Bug 438723
| Summary: | 32bit NFS server returns -EIO for readdirplus request when backing file system has 32bit inodes | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 4 | Reporter: | Jeff Moyer <jmoyer> | ||||
| Component: | kernel | Assignee: | Peter Staubach <staubach> | ||||
| Status: | CLOSED ERRATA | QA Contact: | Martin Jenner <mjenner> | ||||
| Severity: | low | Docs Contact: | |||||
| Priority: | low | ||||||
| Version: | 4.6 | CC: | jburke, jlayton, jmoyer, k.georgiou, steved, vgoyal | ||||
| Target Milestone: | rc | ||||||
| Target Release: | --- | ||||||
| Hardware: | All | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | RHSA-2008-0665 | Doc Type: | Bug Fix | ||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2008-07-24 19:27:50 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: |
|
||||||
The problem is that the NFS server was just invoking vfs_readdir64() and expecting it to correctly handle file systems which don't support the 64 bit readdir entry point. It wasn't and was just returning ENOSYS to the NFS server. The NFS server needs to be handle backing off to just using vfs_readdir(). This means that it must be capable of handling having its filldir entry points called with a 32 bit ino and not just a 64 bit ino. This makes the changes a bit more voluminous, but still relatively straightforward. Created attachment 299222 [details]
Proposed patch
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release. Committed in 68.30.EL . RPMS are available at http://people.redhat.com/vgoyal/rhel4/ 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 therefore 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-2008-0665.html |
Description of problem: Running a client with 2.6.9-68.15.EL and a server with 2.6.9-67.EL does not exhibit the problem. Running both client and server on 2.6.9-68.15.EL or later (tested up to 68.19) exibits the problem. I eliminated the automounter, doing just a mount from the server. Then, strace shows the following: stat64("/mnt/1", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 close(3) = 0 ... open("/mnt/1", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 3 fstat64(3, {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0 fcntl64(3, F_SETFD, FD_CLOEXEC) = 0 getdents64(3, 0x897cdcc, 32768) = -1 EIO (Input/output error) close(3) = 0 The systems are both i686, but I've seen the problem running a 64bit client and a 32bit server as well. The wireshark trace shows EIO being returned from the readdirplus call. Version-Release number of selected component (if applicable): kernel: 2.6.9-68.15.EL How reproducible: 100% Steps to Reproduce: 1. Create an NFS export on a 32bit machine running the aflicted kernel 2. Mount the exported file system from the client 3. Perform an ls of the mounted file system Actual results: -EIO Expected results: directory listing Additional info: You can also reproduce this by running the autofs regression tests: ./autofs_workflow -a x86_64 -d RHEL4-U6 -A http://porkchop.devel.redhat.com/~jburke/dist-cvs-repos/kernel-2.6.9-68.15.EL -i kernel-smp-2.6.9-68.15 -S rhts.redhat.com -u <username>@redhat.com -s bz130467 -n 1