[PATCH] drm/vgm: replace idr_init() by idr_init_base()

[PATCH] drm/vgm: replace idr_init() by idr_init_base()

[Deepak R Varma]

idr_init() uses base 0 which is an invalid identifier. The new function
idr_init_base allows IDR to set the ID lookup from base 1. This avoids
all lookups that otherwise starts from 0 since 0 is always unused.

References: commit 6ce711f27500 ("idr: Make 1-based IDRs more efficient")

Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
---
 drivers/gpu/drm/vgem/vgem_fence.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c
index 17f32f550dd9..2902dc6e64fa 100644
--- a/drivers/gpu/drm/vgem/vgem_fence.c
+++ b/drivers/gpu/drm/vgem/vgem_fence.c
@@ -233,7 +233,7 @@ int vgem_fence_signal_ioctl(struct drm_device *dev,
 int vgem_fence_open(struct vgem_file *vfile)
 {
 	mutex_init(&vfile->fence_mutex);
-	idr_init(&vfile->fence_idr);
+	idr_init_base(&vfile->fence_idr, 1);
 
 	return 0;
 }
-- 
2.25.1

Re: [PATCH] drm/vgm: replace idr_init() by idr_init_base()

[Daniel Vetter]

On Wed, Nov 04, 2020 at 04:53:38PM +0530, Deepak R Varma wrote:
> idr_init() uses base 0 which is an invalid identifier. The new function
> idr_init_base allows IDR to set the ID lookup from base 1. This avoids
> all lookups that otherwise starts from 0 since 0 is always unused.
> 
> References: commit 6ce711f27500 ("idr: Make 1-based IDRs more efficient")
> 
> Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>

Tiny typo in the commit message summary: s/vgm/vgem/

Also can you pls resbumit this with intel-gfx mailing list on cc (like for
i915)? There's a CI bot there which runs a few vgem tests, would be good
to confirm nothing has been broken.

Otherwise lgtm.

Thanks, Daniel

> ---
>  drivers/gpu/drm/vgem/vgem_fence.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c
> index 17f32f550dd9..2902dc6e64fa 100644
> --- a/drivers/gpu/drm/vgem/vgem_fence.c
> +++ b/drivers/gpu/drm/vgem/vgem_fence.c
> @@ -233,7 +233,7 @@ int vgem_fence_signal_ioctl(struct drm_device *dev,
>  int vgem_fence_open(struct vgem_file *vfile)
>  {
>  	mutex_init(&vfile->fence_mutex);
> -	idr_init(&vfile->fence_idr);
> +	idr_init_base(&vfile->fence_idr, 1);
>  
>  	return 0;
>  }
> -- 
> 2.25.1
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Re: [PATCH] drm/vgm: replace idr_init() by idr_init_base()

[Deepak R Varma]

On Thu, Nov 05, 2020 at 10:42:15AM +0100, Daniel Vetter wrote:
> On Wed, Nov 04, 2020 at 04:53:38PM +0530, Deepak R Varma wrote:
> > idr_init() uses base 0 which is an invalid identifier. The new function
> > idr_init_base allows IDR to set the ID lookup from base 1. This avoids
> > all lookups that otherwise starts from 0 since 0 is always unused.
> > 
> > References: commit 6ce711f27500 ("idr: Make 1-based IDRs more efficient")
> > 
> > Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
> 
> Tiny typo in the commit message summary: s/vgm/vgem/
> 
> Also can you pls resbumit this with intel-gfx mailing list on cc (like for
> i915)? There's a CI bot there which runs a few vgem tests, would be good
> to confirm nothing has been broken.

Hi Daniel,
sure. I will correct the summary typo and also feed it to the CI bot.

Also, according to Felix Kuehling's comment on a similar patch for
drm/amdkfd driver, an ID can be 0. The change I am proposing is more
efficient for conditions that do not want to use ID as 0. Otherwise,
id = 0 is an acceptable possibility. So, my statement that "Id 0 is an invalid
identifier" is not true.

Can you please comment if this is accurate and I should reword my log
message as well?

Thank you.
./drv

> 
> Otherwise lgtm.
> 
> Thanks, Daniel
> 
> > ---
> >  drivers/gpu/drm/vgem/vgem_fence.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c
> > index 17f32f550dd9..2902dc6e64fa 100644
> > --- a/drivers/gpu/drm/vgem/vgem_fence.c
> > +++ b/drivers/gpu/drm/vgem/vgem_fence.c
> > @@ -233,7 +233,7 @@ int vgem_fence_signal_ioctl(struct drm_device *dev,
> >  int vgem_fence_open(struct vgem_file *vfile)
> >  {
> >  	mutex_init(&vfile->fence_mutex);
> > -	idr_init(&vfile->fence_idr);
> > +	idr_init_base(&vfile->fence_idr, 1);
> >  
> >  	return 0;
> >  }
> > -- 
> > 2.25.1
> > 
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch

Re: [PATCH] drm/vgm: replace idr_init() by idr_init_base()

[Daniel Vetter]

