• Overview
• Basic description of FOLL_PIN
• Which flags are set by each wrapper
• Tracking dma-pinned pages
• FOLL_PIN, FOLL_GET, FOLL_LONGTERM: when to use which flags
  • CASE 1: Direct IO (DIO)
  • CASE 2: RDMA
  • CASE 3: MMU notifier registration, with or without page faulting hardware
  • CASE 4: Pinning for struct page manipulation only
  • CASE 5: Pinning in order to write to the data within the page
• page_maybe_dma_pinned(): the whole point of pinning
• Another way of thinking about FOLL_GET, FOLL_PIN, and FOLL_LONGTERM
• Unit testing
• Other diagnostics
• References

Overview¶

This document describes the following functions:

pin_user_pages()pin_user_pages_fast()pin_user_pages_remote()

Basic description of FOLL_PIN¶

FOLL_PIN and FOLL_LONGTERM are flags that can be passed to the get_user_pages*()(“gup”) family of functions. FOLL_PIN has significant interactions andinterdependencies with FOLL_LONGTERM, so both are covered here.

FOLL_PIN is internal to gup, meaning that it should not appear at the gup callsites. This allows the associated wrapper functions (pin_user_pages*() andothers) to set the correct combination of these flags, and to check for problemsas well.

FOLL_LONGTERM, on the other hand,is allowed to be set at the gup call sites.This is in order to avoid creating a large number of wrapper functions to coverall combinations of get*(), pin*(), FOLL_LONGTERM, and more. Also, thepin_user_pages*() APIs are clearly distinct from the get_user_pages*() APIs, sothat’s a natural dividing line, and a good point to make separate wrapper calls.In other words, use pin_user_pages*() for DMA-pinned pages, andget_user_pages*() for other cases. There are five cases described later on inthis document, to further clarify that concept.

FOLL_PIN and FOLL_GET are mutually exclusive for a given gup call. However,multiple threads and call sites are free to pin the same struct pages, via bothFOLL_PIN and FOLL_GET. It’s just the call site that needs to choose one or theother, not the struct page(s).

The FOLL_PIN implementation is nearly the same as FOLL_GET, except that FOLL_PINuses a different reference counting technique.

FOLL_PIN is a prerequisite to FOLL_LONGTERM. Another way of saying that is,FOLL_LONGTERM is a specific case, more restrictive case of FOLL_PIN.

Which flags are set by each wrapper¶

For these pin_user_pages*() functions, FOLL_PIN is OR’d in with whatever gupflags the caller provides. The caller is required to pass in a non-null structpages* array, and the function then pins pages by incrementing each by a specialvalue: GUP_PIN_COUNTING_BIAS.

For huge pages (and in fact, any compound page of more than 2 pages), theGUP_PIN_COUNTING_BIAS scheme is not used. Instead, an exact form of pin countingis achieved, by using the 3rd struct page in the compound page. A new structpage field, hpage_pinned_refcount, has been added in order to support this.

This approach for compound pages avoids the counting upper limit problems thatare discussed below. Those limitations would have been aggravated severely byhuge pages, because each tail page adds a refcount to the head page. And infact, testing revealed that, without a separate hpage_pinned_refcount field,page overflows were seen in some huge page stress tests.

This also means that huge pages and compound pages (of order > 1) do not sufferfrom the false positives problem that is mentioned below.:

Function--------pin_user_pages          FOLL_PIN is always set internally by this function.pin_user_pages_fast     FOLL_PIN is always set internally by this function.pin_user_pages_remote   FOLL_PIN is always set internally by this function.

For these get_user_pages*() functions, FOLL_GET might not even be specified.Behavior is a little more complex than above. If FOLL_GET wasnot specified,but the caller passed in a non-null struct pages* array, then the functionsets FOLL_GET for you, and proceeds to pin pages by incrementing the refcountof each page by +1.:

Function--------get_user_pages           FOLL_GET is sometimes set internally by this function.get_user_pages_fast      FOLL_GET is sometimes set internally by this function.get_user_pages_remote    FOLL_GET is sometimes set internally by this function.

