RFR: 8225035: Thread stack size issue caused by large TLS size

[Jiangli Zhou]

Hi Andrew and Florian,

With the stack size needs to be aligned at page boundary, it seems that the
effect of using large page size is the same for the following two
approaches:

   1. Using a JVM knob which specifies a minimum stack size for stack size
   adjustment (suggested by Florian)
   2. Using the value obtained from __pthread_get_minstack for stack size
   adjustment

The difference is that the second approach gets the adjustment value
without guessing, which seems to be more user friendly. What are your
thoughts on that? If there is anything that I'm missing, please correct.

After thinking more, it seems that the simple approach with an on/off
switch (suggested by David initially) that adjusts all threads' stack size
(for the on-stack TLS blocks) when the switch is enabled probably is
better. Any objections?

Thanks a lot for all the feedback and suggestions!!

Best regards,

Jiangli

On Mon, Jun 17, 2019 at 3:15 AM Andrew Haley <aph at redhat.com> wrote:

> On 6/17/19 11:08 AM, Florian Weimer wrote:
> >> I think I suggested before to have a JVM knob which specifies a minimum
> >> stack size for passing to libpthread.  This would be a rather
> >> straightforward way to deal with applications which specify impossibly
> >> small thread stacks without having to recompile them.  If users
> >> encounter a problem and they are on 64-bit architectures, they can just
> >> set that knob to 1 MiB because it's unlikely that they have to conserve
> >> address space.
> >
> > I still stand behind this suggestion.
>
> This is simple and has the considerable virtue of not needing any
> particular
> handling of large page sizes.
>
> --
> Andrew Haley
> Java Platform Lead Engineer
> Red Hat UK Ltd. <https://www.redhat.com>
> https://keybase.io/andrewhaley
> EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
>