LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Dan Williams <dan.j.williams@intel.com>
To: Shiyang Ruan <ruansy.fnst@fujitsu.com>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	linux-xfs <linux-xfs@vger.kernel.org>,
	Linux NVDIMM <nvdimm@lists.linux.dev>,
	Linux MM <linux-mm@kvack.org>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>,
	device-mapper development <dm-devel@redhat.com>,
	"Darrick J. Wong" <djwong@kernel.org>,
	david <david@fromorbit.com>, Christoph Hellwig <hch@lst.de>,
	Alasdair Kergon <agk@redhat.com>,
	Mike Snitzer <snitzer@redhat.com>
Subject: Re: [PATCH RESEND v6 5/9] mm: Introduce mf_dax_kill_procs() for fsdax case
Date: Fri, 20 Aug 2021 15:40:49 -0700	[thread overview]
Message-ID: <CAPcyv4gRGryFeY8N-gFHWyNi2Dhyb2QXizX3nUFXLsw6h1Z6PA@mail.gmail.com> (raw)
In-Reply-To: <20210730100158.3117319-6-ruansy.fnst@fujitsu.com>

On Fri, Jul 30, 2021 at 3:02 AM Shiyang Ruan <ruansy.fnst@fujitsu.com> wrote:
>
> This function is called at the end of RMAP routine, i.e. filesystem
> recovery function.  The difference between mf_generic_kill_procs() is,
> mf_dax_kill_procs() accepts file mapping and offset instead of struct
> page.  It is because that different file mappings and offsets may share
> the same page in fsdax mode.  So, it is called when filesystem RMAP
> results are found.
>
> Signed-off-by: Shiyang Ruan <ruansy.fnst@fujitsu.com>
> ---
>  fs/dax.c            | 45 ++++++++++++++++++++++++-------
>  include/linux/dax.h | 16 ++++++++++++
>  include/linux/mm.h  | 10 +++++++
>  mm/memory-failure.c | 64 +++++++++++++++++++++++++++++++++------------
>  4 files changed, 109 insertions(+), 26 deletions(-)
>
> diff --git a/fs/dax.c b/fs/dax.c
> index da41f9363568..dce6307a12eb 100644
> --- a/fs/dax.c
> +++ b/fs/dax.c
> @@ -389,6 +389,41 @@ static struct page *dax_busy_page(void *entry)
>         return NULL;
>  }
>
> +/**
> + * dax_load_pfn - Load pfn of the DAX entry corresponding to a page
> + * @mapping:   The file whose entry we want to load
> + * @index:     offset where the DAX entry located in
> + *
> + * Return:     pfn number of the DAX entry
> + */
> +unsigned long dax_load_pfn(struct address_space *mapping, unsigned long index)
> +{
> +       XA_STATE(xas, &mapping->i_pages, index);
> +       void *entry;
> +       unsigned long pfn;
> +
> +       rcu_read_lock();
> +       for (;;) {
> +               xas_lock_irq(&xas);
> +               entry = xas_load(&xas);
> +               if (dax_is_locked(entry)) {
> +                       rcu_read_unlock();
> +                       wait_entry_unlocked(&xas, entry);
> +                       rcu_read_lock();
> +                       continue;
> +               }
> +
> +               if (dax_is_zero_entry(entry) || dax_is_empty_entry(entry))
> +                       pfn = 0;
> +               else
> +                       pfn = dax_to_pfn(entry);
> +               xas_unlock_irq(&xas);
> +               break;
> +       }
> +       rcu_read_unlock();
> +       return pfn;

Instead of this I think you want a version of dax_lock_page() that
takes a mapping and index. Otherwise I don't see how this function
protects against races to teardown mapping->host, or to invalidate the
association of the mapping to the pfn.

