https://sourceware.org/bugzilla/show_bug.cgi?id=28978
Bug ID: 28978
Summary: [2.38 Regression] O(n²) when parsing DWARF2 info
Product: binutils
Version: 2.38
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: binutils
Assignee: unassigned at sourceware dot org
Reporter: steinar+sourceware at gunderson dot no
Target Milestone: ---
Hi,
There is a recent significant slowdown in running “perf report” on DWARF debug
info, which seems to be due to
commit ca8f6bc629cb27792ce449e7253c74a3f6f75fda
Author: Nick Clifton <[email protected]>
Date: Tue Mar 2 16:08:23 2021 +0000
Fix the BFD library's parsing of DIEs where specification attributes can
refer to variables that are defined later on.
PR 27484
* dwarf2.c (scan_unit_for_symbols): Scan twice, once to accumulate
function and variable tags and a second time to resolve their
attributes.
On loading line number information for a compilation unit for the first time,
the second pass calls lookup_func_by_offset() for each function, and
lookup_var_by_offset() for each variable. The problem is that each of these
scan through a linked list containing all functions/variables, which means
there's an O(n²) in the number of functions. With larger C++ projects easily
having 50k+ functions in a compilation unit, this causes significant slowdown.
Mostly, it seems to be completely unneccessary; the functions should be in the
same order for both traversals, so one can probably just keep func through the
iteration and do func = func->prev_func each time an appropriate DW_TAG_* is
seen?
--
You are receiving this mail because:
You are on the CC list for the bug.
