Merge pull request #7562 from raffenet/4.3.x-fixes

[4.3.x] backports in prep for 4.3.2rc1

Approved-by: Hui Zhou <[email protected]>

Author: raffenet

CHANGES

@@ -1,3 +1,50 @@
+===============================================================================
+                               Changes in 4.3.2
+===============================================================================
+# Improve libfabric provider selection when available providers have
+  negative internal score
+
+# Improve error messages when Level Zero failures are detected
+
+# Improve localhost detection in Hydra
+
+# Update libfabric usage to silence deprecation warnings
+
+# Update embedded UCX to v1.19.0
+
+# Add compatibility with CUDA 13
+
+# Fix crash with GPU-aware build when running on systems with no GPUs
+
+# Fix singleton init with Hydra
+
+# Fix thread safety for Level Zero memcpy functions
+
+# Fix potential crash when release gather collective initialization fails
+
+# Fix ch3 connect/accept protocol handling when a discard event arrives
+  after a connection is already established
+
+# Fix inlining for posix eager modules in ch4/shm
+
+# Fix weak attribute usage in MPI ABI build
+
+# Fix potential use-after-free bug in Hydra during spawn operations
+
+# Fix bug in persistent bcast algorithm
+
+# Fix bug in fabric coordinate retrieval with PMIx
+
+# Fix integer overflow and signed/unsigned bugs in ROMIO
+
+# Fix Quobtye ROMIO driver build error
+
+# Fix broken string conversion in mpi_f08 module
+
+# Fix compilation issue with mpi_f08 and NAG compiler
+
+# Fixes for various test program bugs
+
 ===============================================================================
                                Changes in 4.3.1
 ===============================================================================

README.vin

@@ -6,13 +6,13 @@ has all MPI 4.1 functions and features required by the standard with
 the exception of support for user-defined data representations for I/O.
 
 This README file should contain enough information to get you started
-with MPICH. More extensive installation and user guides can be found
-in the doc/installguide/install.pdf and doc/userguide/user.pdf files
+with MPICH. More extensive installation and user guides can be found in
+the doc/installguide/install.pdf and doc/userguide/user.pdf files
 respectively. Additional information regarding the contents of the
-release can be found in the CHANGES file in the top-level directory,
-and in the RELEASE_NOTES file, where certain restrictions are
-detailed. Finally, the MPICH web site, http://www.mpich.org, contains
-information on bug fixes and new releases.
+release can be found in the CHANGES file in the top-level
+directory. Finally, see the official MPICH GitHub page,
+https://github.com/pmodels/mpich, for active development, bug tracking,
+and up-to-date release information.
 
 
 1.  Getting Started

RELEASE_NOTES

@@ -49,9 +49,6 @@ Check all tests. This includes
       - Before committing `CHANGES`, it is recommended to send
         `CHANGES`, at least, to [email protected] in order to make sure
         everything is clear.
-  - Update the `RELEASE_NOTES` file. This requires input from everyone,
-    and generally requires asking each person (in person) if the current
-    restrictions still apply, and if any new ones should be added.
   - Update the `README.vin` if necessary.
 
 ## Create a final tarball for the release using the release.pl script

doc/wiki/source_code/Creating_A_Release.md

@@ -1042,7 +1042,7 @@ def out_can_be_undefined(p):
                 return True
         return False
     # ----
-    re_Handle = r'MPI_(Comm|Datatype|Errhandler|Group|Info|Message|Op|Request|Session|Win|File)\b'
+    re_Handle = get_abi_re_Handle()
     for p in func['c_parameters']:
         skip_abi_swap = False
         param_type = mapping[p['kind']]
@@ -2801,6 +2801,10 @@ def dump_validate_get_topo_size(func):
 
