diff mbox

[v5,08/11] mm, memory_failure: Teach memory_failure() about dev_pagemap pages

Message ID 153074046610.27838.329669845580014251.stgit@dwillia2-desk3.amr.corp.intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Dan Williams July 4, 2018, 9:41 p.m. UTC
mce: Uncorrected hardware memory error in user-access at af34214200
    {1}[Hardware Error]: It has been corrected by h/w and requires no further action
    mce: [Hardware Error]: Machine check events logged
    {1}[Hardware Error]: event severity: corrected
    Memory failure: 0xaf34214: reserved kernel page still referenced by 1 users
    [..]
    Memory failure: 0xaf34214: recovery action for reserved kernel page: Failed
    mce: Memory error not recovered

In contrast to typical memory, dev_pagemap pages may be dax mapped. With
dax there is no possibility to map in another page dynamically since dax
establishes 1:1 physical address to file offset associations. Also
dev_pagemap pages associated with NVDIMM / persistent memory devices can
internal remap/repair addresses with poison. While memory_failure()
assumes that it can discard typical poisoned pages and keep them
unmapped indefinitely, dev_pagemap pages may be returned to service
after the error is cleared.

Teach memory_failure() to detect and handle MEMORY_DEVICE_HOST
dev_pagemap pages that have poison consumed by userspace. Mark the
memory as UC instead of unmapping it completely to allow ongoing access
via the device driver (nd_pmem). Later, nd_pmem will grow support for
marking the page back to WB when the error is cleared.

Cc: Jan Kara <jack@suse.cz>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 include/linux/mm.h  |    1 
 mm/memory-failure.c |  124 ++++++++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 123 insertions(+), 2 deletions(-)

Comments

Naoya Horiguchi July 13, 2018, 8:52 a.m. UTC | #1
On Wed, Jul 04, 2018 at 02:41:06PM -0700, Dan Williams wrote:
>     mce: Uncorrected hardware memory error in user-access at af34214200

>     {1}[Hardware Error]: It has been corrected by h/w and requires no further action

>     mce: [Hardware Error]: Machine check events logged

>     {1}[Hardware Error]: event severity: corrected

>     Memory failure: 0xaf34214: reserved kernel page still referenced by 1 users

>     [..]

>     Memory failure: 0xaf34214: recovery action for reserved kernel page: Failed

>     mce: Memory error not recovered

>

> In contrast to typical memory, dev_pagemap pages may be dax mapped. With

> dax there is no possibility to map in another page dynamically since dax

> establishes 1:1 physical address to file offset associations. Also

> dev_pagemap pages associated with NVDIMM / persistent memory devices can

> internal remap/repair addresses with poison. While memory_failure()

> assumes that it can discard typical poisoned pages and keep them

> unmapped indefinitely, dev_pagemap pages may be returned to service

> after the error is cleared.

>

> Teach memory_failure() to detect and handle MEMORY_DEVICE_HOST

> dev_pagemap pages that have poison consumed by userspace. Mark the

> memory as UC instead of unmapping it completely to allow ongoing access

> via the device driver (nd_pmem). Later, nd_pmem will grow support for

> marking the page back to WB when the error is cleared.


By the way, what happens if madvise(MADV_SOFT_OFFLINE) is called on
a dev_pagemap page?  I'm not sure that cmci can be triggered on the
nvdimm device, but this injection interface is open for such a case.
Maybe simply ignoring the event is an expected behavior?

A few comments/quetions below ...

>

> Cc: Jan Kara <jack@suse.cz>

> Cc: Christoph Hellwig <hch@lst.de>

> Cc: Jérôme Glisse <jglisse@redhat.com>

> Cc: Matthew Wilcox <mawilcox@microsoft.com>

> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>

> Cc: Ross Zwisler <ross.zwisler@linux.intel.com>

> Signed-off-by: Dan Williams <dan.j.williams@intel.com>

> ---

>  include/linux/mm.h  |    1

>  mm/memory-failure.c |  124 ++++++++++++++++++++++++++++++++++++++++++++++++++-

>  2 files changed, 123 insertions(+), 2 deletions(-)

>

> diff --git a/include/linux/mm.h b/include/linux/mm.h

> index a0fbb9ffe380..374e5e9284f7 100644

> --- a/include/linux/mm.h

> +++ b/include/linux/mm.h

> @@ -2725,6 +2725,7 @@ enum mf_action_page_type {

>  	MF_MSG_TRUNCATED_LRU,

>  	MF_MSG_BUDDY,

>  	MF_MSG_BUDDY_2ND,

> +	MF_MSG_DAX,

>  	MF_MSG_UNKNOWN,

>  };

>

> diff --git a/mm/memory-failure.c b/mm/memory-failure.c

> index 4d70753af59c..161aa1b70212 100644

> --- a/mm/memory-failure.c

> +++ b/mm/memory-failure.c

> @@ -55,6 +55,7 @@

>  #include <linux/hugetlb.h>

>  #include <linux/memory_hotplug.h>

>  #include <linux/mm_inline.h>

> +#include <linux/memremap.h>

>  #include <linux/kfifo.h>

>  #include <linux/ratelimit.h>

>  #include "internal.h"

> @@ -263,6 +264,39 @@ void shake_page(struct page *p, int access)

>  }

>  EXPORT_SYMBOL_GPL(shake_page);

>

> +static unsigned long mapping_size(struct page *page, struct vm_area_struct *vma)

> +{

> +	unsigned long address = vma_address(page, vma);

> +	pgd_t *pgd;

> +	p4d_t *p4d;

> +	pud_t *pud;

> +	pmd_t *pmd;

> +	pte_t *pte;

> +

> +	pgd = pgd_offset(vma->vm_mm, address);

> +	if (!pgd_present(*pgd))

> +		return 0;

> +	p4d = p4d_offset(pgd, address);

> +	if (!p4d_present(*p4d))

> +		return 0;

> +	pud = pud_offset(p4d, address);

> +	if (!pud_present(*pud))

> +		return 0;

> +	if (pud_devmap(*pud))

> +		return PUD_SIZE;

> +	pmd = pmd_offset(pud, address);

> +	if (!pmd_present(*pmd))

> +		return 0;

> +	if (pmd_devmap(*pmd))

> +		return PMD_SIZE;

> +	pte = pte_offset_map(pmd, address);

> +	if (!pte_present(*pte))

> +		return 0;

> +	if (pte_devmap(*pte))

> +		return PAGE_SIZE;

> +	return 0;

> +}

> +


The function name looks generic, but this function seems to focus on
devmap thing, so could you include the word 'devmap' in the name?

>  /*

>   * Failure handling: if we can't find or can't kill a process there's

>   * not much we can do.	We just print a message and ignore otherwise.

> @@ -292,7 +326,10 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,

>  	}

>  	tk->addr = page_address_in_vma(p, vma);

>  	tk->addr_valid = 1;

> -	tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;

> +	if (is_zone_device_page(p))

> +		tk->size_shift = ilog2(mapping_size(p, vma));

> +	else

> +		tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;

>

>  	/*

>  	 * In theory we don't have to kill when the page was

> @@ -300,7 +337,7 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,

>  	 * likely very rare kill anyways just out of paranoia, but use

>  	 * a SIGKILL because the error is not contained anymore.

>  	 */

> -	if (tk->addr == -EFAULT) {

> +	if (tk->addr == -EFAULT || tk->size_shift == 0) {

>  		pr_info("Memory failure: Unable to find user space address %lx in %s\n",

>  			page_to_pfn(p), tsk->comm);

>  		tk->addr_valid = 0;

> @@ -514,6 +551,7 @@ static const char * const action_page_types[] = {

>  	[MF_MSG_TRUNCATED_LRU]		= "already truncated LRU page",

>  	[MF_MSG_BUDDY]			= "free buddy page",

>  	[MF_MSG_BUDDY_2ND]		= "free buddy page (2nd try)",

> +	[MF_MSG_DAX]			= "dax page",

>  	[MF_MSG_UNKNOWN]		= "unknown page",

>  };

>

> @@ -1111,6 +1149,83 @@ static int memory_failure_hugetlb(unsigned long pfn, int flags)

>  	return res;

>  }

>

> +static int memory_failure_dev_pagemap(unsigned long pfn, int flags,

> +		struct dev_pagemap *pgmap)

> +{

> +	struct page *page = pfn_to_page(pfn);

> +	const bool unmap_success = true;

> +	unsigned long size = 0;

> +	struct to_kill *tk;

> +	LIST_HEAD(tokill);

> +	int rc = -EBUSY;

> +	loff_t start;

> +

> +	/*

> +	 * Prevent the inode from being freed while we are interrogating

> +	 * the address_space, typically this would be handled by

> +	 * lock_page(), but dax pages do not use the page lock. This

> +	 * also prevents changes to the mapping of this pfn until

> +	 * poison signaling is complete.

> +	 */

> +	if (!dax_lock_mapping_entry(page))

> +		goto out;

> +

> +	if (hwpoison_filter(page)) {

> +		rc = 0;

> +		goto unlock;

> +	}

> +

> +	switch (pgmap->type) {

> +	case MEMORY_DEVICE_PRIVATE:

> +	case MEMORY_DEVICE_PUBLIC:

> +		/*

> +		 * TODO: Handle HMM pages which may need coordination

> +		 * with device-side memory.

> +		 */

> +		goto unlock;

> +	default:

> +		break;

> +	}

> +

> +	/*

> +	 * Use this flag as an indication that the dax page has been

> +	 * remapped UC to prevent speculative consumption of poison.

> +	 */

> +	SetPageHWPoison(page);


The number of hwpoison pages is maintained by num_poisoned_pages,
so you can call num_poisoned_pages_inc()?

Related to this, I'm interested in whether/how unpoison_page() works
on a hwpoisoned dev_pagemap page.

Thanks,
Naoya Horiguchi

> +

> +	/*

> +	 * 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(page, &tokill, flags & MF_ACTION_REQUIRED);

> +

> +	list_for_each_entry(tk, &tokill, nd)

> +		if (tk->size_shift)

> +			size = max(size, 1UL << tk->size_shift);

> +	if (size) {

> +		/*

> +		 * Unmap the largest mapping to avoid breaking up

> +		 * device-dax mappings which are constant size. The

> +		 * actual size of the mapping being torn down is

> +		 * communicated in siginfo, see kill_proc()

> +		 */

> +		start = (page->index << PAGE_SHIFT) & ~(size - 1);

> +		unmap_mapping_range(page->mapping, start, start + size, 0);

> +	}

> +	kill_procs(&tokill, flags & MF_MUST_KILL, !unmap_success, pfn, flags);

> +	rc = 0;

> +unlock:

> +	dax_unlock_mapping_entry(page);

> +out:

> +	/* drop pgmap ref acquired in caller */

> +	put_dev_pagemap(pgmap);

> +	action_result(pfn, MF_MSG_DAX, rc ? MF_FAILED : MF_RECOVERED);

> +	return rc;

> +}

> +

>  /**

>   * memory_failure - Handle memory failure of a page.

>   * @pfn: Page Number of the corrupted page

> @@ -1133,6 +1248,7 @@ int memory_failure(unsigned long pfn, int flags)

>  	struct page *p;

>  	struct page *hpage;

>  	struct page *orig_head;

> +	struct dev_pagemap *pgmap;

>  	int res;

>  	unsigned long page_flags;

>

> @@ -1145,6 +1261,10 @@ int memory_failure(unsigned long pfn, int flags)

>  		return -ENXIO;

>  	}

>

> +	pgmap = get_dev_pagemap(pfn, NULL);

> +	if (pgmap)

> +		return memory_failure_dev_pagemap(pfn, flags, pgmap);

> +

>  	p = pfn_to_page(pfn);

>  	if (PageHuge(p))

>  		return memory_failure_hugetlb(pfn, flags);

>
Dan Williams July 14, 2018, 12:28 a.m. UTC | #2
On Fri, Jul 13, 2018 at 1:52 AM, Naoya Horiguchi
<n-horiguchi@ah.jp.nec.com> wrote:
> On Wed, Jul 04, 2018 at 02:41:06PM -0700, Dan Williams wrote:
>>     mce: Uncorrected hardware memory error in user-access at af34214200
>>     {1}[Hardware Error]: It has been corrected by h/w and requires no further action
>>     mce: [Hardware Error]: Machine check events logged
>>     {1}[Hardware Error]: event severity: corrected
>>     Memory failure: 0xaf34214: reserved kernel page still referenced by 1 users
>>     [..]
>>     Memory failure: 0xaf34214: recovery action for reserved kernel page: Failed
>>     mce: Memory error not recovered
>>
>> In contrast to typical memory, dev_pagemap pages may be dax mapped. With
>> dax there is no possibility to map in another page dynamically since dax
>> establishes 1:1 physical address to file offset associations. Also
>> dev_pagemap pages associated with NVDIMM / persistent memory devices can
>> internal remap/repair addresses with poison. While memory_failure()
>> assumes that it can discard typical poisoned pages and keep them
>> unmapped indefinitely, dev_pagemap pages may be returned to service
>> after the error is cleared.
>>
>> Teach memory_failure() to detect and handle MEMORY_DEVICE_HOST
>> dev_pagemap pages that have poison consumed by userspace. Mark the
>> memory as UC instead of unmapping it completely to allow ongoing access
>> via the device driver (nd_pmem). Later, nd_pmem will grow support for
>> marking the page back to WB when the error is cleared.
>
> By the way, what happens if madvise(MADV_SOFT_OFFLINE) is called on
> a dev_pagemap page?  I'm not sure that cmci can be triggered on the
> nvdimm device, but this injection interface is open for such a case.
> Maybe simply ignoring the event is an expected behavior?

Yeah, I should explicitly block attempts to MADV_SOFT_OFFLINE a
dev_pagemap page because they are never onlined. They'll never appear
on an lru and never be available via the page allocator.

>
> A few comments/quetions below ...
>
>>
>> Cc: Jan Kara <jack@suse.cz>
>> Cc: Christoph Hellwig <hch@lst.de>
>> Cc: Jérôme Glisse <jglisse@redhat.com>
>> Cc: Matthew Wilcox <mawilcox@microsoft.com>
>> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
>> Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
>> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
>> ---
>>  include/linux/mm.h  |    1
>>  mm/memory-failure.c |  124 ++++++++++++++++++++++++++++++++++++++++++++++++++-
>>  2 files changed, 123 insertions(+), 2 deletions(-)
>>
>> diff --git a/include/linux/mm.h b/include/linux/mm.h
>> index a0fbb9ffe380..374e5e9284f7 100644
>> --- a/include/linux/mm.h
>> +++ b/include/linux/mm.h
>> @@ -2725,6 +2725,7 @@ enum mf_action_page_type {
>>       MF_MSG_TRUNCATED_LRU,
>>       MF_MSG_BUDDY,
>>       MF_MSG_BUDDY_2ND,
>> +     MF_MSG_DAX,
>>       MF_MSG_UNKNOWN,
>>  };
>>
>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>> index 4d70753af59c..161aa1b70212 100644
>> --- a/mm/memory-failure.c
>> +++ b/mm/memory-failure.c
>> @@ -55,6 +55,7 @@
>>  #include <linux/hugetlb.h>
>>  #include <linux/memory_hotplug.h>
>>  #include <linux/mm_inline.h>
>> +#include <linux/memremap.h>
>>  #include <linux/kfifo.h>
>>  #include <linux/ratelimit.h>
>>  #include "internal.h"
>> @@ -263,6 +264,39 @@ void shake_page(struct page *p, int access)
>>  }
>>  EXPORT_SYMBOL_GPL(shake_page);
>>
>> +static unsigned long mapping_size(struct page *page, struct vm_area_struct *vma)
>> +{
>> +     unsigned long address = vma_address(page, vma);
>> +     pgd_t *pgd;
>> +     p4d_t *p4d;
>> +     pud_t *pud;
>> +     pmd_t *pmd;
>> +     pte_t *pte;
>> +
>> +     pgd = pgd_offset(vma->vm_mm, address);
>> +     if (!pgd_present(*pgd))
>> +             return 0;
>> +     p4d = p4d_offset(pgd, address);
>> +     if (!p4d_present(*p4d))
>> +             return 0;
>> +     pud = pud_offset(p4d, address);
>> +     if (!pud_present(*pud))
>> +             return 0;
>> +     if (pud_devmap(*pud))
>> +             return PUD_SIZE;
>> +     pmd = pmd_offset(pud, address);
>> +     if (!pmd_present(*pmd))
>> +             return 0;
>> +     if (pmd_devmap(*pmd))
>> +             return PMD_SIZE;
>> +     pte = pte_offset_map(pmd, address);
>> +     if (!pte_present(*pte))
>> +             return 0;
>> +     if (pte_devmap(*pte))
>> +             return PAGE_SIZE;
>> +     return 0;
>> +}
>> +
>
> The function name looks generic, but this function seems to focus on
> devmap thing, so could you include the word 'devmap' in the name?

Ok.

>
>>  /*
>>   * Failure handling: if we can't find or can't kill a process there's
>>   * not much we can do.       We just print a message and ignore otherwise.
>> @@ -292,7 +326,10 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,
>>       }
>>       tk->addr = page_address_in_vma(p, vma);
>>       tk->addr_valid = 1;
>> -     tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;
>> +     if (is_zone_device_page(p))
>> +             tk->size_shift = ilog2(mapping_size(p, vma));
>> +     else
>> +             tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;
>>
>>       /*
>>        * In theory we don't have to kill when the page was
>> @@ -300,7 +337,7 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,
>>        * likely very rare kill anyways just out of paranoia, but use
>>        * a SIGKILL because the error is not contained anymore.
>>        */
>> -     if (tk->addr == -EFAULT) {
>> +     if (tk->addr == -EFAULT || tk->size_shift == 0) {
>>               pr_info("Memory failure: Unable to find user space address %lx in %s\n",
>>                       page_to_pfn(p), tsk->comm);
>>               tk->addr_valid = 0;
>> @@ -514,6 +551,7 @@ static const char * const action_page_types[] = {
>>       [MF_MSG_TRUNCATED_LRU]          = "already truncated LRU page",
>>       [MF_MSG_BUDDY]                  = "free buddy page",
>>       [MF_MSG_BUDDY_2ND]              = "free buddy page (2nd try)",
>> +     [MF_MSG_DAX]                    = "dax page",
>>       [MF_MSG_UNKNOWN]                = "unknown page",
>>  };
>>
>> @@ -1111,6 +1149,83 @@ static int memory_failure_hugetlb(unsigned long pfn, int flags)
>>       return res;
>>  }
>>
>> +static int memory_failure_dev_pagemap(unsigned long pfn, int flags,
>> +             struct dev_pagemap *pgmap)
>> +{
>> +     struct page *page = pfn_to_page(pfn);
>> +     const bool unmap_success = true;
>> +     unsigned long size = 0;
>> +     struct to_kill *tk;
>> +     LIST_HEAD(tokill);
>> +     int rc = -EBUSY;
>> +     loff_t start;
>> +
>> +     /*
>> +      * Prevent the inode from being freed while we are interrogating
>> +      * the address_space, typically this would be handled by
>> +      * lock_page(), but dax pages do not use the page lock. This
>> +      * also prevents changes to the mapping of this pfn until
>> +      * poison signaling is complete.
>> +      */
>> +     if (!dax_lock_mapping_entry(page))
>> +             goto out;
>> +
>> +     if (hwpoison_filter(page)) {
>> +             rc = 0;
>> +             goto unlock;
>> +     }
>> +
>> +     switch (pgmap->type) {
>> +     case MEMORY_DEVICE_PRIVATE:
>> +     case MEMORY_DEVICE_PUBLIC:
>> +             /*
>> +              * TODO: Handle HMM pages which may need coordination
>> +              * with device-side memory.
>> +              */
>> +             goto unlock;
>> +     default:
>> +             break;
>> +     }
>> +
>> +     /*
>> +      * Use this flag as an indication that the dax page has been
>> +      * remapped UC to prevent speculative consumption of poison.
>> +      */
>> +     SetPageHWPoison(page);
>
> The number of hwpoison pages is maintained by num_poisoned_pages,
> so you can call num_poisoned_pages_inc()?

I don't think we want these pages accounted in num_poisoned_pages().
We have the badblocks infrastructure in libnvdimm to track how many
errors and where they are located, and since they can be repaired via
driver actions I think we should track them separately.

> Related to this, I'm interested in whether/how unpoison_page() works
> on a hwpoisoned dev_pagemap page.

unpoison_page() is only triggered via freeing pages to the page
allocator, and that never happens for dev_pagemap / ZONE_DEVICE pages.
Naoya Horiguchi July 17, 2018, 6:36 a.m. UTC | #3
On Fri, Jul 13, 2018 at 05:28:05PM -0700, Dan Williams wrote:
> On Fri, Jul 13, 2018 at 1:52 AM, Naoya Horiguchi
> <n-horiguchi@ah.jp.nec.com> wrote:
> > On Wed, Jul 04, 2018 at 02:41:06PM -0700, Dan Williams wrote:
...
> >> +
> >> +     /*
> >> +      * Use this flag as an indication that the dax page has been
> >> +      * remapped UC to prevent speculative consumption of poison.
> >> +      */
> >> +     SetPageHWPoison(page);
> >
> > The number of hwpoison pages is maintained by num_poisoned_pages,
> > so you can call num_poisoned_pages_inc()?
> 
> I don't think we want these pages accounted in num_poisoned_pages().
> We have the badblocks infrastructure in libnvdimm to track how many
> errors and where they are located, and since they can be repaired via
> driver actions I think we should track them separately.

OK.

> > Related to this, I'm interested in whether/how unpoison_page() works
> > on a hwpoisoned dev_pagemap page.
> 
> unpoison_page() is only triggered via freeing pages to the page
> allocator, and that never happens for dev_pagemap / ZONE_DEVICE pages.

sorry, my bad comment.
I meant unpoison_memory() in mm/memory-failure.c, which is triggered
via debugfs:hwpoison/unpoison-pfn. This interface looks like below

  int unpoison_memory(unsigned long pfn)
  {
          struct page *page;
          struct page *p;
          int freeit = 0;
          static DEFINE_RATELIMIT_STATE(unpoison_rs, DEFAULT_RATELIMIT_INTERVAL,
                                          DEFAULT_RATELIMIT_BURST);

          if (!pfn_valid(pfn))
                  return -ENXIO;

          p = pfn_to_page(pfn);
          page = compound_head(p);

          if (!PageHWPoison(p)) {
                  unpoison_pr_info("Unpoison: Page was already unpoisoned %#lx\n",
                                   pfn, &unpoison_rs);
                  return 0;
          }
  ...

so I think that we can add is_zone_device_page() check at the beginning
of this function to call hwpoison_clear() introduced in patch 13/13?
Otherwise maybe compound_head() will cause some critical issue like
general protection fault.

Thanks,
Naoya Horiguchi
diff mbox

Patch

diff --git a/include/linux/mm.h b/include/linux/mm.h
index a0fbb9ffe380..374e5e9284f7 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2725,6 +2725,7 @@  enum mf_action_page_type {
 	MF_MSG_TRUNCATED_LRU,
 	MF_MSG_BUDDY,
 	MF_MSG_BUDDY_2ND,
+	MF_MSG_DAX,
 	MF_MSG_UNKNOWN,
 };
 
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 4d70753af59c..161aa1b70212 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -55,6 +55,7 @@ 
 #include <linux/hugetlb.h>
 #include <linux/memory_hotplug.h>
 #include <linux/mm_inline.h>
+#include <linux/memremap.h>
 #include <linux/kfifo.h>
 #include <linux/ratelimit.h>
 #include "internal.h"
@@ -263,6 +264,39 @@  void shake_page(struct page *p, int access)
 }
 EXPORT_SYMBOL_GPL(shake_page);
 
+static unsigned long mapping_size(struct page *page, struct vm_area_struct *vma)
+{
+	unsigned long address = vma_address(page, vma);
+	pgd_t *pgd;
+	p4d_t *p4d;
+	pud_t *pud;
+	pmd_t *pmd;
+	pte_t *pte;
+
+	pgd = pgd_offset(vma->vm_mm, address);
+	if (!pgd_present(*pgd))
+		return 0;
+	p4d = p4d_offset(pgd, address);
+	if (!p4d_present(*p4d))
+		return 0;
+	pud = pud_offset(p4d, address);
+	if (!pud_present(*pud))
+		return 0;
+	if (pud_devmap(*pud))
+		return PUD_SIZE;
+	pmd = pmd_offset(pud, address);
+	if (!pmd_present(*pmd))
+		return 0;
+	if (pmd_devmap(*pmd))
+		return PMD_SIZE;
+	pte = pte_offset_map(pmd, address);
+	if (!pte_present(*pte))
+		return 0;
+	if (pte_devmap(*pte))
+		return PAGE_SIZE;
+	return 0;
+}
+
 /*
  * Failure handling: if we can't find or can't kill a process there's
  * not much we can do.	We just print a message and ignore otherwise.
@@ -292,7 +326,10 @@  static void add_to_kill(struct task_struct *tsk, struct page *p,
 	}
 	tk->addr = page_address_in_vma(p, vma);
 	tk->addr_valid = 1;
-	tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;
+	if (is_zone_device_page(p))
+		tk->size_shift = ilog2(mapping_size(p, vma));
+	else
+		tk->size_shift = compound_order(compound_head(p)) + PAGE_SHIFT;
 
 	/*
 	 * In theory we don't have to kill when the page was
@@ -300,7 +337,7 @@  static void add_to_kill(struct task_struct *tsk, struct page *p,
 	 * likely very rare kill anyways just out of paranoia, but use
 	 * a SIGKILL because the error is not contained anymore.
 	 */
-	if (tk->addr == -EFAULT) {
+	if (tk->addr == -EFAULT || tk->size_shift == 0) {
 		pr_info("Memory failure: Unable to find user space address %lx in %s\n",
 			page_to_pfn(p), tsk->comm);
 		tk->addr_valid = 0;
@@ -514,6 +551,7 @@  static const char * const action_page_types[] = {
 	[MF_MSG_TRUNCATED_LRU]		= "already truncated LRU page",
 	[MF_MSG_BUDDY]			= "free buddy page",
 	[MF_MSG_BUDDY_2ND]		= "free buddy page (2nd try)",
+	[MF_MSG_DAX]			= "dax page",
 	[MF_MSG_UNKNOWN]		= "unknown page",
 };
 
@@ -1111,6 +1149,83 @@  static int memory_failure_hugetlb(unsigned long pfn, int flags)
 	return res;
 }
 
+static int memory_failure_dev_pagemap(unsigned long pfn, int flags,
+		struct dev_pagemap *pgmap)
+{
+	struct page *page = pfn_to_page(pfn);
+	const bool unmap_success = true;
+	unsigned long size = 0;
+	struct to_kill *tk;
+	LIST_HEAD(tokill);
+	int rc = -EBUSY;
+	loff_t start;
+
+	/*
+	 * Prevent the inode from being freed while we are interrogating
+	 * the address_space, typically this would be handled by
+	 * lock_page(), but dax pages do not use the page lock. This
+	 * also prevents changes to the mapping of this pfn until
+	 * poison signaling is complete.
+	 */
+	if (!dax_lock_mapping_entry(page))
+		goto out;
+
+	if (hwpoison_filter(page)) {
+		rc = 0;
+		goto unlock;
+	}
+
+	switch (pgmap->type) {
+	case MEMORY_DEVICE_PRIVATE:
+	case MEMORY_DEVICE_PUBLIC:
+		/*
+		 * TODO: Handle HMM pages which may need coordination
+		 * with device-side memory.
+		 */
+		goto unlock;
+	default:
+		break;
+	}
+
+	/*
+	 * Use this flag as an indication that the dax page has been
+	 * remapped UC to prevent speculative consumption of poison.
+	 */
+	SetPageHWPoison(page);
+
+	/*
+	 * 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(page, &tokill, flags & MF_ACTION_REQUIRED);
+
+	list_for_each_entry(tk, &tokill, nd)
+		if (tk->size_shift)
+			size = max(size, 1UL << tk->size_shift);
+	if (size) {
+		/*
+		 * Unmap the largest mapping to avoid breaking up
+		 * device-dax mappings which are constant size. The
+		 * actual size of the mapping being torn down is
+		 * communicated in siginfo, see kill_proc()
+		 */
+		start = (page->index << PAGE_SHIFT) & ~(size - 1);
+		unmap_mapping_range(page->mapping, start, start + size, 0);
+	}
+	kill_procs(&tokill, flags & MF_MUST_KILL, !unmap_success, pfn, flags);
+	rc = 0;
+unlock:
+	dax_unlock_mapping_entry(page);
+out:
+	/* drop pgmap ref acquired in caller */
+	put_dev_pagemap(pgmap);
+	action_result(pfn, MF_MSG_DAX, rc ? MF_FAILED : MF_RECOVERED);
+	return rc;
+}
+
 /**
  * memory_failure - Handle memory failure of a page.
  * @pfn: Page Number of the corrupted page
@@ -1133,6 +1248,7 @@  int memory_failure(unsigned long pfn, int flags)
 	struct page *p;
 	struct page *hpage;
 	struct page *orig_head;
+	struct dev_pagemap *pgmap;
 	int res;
 	unsigned long page_flags;
 
@@ -1145,6 +1261,10 @@  int memory_failure(unsigned long pfn, int flags)
 		return -ENXIO;
 	}
 
+	pgmap = get_dev_pagemap(pfn, NULL);
+	if (pgmap)
+		return memory_failure_dev_pagemap(pfn, flags, pgmap);
+
 	p = pfn_to_page(pfn);
 	if (PageHuge(p))
 		return memory_failure_hugetlb(pfn, flags);