An upcoming patch uses try_get_compound_head() more widely, so move it to the top of gup.c.

Also fix a tiny spelling error and a checkpatch.pl warning.

Reviewed-by: Christoph Hellwig hch@lst.de Reviewed-by: Jan Kara jack@suse.cz Reviewed-by: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- mm/gup.c | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c index f764432914c4..3ecce297a47f 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -29,6 +29,21 @@ struct follow_page_context { unsigned int page_mask; };

+/* + * Return the compound head page with ref appropriately incremented, + * or NULL if that failed. + */ +static inline struct page *try_get_compound_head(struct page *page, int refs) +{ + struct page *head = compound_head(page); + + if (WARN_ON_ONCE(page_ref_count(head) < 0)) + return NULL; + if (unlikely(!page_cache_add_speculative(head, refs))) + return NULL; + return head; +} + /** * put_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages * @pages: array of pages to be maybe marked dirty, and definitely released. @@ -1807,20 +1822,6 @@ static void __maybe_unused undo_dev_pagemap(int *nr, int nr_start, } }

-/* - * Return the compund head page with ref appropriately incremented, - * or NULL if that failed. - */ -static inline struct page *try_get_compound_head(struct page *page, int refs) -{ - struct page *head = compound_head(page); - if (WARN_ON_ONCE(page_ref_count(head) < 0)) - return NULL; - if (unlikely(!page_cache_add_speculative(head, refs))) - return NULL; - return head; -} - #ifdef CONFIG_ARCH_HAS_PTE_SPECIAL static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr)

An upcoming patch changes and complicates the refcounting and especially the "put page" aspects of it. In order to keep everything clean, refactor the devmap page release routines:

* Rename put_devmap_managed_page() to page_is_devmap_managed(), and limit the functionality to "read only": return a bool, with no side effects.

* Add a new routine, put_devmap_managed_page(), to handle checking what kind of page it is, and what kind of refcount handling it requires.

* Rename __put_devmap_managed_page() to free_devmap_managed_page(), and limit the functionality to unconditionally freeing a devmap page.

This is originally based on a separate patch by Ira Weiny, which applied to an early version of the put_user_page() experiments. Since then, Jérôme Glisse suggested the refactoring described above.

Cc: Christoph Hellwig hch@lst.de Suggested-by: Jérôme Glisse jglisse@redhat.com Reviewed-by: Dan Williams dan.j.williams@intel.com Reviewed-by: Jan Kara jack@suse.cz Signed-off-by: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- include/linux/mm.h | 17 +++++++++++++---- mm/memremap.c | 16 ++-------------- mm/swap.c | 24 ++++++++++++++++++++++++ 3 files changed, 39 insertions(+), 18 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h index c97ea3b694e6..77a4df06c8a7 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -952,9 +952,10 @@ static inline bool is_zone_device_page(const struct page *page) #endif

#ifdef CONFIG_DEV_PAGEMAP_OPS -void __put_devmap_managed_page(struct page *page); +void free_devmap_managed_page(struct page *page); DECLARE_STATIC_KEY_FALSE(devmap_managed_key); -static inline bool put_devmap_managed_page(struct page *page) + +static inline bool page_is_devmap_managed(struct page *page) { if (!static_branch_unlikely(&devmap_managed_key)) return false; @@ -963,7 +964,6 @@ static inline bool put_devmap_managed_page(struct page *page) switch (page->pgmap->type) { case MEMORY_DEVICE_PRIVATE: case MEMORY_DEVICE_FS_DAX: - __put_devmap_managed_page(page); return true; default: break; @@ -971,7 +971,14 @@ static inline bool put_devmap_managed_page(struct page *page) return false; }

+bool put_devmap_managed_page(struct page *page); + #else /* CONFIG_DEV_PAGEMAP_OPS */ +static inline bool page_is_devmap_managed(struct page *page) +{ + return false; +} + static inline bool put_devmap_managed_page(struct page *page) { return false; @@ -1028,8 +1035,10 @@ static inline void put_page(struct page *page) * need to inform the device driver through callback. See * include/linux/memremap.h and HMM for details. */ - if (put_devmap_managed_page(page)) + if (page_is_devmap_managed(page)) { + put_devmap_managed_page(page); return; + }

if (put_page_testzero(page)) __put_page(page); diff --git a/mm/memremap.c b/mm/memremap.c index e899fa876a62..2ba773859031 100644 --- a/mm/memremap.c +++ b/mm/memremap.c @@ -411,20 +411,8 @@ struct dev_pagemap *get_dev_pagemap(unsigned long pfn, EXPORT_SYMBOL_GPL(get_dev_pagemap);

#ifdef CONFIG_DEV_PAGEMAP_OPS -void __put_devmap_managed_page(struct page *page) +void free_devmap_managed_page(struct page *page) { - int count = page_ref_dec_return(page); - - /* still busy */ - if (count > 1) - return; - - /* only triggered by the dev_pagemap shutdown path */ - if (count == 0) { - __put_page(page); - return; - } - /* notify page idle for dax */ if (!is_device_private_page(page)) { wake_up_var(&page->_refcount); @@ -461,5 +449,5 @@ void __put_devmap_managed_page(struct page *page) page->mapping = NULL; page->pgmap->ops->page_free(page); } -EXPORT_SYMBOL(__put_devmap_managed_page); +EXPORT_SYMBOL(free_devmap_managed_page); #endif /* CONFIG_DEV_PAGEMAP_OPS */ diff --git a/mm/swap.c b/mm/swap.c index 5341ae93861f..49f7c2eea0ba 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -1102,3 +1102,27 @@ void __init swap_setup(void) * _really_ don't want to cluster much more */ } + +#ifdef CONFIG_DEV_PAGEMAP_OPS +bool put_devmap_managed_page(struct page *page) +{ + bool is_devmap = page_is_devmap_managed(page); + + if (is_devmap) { + int count = page_ref_dec_return(page); + + /* + * devmap page refcounts are 1-based, rather than 0-based: if + * refcount is 1, then the page is free and the refcount is + * stable because nobody holds a reference on the page. + */ + if (count == 1) + free_devmap_managed_page(page); + else if (!count) + __put_page(page); + } + + return is_devmap; +} +EXPORT_SYMBOL(put_devmap_managed_page); +#endif

As it says in the updated comment in gup.c: current FOLL_LONGTERM behavior is incompatible with FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on vmas.

However, the corresponding restriction in get_user_pages_remote() was slightly stricter than is actually required: it forbade all FOLL_LONGTERM callers, but we can actually allow FOLL_LONGTERM callers that do not set the "locked" arg.

Update the code and comments to loosen the restriction, allowing FOLL_LONGTERM in some cases.

Also, copy the DAX check ("if a VMA is DAX, don't allow long term pinning") from the VFIO call site, all the way into the internals of get_user_pages_remote() and __gup_longterm_locked(). That is: get_user_pages_remote() calls __gup_longterm_locked(), which in turn calls check_dax_vmas(). This check will then be removed from the VFIO call site in a subsequent patch.

Thanks to Jason Gunthorpe for pointing out a clean way to fix this, and to Dan Williams for helping clarify the DAX refactoring.

Tested-by: Alex Williamson alex.williamson@redhat.com Acked-by: Alex Williamson alex.williamson@redhat.com Reviewed-by: Jason Gunthorpe jgg@mellanox.com Reviewed-by: Ira Weiny ira.weiny@intel.com Suggested-by: Jason Gunthorpe jgg@ziepe.ca Cc: Dan Williams dan.j.williams@intel.com Cc: Jerome Glisse jglisse@redhat.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- mm/gup.c | 27 ++++++++++++++++++++++----- 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c index 3ecce297a47f..c0c56888e7cc 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -29,6 +29,13 @@ struct follow_page_context { unsigned int page_mask; };

+static __always_inline long __gup_longterm_locked(struct task_struct *tsk, + struct mm_struct *mm, + unsigned long start, + unsigned long nr_pages, + struct page **pages, + struct vm_area_struct **vmas, + unsigned int flags); /* * Return the compound head page with ref appropriately incremented, * or NULL if that failed. @@ -1179,13 +1186,23 @@ long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, struct vm_area_struct **vmas, int *locked) { /* - * FIXME: Current FOLL_LONGTERM behavior is incompatible with + * Parts of FOLL_LONGTERM behavior are incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on - * vmas. As there are no users of this flag in this call we simply - * disallow this option for now. + * vmas. However, this only comes up if locked is set, and there are + * callers that do request FOLL_LONGTERM, but do not set locked. So, + * allow what we can. */ - if (WARN_ON_ONCE(gup_flags & FOLL_LONGTERM)) - return -EINVAL; + if (gup_flags & FOLL_LONGTERM) { + if (WARN_ON_ONCE(locked)) + return -EINVAL; + /* + * This will check the vmas (even if our vmas arg is NULL) + * and return -ENOTSUPP if DAX isn't allowed in this case: + */ + return __gup_longterm_locked(tsk, mm, start, nr_pages, pages, + vmas, gup_flags | FOLL_TOUCH | + FOLL_REMOTE); + }

return __get_user_pages_locked(tsk, mm, start, nr_pages, pages, vmas, locked,

Commit 817be129e6f2 ("mm: validate get_user_pages_fast flags") allowed only FOLL_WRITE and FOLL_LONGTERM to be passed to get_user_pages_fast(). This, combined with the fact that get_user_pages_fast() falls back to "slow gup", which *does* accept FOLL_FORCE, leads to an odd situation: if you need FOLL_FORCE, you cannot call get_user_pages_fast().

There does not appear to be any reason for filtering out FOLL_FORCE. There is nothing in the _fast() implementation that requires that we avoid writing to the pages. So it appears to have been an oversight.

Fix by allowing FOLL_FORCE to be set for get_user_pages_fast().

Fixes: 817be129e6f2 ("mm: validate get_user_pages_fast flags") Cc: Christoph Hellwig hch@lst.de Reviewed-by: Leon Romanovsky leonro@mellanox.com Reviewed-by: Jan Kara jack@suse.cz Signed-off-by: John Hubbard jhubbard@nvidia.com --- mm/gup.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/mm/gup.c b/mm/gup.c index c0c56888e7cc..958ab0757389 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -2414,7 +2414,8 @@ int get_user_pages_fast(unsigned long start, int nr_pages, unsigned long addr, len, end; int nr = 0, ret = 0;

- if (WARN_ON_ONCE(gup_flags & ~(FOLL_WRITE | FOLL_LONGTERM))) + if (WARN_ON_ONCE(gup_flags & ~(FOLL_WRITE | FOLL_LONGTERM | + FOLL_FORCE))) return -EINVAL;

start = untagged_addr(start) & PAGE_MASK;

Introduce pin_user_pages*() variations of get_user_pages*() calls, and also pin_longterm_pages*() variations.