Tracking dma-pinned pages¶

Some of the key design constraints, and solutions, for tracking dma-pinnedpages:

• An actual reference count, per struct page, is required. This is becausemultiple processes may pin and unpin a page.
• False positives (reporting that a page is dma-pinned, when in fact it is not)are acceptable, but false negatives are not.
• struct page may not be increased in size for this, and all fields are alreadyused.
• Given the above, we can overload the page->_refcount field by using, sort of,the upper bits in that field for a dma-pinned count. “Sort of”, means that,rather than dividing page->_refcount into bit fields, we simple add a medium-large value (GUP_PIN_COUNTING_BIAS, initially chosen to be 1024: 10 bits) topage->_refcount. This provides fuzzy behavior: if a page has get_page() calledon it 1024 times, then it will appear to have a single dma-pinned count.And again, that’s acceptable.

This also leads to limitations: there are only 31-10==21 bits available for acounter that increments 10 bits at a time.

• Callers must specifically request “dma-pinned tracking of pages”. In otherwords, just calling get_user_pages() will not suffice; a new set of functions,pin_user_page() and related, must be used.

FOLL_PIN, FOLL_GET, FOLL_LONGTERM: when to use which flags¶

Thanks to Jan Kara, Vlastimil Babka and several other -mm people, for describingthese categories:

FOLL_PIN

FOLL_PIN | FOLL_LONGTERM

page_maybe_dma_pinned(): the whole point of pinning¶

The whole point of marking pages as “DMA-pinned” or “gup-pinned” is to be ableto query, “is this page DMA-pinned?” That allows code such as page_mkclean()(and file system writeback code in general) to make informed decisions aboutwhat to do when a page cannot be unmapped due to such pins.

What to do in those cases is the subject of a years-long series of discussionsand debates (see the References at the end of this document). It’s a TODO itemhere: fill in the details once that’s worked out. Meanwhile, it’s safe to saythat having this available:

static inline bool page_maybe_dma_pinned(struct page *page)

…is a prerequisite to solving the long-running gup+DMA problem.

Another way of thinking about FOLL_GET, FOLL_PIN, and FOLL_LONGTERM¶

Another way of thinking about these flags is as a progression of restrictions:FOLL_GET is for struct page manipulation, without affecting the data that thestruct page refers to. FOLL_PIN is areplacement for FOLL_GET, and is forshort term pins on pages whose datawill get accessed. As such, FOLL_PIN isa “more severe” form of pinning. And finally, FOLL_LONGTERM is an even morerestrictive case that has FOLL_PIN as a prerequisite: this is for pages thatwill be pinned longterm, and whose data will be accessed.

Unit testing¶

This file:

tools/testing/selftests/vm/gup_benchmark.c

has the following new calls to exercise the new pin*() wrapper functions:

• PIN_FAST_BENCHMARK (./gup_benchmark -a)
• PIN_BENCHMARK (./gup_benchmark -b)

You can monitor how many total dma-pinned pages have been acquired and releasedsince the system was booted, via two new /proc/vmstat entries:

/proc/vmstat/nr_foll_pin_acquired/proc/vmstat/nr_foll_pin_released

Under normal conditions, these two values will be equal unless there are anylong-term [R]DMA pins in place, or during pin/unpin transitions.

• nr_foll_pin_acquired: This is the number of logical pins that have beenacquired since the system was powered on. For huge pages, the head page ispinned once for each page (head page and each tail page) within the huge page.This follows the same sort of behavior that get_user_pages() uses for hugepages: the head page is refcounted once for each tail or head page in the hugepage, when get_user_pages() is applied to a huge page.

• nr_foll_pin_released: The number of logical pins that have been released sincethe system was powered on. Note that pages are released (unpinned) on aPAGE_SIZE granularity, even if the original pin was applied to a huge page.Becaused of the pin count behavior described above in “nr_foll_pin_acquired”,the accounting balances out, so that after doing this:

  pin_user_pages(huge_page);for (each page in huge_page)    unpin_user_page(page);

…the following is expected:

nr_foll_pin_released == nr_foll_pin_acquired

(…unless it was already out of balance due to a long-term RDMA pin being inplace.)

Other diagnostics¶

dump_page() has been enhanced slightly, to handle these new counting fields, andto better report on compound pages in general. Specifically, for compound pageswith order > 1, the exact (hpage_pinned_refcount) pincount is reported.

References¶

• Some slow progress on get_user_pages() (Apr 2, 2019)
• DMA and get_user_pages() (LPC: Dec 12, 2018)
• The trouble with get_user_pages() (Apr 30, 2018)
• LWN kernel index: get_user_pages()

John Hubbard, October, 2019