Bug 510384 - kernel link is not produce proper results for _end symbol
Summary: kernel link is not produce proper results for _end symbol
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: binutils
Version: 11
Hardware: powerpc
OS: Linux
low
medium
Target Milestone: ---
Assignee: Nick Clifton
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-07-08 21:44 UTC by Kumar Gala
Modified: 2009-08-12 20:55 UTC (History)
4 users (show)

Fixed In Version: 2.19.51.0.14-1.fc11
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2009-07-17 11:20:37 UTC
Type: ---


Attachments (Terms of Use)

Description Kumar Gala 2009-07-08 21:44:39 UTC
The _end symbol in a kernel build is not correct.  When building the 'mpc85xx_defconfig' on a v2.6.30 kernel the _end symbol is:

vmlinux:00001000 A _end

Building this same kernel w/gcc4.3 & binutils 2.18.50.0.9-8.fc10 from FC10 produce:

vmlinux:c0599000 A _end

It appears to be some issue with the ALIGN() directive in the linker script (arch/powerpc/kernel/vmlinux.lds) before the _end.  If we remove the ALIGN directive the _end symbol has a reasonable value of:

vmlinux:c05d3018 A _end

Comment 1 Nick Clifton 2009-07-10 09:21:15 UTC
Hi Kumar,

  I assume that this problem is the same as the one that you have reported on the bug-binutils mailing list:

  http://lists.gnu.org/archive/html/bug-binutils/2009-07/msg00059.h

It sounds like you have found a workaround for the problem.  Do you still want this bug report left open ?

Cheers
  Nick

Comment 2 Kumar Gala 2009-07-10 14:53:29 UTC
Yes I'd like it still open.  While we have a workaround I dont consider it sufficient since the fact that we have a chicken/egg scenario here.

Older kernel linker scripts aren't going to know about all sections and thus have orphans.  If newer binutils are going to place orphan sections at places that break things like _end we have an issue.

Comment 3 Nick Clifton 2009-07-10 17:56:38 UTC
Hi Kumar,

  In which case, please could you put together a small testcase that reproduces the problem so that I can investigate further.

Cheers
  Nick

Comment 4 Kumar Gala 2009-07-11 16:45:11 UTC
It looks like Alan Morda posted a patch and has applied it to mainline binutils CVS for this issue.

Do you mind back porting that to the fedora binutils and making a bug fix release

The commit info is here:

http://sourceware.org/ml/binutils-cvs/2009-07/msg00069.html

Comment 5 Nick Clifton 2009-07-14 10:01:53 UTC
Hi Kumar,

> Do you mind back porting that to the fedora binutils and making a bug fix
> release ?

Done.  binutils-2_19_51_0_11-26_fc12.

Cheers
  Nick

Comment 6 Kumar Gala 2009-07-14 14:27:45 UTC
Thanks.  Any ability to get an updated package for fc11?

- k

Comment 7 Kumar Gala 2009-07-14 14:37:46 UTC
At Josh Boyer's recommendation I'm reopening this and pointing out that it was against F11.

Comment 8 Fedora Update System 2009-07-17 11:19:00 UTC
binutils-2.19.51.0.2-18.fc11 has been submitted as an update for Fedora 11.
http://admin.fedoraproject.org/updates/binutils-2.19.51.0.2-18.fc11

Comment 9 Nick Clifton 2009-07-17 11:20:37 UTC
Hi Kumar,

> Any ability to get an updated package for fc11?

Done.  Update to binutils-2.19.51.0.2.18.fc11.

Cheers
  Nick

Comment 10 Fedora Update System 2009-07-27 11:09:24 UTC
binutils-2.19.51.0.14-1.fc11 has been submitted as an update for Fedora 11.
http://admin.fedoraproject.org/updates/binutils-2.19.51.0.14-1.fc11

Comment 11 Fedora Update System 2009-08-12 20:55:02 UTC
binutils-2.19.51.0.14-1.fc11 has been pushed to the Fedora 11 stable repository.  If problems still persist, please make note of it in this bug report.


Note You need to log in before you can comment on or make changes to this bug.