For now, these are placeholder calls, until the various call sites are converted to use the correct get_user_pages*() or pin_user_pages*() API.

These variants will eventually all set FOLL_PIN, which is also introduced, and thoroughly documented.

pin_user_pages() pin_user_pages_remote() pin_user_pages_fast()

All pages that are pinned via the above calls, must be unpinned via put_user_page().

The underlying rules are:

* FOLL_PIN is a gup-internal flag, so the call sites should not directly set it. That behavior is enforced with assertions.

* Call sites that want to indicate that they are going to do DirectIO ("DIO") or something with similar characteristics, should call a get_user_pages()-like wrapper call that sets FOLL_PIN. These wrappers will: * Start with "pin_user_pages" instead of "get_user_pages". That makes it easy to find and audit the call sites. * Set FOLL_PIN

* For pages that are received via FOLL_PIN, those pages must be returned via put_user_page().

Thanks to Jan Kara and Vlastimil Babka for explaining the 4 cases in this documentation. (I've reworded it and expanded upon it.)

Reviewed-by: Jan Kara jack@suse.cz Reviewed-by: Mike Rapoport rppt@linux.ibm.com # Documentation Reviewed-by: Jérôme Glisse jglisse@redhat.com Cc: Jonathan Corbet corbet@lwn.net Cc: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- Documentation/core-api/index.rst | 1 + Documentation/core-api/pin_user_pages.rst | 232 ++++++++++++++++++++++ include/linux/mm.h | 63 ++++-- mm/gup.c | 161 +++++++++++++-- 4 files changed, 423 insertions(+), 34 deletions(-) create mode 100644 Documentation/core-api/pin_user_pages.rst

diff --git a/Documentation/core-api/index.rst b/Documentation/core-api/index.rst index ab0eae1c153a..413f7d7c8642 100644 --- a/Documentation/core-api/index.rst +++ b/Documentation/core-api/index.rst @@ -31,6 +31,7 @@ Core utilities generic-radix-tree memory-allocation mm-api + pin_user_pages gfp_mask-from-fs-io timekeeping boot-time-mm diff --git a/Documentation/core-api/pin_user_pages.rst b/Documentation/core-api/pin_user_pages.rst new file mode 100644 index 000000000000..a7a261d869f1 --- /dev/null +++ b/Documentation/core-api/pin_user_pages.rst @@ -0,0 +1,232 @@ +.. SPDX-License-Identifier: GPL-2.0 + +==================================================== +pin_user_pages() and related calls +==================================================== + +.. contents:: :local: + +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 and +interdependencies with FOLL_LONGTERM, so both are covered here. + +FOLL_PIN is internal to gup, meaning that it should not appear at the gup call +sites. This allows the associated wrapper functions (pin_user_pages*() and +others) to set the correct combination of these flags, and to check for problems +as 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 cover +all combinations of get*(), pin*(), FOLL_LONGTERM, and more. Also, the +pin_user_pages*() APIs are clearly distinct from the get_user_pages*() APIs, so +that'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, and +get_user_pages*() for other cases. There are four cases described later on in +this 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 both +FOLL_PIN and FOLL_GET. It's just the call site that needs to choose one or the +other, not the struct page(s). + +The FOLL_PIN implementation is nearly the same as FOLL_GET, except that FOLL_PIN +uses a different reference counting technique. + +FOLL_PIN is a prerequisite to FOLL_LONGTGERM. 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 gup +flags the caller provides. The caller is required to pass in a non-null struct +pages* array, and the function then pin pages by incrementing each by a special +value. For now, that value is +1, just like get_user_pages*().:: + + 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 was *not* specified, +but the caller passed in a non-null struct pages* array, then the function +sets FOLL_GET for you, and proceeds to pin pages by incrementing the refcount +of 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-pinned +pages: + +* An actual reference count, per struct page, is required. This is because + multiple 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 already + used. + +* 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) to + page->_refcount. This provides fuzzy behavior: if a page has get_page() called + on 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 a +counter that increments 10 bits at a time. + +TODO: for 1GB and larger huge pages, this is cutting it close. That's because +when pin_user_pages() follows such pages, it increments the head page by "1" +(where "1" used to mean "+1" for get_user_pages(), but now means "+1024" for +pin_user_pages()) for each tail page. So if you have a 1GB huge page: + +* There are 256K (18 bits) worth of 4 KB tail pages. +* There are 21 bits available to count up via GUP_PIN_COUNTING_BIAS (that is, + 10 bits at a time) +* There are 21 - 18 == 3 bits available to count. Except that there aren't, + because you need to allow for a few normal get_page() calls on the head page, + as well. Fortunately, the approach of using addition, rather than "hard" + bitfields, within page->_refcount, allows for sharing these bits gracefully. + But we're still looking at about 8 references. + +This, however, is a missing feature more than anything else, because it's easily +solved by addressing an obvious inefficiency in the original get_user_pages() +approach of retrieving pages: stop treating all the pages as if they were +PAGE_SIZE. Retrieve huge pages as huge pages. The callers need to be aware of +this, so some work is required. Once that's in place, this limitation mostly +disappears from view, because there will be ample refcounting range available. + +* Callers must specifically request "dma-pinned tracking of pages". In other + words, 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 describing +these categories: + +CASE 1: Direct IO (DIO) +----------------------- +There are GUP references to pages that are serving +as DIO buffers. These buffers are needed for a relatively short time (so they +are not "long term"). No special synchronization with page_mkclean() or +munmap() is provided. Therefore, flags to set at the call site are: :: + + FOLL_PIN + +...but rather than setting FOLL_PIN directly, call sites should use one of +the pin_user_pages*() routines that set FOLL_PIN. + +CASE 2: RDMA +------------ +There are GUP references to pages that are serving as DMA +buffers. These buffers are needed for a long time ("long term"). No special +synchronization with page_mkclean() or munmap() is provided. Therefore, flags +to set at the call site are: :: + + FOLL_PIN | FOLL_LONGTERM + +NOTE: Some pages, such as DAX pages, cannot be pinned with longterm pins. That's +because DAX pages do not have a separate page cache, and so "pinning" implies +locking down file system blocks, which is not (yet) supported in that way. + +CASE 3: Hardware with page faulting support +------------------------------------------- +Here, a well-written driver doesn't normally need to pin pages at all. However, +if the driver does choose to do so, it can register MMU notifiers for the range, +and will be called back upon invalidation. Either way (avoiding page pinning, or +using MMU notifiers to unpin upon request), there is proper synchronization with +both filesystem and mm (page_mkclean(), munmap(), etc). + +Therefore, neither flag needs to be set. + +In this case, ideally, neither get_user_pages() nor pin_user_pages() should be +called. Instead, the software should be written so that it does not pin pages. +This allows mm and filesystems to operate more efficiently and reliably. + +CASE 4: Pinning for struct page manipulation only +------------------------------------------------- +Here, normal GUP calls are sufficient, so neither flag needs to be set. + +page_dma_pinned(): the whole point of pinning +============================================= + +The whole point of marking pages as "DMA-pinned" or "gup-pinned" is to be able +to query, "is this page DMA-pinned?" That allows code such as page_mkclean() +(and file system writeback code in general) to make informed decisions about +what 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 discussions +and debates (see the References at the end of this document). It's a TODO item +here: fill in the details once that's worked out. Meanwhile, it's safe to say +that having this available: :: + + static inline bool page_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 the +struct page refers to. FOLL_PIN is a *replacement* for FOLL_GET, and is for +short term pins on pages whose data *will* get accessed. As such, FOLL_PIN is +a "more severe" form of pinning. And finally, FOLL_LONGTERM is an even more +restrictive case that has FOLL_PIN as a prerequisite: this is for pages that +will 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 released +since the system was booted, via two new /proc/vmstat entries: :: + + /proc/vmstat/nr_foll_pin_requested + /proc/vmstat/nr_foll_pin_requested + +Those are both going to show zero, unless CONFIG_DEBUG_VM is set. This is +because there is a noticeable performance drop in put_user_page(), when they +are activated. + +References +========== + +* `Some slow progress on get_user_pages() (Apr 2, 2019) https://lwn.net/Articles/784574/`_ +* `DMA and get_user_pages() (LPC: Dec 12, 2018) https://lwn.net/Articles/774411/`_ +* `The trouble with get_user_pages() (Apr 30, 2018) https://lwn.net/Articles/753027/`_ + +John Hubbard, October, 2019 diff --git a/include/linux/mm.h b/include/linux/mm.h index 77a4df06c8a7..0fb9929e00af 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1048,16 +1048,14 @@ static inline void put_page(struct page *page) * put_user_page() - release a gup-pinned page * @page: pointer to page to be released * - * Pages that were pinned via get_user_pages*() must be released via - * either put_user_page(), or one of the put_user_pages*() routines - * below. This is so that eventually, pages that are pinned via - * get_user_pages*() can be separately tracked and uniquely handled. In - * particular, interactions with RDMA and filesystems need special - * handling. + * Pages that were pinned via pin_user_pages*() must be released via either + * put_user_page(), or one of the put_user_pages*() routines. This is so that + * eventually such pages can be separately tracked and uniquely handled. In + * particular, interactions with RDMA and filesystems need special handling. * * put_user_page() and put_page() are not interchangeable, despite this early * implementation that makes them look the same. put_user_page() calls must - * be perfectly matched up with get_user_page() calls. + * be perfectly matched up with pin*() calls. */ static inline void put_user_page(struct page *page) { @@ -1515,9 +1513,16 @@ long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked); +long pin_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, + unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas, int *locked); long get_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas); +long pin_user_pages(unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas); long get_user_pages_locked(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, int *locked); long get_user_pages_unlocked(unsigned long start, unsigned long nr_pages, @@ -1525,6 +1530,8 @@ long get_user_pages_unlocked(unsigned long start, unsigned long nr_pages,

int get_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages); +int pin_user_pages_fast(unsigned long start, int nr_pages, + unsigned int gup_flags, struct page **pages);

int account_locked_vm(struct mm_struct *mm, unsigned long pages, bool inc); int __account_locked_vm(struct mm_struct *mm, unsigned long pages, bool inc, @@ -2588,13 +2595,15 @@ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, #define FOLL_ANON 0x8000 /* don't do file mappings */ #define FOLL_LONGTERM 0x10000 /* mapping lifetime is indefinite: see below */ #define FOLL_SPLIT_PMD 0x20000 /* split huge pmd before returning */ +#define FOLL_PIN 0x40000 /* pages must be released via put_user_page() */

