linux-next: build failure after merge of the hmm tree

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2538 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

drivers/gpu/drm/i915/gem/i915_gem_ttm.c: In function 'i915_ttm_tt_get_st':
drivers/gpu/drm/i915/gem/i915_gem_ttm.c:396:7: error: implicit declaration of function '__sg_alloc_table_from_pages'; did you mean 'sg_alloc_table_from_pages'? [-Werror=implicit-function-declaration]
  396 |  sg = __sg_alloc_table_from_pages
      |       ^~~~~~~~~~~~~~~~~~~~~~~~~~~
      |       sg_alloc_table_from_pages
drivers/gpu/drm/i915/gem/i915_gem_ttm.c:396:5: warning: assignment to 'struct scatterlist *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
  396 |  sg = __sg_alloc_table_from_pages
      |     ^

Caused by commit

  fcbfe956561b ("lib/scatterlist: Provide a dedicated function to support table append")

interacting with commit

  213d50927763 ("drm/i915/ttm: Introduce a TTM i915 gem object backend")

from the drm tree.

I have applied the following merge resolution patch.

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 23 Aug 2021 17:46:27 +1000
Subject: [PATCH] drm/i915/ttm: fix up for "lib/scatterlist: Provide a
 dedicated function to support tableappend"

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
index 771eb2963123..d3d95934a047 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -382,7 +382,6 @@ i915_ttm_region(struct ttm_device *bdev, int ttm_mem_type)
 static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
 {
 	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
-	struct scatterlist *sg;
 	struct sg_table *st;
 	int ret;
 
@@ -393,13 +392,13 @@ static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
 	if (!st)
 		return ERR_PTR(-ENOMEM);
 
-	sg = __sg_alloc_table_from_pages
+	ret = sg_alloc_table_from_pages_segment
 		(st, ttm->pages, ttm->num_pages, 0,
 		 (unsigned long)ttm->num_pages << PAGE_SHIFT,
-		 i915_sg_segment_size(), NULL, 0, GFP_KERNEL);
-	if (IS_ERR(sg)) {
+		 i915_sg_segment_size(), GFP_KERNEL);
+	if (ret) {
 		kfree(st);
-		return ERR_CAST(sg);
+		return ERR_PTR(ret);
 	}
 
 	ret = dma_map_sgtable(i915_tt->dev, st, DMA_BIDIRECTIONAL, 0);
-- 
2.32.0

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Mon, Aug 23, 2021 at 05:50:27PM +1000, Stephen Rothwell wrote:
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Mon, 23 Aug 2021 17:46:27 +1000
> Subject: [PATCH] drm/i915/ttm: fix up for "lib/scatterlist: Provide a
>  dedicated function to support tableappend"
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
>  drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 9 ++++-----
>  1 file changed, 4 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> index 771eb2963123..d3d95934a047 100644
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> @@ -382,7 +382,6 @@ i915_ttm_region(struct ttm_device *bdev, int ttm_mem_type)
>  static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
>  {
>  	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
> -	struct scatterlist *sg;
>  	struct sg_table *st;
>  	int ret;
>  
> @@ -393,13 +392,13 @@ static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
>  	if (!st)
>  		return ERR_PTR(-ENOMEM);
>  
> -	sg = __sg_alloc_table_from_pages
> +	ret = sg_alloc_table_from_pages_segment
>  		(st, ttm->pages, ttm->num_pages, 0,
>  		 (unsigned long)ttm->num_pages << PAGE_SHIFT,
> -		 i915_sg_segment_size(), NULL, 0, GFP_KERNEL);
> -	if (IS_ERR(sg)) {
> +		 i915_sg_segment_size(), GFP_KERNEL);
> +	if (ret) {
>  		kfree(st);
> -		return ERR_CAST(sg);
> +		return ERR_PTR(ret);
>  	}

Looks right, thanks

Jason

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2905 bytes --]

