Bug#551448: It's because IDE support is not reimplemented somewhere in udev-143
tag 551448 wontfix thanks On Nov 12, Marco d'Itri wrote: > Good point. I understand that the most recent Debian kernels will > replace the old IDE driver with libata and this will take care of the > problem. > If anybody still wants to fix this, please send a patch for path_id. The kernel packages switched to libata for all architectures and nobody sent a patch, so I am tagging this wontfix. -- ciao, Marco signature.asc Description: Digital signature
Bug#551448: It's because IDE support is not reimplemented somewhere in udev-143
Marco d'Itri wrote: Good point. I understand that the most recent Debian kernels will replace the old IDE driver with libata and this will take care of the problem. I just tried with stock 2.6.32-rc6 but couldn't get access to the cdrom & dvd drives. lsmod included:- Module Size Used by scsi_wait_scan 471 0 sg 21152 0 sr_mod 10022 0 cdrom 28055 1 sr_mod ata_generic 2179 0 ata_piix 17761 0 Maybe I didn't properly configure the kernel. ael -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#551448: It's because IDE support is not reimplemented somewhere in udev-143
retitle 551448 path_id.c does not support PATA devices tag 551448 upstream help thanks On Nov 12, Gert Robben wrote: > So, the problem seems to be that since version 143, udev doesn't put the IDE > device numbers in ID_PATH. > Then, if a few IDE devices belong to the same PCI ID (as can be seen in the > 70-persistent-cd.rules files from the affected people), they can't be > differentiated by udev anymore, breaking the persistance. Good point. I understand that the most recent Debian kernels will replace the old IDE driver with libata and this will take care of the problem. If anybody still wants to fix this, please send a patch for path_id. > If you use libata (instead of the old BLK_DEV_IDE drivers), you wouldn't be > affected, I guess. But will the path stay the same when the IDE devices will appear as SCSI devices? -- ciao, Marco signature.asc Description: Digital signature
Bug#551448: It's because IDE support is not reimplemented somewhere in udev-143
AFAICS the same bug is reported here: http://www.spinics.net/lists/hotplug/msg02501.html So, the problem seems to be that since version 143, udev doesn't put the IDE device numbers in ID_PATH. Then, if a few IDE devices belong to the same PCI ID (as can be seen in the 70-persistent-cd.rules files from the affected people), they can't be differentiated by udev anymore, breaking the persistance. I have the problem too, resulting in wrong symlinks, varying everytime (or sometimes) udev is restarted. I tried udev-141 from Ubuntu, which (as I expected) _does_ put the IDE information in ID_PATH, which makes correct symlinks in /dev. If you use libata (instead of the old BLK_DEV_IDE drivers), you wouldn't be affected, I guess. And I assume that if you put your 2 CD drives on different IDE channels, there would be no problem too because the PCI IDs would be different. Greetings, Gert Robben -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org