Bug 175988 - hang-check timer needs to be build on S390/S390x
hang-check timer needs to be build on S390/S390x
Product: Red Hat Enterprise Linux 4
Classification: Red Hat
Component: kernel (Show other bugs)
s390 Linux
medium Severity medium
: ---
: ---
Assigned To: Jan Glauber
Brian Brock
Depends On:
Blocks: 168430
  Show dependency treegraph
Reported: 2005-12-16 17:35 EST by Linda Wang
Modified: 2007-11-30 17:07 EST (History)
1 user (show)

See Also:
Fixed In Version: RHSA-2006-0132
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2006-03-07 16:07:01 EST
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)
hangcheck-timer.diff (3.44 KB, patch)
2006-01-11 16:27 EST, Jan Glauber
no flags Details | Diff

  None (edit)
Description Linda Wang 2005-12-16 17:35:40 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.8) Gecko/20050512 Red Hat/1.0.4-1.4.1 Firefox/1.0.4

Description of problem:
hangcheck-timer kernel modules are currently not build on s390 and s390x

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1.install RHEL4 U3 kernel on s390 machine
2.cd /lib/modules/2.6.9-25/
3.search for hangcheck-timer.ko

Additional info:
Comment 2 Jan Glauber 2005-12-20 11:18:47 EST
This is just a matter of including the Kconfig where HANGCHECK_TIMER
is configured. But this is drivers/char/Kconfig which we currently do not
include from s390 Kconfig since there is too much stuff there that don't makes
sense on s390.

I wonder why the hangcheck timer is not in the watchdog directory?
(The watchdog/Kconfig is included from s390 Kconfig).

Comment 4 Jan Glauber 2006-01-09 10:00:01 EST
The hangcheck timer also needs a monotonic_clock function.
This is currently not implemented on s390 but it could be easily
done by using get_clock().
Comment 5 Jan Glauber 2006-01-10 07:36:30 EST
Um, just figured out that Jason already added the monotonic_clock function,
so this patch is not upstream.

The comment from the archs that have the monotonic_clock() says:
/* monotonic_clock(): returns # of nanoseconds passed since time_init()

Your code just returns get_clock() which is nanoseconds since 1900.
But time_init should refer to the time the machine booted.
Therfore it should be get_clock() - jiffies_timer_cc, or?
Comment 6 Jason Baron 2006-01-10 17:33:28 EST
well, i agree it doesn't meet the spec, but the way its implemented should be
fine since we subtract out the previous value that monotonic_clock() returned.
thus, the starting point is irrelevant as long as its conssitent.
Comment 7 Jan Glauber 2006-01-11 16:26:37 EST
Correct, it should work if it only takes the difference.
But I think this could be easier done by adding the monotonic_clock
interface to s390 arch code. See my attached patch.

BTW, your definition of the s390 second had a typo ;-
Comment 8 Jan Glauber 2006-01-11 16:27:52 EST
Created attachment 123076 [details]
Comment 9 Jason Baron 2006-01-12 17:02:25 EST
look good to me...probably worth sending upstream as well.
Comment 10 Jan Glauber 2006-01-13 03:56:57 EST
I've talked with Martin Schwidefsky, he has no objections
to the monotonic_clock in our arch code, so we will send this
upstream too.
Comment 15 Red Hat Bugzilla 2006-03-07 16:07:02 EST
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on the solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.


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