[PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

In the following patches, we can convert the calls of kmap_atomic() /
kunmap_atomic() to kmap_local_page() / kunmap_local(), which can
instead do the mapping / unmapping regardless of the context.

With kmap_local_page(), the mapping is per thread, CPU local and not
globally visible.

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
---
Zhao Liu (9):
  drm/i915: Use kmap_local_page() in gem/i915_gem_object.c
  drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c
  drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c
  drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c
  drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_coherency.c
  drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_context.c
  drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c
  drm/i915: Use kmap_local_page() in i915_cmd_parser.c
  drm/i915: Use kmap_local_page() in gem/i915_gem_execbuffer.c

 drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c       | 10 +++++-----
 drivers/gpu/drm/i915/gem/i915_gem_object.c           |  8 +++-----
 drivers/gpu/drm/i915/gem/i915_gem_phys.c             |  8 ++++----
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c            |  6 ++++--
 drivers/gpu/drm/i915/gem/selftests/huge_pages.c      |  6 +++---
 .../gpu/drm/i915/gem/selftests/i915_gem_coherency.c  | 12 ++++--------
 .../gpu/drm/i915/gem/selftests/i915_gem_context.c    |  8 ++++----
 drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c             |  5 +----
 drivers/gpu/drm/i915/i915_cmd_parser.c               |  4 ++--
 9 files changed, 30 insertions(+), 37 deletions(-)

-- 
2.34.1

[PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
need to disable pagefaults and preemption for mapping:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. In drm/i915/gem/i915_gem_object.c, the function
i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range() to
use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
operation is global and any issue with cpu's being added or removed
can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, i915_gem_object_read_from_page_kmap() is a function where
the use of kmap_local_page() in place of kmap_atomic() is correctly
suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

And remove the redundant variable that stores the address of the mapped
page since kunmap_local() can accept any pointer within the page.

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/gem/i915_gem_object.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c
index 369006c5317f..a0072abed75e 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
@@ -413,17 +413,15 @@ void __i915_gem_object_invalidate_frontbuffer(struct drm_i915_gem_object *obj,
 static void
 i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, u64 offset, void *dst, int size)
 {
-	void *src_map;
 	void *src_ptr;
 
-	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >> PAGE_SHIFT));
-
-	src_ptr = src_map + offset_in_page(offset);
+	src_ptr = kmap_local_page(i915_gem_object_get_page(obj, offset >> PAGE_SHIFT))
+	          + offset_in_page(offset);
 	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
 		drm_clflush_virt_range(src_ptr, size);
 	memcpy(dst, src_ptr, size);
 
-	kunmap_atomic(src_map);
+	kunmap_local(src_ptr);
 }
 
 static void
-- 
2.34.1

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Fabio M. De Francesco]

On lunedì 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].
> 
> The main difference between atomic and local mappings is that local
> mappings doesn't disable page faults or preemption.

You are right about about page faults which are never disabled by 
kmap_local_page(). However kmap_atomic might not disable preemption. It 
depends on CONFIG_PREEMPT_RT.

Please refer to how kmap_atomic_prot() works (this function is called by 
kmap_atomic() when kernels have HIGHMEM enabled).

> 
> There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> need to disable pagefaults and preemption for mapping:
> 
> 1. The flush operation is safe for CPU hotplug when preemption is not
> disabled. 

I'm confused here. Why are you talking about CPU hotplug?
In any case, developers should never rely on implicit calls of 
preempt_disable() for the reasons said above. Therefore, flush operations 
should be allowed regardless that kmap_atomic() potential side effect.

> In drm/i915/gem/i915_gem_object.c, the function
> i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range()

If I recall correctly, drm_clflush_virt_range() can always be called with page 
faults and preemption enabled. If so, this is enough to say that the 
conversion is safe. 

Is this code explicitly related to flushing the cache lines before removing / 
adding CPUs? If I recall correctly, there are several other reasons behind the 
need to issue cache lines flushes. Am I wrong about this?

Can you please say more about what I'm missing here?

> to
> use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
> and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
> operation is global and any issue with cpu's being added or removed
> can be handled safely.

Again your main concern is about CPU hotplug.

Even if I'm missing something, do we really need all these details about the 
inner workings of drm_clflush_virt_range()? 

I'm not an expert, so may be that I'm wrong about all I wrote above.

Therefore, can you please elaborate a little more for readers with very little 
knowledge of these kinds of things (like me and perhaps others)?
 
> 2. Any context switch caused by preemption or sleep (pagefault may
> cause sleep) doesn't affect the validity of local mapping.

I'd replace "preemption or sleep" with "preemption and page faults" since 
yourself then added that page faults lead to tasks being put to sleep.  

> Therefore, i915_gem_object_read_from_page_kmap() is a function where
> the use of kmap_local_page() in place of kmap_atomic() is correctly
> suited.
> 
> Convert the calls of kmap_atomic() / kunmap_atomic() to
> kmap_local_page() / kunmap_local().
> 
> And remove the redundant variable that stores the address of the mapped
> page since kunmap_local() can accept any pointer within the page.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> 
> Suggested-by: Dave Hansen <dave.hansen@intel.com>
> Suggested-by: Ira Weiny <ira.weiny@intel.com>
> Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> ---
> Suggested by credits:
>   Dave: Referred to his explanation about cache flush.
>   Ira: Referred to his task document, review comments and explanation about
>        cache flush.
>   Fabio: Referred to his boiler plate commit message.
> ---
>  drivers/gpu/drm/i915/gem/i915_gem_object.c | 8 +++-----
>  1 file changed, 3 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> b/drivers/gpu/drm/i915/gem/i915_gem_object.c index 
369006c5317f..a0072abed75e 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
> @@ -413,17 +413,15 @@ void __i915_gem_object_invalidate_frontbuffer(struct
> drm_i915_gem_object *obj, static void
>  i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, u64 
offset, void
> *dst, int size) {
> -	void *src_map;
>  	void *src_ptr;
> 
> -	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >> 
PAGE_SHIFT));
> -
> -	src_ptr = src_map + offset_in_page(offset);
> +	src_ptr = kmap_local_page(i915_gem_object_get_page(obj, offset >> 
PAGE_SHIFT))
> +	          + offset_in_page(offset);
>  	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
>  		drm_clflush_virt_range(src_ptr, size);
>  	memcpy(dst, src_ptr, size);
> 
> -	kunmap_atomic(src_map);
> +	kunmap_local(src_ptr);
>  }
> 
>  static void

The changes look good, but I'd like to better understand the commit message.

Thanks,

Fabio 

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Ira Weiny]

On Sat, Oct 29, 2022 at 01:17:03PM +0200, Fabio M. De Francesco wrote:
> On lunedì 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> > From: Zhao Liu <zhao1.liu@intel.com>
> > 
> > The use of kmap_atomic() is being deprecated in favor of
> > kmap_local_page()[1].
> > 
> > The main difference between atomic and local mappings is that local
> > mappings doesn't disable page faults or preemption.
> 
> You are right about about page faults which are never disabled by 
> kmap_local_page(). However kmap_atomic might not disable preemption. It 
> depends on CONFIG_PREEMPT_RT.
> 
> Please refer to how kmap_atomic_prot() works (this function is called by 
> kmap_atomic() when kernels have HIGHMEM enabled).
> 
> > 
> > There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> > need to disable pagefaults and preemption for mapping:
> > 
> > 1. The flush operation is safe for CPU hotplug when preemption is not
> > disabled. 
> 
> I'm confused here. Why are you talking about CPU hotplug?

I agree with Fabio here.  I'm not making the connection between cpu hotplug and
this code path.