/* - * NOTE on FOLL_LONGTERM: + * FOLL_PIN and FOLL_LONGTERM may be used in various combinations with each + * other. Here is what they mean, and how to use them: * * FOLL_LONGTERM indicates that the page will be held for an indefinite time - * period _often_ under userspace control. This is contrasted with - * iov_iter_get_pages() where usages which are transient. + * period _often_ under userspace control. This is in contrast to + * iov_iter_get_pages(), whose usages are transient. * * FIXME: For pages which are part of a filesystem, mappings are subject to the * lifetime enforced by the filesystem and we need guarantees that longterm @@ -2609,11 +2618,39 @@ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, * Currently only get_user_pages() and get_user_pages_fast() support this flag * and calls to get_user_pages_[un]locked are specifically not allowed. This * is due to an incompatibility with the FS DAX check and - * FAULT_FLAG_ALLOW_RETRY + * FAULT_FLAG_ALLOW_RETRY. * - * In the CMA case: longterm pins in a CMA region would unnecessarily fragment - * that region. And so CMA attempts to migrate the page before pinning when + * In the CMA case: long term pins in a CMA region would unnecessarily fragment + * that region. And so, CMA attempts to migrate the page before pinning, when * FOLL_LONGTERM is specified. + * + * FOLL_PIN indicates that a special kind of tracking (not just page->_refcount, + * but an additional pin counting system) will be invoked. This is intended for + * anything that gets a page reference and then touches page data (for example, + * Direct IO). This lets the filesystem know that some non-file-system entity is + * potentially changing the pages' data. In contrast to FOLL_GET (whose pages + * are released via put_page()), FOLL_PIN pages must be released, ultimately, by + * a call to put_user_page(). + * + * FOLL_PIN is similar to FOLL_GET: both of these pin pages. They use different + * and separate refcounting mechanisms, however, and that means that each has + * its own acquire and release mechanisms: + * + * FOLL_GET: get_user_pages*() to acquire, and put_page() to release. + * + * FOLL_PIN: pin_user_pages*() to acquire, and put_user_pages to release. + * + * FOLL_PIN and FOLL_GET are mutually exclusive for a given function call. + * (The underlying pages may experience both FOLL_GET-based and FOLL_PIN-based + * calls applied to them, and that's perfectly OK. This is a constraint on the + * callers, not on the pages.) + * + * FOLL_PIN should be set internally by the pin_user_pages*() APIs, never + * directly by the caller. That's in order to help avoid mismatches when + * releasing pages: get_user_pages*() pages must be released via put_page(), + * while pin_user_pages*() pages must be released via put_user_page(). + * + * Please see Documentation/vm/pin_user_pages.rst for more information. */

static inline int vm_fault_to_errno(vm_fault_t vm_fault, int foll_flags) diff --git a/mm/gup.c b/mm/gup.c index 958ab0757389..4862ff982bc3 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -201,6 +201,10 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, spinlock_t *ptl; pte_t *ptep, pte;

+ /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) == + (FOLL_PIN | FOLL_GET))) + return ERR_PTR(-EINVAL); retry: if (unlikely(pmd_bad(*pmd))) return no_page_table(vma, flags); @@ -818,7 +822,7 @@ static long __get_user_pages(struct task_struct *tsk, struct mm_struct *mm,

start = untagged_addr(start);

- VM_BUG_ON(!!pages != !!(gup_flags & FOLL_GET)); + VM_BUG_ON(!!pages != !!(gup_flags & (FOLL_GET | FOLL_PIN)));

/* * If FOLL_FORCE is set then do not force a full fault as the hinting @@ -1042,7 +1046,16 @@ static __always_inline long __get_user_pages_locked(struct task_struct *tsk, BUG_ON(*locked != 1); }

- if (pages) + /* + * FOLL_PIN and FOLL_GET are mutually exclusive. Traditional behavior + * is to set FOLL_GET if the caller wants pages[] filled in (but has + * carelessly failed to specify FOLL_GET), so keep doing that, but only + * for FOLL_GET, not for the newer FOLL_PIN. + * + * FOLL_PIN always expects pages to be non-null, but no need to assert + * that here, as any failures will be obvious enough. + */ + if (pages && !(flags & FOLL_PIN)) flags |= FOLL_GET;

pages_done = 0; @@ -1185,6 +1198,13 @@ long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { + /* + * FOLL_PIN must only be set internally by the pin_user_pages*() APIs, + * never directly by the caller, so enforce that with an assertion: + */ + if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) + return -EINVAL; + /* * Parts of FOLL_LONGTERM behavior are incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on @@ -1400,6 +1420,14 @@ static long __get_user_pages_locked(struct task_struct *tsk, finish_or_fault: return i ? : -EFAULT; } + +long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, + unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas, int *locked) +{ + return 0; +} #endif /* !CONFIG_MMU */

#if defined(CONFIG_FS_DAX) || defined (CONFIG_CMA) @@ -1654,6 +1682,13 @@ long get_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas) { + /* + * FOLL_PIN must only be set internally by the pin_user_pages*() APIs, + * never directly by the caller, so enforce that with an assertion: + */ + if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) + return -EINVAL; + return __gup_longterm_locked(current, current->mm, start, nr_pages, pages, vmas, gup_flags | FOLL_TOUCH); } @@ -2392,30 +2427,15 @@ static int __gup_longterm_unlocked(unsigned long start, int nr_pages, return ret; }

-/** - * get_user_pages_fast() - pin user pages in memory - * @start: starting user address - * @nr_pages: number of pages from start to pin - * @gup_flags: flags modifying pin behaviour - * @pages: array that receives pointers to the pages pinned. - * Should be at least nr_pages long. - * - * Attempt to pin user pages in memory without taking mm->mmap_sem. - * If not successful, it will fall back to taking the lock and - * calling get_user_pages(). - * - * Returns number of pages pinned. This may be fewer than the number - * requested. If nr_pages is 0 or negative, returns 0. If no pages - * were pinned, returns -errno. - */ -int get_user_pages_fast(unsigned long start, int nr_pages, - unsigned int gup_flags, struct page **pages) +static int internal_get_user_pages_fast(unsigned long start, int nr_pages, + unsigned int gup_flags, + struct page **pages) { unsigned long addr, len, end; int nr = 0, ret = 0;

if (WARN_ON_ONCE(gup_flags & ~(FOLL_WRITE | FOLL_LONGTERM | - FOLL_FORCE))) + FOLL_FORCE | FOLL_PIN))) return -EINVAL;

start = untagged_addr(start) & PAGE_MASK; @@ -2455,4 +2475,103 @@ int get_user_pages_fast(unsigned long start, int nr_pages,

return ret; } + +/** + * get_user_pages_fast() - pin user pages in memory + * @start: starting user address + * @nr_pages: number of pages from start to pin + * @gup_flags: flags modifying pin behaviour + * @pages: array that receives pointers to the pages pinned. + * Should be at least nr_pages long. + * + * Attempt to pin user pages in memory without taking mm->mmap_sem. + * If not successful, it will fall back to taking the lock and + * calling get_user_pages(). + * + * Returns number of pages pinned. This may be fewer than the number requested. + * If nr_pages is 0 or negative, returns 0. If no pages were pinned, returns + * -errno. + */ +int get_user_pages_fast(unsigned long start, int nr_pages, + unsigned int gup_flags, struct page **pages) +{ + /* + * FOLL_PIN must only be set internally by the pin_user_pages*() APIs, + * never directly by the caller, so enforce that: + */ + if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) + return -EINVAL; + + return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); +} EXPORT_SYMBOL_GPL(get_user_pages_fast); + +/** + * pin_user_pages_fast() - pin user pages in memory without taking locks + * + * For now, this is a placeholder function, until various call sites are + * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, + * this is identical to get_user_pages_fast(). + * + * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It + * is NOT intended for Case 2 (RDMA: long-term pins). + */ +int pin_user_pages_fast(unsigned long start, int nr_pages, + unsigned int gup_flags, struct page **pages) +{ + /* + * This is a placeholder, until the pin functionality is activated. + * Until then, just behave like the corresponding get_user_pages*() + * routine. + */ + return get_user_pages_fast(start, nr_pages, gup_flags, pages); +} +EXPORT_SYMBOL_GPL(pin_user_pages_fast); + +/** + * pin_user_pages_remote() - pin pages of a remote process (task != current) + * + * For now, this is a placeholder function, until various call sites are + * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, + * this is identical to get_user_pages_remote(). + * + * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It + * is NOT intended for Case 2 (RDMA: long-term pins). + */ +long pin_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, + unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas, int *locked) +{ + /* + * This is a placeholder, until the pin functionality is activated. + * Until then, just behave like the corresponding get_user_pages*() + * routine. + */ + return get_user_pages_remote(tsk, mm, start, nr_pages, gup_flags, pages, + vmas, locked); +} +EXPORT_SYMBOL(pin_user_pages_remote); + +/** + * pin_user_pages() - pin user pages in memory for use by other devices + * + * For now, this is a placeholder function, until various call sites are + * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, + * this is identical to get_user_pages(). + * + * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It + * is NOT intended for Case 2 (RDMA: long-term pins). + */ +long pin_user_pages(unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas) +{ + /* + * This is a placeholder, until the pin functionality is activated. + * Until then, just behave like the corresponding get_user_pages*() + * routine. + */ + return get_user_pages(start, nr_pages, gup_flags, pages, vmas); +} +EXPORT_SYMBOL(pin_user_pages);

Convert infiniband to use the new pin_user_pages*() calls.

Also, revert earlier changes to Infiniband ODP that had it using put_user_page(). ODP is "Case 3" in Documentation/core-api/pin_user_pages.rst, which is to say, normal get_user_pages() and put_page() is the API to use there.

The new pin_user_pages*() calls replace corresponding get_user_pages*() calls, and set the FOLL_PIN flag. The FOLL_PIN flag requires that the caller must return the pages via put_user_page*() calls, but infiniband was already doing that as part of an earlier commit.