> +}
> +
>  /*
>   * dax_lock_mapping_entry - Lock the DAX entry corresponding to a page
>   * @page: The page whose entry we want to lock
> @@ -790,16 +825,6 @@ static void *dax_insert_entry(struct xa_state *xas,
>         return entry;
>  }
>
> -static inline
> -unsigned long pgoff_address(pgoff_t pgoff, struct vm_area_struct *vma)
> -{
> -       unsigned long address;
> -
> -       address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT);
> -       VM_BUG_ON_VMA(address < vma->vm_start || address >= vma->vm_end, vma);
> -       return address;
> -}
> -
>  /* Walk all mappings of a given index of a file and writeprotect them */
>  static void dax_entry_mkclean(struct address_space *mapping, pgoff_t index,
>                 unsigned long pfn)
> diff --git a/include/linux/dax.h b/include/linux/dax.h
> index 6f4b5c97ceb0..359e809516b8 100644
> --- a/include/linux/dax.h
> +++ b/include/linux/dax.h
> @@ -165,6 +165,7 @@ int dax_writeback_mapping_range(struct address_space *mapping,
>
>  struct page *dax_layout_busy_page(struct address_space *mapping);
>  struct page *dax_layout_busy_page_range(struct address_space *mapping, loff_t start, loff_t end);
> +unsigned long dax_load_pfn(struct address_space *mapping, unsigned long index);
>  dax_entry_t dax_lock_page(struct page *page);
>  void dax_unlock_page(struct page *page, dax_entry_t cookie);
>  #else
> @@ -206,6 +207,12 @@ static inline int dax_writeback_mapping_range(struct address_space *mapping,
>         return -EOPNOTSUPP;
>  }
>
> +static inline unsigned long dax_load_pfn(struct address_space *mapping,
> +               unsigned long index)
> +{
> +       return 0;
> +}
> +
>  static inline dax_entry_t dax_lock_page(struct page *page)
>  {
>         if (IS_DAX(page->mapping->host))
> @@ -259,6 +266,15 @@ static inline bool dax_mapping(struct address_space *mapping)
>  {
>         return mapping->host && IS_DAX(mapping->host);
>  }
> +static inline unsigned long pgoff_address(pgoff_t pgoff,
> +               struct vm_area_struct *vma)
> +{
> +       unsigned long address;
> +
> +       address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT);
> +       VM_BUG_ON_VMA(address < vma->vm_start || address >= vma->vm_end, vma);
> +       return address;
> +}
>
>  #ifdef CONFIG_DEV_DAX_HMEM_DEVICES
>  void hmem_register_device(int target_nid, struct resource *r);
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 7ca22e6e694a..530aaf7a6eb2 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -1190,6 +1190,14 @@ static inline bool is_device_private_page(const struct page *page)
>                 page->pgmap->type == MEMORY_DEVICE_PRIVATE;
>  }
>
> +static inline bool is_device_fsdax_page(const struct page *page)
> +{
> +       return IS_ENABLED(CONFIG_DEV_PAGEMAP_OPS) &&
> +               IS_ENABLED(CONFIG_FS_DAX) &&
> +               is_zone_device_page(page) &&
> +               page->pgmap->type == MEMORY_DEVICE_FS_DAX;
> +}

The value of this helper is unclear to me. The MEMORY_DEVICE_FS_DAX
indication is for communicating page-idle notifications to filesystem
code.

> +
>  static inline bool is_pci_p2pdma_page(const struct page *page)
>  {
>         return IS_ENABLED(CONFIG_DEV_PAGEMAP_OPS) &&
> @@ -3113,6 +3121,8 @@ enum mf_flags {
>         MF_MUST_KILL = 1 << 2,
>         MF_SOFT_OFFLINE = 1 << 3,
>  };
> +extern int mf_dax_kill_procs(struct address_space *mapping, pgoff_t index,
> +                            int flags);
>  extern int memory_failure(unsigned long pfn, int flags);
>  extern void memory_failure_queue(unsigned long pfn, int flags);
>  extern void memory_failure_queue_kick(int cpu);
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index ab3eda335acd..520664c405fc 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -134,6 +134,12 @@ static int hwpoison_filter_dev(struct page *p)
>         if (PageSlab(p))
>                 return -EINVAL;
>
> +       if (pfn_valid(page_to_pfn(p))) {
> +               if (is_device_fsdax_page(p))

hwpoison_filter was built to test triggering failures in hard to reach
places of the page cache. I think you can make an argument the
hwpoison_filter does not apply to DAX since DAX by definition
eliminates page cache. So, I'd consult pgmap->memory_failure()
*instead* of the hwpoison_filter, don't teach the filter to ignore
fsdax pages. By definition if the ->memory_failure() says -EOPNOTSUPP
then hwpoison_filter can be consulted per usual.