Hi ,

On Mon, 23 Aug 2021 17:50:27 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Hi all,
> 
> After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> drivers/gpu/drm/i915/gem/i915_gem_ttm.c: In function 'i915_ttm_tt_get_st':
> drivers/gpu/drm/i915/gem/i915_gem_ttm.c:396:7: error: implicit declaration of function '__sg_alloc_table_from_pages'; did you mean 'sg_alloc_table_from_pages'? [-Werror=implicit-function-declaration]
>   396 |  sg = __sg_alloc_table_from_pages
>       |       ^~~~~~~~~~~~~~~~~~~~~~~~~~~
>       |       sg_alloc_table_from_pages
> drivers/gpu/drm/i915/gem/i915_gem_ttm.c:396:5: warning: assignment to 'struct scatterlist *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
>   396 |  sg = __sg_alloc_table_from_pages
>       |     ^
> 
> Caused by commit
> 
>   fcbfe956561b ("lib/scatterlist: Provide a dedicated function to support table append")
> 
> interacting with commit
> 
>   213d50927763 ("drm/i915/ttm: Introduce a TTM i915 gem object backend")
> 
> from the drm tree.
> 
> I have applied the following merge resolution patch.
> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Mon, 23 Aug 2021 17:46:27 +1000
> Subject: [PATCH] drm/i915/ttm: fix up for "lib/scatterlist: Provide a
>  dedicated function to support tableappend"
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>  drivers/gpu/drm/i915/gem/i915_gem_ttm.c | 9 ++++-----
>  1 file changed, 4 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> index 771eb2963123..d3d95934a047 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
> @@ -382,7 +382,6 @@ i915_ttm_region(struct ttm_device *bdev, int ttm_mem_type)
>  static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
>  {
>  	struct i915_ttm_tt *i915_tt = container_of(ttm, typeof(*i915_tt), ttm);
> -	struct scatterlist *sg;
>  	struct sg_table *st;
>  	int ret;
>  
> @@ -393,13 +392,13 @@ static struct sg_table *i915_ttm_tt_get_st(struct ttm_tt *ttm)
>  	if (!st)
>  		return ERR_PTR(-ENOMEM);
>  
> -	sg = __sg_alloc_table_from_pages
> +	ret = sg_alloc_table_from_pages_segment
>  		(st, ttm->pages, ttm->num_pages, 0,
>  		 (unsigned long)ttm->num_pages << PAGE_SHIFT,
> -		 i915_sg_segment_size(), NULL, 0, GFP_KERNEL);
> -	if (IS_ERR(sg)) {
> +		 i915_sg_segment_size(), GFP_KERNEL);
> +	if (ret) {
>  		kfree(st);
> -		return ERR_CAST(sg);
> +		return ERR_PTR(ret);
>  	}
>  
>  	ret = dma_map_sgtable(i915_tt->dev, st, DMA_BIDIRECTIONAL, 0);

I am now applying this merge fix patch to the merge of the rdma tree
(since that has merged the hmm tree and the drm tree has been merged
upstream).

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2638 bytes --]

Hi all,

On Mon, 12 Oct 2020 15:19:48 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> On Tue, 6 Oct 2020 13:41:20 -0300 Jason Gunthorpe <jgg@nvidia.com> wrote:
> >
> > On Tue, Oct 06, 2020 at 08:35:08PM +1100, Stephen Rothwell wrote:  
> > > Hi all,
> > > 
> > > After merging the hmm tree, today's linux-next build (arm
> > > multi_v7_defconfig) failed like this:
> > > 
> > > 
> > > Caused by commit
> > > 
> > >   07da1223ec93 ("lib/scatterlist: Add support in dynamic allocation of SG table from pages")
> > > 
> > > interacting with commit
> > > 
> > >   707d561f77b5 ("drm: allow limiting the scatter list size.")
> > > 
> > > from the drm tree.
> > > 
> > > I have added the following merge fix patch
> > > 
> > > From: Stephen Rothwell <sfr@canb.auug.org.au>
> > > Date: Tue, 6 Oct 2020 20:22:51 +1100
> > > Subject: [PATCH] lib/scatterlist: merge fix for "drm: allow limiting the
> > >  scatter list size."
> > > 
> > > Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> > >  drivers/gpu/drm/drm_prime.c | 9 ++++++---
> > >  1 file changed, 6 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> > > index 11fe9ff76fd5..83ac901b65a2 100644
> > > +++ b/drivers/gpu/drm/drm_prime.c
> > > @@ -807,6 +807,7 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
> > >  				       struct page **pages, unsigned int nr_pages)
> > >  {
> > >  	struct sg_table *sg = NULL;
> > > +	struct scatterlist *sl;
> > >  	size_t max_segment = 0;
> > >  	int ret;
> > >  
> > > @@ -820,11 +821,13 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
> > >  		max_segment = dma_max_mapping_size(dev->dev);
> > >  	if (max_segment == 0 || max_segment > SCATTERLIST_MAX_SEGMENT)
> > >  		max_segment = SCATTERLIST_MAX_SEGMENT;
> > > -	ret = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
> > > +	sl = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
> > >  					  nr_pages << PAGE_SHIFT,
> > > -					  max_segment, GFP_KERNEL);
> > > -	if (ret)
> > > +					  max_segment, NULL, 0, GFP_KERNEL);
> > > +	if (IS_ERR(sl)) {
> > > +		ret = PTR_ERR(sl);
> > >  		goto out;
> > > +	}
> > >  
> > >  	return sg;
> > >  out:    
> > 
> > This looks OK to me, thanks  
> 
> This merge fix patch is now being applied to the merge of the drm tree
> since the rdma tree (that is merged before the drm tree) has merged the
> hmm tree.