Reviewed-by: Jason Gunthorpe jgg@mellanox.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- drivers/infiniband/core/umem.c | 2 +- drivers/infiniband/core/umem_odp.c | 13 ++++++------- drivers/infiniband/hw/hfi1/user_pages.c | 2 +- drivers/infiniband/hw/mthca/mthca_memfree.c | 2 +- drivers/infiniband/hw/qib/qib_user_pages.c | 2 +- drivers/infiniband/hw/qib/qib_user_sdma.c | 2 +- drivers/infiniband/hw/usnic/usnic_uiom.c | 2 +- drivers/infiniband/sw/siw/siw_mem.c | 2 +- 8 files changed, 13 insertions(+), 14 deletions(-)

diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c index 214e87aa609d..55daefaa9b88 100644 --- a/drivers/infiniband/core/umem.c +++ b/drivers/infiniband/core/umem.c @@ -266,7 +266,7 @@ struct ib_umem *ib_umem_get(struct ib_udata *udata, unsigned long addr, sg = umem->sg_head.sgl;

while (npages) { - ret = get_user_pages_fast(cur_base, + ret = pin_user_pages_fast(cur_base, min_t(unsigned long, npages, PAGE_SIZE / sizeof(struct page *)), diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/umem_odp.c index e42d44e501fd..abc3bb6578cc 100644 --- a/drivers/infiniband/core/umem_odp.c +++ b/drivers/infiniband/core/umem_odp.c @@ -308,9 +308,8 @@ EXPORT_SYMBOL(ib_umem_odp_release); * The function returns -EFAULT if the DMA mapping operation fails. It returns * -EAGAIN if a concurrent invalidation prevents us from updating the page. * - * The page is released via put_user_page even if the operation failed. For - * on-demand pinning, the page is released whenever it isn't stored in the - * umem. + * The page is released via put_page even if the operation failed. For on-demand + * pinning, the page is released whenever it isn't stored in the umem. */ static int ib_umem_odp_map_dma_single_page( struct ib_umem_odp *umem_odp, @@ -363,7 +362,7 @@ static int ib_umem_odp_map_dma_single_page( }

out: - put_user_page(page); + put_page(page); return ret; }

@@ -473,7 +472,7 @@ int ib_umem_odp_map_dma_pages(struct ib_umem_odp *umem_odp, u64 user_virt, ret = -EFAULT; break; } - put_user_page(local_page_list[j]); + put_page(local_page_list[j]); continue; }

@@ -500,8 +499,8 @@ int ib_umem_odp_map_dma_pages(struct ib_umem_odp *umem_odp, u64 user_virt, * ib_umem_odp_map_dma_single_page(). */ if (npages - (j + 1) > 0) - put_user_pages(&local_page_list[j+1], - npages - (j + 1)); + release_pages(&local_page_list[j+1], + npages - (j + 1)); break; } } diff --git a/drivers/infiniband/hw/hfi1/user_pages.c b/drivers/infiniband/hw/hfi1/user_pages.c index 469acb961fbd..9a94761765c0 100644 --- a/drivers/infiniband/hw/hfi1/user_pages.c +++ b/drivers/infiniband/hw/hfi1/user_pages.c @@ -106,7 +106,7 @@ int hfi1_acquire_user_pages(struct mm_struct *mm, unsigned long vaddr, size_t np int ret; unsigned int gup_flags = FOLL_LONGTERM | (writable ? FOLL_WRITE : 0);

- ret = get_user_pages_fast(vaddr, npages, gup_flags, pages); + ret = pin_user_pages_fast(vaddr, npages, gup_flags, pages); if (ret < 0) return ret;

diff --git a/drivers/infiniband/hw/mthca/mthca_memfree.c b/drivers/infiniband/hw/mthca/mthca_memfree.c index edccfd6e178f..8269ab040c21 100644 --- a/drivers/infiniband/hw/mthca/mthca_memfree.c +++ b/drivers/infiniband/hw/mthca/mthca_memfree.c @@ -472,7 +472,7 @@ int mthca_map_user_db(struct mthca_dev *dev, struct mthca_uar *uar, goto out; }

- ret = get_user_pages_fast(uaddr & PAGE_MASK, 1, + ret = pin_user_pages_fast(uaddr & PAGE_MASK, 1, FOLL_WRITE | FOLL_LONGTERM, pages); if (ret < 0) goto out; diff --git a/drivers/infiniband/hw/qib/qib_user_pages.c b/drivers/infiniband/hw/qib/qib_user_pages.c index 6bf764e41891..7fc4b5f81fcd 100644 --- a/drivers/infiniband/hw/qib/qib_user_pages.c +++ b/drivers/infiniband/hw/qib/qib_user_pages.c @@ -108,7 +108,7 @@ int qib_get_user_pages(unsigned long start_page, size_t num_pages,

down_read(&current->mm->mmap_sem); for (got = 0; got < num_pages; got += ret) { - ret = get_user_pages(start_page + got * PAGE_SIZE, + ret = pin_user_pages(start_page + got * PAGE_SIZE, num_pages - got, FOLL_LONGTERM | FOLL_WRITE | FOLL_FORCE, p + got, NULL); diff --git a/drivers/infiniband/hw/qib/qib_user_sdma.c b/drivers/infiniband/hw/qib/qib_user_sdma.c index 05190edc2611..1a3cc2957e3a 100644 --- a/drivers/infiniband/hw/qib/qib_user_sdma.c +++ b/drivers/infiniband/hw/qib/qib_user_sdma.c @@ -670,7 +670,7 @@ static int qib_user_sdma_pin_pages(const struct qib_devdata *dd, else j = npages;

- ret = get_user_pages_fast(addr, j, FOLL_LONGTERM, pages); + ret = pin_user_pages_fast(addr, j, FOLL_LONGTERM, pages); if (ret != j) { i = 0; j = ret; diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index 62e6ffa9ad78..600896727d34 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -141,7 +141,7 @@ static int usnic_uiom_get_pages(unsigned long addr, size_t size, int writable, ret = 0;

while (npages) { - ret = get_user_pages(cur_base, + ret = pin_user_pages(cur_base, min_t(unsigned long, npages, PAGE_SIZE / sizeof(struct page *)), gup_flags | FOLL_LONGTERM, diff --git a/drivers/infiniband/sw/siw/siw_mem.c b/drivers/infiniband/sw/siw/siw_mem.c index e99983f07663..e53b07dcfed5 100644 --- a/drivers/infiniband/sw/siw/siw_mem.c +++ b/drivers/infiniband/sw/siw/siw_mem.c @@ -426,7 +426,7 @@ struct siw_umem *siw_umem_get(u64 start, u64 len, bool writable) while (nents) { struct page **plist = &umem->page_chunk[i].plist[got];

- rv = get_user_pages(first_page_va, nents, + rv = pin_user_pages(first_page_va, nents, foll_flags | FOLL_LONGTERM, plist, NULL); if (rv < 0)

Convert drm/via to use the new pin_user_pages_fast() call, which sets FOLL_PIN. Setting FOLL_PIN is now required for code that requires tracking of pinned pages, and therefore for any code that calls put_user_page().

In partial anticipation of this work, the drm/via driver was already calling put_user_page() instead of put_page(). Therefore, in order to convert from the get_user_pages()/put_page() model, to the pin_user_pages()/put_user_page() model, the only change required is to change get_user_pages() to pin_user_pages().

Acked-by: Daniel Vetter daniel.vetter@ffwll.ch Reviewed-by: Jérôme Glisse jglisse@redhat.com Reviewed-by: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- drivers/gpu/drm/via/via_dmablit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/via/via_dmablit.c b/drivers/gpu/drm/via/via_dmablit.c index 3db000aacd26..37c5e572993a 100644 --- a/drivers/gpu/drm/via/via_dmablit.c +++ b/drivers/gpu/drm/via/via_dmablit.c @@ -239,7 +239,7 @@ via_lock_all_dma_pages(drm_via_sg_info_t *vsg, drm_via_dmablit_t *xfer) vsg->pages = vzalloc(array_size(sizeof(struct page *), vsg->num_pages)); if (NULL == vsg->pages) return -ENOMEM; - ret = get_user_pages_fast((unsigned long)xfer->mem_addr, + ret = pin_user_pages_fast((unsigned long)xfer->mem_addr, vsg->num_pages, vsg->direction == DMA_FROM_DEVICE ? FOLL_WRITE : 0, vsg->pages);

Convert net/xdp to use the new pin_longterm_pages() call, which sets FOLL_PIN. Setting FOLL_PIN is now required for code that requires tracking of pinned pages.

In partial anticipation of this work, the net/xdp code was already calling put_user_page() instead of put_page(). Therefore, in order to convert from the get_user_pages()/put_page() model, to the pin_user_pages()/put_user_page() model, the only change required here is to change get_user_pages() to pin_user_pages().

Acked-by: Björn Töpel bjorn.topel@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- net/xdp/xdp_umem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/xdp/xdp_umem.c b/net/xdp/xdp_umem.c index 3049af269fbf..d071003b5e76 100644 --- a/net/xdp/xdp_umem.c +++ b/net/xdp/xdp_umem.c @@ -291,7 +291,7 @@ static int xdp_umem_pin_pages(struct xdp_umem *umem) return -ENOMEM;

down_read(&current->mm->mmap_sem); - npgs = get_user_pages(umem->address, umem->npgs, + npgs = pin_user_pages(umem->address, umem->npgs, gup_flags | FOLL_LONGTERM, &umem->pgs[0], NULL); up_read(&current->mm->mmap_sem);

1. Change v4l2 from get_user_pages() to pin_user_pages().

2. Because all FOLL_PIN-acquired pages must be released via put_user_page(), also convert the put_page() call over to put_user_pages_dirty_lock().

Acked-by: Hans Verkuil hverkuil-cisco@xs4all.nl Cc: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- drivers/media/v4l2-core/videobuf-dma-sg.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/drivers/media/v4l2-core/videobuf-dma-sg.c b/drivers/media/v4l2-core/videobuf-dma-sg.c index 28262190c3ab..162a2633b1e3 100644 --- a/drivers/media/v4l2-core/videobuf-dma-sg.c +++ b/drivers/media/v4l2-core/videobuf-dma-sg.c @@ -183,12 +183,12 @@ static int videobuf_dma_init_user_locked(struct videobuf_dmabuf *dma, dprintk(1, "init user [0x%lx+0x%lx => %d pages]\n", data, size, dma->nr_pages);

- err = get_user_pages(data & PAGE_MASK, dma->nr_pages, + err = pin_user_pages(data & PAGE_MASK, dma->nr_pages, flags | FOLL_LONGTERM, dma->pages, NULL);

if (err != dma->nr_pages) { dma->nr_pages = (err >= 0) ? err : 0; - dprintk(1, "get_user_pages: err=%d [%d]\n", err, + dprintk(1, "pin_user_pages: err=%d [%d]\n", err, dma->nr_pages); return err < 0 ? err : -EINVAL; } @@ -349,11 +349,8 @@ int videobuf_dma_free(struct videobuf_dmabuf *dma) BUG_ON(dma->sglen);

if (dma->pages) { - for (i = 0; i < dma->nr_pages; i++) { - if (dma->direction == DMA_FROM_DEVICE) - set_page_dirty_lock(dma->pages[i]); - put_page(dma->pages[i]); - } + put_user_pages_dirty_lock(dma->pages, dma->nr_pages, + dma->direction == DMA_FROM_DEVICE); kfree(dma->pages); dma->pages = NULL; }

1. Convert from get_user_pages() to pin_user_pages().

2. As required by pin_user_pages(), release these pages via put_user_page().

Cc: Jan Kara jack@suse.cz Signed-off-by: John Hubbard jhubbard@nvidia.com --- arch/powerpc/mm/book3s64/iommu_api.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/mm/book3s64/iommu_api.c b/arch/powerpc/mm/book3s64/iommu_api.c index 56cc84520577..a86547822034 100644 --- a/arch/powerpc/mm/book3s64/iommu_api.c +++ b/arch/powerpc/mm/book3s64/iommu_api.c @@ -103,7 +103,7 @@ static long mm_iommu_do_alloc(struct mm_struct *mm, unsigned long ua, for (entry = 0; entry < entries; entry += chunk) { unsigned long n = min(entries - entry, chunk);

- ret = get_user_pages(ua + (entry << PAGE_SHIFT), n, + ret = pin_user_pages(ua + (entry << PAGE_SHIFT), n, FOLL_WRITE | FOLL_LONGTERM, mem->hpages + entry, NULL); if (ret == n) { @@ -167,9 +167,8 @@ static long mm_iommu_do_alloc(struct mm_struct *mm, unsigned long ua, return 0;

free_exit: - /* free the reference taken */ - for (i = 0; i < pinned; i++) - put_page(mem->hpages[i]); + /* free the references taken */ + put_user_pages(mem->hpages, pinned);

vfree(mem->hpas); kfree(mem); @@ -215,7 +214,8 @@ static void mm_iommu_unpin(struct mm_iommu_table_group_mem_t *mem) if (mem->hpas[i] & MM_IOMMU_TABLE_GROUP_PAGE_DIRTY) SetPageDirty(page);

- put_page(page); + put_user_page(page); + mem->hpas[i] = 0; } }

Fix the gup benchmark flags to use the symbolic FOLL_WRITE, instead of a hard-coded "1" value.

