Adam Litke wrote on Monday, August 08, 2005 3:17 PM
> The reason for the VM_FAULT_SIGBUS default return is because I thought a
> fault on a pte_present hugetlb page was an invalid/unhandled fault.
> I'll have another think about races to the fault handler though.
Two threads fault on the same
On Fri, 2005-08-05 at 17:05, Chen, Kenneth W wrote:
> Adam Litke wrote on Friday, August 05, 2005 8:22 AM
> > Below is a patch to implement demand faulting for huge pages. The main
> > motivation for changing from prefaulting to demand faulting is so that
> > huge page allocations can follow the
On Fri, 2005-08-05 at 17:05, Chen, Kenneth W wrote:
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA
Adam Litke wrote on Monday, August 08, 2005 3:17 PM
The reason for the VM_FAULT_SIGBUS default return is because I thought a
fault on a pte_present hugetlb page was an invalid/unhandled fault.
I'll have another think about races to the fault handler though.
Two threads fault on the same pte,
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
> Below is a patch to implement demand faulting for huge pages. The main
> motivation for changing from prefaulting to demand faulting is so that
> huge page allocations can follow the NUMA API. Currently, huge pages
> are allocated round-robin
On Fri, Aug 05, 2005 at 02:05:42PM -0700, Chen, Kenneth W wrote:
> Adam Litke wrote on Friday, August 05, 2005 8:22 AM
> > Below is a patch to implement demand faulting for huge pages. The main
> > motivation for changing from prefaulting to demand faulting is so that
> > huge page allocations
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
> +int hugetlb_pte_fault(struct mm_struct *mm, struct vm_area_struct *vma,
> + unsigned long address, int write_access)
> +{
> + int ret = VM_FAULT_MINOR;
> + unsigned long idx;
> + pte_t *pte;
> + struct page
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
> Below is a patch to implement demand faulting for huge pages. The main
> motivation for changing from prefaulting to demand faulting is so that
> huge page allocations can follow the NUMA API. Currently, huge pages
> are allocated round-robin
On Fri, Aug 05, 2005 at 12:00:00PM -0500, Adam Litke wrote:
> On Fri, 2005-08-05 at 11:47, Andi Kleen wrote:
> > On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
> > > On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
> > > > On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
> >
On Fri, 5 Aug 2005, Andi Kleen wrote:
> > Unless I am missing something, the call to follow_hugetlb_page() in
> > get_user_pages() is just an optimization. Removing it means
> > follow_page() will be called individually for each PAGE_SIZE page in the
> > huge page. We can probably do better but
On Fri, 2005-08-05 at 11:47, Andi Kleen wrote:
> On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
> > On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
> > > On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
> > > > Below is a patch to implement demand faulting for huge pages.
On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
> On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
> > On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
> > > Below is a patch to implement demand faulting for huge pages. The main
> > > motivation for changing from prefaulting
On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
> On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
> > Below is a patch to implement demand faulting for huge pages. The main
> > motivation for changing from prefaulting to demand faulting is so that
> > huge page allocations can follow
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
> Below is a patch to implement demand faulting for huge pages. The main
> motivation for changing from prefaulting to demand faulting is so that
> huge page allocations can follow the NUMA API. Currently, huge pages
> are allocated
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA API. Currently, huge pages
are allocated round-robin from all NUMA nodes.
The default behavior in SLES9 for
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA API. Currently, huge pages
are allocated round-robin from all NUMA nodes.
The default behavior in SLES9 for
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA API. Currently, huge pages
are allocated
On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA
On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand
On Fri, 2005-08-05 at 11:47, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
Below is a patch to implement demand faulting for huge pages. The main
On Fri, 5 Aug 2005, Andi Kleen wrote:
Unless I am missing something, the call to follow_hugetlb_page() in
get_user_pages() is just an optimization. Removing it means
follow_page() will be called individually for each PAGE_SIZE page in the
huge page. We can probably do better but I
On Fri, Aug 05, 2005 at 12:00:00PM -0500, Adam Litke wrote:
On Fri, 2005-08-05 at 11:47, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 11:37:27AM -0500, Adam Litke wrote:
On Fri, 2005-08-05 at 10:53, Andi Kleen wrote:
On Fri, Aug 05, 2005 at 10:21:38AM -0500, Adam Litke wrote:
Below is
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA API. Currently, huge pages
are allocated round-robin
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
+int hugetlb_pte_fault(struct mm_struct *mm, struct vm_area_struct *vma,
+ unsigned long address, int write_access)
+{
+ int ret = VM_FAULT_MINOR;
+ unsigned long idx;
+ pte_t *pte;
+ struct page *page;
On Fri, Aug 05, 2005 at 02:05:42PM -0700, Chen, Kenneth W wrote:
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can
Adam Litke wrote on Friday, August 05, 2005 8:22 AM
Below is a patch to implement demand faulting for huge pages. The main
motivation for changing from prefaulting to demand faulting is so that
huge page allocations can follow the NUMA API. Currently, huge pages
are allocated round-robin
26 matches
Mail list logo