This merge fix patch is now being applied to the merge of the rdma tree
since the Linus has merged the drm tree.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2301 bytes --]

Hi all,

On Tue, 6 Oct 2020 13:41:20 -0300 Jason Gunthorpe <jgg@nvidia.com> wrote:
>
> On Tue, Oct 06, 2020 at 08:35:08PM +1100, Stephen Rothwell wrote:
> > Hi all,
> > 
> > After merging the hmm tree, today's linux-next build (arm
> > multi_v7_defconfig) failed like this:
> > 
> > 
> > Caused by commit
> > 
> >   07da1223ec93 ("lib/scatterlist: Add support in dynamic allocation of SG table from pages")
> > 
> > interacting with commit
> > 
> >   707d561f77b5 ("drm: allow limiting the scatter list size.")
> > 
> > from the drm tree.
> > 
> > I have added the following merge fix patch
> > 
> > From: Stephen Rothwell <sfr@canb.auug.org.au>
> > Date: Tue, 6 Oct 2020 20:22:51 +1100
> > Subject: [PATCH] lib/scatterlist: merge fix for "drm: allow limiting the
> >  scatter list size."
> > 
> > Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> >  drivers/gpu/drm/drm_prime.c | 9 ++++++---
> >  1 file changed, 6 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> > index 11fe9ff76fd5..83ac901b65a2 100644
> > +++ b/drivers/gpu/drm/drm_prime.c
> > @@ -807,6 +807,7 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
> >  				       struct page **pages, unsigned int nr_pages)
> >  {
> >  	struct sg_table *sg = NULL;
> > +	struct scatterlist *sl;
> >  	size_t max_segment = 0;
> >  	int ret;
> >  
> > @@ -820,11 +821,13 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
> >  		max_segment = dma_max_mapping_size(dev->dev);
> >  	if (max_segment == 0 || max_segment > SCATTERLIST_MAX_SEGMENT)
> >  		max_segment = SCATTERLIST_MAX_SEGMENT;
> > -	ret = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
> > +	sl = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
> >  					  nr_pages << PAGE_SHIFT,
> > -					  max_segment, GFP_KERNEL);
> > -	if (ret)
> > +					  max_segment, NULL, 0, GFP_KERNEL);
> > +	if (IS_ERR(sl)) {
> > +		ret = PTR_ERR(sl);
> >  		goto out;
> > +	}
> >  
> >  	return sg;
> >  out:  
> 
> This looks OK to me, thanks

This merge fix patch is now being applied to the merge of the drm tree
since the rdma tree (that is merged before the drm tree) has merged the
hmm tree.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Tue, Oct 06, 2020 at 08:35:08PM +1100, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the hmm tree, today's linux-next build (arm
> multi_v7_defconfig) failed like this:
> 
> 
> Caused by commit
> 
>   07da1223ec93 ("lib/scatterlist: Add support in dynamic allocation of SG table from pages")
> 
> interacting with commit
> 
>   707d561f77b5 ("drm: allow limiting the scatter list size.")
> 
> from the drm tree.
> 
> I have added the following merge fix patch
> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Tue, 6 Oct 2020 20:22:51 +1100
> Subject: [PATCH] lib/scatterlist: merge fix for "drm: allow limiting the
>  scatter list size."
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
>  drivers/gpu/drm/drm_prime.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
> index 11fe9ff76fd5..83ac901b65a2 100644
> +++ b/drivers/gpu/drm/drm_prime.c
> @@ -807,6 +807,7 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
>  				       struct page **pages, unsigned int nr_pages)
>  {
>  	struct sg_table *sg = NULL;
> +	struct scatterlist *sl;
>  	size_t max_segment = 0;
>  	int ret;
>  
> @@ -820,11 +821,13 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
>  		max_segment = dma_max_mapping_size(dev->dev);
>  	if (max_segment == 0 || max_segment > SCATTERLIST_MAX_SEGMENT)
>  		max_segment = SCATTERLIST_MAX_SEGMENT;
> -	ret = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
> +	sl = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
>  					  nr_pages << PAGE_SHIFT,
> -					  max_segment, GFP_KERNEL);
> -	if (ret)
> +					  max_segment, NULL, 0, GFP_KERNEL);
> +	if (IS_ERR(sl)) {
> +		ret = PTR_ERR(sl);
>  		goto out;
> +	}
>  
>  	return sg;
>  out:

This looks OK to me, thanks

Jason

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1826 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (arm
multi_v7_defconfig) failed like this:


Caused by commit

  07da1223ec93 ("lib/scatterlist: Add support in dynamic allocation of SG table from pages")

interacting with commit

  707d561f77b5 ("drm: allow limiting the scatter list size.")

from the drm tree.

I have added the following merge fix patch

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Tue, 6 Oct 2020 20:22:51 +1100
Subject: [PATCH] lib/scatterlist: merge fix for "drm: allow limiting the
 scatter list size."

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/gpu/drm/drm_prime.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 11fe9ff76fd5..83ac901b65a2 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -807,6 +807,7 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
 				       struct page **pages, unsigned int nr_pages)
 {
 	struct sg_table *sg = NULL;
+	struct scatterlist *sl;
 	size_t max_segment = 0;
 	int ret;
 
