Bug 26252 - [odr] dwz.c:11404: write_die: Assertion `value && refdcu->cu_kind != CU_ALT' failed. #2
Summary: [odr] dwz.c:11404: write_die: Assertion `value && refdcu->cu_kind != CU_ALT' ...
Status: RESOLVED FIXED
Alias: None
Product: dwz
Classification: Unclassified
Component: default (show other bugs)
Version: unspecified
: P2 normal
Target Milestone: ---
Assignee: Nobody
URL:
Keywords:
: 25439 (view as bug list)
Depends on:
Blocks:
 
Reported: 2020-07-16 15:40 UTC by Jan Kratochvil
Modified: 2021-02-23 13:11 UTC (History)
4 users (show)

See Also:
Host:
Target:
Build:
Last reconfirmed:

Attachments
.tar.xz (1.48 MB, application/octet-stream)
2020-08-19 19:05 UTC, Jan Kratochvil 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Jan Kratochvil 2020-07-16 15:40:51 UTC 
With b7111689a2ccec2f57343f1051ec8f1df5e89e5c = commit 2020-02-19
getting $subject which should have been fixed by PR 25109 but it is not.

It is reproducible at least with these 233 packages:
  http://people.redhat.com/jkratoch/dwz-assert.pkglist
Comment 1 Tom de Vries 2020-08-18 09:23:05 UTC 
(In reply to Jan Kratochvil from comment #0)
> With b7111689a2ccec2f57343f1051ec8f1df5e89e5c = commit 2020-02-19
> getting $subject which should have been fixed by PR 25109 but it is not.
> 
> It is reproducible at least with these 233 packages:
>   http://people.redhat.com/jkratoch/dwz-assert.pkglist

Can you attach a reproducer?
Comment 2 Tom de Vries 2020-08-18 09:54:05 UTC 
(In reply to Jan Kratochvil from comment #0)
> It is reproducible at least with these 233 packages:
>   http://people.redhat.com/jkratoch/dwz-assert.pkglist

FWIW, I've tried to make a reproducer myself by building binutils, which is in this list, but no luck.
Comment 3 Jan Kratochvil 2020-08-19 19:05:13 UTC 
Created attachment 12782 [details]
.tar.xz

b7111689a2ccec2f57343f1051ec8f1df5e89e5c = 2020-02-19
~/redhat/dwz-git/dwz -m dwz.debug --odr 1.debug 2.debug 
~/redhat/dwz-git/dwz: 1.debug: DWARF compression not beneficial - old size 317760 new size 317760
dwz: dwz.c:11404: write_die: Assertion `value && refdcu->cu_kind != CU_ALT' failed.
Aborted (core dumped)

Although I find more a blocker for DWZ after 3+ years it still does not support DWARF-5; and not just its part as produced by GCC but also the DWARF-5 features produced by LLVM: PR 24726
Comment 4 Mark Wielaard 2020-09-13 14:12:30 UTC 
This only reproduces with --odr. Without things work fine.
Comment 5 Tom de Vries 2021-02-16 09:14:06 UTC 
Posted fix: https://sourceware.org/pipermail/dwz/2021q1/000890.html
Comment 6 Tom de Vries 2021-02-16 09:21:55 UTC 
(In reply to Jan Kratochvil from comment #0)
> With b7111689a2ccec2f57343f1051ec8f1df5e89e5c = commit 2020-02-19
> getting $subject which should have been fixed by PR 25109 but it is not.
> 

PR25109 is related to die_no_multifile propagation, and the problem occurs independent of odr.

This problem is related to CK_BAD propagation, and occurs only with odr.

The two root causes are different, they just happen to trigger the same assert.
Comment 8 Tom de Vries 2021-02-23 13:11:24 UTC 
*** Bug 25439 has been marked as a duplicate of this bug. ***