| Summary: | qemuarm and qemuppc images cannot shutdown entirely | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] Meta-yocto | Reporter: | Jiajun Xu <jiajun.xu> |
| Component: | meta-yocto | Assignee: | Bruce Ashfield <bruce.ashfield> |
| Status: | VERIFIED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium | CC: | david.c.stewart, jessica.zhang, josh, kevin.tian, mark.hatle, poky.bs.watcher, poky.watcher, richard.purdie, scott.a.garman, sgw, xianchao.zhang, yongkang.you |
| Version: | 0.9 | ||
| Target Milestone: | 0.9 M4 | ||
| Hardware: | x86 | ||
| OS: | arm | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Program Management | Documentation change: | --- |
|
Description
Jiajun Xu
2010-07-08 06:27:27 UTC
I tested with images with latest M1 build, http://autobuilder.pokylinux.org/milestone/20100710-1/. qemuppc image has the same issue also. qemu stops after printing following message: ###### qemuppc login: md: stopping all md devices. Shutdown: hda Power down. System Halted, OK to turn off power ###### The message indicates that the Linux system DID shut down. However, either Linux never signaled QEMU to terminate, or QEMU wasn't listening for the termination signal. This is between the handoff between the kernel and qemu, and for ppc, it may not be fixable without a medium sized hack to qemu. These will be fixed post M2, when we have some dedicated time to look at them. I don't see these as high defects, since it doesn't impact fundamental operation of the target, and indeed we have many hardware platforms that behave this same way. (In reply to comment #1) > I tested with images with latest M1 build, > http://autobuilder.pokylinux.org/milestone/20100710-1/. qemuppc image has the > same issue also. > > qemu stops after printing following message: > ###### > qemuppc login: md: stopping all md devices. > Shutdown: hda > Power down. > System Halted, OK to turn off power > ###### I tested with images with latest M3 build, http://autobuilder.pokylinux.org/milestone/20100905-1/ qemuppc image has the same issue too. qemu stops after printing following message: ############# The system is going down for system halt NOW! root@qemuppc:~# md: stopping all md devices. Shutdown: hdb Shutdown: hda Power down. System Halted, OK to turn off power ############## We've got some new resources looking at this now. will update when we know more. I tested with images with latest M3 build, http://autobuilder.pokylinux.org/milestone/20100911-1/ qemuarm and qemuppc image have the same issue too. M3 should be unchanged for this. We didn't do any fixes for it. Although there are people looking into this at the moment. So we should have fixes for M4. *** Bug 320 has been marked as a duplicate of this bug. *** I believe this should now work for qemuppc. qemuarm still has the problem so leaving the bug open. qemuppc is fixed, please verify and provide an update (In reply to comment #10) > qemuppc is fixed, please verify and provide an update Xianchao, Could you help try latest build and update the bug? (In reply to comment #11) > (In reply to comment #10) > > qemuppc is fixed, please verify and provide an update > > Xianchao, > Could you help try latest build and update the bug? This issue has not been fixed or not merged with the milestone build 10/11. can someone try 'reboot' instead of shutdown ? This arm/powerpc boards have h/w limitations that prevent shutdown and qemu emulates that. If you issue 'reboot' or 'shutdown -r now', what is the behaviour ? (In reply to comment #13) > can someone try 'reboot' instead of shutdown ? This arm/powerpc boards > have h/w limitations that prevent shutdown and qemu emulates that. If you > issue 'reboot' or 'shutdown -r now', what is the behaviour ? If issue 'reboot', the system can shutdown both qemuarm and qemuppc, but 'shutdown -r now' will cause system halt with the two arch. The 'reboot' working, and shutdown -r now (or halt) not working is the expected behaviour. The ARM versatile h/w doesn't fault back to the boot monitor in our experience, and hence the simulation has the same behaviour. This is much the same with qemuppc, we could only hook up the restart behaviour. Can we modify any scripts or wigets that were using shutdown or halt to exit the simulation to use reboot, and call this one 'good enough' for the simulated machines ? Commit 9a2ae44380a5ef81fb099075af99351db06a3252 makes the -no-reboot option to QEMU mandatory in our poky-qemu control scripts. I believe it is agreed that that should resolve this bug. I agree that if reboot is working we can close this bug and that the automated testing infrastructure can use the reboot command. Scott has made sure qemu will terminate if the guest reboots in commit: http://git.pokylinux.org/cgit.cgi/poky/commit/?id=9a2ae44380a5ef81fb099075af99351db06a3252 reboot can make qemu terminate both qemuarm and qemuppc with the milestone build 10/18 |