mm/gup: Unify hugetlb, speed up thp
Subsystems (6)
Peter Xu
reviewerv2 cover mm/gup: Unify hugetlb, speed up thp
v2:
- Added R-bs
- Merged patch 2 & 4 into a single patch [David, Mike]
- Remove redundant compound_head() [Matthew]
- Still handle failure cases of try_get_folio/page() [Lorenzo]
- Modified more comments in the last removal patch [Lorenzo]
- Fixed hugetlb npages>1 longterm pin issue
- Added two testcase patches at last
Hugetlb has a special path for slow gup that follow_page_mask() is actually skipped completely along with faultin_page(). It's not only confusing, but also duplicating a lot of logics that generic gup already has, making hugetlb slightly special.
This patchset tries to dedup the logic, by first touching up the slow gup code to be able to handle hugetlb pages correctly with the current follow page and faultin routines (where we're mostly there.. due to 10 years ago we did try to optimize thp, but half way done; more below), then at the last patch drop the special path, then the hugetlb gup will always go the generic routine too via faultin_page().
Note that hugetlb is still special for gup, mostly due to the pgtable walking (hugetlb_walk()) that we rely on which is currently per-arch. But this is still one small step forward, and the diffstat might be a proof too that this might be worthwhile.
Then for the "speed up thp" side: as a side effect, when I'm looking at the chunk of code, I found that thp support is actually partially done. It doesn't mean that thp won't work for gup, but as long as **pages pointer passed over, the optimization will be skipped too. Patch 6 should address that, so for thp we now get full speed gup.
For a quick number, "chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10" gives me 13992.50us -> 378.50us. Gup_test is an extreme case, but just to show how it affects thp gups.
Patch 1-5: prepares for the switch Patch 6: switchover to the new code and remove the old Patch 7-8: added some gup test matrix into run_vmtests.sh
Please review, thanks.
Peter Xu (8):
mm/hugetlb: Handle FOLL_DUMP well in follow_page_mask() mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN mm/hugetlb: Add page_mask for hugetlb_follow_page_mask() mm/gup: Cleanup next_page handling
mm/gup: Accelerate thp gup even for "pages != NULL" mm/gup: Retire follow_hugetlb_page() selftests/mm: Add -a to run_vmtests.sh selftests/mm: Add gup test matrix in run_vmtests.sh
fs/userfaultfd.c | 2 +- include/linux/hugetlb.h | 20 +- mm/gup.c | 83 ++++--- mm/hugetlb.c | 256 +++------------------- tools/testing/selftests/mm/run_vmtests.sh | 48 +++- 5 files changed, 119 insertions(+), 290 deletions(-)
Signature · 1 line
2.40.1
Peter Xu
reviewerPATCH v2 1/8 mm/hugetlb: Handle FOLL_DUMP well in follow_page_mask()
Firstly, the no_page_table() is meaningless for hugetlb which is a no-op there, because a hugetlb page always satisfies:
- vma_is_anonymous() == false
- vma->vm_ops->fault != NULL
So we can already safely remove it in hugetlb_follow_page_mask(), alongside with the page* variable.
Meanwhile, what we do in follow_hugetlb_page() actually makes sense for a
dump: we try to fault in the page only if the page cache is already allocated. Let's do the same here for follow_page_mask() on hugetlb.
It should so far has zero effect on real dumps, because that still goes into follow_hugetlb_page(). But this may start to influence a bit on follow_page() users who mimics a "dump page" scenario, but hopefully in a good way. This also paves way for unifying the hugetlb gup-slow.
--- mm/gup.c | 9 ++------- mm/hugetlb.c | 9 +++++++++ 2 files changed, 11 insertions(+), 7 deletions(-)
mm/gup.c +2 -7
@@ -767,7 +767,6 @@ static struct page *follow_page_mask(struct vm_area_struct *vma, struct follow_page_context *ctx) {pgd_t *pgd; -struct page *page;struct mm_struct *mm = vma->vm_mm; ctx->page_mask =0 ;
@@ -780,12 +779,8 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* hugetlb_follow_page_mask is only for follow_page() handling here. * Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ -if (is_vm_hugetlb_page (vma)) { - page =hugetlb_follow_page_mask (vma, address, flags); -if (!page) - page =no_page_table (vma, flags); -return page; - } +if (is_vm_hugetlb_page (vma)) +return hugetlb_follow_page_mask (vma, address, flags); pgd =pgd_offset (mm, address);
mm/hugetlb.c +9 -0
@@ -6498,6 +6498,15 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,spin_unlock (ptl); out_unlock:hugetlb_vma_unlock_read (vma); + +/* +* Fixup retval for dump requests: if pagecache doesn't exist, +* don't try to allocate a new page but just skip it. +*/ +if (!page && (flags &FOLL_DUMP ) && + !hugetlbfs_pagecache_present (h, vma, address)) + page =ERR_PTR (-EFAULT ); +return page; }
Signature · 1 line
2.40.1
Peter Xu
reviewerPATCH v2 2/8 mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN
In reply to Peter Xufollow_page() doesn't use FOLL_PIN, meanwhile hugetlb seems to not be the target of FOLL_WRITE either. However add the checks.
Namely, either the need to CoW due to missing write bit, or proper CoR on !AnonExclusive pages over R/O pins to reject the follow page. That brings this function closer to follow_hugetlb_page().
So we don't care before, and also for now. But we'll care if we switch over slow-gup to use hugetlb_follow_page_mask(). We'll also care when to return -EMLINK properly, as that's the gup internal api to mean "we should do CoR". Not really needed for follow page path, though.
When at it, switching the try_grab_page() to use WARN_ON_ONCE(), to be clear that it just should never fail.
--- mm/hugetlb.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-)
mm/hugetlb.c +15 -9
@@ -6463,13 +6463,6 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,spinlock_t *ptl;pte_t *pte, entry; -/* -* FOLL_PIN is not supported for follow_page(). Ordinary GUP goes via -* follow_hugetlb_page(). -*/ -if (WARN_ON_ONCE (flags &FOLL_PIN )) -return NULL ; -hugetlb_vma_lock_read (vma); pte =hugetlb_walk (vma, haddr,huge_page_size (h));if (!pte)
@@ -6478,8 +6471,21 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, ptl =huge_pte_lock (h, mm, pte); entry =huge_ptep_get (pte);if (pte_present (entry)) { - page =pte_page (entry) + - ((address & ~huge_page_mask (h)) >>PAGE_SHIFT ); + page =pte_page (entry); + +if (gup_must_unshare (vma, flags, page)) { +/* Tell the caller to do Copy-On-Read */ + page =ERR_PTR (-EMLINK ); + goto out; + } + +if ((flags &FOLL_WRITE ) && !pte_write (entry)) { + page =NULL ; + goto out; + } + + page += ((address & ~huge_page_mask (h)) >>PAGE_SHIFT ); +/* * Note that page may be a sub-page, and with vmemmap * optimizations the page struct may be read only.
