From: John Hubbard <jhubbard@nvidia.com>
To: <ira.weiny@intel.com>, Andrew Morton <akpm@linux-foundation.org>
Cc: Jason Gunthorpe <jgg@ziepe.ca>,
Dan Williams <dan.j.williams@intel.com>,
Matthew Wilcox <willy@infradead.org>, Jan Kara <jack@suse.cz>,
Theodore Ts'o <tytso@mit.edu>, Michal Hocko <mhocko@suse.com>,
Dave Chinner <david@fromorbit.com>, <linux-xfs@vger.kernel.org>,
<linux-rdma@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<linux-fsdevel@vger.kernel.org>, <linux-nvdimm@lists.01.org>,
<linux-ext4@vger.kernel.org>, <linux-mm@kvack.org>
Subject: Re: [RFC PATCH v2 12/19] mm/gup: Prep put_user_pages() to take an vaddr_pin struct
Date: Fri, 9 Aug 2019 17:30:00 -0700 [thread overview]
Message-ID: <12b6a576-7a64-102c-f4d7-7a4ad34df710@nvidia.com> (raw)
In-Reply-To: <20190809225833.6657-13-ira.weiny@intel.com>
On 8/9/19 3:58 PM, ira.weiny@intel.com wrote:
> From: Ira Weiny <ira.weiny@intel.com>
>
> Once callers start to use vaddr_pin the put_user_pages calls will need
> to have access to this data coming in. Prep put_user_pages() for this
> data.
>
> Signed-off-by: Ira Weiny <ira.weiny@intel.com>
> ---
> include/linux/mm.h | 20 +-------
> mm/gup.c | 122 ++++++++++++++++++++++++++++++++-------------
> 2 files changed, 88 insertions(+), 54 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index befe150d17be..9d37cafbef9a 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1064,25 +1064,7 @@ static inline void put_page(struct page *page)
> __put_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.
> - *
> - * 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.
> - */
> -static inline void put_user_page(struct page *page)
> -{
> - put_page(page);
> -}
> +void put_user_page(struct page *page);
>
> void put_user_pages_dirty_lock(struct page **pages, unsigned long npages,
> bool make_dirty);
> diff --git a/mm/gup.c b/mm/gup.c
> index a7a9d2f5278c..10cfd30ff668 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -24,30 +24,41 @@
>
> #include "internal.h"
>
> -/**
> - * put_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages
> - * @pages: array of pages to be maybe marked dirty, and definitely released.
A couple comments from our circular review chain: some fellow with the same
last name as you, recommended wording it like this:
@pages: array of pages to be put
> - * @npages: number of pages in the @pages array.
> - * @make_dirty: whether to mark the pages dirty
> - *
> - * "gup-pinned page" refers to a page that has had one of the get_user_pages()
> - * variants called on that page.
> - *
> - * For each page in the @pages array, make that page (or its head page, if a
> - * compound page) dirty, if @make_dirty is true, and if the page was previously
> - * listed as clean. In any case, releases all pages using put_user_page(),
> - * possibly via put_user_pages(), for the non-dirty case.
> - *
> - * Please see the put_user_page() documentation for details.
> - *
> - * set_page_dirty_lock() is used internally. If instead, set_page_dirty() is
> - * required, then the caller should a) verify that this is really correct,
> - * because _lock() is usually required, and b) hand code it:
> - * set_page_dirty_lock(), put_user_page().
> - *
> - */
> -void put_user_pages_dirty_lock(struct page **pages, unsigned long npages,
> - bool make_dirty)
> +static void __put_user_page(struct vaddr_pin *vaddr_pin, 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 (put_devmap_managed_page(page))
> + return;
> +
> + if (put_page_testzero(page))
> + __put_page(page);
> +}
> +
> +static void __put_user_pages(struct vaddr_pin *vaddr_pin, struct page **pages,
> + unsigned long npages)
> +{
> + unsigned long index;
> +
> + /*
> + * TODO: this can be optimized for huge pages: if a series of pages is
> + * physically contiguous and part of the same compound page, then a
> + * single operation to the head page should suffice.
