Re: [PATCH 2/2] mm/hmm: make full use of walk_page_range()
On 7/24/19 4:53 AM, Jason Gunthorpe wrote: On Wed, Jul 24, 2019 at 08:51:46AM +0200, Christoph Hellwig wrote: On Tue, Jul 23, 2019 at 04:30:16PM -0700, Ralph Campbell wrote: hmm_range_snapshot() and hmm_range_fault() both call find_vma() and walk_page_range() in a loop. This is unnecessary duplication since walk_page_range() calls find_vma() in a loop already. Simplify hmm_range_snapshot() and hmm_range_fault() by defining a walk_test() callback function to filter unhandled vmas. I like the approach a lot! But we really need to sort out the duplication between hmm_range_fault and hmm_range_snapshot first, as they are basically the same code. I have patches here: http://git.infradead.org/users/hch/misc.git/commitdiff/a34ccd30ee8a8a3111d9e91711c12901ed7dea74 http://git.infradead.org/users/hch/misc.git/commitdiff/81f442ebac7170815af7770a1efa9c4ab662137e Yeah, that is a straightforward improvement, maybe Ralph should grab these two as part of his series? Sure, no problem. I'll add them in v2 when I fix the other issues in the series. That being said we don't really have any users for the snapshot mode or non-blocking faults, and I don't see any in the immediate pipeline either. If this code was production ready I'd use it in ODP right away. When we first create a ODP MR we'd want to snapshot to pre-load the NIC tables with something other than page fault, but not fault anything. This would be a big performance improvement for ODP. Jason
Re: [PATCH 2/2] mm/hmm: make full use of walk_page_range()
On Wed, Jul 24, 2019 at 08:51:46AM +0200, Christoph Hellwig wrote: > On Tue, Jul 23, 2019 at 04:30:16PM -0700, Ralph Campbell wrote: > > hmm_range_snapshot() and hmm_range_fault() both call find_vma() and > > walk_page_range() in a loop. This is unnecessary duplication since > > walk_page_range() calls find_vma() in a loop already. > > Simplify hmm_range_snapshot() and hmm_range_fault() by defining a > > walk_test() callback function to filter unhandled vmas. > > I like the approach a lot! > > But we really need to sort out the duplication between hmm_range_fault > and hmm_range_snapshot first, as they are basically the same code. I > have patches here: > > http://git.infradead.org/users/hch/misc.git/commitdiff/a34ccd30ee8a8a3111d9e91711c12901ed7dea74 > > http://git.infradead.org/users/hch/misc.git/commitdiff/81f442ebac7170815af7770a1efa9c4ab662137e Yeah, that is a straightforward improvement, maybe Ralph should grab these two as part of his series? > That being said we don't really have any users for the snapshot mode > or non-blocking faults, and I don't see any in the immediate pipeline > either. If this code was production ready I'd use it in ODP right away. When we first create a ODP MR we'd want to snapshot to pre-load the NIC tables with something other than page fault, but not fault anything. This would be a big performance improvement for ODP. Jason
Re: [PATCH 2/2] mm/hmm: make full use of walk_page_range()
On Tue, Jul 23, 2019 at 04:30:16PM -0700, Ralph Campbell wrote: > hmm_range_snapshot() and hmm_range_fault() both call find_vma() and > walk_page_range() in a loop. This is unnecessary duplication since > walk_page_range() calls find_vma() in a loop already. > Simplify hmm_range_snapshot() and hmm_range_fault() by defining a > walk_test() callback function to filter unhandled vmas. I like the approach a lot! But we really need to sort out the duplication between hmm_range_fault and hmm_range_snapshot first, as they are basically the same code. I have patches here: http://git.infradead.org/users/hch/misc.git/commitdiff/a34ccd30ee8a8a3111d9e91711c12901ed7dea74 http://git.infradead.org/users/hch/misc.git/commitdiff/81f442ebac7170815af7770a1efa9c4ab662137e That being said we don't really have any users for the snapshot mode or non-blocking faults, and I don't see any in the immediate pipeline either. It might actually be a better idea to just kill that stuff off for now until we have a user, as code without users is per definition untested and will just bitrot and break. > + const unsigned long device_vma = VM_IO | VM_PFNMAP | VM_MIXEDMAP; > + struct hmm_vma_walk *hmm_vma_walk = walk->private; > + struct hmm_range *range = hmm_vma_walk->range; > + struct vm_area_struct *vma = walk->vma; > + > + /* If range is no longer valid, force retry. */ > + if (!range->valid) > + return -EBUSY; > + > + if (vma->vm_flags & device_vma) Can we just kill off this odd device_vma variable? if (vma->vm_flags & (VM_IO | VM_PFNMAP | VM_MIXEDMAP)) and maybe add a comment on why we are skipping them (because they don't have struct page backing I guess..)
[PATCH 2/2] mm/hmm: make full use of walk_page_range()
hmm_range_snapshot() and hmm_range_fault() both call find_vma() and
walk_page_range() in a loop. This is unnecessary duplication since
walk_page_range() calls find_vma() in a loop already.