Signature · 1 line
2.40.1
Re: [PATCH v2 2/8] mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN
Quoted 4 lines · Peter Xu
follow_page()doesn't useFOLL_PIN, meanwhile hugetlb seems to not be the target ofFOLL_WRITEeither. However add the checks.Namely, either the need to CoW due to missing write bit, or proper CoR on
s/CoR/unsharing/
Quoted 7 lines
!AnonExclusive pages over R/O pins to reject the follow page. That brings this function closer to
follow_hugetlb_page().So we don't care before, and also for now. But we'll care if we switch over slow-gup to use
hugetlb_follow_page_mask(). We'll also care when to return-EMLINKproperly, as that's the gup internal api to mean "we should do CoR". Not really needed for follow page path, though.
"we should unshare".
Quoted 29 lines
When at it, switching the
try_grab_page()to useWARN_ON_ONCE(), to be clear that it just should never fail.Reviewed-by Mike Kravetz <[email protected]>Signed-off-by Peter Xu <[email protected]>--- mm/hugetlb.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-)mm/hugetlb.c +0
-7@@ -6463,13 +6463,6 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,spinlock_t *ptl;pte_t *pte, entry; -/* -* FOLL_PIN is not supported for follow_page(). Ordinary GUP goes via -* follow_hugetlb_page(). -*/ -if (WARN_ON_ONCE (flags &FOLL_PIN )) -return NULL ; -hugetlb_vma_lock_read (vma); pte =hugetlb_walk (vma, haddr,huge_page_size (h));if (!pte)
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6476
Quoted 8 lines
mm/hugetlb.c +3
-2@@ -6478,8 +6471,21 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, ptl =huge_pte_lock (h, mm, pte); entry =huge_ptep_get (pte);if (pte_present (entry)) { - page =pte_page (entry) + - ((address & ~huge_page_mask (h)) >>PAGE_SHIFT ); + page =pte_page (entry); + +if (gup_must_unshare (vma, flags, page)) {
All other callers (like follow_page_pte(), including __follow_hugetlb_must_fault())
(a) check for write permissions first.
(b) check for gup_must_unshare() only if !pte_write(entry)
I'd vote to keep these checks as similar as possible to the other GUP code.
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6477
Quoted 1 line
mm/hugetlb.c +1
-0+/* Tell the caller to do Copy-On-Read */
"Tell the caller to unshare".
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6485
Quoted 8 lines
mm/hugetlb.c +8
-0+ page =ERR_PTR (-EMLINK ); + goto out; + } + +if ((flags &FOLL_WRITE ) && !pte_write (entry)) { + page =NULL ; + goto out; + }
I'm confused about pte_write() vs. huge_pte_write(), and I don't know what's right or wrong here.
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerRe: [PATCH v2 2/8] mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN
Quoted 58 lines · David Hildenbrand
Quoted 4 lines · Peter Xu
follow_page()doesn't useFOLL_PIN, meanwhile hugetlb seems to not be the target ofFOLL_WRITEeither. However add the checks.Namely, either the need to CoW due to missing write bit, or proper CoR on
s/CoR/unsharing/
Quoted 7 lines
!AnonExclusive pages over R/O pins to reject the follow page. That brings this function closer to
follow_hugetlb_page().So we don't care before, and also for now. But we'll care if we switch over slow-gup to use
hugetlb_follow_page_mask(). We'll also care when to return-EMLINKproperly, as that's the gup internal api to mean "we should do CoR". Not really needed for follow page path, though."we should unshare".
Quoted 28 lines
When at it, switching the
try_grab_page()to useWARN_ON_ONCE(), to be clear that it just should never fail.Reviewed-by Mike Kravetz <[email protected]>Signed-off-by Peter Xu <[email protected]>--- mm/hugetlb.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-)mm/hugetlb.c +0
-7@@ -6463,13 +6463,6 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,spinlock_t *ptl;pte_t *pte, entry; -/* -* FOLL_PIN is not supported for follow_page(). Ordinary GUP goes via -* follow_hugetlb_page(). -*/ -if (WARN_ON_ONCE (flags &FOLL_PIN )) -return NULL ; -hugetlb_vma_lock_read (vma); pte =hugetlb_walk (vma, haddr,huge_page_size (h));if (!pte)mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6476
Quoted 8 lines
mm/hugetlb.c +3
-2@@ -6478,8 +6471,21 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, ptl =huge_pte_lock (h, mm, pte); entry =huge_ptep_get (pte);if (pte_present (entry)) { - page =pte_page (entry) + - ((address & ~huge_page_mask (h)) >>PAGE_SHIFT ); + page =pte_page (entry); + +if (gup_must_unshare (vma, flags, page)) {All other callers (like
follow_page_pte(), including__follow_hugetlb_must_fault())
Quoted 5 lines
(a) check for write permissions first.
(b) check for
gup_must_unshare()only if !pte_write(entry)I'd vote to keep these checks as similar as possible to the other GUP code.
I'm pretty sure the order doesn't matter here since one for read and one for write.. but sure I can switch the order.
Quoted 5 lines
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6477
Quoted 2 lines
mm/hugetlb.c +1
-0+/* Tell the caller to do Copy-On-Read */ "Tell the caller to unshare".
Quoted 12 lines
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6485
Quoted 8 lines
mm/hugetlb.c +8
-0+ page =ERR_PTR (-EMLINK ); + goto out; + } + +if ((flags &FOLL_WRITE ) && !pte_write (entry)) { + page =NULL ; + goto out; + }I'm confused about
pte_write()vs.huge_pte_write(), and I don't know what's right or wrong here.
AFAICT, they should always be identical in code. But yeah.. I should just use the huge_ version.
Thanks,
Signature · 1 line
Peter Xu
Re: [PATCH v2 2/8] mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN
In reply to Peter XuQuoted 7 lines · Peter Xu
follow_page()doesn't useFOLL_PIN, meanwhile hugetlb seems to not be the target ofFOLL_WRITEeither. However add the checks.Namely, either the need to CoW due to missing write bit, or proper CoR on !AnonExclusive pages over R/O pins to reject the follow page. That brings this function closer to
follow_hugetlb_page().
Quoted 7 lines
So we don't care before, and also for now. But we'll care if we switch over slow-gup to use
hugetlb_follow_page_mask(). We'll also care when to return-EMLINKproperly, as that's the gup internal api to mean "we should do CoR". Not really needed for follow page path, though.When at it, switching the
try_grab_page()to useWARN_ON_ONCE(), to be clear that it just should never fail.
Oh, and does this comment really belong into this patch or am I confused?
