netbsd-bugs@netbsd.org
[Top] [All Lists]

Re: kern/36358: Page fault trap at malloc() in pathname_get() when verie

Subject: Re: kern/36358: Page fault trap at malloc() in pathname_get() when veriexec is enabled
From: Christos Zoulas
Date: Fri, 18 May 2007 12:03:49 -0400
On May 18, 11:15am, blair.sadewitz@xxxxxxxxx (blair.sadewitz@xxxxxxxxx) wrote:
-- Subject: kern/36358: Page fault trap at malloc() in pathname_get() when ve

| >Number:         36358
| >Category:       kern
| >Synopsis:       Page fault trap at malloc() in pathname_get() when veriexec 
is enabled
| >Confidential:   no
| >Severity:       critical
| >Priority:       high
| >Responsible:    kern-bug-people
| >State:          open
| >Class:          sw-bug
| >Submitter-Id:   net
| >Arrival-Date:   Fri May 18 11:15:00 +0000 2007
| >Originator:     Blair Sadewitz
| >Release:        4.99.20/i386
| >Organization:
| >Environment:
| NetBSD/i386 4.99.20 with custom kernel
| >Description:
| I can reliably reproduce this problem by building the kernel with make -j.  I 
don't have a backtrace, but the path is something like:
| 
| vn_open() -> pathname_get() -> malloc()
| 
| I get a page fault trap when malloc tries to allocate memory.  I haven't 
looked into this much yet as I've got to get to bed, but this problem doesn't 
occur with a snapshot kernel (generic.mp) from 5/13.
| 
| If I don't build veriexec into the kernel, it doesn't occur.  Another user 
reported this problem to current-users within the last day or two.
| >How-To-Repeat:
| Follow the above code path. ;)
| >Fix:
| Disabling verified exec is a workaround, but I'm not sure what the actual 
problem is.


Can you post the full trace?

christos

<Prev in Thread] Current Thread [Next in Thread>