On Fri, 2012-12-14 at 16:53 -0800, Andy Grover wrote: > On 12/14/2012 04:26 PM, Nicholas A. Bellinger wrote: > > On Wed, 2012-12-12 at 16:34 -0800, Andy Grover wrote: > >> It doesn't appear to be used anywhere, and is the site of an oops (deve is > >> invalid in core_dec_lacl_count), see: > >> > >> https://bugzilla.redhat.com/show_bug.cgi?id=886698 > >> > >> ...so there may still be a bug lurking, but maybe it'll be easier to spot > >> with this gone. > >> > > > > I think this old code is probably safe to remove now , but I'm curious > > if the tcm_fc driver that's being used in 2.6.32-344.el6.x86_64 code has > > been converted to use target_submit_cmd() w/ target_get_sess_cmd() + > > se_cmd->cmd_kref..? > > No, we're still using drivers/target from 3.2.6. No critical issues to > justify rebasing (until now, maybe). > > > Also, the OOPs seems to indicate that se_node_acl has been released > > while there is still active I/O moving around. > > > > Are you able to reproduce with recent mainline code..? > > Good q, still unknown. > Ok, I'm mentioning this because the Intel + Open-FCoE folks have spent quite a bit of time this year testing tcm_fc w/ active I/O shutdown post conversion to target_submit_cmd() + se_cmd->cmd_kref accounting. IIRC this conversion did go in for v3.2-rc5 code, so I'm thinking there is some else in >= 3.3.0 that is missing here.. --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