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerRe: [PATCH v2 2/8] mm/hugetlb: Prepare hugetlb_follow_page_mask() for FOLL_PIN
Quoted 13 lines · David Hildenbrand
Quoted 12 lines · Peter Xu
follow_page()doesn't useFOLL_PIN, meanwhile hugetlb seems to not be the target ofFOLL_WRITEeither. However add the checks.Namely, either the need to CoW due to missing write bit, or proper CoR on !AnonExclusive pages over R/O pins to reject the follow page. That brings this function closer to
follow_hugetlb_page().So we don't care before, and also for now. But we'll care if we switch over slow-gup to use
hugetlb_follow_page_mask(). We'll also care when to return-EMLINKproperly, as that's the gup internal api to mean "we should do CoR". Not really needed for follow page path, though.
Quoted 4 lines
Quoted 2 lines
When at it, switching the
try_grab_page()to useWARN_ON_ONCE(), to be clear that it just should never fail.Oh, and does this comment really belong into this patch or am I confused?
Ah yeh, thanks for spotting. I used to have it in v1 but I kept the old failure path here to partly address Lorenzo's worry; but then I forgot to add the WARN_ON_ONCE back to guard. I'll remember to add that in v3.
Signature · 1 line
Peter Xu
Peter Xu
reviewerPATCH v2 3/8 mm/hugetlb: Add page_mask for hugetlb_follow_page_mask()
In reply to Peter Xufollow_page() doesn't need it, but we'll start to need it when unifying gup for hugetlb.
--- include/linux/hugetlb.h | 8 +++++--- mm/gup.c | 3 ++- mm/hugetlb.c | 5 ++++- 3 files changed, 11 insertions(+), 5 deletions(-)
include/linux/hugetlb.h +5 -3
@@ -131,7 +131,8 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma,int copy_hugetlb_page_range (struct mm_struct *,struct mm_struct *,struct vm_area_struct *,struct vm_area_struct *);struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags); +unsigned long address,unsigned int flags, +unsigned int *page_mask);long follow_hugetlb_page (struct mm_struct *,struct vm_area_struct *,struct page **,unsigned long *,unsigned long *,long ,unsigned int ,int *);
@@ -297,8 +298,9 @@ static inline void adjust_range_if_pmd_sharing_possible( { } -static inline struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +static inline struct page *hugetlb_follow_page_mask ( +struct vm_area_struct *vma,unsigned long address,unsigned int flags, +unsigned int *page_mask) {BUILD_BUG ();/* should never be compiled in if !CONFIG_HUGETLB_PAGE*/ }
mm/gup.c +2 -1
@@ -780,7 +780,8 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ if (is_vm_hugetlb_page (vma)) -return hugetlb_follow_page_mask (vma, address, flags); +return hugetlb_follow_page_mask (vma, address, flags, + &ctx->page_mask); pgd =pgd_offset (mm, address);
mm/hugetlb.c +4 -1
@@ -6454,7 +6454,8 @@ static inline bool __follow_hugetlb_must_fault(struct vm_area_struct *vma, }struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +unsigned long address,unsigned int flags, +unsigned int *page_mask) {struct hstate *h =hstate_vma (vma);struct mm_struct *mm = vma->vm_mm;
@@ -6499,6 +6500,8 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, page =NULL ; goto out; } + + *page_mask = ~huge_page_mask (h) >>PAGE_SHIFT ; } out:spin_unlock (ptl);
Signature · 1 line
2.40.1
Re: [PATCH v2 3/8] mm/hugetlb: Add page_mask for hugetlb_follow_page_mask()
Quoted 65 lines · Peter Xu
follow_page()doesn't need it, but we'll start to need it when unifying gup for hugetlb.Signed-off-by Peter Xu <[email protected]>--- include/linux/hugetlb.h | 8 +++++--- mm/gup.c | 3 ++- mm/hugetlb.c | 5 ++++- 3 files changed, 11 insertions(+), 5 deletions(-)include/linux/hugetlb.h +5
-3@@ -131,7 +131,8 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma,int copy_hugetlb_page_range (struct mm_struct *,struct mm_struct *,struct vm_area_struct *,struct vm_area_struct *);struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags); +unsigned long address,unsigned int flags, +unsigned int *page_mask);long follow_hugetlb_page (struct mm_struct *,struct vm_area_struct *,struct page **,unsigned long *,unsigned long *,long ,unsigned int ,int *);@@ -297,8 +298,9 @@ static inline void adjust_range_if_pmd_sharing_possible( { } -static inline struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +static inline struct page *hugetlb_follow_page_mask ( +struct vm_area_struct *vma,unsigned long address,unsigned int flags, +unsigned int *page_mask) {BUILD_BUG ();/* should never be compiled in if !CONFIG_HUGETLB_PAGE*/ }mm/gup.c +2
-1@@ -780,7 +780,8 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ if (is_vm_hugetlb_page (vma)) -return hugetlb_follow_page_mask (vma, address, flags); +return hugetlb_follow_page_mask (vma, address, flags, + &ctx->page_mask); pgd =pgd_offset (mm, address);mm/hugetlb.c +2
-1@@ -6454,7 +6454,8 @@ static inline bool __follow_hugetlb_must_fault(struct vm_area_struct *vma, }struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +unsigned long address,unsigned int flags, +unsigned int *page_mask) {struct hstate *h =hstate_vma (vma);struct mm_struct *mm = vma->vm_mm;
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6504
Quoted 5 lines
mm/hugetlb.c +2
-0@@ -6499,6 +6500,8 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, page =NULL ; goto out; } + + *page_mask = ~huge_page_mask (h) >>PAGE_SHIFT ;
As discussed, can be simplified. But can be done on top (or not at all, but it is confusing code).
