Thu Dec 11 16:01:32 2008 Google Inc. <opensource@google.com>

	* google-perftools: version 1.0rc1 release
	* Replace API for selectively disabling heap-checker in code (sanjay)
	* Add a pre-mmap hook (daven, adlr)
	* Add MallocExtension interface to set memory-releasing rate (fikes)
	* Augment pprof to allow any string ending in /pprof/profile (csilvers)
	* PORTING: Rewrite -- and fix --  malloc patching for windows (dvitek)
	* PORTING: Add nm-pdb and addr2line-pdb for use by pprof (dvitek)
	* PORTING: Improve cygwin and mingw support (jperkins, csilvers)
	* PORTING: Fix pprof for mac os x, other pprof improvements (csilvers)
	* PORTING: Fix some PPC bugs in our locking code (anton.blanchard)
	* A new unittest, smapling_test, to verify tcmalloc-profiles (csilvers)
	* Turn off TLS for gcc < 4.1.2, due to a TLS + -fPIC bug (csilvers)
	* Prefer __builtin_frame_address to assembly for stacktraces (nlewycky)
	* Separate tcmalloc.cc out into multiple files -- finally! (kash)
	* Make our locking code work with -fPIC on 32-bit x86 (aruns)
	* Fix an initialization-ordering bug for tcmalloc/profiling (csilvers)
	* Use "initial exec" model of TLS to speed up tcmalloc (csilvers)
	* Enforce 16-byte alignment for tcmalloc, for SSE (sanjay)


git-svn-id: http://gperftools.googlecode.com/svn/trunk@60 6b5cf1ce-ec42-a296-1ba9-69fdba395a50

1 files changed, 42 insertions, 11 deletions

diff --git a/INSTALL b/INSTALL
index 71eaac8..039d383 100644
--- a/INSTALL
+++ b/INSTALL
@@ -57,6 +57,7 @@ Even with the use of libunwind, there are still known problems with
 stack unwinding on 64-bit systems, particularly x86-64.  See the
 "64-BIT ISSUES" section in README.
 
+
 *** NOTE FOR ___tls_get_addr ERROR
 
 When compiling perftools on some old systems, like RedHat 8, you may
@@ -66,11 +67,23 @@ get an error like this:
 This means that you have a system where some parts are updated enough
 to support Thread Local Storage, but others are not.  The perftools
 configure script can't always detect this kind of case, leading to
-that error.  To fix it, just uncomment the line
+that error.  To fix it, just comment out the line
    #define HAVE_TLS 1
 in your config.h file before building.
 
 
+*** TCMALLOC AND DLOPEN
+
+To improve performance, we use the "initial exec" model of Thread
+Local Storage in tcmalloc.  The price for this is the library will not
+work correctly if it is loaded via dlopen().  This should not be a
+problem, since loading a malloc-replacement library via dlopen is
+asking for trouble in any case: some data will be allocated with one
+malloc, some with another.  If, for some reason, you *do* need to use
+dlopen on tcmalloc, the easiest way is to use a version of tcmalloc
+with TLS turned off; see the ___tls_get_addr note above.
+
+
 *** COMPILING ON NON-LINUX SYSTEMS
 
 Perftools has been tested on the following systems:
@@ -91,6 +104,7 @@ Perftools has been tested on the following systems:
    Windows XP, Visual Studio 2003 (VC++ 7) (x86)
    Windows XP, Visual Studio 2005 (VC++ 8) (x86)
    Windows XP, MinGW 5.1.3 (x86)
+   Windows XP, Cygwin 5.1 (x86)
 
 It works in its full generality on the Linux systems
 tested (though see 64-bit notes above).  Portions of perftools work on
@@ -172,16 +186,33 @@ above, by linking in libtcmalloc_minimal.
 
 ** Windows:
 
-   Work on Windows is rather preliminary: we have to #define a special
-   symbol that turns off the stacktrace functionality, since that is
-   not yet implemented (see TODOs in src/windows/port.cc if you're
-   interested in playing around with this).  But as in other ports,
-   the basic tcmalloc library functionality, overriding malloc and new
-   and such, is working fine at least with VC++ 8.0 (Visual Studio
-   2005) when compiled in debug mode.  See README.windows for
-   instructions on how to install on Windows using Visual Studio.
-
-   This Windows functionality is also available using MinGW and Msys.
+   Work on Windows is rather preliminary: we haven't found a good way
+   to get stack traces in release mode on windows (that is, when FPO
+   is enabled), so the heap profiling may not be reliable in that
+   case.  Also, heap-checking and CPU profiling do not yet work at
+   all.  But as in other ports, the basic tcmalloc library
+   functionality, overriding malloc and new and such (and even
+   windows-specific functions like _aligned_malloc!), is working fine,
+   at least with VC++ 8.0 (Visual Studio 2005), in both debug and
+   release modes.  See README.windows for instructions on how to
+   install on Windows using Visual Studio.
+
+   Cygwin can compile some but not all of perftools.  Furthermore,
+   there is a problem with exception-unwinding in cygwin (it can call
+   malloc, which can call the exception-unwinding-setup code, which
+   can lead to an infinite loop).  I've comitted a workaround to the
+   exception unwinding problem, but it only works in debug mode.  I
+   hope to have a more proper fix in a later release.  To configure
+   under cygwin, run
+      ./configure CXXFLAGS=-g && make -k
+   Most of cygwin will compile (cygwin doesn't allow weak symbols, so
+   the heap-checker and a few other pieces of functionality will not
+   compile).  'make -k' will compile those libraries and tests that
+   can be compiled.  You can run
+      ./tcmalloc_minimal_unittest
+   to make sure the basic functionality is working.
+
+   This Windows functionality is also available using MinGW and Msys,
    In this case, you can use the regular './configure && make'
    process.  'make install' should also work.  The Makefile will limit
    itself to those libraries and binaries that work on windows.