public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
From: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
To: Matthew Brost <matthew.brost@intel.com>
Cc: intel-xe@lists.freedesktop.org,
	Matthew Auld <matthew.auld@intel.com>,
	 Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>,
		dri-devel@lists.freedesktop.org, stable@vger.kernel.org
Subject: Re: [PATCH v2 1/2] drm, drm/xe: Fix xe userptr in the absence of CONFIG_DEVICE_PRIVATE
Date: Fri, 23 Jan 2026 14:34:19 +0100	[thread overview]
Message-ID: <34532068b6414d35264bc1d2f0a33b0a1be8fb3e.camel@linux.intel.com> (raw)
In-Reply-To: <aW/D6ehR6DB2SVlf@lstrano-desk.jf.intel.com>

On Tue, 2026-01-20 at 10:05 -0800, Matthew Brost wrote:
> On Tue, Jan 20, 2026 at 03:34:58PM +0100, Thomas Hellström wrote:
> > CONFIG_DEVICE_PRIVATE is not selected by default by some distros,
> > for example Fedora, and that leads to a regression in the xe driver
> > since userptr support gets compiled out.
> > 
> > It turns out that DRM_GPUSVM, which is needed for xe userptr
> > support
> > compiles also without CONFIG_DEVICE_PRIVATE, but doesn't compile
> > without CONFIG_ZONE_DEVICE.
> > Exclude the drm_pagemap files from compilation with
> > !CONFIG_ZONE_DEVICE,
> > and remove the CONFIG_DEVICE_PRIVATE dependency from
> > CONFIG_DRM_GPUSVM and
> > the xe driver's selection of it, re-enabling xe userptr for those
> > configs.
> > 
> > v2:
> > - Don't compile the drm_pagemap files unless CONFIG_ZONE_DEVICE is
> > set.
> > - Adjust the drm_pagemap.h header accordingly.
> > 
> > Fixes: 9e9787414882 ("drm/xe/userptr: replace xe_hmm with gpusvm")
> > Cc: Matthew Auld <matthew.auld@intel.com>
> > Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
> > Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> > Cc: Matthew Brost <matthew.brost@intel.com>
> > Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
> > Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > Cc: dri-devel@lists.freedesktop.org
> > Cc: <stable@vger.kernel.org> # v6.18+
> > Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> > ---
> >  drivers/gpu/drm/Kconfig    |  2 +-
> >  drivers/gpu/drm/Makefile   |  4 +++-
> >  drivers/gpu/drm/xe/Kconfig |  2 +-
> >  include/drm/drm_pagemap.h  | 18 ++++++++++++++----
> >  4 files changed, 19 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> > index a33b90251530..d3d52310c9cc 100644
> > --- a/drivers/gpu/drm/Kconfig
> > +++ b/drivers/gpu/drm/Kconfig
> > @@ -210,7 +210,7 @@ config DRM_GPUVM
> >  
> >  config DRM_GPUSVM
> >  	tristate
> > -	depends on DRM && DEVICE_PRIVATE
> > +	depends on DRM
> >  	select HMM_MIRROR
> >  	select MMU_NOTIFIER
> >  	help
> > diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> > index 0deee72ef935..0c21029c446f 100644
> > --- a/drivers/gpu/drm/Makefile
> > +++ b/drivers/gpu/drm/Makefile
> > @@ -108,9 +108,11 @@ obj-$(CONFIG_DRM_EXEC) += drm_exec.o
> >  obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> >  
> >  drm_gpusvm_helper-y := \
> > -	drm_gpusvm.o\
> > +	drm_gpusvm.o
> > +drm_gpusvm_helper-$(CONFIG_ZONE_DEVICE) += \
> >  	drm_pagemap.o\
> >  	drm_pagemap_util.o
> > +
> >  obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm_helper.o
> >  
> >  obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
> > diff --git a/drivers/gpu/drm/xe/Kconfig
> > b/drivers/gpu/drm/xe/Kconfig
> > index 4b288eb3f5b0..c34be1be155b 100644
> > --- a/drivers/gpu/drm/xe/Kconfig
> > +++ b/drivers/gpu/drm/xe/Kconfig
> > @@ -39,7 +39,7 @@ config DRM_XE
> >  	select DRM_TTM
> >  	select DRM_TTM_HELPER
> >  	select DRM_EXEC
> > -	select DRM_GPUSVM if !UML && DEVICE_PRIVATE
> > +	select DRM_GPUSVM if !UML
> >  	select DRM_GPUVM
> >  	select DRM_SCHED
> >  	select MMU_NOTIFIER
> > diff --git a/include/drm/drm_pagemap.h b/include/drm/drm_pagemap.h
> > index 46e9c58f09e0..2baf0861f78f 100644
> > --- a/include/drm/drm_pagemap.h
> > +++ b/include/drm/drm_pagemap.h
> > @@ -243,6 +243,8 @@ struct drm_pagemap_devmem_ops {
> >  			   struct dma_fence *pre_migrate_fence);
> >  };
> >  
> > +#if IS_ENABLED(CONFIG_ZONE_DEVICE)
> > +
> 
> What happens if we select ZONE_DEVICE but not DEVICE_PRIVATE?