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerRe: [PATCH v2 3/8] mm/hugetlb: Add page_mask for hugetlb_follow_page_mask()
Quoted 62 lines · David Hildenbrand
Quoted 61 lines · Peter Xu
follow_page()doesn't need it, but we'll start to need it when unifying gup for hugetlb.Signed-off-by Peter Xu <[email protected]>--- include/linux/hugetlb.h | 8 +++++--- mm/gup.c | 3 ++- mm/hugetlb.c | 5 ++++- 3 files changed, 11 insertions(+), 5 deletions(-)include/linux/hugetlb.h +5
-3@@ -131,7 +131,8 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma,int copy_hugetlb_page_range (struct mm_struct *,struct mm_struct *,struct vm_area_struct *,struct vm_area_struct *);struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags); +unsigned long address,unsigned int flags, +unsigned int *page_mask);long follow_hugetlb_page (struct mm_struct *,struct vm_area_struct *,struct page **,unsigned long *,unsigned long *,long ,unsigned int ,int *);@@ -297,8 +298,9 @@ static inline void adjust_range_if_pmd_sharing_possible( { } -static inline struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +static inline struct page *hugetlb_follow_page_mask ( +struct vm_area_struct *vma,unsigned long address,unsigned int flags, +unsigned int *page_mask) {BUILD_BUG ();/* should never be compiled in if !CONFIG_HUGETLB_PAGE*/ }mm/gup.c +2
-1@@ -780,7 +780,8 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ if (is_vm_hugetlb_page (vma)) -return hugetlb_follow_page_mask (vma, address, flags); +return hugetlb_follow_page_mask (vma, address, flags, + &ctx->page_mask); pgd =pgd_offset (mm, address);mm/hugetlb.c +2
-1@@ -6454,7 +6454,8 @@ static inline bool __follow_hugetlb_must_fault(struct vm_area_struct *vma, }struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +unsigned long address,unsigned int flags, +unsigned int *page_mask) {struct hstate *h =hstate_vma (vma);struct mm_struct *mm = vma->vm_mm;@@ -6499,6 +6500,8 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,
Quoted 8 lines
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6504
Quoted 5 lines
mm/hugetlb.c +2
-0page =NULL ; goto out; } + + *page_mask = ~huge_page_mask (h) >>PAGE_SHIFT ;As discussed, can be simplified. But can be done on top (or not at all, but it is confusing code).
Since we decided to make this prettier.. At last I decided to go with this:
*page_mask = (1U << huge_page_order(h)) - 1;
The previous suggestion of PHYS_PFN() will do two shifts over PAGE_SIZE (the other one in huge_page_size()) which might be unnecessary, also, PHYS_ can be slightly misleading too as prefix.
Quoted 2 lines
Reviewed-by David Hildenbrand <[email protected]>
I'll take this with above change, please shoot if not applicable. Thanks,
Signature · 1 line
Peter Xu
Re: [PATCH v2 3/8] mm/hugetlb: Add page_mask for hugetlb_follow_page_mask()
Quoted 76 lines · Peter Xu
Quoted 62 lines · David Hildenbrand
Quoted 61 lines · Peter Xu
follow_page()doesn't need it, but we'll start to need it when unifying gup for hugetlb.Signed-off-by Peter Xu <[email protected]>--- include/linux/hugetlb.h | 8 +++++--- mm/gup.c | 3 ++- mm/hugetlb.c | 5 ++++- 3 files changed, 11 insertions(+), 5 deletions(-)include/linux/hugetlb.h +5
-3@@ -131,7 +131,8 @@ int move_hugetlb_page_tables(struct vm_area_struct *vma,int copy_hugetlb_page_range (struct mm_struct *,struct mm_struct *,struct vm_area_struct *,struct vm_area_struct *);struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags); +unsigned long address,unsigned int flags, +unsigned int *page_mask);long follow_hugetlb_page (struct mm_struct *,struct vm_area_struct *,struct page **,unsigned long *,unsigned long *,long ,unsigned int ,int *);@@ -297,8 +298,9 @@ static inline void adjust_range_if_pmd_sharing_possible( { } -static inline struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +static inline struct page *hugetlb_follow_page_mask ( +struct vm_area_struct *vma,unsigned long address,unsigned int flags, +unsigned int *page_mask) {BUILD_BUG ();/* should never be compiled in if !CONFIG_HUGETLB_PAGE*/ }mm/gup.c +2
-1@@ -780,7 +780,8 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ if (is_vm_hugetlb_page (vma)) -return hugetlb_follow_page_mask (vma, address, flags); +return hugetlb_follow_page_mask (vma, address, flags, + &ctx->page_mask); pgd =pgd_offset (mm, address);mm/hugetlb.c +2
-1@@ -6454,7 +6454,8 @@ static inline bool __follow_hugetlb_must_fault(struct vm_area_struct *vma, }struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma, -unsigned long address,unsigned int flags) +unsigned long address,unsigned int flags, +unsigned int *page_mask) {struct hstate *h =hstate_vma (vma);struct mm_struct *mm = vma->vm_mm;@@ -6499,6 +6500,8 @@ struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma,Quoted 8 lines
mm/hugetlb.c struct page *hugetlb_follow_page_mask(struct vm_area_struct *vma, · line 6504
Quoted 5 lines
mm/hugetlb.c +2
-0page =NULL ; goto out; } + + *page_mask = ~huge_page_mask (h) >>PAGE_SHIFT ;As discussed, can be simplified. But can be done on top (or not at all, but it is confusing code).
Since we decided to make this prettier.. At last I decided to go with this:
*page_mask = (1U << huge_page_order(h)) - 1;
Quoted 9 lines
The previous suggestion of
PHYS_PFN()will do two shifts overPAGE_SIZE(the other one inhuge_page_size()) which might be unnecessary, also,PHYS_can be slightly misleading too as prefix.Quoted 2 lines
Reviewed-by David Hildenbrand <[email protected]>I'll take this with above change, please shoot if not applicable. Thanks,
Perfectly fine :)
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerPATCH v2 4/8 mm/gup: Cleanup next_page handling
In reply to Peter XuThe only path that doesn't use generic "**pages" handling is the gate vma. Make it use the same path, meanwhile tune the next_page label upper to cover "**pages" handling. This prepares for THP handling for "**pages".
--- mm/gup.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)
mm/gup.c +3 -4
@@ -1124,7 +1124,7 @@ static long __get_user_pages(struct mm_struct *mm,if (!vma &&in_gate_area (mm, start)) { ret =get_gate_page (mm, start &PAGE_MASK , gup_flags, &vma, - pages ? &pages[i] :NULL ); + pages ? &page :NULL );if (ret) goto out; ctx.page_mask =0 ;
@@ -1194,19 +1194,18 @@ static long __get_user_pages(struct mm_struct *mm, ret =PTR_ERR (page); goto out; } - - goto next_page; }else if (IS_ERR (page)) { ret =PTR_ERR (page); goto out; } +next_page:if (pages) { pages[i] = page;flush_anon_page (vma, page, start);flush_dcache_page (page); ctx.page_mask =0 ; } -next_page: + page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages;
Signature · 1 line
2.40.1
Re: [PATCH v2 4/8] mm/gup: Cleanup next_page handling
Quoted 35 lines · Peter Xu
The only path that doesn't use generic "**pages" handling is the gate vma. Make it use the same path, meanwhile tune the
next_pagelabel upper to cover "**pages" handling. This prepares for THP handling for "**pages".Reviewed-by Lorenzo Stoakes <[email protected]>Signed-off-by Peter Xu <[email protected]>--- mm/gup.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)mm/gup.c +2
-3@@ -1124,7 +1124,7 @@ static long __get_user_pages(struct mm_struct *mm,if (!vma &&in_gate_area (mm, start)) { ret =get_gate_page (mm, start &PAGE_MASK , gup_flags, &vma, - pages ? &pages[i] :NULL ); + pages ? &page :NULL );if (ret) goto out; ctx.page_mask =0 ;@@ -1194,19 +1194,18 @@ static long __get_user_pages(struct mm_struct *mm, ret =PTR_ERR (page); goto out; } - - goto next_page; }else if (IS_ERR (page)) { ret =PTR_ERR (page); goto out; } +next_page:if (pages) {
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1211
Quoted 10 lines
mm/gup.c +1
-1pages[i] = page;flush_anon_page (vma, page, start);flush_dcache_page (page); ctx.page_mask =0 ; } -next_page: + page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages;
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerPATCH v2 5/8 mm/gup: Accelerate thp gup even for "pages != NULL"
In reply to Peter XuThe acceleration of THP was done with ctx.page_mask, however it'll be ignored if **pages is non-NULL.
