[Bug tools/28032] New: eu-readelf --debug-dump=line stops after finding a table without line number statements.

[Bug tools/28032] New: eu-readelf --debug-dump=line stops after finding a table without line number statements.

[vifoxa2454 at advew dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=28032

            Bug ID: 28032
           Summary: eu-readelf --debug-dump=line stops after finding a
                    table without line number statements.
           Product: elfutils
           Version: unspecified
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: tools
          Assignee: unassigned at sourceware dot org
          Reporter: vifoxa2454 at advew dot com
                CC: elfutils-devel at sourceware dot org
  Target Milestone: ---

We use an ABI dumper (https://github.com/lvc/abi-dumper) to make sure we don't
accidentally release binary incompatible changes by mistake. This tool uses
"eu-readelf --debug-dump=line" to get some of the information it needs.

In version 0.180 this worked fine, but when we upgraded to 0.182 it stopped
working. After some digging I found that a change in 0.181 caused the issue.

I found this link describing the change:
https://www.mail-archive.com/search?l=elfutils-devel@sourceware.org&q=subject:%22%5C%5BPATCH%5C%5D+readelf%5C%3A+It+is+not+an+error+if+there+are+no+line+number+statements%22&o=newest&f=1

Running "eu-readelf --debug-dump=line" on one of our shared libraries I get 92
tables on version 0.180, but only 6 tables on version 0.182. The last line in
the output of the 0.182 version is "No line number statements". The
corresponding table in version 0.180 says only "Line number statements:" and
then nothing. That table is very small (length 317, prologue length 311).

I believe that the bug is that the tool stops at this point instead of
continuing with the next table.

The problematic code (in src/readelf.c taken from version 0.185):

    if (linep == lineendp)
    {
        puts (_("\nNo line number statements."));
        return;
    }


If I comment out that block the tool work again for our use case.

Maybe some sort if if/else would be better?
I am not sure what would fit that code best as I have not done any deep diving
into it.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug tools/28032] eu-readelf --debug-dump=line stops after finding a table without line number statements.

[mark at klomp dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=28032

Mark Wielaard <mark at klomp dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |mark at klomp dot org
         Resolution|---                         |FIXED

--- Comment #1 from Mark Wielaard <mark at klomp dot org> ---
(In reply to vifoxa2454 from comment #0)
> I believe that the bug is that the tool stops at this point instead of
> continuing with the next table.
> 
> The problematic code (in src/readelf.c taken from version 0.185):
> 
>     if (linep == lineendp)
>     {
>         puts (_("\nNo line number statements."));
>         return;
>     }
> 
> 
> If I comment out that block the tool work again for our use case.

You are right, that return should have been a continue statement.
I fixed it and added a testcase for this:


commit 6648d378bd94f50b455a8550db5f43ef3690b451
Author: Mark Wielaard <mark@klomp.org>
Date:   Sun Jul 4 00:08:32 2021 +0200

    readelf: Handle line tables without line number statements correctly

    When we see a line table without line number statements we need to
    continue with the next table. Add a testcase for this situation.

    https://sourceware.org/bugzilla/show_bug.cgi?id=28032

    Signed-off-by: Mark Wielaard <mark@klomp.org>

-- 
You are receiving this mail because:
You are on the CC list for the bug.