linux-next: manual merge of the akpm-current tree with the xen-tip tree

linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2516 bytes --]

Hi all,

Today's linux-next merge of the akpm-current tree got a conflict in:

  drivers/xen/gntdev.c

between commit:

  1d3145675538 ("xen/gntdev: Make private routines/structures accessible")

from the xen-tip tree and commit:

  aaefcabe9c25 ("mm, oom: distinguish blockable mode for mmu notifiers")

from the akpm-current tree.

I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging.  You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.

-- 
Cheers,
Stephen Rothwell

diff --cc drivers/xen/gntdev.c
index c866a62f766d,55b4f0e3f4d6..000000000000
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@@ -479,7 -441,20 +479,20 @@@ static const struct vm_operations_struc
  
  /* ------------------------------------------------------------------ */
  
 -static bool in_range(struct grant_map *map,
++static bool in_range(struct gntdev_grant_map *map,
+ 			      unsigned long start, unsigned long end)
+ {
+ 	if (!map->vma)
+ 		return false;
+ 	if (map->vma->vm_start >= end)
+ 		return false;
+ 	if (map->vma->vm_end <= start)
+ 		return false;
+ 
+ 	return true;
+ }
+ 
 -static void unmap_if_in_range(struct grant_map *map,
 +static void unmap_if_in_range(struct gntdev_grant_map *map,
  			      unsigned long start, unsigned long end)
  {
  	unsigned long mstart, mend;
@@@ -503,15 -472,26 +510,26 @@@
  	WARN_ON(err);
  }
  
- static void mn_invl_range_start(struct mmu_notifier *mn,
+ static int mn_invl_range_start(struct mmu_notifier *mn,
  				struct mm_struct *mm,
- 				unsigned long start, unsigned long end)
+ 				unsigned long start, unsigned long end,
+ 				bool blockable)
  {
  	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
 -	struct grant_map *map;
 +	struct gntdev_grant_map *map;
+ 	int ret = 0;
+ 
+ 	/* TODO do we really need a mutex here? */
+ 	if (blockable)
+ 		mutex_lock(&priv->lock);
+ 	else if (!mutex_trylock(&priv->lock))
+ 		return -EAGAIN;
  
- 	mutex_lock(&priv->lock);
  	list_for_each_entry(map, &priv->maps, next) {
+ 		if (in_range(map, start, end)) {
+ 			ret = -EAGAIN;
+ 			goto out_unlock;
+ 		}
  		unmap_if_in_range(map, start, end);
  	}
  	list_for_each_entry(map, &priv->freeable_maps, next) {

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Boris Ostrovsky]

[-- Attachment #1.1: Type: text/plain, Size: 2786 bytes --]

On 07/30/2018 05:02 AM, Stephen Rothwell wrote:
> Hi all,
>
> Today's linux-next merge of the akpm-current tree got a conflict in:
>
>   drivers/xen/gntdev.c
>
> between commit:
>
>   1d3145675538 ("xen/gntdev: Make private routines/structures accessible")
>
> from the xen-tip tree and commit:
>
>   aaefcabe9c25 ("mm, oom: distinguish blockable mode for mmu notifiers")
>
> from the akpm-current tree.
>
> I fixed it up (see below) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging.  You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.
>
> -- Cheers, Stephen Rothwell diff --cc drivers/xen/gntdev.c index
> c866a62f766d,55b4f0e3f4d6..000000000000 --- a/drivers/xen/gntdev.c +++
> b/drivers/xen/gntdev.c @@@ -479,7 -441,20 +479,20 @@@ static const
> struct vm_operations_struc /*
> ------------------------------------------------------------------ */
> -static bool in_range(struct grant_map *map, ++static bool
> in_range(struct gntdev_grant_map *map, + unsigned long start, unsigned
> long end) + { + if (!map->vma) + return false; + if
> (map->vma->vm_start >= end) + return false; + if (map->vma->vm_end <=
> start) + return false; + + return true; + } + -static void
> unmap_if_in_range(struct grant_map *map, +static void
> unmap_if_in_range(struct gntdev_grant_map *map, unsigned long start,
> unsigned long end) { unsigned long mstart, mend; @@@ -503,15 -472,26
> +510,26 @@@ WARN_ON(err); } - static void mn_invl_range_start(struct
> mmu_notifier *mn, + static int mn_invl_range_start(struct mmu_notifier
> *mn, struct mm_struct *mm, - unsigned long start, unsigned long end) +
> unsigned long start, unsigned long end, + bool blockable) { struct
> gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn); - struct
> grant_map *map; + struct gntdev_grant_map *map; + int ret = 0; + + /*
> TODO do we really need a mutex here? */ + if (blockable) +
> mutex_lock(&priv->lock); + else if (!mutex_trylock(&priv->lock)) +
> return -EAGAIN; - mutex_lock(&priv->lock); list_for_each_entry(map,
> &priv->maps, next) { + if (in_range(map, start, end)) { + ret =
> -EAGAIN; + goto out_unlock; + } unmap_if_in_range(map, start, end); }
> list_for_each_entry(map, &priv->freeable_maps, next) {