@@ -820,11 +821,13 @@ struct sg_table *drm_prime_pages_to_sg(struct drm_device *dev,
 		max_segment = dma_max_mapping_size(dev->dev);
 	if (max_segment == 0 || max_segment > SCATTERLIST_MAX_SEGMENT)
 		max_segment = SCATTERLIST_MAX_SEGMENT;
-	ret = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
+	sl = __sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
 					  nr_pages << PAGE_SHIFT,
-					  max_segment, GFP_KERNEL);
-	if (ret)
+					  max_segment, NULL, 0, GFP_KERNEL);
+	if (IS_ERR(sl)) {
+		ret = PTR_ERR(sl);
 		goto out;
+	}
 
 	return sg;
 out:
-- 
2.28.0


I assume that there may be more needed.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Thu, Nov 14, 2019 at 04:34:35PM +1100, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> drivers/infiniband/hw/hfi1/user_exp_rcv.c: In function 'set_rcvarray_entry':
> drivers/infiniband/hw/hfi1/user_exp_rcv.c:768:33: warning: passing argument 2 of 'mmu_interval_notifier_insert' makes pointer from integer without a cast [-Wint-conversion]
>   768 |    &node->notifier, tbuf->vaddr + (pageidx * PAGE_SIZE),
>       |                     ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~
>       |                                 |
>       |                                 long unsigned int
> In file included from include/rdma/ib_verbs.h:59,
>                  from include/rdma/ib_hdrs.h:53,
>                  from drivers/infiniband/hw/hfi1/hfi.h:68,
>                  from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
>                  from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
> include/linux/mmu_notifier.h:295:24: note: expected 'struct mm_struct *' but argument is of type 'long unsigned int'
>   295 |      struct mm_struct *mm, unsigned long start,
>       |      ~~~~~~~~~~~~~~~~~~^~
> drivers/infiniband/hw/hfi1/user_exp_rcv.c:769:26: warning: passing argument 4 of 'mmu_interval_notifier_insert' makes integer from pointer without a cast [-Wint-conversion]
>   769 |    npages * PAGE_SIZE, fd->mm);
>       |                        ~~^~~~
>       |                          |
>       |                          struct mm_struct *
> In file included from include/rdma/ib_verbs.h:59,
>                  from include/rdma/ib_hdrs.h:53,
>                  from drivers/infiniband/hw/hfi1/hfi.h:68,
>                  from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
>                  from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
> include/linux/mmu_notifier.h:296:20: note: expected 'long unsigned int' but argument is of type 'struct mm_struct *'
>   296 |      unsigned long length,
>       |      ~~~~~~~~~~~~~~^~~~~~
> drivers/infiniband/hw/hfi1/user_exp_rcv.c:767:9: error: too few arguments to function 'mmu_interval_notifier_insert'
>   767 |   ret = mmu_interval_notifier_insert(
>       |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> In file included from include/rdma/ib_verbs.h:59,
>                  from include/rdma/ib_hdrs.h:53,
>                  from drivers/infiniband/hw/hfi1/hfi.h:68,
>                  from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
>                  from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
> include/linux/mmu_notifier.h:294:5: note: declared here
>   294 | int mmu_interval_notifier_insert(struct mmu_interval_notifier *mni,
>       |     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> Caused by commit
> 
>   c90dad714405 ("RDMA/hfi1: Use mmu_interval_notifier_insert for user_exp_rcv")
> 
> I have used the hmm tree from next-20191113 for today.

Another case of 0-day missing stuff, it gave this branch an OK :( 

I'll fix it for tomorrow

Thanks,
Jason

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2796 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

drivers/infiniband/hw/hfi1/user_exp_rcv.c: In function 'set_rcvarray_entry':
drivers/infiniband/hw/hfi1/user_exp_rcv.c:768:33: warning: passing argument 2 of 'mmu_interval_notifier_insert' makes pointer from integer without a cast [-Wint-conversion]
  768 |    &node->notifier, tbuf->vaddr + (pageidx * PAGE_SIZE),
      |                     ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~
      |                                 |
      |                                 long unsigned int
In file included from include/rdma/ib_verbs.h:59,
                 from include/rdma/ib_hdrs.h:53,
                 from drivers/infiniband/hw/hfi1/hfi.h:68,
                 from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
                 from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
include/linux/mmu_notifier.h:295:24: note: expected 'struct mm_struct *' but argument is of type 'long unsigned int'
  295 |      struct mm_struct *mm, unsigned long start,
      |      ~~~~~~~~~~~~~~~~~~^~
drivers/infiniband/hw/hfi1/user_exp_rcv.c:769:26: warning: passing argument 4 of 'mmu_interval_notifier_insert' makes integer from pointer without a cast [-Wint-conversion]
  769 |    npages * PAGE_SIZE, fd->mm);
      |                        ~~^~~~
      |                          |
      |                          struct mm_struct *
In file included from include/rdma/ib_verbs.h:59,
                 from include/rdma/ib_hdrs.h:53,
                 from drivers/infiniband/hw/hfi1/hfi.h:68,
                 from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
                 from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
include/linux/mmu_notifier.h:296:20: note: expected 'long unsigned int' but argument is of type 'struct mm_struct *'
  296 |      unsigned long length,
      |      ~~~~~~~~~~~~~~^~~~~~
drivers/infiniband/hw/hfi1/user_exp_rcv.c:767:9: error: too few arguments to function 'mmu_interval_notifier_insert'
  767 |   ret = mmu_interval_notifier_insert(
      |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from include/rdma/ib_verbs.h:59,
                 from include/rdma/ib_hdrs.h:53,
                 from drivers/infiniband/hw/hfi1/hfi.h:68,
                 from drivers/infiniband/hw/hfi1/mmu_rb.h:50,
                 from drivers/infiniband/hw/hfi1/user_exp_rcv.c:50:
include/linux/mmu_notifier.h:294:5: note: declared here
  294 | int mmu_interval_notifier_insert(struct mmu_interval_notifier *mni,
      |     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~

