Bug 432052
| Summary: | [LSI] CPU lockup with kill_fasync interface | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 5 | Reporter: | Atul Mukker <atul.mukker> | ||||
| Component: | kernel | Assignee: | Tom Coughlan <coughlan> | ||||
| Status: | CLOSED CANTFIX | QA Contact: | Martin Jenner <mjenner> | ||||
| Severity: | low | Docs Contact: | |||||
| Priority: | low | ||||||
| Version: | 5.1 | CC: | andriusb, bibhu.das, coldwell | ||||
| Target Milestone: | rc | Keywords: | OtherQA | ||||
| Target Release: | --- | ||||||
| Hardware: | x86_64 | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2008-04-03 17:29:19 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: | 217105 | ||||||
| Attachments: |
|
||||||
|
Description
Atul Mukker
2008-02-08 16:11:23 UTC
Created attachment 294370 [details]
Kernel Log
Any update on this? (In reply to comment #0) > Pid: 4609, comm: dd Tainted: P 2.6.18-prep #1 To recap earlier mail discussion: 1) Chip indicated that it is ok to call kill_fasync() from ISR. 2) The trace indicates that there is a deadlock involving your code. He suggested turning on lockdep, by building against the RHEL-5 kernel-debug package. The additional debug information might help you to identify the problem in your closed source driver code. This is about as far as we are prepared to go in helping you to solve this problem in your driver. If you can reproduce this problem on a kernel that we ship, or even with a driver whose source code is available to us, we maybe able to justify spending some time analyzing this. Otherwise we will consult with you on this as time permits, but we will not have the time to do any detailed analysis. I will leave this BZ open for a while, in case you have specific follow-up questions resulting from your analysis of the trace. Hi, We already run the test with the RHEL-5 kernel debug package and the log is already attached here with. Could you please have a look at the log? (In reply to comment #6) > Could you please have a look at the log? No. We expect you to debug your closed source driver. Atul:
Please run the megasr driver with the CL_IOCTL log flag, and try to generate a
kdump when the deadlock has occured. In the stack it appears that the kernel is
trying to do
472 void send_sigio(struct fown_struct *fown, int fd, int band)
473 {
474 struct task_struct *p;
475 int pid;
476
477 read_lock(&fown->lock); <====== deadlock here
478 pid = fown->pid;
This is strange because that lock is hardly ever taken in write mode. We need a
dump to see what's really going on. Please use a stock Red Hat kernel, so that
we can analyze more easily.
Martin
|