I like this approach. It certainly seems simpler than the library-based
approach I was pursuing.
Since (1) this approach does not (yet) include a method to specify the
size of the hugepage stack and (2) we're in the kernel anyway, I'd
suggest extending the feature to include automatic shrink/g
On 01.04.2008 [14:45:11 +0100], Mel Gorman wrote:
> On (31/03/08 15:19), Nishanth Aravamudan didst pronounce:
> > On 31.03.2008 [23:07:52 +0800], Eric B Munson wrote:
> > > First pass at enabling huge page backed stacks. This simply enables it
> > > in kernel, to setup a huge page backed stack you
On (31/03/08 15:19), Nishanth Aravamudan didst pronounce:
> On 31.03.2008 [23:07:52 +0800], Eric B Munson wrote:
> > First pass at enabling huge page backed stacks. This simply enables it
> > in kernel, to setup a huge page backed stack you need a user-space
> > utility that sets the HUGE_PAGE_STA
On 31.03.2008 [23:07:52 +0800], Eric B Munson wrote:
> First pass at enabling huge page backed stacks. This simply enables it
> in kernel, to setup a huge page backed stack you need a user-space
> utility that sets the HUGE_PAGE_STACK personality flag for a process.
> This patch does not enable fu
First pass at enabling huge page backed stacks. This simply enables it
in kernel, to setup a huge page backed stack you need a user-space
utility that sets the HUGE_PAGE_STACK personality flag for a process.
This patch does not enable full huge page backed stack functionality,
it will fail if a pr