The old optimization was introduced in 2013 in 240aadeedc4a ("mm:
accelerate mm_populate() treatment of THP pages"). It didn't explain why we can't optimize the **pages non-NULL case. It's possible that at that time the major goal was for mm_populate() which should be enough back then.
Optimize thp for all cases, by properly looping over each subpage, doing cache flushes, and boost refcounts / pincounts where needed in one go.
This can be verified using gup_test below:
# chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10
Before: 13992.50 ( +-8.75%) After: 378.50 (+-69.62%)
--- mm/gup.c | 51 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-)
mm/gup.c +44 -7
@@ -1199,16 +1199,53 @@ static long __get_user_pages(struct mm_struct *mm, goto out; } next_page: -if (pages) { - pages[i] = page; -flush_anon_page (vma, page, start); -flush_dcache_page (page); - ctx.page_mask =0 ; - } - page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages; + +if (pages) { +struct page *subpage; +unsigned int j; + +/* +* This must be a large folio (and doesn't need to +* be the whole folio; it can be part of it), do +* the refcount work for all the subpages too. +* +* NOTE: here the page may not be the head page +* e.g. when start addr is not thp-size aligned. +* try_grab_folio() should have taken care of tail +* pages. +*/ +if (page_increm >1 ) { +struct folio *folio; + +/* +* Since we already hold refcount on the +* large folio, this should never fail. +*/ + folio =try_grab_folio (page, page_increm -1 , + foll_flags); +if (WARN_ON_ONCE (!folio)) { +/* +* Release the 1st page ref if the +* folio is problematic, fail hard. +*/ +gup_put_folio (page_folio (page),1 , + foll_flags); + ret = -EFAULT ; + goto out; + } + } + +for (j =0 ; j < page_increm; j++) { + subpage =nth_page (page, j); + pages[i+j] = subpage; +flush_anon_page (vma, subpage, start + j *PAGE_SIZE ); +flush_dcache_page (subpage); + } + } + i += page_increm; start += page_increm *PAGE_SIZE ; nr_pages -= page_increm;
Signature · 1 line
2.40.1
Re: [PATCH v2 5/8] mm/gup: Accelerate thp gup even for "pages != NULL"
Quoted 7 lines · Peter Xu
The acceleration of THP was done with ctx.page_mask, however it'll be ignored if **pages is non-NULL.
The old optimization was introduced in 2013 in
240aadeedc4a("mm:
acceleratemm_populate()treatment of THP pages"). It didn't explain why we can't optimize the **pages non-NULL case. It's possible that at that time the major goal was formm_populate()which should be enough back then.
In the past we had these sub-page refcounts for THP. My best guess (and I didn't check if that was still the case in 2013) would be that it was simpler regarding refcount handling to to do it one-subpage at a time.
But I might be just wrong.
Quoted 63 lines
Optimize thp for all cases, by properly looping over each subpage, doing cache flushes, and boost refcounts / pincounts where needed in one go.
This can be verified using
gup_testbelow:# chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10Before: 13992.50 ( +-8.75%) After: 378.50 (+-69.62%)Signed-off-by Peter Xu <[email protected]>--- mm/gup.c | 51 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-)mm/gup.c +29
-7@@ -1199,16 +1199,53 @@ static long __get_user_pages(struct mm_struct *mm, goto out; } next_page: -if (pages) { - pages[i] = page; -flush_anon_page (vma, page, start); -flush_dcache_page (page); - ctx.page_mask =0 ; - } - page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages; + +if (pages) { +struct page *subpage; +unsigned int j; + +/* +* This must be a large folio (and doesn't need to +* be the whole folio; it can be part of it), do +* the refcount work for all the subpages too. +* +* NOTE: here the page may not be the head page +* e.g. when start addr is not thp-size aligned. +* try_grab_folio() should have taken care of tail +* pages. +*/ +if (page_increm >1 ) { +struct folio *folio; + +/* +* Since we already hold refcount on the +* large folio, this should never fail. +*/ + folio =try_grab_folio (page, page_increm -1 , + foll_flags); +if (WARN_ON_ONCE (!folio)) { +/* +* Release the 1st page ref if the +* folio is problematic, fail hard. +*/
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1243
Quoted 10 lines
mm/gup.c +10
-0+gup_put_folio (page_folio (page),1 , + foll_flags); + ret = -EFAULT ; + goto out; + } + } + +for (j =0 ; j < page_increm; j++) { + subpage =nth_page (page, j); + pages[i+j] = subpage;
Doe checkpatch like pages[i+j]? I'd have used spaces around the +.
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1251
Quoted 8 lines
mm/gup.c +5
-0+flush_anon_page (vma, subpage, start + j *PAGE_SIZE ); +flush_dcache_page (subpage); + } + } + i += page_increm; start += page_increm *PAGE_SIZE ; nr_pages -= page_increm;
So, we did the first try_grab_folio() while our page was PMD-mapped udner the PT lock and we had sufficient permissions (e.g., mapped writable, no unsharing required). With FOLL_PIN, we incremented the pincount.
I was wondering if something could have happened ever since we unlocked the PT table lock and possibly PTE-mapped the THP. ... but as it's already pinned, it cannot get shared during fork() [will stay exclusive].
So we can just take additional pins on that folio.
LGTM, although I do like the GUP-fast way of recording+ref'ing it at a central place (see gup_huge_pmd() with record_subpages() and friends), not after the effects.
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerRe: [PATCH v2 5/8] mm/gup: Accelerate thp gup even for "pages != NULL"
Quoted 82 lines · David Hildenbrand
Quoted 7 lines · Peter Xu
The acceleration of THP was done with ctx.page_mask, however it'll be ignored if **pages is non-NULL.