Caused by commit

  c90dad714405 ("RDMA/hfi1: Use mmu_interval_notifier_insert for user_exp_rcv")

I have used the hmm tree from next-20191113 for today.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Dan Williams]

On Mon, Sep 2, 2019 at 10:42 PM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Hi Dan,
>
> On Mon, 2 Sep 2019 22:31:00 -0700 Dan Williams <dan.j.williams@intel.com> wrote:
> >
> > On Mon, Sep 2, 2019 at 4:45 PM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> > >
> > > Hi Jason,
> > >
> > > On Mon, 2 Sep 2019 10:51:41 +0000 Jason Gunthorpe <jgg@mellanox.com> wrote:
> > > >
> > > > On Mon, Sep 02, 2019 at 08:50:17PM +1000, Stephen Rothwell wrote:
> > > >
> > > > > ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > > ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > > ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > > ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > >
> > > > > Caused by commit
> > > > >
> > > > >   126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")
> > > > >
> > > > > I have reverted that commit for today.
> > > >
> > > > Looks like more kconfig trouble, can you send Dan your kconfig? I'll
> > > > drop this patch again
> > > >
> > >
> > > Thanks.  It was just an x86_64 allmodconfig build.  I don't actually
> > > have the .config file (it gets cleaned up, sorry).
> >
> > Strange. x86_64 allmodconfig is certainly a 0day build target. Could
> > this be toolchain dependent?
>
> Possible, I guess.  I am cross compiling on a PowerPC LE host
>
> $ x86_64-linux-gnu-gcc --version
> x86_64-linux-gnu-gcc (Debian 9.2.1-4) 9.2.1 20190821
>
> $ x86_64-linux-gnu-ld --version
> GNU ld (GNU Binutils for Debian) 2.32.51.20190821
>
> It could also be an interaction with something else in linux-next.

I was able to reproduce and it seems to be fixed with this change:

diff --git a/drivers/nvdimm/Makefile b/drivers/nvdimm/Makefile
index 6557e126892f..29203f3d3069 100644
--- a/drivers/nvdimm/Makefile
+++ b/drivers/nvdimm/Makefile
@@ -32,4 +32,4 @@ libnvdimm-$(CONFIG_NVDIMM_KEYS) += security.o

 TOOLS := ../../tools
 TEST_SRC := $(TOOLS)/testing/nvdimm/test
-obj-$(CONFIG_NVDIMM_TEST_BUILD) := $(TEST_SRC)/iomap.o
+obj-$(CONFIG_NVDIMM_TEST_BUILD) += $(TEST_SRC)/iomap.o

I'll send it through one more 0day run before asking Jason to try again.

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1589 bytes --]

Hi Dan,

On Mon, 2 Sep 2019 22:31:00 -0700 Dan Williams <dan.j.williams@intel.com> wrote:
>
> On Mon, Sep 2, 2019 at 4:45 PM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > Hi Jason,
> >
> > On Mon, 2 Sep 2019 10:51:41 +0000 Jason Gunthorpe <jgg@mellanox.com> wrote:  
> > >
> > > On Mon, Sep 02, 2019 at 08:50:17PM +1000, Stephen Rothwell wrote:  
> > >  
> > > > ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
> > > > ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
> > > >
> > > > Caused by commit
> > > >
> > > >   126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")
> > > >
> > > > I have reverted that commit for today.  
> > >
> > > Looks like more kconfig trouble, can you send Dan your kconfig? I'll
> > > drop this patch again
> > >  
> >
> > Thanks.  It was just an x86_64 allmodconfig build.  I don't actually
> > have the .config file (it gets cleaned up, sorry).  
> 
> Strange. x86_64 allmodconfig is certainly a 0day build target. Could
> this be toolchain dependent?

Possible, I guess.  I am cross compiling on a PowerPC LE host

$ x86_64-linux-gnu-gcc --version
x86_64-linux-gnu-gcc (Debian 9.2.1-4) 9.2.1 20190821

$ x86_64-linux-gnu-ld --version
GNU ld (GNU Binutils for Debian) 2.32.51.20190821

It could also be an interaction with something else in linux-next.
-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Dan Williams]

On Mon, Sep 2, 2019 at 4:45 PM Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> Hi Jason,
>
> On Mon, 2 Sep 2019 10:51:41 +0000 Jason Gunthorpe <jgg@mellanox.com> wrote:
> >
> > On Mon, Sep 02, 2019 at 08:50:17PM +1000, Stephen Rothwell wrote:
> > > Hi all,
> >
> > > ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
> > > ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
> > > ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
> > > ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
> > >
> > > Caused by commit
> > >
> > >   126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")
> > >
> > > I have reverted that commit for today.
> >
> > Looks like more kconfig trouble, can you send Dan your kconfig? I'll
> > drop this patch again
> >
>
> Thanks.  It was just an x86_64 allmodconfig build.  I don't actually
> have the .config file (it gets cleaned up, sorry).