> +                       return 0;
> +       } else
> +               return -EINVAL;
> +
>         mapping = page_mapping(p);
>         if (mapping == NULL || mapping->host == NULL)
>                 return -EINVAL;
> @@ -304,10 +310,9 @@ void shake_page(struct page *p, int access)
>  }
>  EXPORT_SYMBOL_GPL(shake_page);
>
> -static unsigned long dev_pagemap_mapping_shift(struct page *page,
> +static unsigned long dev_pagemap_mapping_shift(unsigned long address,
>                 struct vm_area_struct *vma)
>  {
> -       unsigned long address = vma_address(page, vma);
>         pgd_t *pgd;
>         p4d_t *p4d;
>         pud_t *pud;
> @@ -347,7 +352,7 @@ static unsigned long dev_pagemap_mapping_shift(struct page *page,
>   * Schedule a process for later kill.
>   * Uses GFP_ATOMIC allocations to avoid potential recursions in the VM.
>   */
> -static void add_to_kill(struct task_struct *tsk, struct page *p,
> +static void add_to_kill(struct task_struct *tsk, struct page *p, pgoff_t pgoff,
>                        struct vm_area_struct *vma,
>                        struct list_head *to_kill)
>  {
> @@ -360,9 +365,14 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,
>         }
>
>         tk->addr = page_address_in_vma(p, vma);
> -       if (is_zone_device_page(p))
> -               tk->size_shift = dev_pagemap_mapping_shift(p, vma);
> -       else
> +       if (is_zone_device_page(p)) {
> +               /* Since page->mapping is no more used for fsdax, we should
> +                * calculate the address in a fsdax way.
> +                */
> +               if (is_device_fsdax_page(p))
> +                       tk->addr = pgoff_address(pgoff, vma);
> +               tk->size_shift = dev_pagemap_mapping_shift(tk->addr, vma);
> +       } else
>                 tk->size_shift = page_shift(compound_head(p));
>
>         /*
> @@ -510,7 +520,7 @@ static void collect_procs_anon(struct page *page, struct list_head *to_kill,
>                         if (!page_mapped_in_vma(page, vma))
>                                 continue;
>                         if (vma->vm_mm == t->mm)
> -                               add_to_kill(t, page, vma, to_kill);
> +                               add_to_kill(t, page, 0, vma, to_kill);
>                 }
>         }
>         read_unlock(&tasklist_lock);
> @@ -520,24 +530,20 @@ static void collect_procs_anon(struct page *page, struct list_head *to_kill,
>  /*
>   * Collect processes when the error hit a file mapped page.
>   */
> -static void collect_procs_file(struct page *page, struct list_head *to_kill,
> -                               int force_early)
> +static void collect_procs_file(struct page *page, struct address_space *mapping,
> +               pgoff_t pgoff, struct list_head *to_kill, int force_early)
>  {
>         struct vm_area_struct *vma;
>         struct task_struct *tsk;
> -       struct address_space *mapping = page->mapping;
> -       pgoff_t pgoff;
>
>         i_mmap_lock_read(mapping);
>         read_lock(&tasklist_lock);
> -       pgoff = page_to_pgoff(page);
>         for_each_process(tsk) {
>                 struct task_struct *t = task_early_kill(tsk, force_early);
>
>                 if (!t)
>                         continue;
> -               vma_interval_tree_foreach(vma, &mapping->i_mmap, pgoff,
> -                                     pgoff) {
> +               vma_interval_tree_foreach(vma, &mapping->i_mmap, pgoff, pgoff) {
>                         /*
>                          * Send early kill signal to tasks where a vma covers
>                          * the page but the corrupted page is not necessarily
> @@ -546,7 +552,7 @@ static void collect_procs_file(struct page *page, struct list_head *to_kill,
>                          * to be informed of all such data corruptions.
>                          */
>                         if (vma->vm_mm == t->mm)
> -                               add_to_kill(t, page, vma, to_kill);
> +                               add_to_kill(t, page, pgoff, vma, to_kill);
>                 }
>         }
>         read_unlock(&tasklist_lock);
> @@ -565,7 +571,8 @@ static void collect_procs(struct page *page, struct list_head *tokill,
>         if (PageAnon(page))
>                 collect_procs_anon(page, tokill, force_early);
>         else
> -               collect_procs_file(page, tokill, force_early);
> +               collect_procs_file(page, page->mapping, page->index, tokill,
> +                                  force_early);
>  }
>
>  struct hwp_walk {
> @@ -1477,6 +1484,31 @@ static int mf_generic_kill_procs(unsigned long long pfn, int flags)
>         return 0;
>  }
>
> +int mf_dax_kill_procs(struct address_space *mapping, pgoff_t index, int flags)
> +{
> +       LIST_HEAD(to_kill);
> +       /* load the pfn of the dax mapping file */
> +       unsigned long pfn = dax_load_pfn(mapping, index);
> +
> +       /* the failure pfn may not actually be mmapped, so no need to
> +        * unmap and kill procs */
> +       if (!pfn)

pfn-0 is a valid pfn. I think you should use a cookie value like
dax_load_page() to indicate failure.

