Re: [PATCH v7 8/8] x86/tlb: just do tlb flush on one of siblingsof SMT

[Peter Zijlstra]

On Thu, 2012-05-24 at 17:27 +0800, Alex Shi wrote:

> I just want to know why it is too big on stack? Since if it causes
> trouble in using, that mean current kernel will fail to run on 4096 CPU
> system.

Because the stack is only 8k or so. If you add functions with 512+ bytes
of footprint you're done very very quickly indeed.

Now if you were able to build a call-graph using a static analysis tool
and were able to weight each edge with the stack usage of every
function, finding the actual max stack would be doable. Lacking this its
a matter of policy and 'luck'.

Our policy is to use as little stack as possible and 512 bytes is
definitely out.

Our 'luck' is that every so often people over-run their stack and we get
to spend time on figuring out wtf happened.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/