Strange. x86_64 allmodconfig is certainly a 0day build target. Could
this be toolchain dependent?

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 906 bytes --]

Hi Jason,

On Mon, 2 Sep 2019 10:51:41 +0000 Jason Gunthorpe <jgg@mellanox.com> wrote:
>
> On Mon, Sep 02, 2019 at 08:50:17PM +1000, Stephen Rothwell wrote:
> > Hi all,  
> 
> > ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
> > ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
> > ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
> > ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
> > 
> > Caused by commit
> > 
> >   126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")
> > 
> > I have reverted that commit for today.
> 
> Looks like more kconfig trouble, can you send Dan your kconfig? I'll
> drop this patch again
>

Thanks.  It was just an x86_64 allmodconfig build.  I don't actually
have the .config file (it gets cleaned up, sorry).
-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Mon, Sep 02, 2019 at 08:50:17PM +1000, Stephen Rothwell wrote:
> Hi all,

> ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
> ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
> ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
> ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
> 
> Caused by commit
> 
>   126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")
> 
> I have reverted that commit for today.
> 

Looks like more kconfig trouble, can you send Dan your kconfig? I'll
drop this patch again

Thanks,
Jason

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 3851 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

ERROR: "__nd_driver_register" [drivers/dax/pmem/dax_pmem_compat.ko] undefined!
ERROR: "nvdimm_namespace_common_probe" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "devm_nsio_disable" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "devm_nsio_enable" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "nvdimm_setup_pfn" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "to_nd_dax" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "to_nd_region" [drivers/dax/pmem/dax_pmem_core.ko] undefined!
ERROR: "__nd_driver_register" [drivers/dax/pmem/dax_pmem.ko] undefined!
ERROR: "nvdimm_cmd_mask" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_blk_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_blk_region_to_dimm" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_bus_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_fletcher64" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_blk_region_set_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_kobj" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_pmem_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_cmd_in_size" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_region_dev" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_bus_register" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_to_bus" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_numa_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_in_overwrite" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_flush" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_region_acquire_lane" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_volatile_region_create" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "devm_nvdimm_memremap" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_region_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nvdimm_bus_dev" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_name" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_blk_memremap_flags" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_bus_unregister" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_cmd_dimm_desc" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_mapping_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_cmd_out_size" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_region_release_lane" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_device_attribute_group" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nd_desc" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_has_flush" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nvdimm_bus" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_bus_check_dimm_count" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nd_region" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "__nvdimm_create" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nvdimm" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_cmd_bus_desc" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_bus_add_badrange" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nd_region_provider_data" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "to_nd_blk_region" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_region_notify" [drivers/acpi/nfit/nfit.ko] undefined!
ERROR: "nvdimm_blk_region_create" [drivers/acpi/nfit/nfit.ko] undefined!

Caused by commit

  126470c8a58b ("libnvdimm: Enable unit test infrastructure compile checks")

I have reverted that commit for today.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Tue, Jul 09, 2019 at 10:21:37AM +1000, Stephen Rothwell wrote:
> Hi all,
> 
> On Mon, 1 Jul 2019 21:08:53 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
> >
> > After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> > failed like this:
> > 
> > mm/hmm.c: In function 'hmm_get_or_create':
> > mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
> >   lockdep_assert_held_exclusive(&mm->mmap_sem);
> >   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >   lockdep_assert_held_once
> > 
> > Caused by commit
> > 
> >   8a9320b7ec5d ("mm/hmm: Simplify hmm_get_or_create and make it reliable")
> > 
> > interacting with commit
> > 
> >   9ffbe8ac05db ("locking/lockdep: Rename lockdep_assert_held_exclusive() -> lockdep_assert_held_write()")
> > 
> > from the tip tree.
> > 
> > I have added the following merge fix.
> > 
> > From: Stephen Rothwell <sfr@canb.auug.org.au>
> > Date: Mon, 1 Jul 2019 21:05:59 +1000
> > Subject: [PATCH] mm/hmm: fixup for "locking/lockdep: Rename
> >  lockdep_assert_held_exclusive() -> lockdep_assert_held_write()"
> > 
> > Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> >  mm/hmm.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/mm/hmm.c b/mm/hmm.c
> > index c1bdcef403ee..2ddbd589b207 100644
> > +++ b/mm/hmm.c
> > @@ -47,7 +47,7 @@ static struct hmm *hmm_get_or_create(struct mm_struct *mm)
> >  {
> >  	struct hmm *hmm;
> >  
> > -	lockdep_assert_held_exclusive(&mm->mmap_sem);
> > +	lockdep_assert_held_write(&mm->mmap_sem);
> >  
> >  	/* Abuse the page_table_lock to also protect mm->hmm. */
> >  	spin_lock(&mm->page_table_lock);
> > @@ -248,7 +248,7 @@ static const struct mmu_notifier_ops hmm_mmu_notifier_ops = {
> >   */
> >  int hmm_mirror_register(struct hmm_mirror *mirror, struct mm_struct *mm)
> >  {
> > -	lockdep_assert_held_exclusive(&mm->mmap_sem);
> > +	lockdep_assert_held_write(&mm->mmap_sem);
> >  
> >  	/* Sanity check */
> >  	if (!mm || !mirror || !mirror->ops)
> 
> I am still getting this conflict (the commit ids may have changed).
> Just a reminder in case you think Linus may need to know.

