Jo, Kzysztof, 3.16/3.17 currently suffer from a change affecting the kernel/module split. I am trying to get behind this. But right now now PV guest works without re-compiling the kernel with CONFIG_RANDOMIZE_BASE_MAX_OFFSET=0x20000000 (resulting in the same layout as before).
The stack trace itself does not directly contain Xen specific functions (actually that is not a surprise since this is transferring to dom0, which uses the real NIC). So we are talking about an issue related to receiving large transfers on a NetXtreme BCM5704 (that would be the tg3 driver). Comment #4 is a bit confusing. I assume it meant a 3.17 kernel works on bare metal but not as dom0 which would be the max offset problem. The other question would be: does the file transfer work, when not booted as dom0? Maybe as a first step for further debugging I would try to upgrade the dom0 to the most current Trusty kernel (3.13.0-34.60) or maybe the proposed version manually from [1]. That would be 3.13.0-35.62 (you would need linux-headers (arch and all) linux-image (arch), and linux- image-extra (arch). For checking 3.17, we need a special kernel. [1] http://archive.ubuntu.com/ubuntu/pool/main/l/linux/ -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1359742 Title: [xen] The kernel panics on transfer To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1359742/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
