fs/ufs: change default locking protocol

The fs/ufs component by default disabled all file locking before
read/write operations (except for NFS file systems). This was based on
the assumption, that the file system itself performs the required
locking operation and hence we don't have to add to it.

This assumption is incorrect when using data sieving. In data sieving,
the code 'ignore' small gaps when we write to a file, and perform
instead a read-modify-write sequence ourselves for performance reasons.
The problem is however that even within a collective operation not all
aggregators might want to use data sieving. Hence, enabling locking just
for the data-sieving routines is insufficient, all processes have to
perform the locking. Therefore, our two options are: a) either disable
write data-sieving by default, or b) enable range-locking by default.

After some testing, I think enabling range-locking be default is the
safer and better approach. It doesn't seem to show any significant
performance impact on my test systems.

Note, that on Lustre file systems, we can keep the default to no-locking
as far as I can see, since the collective algorithm used by Lustre is
unlikely to produce this pattern. I did add in however an mca parameter
that allows us to control the locking algorithm used by the Lustre
component as well, in case we need to change that for a particular
use-case or platform.

Fixes Issue open-mpi#12718

Signed-off-by: Edgar Gabriel <[email protected]>

Author: edgargabriel

ompi/mca/fs/lustre/fs_lustre.h

@@ -13,6 +13,7 @@
  * Copyright (c) 2015-2018 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * Copyright (c) 2016-2017 IBM Corporation. All rights reserved.
+ * Copyright (c) 2024      Advanced Micro Devices, Inc. All rights reserverd.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -31,6 +32,12 @@
 extern int mca_fs_lustre_priority;
 extern int mca_fs_lustre_stripe_size;
 extern int mca_fs_lustre_stripe_width;
+extern int mca_fs_lustre_lock_algorithm;
+
+#define FS_LUSTRE_LOCK_AUTO        0
+#define FS_LUSTRE_LOCK_NEVER       1
+#define FS_LUSTRE_LOCK_ENTIRE_FILE 2
+#define FS_LUSTRE_LOCK_RANGES      3
 
 BEGIN_C_DECLS
 

ompi/mca/fs/lustre/fs_lustre_component.c

@@ -13,6 +13,7 @@
  * Copyright (c) 2008-2011 University of Houston. All rights reserved.
  * Copyright (c) 2015      Los Alamos National Security, LLC. All rights
  *                         reserved.
+ * Copyright (c) 2024      Advanced Micro Devices, Inc. All rights reserverd.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -45,6 +46,7 @@ int mca_fs_lustre_priority = 20;
    runtime also*/
 int mca_fs_lustre_stripe_size = 0;
 int mca_fs_lustre_stripe_width = 0;
+int mca_fs_lustre_lock_algorithm = 0; /* auto */
 /*
  * Instantiate the public struct with all of our public information
  * and pointers to our public functions in it
@@ -93,6 +95,15 @@ lustre_register(void)
                                            MCA_BASE_VAR_TYPE_INT, NULL, 0, 0,
                                            OPAL_INFO_LVL_9,
                                            MCA_BASE_VAR_SCOPE_READONLY, &mca_fs_lustre_stripe_width);
+    mca_fs_lustre_lock_algorithm = 0;
+    (void) mca_base_component_var_register(&mca_fs_lustre_component.fsm_version,
+                                           "lock_algorithm", "Locking algorithm used by the fs ufs component. "
+                                           " 0: auto (default), 1: skip locking, 2: always lock entire file, "
+                                           "3: lock only specific ranges",
+                                           MCA_BASE_VAR_TYPE_INT, NULL, 0, 0,
+                                           OPAL_INFO_LVL_9,
+                                           MCA_BASE_VAR_SCOPE_READONLY,
+                                           &mca_fs_lustre_lock_algorithm );
 
     return OMPI_SUCCESS;
 }

ompi/mca/fs/lustre/fs_lustre_file_open.c

@@ -13,6 +13,7 @@
  * Copyright (c) 2015-2020 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * Copyright (c) 2016-2017 IBM Corporation. All rights reserved.
+ * Copyright (c) 2024      Advanced Micro Devices, Inc. All rights reserverd.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -175,8 +176,22 @@ mca_fs_lustre_file_open (struct ompi_communicator_t *comm,
     fh->f_stripe_size   = lump->lmm_stripe_size;
     fh->f_stripe_count  = lump->lmm_stripe_count;
     fh->f_fs_block_size = lump->lmm_stripe_size;
-    fh->f_flags |= OMPIO_LOCK_NEVER;
     free(lump);
 
+    if (FS_LUSTRE_LOCK_AUTO == mca_fs_lustre_lock_algorithm ||
+        FS_LUSTRE_LOCK_NEVER == mca_fs_lustre_lock_algorithm ) {
+        fh->f_flags |= OMPIO_LOCK_NEVER;
+    }
+    else if (FS_LUSTRE_LOCK_ENTIRE_FILE == mca_fs_lustre_lock_algorithm) {
+        fh->f_flags |= OMPIO_LOCK_ENTIRE_FILE;
+    }
+    else if (FS_LUSTRE_LOCK_RANGES == mca_fs_lustre_lock_algorithm) {
+        /* Nothing to be done. This is what the posix fbtl component would do
+           anyway without additional information . */
+    }
+    else {
+        opal_output ( 1, "Invalid value for mca_fs_lustre_lock_algorithm %d", mca_fs_lustre_lock_algorithm );
+    }
+
     return OMPI_SUCCESS;
 }

ompi/mca/fs/ufs/fs_ufs_file_open.c

@@ -13,6 +13,7 @@
  * Copyright (c) 2015-2018 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * Copyright (c) 2016-2017 IBM Corporation. All rights reserved.
+ * Copyright (c) 2024      Advanced Micro Devices, Inc. All rights reserverd.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -105,12 +106,6 @@ mca_fs_ufs_file_open (struct ompi_communicator_t *comm,
                    component. */
                 fh->f_flags |= OMPIO_LOCK_ENTIRE_FILE;
             }
-            else {
-                fh->f_flags |= OMPIO_LOCK_NEVER;
-            }
-        }
-        else {
-            fh->f_flags |= OMPIO_LOCK_NEVER;
         }
         free (fstype);
     }