[PATCH] drm/i915: Do not drop pagetables when empty

[Stefan Bader]

On 27.07.2017 13:05, Rex Tsai (蔡志展) wrote:
> From: Chris Wilson <chris at chris-wilson.co.uk>
> 
> From: Chris Wilson <chris at chris-wilson.co.uk>
> 
> BugLink: https://bugs.launchpad.net/bugs/1680904
> 
> This is the minimal backport for stable of the upstream commit:
> 
> commit dd19674bacba227ae5d3ce680cbc5668198894dc
> Author: Chris Wilson <chris at chris-wilson.co.uk>
> Date:   Wed Feb 15 08:43:46 2017 +0000
> 
>     drm/i915: Remove bitmap tracking for used-ptes
> 
> Due to a race with the shrinker, when we try to allocate a pagetable, we
> may end up shrinking it instead. This comes as a nasty surprise as we
> try to dereference it to fill in the pagetable entries for the object.
> 
> In linus/master this is fixed by pinning the pagetables prior to
> allocation, but that backport is roughly
>  drivers/gpu/drm/i915/i915_gem_gtt.c |   10 ----------
>  1 file changed, 10 deletions(-)

^Funnily that is the size of this backport...

> i.e. unsuitable for stable. Instead we neuter the code that tried to
> free the pagetables.
> 
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99295
> Fixes: 2ce5179fe826 ("drm/i915/gtt: Free unused lower-level page tables")
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Michel Thierry <michel.thierry at intel.com>
> Cc: Mika Kuoppala <mika.kuoppala at intel.com>
> Cc: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
> Cc: Michał Winiarski <michal.winiarski at intel.com>
> Cc: Daniel Vetter <daniel.vetter at intel.com>
> Cc: Jani Nikula <jani.nikula at linux.intel.com>
> Cc: intel-gfx at lists.freedesktop.org
> Cc: <stable at vger.kernel.org> # v4.10+
> Tested-by: Maël Lavault <mael.lavault at protonmail.com>
> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> Signed-off-by: Greg Kroah-Hartman <gregkh at linuxfoundation.org>
> (cherry picked from commit 92220696d5d07525443d9280c08c498e77d0386c)

Not upstream linux tree, where was the patch picked from? I assume one of the
stable trees, but which one?

> Signed-off-by: Rex Tsai (蔡志展) <rex.tsai at canonical.com>
> ---
>  drivers/gpu/drm/i915/i915_gem_gtt.c | 10 ----------
>  1 file changed, 10 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
> index 6924a8e..b2d57d1 100644
> --- a/drivers/gpu/drm/i915/i915_gem_gtt.c
> +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
> @@ -735,10 +735,6 @@ static bool gen8_ppgtt_clear_pt(struct i915_address_space *vm,
>  	GEM_BUG_ON(pte_end > GEN8_PTES);
>  
>  	bitmap_clear(pt->used_ptes, pte, num_entries);
> -	if (USES_FULL_PPGTT(vm->i915)) {
> -		if (bitmap_empty(pt->used_ptes, GEN8_PTES))
> -			return true;
> -	}
>  
>  	pt_vaddr = kmap_px(pt);
>  
> @@ -778,9 +774,6 @@ static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
>  		}
>  	}
>  
> -	if (bitmap_empty(pd->used_pdes, I915_PDES))
> -		return true;
> -
>  	return false;
>  }
>  
> @@ -816,9 +809,6 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm,
>  
>  	mark_tlbs_dirty(ppgtt);
>  
> -	if (bitmap_empty(pdp->used_pdpes, I915_PDPES_PER_PDP(dev_priv)))
> -		return true;
> -
>  	return false;
>  }
>  
> 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20170727/c9e49d44/attachment.sig>