Ingo already sent the PR to Linus with the function rename, so I will
take care of it.

Thanks,
Jason

Re: linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 2164 bytes --]

Hi all,

On Mon, 1 Jul 2019 21:08:53 +1000 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> mm/hmm.c: In function 'hmm_get_or_create':
> mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
>   lockdep_assert_held_exclusive(&mm->mmap_sem);
>   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   lockdep_assert_held_once
> 
> Caused by commit
> 
>   8a9320b7ec5d ("mm/hmm: Simplify hmm_get_or_create and make it reliable")
> 
> interacting with commit
> 
>   9ffbe8ac05db ("locking/lockdep: Rename lockdep_assert_held_exclusive() -> lockdep_assert_held_write()")
> 
> from the tip tree.
> 
> I have added the following merge fix.
> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Mon, 1 Jul 2019 21:05:59 +1000
> Subject: [PATCH] mm/hmm: fixup for "locking/lockdep: Rename
>  lockdep_assert_held_exclusive() -> lockdep_assert_held_write()"
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
>  mm/hmm.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/hmm.c b/mm/hmm.c
> index c1bdcef403ee..2ddbd589b207 100644
> --- a/mm/hmm.c
> +++ b/mm/hmm.c
> @@ -47,7 +47,7 @@ static struct hmm *hmm_get_or_create(struct mm_struct *mm)
>  {
>  	struct hmm *hmm;
>  
> -	lockdep_assert_held_exclusive(&mm->mmap_sem);
> +	lockdep_assert_held_write(&mm->mmap_sem);
>  
>  	/* Abuse the page_table_lock to also protect mm->hmm. */
>  	spin_lock(&mm->page_table_lock);
> @@ -248,7 +248,7 @@ static const struct mmu_notifier_ops hmm_mmu_notifier_ops = {
>   */
>  int hmm_mirror_register(struct hmm_mirror *mirror, struct mm_struct *mm)
>  {
> -	lockdep_assert_held_exclusive(&mm->mmap_sem);
> +	lockdep_assert_held_write(&mm->mmap_sem);
>  
>  	/* Sanity check */
>  	if (!mm || !mirror || !mirror->ops)

I am still getting this conflict (the commit ids may have changed).
Just a reminder in case you think Linus may need to know.

-- 
Cheers,
Stephen Rothwell

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

Re: linux-next: build failure after merge of the hmm tree

[Alex Deucher]

On Tue, Jul 2, 2019 at 3:24 AM Jason Gunthorpe <jgg@mellanox.com> wrote:
>
> On Mon, Jul 01, 2019 at 09:33:04PM +1000, Stephen Rothwell wrote:
> > Hi all,
> >
> > After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> > failed like this:
> >
> > mm/hmm.c: In function 'hmm_get_or_create':
> > mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
> >   lockdep_assert_held_exclusive(&mm->mmap_sem);
> >   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >   lockdep_assert_held_once
> > drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c: In function 'amdgpu_ttm_tt_get_user_pages':
> > drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:778:28: error: passing argument 2 of 'hmm_range_register' from incompatible pointer type [-Werror=incompatible-pointer-types]
> >   hmm_range_register(range, mm, start,
> >                             ^~
> > In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:35:
> > include/linux/hmm.h:464:29: note: expected 'struct hmm_mirror *' but argument is of type 'struct mm_struct *'
> >           struct hmm_mirror *mirror,
> >           ~~~~~~~~~~~~~~~~~~~^~~~~~
> >
> > Caused by commit
> >
> >   e36acfe6c86d ("mm/hmm: Use hmm_mirror not mm as an argument for hmm_range_register")
> >
> > interacting with commit
> >
> >   66c45500bfdc ("drm/amdgpu: use new HMM APIs and helpers")
> >
> > from the drm tree.
> >
> > All I could do for now was to mark the AMDGPU driver broken.  Please
> > submit a merge for for me (and later Linus) to use.
>
> This is expected, the AMD guys have the resolution for this from when
> they tested hmm.git..
>
> I think we are going to have to merge hmm.git into the amdgpu tree and
> push the resolution forward, as it looks kind of complicated to shift
> onto Linus or DRM.
>
> Probably amdgpu needs to gain a few patches making the hmm_mirror
> visible to amdgpu_ttm.c and then the merge resolution will be simple?
>
> AMD/DRM we have a few days left to decide on how best to handle the
> conflict, thanks.

Philip and Felix have been working on a branch with hmm merged into
drm-next with all the conflicts fixes up.  I'll post it out tomorrow
once I get Felix's latest revisions.

