Hi all,
I've been working on testing images under QEMU with the lava-dispatcher (mostly this is preparation for testing A15 using a fast model, but it's interesting in its own right too). I think I have made the changes to the dispatcher I need (more about this in another mail) but I can't do an end-to-end test because QEMU doesn't seem to be working very well on my system (64-bit Oneiric, if that matters) -- it segfaults a few seconds into booting the beagle model.
Here's what I've done today to check that it's not entirely the fault of how I'm doing things in the dispatcher:
1) Prepare images
a) Latest image
Dowloaded:
http://snapshots.linaro.org/oneiric/linaro-o-developer/20111122/0/images/tar... http://snapshots.linaro.org/oneiric/omap3-oneiric/20111122/2/images/hwpack/h...
Ran:
sudo linaro-media-create --hwpack ~/Downloads/hwpack_linaro-omap3_20111122-2_armel_supported.tar.gz --binary ~/Downloads/linaro-o-developer-tar-20111122-0.tar.gz --dev beagle --image_file beagle-latest.img
b) 2011.10 image
Downloaded:
http://releases.linaro.org/11.10/ubuntu/oneiric-images/developer/linaro-o-de... http://releases.linaro.org/11.10/ubuntu/oneiric-hwpacks/hwpack_linaro-omap3_...
Ran:
sudo linaro-media-create --hwpack ~/Downloads/hwpack_linaro-omap3_20111024-0_armel_supported.tar.gz --binary ~/Downloads/linaro-o-developer-tar-20111024-0.tar.gz --dev beagle --image_file beagle-2011.10.img --hwpack-force-yes
2) Run images
$ qemu-system-arm -M beaglexm -drive if=sd,cache=writeback,file=beagle-2011.10.img -clock unix -device usb-kbd -device usb-mouse -usb -device usb-net,netdev=mynet -netdev user,id=mynet -nographic
This gets into and through uboot, but segfaults soon after (full log in https://pastebin.linaro.org/327/).
I think I'm running the 2011.10 release of qemu-linaro:
$ qemu-system-arm -version QEMU emulator version 0.15.50 (Debian 0.15.50-2011.10-0ubuntu1~ppa11.10.1), Copyright (c) 2003-2008 Fabrice Bellard
Am I just Doing It Wrong? Or should I be reporting all this as a bug?
Cheers, mwh
On 23 November 2011 00:48, Michael Hudson-Doyle michael.hudson@linaro.org wrote:
I've been working on testing images under QEMU with the lava-dispatcher (mostly this is preparation for testing A15 using a fast model, but it's interesting in its own right too).
Ah, this is cool, more automated testing of QEMU would definitely be useful.
$ qemu-system-arm -M beaglexm -drive if=sd,cache=writeback,file=beagle-2011.10.img -clock unix -device usb-kbd -device usb-mouse -usb -device usb-net,netdev=mynet -netdev user,id=mynet -nographic
This gets into and through uboot, but segfaults soon after (full log in https://pastebin.linaro.org/327/).
QEMU segfaulting is a bug, so please report that. (If you can run it under gdb and put the backtrace in the report that would be helpful.)
(NB that graphics on beagle is currently broken due to some change in the 2011.10 kernel which we haven't investigated yet. 2011.09 works fine.)
-- PMM
On 23 November 2011 09:05, Peter Maydell peter.maydell@linaro.org wrote:
On 23 November 2011 00:48, Michael Hudson-Doyle michael.hudson@linaro.org wrote:
I've been working on testing images under QEMU with the lava-dispatcher (mostly this is preparation for testing A15 using a fast model, but it's interesting in its own right too).
Ah, this is cool, more automated testing of QEMU would definitely be useful.
$ qemu-system-arm -M beaglexm -drive if=sd,cache=writeback,file=beagle-2011.10.img -clock unix -device usb-kbd -device usb-mouse -usb -device usb-net,netdev=mynet -netdev user,id=mynet -nographic
This gets into and through uboot, but segfaults soon after (full log in https://pastebin.linaro.org/327/).
QEMU segfaulting is a bug, so please report that.
Actually, I've just investigated, and this is a bug we've already fixed in qemu-linaro 2011.11, so no need to file a bug. Can you try using that version instead?
(commit fixing the bug: http://git.linaro.org/gitweb?p=qemu/qemu-linaro.git%3Ba=commitdiff%3Bh=520b5... )
thanks -- PMM
On Wed, 23 Nov 2011 14:50:58 +0000, Peter Maydell peter.maydell@linaro.org wrote:
On 23 November 2011 09:05, Peter Maydell peter.maydell@linaro.org wrote:
On 23 November 2011 00:48, Michael Hudson-Doyle michael.hudson@linaro.org wrote:
I've been working on testing images under QEMU with the lava-dispatcher (mostly this is preparation for testing A15 using a fast model, but it's interesting in its own right too).
Ah, this is cool, more automated testing of QEMU would definitely be useful.
I'm glad all of this is going to be of some use :)
$ qemu-system-arm -M beaglexm -drive if=sd,cache=writeback,file=beagle-2011.10.img -clock unix -device usb-kbd -device usb-mouse -usb -device usb-net,netdev=mynet -netdev user,id=mynet -nographic
This gets into and through uboot, but segfaults soon after (full log in https://pastebin.linaro.org/327/).
QEMU segfaulting is a bug, so please report that.
Actually, I've just investigated, and this is a bug we've already fixed in qemu-linaro 2011.11, so no need to file a bug. Can you try using that version instead?
(commit fixing the bug: http://git.linaro.org/gitweb?p=qemu/qemu-linaro.git%3Ba=commitdiff%3Bh=520b5... )
That's a pretty funny bug :-) but happily it does fix things, and I can now actually build an image, install tests and run them (but not retrieve the results yet, oh well) on QEMU! Getting there!
Cheers, mwh