The old optimization was introduced in 2013 in
240aadeedc4a("mm:
acceleratemm_populate()treatment of THP pages"). It didn't explain why we can't optimize the **pages non-NULL case. It's possible that at that time the major goal was formm_populate()which should be enough back then.In the past we had these sub-page refcounts for THP. My best guess (and I didn't check if that was still the case in 2013) would be that it was simpler regarding refcount handling to to do it one-subpage at a time.
But I might be just wrong.
Quoted 67 lines
Optimize thp for all cases, by properly looping over each subpage, doing cache flushes, and boost refcounts / pincounts where needed in one go.
This can be verified using
gup_testbelow:# chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10Before: 13992.50 ( +-8.75%) After: 378.50 (+-69.62%)Signed-off-by Peter Xu <[email protected]>--- mm/gup.c | 51 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-)mm/gup.c +33
-7@@ -1199,16 +1199,53 @@ static long __get_user_pages(struct mm_struct *mm, goto out; } next_page: -if (pages) { - pages[i] = page; -flush_anon_page (vma, page, start); -flush_dcache_page (page); - ctx.page_mask =0 ; - } - page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages; + +if (pages) { +struct page *subpage; +unsigned int j; + +/* +* This must be a large folio (and doesn't need to +* be the whole folio; it can be part of it), do +* the refcount work for all the subpages too. +* +* NOTE: here the page may not be the head page +* e.g. when start addr is not thp-size aligned. +* try_grab_folio() should have taken care of tail +* pages. +*/ +if (page_increm >1 ) { +struct folio *folio; + +/* +* Since we already hold refcount on the +* large folio, this should never fail. +*/ + folio =try_grab_folio (page, page_increm -1 , + foll_flags); +if (WARN_ON_ONCE (!folio)) { +/* +* Release the 1st page ref if the +* folio is problematic, fail hard. +*/ +gup_put_folio (page_folio (page),1 , + foll_flags); + ret = -EFAULT ; + goto out;
Quoted 8 lines
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1243
Quoted 6 lines
mm/gup.c +6
-0+ } + } + +for (j =0 ; j < page_increm; j++) { + subpage =nth_page (page, j); + pages[i+j] = subpage;Doe checkpatch like pages[i+j]? I'd have used spaces around the +.
Can do.
Quoted 20 lines
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1251
Quoted 9 lines
mm/gup.c +5
-0+flush_anon_page (vma, subpage, start + j *PAGE_SIZE ); +flush_dcache_page (subpage); + } + } + i += page_increm; start += page_increm *PAGE_SIZE ; nr_pages -= page_increm;So, we did the first
try_grab_folio()while our page was PMD-mapped udner the PT lock and we had sufficient permissions (e.g., mapped writable, no unsharing required). WithFOLL_PIN, we incremented the pincount.I was wondering if something could have happened ever since we unlocked the PT table lock and possibly PTE-mapped the THP. ... but as it's already pinned, it cannot get shared during
fork()[will stay exclusive].
Quoted 6 lines
So we can just take additional pins on that folio.
LGTM, although I do like the GUP-fast way of recording+ref'ing it at a central place (see
gup_huge_pmd()withrecord_subpages()and friends), not after the effects.
My read on this is follow_page_mask() is also used in follow page, which does not need page*.
No strong opinion here. Maybe we leave this as a follow up even if it can be justified? This patch is probably still the smallest (and still clean) change to speed this whole thing up over either thp or hugetlb.
Signature · 1 line
Peter Xu
Re: [PATCH v2 5/8] mm/gup: Accelerate thp gup even for "pages != NULL"
Quoted 117 lines · Peter Xu
Quoted 82 lines · David Hildenbrand
Quoted 7 lines · Peter Xu
The acceleration of THP was done with ctx.page_mask, however it'll be ignored if **pages is non-NULL.
The old optimization was introduced in 2013 in
240aadeedc4a("mm:
acceleratemm_populate()treatment of THP pages"). It didn't explain why we can't optimize the **pages non-NULL case. It's possible that at that time the major goal was formm_populate()which should be enough back then.In the past we had these sub-page refcounts for THP. My best guess (and I didn't check if that was still the case in 2013) would be that it was simpler regarding refcount handling to to do it one-subpage at a time.
But I might be just wrong.
Quoted 67 lines
Optimize thp for all cases, by properly looping over each subpage, doing cache flushes, and boost refcounts / pincounts where needed in one go.
This can be verified using
gup_testbelow:# chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10Before: 13992.50 ( +-8.75%) After: 378.50 (+-69.62%)Signed-off-by Peter Xu <[email protected]>--- mm/gup.c | 51 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-)mm/gup.c +33
-7@@ -1199,16 +1199,53 @@ static long __get_user_pages(struct mm_struct *mm, goto out; } next_page: -if (pages) { - pages[i] = page; -flush_anon_page (vma, page, start); -flush_dcache_page (page); - ctx.page_mask =0 ; - } - page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages; + +if (pages) { +struct page *subpage; +unsigned int j; + +/* +* This must be a large folio (and doesn't need to +* be the whole folio; it can be part of it), do +* the refcount work for all the subpages too. +* +* NOTE: here the page may not be the head page +* e.g. when start addr is not thp-size aligned. +* try_grab_folio() should have taken care of tail +* pages. +*/ +if (page_increm >1 ) { +struct folio *folio; + +/* +* Since we already hold refcount on the +* large folio, this should never fail. +*/ + folio =try_grab_folio (page, page_increm -1 , + foll_flags); +if (WARN_ON_ONCE (!folio)) { +/* +* Release the 1st page ref if the +* folio is problematic, fail hard. +*/ +gup_put_folio (page_folio (page),1 , + foll_flags); + ret = -EFAULT ; + goto out;Quoted 8 lines
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1243
Quoted 6 lines
mm/gup.c +6
-0+ } + } + +for (j =0 ; j < page_increm; j++) { + subpage =nth_page (page, j); + pages[i+j] = subpage;Doe checkpatch like pages[i+j]? I'd have used spaces around the +.
Can do.
Quoted 23 lines
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1251
Quoted 9 lines
mm/gup.c +5
-0+flush_anon_page (vma, subpage, start + j *PAGE_SIZE ); +flush_dcache_page (subpage); + } + } + i += page_increm; start += page_increm *PAGE_SIZE ; nr_pages -= page_increm;So, we did the first
try_grab_folio()while our page was PMD-mapped udner the PT lock and we had sufficient permissions (e.g., mapped writable, no unsharing required). WithFOLL_PIN, we incremented the pincount.I was wondering if something could have happened ever since we unlocked the PT table lock and possibly PTE-mapped the THP. ... but as it's already pinned, it cannot get shared during
fork()[will stay exclusive].So we can just take additional pins on that folio.