Ira

> In any case, developers should never rely on implicit calls of 
> preempt_disable() for the reasons said above. Therefore, flush operations 
> should be allowed regardless that kmap_atomic() potential side effect.
> 
> > In drm/i915/gem/i915_gem_object.c, the function
> > i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range()
> 
> If I recall correctly, drm_clflush_virt_range() can always be called with page 
> faults and preemption enabled. If so, this is enough to say that the 
> conversion is safe. 
> 
> Is this code explicitly related to flushing the cache lines before removing / 
> adding CPUs? If I recall correctly, there are several other reasons behind the 
> need to issue cache lines flushes. Am I wrong about this?
> 
> Can you please say more about what I'm missing here?
> 
> > to
> > use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
> > and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
> > operation is global and any issue with cpu's being added or removed
> > can be handled safely.
> 
> Again your main concern is about CPU hotplug.
> 
> Even if I'm missing something, do we really need all these details about the 
> inner workings of drm_clflush_virt_range()? 
> 
> I'm not an expert, so may be that I'm wrong about all I wrote above.
> 
> Therefore, can you please elaborate a little more for readers with very little 
> knowledge of these kinds of things (like me and perhaps others)?
>  
> > 2. Any context switch caused by preemption or sleep (pagefault may
> > cause sleep) doesn't affect the validity of local mapping.
> 
> I'd replace "preemption or sleep" with "preemption and page faults" since 
> yourself then added that page faults lead to tasks being put to sleep.  
> 
> > Therefore, i915_gem_object_read_from_page_kmap() is a function where
> > the use of kmap_local_page() in place of kmap_atomic() is correctly
> > suited.
> > 
> > Convert the calls of kmap_atomic() / kunmap_atomic() to
> > kmap_local_page() / kunmap_local().
> > 
> > And remove the redundant variable that stores the address of the mapped
> > page since kunmap_local() can accept any pointer within the page.
> > 
> > [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> > 
> > Suggested-by: Dave Hansen <dave.hansen@intel.com>
> > Suggested-by: Ira Weiny <ira.weiny@intel.com>
> > Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> > ---
> > Suggested by credits:
> >   Dave: Referred to his explanation about cache flush.
> >   Ira: Referred to his task document, review comments and explanation about
> >        cache flush.
> >   Fabio: Referred to his boiler plate commit message.
> > ---
> >  drivers/gpu/drm/i915/gem/i915_gem_object.c | 8 +++-----
> >  1 file changed, 3 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > b/drivers/gpu/drm/i915/gem/i915_gem_object.c index 
> 369006c5317f..a0072abed75e 100644
> > --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > @@ -413,17 +413,15 @@ void __i915_gem_object_invalidate_frontbuffer(struct
> > drm_i915_gem_object *obj, static void
> >  i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, u64 
> offset, void
> > *dst, int size) {
> > -	void *src_map;
> >  	void *src_ptr;
> > 
> > -	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >> 
> PAGE_SHIFT));
> > -
> > -	src_ptr = src_map + offset_in_page(offset);
> > +	src_ptr = kmap_local_page(i915_gem_object_get_page(obj, offset >> 
> PAGE_SHIFT))
> > +	          + offset_in_page(offset);
> >  	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
> >  		drm_clflush_virt_range(src_ptr, size);
> >  	memcpy(dst, src_ptr, size);
> > 
> > -	kunmap_atomic(src_map);
> > +	kunmap_local(src_ptr);
> >  }
> > 
> >  static void
> 
> The changes look good, but I'd like to better understand the commit message.
> 
> Thanks,
> 
> Fabio 
> 
> 

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Fabio M. De Francesco]

On giovedì 3 novembre 2022 17:51:23 CET Ira Weiny wrote:
> On Sat, Oct 29, 2022 at 01:17:03PM +0200, Fabio M. De Francesco wrote:
> > On lunedì 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> > > From: Zhao Liu <zhao1.liu@intel.com>
> > > 
> > > The use of kmap_atomic() is being deprecated in favor of
> > > kmap_local_page()[1].
> > > 
> > > The main difference between atomic and local mappings is that local
> > > mappings doesn't disable page faults or preemption.
> > 
> > You are right about about page faults which are never disabled by
> > kmap_local_page(). However kmap_atomic might not disable preemption. It
> > depends on CONFIG_PREEMPT_RT.
> > 
> > Please refer to how kmap_atomic_prot() works (this function is called by
> > kmap_atomic() when kernels have HIGHMEM enabled).
> > 
> > > There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> > > need to disable pagefaults and preemption for mapping:
> > > 
> > > 1. The flush operation is safe for CPU hotplug when preemption is not
> > > disabled.
> > 
> > I'm confused here. Why are you talking about CPU hotplug?
> 
> I agree with Fabio here.  I'm not making the connection between cpu hotplug 
and
> this code path.
> 
> Ira

@Zhao,

I'd like to add that I was about to put my reviewed-by tag. The other things I 
objected are minor nits. Please just clarify this connection.

Your code is good and deserves to be applied.

Fabio

> 
> > In any case, developers should never rely on implicit calls of
> > preempt_disable() for the reasons said above. Therefore, flush operations
> > should be allowed regardless that kmap_atomic() potential side effect.
> > 
> > > In drm/i915/gem/i915_gem_object.c, the function
> > > i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range()
> > 
> > If I recall correctly, drm_clflush_virt_range() can always be called with 
page
> > faults and preemption enabled. If so, this is enough to say that the
> > conversion is safe.
> > 
> > Is this code explicitly related to flushing the cache lines before 
removing /
> > adding CPUs? If I recall correctly, there are several other reasons behind 
the
> > need to issue cache lines flushes. Am I wrong about this?
> > 
> > Can you please say more about what I'm missing here?
> > 
> > > to
> > > use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
> > > and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
> > > operation is global and any issue with cpu's being added or removed
> > > can be handled safely.
> > 
> > Again your main concern is about CPU hotplug.
> > 
> > Even if I'm missing something, do we really need all these details about 
the
> > inner workings of drm_clflush_virt_range()?
> > 
> > I'm not an expert, so may be that I'm wrong about all I wrote above.
> > 
> > Therefore, can you please elaborate a little more for readers with very 
little
> > knowledge of these kinds of things (like me and perhaps others)?
> > 
> > > 2. Any context switch caused by preemption or sleep (pagefault may
> > > cause sleep) doesn't affect the validity of local mapping.
> > 
> > I'd replace "preemption or sleep" with "preemption and page faults" since
> > yourself then added that page faults lead to tasks being put to sleep.
> > 
> > > Therefore, i915_gem_object_read_from_page_kmap() is a function where
> > > the use of kmap_local_page() in place of kmap_atomic() is correctly
> > > suited.
> > > 
> > > Convert the calls of kmap_atomic() / kunmap_atomic() to
> > > kmap_local_page() / kunmap_local().
> > > 
> > > And remove the redundant variable that stores the address of the mapped
> > > page since kunmap_local() can accept any pointer within the page.
> > > 
> > > [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> > > 
> > > Suggested-by: Dave Hansen <dave.hansen@intel.com>
> > > Suggested-by: Ira Weiny <ira.weiny@intel.com>
> > > Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > > Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> > > ---
> > > 
> > > Suggested by credits:
> > >   Dave: Referred to his explanation about cache flush.
> > >   Ira: Referred to his task document, review comments and explanation 
about
> > >   
> > >        cache flush.
> > >   
> > >   Fabio: Referred to his boiler plate commit message.
> > > 
> > > ---
> > > 
> > >  drivers/gpu/drm/i915/gem/i915_gem_object.c | 8 +++-----
> > >  1 file changed, 3 insertions(+), 5 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > > b/drivers/gpu/drm/i915/gem/i915_gem_object.c index
> > 
> > 369006c5317f..a0072abed75e 100644
> > 
> > > --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > > @@ -413,17 +413,15 @@ void 
__i915_gem_object_invalidate_frontbuffer(struct
> > > drm_i915_gem_object *obj, static void
> > > 
> > >  i915_gem_object_read_from_page_kmap(struct drm_i915_gem_object *obj, 
u64
> > 
> > offset, void
> > 
> > > *dst, int size) {
> > > -	void *src_map;
> > > 
> > >  	void *src_ptr;
> > > 
> > > -	src_map = kmap_atomic(i915_gem_object_get_page(obj, offset >>
> > 
> > PAGE_SHIFT));
> > 
> > > -
> > > -	src_ptr = src_map + offset_in_page(offset);
> > > +	src_ptr = kmap_local_page(i915_gem_object_get_page(obj, offset >>
> > 
> > PAGE_SHIFT))
> > 
> > > +	          + offset_in_page(offset);
> > > 
> > >  	if (!(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
> > >  	
> > >  		drm_clflush_virt_range(src_ptr, size);
> > >  	
> > >  	memcpy(dst, src_ptr, size);
> > > 
> > > -	kunmap_atomic(src_map);
> > > +	kunmap_local(src_ptr);
> > > 
> > >  }
> > >  
> > >  static void
> > 
> > The changes look good, but I'd like to better understand the commit 
message.
> > 
> > Thanks,
> > 
> > Fabio

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Zhao Liu]

On Thu, Nov 03, 2022 at 08:22:04PM +0100, Fabio M. De Francesco wrote:
> Date: Thu, 03 Nov 2022 20:22:04 +0100
> From: "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
> Subject: Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in
>  gem/i915_gem_object.c
> 
> On gioved? 3 novembre 2022 17:51:23 CET Ira Weiny wrote:
> > On Sat, Oct 29, 2022 at 01:17:03PM +0200, Fabio M. De Francesco wrote:
> > > On luned? 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> > > > From: Zhao Liu <zhao1.liu@intel.com>
> > > > 
> > > > The use of kmap_atomic() is being deprecated in favor of
> > > > kmap_local_page()[1].
> > > > 
> > > > The main difference between atomic and local mappings is that local
> > > > mappings doesn't disable page faults or preemption.
> > > 
> > > You are right about about page faults which are never disabled by
> > > kmap_local_page(). However kmap_atomic might not disable preemption. It
> > > depends on CONFIG_PREEMPT_RT.
> > > 
> > > Please refer to how kmap_atomic_prot() works (this function is called by
> > > kmap_atomic() when kernels have HIGHMEM enabled).
> > > 
> > > > There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> > > > need to disable pagefaults and preemption for mapping:
> > > > 
> > > > 1. The flush operation is safe for CPU hotplug when preemption is not
> > > > disabled.
> > > 
> > > I'm confused here. Why are you talking about CPU hotplug?
> > 
> > I agree with Fabio here.  I'm not making the connection between cpu hotplug 
> and
> > this code path.
> > 
> > Ira
> 
> @Zhao,
> 
> I'd like to add that I was about to put my reviewed-by tag. The other things I 
> objected are minor nits. Please just clarify this connection.