> + */
As discussed in the other review thread (""), let's just delete that comment,
as long as you're moving things around.
> + for (index = 0; index < npages; index++)
> + __put_user_page(vaddr_pin, pages[index]);
> +}
> +
> +static void __put_user_pages_dirty_lock(struct vaddr_pin *vaddr_pin,
> + struct page **pages,
> + unsigned long npages,
> + bool make_dirty)
Elsewhere in this series, we pass vaddr_pin at the end of the arg list.
Here we pass it at the beginning, and it caused a minor jar when reading it.
Obviously just bike shedding at this point, though. Either way. :)
> {
> unsigned long index;
>
> @@ -58,7 +69,7 @@ void put_user_pages_dirty_lock(struct page **pages, unsigned long npages,
> */
>
> if (!make_dirty) {
> - put_user_pages(pages, npages);
> + __put_user_pages(vaddr_pin, pages, npages);
> return;
> }
>
> @@ -86,9 +97,58 @@ void put_user_pages_dirty_lock(struct page **pages, unsigned long npages,
> */
> if (!PageDirty(page))
> set_page_dirty_lock(page);
> - put_user_page(page);
> + __put_user_page(vaddr_pin, 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.
> + *
> + * 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.
> + */
> +void put_user_page(struct page *page)
> +{
> + __put_user_page(NULL, page);
> +}
> +EXPORT_SYMBOL(put_user_page);
> +
> +/**
> + * put_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages
> + * @pages: array of pages to be maybe marked dirty, and definitely released.
Same here:
@pages: array of pages to be put
> + * @npages: number of pages in the @pages array.
> + * @make_dirty: whether to mark the pages dirty
> + *
> + * "gup-pinned page" refers to a page that has had one of the get_user_pages()
> + * variants called on that page.
> + *
> + * For each page in the @pages array, make that page (or its head page, if a
> + * compound page) dirty, if @make_dirty is true, and if the page was previously
> + * listed as clean. In any case, releases all pages using put_user_page(),
> + * possibly via put_user_pages(), for the non-dirty case.
> + *
> + * Please see the put_user_page() documentation for details.
> + *
> + * set_page_dirty_lock() is used internally. If instead, set_page_dirty() is
> + * required, then the caller should a) verify that this is really correct,
> + * because _lock() is usually required, and b) hand code it:
> + * set_page_dirty_lock(), put_user_page().
> + *
> + */
> +void put_user_pages_dirty_lock(struct page **pages, unsigned long npages,
> + bool make_dirty)
> +{
> + __put_user_pages_dirty_lock(NULL, pages, npages, make_dirty);
> +}
> EXPORT_SYMBOL(put_user_pages_dirty_lock);
>
> /**
> @@ -102,15 +162,7 @@ EXPORT_SYMBOL(put_user_pages_dirty_lock);
> */
> void put_user_pages(struct page **pages, unsigned long npages)
> {
> - unsigned long index;
> -
> - /*
> - * TODO: this can be optimized for huge pages: if a series of pages is
> - * physically contiguous and part of the same compound page, then a
> - * single operation to the head page should suffice.
