Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-03-18 Thread Dave Airlie
> the new fault hander made the memory manager code a lot cleaner and > very less hacky in a lot of cases. so I'd rather merge the clean code > than have to fight with the current code... Note that you can probably get away with NOPFN_REFAULT etc... like I did for the SPEs in the meantime.

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-03-18 Thread Dave Airlie
the new fault hander made the memory manager code a lot cleaner and very less hacky in a lot of cases. so I'd rather merge the clean code than have to fight with the current code... Note that you can probably get away with NOPFN_REFAULT etc... like I did for the SPEs in the meantime.

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-27 Thread Nick Piggin
On Mon, Feb 26, 2007 at 09:32:04PM -0800, Andrew Morton wrote: > > On Tue, 27 Feb 2007 15:36:03 +1100 "Dave Airlie" <[EMAIL PROTECTED]> wrote: > > > > > > I've also got rid of the horrible populate API, and integrated nonlinear > > > pages > > > properly with the page fault path. > > > > > >

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-27 Thread Nick Piggin
On Mon, Feb 26, 2007 at 09:32:04PM -0800, Andrew Morton wrote: On Tue, 27 Feb 2007 15:36:03 +1100 Dave Airlie [EMAIL PROTECTED] wrote: I've also got rid of the horrible populate API, and integrated nonlinear pages properly with the page fault path. Downside is that this adds

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Benjamin Herrenschmidt
> the new fault hander made the memory manager code a lot cleaner and > very less hacky in a lot of cases. so I'd rather merge the clean code > than have to fight with the current code... Note that you can probably get away with NOPFN_REFAULT etc... like I did for the SPEs in the meantime. Ben.

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Dave Airlie
On 2/27/07, Andrew Morton <[EMAIL PROTECTED]> wrote: > On Tue, 27 Feb 2007 15:36:03 +1100 "Dave Airlie" <[EMAIL PROTECTED]> wrote: > > > > I've also got rid of the horrible populate API, and integrated nonlinear pages > > properly with the page fault path. > > > > Downside is that this adds one

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Andrew Morton
> On Tue, 27 Feb 2007 15:36:03 +1100 "Dave Airlie" <[EMAIL PROTECTED]> wrote: > > > > I've also got rid of the horrible populate API, and integrated nonlinear > > pages > > properly with the page fault path. > > > > Downside is that this adds one more vector through which the buffered write > >

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Dave Airlie
I've also got rid of the horrible populate API, and integrated nonlinear pages properly with the page fault path. Downside is that this adds one more vector through which the buffered write deadlock can occur. However this is just a very tiny one (pte being unmapped for reclaim), compared to

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Dave Airlie
I've also got rid of the horrible populate API, and integrated nonlinear pages properly with the page fault path. Downside is that this adds one more vector through which the buffered write deadlock can occur. However this is just a very tiny one (pte being unmapped for reclaim), compared to

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Andrew Morton
On Tue, 27 Feb 2007 15:36:03 +1100 Dave Airlie [EMAIL PROTECTED] wrote: I've also got rid of the horrible populate API, and integrated nonlinear pages properly with the page fault path. Downside is that this adds one more vector through which the buffered write deadlock can occur.

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Dave Airlie
On 2/27/07, Andrew Morton [EMAIL PROTECTED] wrote: On Tue, 27 Feb 2007 15:36:03 +1100 Dave Airlie [EMAIL PROTECTED] wrote: I've also got rid of the horrible populate API, and integrated nonlinear pages properly with the page fault path. Downside is that this adds one more vector

Re: [patch 0/6] fault vs truncate/invalidate race fix

2007-02-26 Thread Benjamin Herrenschmidt
the new fault hander made the memory manager code a lot cleaner and very less hacky in a lot of cases. so I'd rather merge the clean code than have to fight with the current code... Note that you can probably get away with NOPFN_REFAULT etc... like I did for the SPEs in the meantime. Ben.