Thanks Fabio for your comments! Sorry I missed the mails that day. This connection
is my misunderstanding. Other thoughts please refer to my reply to your first email
in this thread.

Thanks,
Zhao

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Zhao Liu]

On Thu, Nov 03, 2022 at 09:51:23AM -0700, Ira Weiny wrote:
> Date: Thu, 3 Nov 2022 09:51:23 -0700
> From: Ira Weiny <ira.weiny@intel.com>
> Subject: Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in
>  gem/i915_gem_object.c
> 
> On Sat, Oct 29, 2022 at 01:17:03PM +0200, Fabio M. De Francesco wrote:
> > On luned? 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> > > From: Zhao Liu <zhao1.liu@intel.com>
> > > 
> > > The use of kmap_atomic() is being deprecated in favor of
> > > kmap_local_page()[1].
> > > 
> > > The main difference between atomic and local mappings is that local
> > > mappings doesn't disable page faults or preemption.
> > 
> > You are right about about page faults which are never disabled by 
> > kmap_local_page(). However kmap_atomic might not disable preemption. It 
> > depends on CONFIG_PREEMPT_RT.
> > 
> > Please refer to how kmap_atomic_prot() works (this function is called by 
> > kmap_atomic() when kernels have HIGHMEM enabled).
> > 
> > > 
> > > There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> > > need to disable pagefaults and preemption for mapping:
> > > 
> > > 1. The flush operation is safe for CPU hotplug when preemption is not
> > > disabled. 
> > 
> > I'm confused here. Why are you talking about CPU hotplug?
> 
> I agree with Fabio here.  I'm not making the connection between cpu hotplug and
> this code path.

Sorry, my misunderstanding. Will delete this wrong explanation.

Thanks,
Zhao

Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in gem/i915_gem_object.c

[Zhao Liu]

On Sat, Oct 29, 2022 at 01:17:03PM +0200, Fabio M. De Francesco wrote:
> Date: Sat, 29 Oct 2022 13:17:03 +0200
> From: "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
> Subject: Re: [PATCH 1/9] drm/i915: Use kmap_local_page() in
>  gem/i915_gem_object.c
> 
> On luned? 17 ottobre 2022 11:37:17 CEST Zhao Liu wrote:
> > From: Zhao Liu <zhao1.liu@intel.com>
> > 
> > The use of kmap_atomic() is being deprecated in favor of
> > kmap_local_page()[1].
> > 
> > The main difference between atomic and local mappings is that local
> > mappings doesn't disable page faults or preemption.
> 
> You are right about about page faults which are never disabled by 
> kmap_local_page(). However kmap_atomic might not disable preemption. It 
> depends on CONFIG_PREEMPT_RT.
> 
> Please refer to how kmap_atomic_prot() works (this function is called by 
> kmap_atomic() when kernels have HIGHMEM enabled).

Yes, there is some ambiguity here. What about "The main difference between
atomic and local mappings is that local mappings never disable page faults
or preemption"?

> 
> > 
> > There're 2 reasons why i915_gem_object_read_from_page_kmap() doesn't
> > need to disable pagefaults and preemption for mapping:
> > 
> > 1. The flush operation is safe for CPU hotplug when preemption is not
> > disabled. 
> 
> I'm confused here. Why are you talking about CPU hotplug?
> In any case, developers should never rely on implicit calls of 
> preempt_disable() for the reasons said above. Therefore, flush operations 
> should be allowed regardless that kmap_atomic() potential side effect.

Sorry, it's my fault, my misunderstanding about the connection between hotplug
and flush here. When mapping exists, the cpu cannot be unplugged via CPU-hotplug.
But whether plug or unplug, it has nothing to do with flush. I will delete this
wrong description.

My initial consideration is that this interface of flush may require an atomic
context, so I want to explain more from the details of its implementation
that cache consistency can be guaranteed without atomic context. Is this
consideration redundant?
Also, do I need to state that migration is still ok for this flush interface
here (since __kmap_local_page_prot() doesn't always disable migration)?

> > In drm/i915/gem/i915_gem_object.c, the function
> > i915_gem_object_read_from_page_kmap() calls drm_clflush_virt_range()
> 
> If I recall correctly, drm_clflush_virt_range() can always be called with page 
> faults and preemption enabled. If so, this is enough to say that the 
> conversion is safe. 
> 
> Is this code explicitly related to flushing the cache lines before removing / 
> adding CPUs? If I recall correctly, there are several other reasons behind the 
> need to issue cache lines flushes. Am I wrong about this?
> 
> Can you please say more about what I'm missing here?
> 
> > to
> > use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
> > and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
> > operation is global and any issue with cpu's being added or removed
> > can be handled safely.
> 
> Again your main concern is about CPU hotplug.
> 
> Even if I'm missing something, do we really need all these details about the 
> inner workings of drm_clflush_virt_range()? 
> 
> I'm not an expert, so may be that I'm wrong about all I wrote above.
> 
> Therefore, can you please elaborate a little more for readers with very little 
> knowledge of these kinds of things (like me and perhaps others)?
>  
> > 2. Any context switch caused by preemption or sleep (pagefault may
> > cause sleep) doesn't affect the validity of local mapping.
> 
> I'd replace "preemption or sleep" with "preemption and page faults" since 
> yourself then added that page faults lead to tasks being put to sleep.  

Thanks, good advice.

Zhao

[PATCH 2/9] drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gem/i915_gem_phys.c, the functions
i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
don't need to disable pagefaults and preemption for mapping because of
these 2 reasons:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. In drm/i915/gem/i915_gem_object.c, the functions
i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush.
Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in
drm_clflush_virt_range(), the flush operation is global and any issue
with cpu's being added or removed can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, i915_gem_object_get_pages_phys() and
i915_gem_object_put_pages_phys() are two functions where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/gem/i915_gem_phys.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
index 0d0e46dae559..d602ba19ecb2 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
@@ -66,10 +66,10 @@ static int i915_gem_object_get_pages_phys(struct drm_i915_gem_object *obj)
 		if (IS_ERR(page))
 			goto err_st;
 
-		src = kmap_atomic(page);
+		src = kmap_local_page(page);
 		memcpy(dst, src, PAGE_SIZE);
 		drm_clflush_virt_range(dst, PAGE_SIZE);
-		kunmap_atomic(src);
+		kunmap_local(src);
 
 		put_page(page);
 		dst += PAGE_SIZE;
@@ -114,10 +114,10 @@ i915_gem_object_put_pages_phys(struct drm_i915_gem_object *obj,
 			if (IS_ERR(page))
 				continue;
 
-			dst = kmap_atomic(page);
+			dst = kmap_local_page(page);
 			drm_clflush_virt_range(src, PAGE_SIZE);
 			memcpy(dst, src, PAGE_SIZE);
-			kunmap_atomic(dst);
+			kunmap_local(dst);
 
 			set_page_dirty(page);
 			if (obj->mm.madv == I915_MADV_WILLNEED)
-- 
2.34.1

Re: [PATCH 2/9] drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c

[Fabio M. De Francesco]

On lunedì 17 ottobre 2022 11:37:18 CEST Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].
> 
> The main difference between atomic and local mappings is that local
> mappings doesn't disable page faults or preemption.
> 
> In drm/i915/gem/i915_gem_phys.c, the functions
> i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
> don't need to disable pagefaults and preemption for mapping because of
> these 2 reasons:
> 
> 1. The flush operation is safe for CPU hotplug when preemption is not
> disabled. In drm/i915/gem/i915_gem_object.c, the functions
> i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
> calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush.
> Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in
> drm_clflush_virt_range(), the flush operation is global and any issue
> with cpu's being added or removed can be handled safely.
> 
> 2. Any context switch caused by preemption or sleep (pagefault may
> cause sleep) doesn't affect the validity of local mapping.
> 
> Therefore, i915_gem_object_get_pages_phys() and
> i915_gem_object_put_pages_phys() are two functions where the use of
> kmap_local_page() in place of kmap_atomic() is correctly suited.
> 
> Convert the calls of kmap_atomic() / kunmap_atomic() to
> kmap_local_page() / kunmap_local().
> 

I have here the same questions as in 1/9.

> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> 
> Suggested-by: Dave Hansen <dave.hansen@intel.com>
> Suggested-by: Ira Weiny <ira.weiny@intel.com>
> Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> ---
> Suggested by credits:
>   Dave: Referred to his explanation about cache flush.
>   Ira: Referred to his task document, review comments and explanation about
>        cache flush.
>   Fabio: Referred to his boiler plate commit message.
> ---
>  drivers/gpu/drm/i915/gem/i915_gem_phys.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> b/drivers/gpu/drm/i915/gem/i915_gem_phys.c index 0d0e46dae559..d602ba19ecb2 
100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> @@ -66,10 +66,10 @@ static int i915_gem_object_get_pages_phys(struct 
drm_i915_gem_object
> *obj) if (IS_ERR(page))
>  			goto err_st;
> 
> -		src = kmap_atomic(page);
> +		src = kmap_local_page(page);
>  		memcpy(dst, src, PAGE_SIZE);
>  		drm_clflush_virt_range(dst, PAGE_SIZE);
> -		kunmap_atomic(src);
> +		kunmap_local(src);

Please use memcpy_from_page() instead of open coding mapping + memcpy() + 
unmapping.

> 
>  		put_page(page);
>  		dst += PAGE_SIZE;
> @@ -114,10 +114,10 @@ i915_gem_object_put_pages_phys(struct 
drm_i915_gem_object *obj,
>  			if (IS_ERR(page))
>  				continue;
> 
> -			dst = kmap_atomic(page);
> +			dst = kmap_local_page(page);
>  			drm_clflush_virt_range(src, PAGE_SIZE);
>  			memcpy(dst, src, PAGE_SIZE);
> -			kunmap_atomic(dst);
> +			kunmap_local(dst);

For the same reasons said above, memcpy_to_page() should be used here and 
avoid open coding of three functions.

Using those helpers forces you to move drm_clflush_virt_range() out of the 
mapping / un-mapping region. I may be wrong, however I'm pretty sure that the 
relative positions of each of those call sites is something that cannot be 
randomly chosen.

Thanks,

Fabio

> 
>  			set_page_dirty(page);
>  			if (obj->mm.madv == I915_MADV_WILLNEED)

Re: [PATCH 2/9] drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c

[Zhao Liu]

On Sat, Oct 29, 2022 at 03:32:08PM +0200, Fabio M. De Francesco wrote:
> Date: Sat, 29 Oct 2022 15:32:08 +0200
> From: "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
> Subject: Re: [PATCH 2/9] drm/i915: Use kmap_local_page() in
>  gem/i915_gem_pyhs.c
> 
> On luned? 17 ottobre 2022 11:37:18 CEST Zhao Liu wrote:
> > From: Zhao Liu <zhao1.liu@intel.com>
> > 
> > The use of kmap_atomic() is being deprecated in favor of
> > kmap_local_page()[1].
> > 
> > The main difference between atomic and local mappings is that local
> > mappings doesn't disable page faults or preemption.
> > 
> > In drm/i915/gem/i915_gem_phys.c, the functions
> > i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
> > don't need to disable pagefaults and preemption for mapping because of
> > these 2 reasons:
> > 
> > 1. The flush operation is safe for CPU hotplug when preemption is not
> > disabled. In drm/i915/gem/i915_gem_object.c, the functions
> > i915_gem_object_get_pages_phys() and i915_gem_object_put_pages_phys()
> > calls drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush.
> > Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu in
> > drm_clflush_virt_range(), the flush operation is global and any issue
> > with cpu's being added or removed can be handled safely.
> > 
> > 2. Any context switch caused by preemption or sleep (pagefault may
> > cause sleep) doesn't affect the validity of local mapping.
> > 
> > Therefore, i915_gem_object_get_pages_phys() and
> > i915_gem_object_put_pages_phys() are two functions where the use of
> > kmap_local_page() in place of kmap_atomic() is correctly suited.
> > 
> > Convert the calls of kmap_atomic() / kunmap_atomic() to
> > kmap_local_page() / kunmap_local().
> > 
> 
> I have here the same questions as in 1/9.
> 
> > [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> > 
> > Suggested-by: Dave Hansen <dave.hansen@intel.com>
> > Suggested-by: Ira Weiny <ira.weiny@intel.com>
> > Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> > Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> > ---
> > Suggested by credits:
> >   Dave: Referred to his explanation about cache flush.
> >   Ira: Referred to his task document, review comments and explanation about
> >        cache flush.
> >   Fabio: Referred to his boiler plate commit message.
> > ---
> >  drivers/gpu/drm/i915/gem/i915_gem_phys.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> > b/drivers/gpu/drm/i915/gem/i915_gem_phys.c index 0d0e46dae559..d602ba19ecb2 
> 100644
> > --- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> > +++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c
> > @@ -66,10 +66,10 @@ static int i915_gem_object_get_pages_phys(struct 
> drm_i915_gem_object
> > *obj) if (IS_ERR(page))
> >  			goto err_st;
> > 
> > -		src = kmap_atomic(page);
> > +		src = kmap_local_page(page);
> >  		memcpy(dst, src, PAGE_SIZE);
> >  		drm_clflush_virt_range(dst, PAGE_SIZE);
> > -		kunmap_atomic(src);
> > +		kunmap_local(src);
> 
> Please use memcpy_from_page() instead of open coding mapping + memcpy() + 
> unmapping.

Ok.

> 
> > 
> >  		put_page(page);
> >  		dst += PAGE_SIZE;
> > @@ -114,10 +114,10 @@ i915_gem_object_put_pages_phys(struct 
> drm_i915_gem_object *obj,
> >  			if (IS_ERR(page))
> >  				continue;
> > 
> > -			dst = kmap_atomic(page);
> > +			dst = kmap_local_page(page);
> >  			drm_clflush_virt_range(src, PAGE_SIZE);
> >  			memcpy(dst, src, PAGE_SIZE);
> > -			kunmap_atomic(dst);
> > +			kunmap_local(dst);
> 
> For the same reasons said above, memcpy_to_page() should be used here and 
> avoid open coding of three functions.
> 
> Using those helpers forces you to move drm_clflush_virt_range() out of the 
> mapping / un-mapping region. I may be wrong, however I'm pretty sure that the 
> relative positions of each of those call sites is something that cannot be 
> randomly chosen.

I agree. Will use memcpy_to_page().

Thanks,
Zhao

> 
> Thanks,
> 
> Fabio
> 
> > 
> >  			set_page_dirty(page);
> >  			if (obj->mm.madv == I915_MADV_WILLNEED)
> 
> 
> 

[PATCH 3/9] drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gem/i915_gem_shmem.c, the function shmem_pwrite() need to
disable pagefault to eliminate the potential recursion fault[2]. But
here __copy_from_user_inatomic() doesn't need to disable preemption and
local mapping is valid for sched in/out.

So it can use kmap_local_page() / kunmap_local() with
pagefault_disable() / pagefault_enable() to replace atomic mapping.

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
[2]: https://patchwork.freedesktop.org/patch/295840/

Suggested-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Ira: Referred to his suggestions about keeping pagefault_disable().
---
 drivers/gpu/drm/i915/gem/i915_gem_shmem.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
index f42ca1179f37..e279a3e30c02 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
@@ -472,11 +472,13 @@ shmem_pwrite(struct drm_i915_gem_object *obj,
 		if (err < 0)
 			return err;
 
-		vaddr = kmap_atomic(page);
+		vaddr = kmap_local_page(page);
+		pagefault_disable();
 		unwritten = __copy_from_user_inatomic(vaddr + pg,
 						      user_data,
 						      len);
-		kunmap_atomic(vaddr);
+		pagefault_enable();
+		kunmap_local(vaddr);
 
 		err = aops->write_end(obj->base.filp, mapping, offset, len,
 				      len - unwritten, page, data);
-- 
2.34.1

Re: [PATCH 3/9] drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c

[Ira Weiny]

On Mon, Oct 17, 2022 at 05:37:19PM +0800, Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].
> 
> The main difference between atomic and local mappings is that local
> mappings doesn't disable page faults or preemption.
> 
> In drm/i915/gem/i915_gem_shmem.c, the function shmem_pwrite() need to
> disable pagefault to eliminate the potential recursion fault[2]. But
> here __copy_from_user_inatomic() doesn't need to disable preemption and
> local mapping is valid for sched in/out.
> 
> So it can use kmap_local_page() / kunmap_local() with
> pagefault_disable() / pagefault_enable() to replace atomic mapping.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> [2]: https://patchwork.freedesktop.org/patch/295840/
> 
> Suggested-by: Ira Weiny <ira.weiny@intel.com>

