On Thu, 19 Jan 2006, Michael Reed wrote: > This causes console input to stop being accepted and the inability to shutdown the system. > > > QLogic Fibre Channel HBA Driver: 8.01.04-k-fw > from the sles10-latest kernel received yesterday, which is essentially > 2.6.15-git12. This has rpfixes_2.diff already applied. Yes, I've yet to submit that variant... > I did a switchdisable followed by a switchenable 5 seconds later. > > Jan 19 10:35:16 duck klogd: qla2300 0012:00:02.1: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0012:00:02.0: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0012:00:01.1: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0012:00:01.0: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0002:01:06.0: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0002:01:04.1: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0002:01:06.1: LOOP DOWN detected (2). > Jan 19 10:35:16 duck klogd: qla2300 0002:01:04.0: LOOP DOWN detected (2). > > Well, okay, SIX seconds later . :) > > Jan 19 10:35:43 duck klogd: qla2300 0012:00:01.1: LOOP UP detected (2 Gbps). > Jan 19 10:35:45 duck klogd: run_workqueue: recursion depth exceeded: 4 > Jan 19 10:35:45 duck klogd: > Jan 19 10:35:45 duck klogd: Call Trace: > Jan 19 10:35:45 duck klogd: [<a000000100012b00>] show_stack+0x40/0xa0 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35f9e0 bsp=e00000b07a3593b0 > Jan 19 10:35:45 duck klogd: [<a000000100012b90>] dump_stack+0x30/0x60 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fbb0 bsp=e00000b07a359398 > Jan 19 10:35:45 duck klogd: [<a0000001000c9d90>] run_workqueue+0x90/0x280 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fbb0 bsp=e00000b07a359358 > Jan 19 10:35:45 duck klogd: [<a0000001000c9fd0>] flush_cpu_workqueue+0x50/0x1e0 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fbb0 bsp=e00000b07a359328 > Jan 19 10:35:45 duck klogd: [<a0000001000ca220>] flush_workqueue+0xc0/0x160 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fc10 bsp=e00000b07a359300 > Jan 19 10:35:45 duck klogd: [<a0000001000ca3e0>] flush_scheduled_work+0x20/0x40 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fc10 bsp=e00000b07a3592e8 > Jan 19 10:35:45 duck klogd: [<a000000100554570>] fc_rport_tgt_remove+0xf0/0x140 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fc10 bsp=e00000b07a3592b8 > Jan 19 10:35:45 duck klogd: [<a000000100554820>] fc_rport_terminate+0xa0/0x180 > Jan 19 10:35:45 duck klogd: sp=e00000b07a35fc10 bsp=e00000b07a359280 > Jan 19 10:35:49 duck klogd: [<a000000100554950>] fc_remote_port_delete+0x50/0xe0 > Jan 19 10:35:49 duck klogd: sp=e00000b07a35fc10 bsp=e00000b07a359258 > Jan 19 10:35:49 duck klogd: run_workqueue: recursion depth exceeded: 4 As this pinging of the workqueue during removal has some real awful side-effects. I have a series of patches which attempt to work around the semantic requirements. I'll post them ASAP, in the interim, Mike could you grab the following mbox patchset (apply with git-am): ftp://ftp.qlogic.com/outgoing/linux/patches/8.x/8.01.04k/pr600.diff Thanks, Andrew - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