Quoted 6 lines
Quoted 3 lines
LGTM, although I do like the GUP-fast way of recording+ref'ing it at a central place (see
gup_huge_pmd()withrecord_subpages()and friends), not after the effects.My read on this is
follow_page_mask()is also used in follow page, which does not need page*.
Right ... maybe one day we can do that "better".
Quoted 4 lines
No strong opinion here. Maybe we leave this as a follow up even if it can be justified? This patch is probably still the smallest (and still clean) change to speed this whole thing up over either thp or hugetlb.
Sure, we can leave that as a follow-up.
Thinking about why we have the flush_anon_page/flush_dcache_page stuff here and not in GUP-fast ... I suspect that all GUP-fast archs don't need that stuff.
I was wondering if there are some possible races with the flush_anon_page() / flush_dcache_page() on a page that might have been unmapped in the meantime (as we dropped the PT lock ...).
Some flush_dcache_page() implementations do some IMHO confusing page_mapcount() things (like in arch/arc/mm/cache.c). But maybe the unmap code handles that as well ... and most likely these archs don't support THP.
Anyhow, just a note that the flush_anon_page/flush_dcache_page left me confused.
Signature · 3 lines
Cheers, David / dhildenb
Peter Xu
reviewerRe: [PATCH v2 5/8] mm/gup: Accelerate thp gup even for "pages != NULL"
Quoted 3 lines · David Hildenbrand
Thinking about why we have the flush_anon_page/flush_dcache_page stuff here and not in GUP-fast ... I suspect that all GUP-fast archs don't need that stuff.
Yeah that's a bit confusing, and I sincerely don't know the answer. Though here I had the other side of the feeling - I feel like gup-fast should also do it.. but maybe it's just get missed.
AFAIU the idea was that the data can be mis-aligned between user / kernel, and if it's needed on slow gup I don't see why it's not needed in fast..
There're still a few archs that implemented flush_dcache_page() but meanwhile has HAVE_FAST_GUP selected, like arm/arm64/powerpc.
It's just getting out of scope of what this series wanted to achieve.
Quoted 7 lines
I was wondering if there are some possible races with the
flush_anon_page()/flush_dcache_page()on a page that might have been unmapped in the meantime (as we dropped the PT lock ...).Some
flush_dcache_page()implementations do some IMHO confusingpage_mapcount()things (like inarch/arc/mm/cache.c). But maybe the unmap code handles that as well ... and most likely these archs don't support THP.
Maybe true.
It seems that the page_mapcount() was mostly used to identify whether a page is mapped in the userspace address space, if so I'd worry less because the only race possible here, iiuc, is when the user unmaps the page concurrently (and since we got it from gup it must have been mapped once).
Then I would assume the caller should be prepared for that, and the flush_dcache_page() won't matter too much in this case I assume, if the userspace dropped all the data anyway - the whole page* can already be invalid for that VA for a completed unmap.
Quoted 3 lines
Anyhow, just a note that the flush_anon_page/flush_dcache_page left me confused.
I share the same confusion. Hopefully, what this series did here was not changing that, at least not making it worse.
Signature · 1 line
Peter Xu
Re: [PATCH v2 5/8] mm/gup: Accelerate thp gup even for "pages != NULL"
In reply to Peter XuQuoted 65 lines · Peter Xu
The acceleration of THP was done with ctx.page_mask, however it'll be ignored if **pages is non-NULL.
The old optimization was introduced in 2013 in
240aadeedc4a("mm:
acceleratemm_populate()treatment of THP pages"). It didn't explain why we can't optimize the **pages non-NULL case. It's possible that at that time the major goal was formm_populate()which should be enough back then.Optimize thp for all cases, by properly looping over each subpage, doing cache flushes, and boost refcounts / pincounts where needed in one go.
This can be verified using
gup_testbelow:# chrt -f 1 ./gup_test -m 512 -t -L -n 1024 -r 10Before: 13992.50 ( +-8.75%) After: 378.50 (+-69.62%)Signed-off-by Peter Xu <[email protected]>--- mm/gup.c | 51 ++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 7 deletions(-)mm/gup.c +24
-7@@ -1199,16 +1199,53 @@ static long __get_user_pages(struct mm_struct *mm, goto out; } next_page: -if (pages) { - pages[i] = page; -flush_anon_page (vma, page, start); -flush_dcache_page (page); - ctx.page_mask =0 ; - } - page_increm =1 + (~(start >>PAGE_SHIFT ) & ctx.page_mask);if (page_increm > nr_pages) page_increm = nr_pages; + +if (pages) { +struct page *subpage; +unsigned int j; + +/* +* This must be a large folio (and doesn't need to +* be the whole folio; it can be part of it), do +* the refcount work for all the subpages too. +* +* NOTE: here the page may not be the head page +* e.g. when start addr is not thp-size aligned. +* try_grab_folio() should have taken care of tail +* pages. +*/ +if (page_increm >1 ) { +struct folio *folio; + +/* +* Since we already hold refcount on the +* large folio, this should never fail. +*/ + folio =try_grab_folio (page, page_increm -1 , + foll_flags);
mm/gup.c static long __get_user_pages(struct mm_struct *mm, · line 1238
Quoted 10 lines
mm/gup.c +10
-0+if (WARN_ON_ONCE (!folio)) { +/* +* Release the 1st page ref if the +* folio is problematic, fail hard. +*/ +gup_put_folio (page_folio (page),1 , + foll_flags); + ret = -EFAULT ; + goto out; + }
Thanks this looks good to me, I agree it'd be quite surprising for us not to retrieve folio here and probably something has gone wrong if so, so not actually too unreasonable to warn, as long as we error out.
Quoted 14 lines
mm/gup.c +10
-1+ } + +for (j =0 ; j < page_increm; j++) { + subpage =nth_page (page, j); + pages[i+j] = subpage; +flush_anon_page (vma, subpage, start + j *PAGE_SIZE ); +flush_dcache_page (subpage); + } + } + i += page_increm; start += page_increm *PAGE_SIZE ; nr_pages -= page_increm; --
Quoted 2 lines
2.40.1
Looks good to me overall,
Peter Xu
reviewerPATCH v2 6/8 mm/gup: Retire follow_hugetlb_page()
In reply to Peter XuNow __get_user_pages() should be well prepared to handle thp completely, as long as hugetlb gup requests even without the hugetlb's special path.
Time to retire follow_hugetlb_page().
Tweak misc comments to reflect reality of follow_hugetlb_page()'s removal.