Reviewed-by: Ira Weiny <ira.weiny@intel.com>

> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> ---
> Suggested by credits:
>   Ira: Referred to his suggestions about keeping pagefault_disable().
> ---
>  drivers/gpu/drm/i915/gem/i915_gem_shmem.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> index f42ca1179f37..e279a3e30c02 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> @@ -472,11 +472,13 @@ shmem_pwrite(struct drm_i915_gem_object *obj,
>  		if (err < 0)
>  			return err;
>  
> -		vaddr = kmap_atomic(page);
> +		vaddr = kmap_local_page(page);
> +		pagefault_disable();
>  		unwritten = __copy_from_user_inatomic(vaddr + pg,
>  						      user_data,
>  						      len);
> -		kunmap_atomic(vaddr);
> +		pagefault_enable();
> +		kunmap_local(vaddr);
>  
>  		err = aops->write_end(obj->base.filp, mapping, offset, len,
>  				      len - unwritten, page, data);
> -- 
> 2.34.1
> 

Re: [PATCH 3/9] drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c

[Fabio M. De Francesco]

On lunedì 17 ottobre 2022 11:37:19 CET Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].
> 
> The main difference between atomic and local mappings is that local
> mappings doesn't disable page faults or preemption.
> 
> In drm/i915/gem/i915_gem_shmem.c, the function shmem_pwrite() need to
> disable pagefault to eliminate the potential recursion fault[2]. But
> here __copy_from_user_inatomic() doesn't need to disable preemption and
> local mapping is valid for sched in/out.
> So it can use kmap_local_page() / kunmap_local() with
> pagefault_disable() / pagefault_enable() to replace atomic mapping.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> [2]: https://patchwork.freedesktop.org/patch/295840/
> 
> Suggested-by: Ira Weiny <ira.weiny@intel.com>
> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> ---
> Suggested by credits:
>   Ira: Referred to his suggestions about keeping pagefault_disable().
> ---
>  drivers/gpu/drm/i915/gem/i915_gem_shmem.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)

The commit message explains clearly and the changes to the code look good. The 
necessity to reuse pagefault_disable() / pagefault_enable() from the main 
kmap_atomic() side effect is a nice catch.

Reviewed-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>

Thanks!

> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c index f42ca1179f37..e279a3e30c02 
100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_shmem.c
> @@ -472,11 +472,13 @@ shmem_pwrite(struct drm_i915_gem_object *obj,
>  		if (err < 0)
>  			return err;
> 
> -		vaddr = kmap_atomic(page);
> +		vaddr = kmap_local_page(page);
> +		pagefault_disable();
>  		unwritten = __copy_from_user_inatomic(vaddr + pg,
>  						      user_data,
>  						      len);
> -		kunmap_atomic(vaddr);
> +		pagefault_enable();
> +		kunmap_local(vaddr);
> 
>  		err = aops->write_end(obj->base.filp, mapping, offset, 
len,
>  				      len - unwritten, page, data);

[PATCH 4/9] drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gem/selftests/huge_pages.c, function __cpu_check_shmem()
mainly uses mapping to flush cache and check the value. There're 2
reasons why __cpu_check_shmem() doesn't need to disable pagefaults
and preemption for mapping:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. Function __cpu_check_shmem() calls drm_clflush_virt_range()
to use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
operation is global and any issue with cpu's being added or removed
can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, __cpu_check_shmem() is a function where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/gem/selftests/huge_pages.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
index c570cf780079..6f4efe905105 100644
--- a/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
+++ b/drivers/gpu/drm/i915/gem/selftests/huge_pages.c
@@ -1022,7 +1022,7 @@ __cpu_check_shmem(struct drm_i915_gem_object *obj, u32 dword, u32 val)
 		goto err_unlock;
 
 	for (n = 0; n < obj->base.size >> PAGE_SHIFT; ++n) {
-		u32 *ptr = kmap_atomic(i915_gem_object_get_page(obj, n));
+		u32 *ptr = kmap_local_page(i915_gem_object_get_page(obj, n));
 
 		if (needs_flush & CLFLUSH_BEFORE)
 			drm_clflush_virt_range(ptr, PAGE_SIZE);
@@ -1030,12 +1030,12 @@ __cpu_check_shmem(struct drm_i915_gem_object *obj, u32 dword, u32 val)
 		if (ptr[dword] != val) {
 			pr_err("n=%lu ptr[%u]=%u, val=%u\n",
 			       n, dword, ptr[dword], val);
-			kunmap_atomic(ptr);
+			kunmap_local(ptr);
 			err = -EINVAL;
 			break;
 		}
 
-		kunmap_atomic(ptr);
+		kunmap_local(ptr);
 	}
 
 	i915_gem_object_finish_access(obj);
-- 
2.34.1

[PATCH 5/9] drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_coherency.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gem/selftests/i915_gem_coherency.c, functions cpu_set()
and cpu_get() mainly uses mapping to flush cache and assign the value.
There're 2 reasons why cpu_set() and cpu_get() don't need to disable
pagefaults and preemption for mapping:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. cpu_set() and cpu_get() call drm_clflush_virt_range() to use
CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86 and
WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
operation is global and any issue with cpu's being added or removed
can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, cpu_set() and cpu_get() are functions where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 .../gpu/drm/i915/gem/selftests/i915_gem_coherency.c  | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
index a666d7e610f5..b12402c74424 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_coherency.c
@@ -24,7 +24,6 @@ static int cpu_set(struct context *ctx, unsigned long offset, u32 v)
 {
 	unsigned int needs_clflush;
 	struct page *page;
-	void *map;
 	u32 *cpu;
 	int err;
 
@@ -34,8 +33,7 @@ static int cpu_set(struct context *ctx, unsigned long offset, u32 v)
 		goto out;
 
 	page = i915_gem_object_get_page(ctx->obj, offset >> PAGE_SHIFT);
-	map = kmap_atomic(page);
-	cpu = map + offset_in_page(offset);
+	cpu = kmap_local_page(page) + offset_in_page(offset);
 
 	if (needs_clflush & CLFLUSH_BEFORE)
 		drm_clflush_virt_range(cpu, sizeof(*cpu));
@@ -45,7 +43,7 @@ static int cpu_set(struct context *ctx, unsigned long offset, u32 v)
 	if (needs_clflush & CLFLUSH_AFTER)
 		drm_clflush_virt_range(cpu, sizeof(*cpu));
 
-	kunmap_atomic(map);
+	kunmap_local(cpu);
 	i915_gem_object_finish_access(ctx->obj);
 
 out:
@@ -57,7 +55,6 @@ static int cpu_get(struct context *ctx, unsigned long offset, u32 *v)
 {
 	unsigned int needs_clflush;
 	struct page *page;
-	void *map;
 	u32 *cpu;
 	int err;
 
@@ -67,15 +64,14 @@ static int cpu_get(struct context *ctx, unsigned long offset, u32 *v)
 		goto out;
 
 	page = i915_gem_object_get_page(ctx->obj, offset >> PAGE_SHIFT);
-	map = kmap_atomic(page);
-	cpu = map + offset_in_page(offset);
+	cpu = kmap_local_page(page) + offset_in_page(offset);
 
 	if (needs_clflush & CLFLUSH_BEFORE)
 		drm_clflush_virt_range(cpu, sizeof(*cpu));
 
 	*v = *cpu;
 
-	kunmap_atomic(map);
+	kunmap_local(cpu);
 	i915_gem_object_finish_access(ctx->obj);
 
 out:
-- 
2.34.1

[PATCH 6/9] drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_context.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gem/selftests/i915_gem_context.c, functions cpu_fill() and
cpu_check() mainly uses mapping to flush cache and check/assign the
value.

