Bug 118412
| Summary: | [untested PATCH] SCSI host numbering messed up by reloading driver | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 2.1 | Reporter: | Martin Wilck <martin.wilck> | ||||
| Component: | kernel | Assignee: | Jason Baron <jbaron> | ||||
| Status: | CLOSED WONTFIX | QA Contact: | Brian Brock <bbrock> | ||||
| Severity: | medium | Docs Contact: | |||||
| Priority: | medium | ||||||
| Version: | 2.1 | CC: | coughlan, dledford, knoel, riel | ||||
| Target Milestone: | --- | ||||||
| Target Release: | --- | ||||||
| Hardware: | All | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2007-10-19 19:23:04 UTC | Type: | --- | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
Created attachment 98571 [details]
UNTESTED patch that should give you the idea
AFAICT this error persists in AS2.1 U5 and U6. It appears that there have been comments that I can't see (comment #2,#3,#4). Would you mind making them visible to me? As it stands, I have the impression that nobody ever looked at this. Patch looks sane to us, has it been tested at all? This bug is filed against RHEL2.1, which is in maintenance phase. During the maintenance phase, only security errata and select mission critical bug fixes will be released for enterprise products. Since this bug does not meet that criteria, it is now being closed. For more information of the RHEL errata support policy, please visit: http://www.redhat.com/security/updates/errata/ If you feel this bug is indeed mission critical, please contact your support representative. You may be asked to provide detailed information on how this bug is affecting you. |
Description of problem: When loading/unloading SCSI drivers according to the following scheme: a) load driver A b) unload driver A c) load driver B d) load driver A SCSI host numbers are messed up. Assuming that there are no other hosts in the system and each driver detects obe host, bost hosts of driver A and B will get host number 0. /proc/scsi/scsi will show two devices with identical /host/bus/id/lun if there are devices with the same bus/id/lun connected to the two hosts. Not surprisingly, this causes trouble. The scenario above may at first seem unlikely to happen in the real world. Unfortunately it happens in a pretty common case: During installation of QU3 if a driver disk is needed and USB devices are used for floppy/CD drives. Version-Release number of selected component (if applicable): All until 2.4.9-e.38. How reproducible: Always Steps to Reproduce: 1.Install RedHat AS2.1 QU3 on a system with SCSI controller and USB floppy, using a driver disk for the SCSI controller 2. The system will load usb-storage in order to access the driver disk, copy the driver disk contents into the RAM disk, unload usb-storage, load the SCSI driver, and reload usb.storage. Actual results: Both the SCSI driver and "usb-storage-0" have SCSI host number 0 ("scsi0"). Expected results: SCSI driver and usb-storage-0 have distinct host numbers. Additional info: The error is in line 157 of drivers/scsi/hosts.c in te function scsi_register() (see HERE below). for (shn = scsi_host_no_list;shn;shn = shn->next) { if (!(shn->host_registered) && (hname_len > 0) && (0 == strncmp(hname, shn->name, hname_len))) { flag_new = 0; retval->host_no = shn->host_no; /* <<<< HERE */ shn->host_registered = 1; shn->loaded_as_module = 1; break; } } It is silently assumed that the old host_no has not been given to another host in the meantime. In the scenario described above, though, this is wrong. I am attaching a small patch I just wrote to show how I think the problem could be fixed. The patch is UNTESTED though.