Bug 2172087
| Summary: | fix accesses of inode->i_flctx | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 9 | Reporter: | Jeff Layton <jlayton> |
| Component: | kernel | Assignee: | Jeff Layton <jlayton> |
| kernel sub component: | VFS | QA Contact: | Kun Wang <kunwan> |
| Status: | CLOSED ERRATA | Docs Contact: | |
| Severity: | medium | ||
| Priority: | unspecified | CC: | bcodding, dhowells, mszeredi, nfs-team, swhiteho, xzhou, yoyang |
| Version: | 9.1 | Keywords: | Triaged |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | kernel-5.14.0-297.el9 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2023-11-07 08:41:33 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Jeff Layton
2023-02-21 13:09:43 UTC
We may be technically breaking kABI with this patch: 2f6076aad206 fs/lock: add 2 callbacks to lock_manager_operations to resolve conflict This grows the lock_manager_operations struct, so if anyone is using an out of tree lock manager (something akin to nfsd or lockd), then this could break for them. @bcodding thinks there might be a chance that GPFS is setting this field in some cases, and is going to try to confirm whether that's the case. That sounds wrong to me, but it's worth investigating before we merge this. Stay tuned. (In reply to Jeff Layton from comment #6) > This grows the lock_manager_operations struct, so if anyone is using an out > of tree lock manager (something akin to nfsd or lockd), then this could > break for them. @bcodding thinks there might be a chance that > GPFS is setting this field in some cases, and is going to try to confirm > whether that's the case. That sounds wrong to me, but it's worth > investigating before we merge this. Stay tuned. We did find gpfs is setting lm_ops in some cases, Jeff has modified the work to set a flag if we're going to be accessing the new offsets in the struct. Its possible we don't need to technically do this due to the new kABI rules for RHEL-9, but doing it also makes things safer if older versions of modules are loaded. It does seem like we should be able to restrict older/non-compatible out-of-tree-modules using structs that have changed with some versioning magic, but I'm not sure how that works. Do we want to try to get a kABI SME to look at this? Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Important: kernel security, bug fix, and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2023:6583 |