There're 2 reasons why cpu_fill() and cpu_check() don't need to disable
pagefaults and preemption for mapping:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. cpu_fill() and cpu_check() call drm_clflush_virt_range() to
use CLFLUSHOPT or WBINVD to flush. Since CLFLUSHOPT is global on x86
and WBINVD is called on each cpu in drm_clflush_virt_range(), the flush
operation is global and any issue with cpu's being added or removed
can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, cpu_fill() and cpu_check() are functions where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
index c6ad67b90e8a..736337f23f78 100644
--- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_context.c
@@ -466,12 +466,12 @@ static int cpu_fill(struct drm_i915_gem_object *obj, u32 value)
 	for (n = 0; n < real_page_count(obj); n++) {
 		u32 *map;
 
-		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		map = kmap_local_page(i915_gem_object_get_page(obj, n));
 		for (m = 0; m < DW_PER_PAGE; m++)
 			map[m] = value;
 		if (!has_llc)
 			drm_clflush_virt_range(map, PAGE_SIZE);
-		kunmap_atomic(map);
+		kunmap_local(map);
 	}
 
 	i915_gem_object_finish_access(obj);
@@ -496,7 +496,7 @@ static noinline int cpu_check(struct drm_i915_gem_object *obj,
 	for (n = 0; n < real_page_count(obj); n++) {
 		u32 *map;
 
-		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		map = kmap_local_page(i915_gem_object_get_page(obj, n));
 		if (needs_flush & CLFLUSH_BEFORE)
 			drm_clflush_virt_range(map, PAGE_SIZE);
 
@@ -522,7 +522,7 @@ static noinline int cpu_check(struct drm_i915_gem_object *obj,
 		}
 
 out_unmap:
-		kunmap_atomic(map);
+		kunmap_local(map);
 		if (err)
 			break;
 	}
-- 
2.34.1

[PATCH 7/9] drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In drm/i915/gt/uc/intel_us_fw.c, the function intel_uc_fw_copy_rsa()
just use the mapping to do memory copy so it doesn't need to disable
pagefaults and preemption for mapping. Thus the local mapping without
atomic context (not disable pagefaults / preemption) is enough.

Therefore, intel_uc_fw_copy_rsa() is a function where the use of
memcpy_from_page() with kmap_local_page() in place of memcpy() with
kmap_atomic() is correctly suited.

Convert the calls of memcpy() with kmap_atomic() / kunmap_atomic() to
memcpy_from_page() which uses local mapping to copy.

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com/T/#u

Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Ira: Referred to his task document and suggestions about using
       memcpy_from_page() directly.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
index b91ad4aede1f..64d56f175d32 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
@@ -962,16 +962,13 @@ size_t intel_uc_fw_copy_rsa(struct intel_uc_fw *uc_fw, void *dst, u32 max_len)
 
 		for_each_sgt_page(page, iter, uc_fw->obj->mm.pages) {
 			u32 len = min_t(u32, size, PAGE_SIZE - offset);
-			void *vaddr;
 
 			if (idx > 0) {
 				idx--;
 				continue;
 			}
 
-			vaddr = kmap_atomic(page);
-			memcpy(dst, vaddr + offset, len);
-			kunmap_atomic(vaddr);
+			memcpy_from_page(dst, page, offset, len);
 
 			offset = 0;
 			dst += len;
-- 
2.34.1

Re: [PATCH 7/9] drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c

[Ira Weiny]

On Mon, Oct 17, 2022 at 05:37:23PM +0800, Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].
> 
> The main difference between atomic and local mappings is that local
> mappings doesn't disable page faults or preemption.
> 
> In drm/i915/gt/uc/intel_us_fw.c, the function intel_uc_fw_copy_rsa()
> just use the mapping to do memory copy so it doesn't need to disable
> pagefaults and preemption for mapping. Thus the local mapping without
> atomic context (not disable pagefaults / preemption) is enough.
> 
> Therefore, intel_uc_fw_copy_rsa() is a function where the use of
> memcpy_from_page() with kmap_local_page() in place of memcpy() with
> kmap_atomic() is correctly suited.
> 
> Convert the calls of memcpy() with kmap_atomic() / kunmap_atomic() to
> memcpy_from_page() which uses local mapping to copy.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com/T/#u
> 
> Suggested-by: Ira Weiny <ira.weiny@intel.com>

Reviewed-by: Ira Weiny <ira.weiny@intel.com>

> Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> ---
> Suggested by credits:
>   Ira: Referred to his task document and suggestions about using
>        memcpy_from_page() directly.
>   Fabio: Referred to his boiler plate commit message.
> ---
>  drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c | 5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> index b91ad4aede1f..64d56f175d32 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
> @@ -962,16 +962,13 @@ size_t intel_uc_fw_copy_rsa(struct intel_uc_fw *uc_fw, void *dst, u32 max_len)
>  
>  		for_each_sgt_page(page, iter, uc_fw->obj->mm.pages) {
>  			u32 len = min_t(u32, size, PAGE_SIZE - offset);
> -			void *vaddr;
>  
>  			if (idx > 0) {
>  				idx--;
>  				continue;
>  			}
>  
> -			vaddr = kmap_atomic(page);
> -			memcpy(dst, vaddr + offset, len);
> -			kunmap_atomic(vaddr);
> +			memcpy_from_page(dst, page, offset, len);
>  
>  			offset = 0;
>  			dst += len;
> -- 
> 2.34.1
> 

[PATCH 8/9] drm/i915: Use kmap_local_page() in i915_cmd_parser.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

There're 2 reasons why function copy_batch() doesn't need to disable
pagefaults and preemption for mapping:

1. The flush operation is safe for CPU hotplug when preemption is not
disabled. In i915_cmd_parser.c, the function copy_batch() calls
drm_clflush_virt_range() to use CLFLUSHOPT or WBINVD to flush.
Since CLFLUSHOPT is global on x86 and WBINVD is called on each cpu
in drm_clflush_virt_range(), the flush operation is global and any
issue with cpu's being added or removed can be handled safely.

2. Any context switch caused by preemption or sleep (pagefault may
cause sleep) doesn't affect the validity of local mapping.

Therefore, copy_batch() is a function where the use of
kmap_local_page() in place of kmap_atomic() is correctly suited.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Dave Hansen <dave.hansen@intel.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
Suggested by credits:
  Dave: Referred to his explanation about cache flush.
  Ira: Referred to his task document, review comments and explanation about
       cache flush.
  Fabio: Referred to his boiler plate commit message.
---
 drivers/gpu/drm/i915/i915_cmd_parser.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_cmd_parser.c b/drivers/gpu/drm/i915/i915_cmd_parser.c
index f93e6122f247..1a56000d7476 100644
--- a/drivers/gpu/drm/i915/i915_cmd_parser.c
+++ b/drivers/gpu/drm/i915/i915_cmd_parser.c
@@ -1211,11 +1211,11 @@ static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
 		for (n = offset >> PAGE_SHIFT; remain; n++) {
 			int len = min(remain, PAGE_SIZE - x);
 
-			src = kmap_atomic(i915_gem_object_get_page(src_obj, n));
+			src = kmap_local_page(i915_gem_object_get_page(src_obj, n));
 			if (src_needs_clflush)
 				drm_clflush_virt_range(src + x, len);
 			memcpy(ptr, src + x, len);
-			kunmap_atomic(src);
+			kunmap_local(src);
 
 			ptr += len;
 			remain -= len;
-- 
2.34.1

[PATCH 9/9] drm/i915: Use kmap_local_page() in gem/i915_gem_execbuffer.c

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

The use of kmap_atomic() is being deprecated in favor of
kmap_local_page()[1].

The main difference between atomic and local mappings is that local
mappings doesn't disable page faults or preemption.

In i915_gem_execbuffer.c, eb->reloc_cache.vaddr is mapped by
kmap_atomic() in eb_relocate_entry(), and is unmapped by
kunmap_atomic() in reloc_cache_reset().

And this mapping/unmapping occurs in two places: one is in
eb_relocate_vma(), and another is in eb_relocate_vma_slow().

The function eb_relocate_vma() or eb_relocate_vma_slow() doesn't
need to disable pagefaults and preemption during the above mapping/
unmapping.

So it can simply use kmap_local_page() / kunmap_local() that can
instead do the mapping / unmapping regardless of the context.

Convert the calls of kmap_atomic() / kunmap_atomic() to
kmap_local_page() / kunmap_local().