> - */
> - for (index = 0; index < npages; index++)
> - put_user_page(pages[index]);
> + __put_user_pages(NULL, pages, npages);
> }
> EXPORT_SYMBOL(put_user_pages);
>
>
This all looks pretty good, so regardless of the outcome of the minor
points above,
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
thanks,
--
John Hubbard
NVIDIA
next prev parent reply other threads:[~2019-08-10 0:30 UTC|newest]
Thread overview: 110+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-09 22:58 [RFC PATCH v2 00/19] RDMA/FS DAX truncate proposal V1,000,002 ;-) ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 01/19] fs/locks: Export F_LAYOUT lease to user space ira.weiny
2019-08-09 23:52 ` Dave Chinner
2019-08-12 17:36 ` Ira Weiny
2019-08-14 8:05 ` Dave Chinner
2019-08-14 11:21 ` Jeff Layton
2019-08-14 11:38 ` Dave Chinner
2019-08-09 22:58 ` [RFC PATCH v2 02/19] fs/locks: Add Exclusive flag to user Layout lease ira.weiny
2019-08-14 14:15 ` Jeff Layton
2019-08-14 21:56 ` Dave Chinner
2019-08-26 10:41 ` Jeff Layton
2019-08-29 23:34 ` Ira Weiny
2019-09-04 12:52 ` Jeff Layton
2019-09-04 23:12 ` John Hubbard
2019-08-09 22:58 ` [RFC PATCH v2 03/19] mm/gup: Pass flags down to __gup_device_huge* calls ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 04/19] mm/gup: Ensure F_LAYOUT lease is held prior to GUP'ing pages ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 05/19] fs/ext4: Teach ext4 to break layout leases ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 06/19] fs/ext4: Teach dax_layout_busy_page() to operate on a sub-range ira.weiny
2019-08-23 15:18 ` Vivek Goyal
2019-08-29 18:52 ` Ira Weiny
2019-08-09 22:58 ` [RFC PATCH v2 07/19] fs/xfs: Teach xfs to use new dax_layout_busy_page() ira.weiny
2019-08-09 23:30 ` Dave Chinner
2019-08-12 18:05 ` Ira Weiny
2019-08-14 8:04 ` Dave Chinner
2019-08-09 22:58 ` [RFC PATCH v2 08/19] fs/xfs: Fail truncate if page lease can't be broken ira.weiny
2019-08-09 23:22 ` Dave Chinner
2019-08-12 18:08 ` Ira Weiny
2019-08-09 22:58 ` [RFC PATCH v2 09/19] mm/gup: Introduce vaddr_pin structure ira.weiny
2019-08-10 0:06 ` John Hubbard
2019-08-09 22:58 ` [RFC PATCH v2 10/19] mm/gup: Pass a NULL vaddr_pin through GUP fast ira.weiny
2019-08-10 0:06 ` John Hubbard
2019-08-09 22:58 ` [RFC PATCH v2 11/19] mm/gup: Pass follow_page_context further down the call stack ira.weiny
2019-08-10 0:18 ` John Hubbard
2019-08-12 19:01 ` Ira Weiny
2019-08-09 22:58 ` [RFC PATCH v2 12/19] mm/gup: Prep put_user_pages() to take an vaddr_pin struct ira.weiny
2019-08-10 0:30 ` John Hubbard [this message]
2019-08-12 20:46 ` Ira Weiny
2019-08-09 22:58 ` [RFC PATCH v2 13/19] {mm,file}: Add file_pins objects ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 14/19] fs/locks: Associate file pins while performing GUP ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 15/19] mm/gup: Introduce vaddr_pin_pages() ira.weiny
2019-08-10 0:09 ` John Hubbard
2019-08-12 21:00 ` Ira Weiny
2019-08-12 21:20 ` John Hubbard
2019-08-11 23:07 ` John Hubbard
2019-08-12 21:01 ` Ira Weiny
