Hi James I've revisited my last patches, and am submitting a new set. They address the concern you had that I was searching for devices by /dev/* names where actually user-mode can change those. Additionally there is a related bugfix. These are the patches [PATCH 1/5] libosd: osd_dev_is_ver1 - Minor API cleanup [PATCH 2/5] libosd: osd_sense: OSD_CFO_PERMISSIONS Same as before, please submit regardless. [PATCH 3/5] osduld: Ref-counting bug fix This is a bug fix. In the long past it used to work, by chance. With this fix it does what it's suppose to do. (Please submit) [PATCH 4/5] osduld: Use device->release instead of internal kref [PATCH 5/5] libosd: osd_dev_info: Unique Identification of an OSD device These two patches are new implementation to the API introduced in previous patchset. (Plus a related addition needed by multy-device exofs support). It will now use the class_find_device() API to iterate through devices looking for the requested one. Using class_find_device forced a ref-counting shift. (Please review and submit) I've tested these patches with multy-device exofs installation also exported to pnfs-objects clients (also in multi-device) and I'm satisfied with the results. There are no apparent new instabilities. I would want to submit a substantial patchset to exofs, supporting multiple devices, mirror operations. The exofs patches are dependent on these patches to osd [I'm putting these patches and a rebased set of exofs in linux-next. There should be no problems merging once they get accepted into scsi-misc. As these patches are based on by exofs and pnfs-objects] Thanks Boaz -- To unsubscribe from this list: 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
