Bug 459478
Summary: | [FOCUS] Long boot time and strange Hardware Clock message | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | Red Hat Enterprise MRG | Reporter: | IBM Bug Proxy <bugproxy> | ||||||
Component: | realtime-kernel | Assignee: | Red Hat Real Time Maintenance <rt-maint> | ||||||
Status: | CLOSED ERRATA | QA Contact: | |||||||
Severity: | high | Docs Contact: | |||||||
Priority: | medium | ||||||||
Version: | 1.0 | CC: | bhu, davids, lgoncalv, pzijlstr, williams | ||||||
Target Milestone: | 1.0.3 | ||||||||
Target Release: | --- | ||||||||
Hardware: | x86_64 | ||||||||
OS: | All | ||||||||
Whiteboard: | |||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||
Doc Text: | Story Points: | --- | |||||||
Clone Of: | Environment: | ||||||||
Last Closed: | 2008-10-07 19:21:02 UTC | Type: | --- | ||||||
Regression: | --- | Mount Type: | --- | ||||||
Documentation: | --- | CRM: | |||||||
Verified Versions: | Category: | --- | |||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||
Embargoed: | |||||||||
Attachments: |
|
Description
IBM Bug Proxy
2008-08-19 09:03:33 UTC
Created attachment 314516 [details]
Normal kernel boot log, where no problem is seen
Created attachment 314517 [details]
Debug kernel boot log, where the problem is seen
Peter Zijlstra pointed me to these two patches: 419ca3f13532793b81aff09f80c60af3eacbb43d d6672c501852d577097f6757c311d937aca0b04b These solve the problem of long boot-up time when CONFIG_PROVE_LOCKING is enabled. We surely need these patches in MRG. Hence please include these patches http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff_plain;h=419ca3f13532793b81aff09f80c60af3eacbb43d and http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff_plain;h=d6672c501852d577097f6757c311d937aca0b04b in MRG. These patches apply on current MRG kernel 2.6.24-7-75el5rt with some offsets. These patches and two other from Peter Zijlstra have been added to 2.6.24.7-78.el5rt. The patches: 0001-lockdep-fix-combinatorial-explosion-in-lock-subgrap.patch 0001-lockdep-build-fix.patch 0002-lockdep-change-scheduler-annotation.patch 0004-lockdep-re-annotate-scheduler-runqueues.patch Ok, Darren agreed that it should go in to -rc3, so don't have any last minute changes that don't get wide testing. I'm going to submit the patches for review on the list, and then try to get acks and commit them so the -rc3 can be built. Ok, the patches are committed and I've rebuilt -rc3. Marking this as accepted. Verified that these patches is included into the mrg-rt-2.6.24.7-81 kernel: * 0001-lockdep-fix-combinatorial-explosion-in-lock-subgrap.patch mrg-rt.git commit 7fdba90ff2a2dfd693fab7b76736334b76d8563e linux-2.6 commit 419ca3f13532793b81aff09f80c60af3eacbb43d * 0001-lockdep-build-fix.patch mrg-rt.git commit 2e3b018b87616c9d622625889d3270b6bf93870c linux-2.6 commit d6672c501852d577097f6757c311d937aca0b04b * 0002-lockdep-change-scheduler-annotation.patch mrg-rt.git commit 2768deaf8c7a1e8bc929303ad897da05fc943389 * 0004-lockdep-re-annotate-scheduler-runqueues.patch mrg-rt.git commit b0a890b69d2a807adab797a0936f54a980bb8988 The mrg-rt-2.6.24.7-81 kernel boots without locking up around udev initialisation as the mrg-rt-2.6.24.7.74 kernel does. Both kernels report the hwclock issue: "Cannot access the Hardware Clock via any known method. Use the --debug option to see the details of our search for an access method." ... this is found on dell-pesc430-03.rhts.bos.redhat.com. Re-tested /sbin/hwclock (without arguments) on stock RHEL5 kernel - 2.6.18-92 - it works as expected. Fails on 2.6.24.7-81.el5rt. One observation was that /proc/devices was missing rtc on 2.6.18-92. On the 2.6.24.7-81rt it was found as a character device with id 254. On both kernels /dev/rtc was found as character device 10 135. I tried to recreate /dev/rtc with as character device 10 254, without any success. Doing cat /dev/rtc also gives "No such device", no matter which id being used. In dmesg rtc seems to be found: rtc_cmos 00:05: rtc core: registered rtc_cmos as rtc0 rtc0: alarms up to one day Noticed that I recreated /dev/rtc with the wrong major/minor id's. By doing this: # rm /dev/rtc ; mknod /dev/rtc c 254 0 After this, /sbin/hwclock will work. Moved bug to verified. The issue regarding hwclock is not solved, but tracked in bug #465837. Long boot time is solved in mrg-rt-2.6.24.7-81rt. 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 therefore 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. http://rhn.redhat.com/errata/RHSA-2008-0857.html The fixes committed to resolve this issue have caused a regression in the non-debug kernel (LTC Bug #49268) Here are the commits in question: * 0001-lockdep-fix-combinatorial-explosion-in-lock-subgrap.patch mrg-rt.git commit 7fdba90ff2a2dfd693fab7b76736334b76d8563e linux-2.6 commit 419ca3f13532793b81aff09f80c60af3eacbb43d * 0001-lockdep-build-fix.patch mrg-rt.git commit 2e3b018b87616c9d622625889d3270b6bf93870c linux-2.6 commit d6672c501852d577097f6757c311d937aca0b04b * 0002-lockdep-change-scheduler-annotation.patch mrg-rt.git commit 2768deaf8c7a1e8bc929303ad897da05fc943389 * 0004-lockdep-re-annotate-scheduler-runqueues.patch mrg-rt.git commit b0a890b69d2a807adab797a0936f54a980bb8988 From a bisection it seems the final patch (lockdep-re-annotate-scheduler-runqueues.patch) is where the regression appears. We're going to revert all 4 patches in our tree for now, but try to sort out the problem shortly. I'm reopening this issue. unfortunately, this BZ (on the RH side) was closed by the errata process and we can't reopen it. We'll need a new BZ on our side. (In reply to comment #46) > ------- Comment From williams 2008-10-22 17:29:19 EDT------- > unfortunately, this BZ (on the RH side) was closed by the errata process and we > can't reopen it. We'll need a new BZ on our side. > Okay, we will open a fresh one. I have opened LTC bug 49534 - RH 469186 to track the regression due to lockdep patches. |