Also, clean up the filtering of gup flags a little, by just doing it once before issuing any of the get_user_pages*() calls. This makes it harder to overlook, instead of having little "gup_flags & 1" phrases in the function calls.

Reviewed-by: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- mm/gup_benchmark.c | 9 ++++++--- tools/testing/selftests/vm/gup_benchmark.c | 6 +++++- 2 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/mm/gup_benchmark.c b/mm/gup_benchmark.c index 7dd602d7f8db..7fc44d25eca7 100644 --- a/mm/gup_benchmark.c +++ b/mm/gup_benchmark.c @@ -48,18 +48,21 @@ static int __gup_benchmark_ioctl(unsigned int cmd, nr = (next - addr) / PAGE_SIZE; }

+ /* Filter out most gup flags: only allow a tiny subset here: */ + gup->flags &= FOLL_WRITE; + switch (cmd) { case GUP_FAST_BENCHMARK: - nr = get_user_pages_fast(addr, nr, gup->flags & 1, + nr = get_user_pages_fast(addr, nr, gup->flags, pages + i); break; case GUP_LONGTERM_BENCHMARK: nr = get_user_pages(addr, nr, - (gup->flags & 1) | FOLL_LONGTERM, + gup->flags | FOLL_LONGTERM, pages + i, NULL); break; case GUP_BENCHMARK: - nr = get_user_pages(addr, nr, gup->flags & 1, pages + i, + nr = get_user_pages(addr, nr, gup->flags, pages + i, NULL); break; default: diff --git a/tools/testing/selftests/vm/gup_benchmark.c b/tools/testing/selftests/vm/gup_benchmark.c index 485cf06ef013..389327e9b30a 100644 --- a/tools/testing/selftests/vm/gup_benchmark.c +++ b/tools/testing/selftests/vm/gup_benchmark.c @@ -18,6 +18,9 @@ #define GUP_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark) #define GUP_BENCHMARK _IOWR('g', 3, struct gup_benchmark)

+/* Just the flags we need, copied from mm.h: */ +#define FOLL_WRITE 0x01 /* check pte is writable */ + struct gup_benchmark { __u64 get_delta_usec; __u64 put_delta_usec; @@ -85,7 +88,8 @@ int main(int argc, char **argv) }

gup.nr_pages_per_call = nr_pages; - gup.flags = write; + if (write) + gup.flags |= FOLL_WRITE;

fd = open("/sys/kernel/debug/gup_benchmark", O_RDWR); if (fd == -1)

Add tracking of pages that were pinned via FOLL_PIN.

As mentioned in the FOLL_PIN documentation, callers who effectively set FOLL_PIN are required to ultimately free such pages via unpin_user_page(). The effect is similar to FOLL_GET, and may be thought of as "FOLL_GET for DIO and/or RDMA use".

Pages that have been pinned via FOLL_PIN are identifiable via a new function call:

bool page_dma_pinned(struct page *page);

What to do in response to encountering such a page, is left to later patchsets. There is discussion about this in [1], [2], and [3].

This also changes a BUG_ON(), to a WARN_ON(), in follow_page_mask().

[1] Some slow progress on get_user_pages() (Apr 2, 2019): https://lwn.net/Articles/784574/ [2] DMA and get_user_pages() (LPC: Dec 12, 2018): https://lwn.net/Articles/774411/ [3] The trouble with get_user_pages() (Apr 30, 2018): https://lwn.net/Articles/753027/

Suggested-by: Jan Kara jack@suse.cz Suggested-by: Jérôme Glisse jglisse@redhat.com Reviewed-by: Jan Kara jack@suse.cz Reviewed-by: Jérôme Glisse jglisse@redhat.com Reviewed-by: Ira Weiny ira.weiny@intel.com Cc: Kirill A. Shutemov kirill.shutemov@linux.intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- Documentation/core-api/pin_user_pages.rst | 2 +- include/linux/mm.h | 77 ++++- include/linux/mmzone.h | 2 + include/linux/page_ref.h | 10 + mm/gup.c | 369 ++++++++++++++++------ mm/huge_memory.c | 26 +- mm/hugetlb.c | 25 +- mm/vmstat.c | 2 + 8 files changed, 397 insertions(+), 116 deletions(-)

diff --git a/Documentation/core-api/pin_user_pages.rst b/Documentation/core-api/pin_user_pages.rst index fd2a19c96189..2640d122de29 100644 --- a/Documentation/core-api/pin_user_pages.rst +++ b/Documentation/core-api/pin_user_pages.rst @@ -53,7 +53,7 @@ Which flags are set by each wrapper For these pin_user_pages*() functions, FOLL_PIN is OR'd in with whatever gup flags the caller provides. The caller is required to pass in a non-null struct pages* array, and the function then pin pages by incrementing each by a special -value. For now, that value is +1, just like get_user_pages*().:: +value: GUP_PIN_COUNTING_BIAS.::

Function -------- diff --git a/include/linux/mm.h b/include/linux/mm.h index 6a1a357e7d86..1765332f27e8 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1005,6 +1005,10 @@ static inline bool is_pci_p2pdma_page(const struct page *page) #define page_ref_zero_or_close_to_overflow(page) \ ((unsigned int) page_ref_count(page) + 127u <= 127u)

+#define page_ref_zero_or_close_to_bias_overflow(page) \ + ((unsigned int) page_ref_count(page) + \ + GUP_PIN_COUNTING_BIAS <= GUP_PIN_COUNTING_BIAS) + static inline void get_page(struct page *page) { page = compound_head(page); @@ -1016,6 +1020,8 @@ static inline void get_page(struct page *page) page_ref_inc(page); }

+bool __must_check try_grab_page(struct page *page, unsigned int flags); + static inline __must_check bool try_get_page(struct page *page) { page = compound_head(page); @@ -1044,29 +1050,70 @@ static inline void put_page(struct page *page) __put_page(page); }

-/** - * unpin_user_page() - release a gup-pinned page - * @page: pointer to page to be released +/* + * GUP_PIN_COUNTING_BIAS, and the associated functions that use it, overload + * the page's refcount so that two separate items are tracked: the original page + * reference count, and also a new count of how many pin_user_pages() calls were + * made against the page. ("gup-pinned" is another term for the latter). + * + * With this scheme, pin_user_pages() becomes special: such pages are marked as + * distinct from normal pages. As such, the unpin_user_page() call (and its + * variants) must be used in order to release gup-pinned pages. * - * Pages that were pinned via pin_user_pages*() must be released via either - * unpin_user_page(), or one of the unpin_user_pages*() routines. This is so - * that eventually such pages can be separately tracked and uniquely handled. In - * particular, interactions with RDMA and filesystems need special handling. + * Choice of value: * - * unpin_user_page() and put_page() are not interchangeable, despite this early - * implementation that makes them look the same. unpin_user_page() calls must - * be perfectly matched up with pin*() calls. + * By making GUP_PIN_COUNTING_BIAS a power of two, debugging of page reference + * counts with respect to pin_user_pages() and unpin_user_page() becomes + * simpler, due to the fact that adding an even power of two to the page + * refcount has the effect of using only the upper N bits, for the code that + * counts up using the bias value. This means that the lower bits are left for + * the exclusive use of the original code that increments and decrements by one + * (or at least, by much smaller values than the bias value). + * + * Of course, once the lower bits overflow into the upper bits (and this is + * OK, because subtraction recovers the original values), then visual inspection + * no longer suffices to directly view the separate counts. However, for normal + * applications that don't have huge page reference counts, this won't be an + * issue. + * + * Locking: the lockless algorithm described in page_cache_get_speculative() + * and page_cache_gup_pin_speculative() provides safe operation for + * get_user_pages and page_mkclean and other calls that race to set up page + * table entries. */ -static inline void unpin_user_page(struct page *page) -{ - put_page(page); -} +#define GUP_PIN_COUNTING_BIAS (1U << 10)

+void unpin_user_page(struct page *page); void unpin_user_pages_dirty_lock(struct page **pages, unsigned long npages, bool make_dirty); - void unpin_user_pages(struct page **pages, unsigned long npages);

+/** + * page_dma_pinned() - report if a page is pinned for DMA. + * + * This function checks if a page has been pinned via a call to + * pin_user_pages*(). + * + * The return value is partially fuzzy: false is not fuzzy, because it means + * "definitely not pinned for DMA", but true means "probably pinned for DMA, but + * possibly a false positive due to having at least GUP_PIN_COUNTING_BIAS worth + * of normal page references". + * + * False positives are OK, because: a) it's unlikely for a page to get that many + * refcounts, and b) all the callers of this routine are expected to be able to + * deal gracefully with a false positive. + * + * For more information, please see Documentation/vm/pin_user_pages.rst. + * + * @page: pointer to page to be queried. + * @Return: True, if it is likely that the page has been "dma-pinned". + * False, if the page is definitely not dma-pinned. + */ +static inline bool page_dma_pinned(struct page *page) +{ + return (page_ref_count(compound_head(page))) >= GUP_PIN_COUNTING_BIAS; +} + #if defined(CONFIG_SPARSEMEM) && !defined(CONFIG_SPARSEMEM_VMEMMAP) #define SECTION_IN_PAGE_FLAGS #endif diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 89d8ff06c9ce..a7418f7a44da 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -244,6 +244,8 @@ enum node_stat_item { NR_DIRTIED, /* page dirtyings since bootup */ NR_WRITTEN, /* page writings since bootup */ NR_KERNEL_MISC_RECLAIMABLE, /* reclaimable non-slab kernel pages */ + NR_FOLL_PIN_REQUESTED, /* via: pin_user_page(), gup flag: FOLL_PIN */ + NR_FOLL_PIN_RETURNED, /* pages returned via unpin_user_page() */ NR_VM_NODE_STAT_ITEMS };

diff --git a/include/linux/page_ref.h b/include/linux/page_ref.h index 14d14beb1f7f..b9cbe553d1e7 100644 --- a/include/linux/page_ref.h +++ b/include/linux/page_ref.h @@ -102,6 +102,16 @@ static inline void page_ref_sub(struct page *page, int nr) __page_ref_mod(page, -nr); }