> +               return 0;
> +
> +       /*
> +        * Unlike System-RAM there is no possibility to swap in a
> +        * different physical page at a given virtual address, so all
> +        * userspace consumption of ZONE_DEVICE memory necessitates
> +        * SIGBUS (i.e. MF_MUST_KILL)
> +        */
> +       flags |= MF_ACTION_REQUIRED | MF_MUST_KILL;
> +       collect_procs_file(pfn_to_page(pfn), mapping, index, &to_kill, true);
> +
> +       unmap_and_kill(&to_kill, pfn, mapping, index, flags);
> +       return 0;
> +}
> +EXPORT_SYMBOL_GPL(mf_dax_kill_procs);
> +
>  static int memory_failure_hugetlb(unsigned long pfn, int flags)
>  {
>         struct page *p = pfn_to_page(pfn);
> --
> 2.32.0
>
>
>

  parent reply	other threads:[~2021-08-20 22:41 UTC|newest]

Thread overview: 40+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-30 10:01 [PATCH RESEND v6 0/9] fsdax: introduce fs query to support reflink Shiyang Ruan
2021-07-30 10:01 ` [PATCH RESEND v6 1/9] pagemap: Introduce ->memory_failure() Shiyang Ruan
2021-08-06  1:17   ` Jane Chu
2021-08-16 17:20     ` Jane Chu
2021-08-17  1:44       ` ruansy.fnst
2021-08-18  5:43       ` Jane Chu
2021-08-18  6:08         ` Jane Chu
2021-08-18  7:52           ` ruansy.fnst
2021-08-18 17:10             ` Dan Williams
2021-08-23 13:21               ` hch
2021-08-18 15:52           ` Darrick J. Wong
2021-08-19  7:18           ` Jane Chu
2021-08-19  8:11             ` Jane Chu
2021-08-19  9:10               ` ruansy.fnst
2021-08-19 20:50                 ` Jane Chu
2021-08-20 16:07   ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 2/9] dax: Introduce holder for dax_device Shiyang Ruan
2021-08-06  1:02   ` Jane Chu
2021-08-17  1:45     ` ruansy.fnst
2021-08-20 16:06   ` Dan Williams
2021-08-20 20:19   ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 3/9] mm: factor helpers for memory_failure_dev_pagemap Shiyang Ruan
2021-08-06  1:00   ` Jane Chu
2021-08-20 16:54     ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 4/9] pmem,mm: Implement ->memory_failure in pmem driver Shiyang Ruan
2021-08-20 20:51   ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 5/9] mm: Introduce mf_dax_kill_procs() for fsdax case Shiyang Ruan
2021-08-06  0:59   ` Jane Chu
2021-08-20 22:40   ` Dan Williams [this message]
2021-07-30 10:01 ` [PATCH RESEND v6 6/9] xfs: Implement ->notify_failure() for XFS Shiyang Ruan
2021-08-06  0:50   ` Jane Chu
2021-08-20 22:56     ` Dan Williams
2021-08-20 22:59   ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 7/9] dm: Introduce ->rmap() to find bdev offset Shiyang Ruan
2021-08-20 23:46   ` Dan Williams
2021-07-30 10:01 ` [PATCH RESEND v6 8/9] md: Implement dax_holder_operations Shiyang Ruan
2021-08-06  0:48   ` Jane Chu
2021-08-17  1:59     ` ruansy.fnst
2021-07-30 10:01 ` [PATCH RESEND v6 9/9] fsdax: add exception for reflinked files Shiyang Ruan
2021-08-06  0:46   ` Jane Chu

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=CAPcyv4gRGryFeY8N-gFHWyNi2Dhyb2QXizX3nUFXLsw6h1Z6PA@mail.gmail.com \
    --to=dan.j.williams@intel.com \
    --cc=agk@redhat.com \
    --cc=david@fromorbit.com \
    --cc=djwong@kernel.org \
    --cc=dm-devel@redhat.com \
    --cc=hch@lst.de \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-xfs@vger.kernel.org \
    --cc=nvdimm@lists.linux.dev \
    --cc=ruansy.fnst@fujitsu.com \
    --cc=snitzer@redhat.com \
    --subject='Re: [PATCH RESEND v6 5/9] mm: Introduce mf_dax_kill_procs() for fsdax case' \
    /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

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).