https://sourceware.org/bugzilla/show_bug.cgi?id=32768
Bug ID: 32768
Summary: tst-gmon failure on x86_64
Product: binutils
Version: 2.45 (HEAD)
Status: NEW
Severity: normal
Priority: P2
Component: gprof
Assignee: unassigned at sourceware dot org
Reporter: amodra at gmail dot com
Target Milestone: ---
I'm seeing the new gprof test copied from glibc fail on x86_64-linux, depending
on compiler and CFLAGS. I haven't fully investigated, but I think the problem
is that glibc mcount for x86_64 assumes that functions are at least 16 bytes
apart (_gmonparam.log_hashfraction = 4 when running tst-gmon under gdb).
The test fails with
--- expected
+++ actual
@@ -1,3 +1,3 @@
-f1 2000
+f1 1000
f2 1000
f3 1
FAIL
Running gprof without -C for more detail
index % time self children called name
1000 f1 [1]
0.00 0.00 1000/1000 f3 [3]
[1] 0.0 0.00 0.00 1000+1000 f1 [1]
1000 f1 [1]
-----------------------------------------------
0.00 0.00 1000/1000 f3 [3]
[2] 0.0 0.00 0.00 1000 f2 [2]
-----------------------------------------------
0.00 0.00 1/1 main [9]
[3] 0.0 0.00 0.00 1 f3 [3]
0.00 0.00 1000/1000 f1 [1]
0.00 0.00 1000/1000 f2 [2]
So gprof thinks f1 makes 1000 self-calls, and it doesn't see the call
from f2 to f1.
Disassembly of f1 and f2 shows the expected calls are there, it's just
that gprof miscategorises the f2->f1 call as f1->f1.
0000000000401196 <f1>:
401196: 55 push %rbp
401197: 48 89 e5 mov %rsp,%rbp
40119a: e8 b1 fe ff ff call 401050 <mcount@plt>
40119f: 5d pop %rbp
4011a0: c3 ret
00000000004011a1 <f2>:
4011a1: 55 push %rbp
4011a2: 48 89 e5 mov %rsp,%rbp
4011a5: e8 a6 fe ff ff call 401050 <mcount@plt>
4011aa: e8 e7 ff ff ff call 401196 <f1>
4011af: 5d pop %rbp
4011b0: c3 ret
With any other address mod 16 (same code) I see the test pass, and a
proper call graph.
index % time self children called name
0.00 0.00 1000/2000 f2 [2]
0.00 0.00 1000/2000 f3 [3]
[1] 0.0 0.00 0.00 2000 f1 [1]
-----------------------------------------------
0.00 0.00 1000/1000 f3 [3]
[2] 0.0 0.00 0.00 1000 f2 [2]
0.00 0.00 1000/2000 f1 [1]
-----------------------------------------------
0.00 0.00 1/1 main [9]
[3] 0.0 0.00 0.00 1 f3 [3]
0.00 0.00 1000/1000 f2 [2]
0.00 0.00 1000/2000 f1 [1]
--
You are receiving this mail because:
You are on the CC list for the bug.
