[Bug debug/103134] New: Redundant DW_AT_entry_pc tags for inlined functions

[Bug debug/103134] New: Redundant DW_AT_entry_pc tags for inlined functions

[wcohen at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103134

            Bug ID: 103134
           Summary: Redundant DW_AT_entry_pc tags for inlined functions
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: wcohen at redhat dot com
  Target Milestone: ---

According to the DWARF5.pdf standard DW_AT_entry_pc can be optional "2.18 Entry
Address":

If no DW_AT_entry_pc attribute is present, then the entry address is assumed to
be the same as the base address of the containing scope.

Eliminating those unneeded DW_AT_entry_pc attributes would save some space in
the generated debuginfo.  To determine the possible saving the following dwgrep
(https://pmachata.github.io/dwgrep/) script was used to find how often this
occurs in the Fedora 34 kernel and showed over 200,000 instances where the
DW_AT_low_pc and DW_AT_entry_pc were the same:


$  dwgrep /usr/lib/debug/lib/modules/5.14.15-200.fc34.x86_64/vmlinux -e '
let A := entry (?TAG_inlined_subroutine) ?DW_AT_low_pc ?DW_AT_entry_pc;
let LOW := ((A ?DW_AT_low_pc @DW_AT_low_pc) || (0xdeadbeef));
let ENTRY := ((A ?DW_AT_entry_pc @DW_AT_entry_pc) || (0xdeadbeef));
(LOW == ENTRY) ([A name, ENTRY, LOW ])' | grep \\[  | wc
 214604  643812 12804153

There were less than 3500 cases where the DW_AT_entry_pc was omitted:

$  dwgrep /usr/lib/debug/lib/modules/5.14.15-200.fc34.x86_64/vmlinux -e '
let A := entry (?TAG_inlined_subroutine);
let RANGE := ((A ?DW_AT_ranges @DW_AT_ranges) || (A ?DW_AT_low_pc
@DW_AT_low_pc) || (0xdeadbeef));
let ENTRY := ((A ?DW_AT_entry_pc @DW_AT_entry_pc) || (0xdeadbeef));
(ENTRY == 0xdeadbeef) ([A name, ENTRY, RANGE ])' | grep \\[  | wc
   3448   12235  278068


There were even fewer cases of DW_AT_entry_pc != DW_AT_low_pc (some might be
due to Bug 99654):

]$  dwgrep /usr/lib/debug/lib/modules/5.14.15-200.fc34.x86_64/vmlinux -e '
let A := entry (?TAG_inlined_subroutine) ?DW_AT_low_pc ?DW_AT_entry_pc;
let LOW := ((A ?DW_AT_low_pc @DW_AT_low_pc) || (0xdeadbeef));
let ENTRY := ((A ?DW_AT_entry_pc @DW_AT_entry_pc) || (0xdeadbeef));
(LOW != ENTRY) ([A name, ENTRY, LOW ])' | grep \\[  | sort > entry_pc_diff.log
$ wc entry_pc_diff.log 
  445  1335 27515 entry_pc_diff.log

According to koji build of the package
(https://koji.fedoraproject.org/koji/buildinfo?buildID=1849253) the kernel was
built with gcc-11.2.1-1.fc34

[Bug debug/103134] Redundant DW_AT_entry_pc tags for inlined functions

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103134

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-11-09

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Might be something for DWZ, I'm not sure it's easy enough to determine this
from dwarf2out.c

[Bug debug/103134] Redundant DW_AT_entry_pc tags for inlined functions

[wcohen at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103134

--- Comment #2 from Will Cohen <wcohen at redhat dot com> ---
Filed a dwz bug (https://sourceware.org/bugzilla/show_bug.cgi?id=28568) as an
alternative was to address the redundant DW_AT_entry_pc.