Alex

>
> Regards,
> Jason
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: linux-next: build failure after merge of the hmm tree

[Jason Gunthorpe]

On Mon, Jul 01, 2019 at 09:33:04PM +1000, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> mm/hmm.c: In function 'hmm_get_or_create':
> mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
>   lockdep_assert_held_exclusive(&mm->mmap_sem);
>   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   lockdep_assert_held_once
> drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c: In function 'amdgpu_ttm_tt_get_user_pages':
> drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:778:28: error: passing argument 2 of 'hmm_range_register' from incompatible pointer type [-Werror=incompatible-pointer-types]
>   hmm_range_register(range, mm, start,
>                             ^~
> In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:35:
> include/linux/hmm.h:464:29: note: expected 'struct hmm_mirror *' but argument is of type 'struct mm_struct *'
>           struct hmm_mirror *mirror,
>           ~~~~~~~~~~~~~~~~~~~^~~~~~
> 
> Caused by commit
> 
>   e36acfe6c86d ("mm/hmm: Use hmm_mirror not mm as an argument for hmm_range_register")
> 
> interacting with commit
> 
>   66c45500bfdc ("drm/amdgpu: use new HMM APIs and helpers")
> 
> from the drm tree.
> 
> All I could do for now was to mark the AMDGPU driver broken.  Please
> submit a merge for for me (and later Linus) to use.

This is expected, the AMD guys have the resolution for this from when
they tested hmm.git..

I think we are going to have to merge hmm.git into the amdgpu tree and
push the resolution forward, as it looks kind of complicated to shift
onto Linus or DRM.

Probably amdgpu needs to gain a few patches making the hmm_mirror
visible to amdgpu_ttm.c and then the merge resolution will be simple?

AMD/DRM we have a few days left to decide on how best to handle the
conflict, thanks.

Regards,
Jason

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1391 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

mm/hmm.c: In function 'hmm_get_or_create':
mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
  lockdep_assert_held_exclusive(&mm->mmap_sem);
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  lockdep_assert_held_once
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c: In function 'amdgpu_ttm_tt_get_user_pages':
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:778:28: error: passing argument 2 of 'hmm_range_register' from incompatible pointer type [-Werror=incompatible-pointer-types]
  hmm_range_register(range, mm, start,
                            ^~
In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:35:
include/linux/hmm.h:464:29: note: expected 'struct hmm_mirror *' but argument is of type 'struct mm_struct *'
          struct hmm_mirror *mirror,
          ~~~~~~~~~~~~~~~~~~~^~~~~~

Caused by commit

  e36acfe6c86d ("mm/hmm: Use hmm_mirror not mm as an argument for hmm_range_register")

interacting with commit

  66c45500bfdc ("drm/amdgpu: use new HMM APIs and helpers")

from the drm tree.

All I could do for now was to mark the AMDGPU driver broken.  Please
submit a merge for for me (and later Linus) to use.

-- 
Cheers,
Stephen Rothwell

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

linux-next: build failure after merge of the hmm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1856 bytes --]

Hi all,

After merging the hmm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

mm/hmm.c: In function 'hmm_get_or_create':
mm/hmm.c:50:2: error: implicit declaration of function 'lockdep_assert_held_exclusive'; did you mean 'lockdep_assert_held_once'? [-Werror=implicit-function-declaration]
  lockdep_assert_held_exclusive(&mm->mmap_sem);
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  lockdep_assert_held_once

Caused by commit

  8a9320b7ec5d ("mm/hmm: Simplify hmm_get_or_create and make it reliable")

interacting with commit

  9ffbe8ac05db ("locking/lockdep: Rename lockdep_assert_held_exclusive() -> lockdep_assert_held_write()")

from the tip tree.

I have added the following merge fix.

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 1 Jul 2019 21:05:59 +1000
Subject: [PATCH] mm/hmm: fixup for "locking/lockdep: Rename
 lockdep_assert_held_exclusive() -> lockdep_assert_held_write()"

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 mm/hmm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/mm/hmm.c b/mm/hmm.c
index c1bdcef403ee..2ddbd589b207 100644
--- a/mm/hmm.c
+++ b/mm/hmm.c
@@ -47,7 +47,7 @@ static struct hmm *hmm_get_or_create(struct mm_struct *mm)
 {
 	struct hmm *hmm;
 
-	lockdep_assert_held_exclusive(&mm->mmap_sem);
+	lockdep_assert_held_write(&mm->mmap_sem);
 
 	/* Abuse the page_table_lock to also protect mm->hmm. */
 	spin_lock(&mm->page_table_lock);
@@ -248,7 +248,7 @@ static const struct mmu_notifier_ops hmm_mmu_notifier_ops = {
  */
 int hmm_mirror_register(struct hmm_mirror *mirror, struct mm_struct *mm)
 {
-	lockdep_assert_held_exclusive(&mm->mmap_sem);
+	lockdep_assert_held_write(&mm->mmap_sem);
 
 	/* Sanity check */
 	if (!mm || !mirror || !mirror->ops)
-- 
2.20.1

-- 
Cheers,
Stephen Rothwell

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