Support profiling of multi-threaded apps.

This patch set modifies Cygwin's profiling support to sample PC values
of all an app's threads, not just the main thread. There is no change
to how profiling is requested: just compile and link the app with "-pg"
as usual. The profiling info is dumped into file gmon.out as always.

A new facility enabled via the environment variable GMON_OUT_PREFIX.
This facility is intended to match an undocumented Linux glibc feature.
Exporting the variable with a non-empty value such as "foo" causes the
profiling info to go to a file named foo.$pid instead of the default.
With that, both resulting processes of a fork() can have their profiling
data captured in separate files. gprof already knows how to accumulate
data from multiple files if they all pertain to the same app.

There is no change to the normal Cygwin execution paths if profiling is
not enabled. And when it is enabled, only the one profiling thread per
profiled app is doing more work than it used to.

	* include/sys/cygwin.h: Add CW_CYGHEAP_PROFTHR_ALL.
	* cygheap.cc (cygheap_profthr_all): New C-callable function that
	runs cygheap's threadlist handing each pthread's thread handle in
	turn to profthr_byhandle().
	* external.cc (cygwin_internal): Add case CW_CYGHEAP_PROFTHR_ALL.
	* gmon.c (_mcleanup): Add support for multiple simultaneous
	gmon.out* files named via environment variable GMON_OUT_PREFIX.
	* gmon.h (struct gmonparam): Make state decl volatile.
	* mcount.c (_MCOUNT_DECL): Change stores into gmonparam.state to use
	Interlocked operations. Add #include "winsup.h", update commentary.
	* profil.c (profthr_byhandle): New function abstracting out the
	updating of profile counters based on a thread handle.
	(profthr_func): Update to call profthr_byhandle() to sample the main
	thread then call cygheap_profthr_all() indirectly through
	cygwin_internal(CW_CYGHEAP_PROFTHR_ALL) to sample all other threads.
	(profile_off): Zero targthr to indicate profiling was turned off.
	(profile_on): Fix handle leak on failure path.
	(profile_child): New callback func to restart profiling in child
	process after a fork if the parent was being profiled.
	(profile_ctl): Call pthread_atfork() to set profile_child callback.

Signed-off-by: Corinna Vinschen <corinna@vinschen.de>

1 files changed, 67 insertions, 19 deletions

diff --git a/winsup/cygwin/profil.c b/winsup/cygwin/profil.c
index eb41c0844..2d7d6e394 100644
--- a/winsup/cygwin/profil.c
+++ b/winsup/cygwin/profil.c
@@ -18,9 +18,10 @@
 #endif
 #include <windows.h>
 #include <stdio.h>
+#include <sys/cygwin.h>
 #include <sys/types.h>
 #include <errno.h>
-#include <math.h>
+#include <pthread.h>
 #include "profil.h"
 
 #define SLEEPTIME (1000 / PROF_HZ)
@@ -65,25 +66,42 @@ print_prof (struct profinfo *p)
 }
 #endif
 
-/* Everytime we wake up use the main thread pc to hash into the cell in the
-   profile buffer ARG. */
+/* Every time we wake up sample the main thread's pc to hash into the cell
+   in the profile buffer ARG.  Then all other pthreads' pc's are sampled.  */
 
-static void CALLBACK profthr_func (LPVOID);
+static void
+profthr_byhandle (HANDLE thr)
+{
+  size_t pc;
+
+  /* Sample the pc of the thread indicated by thr; bail if anything amiss. */
+  if (thr == INVALID_HANDLE_VALUE)
+    return;
+  pc = get_thrpc (thr);
+  if (pc == -1)
+    return;
+
+  /* Code assumes there's only one profinfo in play: the static prof above. */
+  if (pc >= prof.lowpc && pc < prof.highpc)
+    {
+      size_t idx = PROFIDX (pc, prof.lowpc, prof.scale);
+      prof.counter[idx]++;
+    }
+}
 
 static void CALLBACK
 profthr_func (LPVOID arg)
 {
   struct profinfo *p = (struct profinfo *) arg;
-  size_t pc, idx;
 