+static inline int page_ref_sub_return(struct page *page, int nr) +{ + int ret = atomic_sub_return(nr, &page->_refcount); + + if (page_ref_tracepoint_active(__tracepoint_page_ref_mod)) + __page_ref_mod(page, -nr); + + return ret; +} + static inline void page_ref_inc(struct page *page) { atomic_inc(&page->_refcount); diff --git a/mm/gup.c b/mm/gup.c index 73aedcefa4bd..038b71165a76 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -36,6 +36,20 @@ static __always_inline long __gup_longterm_locked(struct task_struct *tsk, struct page **pages, struct vm_area_struct **vmas, unsigned int flags); + +#ifdef CONFIG_DEBUG_VM +static inline void __update_proc_vmstat(struct page *page, + enum node_stat_item item, int count) +{ + mod_node_page_state(page_pgdat(page), item, count); +} +#else +static inline void __update_proc_vmstat(struct page *page, + enum node_stat_item item, int count) +{ +} +#endif + /* * Return the compound head page with ref appropriately incremented, * or NULL if that failed. @@ -51,6 +65,145 @@ static inline struct page *try_get_compound_head(struct page *page, int refs) return head; }

+/** + * try_pin_compound_head() - mark a compound page as being used by + * pin_user_pages*(). + * + * This is the FOLL_PIN counterpart to try_get_compound_head(). + * + * @page: pointer to page to be marked + * @Return: the compound head page, with ref appropriately incremented, + * or NULL upon failure. + */ +__must_check struct page *try_pin_compound_head(struct page *page, int refs) +{ + struct page *head = try_get_compound_head(page, + GUP_PIN_COUNTING_BIAS * refs); + if (!head) + return NULL; + + __update_proc_vmstat(page, NR_FOLL_PIN_REQUESTED, refs); + return head; +} + +/* + * try_grab_compound_head() - attempt to elevate a page's refcount, by a + * flags-dependent amount. + * + * This has a default assumption of "use FOLL_GET behavior, if FOLL_PIN is not + * set". + * + * "grab" names in this file mean, "look at flags to decide whether to use + * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount. + */ +static __maybe_unused struct page *try_grab_compound_head(struct page *page, + int refs, + unsigned int flags) +{ + if (flags & FOLL_PIN) + return try_pin_compound_head(page, refs); + + return try_get_compound_head(page, refs); +} + +/** + * try_grab_page() - elevate a page's refcount by a flag-dependent amount + * + * This might not do anything at all, depending on the flags argument. + * + * "grab" names in this file mean, "look at flags to decide whether to use + * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount. + * + * @page: pointer to page to be grabbed + * @flags: gup flags: these are the FOLL_* flag values. + * + * Either FOLL_PIN or FOLL_GET (or neither) may be set, but not both at the same + * time. (That's true throughout the get_user_pages*() and pin_user_pages*() + * APIs.) Cases: + * + * FOLL_GET: page's refcount will be incremented by 1. + * FOLL_PIN: page's refcount will be incremented by GUP_PIN_COUNTING_BIAS. + * + * Return: true for success, or if no action was required (if neither FOLL_PIN + * nor FOLL_GET was set, nothing is done). False for failure: FOLL_GET or + * FOLL_PIN was set, but the page could not be grabbed. + */ +bool __must_check try_grab_page(struct page *page, unsigned int flags) +{ + if (flags & FOLL_GET) + return try_get_page(page); + else if (flags & FOLL_PIN) { + page = compound_head(page); + WARN_ON_ONCE(flags & FOLL_GET); + + if (WARN_ON_ONCE(page_ref_zero_or_close_to_bias_overflow(page))) + return false; + + page_ref_add(page, GUP_PIN_COUNTING_BIAS); + __update_proc_vmstat(page, NR_FOLL_PIN_REQUESTED, 1); + } + + return true; +} + +#ifdef CONFIG_DEV_PAGEMAP_OPS +static bool __unpin_devmap_managed_user_page(struct page *page) +{ + bool is_devmap = page_is_devmap_managed(page); + + if (is_devmap) { + int count = page_ref_sub_return(page, GUP_PIN_COUNTING_BIAS); + + __update_proc_vmstat(page, NR_FOLL_PIN_RETURNED, 1); + /* + * devmap page refcounts are 1-based, rather than 0-based: if + * refcount is 1, then the page is free and the refcount is + * stable because nobody holds a reference on the page. + */ + if (count == 1) + free_devmap_managed_page(page); + else if (!count) + __put_page(page); + } + + return is_devmap; +} +#else +static bool __unpin_devmap_managed_user_page(struct page *page) +{ + return false; +} +#endif /* CONFIG_DEV_PAGEMAP_OPS */ + +/** + * unpin_user_page() - release a dma-pinned page + * @page: pointer to page to be released + * + * Pages that were pinned via pin_user_pages*() must be released via either + * unpin_user_page(), or one of the unpin_user_pages*() routines. This is so + * that such pages can be separately tracked and uniquely handled. In + * particular, interactions with RDMA and filesystems need special handling. + */ +void unpin_user_page(struct page *page) +{ + page = compound_head(page); + + /* + * For devmap managed pages we need to catch refcount transition from + * GUP_PIN_COUNTING_BIAS to 1, when refcount reach one it means the + * page is free and we need to inform the device driver through + * callback. See include/linux/memremap.h and HMM for details. + */ + if (__unpin_devmap_managed_user_page(page)) + return; + + if (page_ref_sub_and_test(page, GUP_PIN_COUNTING_BIAS)) + __put_page(page); + + __update_proc_vmstat(page, NR_FOLL_PIN_RETURNED, 1); +} +EXPORT_SYMBOL(unpin_user_page); + /** * unpin_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages * @pages: array of pages to be maybe marked dirty, and definitely released. @@ -237,10 +390,11 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, }

page = vm_normal_page(vma, address, pte); - if (!page && pte_devmap(pte) && (flags & FOLL_GET)) { + if (!page && pte_devmap(pte) && (flags & (FOLL_GET | FOLL_PIN))) { /* - * Only return device mapping pages in the FOLL_GET case since - * they are only valid while holding the pgmap reference. + * Only return device mapping pages in the FOLL_GET or FOLL_PIN + * case since they are only valid while holding the pgmap + * reference. */ *pgmap = get_dev_pagemap(pte_pfn(pte), *pgmap); if (*pgmap) @@ -278,11 +432,10 @@ static struct page *follow_page_pte(struct vm_area_struct *vma, goto retry; }

- if (flags & FOLL_GET) { - if (unlikely(!try_get_page(page))) { - page = ERR_PTR(-ENOMEM); - goto out; - } + /* try_grab_page() does nothing unless FOLL_GET or FOLL_PIN is set. */ + if (unlikely(!try_grab_page(page, flags))) { + page = ERR_PTR(-ENOMEM); + goto out; } if (flags & FOLL_TOUCH) { if ((flags & FOLL_WRITE) && @@ -544,7 +697,7 @@ static struct page *follow_page_mask(struct vm_area_struct *vma, /* make this handle hugepd */ page = follow_huge_addr(mm, address, flags & FOLL_WRITE); if (!IS_ERR(page)) { - BUG_ON(flags & FOLL_GET); + WARN_ON_ONCE(flags & (FOLL_GET | FOLL_PIN)); return page; }

@@ -1131,6 +1284,36 @@ static __always_inline long __get_user_pages_locked(struct task_struct *tsk, return pages_done; }

+static long __get_user_pages_remote(struct task_struct *tsk, + struct mm_struct *mm, + unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas, int *locked) +{ + /* + * Parts of FOLL_LONGTERM behavior are incompatible with + * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on + * vmas. However, this only comes up if locked is set, and there are + * callers that do request FOLL_LONGTERM, but do not set locked. So, + * allow what we can. + */ + if (gup_flags & FOLL_LONGTERM) { + if (WARN_ON_ONCE(locked)) + return -EINVAL; + /* + * This will check the vmas (even if our vmas arg is NULL) + * and return -ENOTSUPP if DAX isn't allowed in this case: + */ + return __gup_longterm_locked(tsk, mm, start, nr_pages, pages, + vmas, gup_flags | FOLL_TOUCH | + FOLL_REMOTE); + } + + return __get_user_pages_locked(tsk, mm, start, nr_pages, pages, vmas, + locked, + gup_flags | FOLL_TOUCH | FOLL_REMOTE); +} + /* * get_user_pages_remote() - pin user pages in memory * @tsk: the task_struct to use for page fault accounting, or @@ -1205,28 +1388,8 @@ long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) return -EINVAL;

- /* - * Parts of FOLL_LONGTERM behavior are incompatible with - * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on - * vmas. However, this only comes up if locked is set, and there are - * callers that do request FOLL_LONGTERM, but do not set locked. So, - * allow what we can. - */ - if (gup_flags & FOLL_LONGTERM) { - if (WARN_ON_ONCE(locked)) - return -EINVAL; - /* - * This will check the vmas (even if our vmas arg is NULL) - * and return -ENOTSUPP if DAX isn't allowed in this case: - */ - return __gup_longterm_locked(tsk, mm, start, nr_pages, pages, - vmas, gup_flags | FOLL_TOUCH | - FOLL_REMOTE); - } - - return __get_user_pages_locked(tsk, mm, start, nr_pages, pages, vmas, - locked, - gup_flags | FOLL_TOUCH | FOLL_REMOTE); + return __get_user_pages_remote(tsk, mm, start, nr_pages, gup_flags, + pages, vmas, locked); } EXPORT_SYMBOL(get_user_pages_remote);

@@ -1421,10 +1584,11 @@ static long __get_user_pages_locked(struct task_struct *tsk, return i ? : -EFAULT; }

-long get_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, - unsigned long start, unsigned long nr_pages, - unsigned int gup_flags, struct page **pages, - struct vm_area_struct **vmas, int *locked) +static long __get_user_pages_remote(struct task_struct *tsk, + struct mm_struct *mm, + unsigned long start, unsigned long nr_pages, + unsigned int gup_flags, struct page **pages, + struct vm_area_struct **vmas, int *locked) { return 0; } @@ -1864,13 +2028,17 @@ static inline pte_t gup_get_pte(pte_t *ptep) #endif /* CONFIG_GUP_GET_PTE_LOW_HIGH */

static void __maybe_unused undo_dev_pagemap(int *nr, int nr_start, + unsigned int flags, struct page **pages) { while ((*nr) - nr_start) { struct page *page = pages[--(*nr)];

