Distribution:Gentoo Linux Hardware Environment:Athlon 2500+ nforce2 (use ohci1394 module) Software Environment: Problem Description: when pluggin ipod at 1394 port, kernel can't detect ipod, kernel message says like this. ieee1394: Node added: ID:BUS[0-01:1023] GUID[000a270002573bda] ieee1394: The root node is not cycle master capable; selecting a new root node and resetting... ohci1394: fw-host0: SelfID received outside of bus reset sequence ieee1394: Node changed: 0-01:1023 -> 0-00:1023 ieee1394: Node changed: 0-00:1023 -> 0-01:1023 so i modified nodemgr_do_irm_duties() in drivers/ieee1394/nodemgr.c if (ne && ne->busopt.cmc) hpsb_send_phy_config(host, root_node, -1); #if 0 else { HPSB_DEBUG(... ... blah blah return 0; } #endif after load modified ieee1394.ko, kernel can detect ipod right now.
A number of bug fixes to ieee1394 and sbp2 have been commited just recently to the linux1394 project's repo, and feedback is necessary how well they work. Please update the drivers with sources from www.linux1394.org / svn.linux1394.org and test. If the problem persists, try a new module load parameter for ieee1394, disable_irm=1. It may be necessary to include this into /etc/modprobe.d/ieee1394 or /etc/modprobe.conf and reboot in order to reproduce the failure conditions properly: options ieee1394 disable_irm=1 Thanks in advance.
Maybe you should try to plug your ipod in another firewire port. I got the same error (I think that it's what the second line mean), and changing port solves the problem for me.
> Maybe you should try to plug your ipod in another > firewire port. I got the same error (I think that > it's what the second line mean), and changing > port solves the problem for me. The log message "The root node [...] resetting" is not an error. It indicates a normal bus management function. But it is only required if audio/video devices were to be used on the bus with isochronous streams. Changing the ports on which devices are connected subtly affects the hardware selfconfiguration phase of the FireWire bus (the "tree identification") and thus may lead to different root selection in a few cases. It seems not to have this effect in the majority of cases though.
Created attachment 5529 [details] patch lets nodemgr scan more thoroughly for devices that may change their config ROM shortly before the IRM bus reset
Comment on attachment 5529 [details] patch lets nodemgr scan more thoroughly for devices that may change their config ROM shortly before the IRM bus reset http://marc.theaimsgroup.com/?l=linux1394-user&m=112337569219356
Meanwhile there was feedback on linux1394-user that this patch does not fix it. In addition to the disable_irm workaround, two other effective workarounds have been found: http://marc.theaimsgroup.com/?l=linux1394-user&m=112401737206778 The _real_ fix is still being worked on.
A patch which avoids the problem is pending for inclusion into Linux 2.6.14. This should make detection of iPods and similarly affected devices reasonably robust. Further code inspection is needed though.
Should be solved by Linux 2.6.14.