mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-26 04:32:44 +01:00
48a07cae49
This updates the instrumentation based profiling format so that when we have multiple functions with the same name (but different function hashes) we keep all of them instead of rejecting the later ones. There are a number of scenarios where this can come up where it's more useful to keep multiple function profiles: * Name collisions in unrelated libraries that are profiled together. * Multiple "main" functions from multiple tools built against a common library. * Combining profiles from different build configurations (ie, asserts and no-asserts) The profile format now stores the number of counters between the hash and the counts themselves, so that multiple sets of counts can be stored. Since this is backwards incompatible, I've bumped the format version and added some trivial logic to skip this when reading the old format. llvm-svn: 214585
38 lines
816 B
Plaintext
38 lines
816 B
Plaintext
# If we see the same function name, but with different hashes, make
|
|
# sure we keep both.
|
|
|
|
# RUN: llvm-profdata merge %s -o %t 2>&1
|
|
# RUN: llvm-profdata show %t -all-functions -counts > %t.out
|
|
|
|
# The function ordering is non-deterministic, so we need to do our
|
|
# checks in multiple runs.
|
|
# RUN: FileCheck -check-prefix=FOO3 -check-prefix=BOTH %s -input-file %t.out
|
|
# RUN: FileCheck -check-prefix=FOO4 -check-prefix=BOTH %s -input-file %t.out
|
|
|
|
# FOO3: Hash: 0x{{0+}}3
|
|
# FOO3-NEXT: Counters: 3
|
|
# FOO3-NEXT: Function count: 1
|
|
# FOO3-NEXT: Block counts: [2, 3]
|
|
foo
|
|
3
|
|
3
|
|
1
|
|
2
|
|
3
|
|
|
|
# FOO4: Hash: 0x{{0+}}4
|
|
# FOO4-NEXT: Counters: 4
|
|
# FOO4-NEXT: Function count: 11
|
|
# FOO4-NEXT: Block counts: [22, 33, 44]
|
|
foo
|
|
4
|
|
4
|
|
11
|
|
22
|
|
33
|
|
44
|
|
|
|
# BOTH: Total functions: 2
|
|
# BOTH: Maximum function count: 11
|
|
# BOTH: Maximum internal block count: 44
|