Re: [PATCH 4/5] UML - Simplify helper stack handling

[Andrew Morton <akpm@linux-foundation.org>]

On Sun, 5 Aug 2007 22:41:14 +0200 Luca Tettamanti <kronos.it@gmail.com> wrote:

> Il Wed, Jun 27, 2007 at 11:37:01PM -0700, Andrew Morton ha scritto: 
> > 
> > So I'm running the generic version of this on i386 with 8k stacks (below),
> > with a quick LTP run.
> > 
> > Holy cow, either we use a _lot_ of stack or these numbers are off:
> > 
> > vmm:/home/akpm> dmesg -s 1000000|grep 'bytes left' 
> > khelper used greatest stack depth: 7176 bytes left
> > khelper used greatest stack depth: 7064 bytes left
> > khelper used greatest stack depth: 6840 bytes left
> > khelper used greatest stack depth: 6812 bytes left
> > hostname used greatest stack depth: 6636 bytes left
> > uname used greatest stack depth: 6592 bytes left
> > uname used greatest stack depth: 6284 bytes left
> > hotplug used greatest stack depth: 5568 bytes left
> > rpc.nfsd used greatest stack depth: 5136 bytes left
> > chown02 used greatest stack depth: 4956 bytes left
> > fchown01 used greatest stack depth: 4892 bytes left
> > 
> > That's the sum of process stack and interrupt stack, but I doubt if this
> > little box is using much interrupt stack space.
> > 
> > No wonder people are still getting stack overflows with 4k stacks...
> 
> Hi Andrew,
> I was a bit worried about stack usage on my setup and google found your
> mail :P
> 
> FYI:
> 
> khelper used greatest stack depth: 3228 bytes left
> khelper used greatest stack depth: 3124 bytes left
> busybox used greatest stack depth: 2808 bytes left
> modprobe used greatest stack depth: 2744 bytes left
> busybox used greatest stack depth: 2644 bytes left
> modprobe used greatest stack depth: 1836 bytes left
> modprobe used greatest stack depth: 1176 bytes left
> java used greatest stack depth: 932 bytes left
> java used greatest stack depth: 540 bytes left
> 
> I'm running git-current, with 4KiB stacks; filesystems are ext3 and XFS
> on LVM (on libata devices).
> Does it make sense to raise STACK_WARN to get a stack trace in do_IRQ?
> Or is 540 bytes still "safe" taking into account the separate IRQ stack?
> 

540 bytes free means that we've used 90% of the stack.  I'd say it is
extremely unsafe.

Unbelieveably unsafe.  I'm suspecting that the instrumentation is lying to
us for some reason.