ruby: recorder: Fix (de-)serializing with different cache block-sizes

Upon aggregating records, serialize system's cache-block size, as the
cache-block size can be different when restoring from a checkpoint. This way,
we can correctly read all records when restoring from a checkpoints, even if
the cache-block size is different.

Note, that it is only possible to restore from a checkpoint if the
desired cache-block size is smaller or equal to the cache-block size
when the checkpoint was taken; we can split one larger request into
multiple small ones, but it is not reliable to do the opposite.

Committed by: Nilay Vaish <nilay@cs.wisc.edu>

1 files changed, 8 insertions, 0 deletions

diff --git a/util/cpt_upgrader.py b/util/cpt_upgrader.py
index fac9e07a7..e14274806 100755
--- a/util/cpt_upgrader.py
+++ b/util/cpt_upgrader.py
@@ -545,6 +545,13 @@ def from_8(cpt):
             cpt.set(new_sec, 'bootUncacheability', 'false')
             cpt.set(new_sec, 'num_entries', '0')
 
+# Version 10 adds block_size_bytes to system.ruby
+def from_9(cpt):
+    for sec in cpt.sections():
+        if sec == 'system.ruby':
+            # Use Gem5's default of 64; this should be changed if the to be
+            # upgraded checkpoints were not taken with block-size 64!
+            cpt.set(sec, 'block_size_bytes', '64')
 
 migrations = []
 migrations.append(from_0)
@@ -556,6 +563,7 @@ migrations.append(from_5)
 migrations.append(from_6)
 migrations.append(from_7)
 migrations.append(from_8)
+migrations.append(from_9)
 
 verbose_print = False