move 32 bit heap hint on sunos

move 32 bit heap hint on sunos

Setting the V8 heap at or near 0x20000000 on 32bit sunos only allows
512 MB of heap space, instead on sunos move this to 0x80000000.

To quote the comment

For our Solaris/illumos mmap hint, we pick a random address in the bottom
half of the top half of the address space (that is, the third quarter).
Because we do not MAP_FIXED, this will be treated only as a hint -- the
system will not fail to mmap() because something else happens to already
be mapped at our random address. We deliberately set the hint high enough
to get well above the system's break (that is, the heap); Solaris and
illumos will try the hint and if that fails allocate as if there were
no hint at all. The high hint prevents the break from getting hemmed in
at low values, ceding half of the address space to the system heap.

BUG=
R=yangguo@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=15870

[tjfontaine, 2013-07-23 02:42:01]

We've been floating this patch for nodejs for quite a while without any
repercussions, it'd be nice to get it upstreamed

[Yang, 2013-07-24 14:04:08]

On 2013/07/23 02:42:01, tjfontaine wrote:
> We've been floating this patch for nodejs for quite a while without any
> repercussions, it'd be nice to get it upstreamed

We do have a platform-solaris.cc, where you can find both callers of this
function. I would prefer adding this change to that platform-specific file. You
could assert there that OS::GetRandomMmapAddr() returns an address in the
expected range and add the suitable offset to it.

[Sven Panne, 2013-07-25 06:10:42]

On 2013/07/24 14:04:08, Yang wrote:
> We do have a platform-solaris.cc, where you can find both callers of this
> function. I would prefer adding this change to that platform-specific file.
You
> could assert there that OS::GetRandomMmapAddr() returns an address in the
> expected range and add the suitable offset to it.

Hmmm, we are currently working merging things to remove the platform-* code
duplication nonsense, so platform-posix actually looks OK to me. Looping in
bmeurer@...

[Benedikt Meurer, 2013-07-25 06:13:14]

Yeah, please don't put anything into the platform-{solaris,linux,macos,*bsd}.cc
files anymore.

[Yang, 2013-07-25 07:59:43]

On 2013/07/25 06:13:14, Benedikt Meurer wrote:
> Yeah, please don't put anything into the
platform-{solaris,linux,macos,*bsd}.cc
> files anymore.

In that case I guess this lgtm.

[Yang, 2013-07-25 08:06:19]

Committed patchset #1 manually as r15870 (presubmit successful).

[tjfontaine, 2013-07-25 16:50:56]

On 2013/07/25 08:06:19, Yang wrote:
> Committed patchset #1 manually as r15870 (presubmit successful).

Thanks everyone!