Simplify hmm_range_snapshot() and hmm_range_fault() by defining a
walk_test() callback function to filter unhandled vmas.
Signed-off-by: Ralph Campbell
Cc: "Jérôme Glisse"
Cc: Jason Gunthorpe
Cc: Christoph Hellwig
---
mm/hmm.c | 197 ---
1 file changed, 70 insertions(+), 127 deletions(-)
diff --git a/mm/hmm.c b/mm/hmm.c
index 8271f110c243..218557451070 100644
--- a/mm/hmm.c
+++ b/mm/hmm.c
@@ -839,13 +839,40 @@ static int hmm_vma_walk_hugetlb_entry(pte_t *pte,
unsigned long hmask,
#endif
}
-static void hmm_pfns_clear(struct hmm_range *range,
- uint64_t *pfns,
- unsigned long addr,
- unsigned long end)
+static int hmm_vma_walk_test(unsigned long start,
+unsigned long end,
+struct mm_walk *walk)
{
- for (; addr < end; addr += PAGE_SIZE, pfns++)
- *pfns = range->values[HMM_PFN_NONE];
+ const unsigned long device_vma = VM_IO | VM_PFNMAP | VM_MIXEDMAP;
+ struct hmm_vma_walk *hmm_vma_walk = walk->private;
+ struct hmm_range *range = hmm_vma_walk->range;
+ struct vm_area_struct *vma = walk->vma;
+
+ /* If range is no longer valid, force retry. */
+ if (!range->valid)
+ return -EBUSY;
+
+ if (vma->vm_flags & device_vma)
+ return -EFAULT;
+
+ if (is_vm_hugetlb_page(vma)) {
+ if (huge_page_shift(hstate_vma(vma)) != range->page_shift &&
+ range->page_shift != PAGE_SHIFT)
+ return -EINVAL;
+ } else {
+ if (range->page_shift != PAGE_SHIFT)
+ return -EINVAL;
+ }
+
+ /*
+* If vma does not allow read access, then assume that it does not
+* allow write access, either. HMM does not support architectures
+* that allow write without read.
+*/
+ if (!(vma->vm_flags & VM_READ))
+ return -EPERM;
+
+ return 0;
}
/*
@@ -949,65 +976,28 @@ EXPORT_SYMBOL(hmm_range_unregister);
*/
long hmm_range_snapshot(struct hmm_range *range)
{
- const unsigned long device_vma = VM_IO | VM_PFNMAP | VM_MIXEDMAP;
- unsigned long start = range->start, end;
- struct hmm_vma_walk hmm_vma_walk;
+ unsigned long start = range->start;
+ struct hmm_vma_walk hmm_vma_walk = {};
struct hmm *hmm = range->hmm;
- struct vm_area_struct *vma;
- struct mm_walk mm_walk;
+ struct mm_walk mm_walk = {};
+ int ret;
lockdep_assert_held(>mm->mmap_sem);
- do {
- /* If range is no longer valid force retry. */
- if (!range->valid)
- return -EBUSY;
- vma = find_vma(hmm->mm, start);
- if (vma == NULL || (vma->vm_flags & device_vma))
- return -EFAULT;
-
- if (is_vm_hugetlb_page(vma)) {
- if (huge_page_shift(hstate_vma(vma)) !=
- range->page_shift &&
- range->page_shift != PAGE_SHIFT)
- return -EINVAL;
- } else {
- if (range->page_shift != PAGE_SHIFT)
- return -EINVAL;
- }
+ hmm_vma_walk.range = range;
+ hmm_vma_walk.last = start;
+ mm_walk.private = _vma_walk;
- if (!(vma->vm_flags & VM_READ)) {
- /*
-* If vma do not allow read access, then assume that it
-* does not allow write access, either. HMM does not
-* support architecture that allow write without read.
-*/
- hmm_pfns_clear(range, range->pfns,
- range->start, range->end);
- return -EPERM;
- }
+ mm_walk.mm = hmm->mm;
+ mm_walk.pud_entry = hmm_vma_walk_pud;
+ mm_walk.pmd_entry = hmm_vma_walk_pmd;
+ mm_walk.pte_hole = hmm_vma_walk_hole;
+ mm_walk.hugetlb_entry = hmm_vma_walk_hugetlb_entry;
+ mm_walk.test_walk = hmm_vma_walk_test;
- range->vma = vma;
- hmm_vma_walk.pgmap = NULL;
- hmm_vma_walk.last = start;
- hmm_vma_walk.fault = false;
- hmm_vma_walk.range = range;
- mm_walk.private = _vma_walk;
- end = min(range->end, vma->vm_end);
-
- mm_walk.vma = vma;
- mm_walk.mm = vma->vm_mm;
- mm_walk.pte_entry = NULL;
- mm_walk.test_walk = NULL;
- mm_walk.hugetlb_entry =