I clearly missed this (aaefcabe9c25) patch but now that I am looking at
it I don't think I understand the logic for changes in
list_for_each_entry() loops.

Aren't we ending up never unmapping grant pages? Michal, can you explain
what you are trying to do here?


-boris




[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Boris Ostrovsky]

[-- Attachment #1.1: Type: text/plain, Size: 5171 bytes --]

On 07/30/2018 01:02 PM, Boris Ostrovsky wrote:
> On 07/30/2018 05:02 AM, Stephen Rothwell wrote:
>> Hi all,
>>
>> Today's linux-next merge of the akpm-current tree got a conflict in:
>>
>>   drivers/xen/gntdev.c
>>
>> between commit:
>>
>>   1d3145675538 ("xen/gntdev: Make private routines/structures accessible")
>>
>> from the xen-tip tree and commit:
>>
>>   aaefcabe9c25 ("mm, oom: distinguish blockable mode for mmu notifiers")
>>
>> from the akpm-current tree.
>>
>> I fixed it up (see below) and can carry the fix as necessary. This
>> is now fixed as far as linux-next is concerned, but any non trivial
>> conflicts should be mentioned to your upstream maintainer when your tree
>> is submitted for merging.  You may also want to consider cooperating
>> with the maintainer of the conflicting tree to minimise any particularly
>> complex conflicts.
>>
>> -- Cheers, Stephen Rothwell diff --cc drivers/xen/gntdev.c index
>> c866a62f766d,55b4f0e3f4d6..000000000000 --- a/drivers/xen/gntdev.c +++
>> b/drivers/xen/gntdev.c @@@ -479,7 -441,20 +479,20 @@@ static const
>> struct vm_operations_struc /*
>> ------------------------------------------------------------------ */
>> -static bool in_range(struct grant_map *map, ++static bool
>> in_range(struct gntdev_grant_map *map, + unsigned long start, unsigned
>> long end) + { + if (!map->vma) + return false; + if
>> (map->vma->vm_start >= end) + return false; + if (map->vma->vm_end <=
>> start) + return false; + + return true; + } + -static void
>> unmap_if_in_range(struct grant_map *map, +static void
>> unmap_if_in_range(struct gntdev_grant_map *map, unsigned long start,
>> unsigned long end) { unsigned long mstart, mend; @@@ -503,15 -472,26
>> +510,26 @@@ WARN_ON(err); } - static void mn_invl_range_start(struct
>> mmu_notifier *mn, + static int mn_invl_range_start(struct mmu_notifier
>> *mn, struct mm_struct *mm, - unsigned long start, unsigned long end) +
>> unsigned long start, unsigned long end, + bool blockable) { struct
>> gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn); - struct
>> grant_map *map; + struct gntdev_grant_map *map; + int ret = 0; + + /*
>> TODO do we really need a mutex here? */ + if (blockable) +
>> mutex_lock(&priv->lock); + else if (!mutex_trylock(&priv->lock)) +
>> return -EAGAIN; - mutex_lock(&priv->lock); list_for_each_entry(map,
>> &priv->maps, next) { + if (in_range(map, start, end)) { + ret =
>> -EAGAIN; + goto out_unlock; + } unmap_if_in_range(map, start, end); }
>> list_for_each_entry(map, &priv->freeable_maps, next) {