On Thu, Nov 5, 2020 at 11:31 AM Deepak R Varma <mh12gx2825@gmail.com> wrote:
>
> On Thu, Nov 05, 2020 at 10:42:15AM +0100, Daniel Vetter wrote:
> > On Wed, Nov 04, 2020 at 04:53:38PM +0530, Deepak R Varma wrote:
> > > idr_init() uses base 0 which is an invalid identifier. The new function
> > > idr_init_base allows IDR to set the ID lookup from base 1. This avoids
> > > all lookups that otherwise starts from 0 since 0 is always unused.
> > >
> > > References: commit 6ce711f27500 ("idr: Make 1-based IDRs more efficient")
> > >
> > > Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
> >
> > Tiny typo in the commit message summary: s/vgm/vgem/
> >
> > Also can you pls resbumit this with intel-gfx mailing list on cc (like for
> > i915)? There's a CI bot there which runs a few vgem tests, would be good
> > to confirm nothing has been broken.
>
> Hi Daniel,
> sure. I will correct the summary typo and also feed it to the CI bot.
>
> Also, according to Felix Kuehling's comment on a similar patch for
> drm/amdkfd driver, an ID can be 0. The change I am proposing is more
> efficient for conditions that do not want to use ID as 0. Otherwise,
> id = 0 is an acceptable possibility. So, my statement that "Id 0 is an invalid
> identifier" is not true.
>
> Can you please comment if this is accurate and I should reword my log
> message as well?

You need to review the vgem code to see whether we're using id 0 as
invalid identifier or not. That's part of the work that needs to be
done here. Best would be to explain the evidence you've found in the
commit message, why id 0 is invalid for this specific code. Since yes
in general that's not true, it depends how the idr is used.
-Daniel

>
> Thank you.
> ./drv
>
> >
> > Otherwise lgtm.
> >
> > Thanks, Daniel
> >
> > > ---
> > >  drivers/gpu/drm/vgem/vgem_fence.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c
> > > index 17f32f550dd9..2902dc6e64fa 100644
> > > --- a/drivers/gpu/drm/vgem/vgem_fence.c
> > > +++ b/drivers/gpu/drm/vgem/vgem_fence.c
> > > @@ -233,7 +233,7 @@ int vgem_fence_signal_ioctl(struct drm_device *dev,
> > >  int vgem_fence_open(struct vgem_file *vfile)
> > >  {
> > >     mutex_init(&vfile->fence_mutex);
> > > -   idr_init(&vfile->fence_idr);
> > > +   idr_init_base(&vfile->fence_idr, 1);
> > >
> > >     return 0;
> > >  }
> > > --
> > > 2.25.1
> > >
> >
> > --
> > Daniel Vetter
> > Software Engineer, Intel Corporation
> > http://blog.ffwll.ch



-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Re: [PATCH] drm/vgm: replace idr_init() by idr_init_base()

[Deepak R Varma]

On Thu, Nov 05, 2020 at 12:16:34PM +0100, Daniel Vetter wrote:
> On Thu, Nov 5, 2020 at 11:31 AM Deepak R Varma <mh12gx2825@gmail.com> wrote:
> >
> > On Thu, Nov 05, 2020 at 10:42:15AM +0100, Daniel Vetter wrote:
> > > On Wed, Nov 04, 2020 at 04:53:38PM +0530, Deepak R Varma wrote:
> > > > idr_init() uses base 0 which is an invalid identifier. The new function
> > > > idr_init_base allows IDR to set the ID lookup from base 1. This avoids
> > > > all lookups that otherwise starts from 0 since 0 is always unused.
> > > >
> > > > References: commit 6ce711f27500 ("idr: Make 1-based IDRs more efficient")
> > > >
> > > > Signed-off-by: Deepak R Varma <mh12gx2825@gmail.com>
> > >
> > > Tiny typo in the commit message summary: s/vgm/vgem/
> > >
> > > Also can you pls resbumit this with intel-gfx mailing list on cc (like for
> > > i915)? There's a CI bot there which runs a few vgem tests, would be good
> > > to confirm nothing has been broken.
> >
> > Hi Daniel,
> > sure. I will correct the summary typo and also feed it to the CI bot.
> >
> > Also, according to Felix Kuehling's comment on a similar patch for
> > drm/amdkfd driver, an ID can be 0. The change I am proposing is more
> > efficient for conditions that do not want to use ID as 0. Otherwise,
> > id = 0 is an acceptable possibility. So, my statement that "Id 0 is an invalid
> > identifier" is not true.
> >
> > Can you please comment if this is accurate and I should reword my log
> > message as well?
> 
> You need to review the vgem code to see whether we're using id 0 as
> invalid identifier or not. That's part of the work that needs to be
> done here. Best would be to explain the evidence you've found in the
> commit message, why id 0 is invalid for this specific code. Since yes
> in general that's not true, it depends how the idr is used.
> -Daniel
> 
You are correct. For the vgem driver, id 0 is not used. The patch
should then apply to this driver.

Thank you very much Daniel. I have just sent v2 of the patch with your
suggestions.

./drv

> >
> > Thank you.
> > ./drv
> >
> > >
> > > Otherwise lgtm.
> > >
> > > Thanks, Daniel
> > >
> > > > ---
> > > >  drivers/gpu/drm/vgem/vgem_fence.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c
> > > > index 17f32f550dd9..2902dc6e64fa 100644
> > > > --- a/drivers/gpu/drm/vgem/vgem_fence.c
> > > > +++ b/drivers/gpu/drm/vgem/vgem_fence.c
> > > > @@ -233,7 +233,7 @@ int vgem_fence_signal_ioctl(struct drm_device *dev,
> > > >  int vgem_fence_open(struct vgem_file *vfile)
> > > >  {
> > > >     mutex_init(&vfile->fence_mutex);
> > > > -   idr_init(&vfile->fence_idr);
> > > > +   idr_init_base(&vfile->fence_idr, 1);
> > > >
> > > >     return 0;
> > > >  }
> > > > --
> > > > 2.25.1
> > > >
> > >
> > > --
> > > Daniel Vetter
> > > Software Engineer, Intel Corporation
> > > http://blog.ffwll.ch
> 
> 
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch