Bug 164438
| Summary: | LTC12403-CMVC482920:I/O errors caused by eeh error injection-drive unavailable | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 3 | Reporter: | Issue Tracker <tao> |
| Component: | kernel | Assignee: | Nathan Lynch <nlynch> |
| Status: | CLOSED ERRATA | QA Contact: | Brian Brock <bbrock> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 3.0 | CC: | bmaly, dhowells, peterm, petrides, tao |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | powerpc | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | RHSA-2006-0144 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2006-03-15 16:18:35 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: | |||
| Bug Depends On: | |||
| Bug Blocks: | 168424 | ||
|
Description
Issue Tracker
2005-07-27 21:34:51 UTC
Reasigning to Nathan Lynch. Nathan - please review the commentary on rhkernel-list circa May, 2005. In that thread there were technical objections that remain to be cleared up. Subsequent to a discussion with Dustin, placing this on the U7 proposed list. The outstanding concerns from Red Hat development seem to be: 1. "the comments above eeh_check_failure() claim that the token should be an "address in the form 0xA....". If that were true, then eeh_token_to_phys() would have translated it into an I/O region address (0xE...) and then walked the page tables to determine the associated physical address (which presumably would have no region bits set). This implies that an invalid token was passed or that the comment was wrong." 2. "the 2.6 implementation of eeh_token_to_phys() does the virtual-to-physical translation in all cases, not just for EEH region addresses like the RHEL3 implementation. So, perhaps the comment was wrong (and tokens can actually be any kernel virtual address), and thus we'd need a new version of eeh_token_to_phys()." 3. There seems to be some continued misapprehension that there is some IBM kernel tree which is relevant to the discussion. There is not. Argh, I don't think my previous update was marked appropriately to mirror to IT. Here goes again: The outstanding concerns from Red Hat development seem to be: 1. "the comments above eeh_check_failure() claim that the token should be an "address in the form 0xA....". If that were true, then eeh_token_to_phys() would have translated it into an I/O region address (0xE...) and then walked the page tables to determine the associated physical address (which presumably would have no region bits set). This implies that an invalid token was passed or that the comment was wrong." 2. "the 2.6 implementation of eeh_token_to_phys() does the virtual-to-physical translation in all cases, not just for EEH region addresses like the RHEL3 implementation. So, perhaps the comment was wrong (and tokens can actually be any kernel virtual address), and thus we'd need a new version of eeh_token_to_phys()." 3. There seems to be some continued misapprehension that there is some IBM kernel tree which is relevant to the discussion. There is not. Nathan, what's happening with this? I still have my U7 patch tracking file "0970.bmaly.ppc64-eeh-check-fail.patch" in unresolved state. I've reproduced the problem, and figured out exactly why we're passing non-token addresses to eeh_check_region. Will be providing an alternate but functionally equivalent patch soon. A fix for this problem has just been committed to the RHEL3 U7 patch pool this evening (in kernel version 2.4.21-37.3.EL). An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on the solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2006-0144.html |