Ugh... That's some interesting whitespace optimization on part of
thundebird. Let me paste the relevant patch hunk here.


diff --git a/drivers/xen/gntdev.c b/drivers/xen/gntdev.c
index bd56653b9bbc..55b4f0e3f4d6 100644
--- a/drivers/xen/gntdev.c
+++ b/drivers/xen/gntdev.c
@@ -441,18 +441,25 @@ static const struct vm_operations_struct gntdev_vmops = {
 
 /* ------------------------------------------------------------------ */
 
+static bool in_range(struct grant_map *map,
+			      unsigned long start, unsigned long end)
+{
+	if (!map->vma)
+		return false;
+	if (map->vma->vm_start >= end)
+		return false;
+	if (map->vma->vm_end <= start)
+		return false;
+
+	return true;
+}
+
 static void unmap_if_in_range(struct grant_map *map,
 			      unsigned long start, unsigned long end)
 {
 	unsigned long mstart, mend;
 	int err;
 
-	if (!map->vma)
-		return;
-	if (map->vma->vm_start >= end)
-		return;
-	if (map->vma->vm_end <= start)
-		return;
 	mstart = max(start, map->vma->vm_start);
 	mend   = min(end,   map->vma->vm_end);
 	pr_debug("map %d+%d (%lx %lx), range %lx %lx, mrange %lx %lx\n",
@@ -465,21 +472,40 @@ static void unmap_if_in_range(struct grant_map *map,
 	WARN_ON(err);
 }
 
-static void mn_invl_range_start(struct mmu_notifier *mn,
+static int mn_invl_range_start(struct mmu_notifier *mn,
 				struct mm_struct *mm,
-				unsigned long start, unsigned long end)
+				unsigned long start, unsigned long end,
+				bool blockable)
 {
 	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
 	struct grant_map *map;
+	int ret = 0;
+
+	/* TODO do we really need a mutex here? */
+	if (blockable)
+		mutex_lock(&priv->lock);
+	else if (!mutex_trylock(&priv->lock))
+		return -EAGAIN;
 
-	mutex_lock(&priv->lock);
 	list_for_each_entry(map, &priv->maps, next) {
+		if (in_range(map, start, end)) {
+			ret = -EAGAIN;
+			goto out_unlock;
+		}
 		unmap_if_in_range(map, start, end);
 	}
 	list_for_each_entry(map, &priv->freeable_maps, next) {
+		if (in_range(map, start, end)) {
+			ret = -EAGAIN;
+			goto out_unlock;
+		}
 		unmap_if_in_range(map, start, end);
 	}
+
+out_unlock:
 	mutex_unlock(&priv->lock);
+
+	return ret;
 }


-boris



>
> I clearly missed this (aaefcabe9c25) patch but now that I am looking at
> it I don't think I understand the logic for changes in
> list_for_each_entry() loops.
>
> Aren't we ending up never unmapping grant pages? Michal, can you explain
> what you are trying to do here?
>
>
> -boris
>
>
>



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2875 bytes --]

