Re: [Qemu-ppc] [Qemu-devel] [PULL 095/107] spapr: clock should count only if vm is running

[Mark Cave-Ayland]

On 02/02/17 14:20, Laurent Vivier wrote:

> I think adding the the PPCTimebase field and the VMSTATE_PPC_TIMEBASE_V
> macro to the PMac machines should fix your issue.
> 
> Do you have a test case I can try?
> 
> Laurent

Hi Laurent,

Yes I'd say that is required, although I still think you need to migrate
the decrementer value as per the comments on
https://lists.gnu.org/archive/html/qemu-devel/2016-01/msg00546.html.

Here's the reproducer from an off-list email I sent last year:

1) Download https://www.ilande.co.uk/tmp/darwin_empty.qcow2.xz and
decompress the image (it's a pre-partitioned empty Apple Partition Map disk)

2) Download https://opensource.apple.com/static/iso/darwinppc-602.cdr.gz
image, gunzip it and rename with .iso extension

3) Start QEMU using the attached "empty" disk like this:

./qemu-system-ppc -hda darwin_empty.qcow2 -cdrom darwinppc-602.iso -boot d

4) Start the installer in the guest and you'll see lots of files with
ASCII progress bars as the various files are copied to disk

Then to see the problem with the progress bar, repeat the following:

5) Pause the VM

6) Issue "savevm foo" in the monitor

7) Exit QEMU

8) Start QEMU again as below:

./qemu-system-ppc -hda darwin_empty.qcow2 -cdrom darwinppc-602.iso -boot
d -loadvm foo

If you do this enough times (maybe 10 or so?) you'll see the progress
bars stop working correctly and get out of sync, i.e. it will freeze for
long periods of time and then "jump" to catch-up but not all the way.

With my above patch applied to include the decrementer in the migration,
the bug was no longer visible in my tests.


HTH,

Mark.