2019-08-12 12:28 ` Jason Gunthorpe
2019-08-12 21:48 ` Ira Weiny
2019-08-13 11:47 ` Jason Gunthorpe
2019-08-13 17:46 ` Ira Weiny
2019-08-13 17:56 ` John Hubbard
2019-08-09 22:58 ` [RFC PATCH v2 16/19] RDMA/uverbs: Add back pointer to system file object ira.weiny
2019-08-12 13:00 ` Jason Gunthorpe
2019-08-12 17:28 ` Ira Weiny
2019-08-12 17:56 ` Jason Gunthorpe
2019-08-12 21:15 ` Ira Weiny
2019-08-13 11:48 ` Jason Gunthorpe
2019-08-13 17:41 ` Ira Weiny
2019-08-13 18:00 ` Jason Gunthorpe
2019-08-13 20:38 ` Ira Weiny
2019-08-14 12:23 ` Jason Gunthorpe
2019-08-14 17:50 ` Ira Weiny
2019-08-14 18:15 ` Jason Gunthorpe
2019-09-04 22:25 ` Ira Weiny
2019-09-11 8:19 ` Jason Gunthorpe
2019-08-09 22:58 ` [RFC PATCH v2 17/19] RDMA/umem: Convert to vaddr_[pin|unpin]* operations ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 18/19] {mm,procfs}: Add display file_pins proc ira.weiny
2019-08-09 22:58 ` [RFC PATCH v2 19/19] mm/gup: Remove FOLL_LONGTERM DAX exclusion ira.weiny
2019-08-14 10:17 ` [RFC PATCH v2 00/19] RDMA/FS DAX truncate proposal V1,000,002 ;-) Jan Kara
2019-08-14 18:08 ` Ira Weiny
2019-08-15 13:05 ` Jan Kara
2019-08-16 19:05 ` Ira Weiny
2019-08-16 23:20 ` [RFC PATCH v2 00/19] RDMA/FS DAX truncate proposal V1,000,002 ; -) Ira Weiny
2019-08-19 6:36 ` Jan Kara
2019-08-17 2:26 ` [RFC PATCH v2 00/19] RDMA/FS DAX truncate proposal V1,000,002 ;-) Dave Chinner
2019-08-19 6:34 ` Jan Kara
2019-08-19 9:24 ` Dave Chinner
2019-08-19 12:38 ` Jason Gunthorpe
2019-08-19 21:53 ` Ira Weiny
2019-08-20 1:12 ` Dave Chinner
2019-08-20 11:55 ` Jason Gunthorpe
2019-08-21 18:02 ` Ira Weiny
2019-08-21 18:13 ` Jason Gunthorpe
2019-08-21 18:22 ` John Hubbard
2019-08-21 18:57 ` Ira Weiny
2019-08-21 19:06 ` Ira Weiny
2019-08-21 19:48 ` Jason Gunthorpe
2019-08-21 20:44 ` Ira Weiny
2019-08-21 23:49 ` Jason Gunthorpe
2019-08-23 3:23 ` Dave Chinner
2019-08-23 12:04 ` Jason Gunthorpe
2019-08-24 0:11 ` Dave Chinner
2019-08-24 5:08 ` Ira Weiny
2019-08-26 5:55 ` Dave Chinner
2019-08-29 2:02 ` Ira Weiny
2019-08-29 3:27 ` John Hubbard
2019-08-29 16:16 ` Ira Weiny
2019-09-02 22:26 ` Dave Chinner
2019-09-04 16:54 ` Ira Weiny
2019-08-25 19:39 ` Jason Gunthorpe
2019-08-24 4:49 ` Ira Weiny
2019-08-25 19:40 ` Jason Gunthorpe
2019-08-23 0:59 ` Dave Chinner
2019-08-23 17:15 ` Ira Weiny
2019-08-24 0:18 ` Dave Chinner
2019-08-20 0:05 ` John Hubbard
2019-08-20 1:20 ` Dave Chinner
2019-08-20 3:09 ` John Hubbard
2019-08-20 3:36 ` Dave Chinner
2019-08-21 18:43 ` John Hubbard
2019-08-21 19:09 ` Ira Weiny
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=12b6a576-7a64-102c-f4d7-7a4ad34df710@nvidia.com \
--to=jhubbard@nvidia.com \
--cc=akpm@linux-foundation.org \
--cc=dan.j.williams@intel.com \
--cc=david@fromorbit.com \
--cc=ira.weiny@intel.com \
--cc=jack@suse.cz \
--cc=jgg@ziepe.ca \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-nvdimm@lists.01.org \
--cc=linux-rdma@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
--cc=mhocko@suse.com \
--cc=tytso@mit.edu \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).