Hi all,

On Mon, 30 Jul 2018 19:02:10 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Today's linux-next merge of the akpm-current tree got a conflict in:
> 
>   drivers/xen/gntdev.c
> 
> between commit:
> 
>   1d3145675538 ("xen/gntdev: Make private routines/structures accessible")
> 
> from the xen-tip tree and commit:
> 
>   aaefcabe9c25 ("mm, oom: distinguish blockable mode for mmu notifiers")
> 
> from the akpm-current tree.
> 
> I fixed it up (see below) and can carry the fix as necessary. This
> is now fixed as far as linux-next is concerned, but any non trivial
> conflicts should be mentioned to your upstream maintainer when your tree
> is submitted for merging.  You may also want to consider cooperating
> with the maintainer of the conflicting tree to minimise any particularly
> complex conflicts.
> 
> -- 
> Cheers,
> Stephen Rothwell
> 
> diff --cc drivers/xen/gntdev.c
> index c866a62f766d,55b4f0e3f4d6..000000000000
> --- a/drivers/xen/gntdev.c
> +++ b/drivers/xen/gntdev.c
> @@@ -479,7 -441,20 +479,20 @@@ static const struct vm_operations_struc
>   
>   /* ------------------------------------------------------------------ */
>   
>  -static bool in_range(struct grant_map *map,
> ++static bool in_range(struct gntdev_grant_map *map,
> + 			      unsigned long start, unsigned long end)
> + {
> + 	if (!map->vma)
> + 		return false;
> + 	if (map->vma->vm_start >= end)
> + 		return false;
> + 	if (map->vma->vm_end <= start)
> + 		return false;
> + 
> + 	return true;
> + }
> + 
>  -static void unmap_if_in_range(struct grant_map *map,
>  +static void unmap_if_in_range(struct gntdev_grant_map *map,
>   			      unsigned long start, unsigned long end)
>   {
>   	unsigned long mstart, mend;
> @@@ -503,15 -472,26 +510,26 @@@
>   	WARN_ON(err);
>   }
>   
> - static void mn_invl_range_start(struct mmu_notifier *mn,
> + static int mn_invl_range_start(struct mmu_notifier *mn,
>   				struct mm_struct *mm,
> - 				unsigned long start, unsigned long end)
> + 				unsigned long start, unsigned long end,
> + 				bool blockable)
>   {
>   	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
>  -	struct grant_map *map;
>  +	struct gntdev_grant_map *map;
> + 	int ret = 0;
> + 
> + 	/* TODO do we really need a mutex here? */
> + 	if (blockable)
> + 		mutex_lock(&priv->lock);
> + 	else if (!mutex_trylock(&priv->lock))
> + 		return -EAGAIN;
>   
> - 	mutex_lock(&priv->lock);
>   	list_for_each_entry(map, &priv->maps, next) {
> + 		if (in_range(map, start, end)) {
> + 			ret = -EAGAIN;
> + 			goto out_unlock;
> + 		}
>   		unmap_if_in_range(map, start, end);
>   	}
>   	list_for_each_entry(map, &priv->freeable_maps, next) {

This is now a conflict between Linus' tree and the akpm-current tree.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [Xen-devel] linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Boris Ostrovsky]

[-- Attachment #1.1: Type: text/plain, Size: 3008 bytes --]

