Re: Are there any beakage of linuxulator (on amd64)?

On Thursday 27 July 2006 09:13, Divacky Roman wrote:
> > > this thursday at work I'll try to provide some more info, what exaclty 
do you
> > > need? is what -DDEBUG prints enough?
> > 
> > Probably.  The changes in question were just in the linux semctl function, 
so you
> > really only need printf's for that function to figure out which case it is 
blowing
> > up one and why.
> 
> soooo....
> 
> I checked the coredump and found this:
> 
> 1) its not acroread what coredumps but bash binary (the binary used for the
> script)
> when I manually tried running the bash and "exec /bin/ls" etc. it worked
> I havent investigated further waht causes the coredump
> 
> 2) I put printf() at the very begining of the linux_semctl() function and
> ran the acroread binary. The printf was not printed (ie. it didnt used the
> linxu_semctl function)

That's odd, because the person who did the binary test claimed it was just the 
change to this file that caused the breakage.  That is, if they reverted 
linux_ipc.c to the revision before the kern_semctl() changes it worked fine.  
Can you test that to see if that's true for you?  (You'll have to revert last 
revision of linux_util.h as well.)

> 3) here is a outpuit od -DDEBUG compiled linuxolator and running of acroread
> (the first 3 lines are output of command line, it might be interesting to 
see
> the VA = 0x0)
> 
> www.stud.fit.vutbr.cz/~xdivac02/linuxamd64

Unfortunately I really can't see what is failing.  Perhaps if you could 
capture the output from starting acroread on a working system so we can see 
where the output starts to change that would help.

-- 
John Baldwin