Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Chris Wilson
To end this particular thread, Daniel made a good point on IRC that his intent is to blow away the contents of the page tables if we need to swap, and then recreate them upon next use. -Chris -- Chris Wilson, Intel Open Source Technology Centre ___ Inte

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Chris Wilson
On Fri, Dec 19, 2014 at 02:29:57PM +0100, Daniel Vetter wrote: > On Fri, Dec 19, 2014 at 01:10:40PM +, Chris Wilson wrote: > > On Fri, Dec 19, 2014 at 11:13:51AM +0100, Daniel Vetter wrote: > > > On Fri, Dec 19, 2014 at 08:50:09AM +, Chris Wilson wrote: > > > > On Fri, Dec 19, 2014 at 09:37

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Daniel Vetter
On Fri, Dec 19, 2014 at 01:10:40PM +, Chris Wilson wrote: > On Fri, Dec 19, 2014 at 11:13:51AM +0100, Daniel Vetter wrote: > > On Fri, Dec 19, 2014 at 08:50:09AM +, Chris Wilson wrote: > > > On Fri, Dec 19, 2014 at 09:37:52AM +0100, Daniel Vetter wrote: > > > > On Fri, Dec 19, 2014 at 08:31

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Chris Wilson
On Fri, Dec 19, 2014 at 11:13:51AM +0100, Daniel Vetter wrote: > On Fri, Dec 19, 2014 at 08:50:09AM +, Chris Wilson wrote: > > On Fri, Dec 19, 2014 at 09:37:52AM +0100, Daniel Vetter wrote: > > > On Fri, Dec 19, 2014 at 08:31:03AM +, Chris Wilson wrote: > > > > On Thu, Dec 18, 2014 at 10:16

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Michel Thierry
On 12/19/2014 10:13 AM, Daniel Vetter wrote: On Fri, Dec 19, 2014 at 08:50:09AM +, Chris Wilson wrote: On Fri, Dec 19, 2014 at 09:37:52AM +0100, Daniel Vetter wrote: On Fri, Dec 19, 2014 at 08:31:03AM +, Chris Wilson wrote: On Thu, Dec 18, 2014 at 10:16:22PM +0100, Daniel Vetter wrote:

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Daniel Vetter
On Fri, Dec 19, 2014 at 08:50:09AM +, Chris Wilson wrote: > On Fri, Dec 19, 2014 at 09:37:52AM +0100, Daniel Vetter wrote: > > On Fri, Dec 19, 2014 at 08:31:03AM +, Chris Wilson wrote: > > > On Thu, Dec 18, 2014 at 10:16:22PM +0100, Daniel Vetter wrote: > > > > On Thu, Dec 18, 2014 at 05:09

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Chris Wilson
On Fri, Dec 19, 2014 at 09:37:52AM +0100, Daniel Vetter wrote: > On Fri, Dec 19, 2014 at 08:31:03AM +, Chris Wilson wrote: > > On Thu, Dec 18, 2014 at 10:16:22PM +0100, Daniel Vetter wrote: > > > On Thu, Dec 18, 2014 at 05:09:57PM +, Michel Thierry wrote: > > > > This new version tries to r

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Daniel Vetter
On Fri, Dec 19, 2014 at 08:31:03AM +, Chris Wilson wrote: > On Thu, Dec 18, 2014 at 10:16:22PM +0100, Daniel Vetter wrote: > > On Thu, Dec 18, 2014 at 05:09:57PM +, Michel Thierry wrote: > > > This new version tries to remove as many unnecessary changes as possible > > > from > > > the pre

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-19 Thread Chris Wilson
On Thu, Dec 18, 2014 at 10:16:22PM +0100, Daniel Vetter wrote: > On Thu, Dec 18, 2014 at 05:09:57PM +, Michel Thierry wrote: > > This new version tries to remove as many unnecessary changes as possible > > from > > the previous RFC. > > > > For GEN8, it has also been extended to work in logi

Re: [Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-18 Thread Daniel Vetter
On Thu, Dec 18, 2014 at 05:09:57PM +, Michel Thierry wrote: > This new version tries to remove as many unnecessary changes as possible from > the previous RFC. > > For GEN8, it has also been extended to work in logical ring submission (lrc) > mode, as it will be the preferred mode of operatio

[Intel-gfx] [PATCH 00/24] PPGTT dynamic page allocations

2014-12-18 Thread Michel Thierry
This new version tries to remove as many unnecessary changes as possible from the previous RFC. For GEN8, it has also been extended to work in logical ring submission (lrc) mode, as it will be the preferred mode of operation. I also tried to update the lrc code at the same time the ppgtt refactor