On 08/15/2018 08:05 PM, Stephen Rothwell wrote:
> Hi all,
>
> On Mon, 30 Jul 2018 19:02:10 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>> Today's linux-next merge of the akpm-current tree got a conflict in:
>>
>>   drivers/xen/gntdev.c
>>
>> between commit:
>>
>>   1d3145675538 ("xen/gntdev: Make private routines/structures accessible")
>>
>> from the xen-tip tree and commit:
>>
>>   aaefcabe9c25 ("mm, oom: distinguish blockable mode for mmu notifiers")
>>
>> from the akpm-current tree.
>>
>> I fixed it up (see below) and can carry the fix as necessary. This
>> is now fixed as far as linux-next is concerned, but any non trivial
>> conflicts should be mentioned to your upstream maintainer when your tree
>> is submitted for merging.  You may also want to consider cooperating
>> with the maintainer of the conflicting tree to minimise any particularly
>> complex conflicts.
>>
>> -- 
>> Cheers,
>> Stephen Rothwell
>>
>> diff --cc drivers/xen/gntdev.c
>> index c866a62f766d,55b4f0e3f4d6..000000000000
>> --- a/drivers/xen/gntdev.c
>> +++ b/drivers/xen/gntdev.c
>> @@@ -479,7 -441,20 +479,20 @@@ static const struct vm_operations_struc
>>   
>>   /* ------------------------------------------------------------------ */
>>   
>>  -static bool in_range(struct grant_map *map,
>> ++static bool in_range(struct gntdev_grant_map *map,
>> + 			      unsigned long start, unsigned long end)
>> + {
>> + 	if (!map->vma)
>> + 		return false;
>> + 	if (map->vma->vm_start >= end)
>> + 		return false;
>> + 	if (map->vma->vm_end <= start)
>> + 		return false;
>> + 
>> + 	return true;
>> + }
>> + 
>>  -static void unmap_if_in_range(struct grant_map *map,
>>  +static void unmap_if_in_range(struct gntdev_grant_map *map,
>>   			      unsigned long start, unsigned long end)
>>   {
>>   	unsigned long mstart, mend;
>> @@@ -503,15 -472,26 +510,26 @@@
>>   	WARN_ON(err);
>>   }
>>   
>> - static void mn_invl_range_start(struct mmu_notifier *mn,
>> + static int mn_invl_range_start(struct mmu_notifier *mn,
>>   				struct mm_struct *mm,
>> - 				unsigned long start, unsigned long end)
>> + 				unsigned long start, unsigned long end,
>> + 				bool blockable)
>>   {
>>   	struct gntdev_priv *priv = container_of(mn, struct gntdev_priv, mn);
>>  -	struct grant_map *map;
>>  +	struct gntdev_grant_map *map;
>> + 	int ret = 0;
>> + 
>> + 	/* TODO do we really need a mutex here? */
>> + 	if (blockable)
>> + 		mutex_lock(&priv->lock);
>> + 	else if (!mutex_trylock(&priv->lock))
>> + 		return -EAGAIN;
>>   
>> - 	mutex_lock(&priv->lock);
>>   	list_for_each_entry(map, &priv->maps, next) {
>> + 		if (in_range(map, start, end)) {
>> + 			ret = -EAGAIN;
>> + 			goto out_unlock;
>> + 		}
>>   		unmap_if_in_range(map, start, end);


I think I mentioned this earlier --- this doesn't look right. Not the
conflict resolution but the original patch.

Should I send a patch against -next? Or -mm?


-boris




[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

linux-next: manual merge of the akpm-current tree with the xen-tip tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1309 bytes --]

Hi Andrew,

Today's linux-next merge of the akpm-current tree got a conflict in
include/linux/mm.h between commit 667a0a06c99d ("mm: provide a
find_special_page vma operation") from the xen-tip tree and commit
cfd00f5aa195 ("mm: drop vm_ops->remap_pages and
generic_file_remap_pages() stub") from the akpm-current tree.

I fixed it up (see below) and can carry the fix as necessary (no action
is required).

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

diff --cc include/linux/mm.h
index 9269af7349fe,b5e28731a293..000000000000
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@@ -287,17 -284,6 +284,14 @@@ struct vm_operations_struct 
  	struct mempolicy *(*get_policy)(struct vm_area_struct *vma,
  					unsigned long addr);
  #endif
- 	/* called by sys_remap_file_pages() to populate non-linear mapping */
- 	int (*remap_pages)(struct vm_area_struct *vma, unsigned long addr,
- 			   unsigned long size, pgoff_t pgoff);
 +
 +	/*
 +	 * Called by vm_normal_page() for special PTEs to find the
 +	 * page for @addr.  This is useful if the default behavior
 +	 * (using pte_page()) would not find the correct page.
 +	 */
 +	struct page *(*find_special_page)(struct vm_area_struct *vma,
 +					  unsigned long addr);
  };
  
  struct mmu_gather;

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]