linux32 breakage in current..

Steve Kargl sgk at troutmask.apl.washington.edu
Wed Aug 16 00:04:25 UTC 2006


On Tue, Aug 15, 2006 at 02:24:15PM -0700, Steve Kargl wrote:
> On Tue, Aug 15, 2006 at 05:01:46PM -0400, John Baldwin wrote:
> > So this went cold.  IIRC, someone had tracked the linux32 breakage down to the 
> > linux_semctl() changes to use kern_semctl(), but then someone else noticed 
> > via printf that linux_semctl() is never called.  So, can someone sit down and 
> > do a binary date exercise on -CURRENT to figure out when linux32 broke?
> > 
> 
> It will take a few minutes.  I know that linux32 works with 01 Aug 06
> sources and doesn't work with 11 Aug 06 sources.
> 

Oh crap, this is more complicated than I thought
With a kernel built from 2006.08.06.12.00.00 sources,
I'm seeing the classic race condition for locking

troutmask:kargl[201] acroread church.pdf
Segmentation fault (core dumped)
troutmask:kargl[202] acroread church.pdf  <-- This one worked.
troutmask:kargl[203] acroread church.pdf
Segmentation fault (core dumped)
troutmask:kargl[204] acroread church.pdf
Segmentation fault (core dumped)
troutmask:kargl[205] acroread church.pdf  <-- This one worked.
troutmask:kargl[206] 

I'll see if I can dig deeper.

-- 
Steve


More information about the freebsd-amd64 mailing list