Created attachment 374916 [details, diff] fix for igb oops in set_interrupt_capability 3.14.0 shipped with bugs in the igb driver that crashes if MSI/MSIX is not enabled or available. There are two patches needed to fix that, here's URLs but I will attach the isolated .patch files as well. http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b709323d2477614823a38c2f2a9a206e087e28fc http://marc.info/?l=linux-netdev&m=139540390526158&w=2 With both of these patches, 3.14.0 boots and works for me. Please consider adding these to the current genpatches set.
Created attachment 374918 [details, diff] fix for igb oops in reset_q_vector
3.14.1 is out, and there's a new genpatches, but both bugs appear to still be present in the igb driver. The same patches apply cleanly to gentoo-sources-3.14.1.
For my reference: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=b709323d2477614823a38c2f2a9a206e087e28fc http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/net/ethernet/intel/igb/igb_main.c?id=cb06d102327eadcd1bdc480bfd9f8876251d1007
hardened-sources-3.14.1 is out now (yay!). But it contains the same broken igb driver, so will also oops at driver init on a system without MSI(X) enabled. It needs the same patches above.
(In reply to Hank Leininger from comment #4) > hardened-sources-3.14.1 is out now (yay!). > > But it contains the same broken igb driver, so will also oops at driver init > on a system without MSI(X) enabled. > > It needs the same patches above. This appears to be included in the latest hardened-sources.