ClearPageReferenced(page); - put_page(page); + if (flags & FOLL_PIN) + unpin_user_page(page); + else + put_page(page); } }

@@ -1903,7 +2071,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,

pgmap = get_dev_pagemap(pte_pfn(pte), pgmap); if (unlikely(!pgmap)) { - undo_dev_pagemap(nr, nr_start, pages); + undo_dev_pagemap(nr, nr_start, flags, pages); goto pte_unmap; } } else if (pte_special(pte)) @@ -1912,7 +2080,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end, VM_BUG_ON(!pfn_valid(pte_pfn(pte))); page = pte_page(pte);

- head = try_get_compound_head(page, 1); + head = try_grab_compound_head(page, 1, flags); if (!head) goto pte_unmap;

@@ -1957,7 +2125,8 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,

#if defined(CONFIG_ARCH_HAS_PTE_DEVMAP) && defined(CONFIG_TRANSPARENT_HUGEPAGE) static int __gup_device_huge(unsigned long pfn, unsigned long addr, - unsigned long end, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { int nr_start = *nr; struct dev_pagemap *pgmap = NULL; @@ -1967,12 +2136,15 @@ static int __gup_device_huge(unsigned long pfn, unsigned long addr,

pgmap = get_dev_pagemap(pfn, pgmap); if (unlikely(!pgmap)) { - undo_dev_pagemap(nr, nr_start, pages); + undo_dev_pagemap(nr, nr_start, flags, pages); return 0; } SetPageReferenced(page); pages[*nr] = page; - get_page(page); + if (unlikely(!try_grab_page(page, flags))) { + undo_dev_pagemap(nr, nr_start, flags, pages); + return 0; + } (*nr)++; pfn++; } while (addr += PAGE_SIZE, addr != end); @@ -1983,48 +2155,52 @@ static int __gup_device_huge(unsigned long pfn, unsigned long addr, }

static int __gup_device_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, - unsigned long end, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { unsigned long fault_pfn; int nr_start = *nr;

fault_pfn = pmd_pfn(orig) + ((addr & ~PMD_MASK) >> PAGE_SHIFT); - if (!__gup_device_huge(fault_pfn, addr, end, pages, nr)) + if (!__gup_device_huge(fault_pfn, addr, end, flags, pages, nr)) return 0;

if (unlikely(pmd_val(orig) != pmd_val(*pmdp))) { - undo_dev_pagemap(nr, nr_start, pages); + undo_dev_pagemap(nr, nr_start, flags, pages); return 0; } return 1; }

static int __gup_device_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, - unsigned long end, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { unsigned long fault_pfn; int nr_start = *nr;

fault_pfn = pud_pfn(orig) + ((addr & ~PUD_MASK) >> PAGE_SHIFT); - if (!__gup_device_huge(fault_pfn, addr, end, pages, nr)) + if (!__gup_device_huge(fault_pfn, addr, end, flags, pages, nr)) return 0;

if (unlikely(pud_val(orig) != pud_val(*pudp))) { - undo_dev_pagemap(nr, nr_start, pages); + undo_dev_pagemap(nr, nr_start, flags, pages); return 0; } return 1; } #else static int __gup_device_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, - unsigned long end, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { BUILD_BUG(); return 0; }

static int __gup_device_huge_pud(pud_t pud, pud_t *pudp, unsigned long addr, - unsigned long end, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { BUILD_BUG(); return 0; @@ -2042,8 +2218,11 @@ static int record_subpages(struct page *page, unsigned long addr, return nr; }

-static void put_compound_head(struct page *page, int refs) +static void put_compound_head(struct page *page, int refs, unsigned int flags) { + if (flags & FOLL_PIN) + refs *= GUP_PIN_COUNTING_BIAS; + /* Do a get_page() first, in case refs == page->_refcount */ get_page(page); page_ref_sub(page, refs); @@ -2083,12 +2262,12 @@ static int gup_hugepte(pte_t *ptep, unsigned long sz, unsigned long addr, page = head + ((addr & (sz-1)) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr);

- head = try_get_compound_head(head, refs); + head = try_grab_compound_head(head, refs, flags); if (!head) return 0;

if (unlikely(pte_val(pte) != pte_val(*ptep))) { - put_compound_head(head, refs); + put_compound_head(head, refs, flags); return 0; }

@@ -2136,18 +2315,19 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, if (pmd_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; - return __gup_device_huge_pmd(orig, pmdp, addr, end, pages, nr); + return __gup_device_huge_pmd(orig, pmdp, addr, end, flags, + pages, nr); }

page = pmd_page(orig) + ((addr & ~PMD_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr);

- head = try_get_compound_head(pmd_page(orig), refs); + head = try_grab_compound_head(pmd_page(orig), refs, flags); if (!head) return 0;

if (unlikely(pmd_val(orig) != pmd_val(*pmdp))) { - put_compound_head(head, refs); + put_compound_head(head, refs, flags); return 0; }

@@ -2157,7 +2337,8 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, }

static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, - unsigned long end, unsigned int flags, struct page **pages, int *nr) + unsigned long end, unsigned int flags, + struct page **pages, int *nr) { struct page *head, *page; int refs; @@ -2168,18 +2349,19 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, if (pud_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; - return __gup_device_huge_pud(orig, pudp, addr, end, pages, nr); + return __gup_device_huge_pud(orig, pudp, addr, end, flags, + pages, nr); }

page = pud_page(orig) + ((addr & ~PUD_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr);

- head = try_get_compound_head(pud_page(orig), refs); + head = try_grab_compound_head(pud_page(orig), refs, flags); if (!head) return 0;

if (unlikely(pud_val(orig) != pud_val(*pudp))) { - put_compound_head(head, refs); + put_compound_head(head, refs, flags); return 0; }

@@ -2203,12 +2385,12 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr, page = pgd_page(orig) + ((addr & ~PGDIR_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr);

- head = try_get_compound_head(pgd_page(orig), refs); + head = try_grab_compound_head(pgd_page(orig), refs, flags); if (!head) return 0;

if (unlikely(pgd_val(orig) != pgd_val(*pgdp))) { - put_compound_head(head, refs); + put_compound_head(head, refs, flags); return 0; }

@@ -2509,9 +2691,12 @@ EXPORT_SYMBOL_GPL(get_user_pages_fast); /** * pin_user_pages_fast() - pin user pages in memory without taking locks * - * For now, this is a placeholder function, until various call sites are - * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, - * this is identical to get_user_pages_fast(). + * Nearly the same as get_user_pages_fast(), except that FOLL_PIN is set. See + * get_user_pages_fast() for documentation on the function arguments, because + * the arguments here are identical. + * + * FOLL_PIN means that the pages must be released via unpin_user_page(). Please + * see Documentation/vm/pin_user_pages.rst for further details. * * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It * is NOT intended for Case 2 (RDMA: long-term pins). @@ -2519,21 +2704,24 @@ EXPORT_SYMBOL_GPL(get_user_pages_fast); int pin_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { - /* - * This is a placeholder, until the pin functionality is activated. - * Until then, just behave like the corresponding get_user_pages*() - * routine. - */ - return get_user_pages_fast(start, nr_pages, gup_flags, pages); + /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE(gup_flags & FOLL_GET)) + return -EINVAL; + + gup_flags |= FOLL_PIN; + return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); } EXPORT_SYMBOL_GPL(pin_user_pages_fast);

/** * pin_user_pages_remote() - pin pages of a remote process (task != current) * - * For now, this is a placeholder function, until various call sites are - * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, - * this is identical to get_user_pages_remote(). + * Nearly the same as get_user_pages_remote(), except that FOLL_PIN is set. See + * get_user_pages_remote() for documentation on the function arguments, because + * the arguments here are identical. + * + * FOLL_PIN means that the pages must be released via unpin_user_page(). Please + * see Documentation/vm/pin_user_pages.rst for details. * * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It * is NOT intended for Case 2 (RDMA: long-term pins). @@ -2543,22 +2731,24 @@ long pin_user_pages_remote(struct task_struct *tsk, struct mm_struct *mm, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { - /* - * This is a placeholder, until the pin functionality is activated. - * Until then, just behave like the corresponding get_user_pages*() - * routine. - */ - return get_user_pages_remote(tsk, mm, start, nr_pages, gup_flags, pages, - vmas, locked); + /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE(gup_flags & FOLL_GET)) + return -EINVAL; + + gup_flags |= FOLL_PIN; + return __get_user_pages_remote(tsk, mm, start, nr_pages, gup_flags, + pages, vmas, locked); } EXPORT_SYMBOL(pin_user_pages_remote);

/** * pin_user_pages() - pin user pages in memory for use by other devices * - * For now, this is a placeholder function, until various call sites are - * converted to use the correct get_user_pages*() or pin_user_pages*() API. So, - * this is identical to get_user_pages(). + * Nearly the same as get_user_pages(), except that FOLL_TOUCH is not set, and + * FOLL_PIN is set. + * + * FOLL_PIN means that the pages must be released via unpin_user_page(). Please + * see Documentation/vm/pin_user_pages.rst for details. * * This is intended for Case 1 (DIO) in Documentation/vm/pin_user_pages.rst. It * is NOT intended for Case 2 (RDMA: long-term pins). @@ -2567,11 +2757,12 @@ long pin_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas) { - /* - * This is a placeholder, until the pin functionality is activated. - * Until then, just behave like the corresponding get_user_pages*() - * routine. - */ - return get_user_pages(start, nr_pages, gup_flags, pages, vmas); + /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE(gup_flags & FOLL_GET)) + return -EINVAL; + + gup_flags |= FOLL_PIN; + return __gup_longterm_locked(current, current->mm, start, nr_pages, + pages, vmas, gup_flags); } EXPORT_SYMBOL(pin_user_pages); diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 41a0fbddc96b..430e55673e35 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -945,6 +945,11 @@ struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr, */ WARN_ONCE(flags & FOLL_COW, "mm: In follow_devmap_pmd with FOLL_COW set");

+ /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) == + (FOLL_PIN | FOLL_GET))) + return NULL; + if (flags & FOLL_WRITE && !pmd_write(*pmd)) return NULL;

@@ -960,7 +965,7 @@ struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr, * device mapped pages can only be returned if the * caller will manage the page reference count. */ - if (!(flags & FOLL_GET)) + if (!(flags & (FOLL_GET | FOLL_PIN))) return ERR_PTR(-EEXIST);

pfn += (addr & ~PMD_MASK) >> PAGE_SHIFT; @@ -968,7 +973,8 @@ struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr, if (!*pgmap) return ERR_PTR(-EFAULT); page = pfn_to_page(pfn); - get_page(page); + if (!try_grab_page(page, flags)) + page = ERR_PTR(-EFAULT);

return page; } @@ -1088,6 +1094,11 @@ struct page *follow_devmap_pud(struct vm_area_struct *vma, unsigned long addr, if (flags & FOLL_WRITE && !pud_write(*pud)) return NULL;