This code builds, and in theory would work for DEVICE_COHERENT but the
device-private checks in the code always return false as I understand
it. 

So should be safe for our userptr code, and SVM still requires /
selects DEVICE_PRIVATE.

Thomas


> 
> Matt
> 
> >  int drm_pagemap_init(struct drm_pagemap *dpagemap,
> >  		     struct dev_pagemap *pagemap,
> >  		     struct drm_device *drm,
> > @@ -252,17 +254,22 @@ struct drm_pagemap *drm_pagemap_create(struct
> > drm_device *drm,
> >  				       struct dev_pagemap
> > *pagemap,
> >  				       const struct
> > drm_pagemap_ops *ops);
> >  
> > -#if IS_ENABLED(CONFIG_DRM_GPUSVM)
> > +struct drm_pagemap *drm_pagemap_page_to_dpagemap(struct page
> > *page);
> >  
> >  void drm_pagemap_put(struct drm_pagemap *dpagemap);
> >  
> >  #else
> >  
> > +static inline struct drm_pagemap
> > *drm_pagemap_page_to_dpagemap(struct page *page)
> > +{
> > +	return NULL;
> > +}
> > +
> >  static inline void drm_pagemap_put(struct drm_pagemap *dpagemap)
> >  {
> >  }
> >  
> > -#endif /* IS_ENABLED(CONFIG_DRM_GPUSVM) */
> > +#endif /* IS_ENABLED(CONFIG_ZONE_DEVICE) */
> >  
> >  /**
> >   * drm_pagemap_get() - Obtain a reference on a struct drm_pagemap
> > @@ -334,6 +341,8 @@ struct drm_pagemap_migrate_details {
> >  	u32 source_peer_migrates : 1;
> >  };
> >  
> > +#if IS_ENABLED(CONFIG_ZONE_DEVICE)
> > +
> >  int drm_pagemap_migrate_to_devmem(struct drm_pagemap_devmem
> > *devmem_allocation,
> >  				  struct mm_struct *mm,
> >  				  unsigned long start, unsigned
> > long end,
> > @@ -343,8 +352,6 @@ int drm_pagemap_evict_to_ram(struct
> > drm_pagemap_devmem *devmem_allocation);
> >  
> >  const struct dev_pagemap_ops *drm_pagemap_pagemap_ops_get(void);
> >  
> > -struct drm_pagemap *drm_pagemap_page_to_dpagemap(struct page
> > *page);
> > -
> >  void drm_pagemap_devmem_init(struct drm_pagemap_devmem
> > *devmem_allocation,
> >  			     struct device *dev, struct mm_struct
> > *mm,
> >  			     const struct drm_pagemap_devmem_ops
> > *ops,
> > @@ -359,4 +366,7 @@ int drm_pagemap_populate_mm(struct drm_pagemap
> > *dpagemap,
> >  void drm_pagemap_destroy(struct drm_pagemap *dpagemap, bool
> > is_atomic_or_reclaim);
> >  
> >  int drm_pagemap_reinit(struct drm_pagemap *dpagemap);
> > +
> > +#endif /* IS_ENABLED(CONFIG_ZONE_DEVICE) */
> > +
> >  #endif
> > -- 
> > 2.52.0
> > 

      reply	other threads:[~2026-01-23 13:34 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20260120143459.9485-1-thomas.hellstrom@linux.intel.com>
2026-01-20 14:34 ` [PATCH v2 1/2] drm, drm/xe: Fix xe userptr in the absence of CONFIG_DEVICE_PRIVATE Thomas Hellström
2026-01-20 18:05   ` Matthew Brost
2026-01-23 13:34     ` Thomas Hellström [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=34532068b6414d35264bc1d2f0a33b0a1be8fb3e.camel@linux.intel.com \
    --to=thomas.hellstrom@linux.intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=himal.prasad.ghimiray@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=matthew.auld@intel.com \
    --cc=matthew.brost@intel.com \
    --cc=rodrigo.vivi@intel.com \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox