Commit 41f0a954 authored by Ralph Campbell's avatar Ralph Campbell Committed by Jonathan Corbet

docs/vm: Minor editorial changes in the THP and hugetlbfs

Some minor wording changes and typo corrections.
Signed-off-by: default avatarRalph Campbell <rcampbell@nvidia.com>
Acked-by: default avatarRandy Dunlap <rdunlap@infradead.org>
Acked-by: default avatarMike Kravetz <mike.kravetz@oracle.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Signed-off-by: default avatarJonathan Corbet <corbet@lwn.net>
parent 7d10bdbd
...@@ -85,10 +85,10 @@ Reservation Map Location (Private or Shared) ...@@ -85,10 +85,10 @@ Reservation Map Location (Private or Shared)
A huge page mapping or segment is either private or shared. If private, A huge page mapping or segment is either private or shared. If private,
it is typically only available to a single address space (task). If shared, it is typically only available to a single address space (task). If shared,
it can be mapped into multiple address spaces (tasks). The location and it can be mapped into multiple address spaces (tasks). The location and
semantics of the reservation map is significantly different for two types semantics of the reservation map is significantly different for the two types
of mappings. Location differences are: of mappings. Location differences are:
- For private mappings, the reservation map hangs off the the VMA structure. - For private mappings, the reservation map hangs off the VMA structure.
Specifically, vma->vm_private_data. This reserve map is created at the Specifically, vma->vm_private_data. This reserve map is created at the
time the mapping (mmap(MAP_PRIVATE)) is created. time the mapping (mmap(MAP_PRIVATE)) is created.
- For shared mappings, the reservation map hangs off the inode. Specifically, - For shared mappings, the reservation map hangs off the inode. Specifically,
...@@ -109,15 +109,15 @@ These operations result in a call to the routine hugetlb_reserve_pages():: ...@@ -109,15 +109,15 @@ These operations result in a call to the routine hugetlb_reserve_pages()::
struct vm_area_struct *vma, struct vm_area_struct *vma,
vm_flags_t vm_flags) vm_flags_t vm_flags)
The first thing hugetlb_reserve_pages() does is check for the NORESERVE The first thing hugetlb_reserve_pages() does is check if the NORESERVE
flag was specified in either the shmget() or mmap() call. If NORESERVE flag was specified in either the shmget() or mmap() call. If NORESERVE
was specified, then this routine returns immediately as no reservation was specified, then this routine returns immediately as no reservations
are desired. are desired.
The arguments 'from' and 'to' are huge page indices into the mapping or The arguments 'from' and 'to' are huge page indices into the mapping or
underlying file. For shmget(), 'from' is always 0 and 'to' corresponds to underlying file. For shmget(), 'from' is always 0 and 'to' corresponds to
the length of the segment/mapping. For mmap(), the offset argument could the length of the segment/mapping. For mmap(), the offset argument could
be used to specify the offset into the underlying file. In such a case be used to specify the offset into the underlying file. In such a case,
the 'from' and 'to' arguments have been adjusted by this offset. the 'from' and 'to' arguments have been adjusted by this offset.
One of the big differences between PRIVATE and SHARED mappings is the way One of the big differences between PRIVATE and SHARED mappings is the way
...@@ -138,7 +138,8 @@ to indicate this VMA owns the reservations. ...@@ -138,7 +138,8 @@ to indicate this VMA owns the reservations.
The reservation map is consulted to determine how many huge page reservations The reservation map is consulted to determine how many huge page reservations
are needed for the current mapping/segment. For private mappings, this is are needed for the current mapping/segment. For private mappings, this is
always the value (to - from). However, for shared mappings it is possible that some reservations may already exist within the range (to - from). See the always the value (to - from). However, for shared mappings it is possible that
some reservations may already exist within the range (to - from). See the
section :ref:`Reservation Map Modifications <resv_map_modifications>` section :ref:`Reservation Map Modifications <resv_map_modifications>`
for details on how this is accomplished. for details on how this is accomplished.
...@@ -165,7 +166,7 @@ these counters. ...@@ -165,7 +166,7 @@ these counters.
If there were enough free huge pages and the global count resv_huge_pages If there were enough free huge pages and the global count resv_huge_pages
was adjusted, then the reservation map associated with the mapping is was adjusted, then the reservation map associated with the mapping is
modified to reflect the reservations. In the case of a shared mapping, a modified to reflect the reservations. In the case of a shared mapping, a
file_region will exist that includes the range 'from' 'to'. For private file_region will exist that includes the range 'from' - 'to'. For private
mappings, no modifications are made to the reservation map as lack of an mappings, no modifications are made to the reservation map as lack of an
entry indicates a reservation exists. entry indicates a reservation exists.
...@@ -239,7 +240,7 @@ subpool accounting when the page is freed. ...@@ -239,7 +240,7 @@ subpool accounting when the page is freed.
The routine vma_commit_reservation() is then called to adjust the reserve The routine vma_commit_reservation() is then called to adjust the reserve
map based on the consumption of the reservation. In general, this involves map based on the consumption of the reservation. In general, this involves
ensuring the page is represented within a file_region structure of the region ensuring the page is represented within a file_region structure of the region
map. For shared mappings where the the reservation was present, an entry map. For shared mappings where the reservation was present, an entry
in the reserve map already existed so no change is made. However, if there in the reserve map already existed so no change is made. However, if there
was no reservation in a shared mapping or this was a private mapping a new was no reservation in a shared mapping or this was a private mapping a new
entry must be created. entry must be created.
......
...@@ -4,8 +4,9 @@ ...@@ -4,8 +4,9 @@
Transparent Hugepage Support Transparent Hugepage Support
============================ ============================
This document describes design principles Transparent Hugepage (THP) This document describes design principles for Transparent Hugepage (THP)
Support and its interaction with other parts of the memory management. support and its interaction with other parts of the memory management
system.
Design principles Design principles
================= =================
...@@ -37,23 +38,23 @@ get_user_pages and follow_page ...@@ -37,23 +38,23 @@ get_user_pages and follow_page
get_user_pages and follow_page if run on a hugepage, will return the get_user_pages and follow_page if run on a hugepage, will return the
head or tail pages as usual (exactly as they would do on head or tail pages as usual (exactly as they would do on
hugetlbfs). Most gup users will only care about the actual physical hugetlbfs). Most GUP users will only care about the actual physical
address of the page and its temporary pinning to release after the I/O address of the page and its temporary pinning to release after the I/O
is complete, so they won't ever notice the fact the page is huge. But is complete, so they won't ever notice the fact the page is huge. But
if any driver is going to mangle over the page structure of the tail if any driver is going to mangle over the page structure of the tail
page (like for checking page->mapping or other bits that are relevant page (like for checking page->mapping or other bits that are relevant
for the head page and not the tail page), it should be updated to jump for the head page and not the tail page), it should be updated to jump
to check head page instead. Taking reference on any head/tail page would to check head page instead. Taking a reference on any head/tail page would
prevent page from being split by anyone. prevent the page from being split by anyone.
.. note:: .. note::
these aren't new constraints to the GUP API, and they match the these aren't new constraints to the GUP API, and they match the
same constrains that applies to hugetlbfs too, so any driver capable same constraints that apply to hugetlbfs too, so any driver capable
of handling GUP on hugetlbfs will also work fine on transparent of handling GUP on hugetlbfs will also work fine on transparent
hugepage backed mappings. hugepage backed mappings.
In case you can't handle compound pages if they're returned by In case you can't handle compound pages if they're returned by
follow_page, the FOLL_SPLIT bit can be specified as parameter to follow_page, the FOLL_SPLIT bit can be specified as a parameter to
follow_page, so that it will split the hugepages before returning follow_page, so that it will split the hugepages before returning
them. them.
...@@ -66,11 +67,11 @@ pmd_offset. It's trivial to make the code transparent hugepage aware ...@@ -66,11 +67,11 @@ pmd_offset. It's trivial to make the code transparent hugepage aware
by just grepping for "pmd_offset" and adding split_huge_pmd where by just grepping for "pmd_offset" and adding split_huge_pmd where
missing after pmd_offset returns the pmd. Thanks to the graceful missing after pmd_offset returns the pmd. Thanks to the graceful
fallback design, with a one liner change, you can avoid to write fallback design, with a one liner change, you can avoid to write
hundred if not thousand of lines of complex code to make your code hundreds if not thousands of lines of complex code to make your code
hugepage aware. hugepage aware.
