.
	// First, code below assumes that we are on curg, while raceGetProcCmd
	// can be executed on g0. Second, it is called frequently, so will
	// benefit from this fast path.
	CBNZ	R0, rest
	MOVD	g, R13
#ifdef TLS_darwin
	MOVD	R27, R12 // save R27 a.k.a. REGTMP (callee-save in C). load_g clobbers it
#endif
	load_g
#ifdef TLS_darwin

// Calls to Keys should be guarded with a lock on the EndpointShards.
func (es *EndpointShards) Keys() []ShardKey {
	// len(shards) ~= number of remote clusters which isn't too large, doing this sort frequently
	// shouldn't be too problematic. If it becomes an issue we can cache it in the EndpointShards struct.
	keys := make([]ShardKey, 0, len(es.Shards))
	for k := range es.Shards {
		keys = append(keys, k)
	}

adding 3 NOPs.

The purpose of this directive is to improve performance for cases like loops
where better alignment (8 or 16 instead of 4) might be helpful. This directive
exists in PPC64 assembler and is frequently used by PPC64 assembler writers.

PCALIGN $16
PCALIGN $8

By default, functions in Go are aligned to 16 bytes, as is the case in all
other compilers for PPC64. If there is a PCALIGN directive requesting alignment

				{"runtime.GOMAXPROCS", 0},
				{"main.main", 0},
			}},
		}
		if !stress {
			// Only check for this stack if !stress because traceAdvance alone could
			// allocate enough memory to trigger a GC if called frequently enough.
			// This might cause the runtime.GC call we're trying to match against to
			// coalesce with an active GC triggered this by traceAdvance. In that case

By compiling only classes whose source code has changed since the previous compilation, and classes affected by these changes, incremental compilation can significantly reduce Scala compilation time. It is particularly effective when frequently compiling small code increments, as is often done at development time.

TEXT	runtime·racecallbackthunk(SB), NOSPLIT|NOFRAME, $0
	// Handle command raceGetProcCmd (0) here.
	// First, code below assumes that we are on curg, while raceGetProcCmd
	// can be executed on g0. Second, it is called frequently, so will
	// benefit from this fast path.
	MOVD	$0, R0		// clear R0 since we came from C code
	CMP	R3, $0
	BNE	rest
	// Inline raceGetProdCmd without clobbering callee-save registers.
	MOVD	runtime·tls_g(SB), R10

type bufferedMarshaller interface {
	proto.Sizer
	runtime.ProtobufMarshaller
}

// Like bufferedMarshaller, but is able to marshal backwards, which is more efficient since it doesn't call Size() as frequently.
type bufferedReverseMarshaller interface {
	proto.Sizer
	runtime.ProtobufReverseMarshaller
}

// estimateUnknownSize returns the expected bytes consumed by a given runtime.Unknown

   * ThreadLocal; however, ThreadLocal is not well optimized for the case where the ThreadLocal is
   * non-static, and is initialized/removed frequently - this causes churn in the Thread specific
   * hashmaps. Using a static ThreadLocal to avoid that overhead would mean that different
   * ExecutionSequencer objects interfere with each other, which would be undesirable, in addition