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
---
 drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index 845023c14eb3..8263d4e6620a 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -1110,7 +1110,7 @@ static void reloc_cache_unmap(struct reloc_cache *cache)
 
 	vaddr = unmask_page(cache->vaddr);
 	if (cache->vaddr & KMAP)
-		kunmap_atomic(vaddr);
+		kunmap_local(vaddr);
 	else
 		io_mapping_unmap_atomic((void __iomem *)vaddr);
 }
@@ -1126,7 +1126,7 @@ static void reloc_cache_remap(struct reloc_cache *cache,
 	if (cache->vaddr & KMAP) {
 		struct page *page = i915_gem_object_get_page(obj, cache->page);
 
-		vaddr = kmap_atomic(page);
+		vaddr = kmap_local_page(page);
 		cache->vaddr = unmask_flags(cache->vaddr) |
 			(unsigned long)vaddr;
 	} else {
@@ -1156,7 +1156,7 @@ static void reloc_cache_reset(struct reloc_cache *cache, struct i915_execbuffer
 		if (cache->vaddr & CLFLUSH_AFTER)
 			mb();
 
-		kunmap_atomic(vaddr);
+		kunmap_local(vaddr);
 		i915_gem_object_finish_access(obj);
 	} else {
 		struct i915_ggtt *ggtt = cache_to_ggtt(cache);
@@ -1188,7 +1188,7 @@ static void *reloc_kmap(struct drm_i915_gem_object *obj,
 	struct page *page;
 
 	if (cache->vaddr) {
-		kunmap_atomic(unmask_page(cache->vaddr));
+		kunmap_local(unmask_page(cache->vaddr));
 	} else {
 		unsigned int flushes;
 		int err;
@@ -1210,7 +1210,7 @@ static void *reloc_kmap(struct drm_i915_gem_object *obj,
 	if (!obj->mm.dirty)
 		set_page_dirty(page);
 
-	vaddr = kmap_atomic(page);
+	vaddr = kmap_local_page(page);
 	cache->vaddr = unmask_flags(cache->vaddr) | (unsigned long)vaddr;
 	cache->page = pageno;
 
-- 
2.34.1

[PATCH v3] x86/hyperv: Replace kmap() with kmap_local_page()

[Zhao Liu]

From: Zhao Liu <zhao1.liu@intel.com>

kmap() is being deprecated in favor of kmap_local_page()[1].

There are two main problems with kmap(): (1) It comes with an overhead as mapping space is restricted and protected by a global lock for synchronization and (2) it also requires global TLB invalidation when the kmap's pool wraps and it might block when the mapping space is fully utilized until a slot becomes available.

With kmap_local_page() the mappings are per thread, CPU local, can take page faults, and can be called from any context (including interrupts).
It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore, the tasks can be preempted and, when they are scheduled to run again, the kernel virtual addresses are restored and are still valid.

Since its use in hyperv/hv_init.c is safe, it should be preferred.

Therefore, replace kmap() with kmap_local_page() in hyperv/hv_init.c.

[1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com

Suggested-by: Ira Weiny <ira.weiny@intel.com>
Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>

---
Suggested by credits.
        Ira: Referred to his task documentation and review comments.
        Fabio: Stole some of his boiler plate commit message.

---
Changelog:
v2:
- Fix wrong incoming parameters in kunmap_local();
- Add Fabio as suggester since I quoted his commit message.

---
 arch/x86/hyperv/hv_init.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/hyperv/hv_init.c b/arch/x86/hyperv/hv_init.c
index 3de6d8b53367..72fe46eb183f 100644
--- a/arch/x86/hyperv/hv_init.c
+++ b/arch/x86/hyperv/hv_init.c
@@ -459,13 +459,13 @@ void __init hyperv_init(void)
                wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);

                pg = vmalloc_to_page(hv_hypercall_pg);
-               dst = kmap(pg);
+               dst = kmap_local_page(pg);
                src = memremap(hypercall_msr.guest_physical_address << PAGE_SHIFT, PAGE_SIZE,
                                MEMREMAP_WB);
                BUG_ON(!(src && dst));
                memcpy(dst, src, HV_HYP_PAGE_SIZE);
                memunmap(src);
-               kunmap(pg);
+               kunmap_local(dst);
        } else {
                hypercall_msr.guest_physical_address = vmalloc_to_pfn(hv_hypercall_pg);
                wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
--
2.34.1

Re: [PATCH v3] x86/hyperv: Replace kmap() with kmap_local_page()

[Zhao Liu]

Sorry, please ignore the last hyperv patch, I made a mistake.

Zhao

On Mon, Oct 17, 2022 at 05:37:26PM +0800, Zhao Liu wrote:
> Date: Mon, 17 Oct 2022 17:37:26 +0800
> From: Zhao Liu <zhao1.liu@linux.intel.com>
> Subject: [PATCH v3] x86/hyperv: Replace kmap() with kmap_local_page()
> X-Mailer: git-send-email 2.34.1
> 
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> kmap() is being deprecated in favor of kmap_local_page()[1].
> 
> There are two main problems with kmap(): (1) It comes with an overhead as mapping space is restricted and protected by a global lock for synchronization and (2) it also requires global TLB invalidation when the kmap's pool wraps and it might block when the mapping space is fully utilized until a slot becomes available.
> 
> With kmap_local_page() the mappings are per thread, CPU local, can take page faults, and can be called from any context (including interrupts).
> It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore, the tasks can be preempted and, when they are scheduled to run again, the kernel virtual addresses are restored and are still valid.
> 
> Since its use in hyperv/hv_init.c is safe, it should be preferred.
> 
> Therefore, replace kmap() with kmap_local_page() in hyperv/hv_init.c.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> 
> Suggested-by: Ira Weiny <ira.weiny@intel.com>
> Suggested-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
> Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
> 
> ---
> Suggested by credits.
>         Ira: Referred to his task documentation and review comments.
>         Fabio: Stole some of his boiler plate commit message.
> 
> ---
> Changelog:
> v2:
> - Fix wrong incoming parameters in kunmap_local();
> - Add Fabio as suggester since I quoted his commit message.
> 
> ---
>  arch/x86/hyperv/hv_init.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/x86/hyperv/hv_init.c b/arch/x86/hyperv/hv_init.c
> index 3de6d8b53367..72fe46eb183f 100644
> --- a/arch/x86/hyperv/hv_init.c
> +++ b/arch/x86/hyperv/hv_init.c
> @@ -459,13 +459,13 @@ void __init hyperv_init(void)
>                 wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
> 
>                 pg = vmalloc_to_page(hv_hypercall_pg);
> -               dst = kmap(pg);
> +               dst = kmap_local_page(pg);
>                 src = memremap(hypercall_msr.guest_physical_address << PAGE_SHIFT, PAGE_SIZE,
>                                 MEMREMAP_WB);
>                 BUG_ON(!(src && dst));
>                 memcpy(dst, src, HV_HYP_PAGE_SIZE);
>                 memunmap(src);
> -               kunmap(pg);
> +               kunmap_local(dst);
>         } else {
>                 hypercall_msr.guest_physical_address = vmalloc_to_pfn(hv_hypercall_pg);
>                 wrmsrl(HV_X64_MSR_HYPERCALL, hypercall_msr.as_uint64);
> --
> 2.34.1
> 

Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Fabio M. De Francesco]

On lunedì 17 ottobre 2022 11:37:16 CEST Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].

Some words to explain why kmap_atomic was deprecated won't hurt. Many 
maintainers and reviewers, and also casual readers might not yet be aware of 
the reasons behind that deprecation.
 
> In the following patches, we can convert the calls of kmap_atomic() /
> kunmap_atomic() to kmap_local_page() / kunmap_local(), which can
> instead do the mapping / unmapping regardless of the context.

Readers are probably much more interested in what you did in the following 
patches and why you did it, instead of being informed about what "we can" do.

I would suggest something like "The following patches convert the calls to 
kmap_atomic() to kmap_local_page() [the rest looks OK]".

This could also be the place to say something about why we prefer 
kmap_local_page() to kmap_atomic(). 

Are you sure that the reasons that motivates your conversions are merely 
summarized to kmap_local_page() being able to do mappings regardless of 
context? I think you are missing the real reasons why. 

What about avoiding the often unwanted side effect of unnecessary page faults 
disables?

> 
> With kmap_local_page(), the mapping is per thread, CPU local and not
> globally visible.

No news here. kmap_atomic() is "per thread, CPU local and not glocally 
visible". I cannot see any difference here between kmap_atomic() and 
kmap_local_page().

> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> ---
> Zhao Liu (9):
>   drm/i915: Use kmap_local_page() in gem/i915_gem_object.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c
>   drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c
>   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_coherency.c
>   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_context.c
>   drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c
>   drm/i915: Use kmap_local_page() in i915_cmd_parser.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_execbuffer.c
> 
>  drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c       | 10 +++++-----
>  drivers/gpu/drm/i915/gem/i915_gem_object.c           |  8 +++-----
>  drivers/gpu/drm/i915/gem/i915_gem_phys.c             |  8 ++++----
>  drivers/gpu/drm/i915/gem/i915_gem_shmem.c            |  6 ++++--
>  drivers/gpu/drm/i915/gem/selftests/huge_pages.c      |  6 +++---
>  .../gpu/drm/i915/gem/selftests/i915_gem_coherency.c  | 12 ++++--------
>  .../gpu/drm/i915/gem/selftests/i915_gem_context.c    |  8 ++++----
>  drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c             |  5 +----
>  drivers/gpu/drm/i915/i915_cmd_parser.c               |  4 ++--
>  9 files changed, 30 insertions(+), 37 deletions(-)

Thanks for helping with kmap_atomic() conversions to kmap_local_page().