If you're not walking pagetables but you run into a physical hugepage If you're not walking pagetables but you run into a physical hugepage
but you can't handle it natively in your code, you can split it by that you can't handle natively in your code, you can split it by
calling split_huge_page(page). This is what the Linux VM does before calling split_huge_page(page). This is what the Linux VM does before
it tries to swapout the hugepage for example. split_huge_page() can fail it tries to swapout the hugepage for example. split_huge_page() can fail
if the page is pinned and you must handle this correctly. if the page is pinned and you must handle this correctly.
...@@ -97,18 +98,18 @@ split_huge_page() or split_huge_pmd() has a cost. ...@@ -97,18 +98,18 @@ split_huge_page() or split_huge_pmd() has a cost.
To make pagetable walks huge pmd aware, all you need to do is to call To make pagetable walks huge pmd aware, all you need to do is to call
pmd_trans_huge() on the pmd returned by pmd_offset. You must hold the pmd_trans_huge() on the pmd returned by pmd_offset. You must hold the
mmap_sem in read (or write) mode to be sure an huge pmd cannot be mmap_sem in read (or write) mode to be sure a huge pmd cannot be
created from under you by khugepaged (khugepaged collapse_huge_page created from under you by khugepaged (khugepaged collapse_huge_page
takes the mmap_sem in write mode in addition to the anon_vma lock). If takes the mmap_sem in write mode in addition to the anon_vma lock). If
pmd_trans_huge returns false, you just fallback in the old code pmd_trans_huge returns false, you just fallback in the old code
paths. If instead pmd_trans_huge returns true, you have to take the paths. If instead pmd_trans_huge returns true, you have to take the
page table lock (pmd_lock()) and re-run pmd_trans_huge. Taking the page table lock (pmd_lock()) and re-run pmd_trans_huge. Taking the
page table lock will prevent the huge pmd to be converted into a page table lock will prevent the huge pmd being converted into a
regular pmd from under you (split_huge_pmd can run in parallel to the regular pmd from under you (split_huge_pmd can run in parallel to the
pagetable walk). If the second pmd_trans_huge returns false, you pagetable walk). If the second pmd_trans_huge returns false, you
should just drop the page table lock and fallback to the old code as should just drop the page table lock and fallback to the old code as
before. Otherwise you can proceed to process the huge pmd and the before. Otherwise, you can proceed to process the huge pmd and the
hugepage natively. Once finished you can drop the page table lock. hugepage natively. Once finished, you can drop the page table lock.
Refcounts and transparent huge pages Refcounts and transparent huge pages
==================================== ====================================
...@@ -116,61 +117,61 @@ Refcounts and transparent huge pages ...@@ -116,61 +117,61 @@ Refcounts and transparent huge pages
Refcounting on THP is mostly consistent with refcounting on other compound Refcounting on THP is mostly consistent with refcounting on other compound
pages: pages:
- get_page()/put_page() and GUP operate in head page's ->_refcount. - get_page()/put_page() and GUP operate on head page's ->_refcount.
- ->_refcount in tail pages is always zero: get_page_unless_zero() never - ->_refcount in tail pages is always zero: get_page_unless_zero() never
succeed on tail pages. succeeds on tail pages.
- map/unmap of the pages with PTE entry increment/decrement ->_mapcount - map/unmap of the pages with PTE entry increment/decrement ->_mapcount
on relevant sub-page of the compound page. on relevant sub-page of the compound page.
- map/unmap of the whole compound page accounted in compound_mapcount - map/unmap of the whole compound page is accounted for in compound_mapcount
(stored in first tail page). For file huge pages, we also increment (stored in first tail page). For file huge pages, we also increment
->_mapcount of all sub-pages in order to have race-free detection of ->_mapcount of all sub-pages in order to have race-free detection of
last unmap of subpages. last unmap of subpages.
PageDoubleMap() indicates that the page is *possibly* mapped with PTEs. PageDoubleMap() indicates that the page is *possibly* mapped with PTEs.
For anonymous pages PageDoubleMap() also indicates ->_mapcount in all For anonymous pages, PageDoubleMap() also indicates ->_mapcount in all
subpages is offset up by one. This additional reference is required to subpages is offset up by one. This additional reference is required to
get race-free detection of unmap of subpages when we have them mapped with get race-free detection of unmap of subpages when we have them mapped with
both PMDs and PTEs. both PMDs and PTEs.