--- fs/userfaultfd.c | 2 +- include/linux/hugetlb.h | 12 --- mm/gup.c | 19 ---- mm/hugetlb.c | 224 ---------------------------------------- 4 files changed, 1 insertion(+), 256 deletions(-)
fs/userfaultfd.c +1 -1
@@ -427,7 +427,7 @@ vm_fault_t handle_userfault(struct vm_fault *vmf, unsigned long reason)* * We also don't do userfault handling during * coredumping. hugetlbfs has the special -* follow_hugetlb_page() to skip missing pages in the +* hugetlb_follow_page_mask() to skip missing pages in the * FOLL_DUMP case, anon memory also checks for FOLL_DUMP with * the no_page_table() helper in follow_page_mask(), but the * shmem_vm_ops->fault method is invoked even during
include/linux/hugetlb.h +0 -12
@@ -133,9 +133,6 @@ int copy_hugetlb_page_range(struct mm_struct *, struct mm_struct *,struct page *hugetlb_follow_page_mask (struct vm_area_struct *vma,unsigned long address,unsigned int flags,unsigned int *page_mask); -long follow_hugetlb_page (struct mm_struct *,struct vm_area_struct *, -struct page **,unsigned long *,unsigned long *, -long ,unsigned int ,int *);void unmap_hugepage_range (struct vm_area_struct *,unsigned long ,unsigned long ,struct page *,zap_flags_t );
@@ -305,15 +302,6 @@ static inline struct page *hugetlb_follow_page_mask(BUILD_BUG ();/* should never be compiled in if !CONFIG_HUGETLB_PAGE*/ } -static inline long follow_hugetlb_page (struct mm_struct *mm, -struct vm_area_struct *vma,struct page **pages, -unsigned long *position,unsigned long *nr_pages, -long i,unsigned int flags,int *nonblocking) -{ -BUG (); -return 0 ; -} - static inline int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, struct vm_area_struct *dst_vma,
mm/gup.c +0 -19
@@ -775,9 +775,6 @@ static struct page *follow_page_mask(struct vm_area_struct *vma,* Call hugetlb_follow_page_mask for hugetlb vmas as it will use * special hugetlb page table walking code. This eliminates the * need to check for hugetlb entries in the general walking code. -* -* hugetlb_follow_page_mask is only for follow_page() handling here. -* Ordinary GUP uses follow_hugetlb_page for hugetlb processing. */ if (is_vm_hugetlb_page (vma)) return hugetlb_follow_page_mask(vma, address, flags,
@@ -1138,22 +1135,6 @@ static long __get_user_pages(struct mm_struct *mm, ret =check_vma_flags (vma, gup_flags);if (ret) goto out; - -if (is_vm_hugetlb_page (vma)) { - i =follow_hugetlb_page (mm, vma, pages, - &start, &nr_pages, i, - gup_flags, locked); -if (!*locked) { -/* -* We've got a VM_FAULT_RETRY -* and we've lost mmap_lock. -* We must stop here. -*/ -BUG_ON (gup_flags &FOLL_NOWAIT ); - goto out; - } -continue ; - } } retry:/*
mm/hugetlb.c +0 -224
242 lines not on this page · raw message
Signature · 1 line
2.40.1
Peter Xu
reviewerPATCH v2 7/8 selftests/mm: Add -a to run_vmtests.sh
In reply to Peter XuAllows to specify optional tests in run_vmtests.sh, where we can run time consuming test matrix only when user specified "-a".
--- tools/testing/selftests/mm/run_vmtests.sh | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
tools/testing/selftests/mm/run_vmtests.sh +8 -3
@@ -12,11 +12,14 @@ exitcode=0 usage() { cat <<EOF -usage: ${BASH_SOURCE[0]:-$0} [ -h | -t "<categories>"] +usage: ${BASH_SOURCE[0]:-$0} [ options ] + + -a: run all tests, including extra ones -t: specify specific categories to tests to run -h: display this message -The default behavior is to run all tests. +The default behavior is to run required tests only. If -a is specified, +will run all tests. Alternatively, specific groups tests can be run by passing a string to the -t argument containing one or more of the following categories
@@ -60,9 +63,11 @@ EOF exit 0 } +RUN_ALL=false -while getopts "ht:" OPT; do +while getopts "aht:" OPT; do case ${OPT} in + "a") RUN_ALL=true ;; "h") usage ;; "t") VM_SELFTEST_ITEMS=${OPTARG} ;; esac
Signature · 1 line
2.40.1
Peter Xu
reviewerPATCH v2 8/8 selftests/mm: Add gup test matrix in run_vmtests.sh
In reply to Peter XuAdd a matrix for testing gup based on the current gup_test. Only run the matrix when -a is specified because it's a bit slow.
It covers:
- Different types of huge pages: thp, hugetlb, or no huge page
- Permissions: Write / Read-only
- Fast-gup, with/without
- Types of the GUP: pin / gup / longterm pins
- Shared / Private memories
- GUP size: 1 / 512 / random page sizes
--- tools/testing/selftests/mm/run_vmtests.sh | 37 ++++++++++++++++++++--- 1 file changed, 32 insertions(+), 5 deletions(-)
tools/testing/selftests/mm/run_vmtests.sh +32 -5
@@ -90,6 +90,30 @@ test_selected() { fi } +run_gup_matrix() { + # -t: thp=on, -T: thp=off, -H: hugetlb=on + local hugetlb_mb=$(( needmem_KB / 1024 )) + + for huge in -t -T "-H -m $hugetlb_mb"; do + # -u: gup-fast, -U: gup-basic, -a: pin-fast, -b: pin-basic, -L: pin-longterm + for test_cmd in -u -U -a -b -L; do + # -w: write=1, -W: write=0 + for write in -w -W; do + # -S: shared + for share in -S " "; do + # -n: How many pages to fetch together? 512 is special + # because it's default thp size (or 2M on x86), 123 to + # just test partial gup when hit a huge in whatever form + for num in "-n 1" "-n 512" "-n 123"; do + CATEGORY="gup_test" run_test ./gup_test \ + $huge $test_cmd $write $share $num + done + done + done + done + done +} + # get huge pagesize and freepages from /proc/meminfo while read -r name size unit; do if [ "$name" = "HugePages_Free:" ]; then
@@ -194,13 +218,16 @@ fi CATEGORY="mmap" run_test ./map_fixed_noreplace -# get_user_pages_fast() benchmark -CATEGORY="gup_test" run_test ./gup_test -u -# pin_user_pages_fast() benchmark -CATEGORY="gup_test" run_test ./gup_test -a +if $RUN_ALL; then + run_gup_matrix +else + # get_user_pages_fast() benchmark + CATEGORY="gup_test" run_test ./gup_test -u + # pin_user_pages_fast() benchmark + CATEGORY="gup_test" run_test ./gup_test -a +fi # Dump pages 0, 19, and 4096, using pin_user_pages: CATEGORY="gup_test" run_test ./gup_test -ct -F 0x1 0 19 0x1000 - CATEGORY="gup_test" run_test ./gup_longterm CATEGORY="userfaultfd" run_test ./uffd-unit-tests
Signature · 1 line
2.40.1