Fabio

Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Zhao Liu]

On Sat, Oct 29, 2022 at 09:12:27AM +0200, Fabio M. De Francesco wrote:
> Date: Sat, 29 Oct 2022 09:12:27 +0200
> From: "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
> Subject: Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with
>  kmap_local_page()

Hi Fabio, thanks for your review!! (I'm sorry I missed the previous mails).

> 
> On luned? 17 ottobre 2022 11:37:16 CEST Zhao Liu wrote:
> > From: Zhao Liu <zhao1.liu@intel.com>
> > 
> > The use of kmap_atomic() is being deprecated in favor of
> > kmap_local_page()[1].
> 
> Some words to explain why kmap_atomic was deprecated won't hurt. Many 
> maintainers and reviewers, and also casual readers might not yet be aware of 
> the reasons behind that deprecation.
>  
> > In the following patches, we can convert the calls of kmap_atomic() /
> > kunmap_atomic() to kmap_local_page() / kunmap_local(), which can
> > instead do the mapping / unmapping regardless of the context.
> 
> Readers are probably much more interested in what you did in the following 
> patches and why you did it, instead of being informed about what "we can" do.
> 
> I would suggest something like "The following patches convert the calls to 
> kmap_atomic() to kmap_local_page() [the rest looks OK]".
> 
> This could also be the place to say something about why we prefer 
> kmap_local_page() to kmap_atomic(). 
> 
> Are you sure that the reasons that motivates your conversions are merely 
> summarized to kmap_local_page() being able to do mappings regardless of 
> context? I think you are missing the real reasons why. 

Thanks for your reminder, I'll emphasize the motivation here.

> What about avoiding the often unwanted side effect of unnecessary page faults 
> disables?

Good suggestion! I'll add this into this cover message.

What I think is that we have two reasons to do the replacement work:
1. (main motication) Avoid unnessary pagefaulta and preemption disabling to gain
performance benefits.
2. We are trying to deprecate the old kmap/kmap_atomic interface. Some maintainer
said it's also a good reason especially for the case that the performance is not
critical [1].

In addition, also from [1], I find in some case people chooses kmap_atomic() for
the consideration that they want the atomic context. So, the explaination about
why the atomic context is not needed is also a reasion? I understand that I need
to make special explaination in each commit depending on the situation (In this
case, it is not suitable to describe in the cover?).

[1]: https://lore.kernel.org/lkml/YzRVaJA0EyfcVisW@liuwe-devbox-debian-v2/#t

> 
> > 
> > With kmap_local_page(), the mapping is per thread, CPU local and not
> > globally visible.
> 
> No news here. kmap_atomic() is "per thread, CPU local and not glocally 
> visible". I cannot see any difference here between kmap_atomic() and 
> kmap_local_page().

What about the below description which refers to your doc?
"kmap_atomic() in the kernel creates a non-preemptible section
and disable pagefaults. This could be a source of unwanted latency.
And kmap_local_page effectively overcomes this issue because it doesn't
disable pagefault and preemption."

Thanks,
Zhao

Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Ira Weiny]

Zhao Liu wrote:
> From: Zhao Liu <zhao1.liu@intel.com>
> 
> The use of kmap_atomic() is being deprecated in favor of
> kmap_local_page()[1].

Zhao,

Was there ever a v2 of this series?  I'm not finding it on Lore.

Thanks,
Ira

> 
> In the following patches, we can convert the calls of kmap_atomic() /
> kunmap_atomic() to kmap_local_page() / kunmap_local(), which can
> instead do the mapping / unmapping regardless of the context.
> 
> With kmap_local_page(), the mapping is per thread, CPU local and not
> globally visible.
> 
> [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> ---
> Zhao Liu (9):
>   drm/i915: Use kmap_local_page() in gem/i915_gem_object.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c
>   drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c
>   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_coherency.c
>   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_context.c
>   drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c
>   drm/i915: Use kmap_local_page() in i915_cmd_parser.c
>   drm/i915: Use kmap_local_page() in gem/i915_gem_execbuffer.c
> 
>  drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c       | 10 +++++-----
>  drivers/gpu/drm/i915/gem/i915_gem_object.c           |  8 +++-----
>  drivers/gpu/drm/i915/gem/i915_gem_phys.c             |  8 ++++----
>  drivers/gpu/drm/i915/gem/i915_gem_shmem.c            |  6 ++++--
>  drivers/gpu/drm/i915/gem/selftests/huge_pages.c      |  6 +++---
>  .../gpu/drm/i915/gem/selftests/i915_gem_coherency.c  | 12 ++++--------
>  .../gpu/drm/i915/gem/selftests/i915_gem_context.c    |  8 ++++----
>  drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c             |  5 +----
>  drivers/gpu/drm/i915/i915_cmd_parser.c               |  4 ++--
>  9 files changed, 30 insertions(+), 37 deletions(-)
> 
> -- 
> 2.34.1
> 

Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Zhao Liu]

On Tue, Feb 14, 2023 at 08:25:08PM -0800, Ira Weiny wrote:
> Date: Tue, 14 Feb 2023 20:25:08 -0800
> From: Ira Weiny <ira.weiny@intel.com>
> Subject: Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with
>  kmap_local_page()
> 
> Zhao Liu wrote:
> > From: Zhao Liu <zhao1.liu@intel.com>
> > 
> > The use of kmap_atomic() is being deprecated in favor of
> > kmap_local_page()[1].
> 
> Zhao,
> 
> Was there ever a v2 of this series?  I'm not finding it on Lore.

Sorry Ira, my delay is too long, I was busy with other patch work,
I will refresh v2 soon, and push this forward!

Best Regards,
Zhao

> 
> Thanks,
> Ira
> 
> > 
> > In the following patches, we can convert the calls of kmap_atomic() /
> > kunmap_atomic() to kmap_local_page() / kunmap_local(), which can
> > instead do the mapping / unmapping regardless of the context.
> > 
> > With kmap_local_page(), the mapping is per thread, CPU local and not
> > globally visible.
> > 
> > [1]: https://lore.kernel.org/all/20220813220034.806698-1-ira.weiny@intel.com
> > ---
> > Zhao Liu (9):
> >   drm/i915: Use kmap_local_page() in gem/i915_gem_object.c
> >   drm/i915: Use kmap_local_page() in gem/i915_gem_pyhs.c
> >   drm/i915: Use kmap_local_page() in gem/i915_gem_shmem.c
> >   drm/i915: Use kmap_local_page() in gem/selftests/huge_pages.c
> >   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_coherency.c
> >   drm/i915: Use kmap_local_page() in gem/selftests/i915_gem_context.c
> >   drm/i915: Use memcpy_from_page() in gt/uc/intel_uc_fw.c
> >   drm/i915: Use kmap_local_page() in i915_cmd_parser.c
> >   drm/i915: Use kmap_local_page() in gem/i915_gem_execbuffer.c
> > 
> >  drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c       | 10 +++++-----
> >  drivers/gpu/drm/i915/gem/i915_gem_object.c           |  8 +++-----
> >  drivers/gpu/drm/i915/gem/i915_gem_phys.c             |  8 ++++----
> >  drivers/gpu/drm/i915/gem/i915_gem_shmem.c            |  6 ++++--
> >  drivers/gpu/drm/i915/gem/selftests/huge_pages.c      |  6 +++---
> >  .../gpu/drm/i915/gem/selftests/i915_gem_coherency.c  | 12 ++++--------
> >  .../gpu/drm/i915/gem/selftests/i915_gem_context.c    |  8 ++++----
> >  drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c             |  5 +----
> >  drivers/gpu/drm/i915/i915_cmd_parser.c               |  4 ++--
> >  9 files changed, 30 insertions(+), 37 deletions(-)
> > 
> > -- 
> > 2.34.1
> > 
> 
> 

Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with kmap_local_page()

[Ira Weiny]

Zhao Liu wrote:
> On Tue, Feb 14, 2023 at 08:25:08PM -0800, Ira Weiny wrote:
> > Date: Tue, 14 Feb 2023 20:25:08 -0800
> > From: Ira Weiny <ira.weiny@intel.com>
> > Subject: Re: [PATCH 0/9] drm/i915: Replace kmap_atomic() with
> >  kmap_local_page()
> > 
> > Zhao Liu wrote:
> > > From: Zhao Liu <zhao1.liu@intel.com>
> > > 
> > > The use of kmap_atomic() is being deprecated in favor of
> > > kmap_local_page()[1].
> > 
> > Zhao,
> > 
> > Was there ever a v2 of this series?  I'm not finding it on Lore.
> 
> Sorry Ira, my delay is too long, I was busy with other patch work,
> I will refresh v2 soon, and push this forward!

Awesome!  Thanks!
Ira

> 
> Best Regards,
> Zhao