 # ---- supporting routines (reusable) ----
 
+def get_abi_re_Handle():
+    # the following handle types need perform ABI-swap for abi bindings
+    return r'MPI_(Comm|Datatype|Errhandler|Group|Info|Message|Op|Request|Session|Win|File)\b'
+
 def get_function_args(func):
     arg_list = []
     for p in func['c_parameters']:
@@ -2812,16 +2816,10 @@ def get_declare_function(func, is_large, kind=""):
     name = get_function_name(func, is_large)
     mapping = get_kind_map('C', is_large)
 
-    ret = "int"
-    if 'return' in func:
-        ret = mapping[func['return']]
-        if func['return'] == 'EXTRA_STATE':
-            ret = 'void *'
+    ret = get_C_return(func, mapping, kind == 'abi')
 
-    params = get_C_params(func, mapping)
+    params = get_C_params(func, mapping, kind == 'abi')
     s_param = ', '.join(params)
-    if kind == 'abi':
-        s_param = re.sub(r'\bMPI_', 'ABI_', s_param)
     s = "%s %s(%s)" % (ret, name, s_param)
 
     if kind == 'proto':
@@ -2831,10 +2829,25 @@ def get_declare_function(func, is_large, kind=""):
         s += " MPICH_API_PUBLIC"
     return s
 
-def get_C_params(func, mapping):
+def get_C_return(func, mapping, is_abi=False):
+    ret = "int"
+    if 'return' in func:
+        ret = mapping[func['return']]
+        if func['return'] == 'EXTRA_STATE':
+            ret = 'void *'
+        if is_abi and func['dir'] != 'io':
+            re_Handle = get_abi_re_Handle()
+            ret = re.sub(re_Handle, r'ABI_\1', ret)
+    return ret
+
+def get_C_params(func, mapping, is_abi=False):
     param_list = []
+    re_Handle = get_abi_re_Handle()
     for p in func['c_parameters']:
-        param_list.append(get_C_param(p, func, mapping))
+        param = get_C_param(p, func, mapping)
+        if is_abi and func['dir'] != 'io':
+            param = re.sub(re_Handle, r'ABI_\1', param)
+        param_list.append(param)
     if not len(param_list):
         return ["void"]
     else:

maint/local_python/binding_c.py

@@ -862,7 +862,12 @@ def dump_interface_function(func, name, c_name, is_large):
         decl_list.append("INTEGER(c_int) :: ierror")
     else:
         ret = "res"
-        decl_list.append("%s :: res" % f08_mapping[func['return']])
+        if func['return'] == 'LOGICAL':
+            # MPIX_Request_is_complete
+            decl_list.append("LOGICAL(c_bool) :: res")
+            uses['c_bool'] = 1
+        else:
+            decl_list.append("%s :: res" % f08_mapping[func['return']])
 
     # ----
     G.out.append("")

maint/local_python/binding_f08.py

@@ -14,7 +14,7 @@
 # changing this by playing with diversions, but then we would probably be
 # playing with autotools-fire.
 
-m4_define([MPICH_VERSION_m4],[4.3.1])dnl
+m4_define([MPICH_VERSION_m4],[4.3.2rc1])dnl
 m4_define([MPICH_RELEASE_DATE_m4],[unreleased development copy])dnl
 
 # For libtool ABI versioning rules see:

maint/version.m4

@@ -7,11 +7,16 @@ MPI_Op_create:
     .poly_impl: separate
 /*
     Notes on the user function:
-    The calling list for the user function type is
+    For MPI_Op_create, the calling list for the user function type is
     .vb
         typedef void (MPI_User_function) (void * a,
                       void * b, int * len, MPI_Datatype *);
     .ve
+    For MPI_Op_create_c, the calling list for the user function type is
+    .vb
+        typedef void (MPI_User_function_c) (void * a,
+                      void * b, MPI_Count * len, MPI_Datatype *);
+    .ve
     where the operation is 'b[i] = a[i] op b[i]', for 'i=0,...,len-1'.  A pointer
     to the datatype given to the MPI collective computation routine (i.e.,
     'MPI_Reduce', 'MPI_Allreduce', 'MPI_Scan', or 'MPI_Reduce_scatter') is also