   for (;;)
     {
-      pc = (size_t) get_thrpc (p->targthr);
-      if (pc >= p->lowpc && pc < p->highpc)
-	{
-	  idx = PROFIDX (pc, p->lowpc, p->scale);
-	  p->counter[idx]++;
-	}
+      /* Record profiling sample for main thread. */
+      profthr_byhandle (p->targthr);
+
+      /* Record profiling samples for other pthreads, if any. */
+      cygwin_internal (CW_CYGHEAP_PROFTHR_ALL, profthr_byhandle);
+
 #if 0
       print_prof (p);
 #endif
@@ -106,6 +124,7 @@ profile_off (struct profinfo *p)
     }
   if (p->targthr)
     CloseHandle (p->targthr);
+  p->targthr = 0;
   return 0;
 }
 
@@ -121,6 +140,7 @@ profile_on (struct profinfo *p)
 			GetCurrentProcess (), &p->targthr, 0, FALSE,
 			DUPLICATE_SAME_ACCESS))
     {
+      p->targthr = 0;
       errno = ESRCH;
       return -1;
     }
@@ -129,7 +149,7 @@ profile_on (struct profinfo *p)
 
   if (!p->quitevt)
     {
-      CloseHandle (p->quitevt);
+      CloseHandle (p->targthr);
       p->targthr = 0;
       errno = EAGAIN;
       return -1;
@@ -148,8 +168,8 @@ profile_on (struct profinfo *p)
     }
 
   /* Set profiler thread priority to highest to be sure that it gets the
-     processor as soon it request it (i.e. when the Sleep terminate) to get
-     the next data out of the profile. */
+     processor as soon it requests it (i.e. when the Sleep terminates) to get
+     the next data sample as quickly as possible. */
 
   SetThreadPriority (p->profthr, THREAD_PRIORITY_TIME_CRITICAL);
 
@@ -157,16 +177,41 @@ profile_on (struct profinfo *p)
 }
 
 /*
- * start or stop profiling
+ * Restart profiling in child after fork.
  *
- * profiling goes into the SAMPLES buffer of size SIZE (which is treated
- * as an array of u_shorts of size size/2)
+ * The profiling control info in prof needs to be selectively updated.
+ * Items counter, lowpc, highpc, and scale are correct as-is.  But items
+ * targthr, profthr, and quitevt need updating because these copied HANDLE
+ * values are only valid in parent process.  We also zero out the sample
+ * buffer so that samples aren't double-counted if multiple gmon.out files
+ * are aggregated.  We calculate buffer's size from other data in prof.
+ */
+static void
+profile_child (void)
+{
+  /* Bail if this was a fork after profiling turned off or was never on. */
+  if (!prof.targthr)
+    return;
+
+  /* Figure out how big the sample buffer is and zero it out. */
+  size_t size = PROFIDX (prof.highpc, prof.lowpc, prof.scale) << 1;
+  memset ((char *) prof.counter, 0, size);
+
+  /* Replace stale HANDLEs in prof and create profiling thread. */
+  profile_on (&prof);
+}
+
+/*
+ * Start or stop profiling.
  *
- * each bin represents a range of pc addresses from OFFSET.  The number
+ * Profiling data goes into the SAMPLES buffer of size SIZE (which is treated
+ * as an array of u_shorts of size SIZE/2).
+ *
+ * Each bin represents a range of pc addresses from OFFSET.  The number
  * of pc addresses in a bin depends on SCALE.  (A scale of 65536 maps
  * each bin to two addresses, A scale of 32768 maps each bin to 4 addresses,
  * a scale of 1 maps each bin to 128k addreses).  Scale may be 1 - 65536,
- * or zero to turn off profiling
+ * or zero to turn off profiling.
  */
 int
 profile_ctl (struct profinfo * p, char *samples, size_t size,
@@ -191,6 +236,9 @@ profile_ctl (struct profinfo * p, char *samples, size_t size,
       prof.highpc = PROFADDR (maxbin, offset, scale);
       prof.scale = scale;
 
+      /* Set up callback to restart profiling in child after fork. */
+      pthread_atfork (NULL, NULL, profile_child);
+
       return profile_on (p);
     }
   return 0;