This is optimization required to lower overhead of per-subpage mapcount This optimization is required to lower the overhead of per-subpage mapcount
tracking. The alternative is alter ->_mapcount in all subpages on each tracking. The alternative is to alter ->_mapcount in all subpages on each
map/unmap of the whole compound page. map/unmap of the whole compound page.
For anonymous pages, we set PG_double_map when a PMD of the page got split For anonymous pages, we set PG_double_map when a PMD of the page is split
for the first time, but still have PMD mapping. The additional references for the first time, but still have a PMD mapping. The additional references
go away with last compound_mapcount. go away with the last compound_mapcount.
File pages get PG_double_map set on first map of the page with PTE and File pages get PG_double_map set on the first map of the page with PTE and
goes away when the page gets evicted from page cache. goes away when the page gets evicted from the page cache.
split_huge_page internally has to distribute the refcounts in the head split_huge_page internally has to distribute the refcounts in the head
page to the tail pages before clearing all PG_head/tail bits from the page page to the tail pages before clearing all PG_head/tail bits from the page
structures. It can be done easily for refcounts taken by page table structures. It can be done easily for refcounts taken by page table
entries. But we don't have enough information on how to distribute any entries, but we don't have enough information on how to distribute any
additional pins (i.e. from get_user_pages). split_huge_page() fails any additional pins (i.e. from get_user_pages). split_huge_page() fails any
requests to split pinned huge page: it expects page count to be equal to requests to split pinned huge pages: it expects page count to be equal to
sum of mapcount of all sub-pages plus one (split_huge_page caller must the sum of mapcount of all sub-pages plus one (split_huge_page caller must
have reference for head page). have a reference to the head page).
split_huge_page uses migration entries to stabilize page->_refcount and split_huge_page uses migration entries to stabilize page->_refcount and
page->_mapcount of anonymous pages. File pages just got unmapped. page->_mapcount of anonymous pages. File pages just get unmapped.
We safe against physical memory scanners too: the only legitimate way We are safe against physical memory scanners too: the only legitimate way
scanner can get reference to a page is get_page_unless_zero(). a scanner can get a reference to a page is get_page_unless_zero().
All tail pages have zero ->_refcount until atomic_add(). This prevents the All tail pages have zero ->_refcount until atomic_add(). This prevents the
scanner from getting a reference to the tail page up to that point. After the scanner from getting a reference to the tail page up to that point. After the
atomic_add() we don't care about the ->_refcount value. We already known how atomic_add() we don't care about the ->_refcount value. We already know how
many references should be uncharged from the head page. many references should be uncharged from the head page.
For head page get_page_unless_zero() will succeed and we don't mind. It's For head page get_page_unless_zero() will succeed and we don't mind. It's
clear where reference should go after split: it will stay on head page. clear where references should go after split: it will stay on the head page.
Note that split_huge_pmd() doesn't have any limitation on refcounting: Note that split_huge_pmd() doesn't have any limitations on refcounting:
pmd can be split at any point and never fails. pmd can be split at any point and never fails.
Partial unmap and deferred_split_huge_page() Partial unmap and deferred_split_huge_page()
...@@ -182,10 +183,10 @@ in page_remove_rmap() and queue the THP for splitting if memory pressure ...@@ -182,10 +183,10 @@ in page_remove_rmap() and queue the THP for splitting if memory pressure
comes. Splitting will free up unused subpages. comes. Splitting will free up unused subpages.
Splitting the page right away is not an option due to locking context in Splitting the page right away is not an option due to locking context in
the place where we can detect partial unmap. It's also might be the place where we can detect partial unmap. It also might be
counterproductive since in many cases partial unmap happens during exit(2) if counterproductive since in many cases partial unmap happens during exit(2) if
a THP crosses a VMA boundary. a THP crosses a VMA boundary.
Function deferred_split_huge_page() is used to queue page for splitting. The function deferred_split_huge_page() is used to queue a page for splitting.
The splitting itself will happen when we get memory pressure via shrinker The splitting itself will happen when we get memory pressure via shrinker
interface. interface.
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment