| Summary: | wdt_i6300esb immediately fires on big endian (ppc64) | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Xu Han <xuhan> |
| Component: | qemu-kvm-rhev | Assignee: | David Gibson <dgibson> |
| Status: | CLOSED ERRATA | QA Contact: | Virtualization Bugs <virt-bugs> |
| Severity: | high | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.2 | CC: | hannsj_uhl, hhuang, knoel, michal.skrivanek, michen, mrezanin, ngu, qzhang, rjones, shuyu, virt-maint, xuhan, ypu |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | ppc64 | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2015-12-04 16:31:13 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Bug Depends On: | |||
| Bug Blocks: | 1152690, Red Hat1154205 | ||
I've finally had a chance to look at this. It looks as if the i6300esb code is setting its timeout to a negative value, due to an integer overflow. What I haven't understood yet, is how it's working on x86, which looks as if it should be subject to exactly the same problem. Hi xu, could you please help a try of this case on x86 to see if it also happens when you have time? Thanks. Xu, never mind. I've found a different, ppc64 only bug that caused the problem here. A side effect of the bug was that it made it easier to trigger the first bug I found, which will also appear on x86, but not with the default configuration of the watchdog driver in the guest side. I will file a different BZ for that bug. Ok, overflow bug filed as 1203914. The ppc64 specific bug reported here (technically it will happen on any platform considered by qemu to be big-endian by default) is much simpler. The IO ports on the i6300esb device are marked in qemu as being NATIVE_ENDIAN, when they should (as a PCI device) be LITTLE_ENDIAN. I've posted an upstream fix for this (and bug 1203914). See: http://lists.gnu.org/archive/html/qemu-devel/2015-03/msg04372.html The fix has now been merged upstream, and was incorporated downstream by the qemu-2.3.0 rebase. Reproduce this bug with BE guest ang LE guest on the "qemu-kvm-rhev-2.1.2-23.el7.ppc64" host,hit the same issue. Verify this bug with BE guest and LE guest on the "qemu-kvm-rhev-2.3.0-12.el7.ppc64" host,not hit this issue. Verify this bug with Be guest and LE guest on the "qemu-kvm-rhev-2.3.0-13.el7.ppc64le" host,not hit this issue. Steps as follows:
1.Boot guest with i6300esb watchdog timer.
/usr/libexec/qemu-kvm ... \
-device i6300esb,id=watchdog0 -watchdog-action pause
2.In the guest:
[root@dhcp106-227 ~]# dmesg |grep -i i6300esb
[ 6.749806] i6300esb: Intel 6300ESB WatchDog Timer Driver v0.05
[ 6.749851] i6300ESB timer 0000:00:00.0: enabling device (0100 -> 0102)
[ 6.750588] i6300esb: initialized (0xd000080080080000). heartbeat=30 sec (nowayout=0)
3.[root@dhcp106-227 ~]# python
Python 2.7.5 (default, Jul 8 2015, 05:06:00)
[GCC 4.8.3 20140911 (Red Hat 4.8.3-9)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> wd=open("/dev/watchdog","w")
Reproduce result:
Guest paused immediately.
(qemu) info status
VM status: paused (watchdog)
Verify result:
Wait about 30s,guest paused
(qemu) info status
VM status: running
(qemu) info status
VM status: running
(qemu) info status
VM status: running
(qemu) info status
VM status: running
(qemu) info status
VM status: paused (watchdog)
Additional info:
CMDline:
/usr/libexec/qemu-kvm -name bug1198936-be-2.3-host-le -machine pseries,accel=kvm,usb=off -m 4096 -cpu POWER8 -smp 4,sockets=2,cores=2,threads=1 -no-user-config -nodefaults -chardev socket,id=charmonitor,path=monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -boot strict=on -device pci-ohci,id=usb,bus=pci.0,addr=0x1 -device spapr-vscsi,id=scsi0,reg=0x2000 -drive file=RHEL-7.2-20150806.1-Server-ppc64.qcow2,if=none,id=drive-scsi0-0-0-0,format=qcow2 -device scsi-hd,bus=scsi0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0-0-0-0,id=scsi0-0-0-0,bootindex=1 -drive file=RHEL-7.2-20150806.1-Server-ppc64-dvd1.iso,if=none,id=drive-scsi0-0-1-0,readonly=on,format=raw -device scsi-cd,bus=scsi0.0,drive=drive-scsi0-0-1-0,bootindex=2,id=scsi0-0-1-0 -netdev tap,id=hostnet0 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:5d:c7:8e,bus=pci.0,addr=0x2 -chardev socket,id=charserial0,path=serial,server,nowait -device spapr-vty,chardev=charserial0,reg=0x30000000 -device usb-kbd,id=input0 -device usb-mouse,id=input1 -device usb-tablet,id=tablet1 -vnc :10 -k en-us -device VGA,id=video0,vgamem_mb=16,bus=pci.0,addr=0x3 -monitor stdio -device i6300esb,id=watchdog0 -watchdog-action pause
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2015-2546.html |
Description of problem: The timer expires immediately just by opening the '/dev/watchdog' interface. Both BE and LE guests hit this issue. Version-Release number of selected component (if applicable): qemu-kvm-rhev-2.2.0-5.el7.ppc64 How reproducible: 100% Steps to Reproduce: 1. Boot guest with i6300esb watchdog timer. /usr/libexec/qemu-kvm ... \ -device i6300esb,id=watchdog0 -watchdog-action pause # dmesg | grep -i i6300esb [ 3.499000] i6300esb: Intel 6300ESB WatchDog Timer Driver v0.05 [ 3.499044] i6300ESB timer 0000:00:00.0: enabling device (0100 -> 0102) [ 3.500952] i6300esb: initialized (0xd000080080080000). heartbeat=30 sec (nowayout=0) 2. Open watchdog interface. # python Python 2.7.5 (default, Dec 19 2014, 13:08:28) [GCC 4.8.3 20140911 (Red Hat 4.8.3-9)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> wd = open("/dev/watchdog", "w") Actual results: The timer expires immediately. (qemu) info status VM status: paused (watchdog) Expected results: This device works properly. Additional info: CMDline: /usr/libexec/qemu-kvm \ -name watchdog-test-xuhan \ -machine pseries,accel=kvm,usb=off \ -m 4096 \ -cpu POWER8 \ -smp 4,sockets=2,cores=2,threads=1 \ -no-user-config \ -nodefaults \ -chardev socket,id=charmonitor,path=monitor,server,nowait \ -mon chardev=charmonitor,id=monitor,mode=control \ -rtc base=utc \ -boot strict=on \ -device pci-ohci,id=usb,bus=pci.0,addr=0x1 \ -device spapr-vscsi,id=scsi0,reg=0x2000 \ -drive file=rhel72.qcow2,if=none,id=drive-scsi0-0-0-0,format=qcow2 \ -device scsi-hd,bus=scsi0.0,channel=0,scsi-id=0,lun=0,drive=drive-scsi0-0-0-0,id=scsi0-0-0-0,bootindex=1 \ -netdev tap,id=hostnet0 \ -device virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:5d:c7:9e,bus=pci.0,addr=0x2 \ -chardev socket,id=charserial0,path=serial,server,nowait \ -device spapr-vty,chardev=charserial0,reg=0x30000000 \ -device usb-kbd,id=input0 \ -device usb-mouse,id=input1 \ -device i6300esb,id=watchdog0 -watchdog-action ${action} \ -vnc 0.0.0.0:50 \ -k en-us \ -device VGA,id=video0,vgamem_mb=16,bus=pci.0,addr=0x3 \ -monitor stdio