+ /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) == + (FOLL_PIN | FOLL_GET))) + return NULL; + if (pud_present(*pud) && pud_devmap(*pud)) /* pass */; else @@ -1099,8 +1110,10 @@ struct page *follow_devmap_pud(struct vm_area_struct *vma, unsigned long addr, /* * device mapped pages can only be returned if the * caller will manage the page reference count. + * + * At least one of FOLL_GET | FOLL_PIN must be set, so assert that here: */ - if (!(flags & FOLL_GET)) + if (!(flags & (FOLL_GET | FOLL_PIN))) return ERR_PTR(-EEXIST);

pfn += (addr & ~PUD_MASK) >> PAGE_SHIFT; @@ -1108,7 +1121,8 @@ struct page *follow_devmap_pud(struct vm_area_struct *vma, unsigned long addr, if (!*pgmap) return ERR_PTR(-EFAULT); page = pfn_to_page(pfn); - get_page(page); + if (!try_grab_page(page, flags)) + page = ERR_PTR(-EFAULT);

return page; } @@ -1522,8 +1536,8 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma, skip_mlock: page += (addr & ~HPAGE_PMD_MASK) >> PAGE_SHIFT; VM_BUG_ON_PAGE(!PageCompound(page) && !is_zone_device_page(page), page); - if (flags & FOLL_GET) - get_page(page); + if (!try_grab_page(page, flags)) + page = ERR_PTR(-EFAULT);

out: return page; diff --git a/mm/hugetlb.c b/mm/hugetlb.c index ac65bb5e38ac..0aab6fe0072f 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -4356,7 +4356,13 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, same_page: if (pages) { pages[i] = mem_map_offset(page, pfn_offset); - get_page(pages[i]); + if (!try_grab_page(pages[i], flags)) { + spin_unlock(ptl); + remainder = 0; + err = -ENOMEM; + WARN_ON_ONCE(1); + break; + } }

if (vmas) @@ -4916,6 +4922,12 @@ follow_huge_pmd(struct mm_struct *mm, unsigned long address, struct page *page = NULL; spinlock_t *ptl; pte_t pte; + + /* FOLL_GET and FOLL_PIN are mutually exclusive. */ + if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) == + (FOLL_PIN | FOLL_GET))) + return NULL; + retry: ptl = pmd_lockptr(mm, pmd); spin_lock(ptl); @@ -4928,8 +4940,11 @@ follow_huge_pmd(struct mm_struct *mm, unsigned long address, pte = huge_ptep_get((pte_t *)pmd); if (pte_present(pte)) { page = pmd_page(*pmd) + ((address & ~PMD_MASK) >> PAGE_SHIFT); - if (flags & FOLL_GET) - get_page(page); + if (unlikely(!try_grab_page(page, flags))) { + WARN_ON_ONCE(1); + page = NULL; + goto out; + } } else { if (is_hugetlb_entry_migration(pte)) { spin_unlock(ptl); @@ -4950,7 +4965,7 @@ struct page * __weak follow_huge_pud(struct mm_struct *mm, unsigned long address, pud_t *pud, int flags) { - if (flags & FOLL_GET) + if (flags & (FOLL_GET | FOLL_PIN)) return NULL;

return pte_page(*(pte_t *)pud) + ((address & ~PUD_MASK) >> PAGE_SHIFT); @@ -4959,7 +4974,7 @@ follow_huge_pud(struct mm_struct *mm, unsigned long address, struct page * __weak follow_huge_pgd(struct mm_struct *mm, unsigned long address, pgd_t *pgd, int flags) { - if (flags & FOLL_GET) + if (flags & (FOLL_GET | FOLL_PIN)) return NULL;

return pte_page(*(pte_t *)pgd) + ((address & ~PGDIR_MASK) >> PAGE_SHIFT); diff --git a/mm/vmstat.c b/mm/vmstat.c index 78d53378db99..b56808bae1b4 100644 --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -1168,6 +1168,8 @@ const char * const vmstat_text[] = { "nr_dirtied", "nr_written", "nr_kernel_misc_reclaimable", + "nr_foll_pin_requested", + "nr_foll_pin_returned",

/* enum writeback_stat_item counters */ "nr_dirty_threshold",

Up until now, gup_benchmark supported testing of the following kernel functions:

* get_user_pages(): via the '-U' command line option * get_user_pages_longterm(): via the '-L' command line option * get_user_pages_fast(): as the default (no options required)

Add test coverage for the new corresponding pin_*() functions:

* pin_user_pages_fast(): via the '-a' command line option * pin_user_pages(): via the '-b' command line option

Also, add an option for clarity: '-u' for what is now (still) the default choice: get_user_pages_fast().

Also, for the commands that set FOLL_PIN, verify that the pages really are dma-pinned, via the new is_dma_pinned() routine. Those commands are:

PIN_FAST_BENCHMARK : calls pin_user_pages_fast() PIN_BENCHMARK : calls pin_user_pages()

In between the calls to pin_*() and unpin_user_pages(), check each page: if page_dma_pinned() returns false, then WARN and return.

Do this outside of the benchmark timestamps, so that it doesn't affect reported times.

Reviewed-by: Ira Weiny ira.weiny@intel.com Signed-off-by: John Hubbard jhubbard@nvidia.com --- mm/gup_benchmark.c | 65 ++++++++++++++++++++-- tools/testing/selftests/vm/gup_benchmark.c | 15 ++++- 2 files changed, 74 insertions(+), 6 deletions(-)

diff --git a/mm/gup_benchmark.c b/mm/gup_benchmark.c index 7fc44d25eca7..76d32db48af8 100644 --- a/mm/gup_benchmark.c +++ b/mm/gup_benchmark.c @@ -8,6 +8,8 @@ #define GUP_FAST_BENCHMARK _IOWR('g', 1, struct gup_benchmark) #define GUP_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark) #define GUP_BENCHMARK _IOWR('g', 3, struct gup_benchmark) +#define PIN_FAST_BENCHMARK _IOWR('g', 4, struct gup_benchmark) +#define PIN_BENCHMARK _IOWR('g', 5, struct gup_benchmark)

struct gup_benchmark { __u64 get_delta_usec; @@ -19,6 +21,42 @@ struct gup_benchmark { __u64 expansion[10]; /* For future use */ };

+static void put_back_pages(int cmd, struct page **pages, unsigned long nr_pages) +{ + int i; + + switch (cmd) { + case GUP_FAST_BENCHMARK: + case GUP_LONGTERM_BENCHMARK: + case GUP_BENCHMARK: + for (i = 0; i < nr_pages; i++) + put_page(pages[i]); + break; + + case PIN_FAST_BENCHMARK: + case PIN_BENCHMARK: + unpin_user_pages(pages, nr_pages); + break; + } +} + +static void verify_dma_pinned(int cmd, struct page **pages, + unsigned long nr_pages) +{ + int i; + + switch (cmd) { + case PIN_FAST_BENCHMARK: + case PIN_BENCHMARK: + for (i = 0; i < nr_pages; i++) { + if (WARN(!page_dma_pinned(pages[i]), + "pages[%d] is NOT dma-pinned\n", i)) + break; + } + break; + } +} + static int __gup_benchmark_ioctl(unsigned int cmd, struct gup_benchmark *gup) { @@ -65,6 +103,14 @@ static int __gup_benchmark_ioctl(unsigned int cmd, nr = get_user_pages(addr, nr, gup->flags, pages + i, NULL); break; + case PIN_FAST_BENCHMARK: + nr = pin_user_pages_fast(addr, nr, gup->flags, + pages + i); + break; + case PIN_BENCHMARK: + nr = pin_user_pages(addr, nr, gup->flags, pages + i, + NULL); + break; default: return -1; } @@ -75,15 +121,22 @@ static int __gup_benchmark_ioctl(unsigned int cmd, } end_time = ktime_get();

+ /* Shifting the meaning of nr_pages: now it is actual number pinned: */ + nr_pages = i; + gup->get_delta_usec = ktime_us_delta(end_time, start_time); gup->size = addr - gup->addr;

+ /* + * Take an un-benchmark-timed moment to verify DMA pinned + * state: print a warning if any non-dma-pinned pages are found: + */ + verify_dma_pinned(cmd, pages, nr_pages); + start_time = ktime_get(); - for (i = 0; i < nr_pages; i++) { - if (!pages[i]) - break; - put_page(pages[i]); - } + + put_back_pages(cmd, pages, nr_pages); + end_time = ktime_get(); gup->put_delta_usec = ktime_us_delta(end_time, start_time);

@@ -101,6 +154,8 @@ static long gup_benchmark_ioctl(struct file *filep, unsigned int cmd, case GUP_FAST_BENCHMARK: case GUP_LONGTERM_BENCHMARK: case GUP_BENCHMARK: + case PIN_FAST_BENCHMARK: + case PIN_BENCHMARK: break; default: return -EINVAL; diff --git a/tools/testing/selftests/vm/gup_benchmark.c b/tools/testing/selftests/vm/gup_benchmark.c index 389327e9b30a..43b4dfe161a2 100644 --- a/tools/testing/selftests/vm/gup_benchmark.c +++ b/tools/testing/selftests/vm/gup_benchmark.c @@ -18,6 +18,10 @@ #define GUP_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark) #define GUP_BENCHMARK _IOWR('g', 3, struct gup_benchmark)

+/* Similar to above, but use FOLL_PIN instead of FOLL_GET. */ +#define PIN_FAST_BENCHMARK _IOWR('g', 4, struct gup_benchmark) +#define PIN_BENCHMARK _IOWR('g', 5, struct gup_benchmark) + /* Just the flags we need, copied from mm.h: */ #define FOLL_WRITE 0x01 /* check pte is writable */

@@ -40,8 +44,14 @@ int main(int argc, char **argv) char *file = "/dev/zero"; char *p;

- while ((opt = getopt(argc, argv, "m:r:n:f:tTLUwSH")) != -1) { + while ((opt = getopt(argc, argv, "m:r:n:f:abtTLUuwSH")) != -1) { switch (opt) { + case 'a': + cmd = PIN_FAST_BENCHMARK; + break; + case 'b': + cmd = PIN_BENCHMARK; + break; case 'm': size = atoi(optarg) * MB; break; @@ -63,6 +73,9 @@ int main(int argc, char **argv) case 'U': cmd = GUP_BENCHMARK; break; + case 'u': + cmd = GUP_FAST_BENCHMARK; + break; case 'w': write = 1; break;