[PATCH 1/2] mm: Handle prefault for size >= 2GB

[PATCH 1/2] mm: Handle prefault for size >= 2GB

[Sagar Arun Kamble]

From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>

i915 gem_exec_reloc subtests *-31 exercise relocation entries object
that is of size >=2GB. This patch will allow prefaulting >=2GB size.

Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
---
 include/linux/pagemap.h | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
index 324c8dba..ec27ba2 100644
--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -538,7 +538,8 @@ static inline void wait_on_page_writeback(struct page *page)
 /*
  * Fault everything in given userspace address range in.
  */
-static inline int fault_in_pages_writeable(char __user *uaddr, int size)
+static inline int fault_in_pages_writeable(char __user *uaddr,
+					   unsigned long size)
 {
 	char __user *end = uaddr + size - 1;
 
@@ -565,7 +566,8 @@ static inline int fault_in_pages_writeable(char __user *uaddr, int size)
 	return 0;
 }
 
-static inline int fault_in_pages_readable(const char __user *uaddr, int size)
+static inline int fault_in_pages_readable(const char __user *uaddr,
+					  unsigned long size)
 {
 	volatile char c;
 	const char __user *end = uaddr + size - 1;
-- 
1.9.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Sagar Arun Kamble]

From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>

With prefaulting extended to support >=2GB, i915 can handle
relocation objects >=2GB. type conversion of length from int to
unsigned long in access_ok was making the validation bail out with
-EFAULT.

Testcase: igt/gem_exec_reloc/wc-31
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
---
 drivers/gpu/drm/i915/i915_gem_execbuffer.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 111b7f8..53a8cc7 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -1172,7 +1172,7 @@ static bool only_mappable_for_reloc(unsigned int flags)
 
 	for (i = 0; i < count; i++) {
 		char __user *ptr = u64_to_user_ptr(exec[i].relocs_ptr);
-		int length; /* limited by fault_in_pages_readable() */
+		unsigned long length;
 
 		if (exec[i].flags & invalid_flags)
 			return -EINVAL;
-- 
1.9.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

✓ Fi.CI.BAT: success for series starting with [1/2] mm: Handle prefault for size >= 2GB

[Patchwork]

== Series Details ==

Series: series starting with [1/2] mm: Handle prefault for size >= 2GB
URL   : https://patchwork.freedesktop.org/series/19533/
State : success

== Summary ==

Series 19533v1 Series without cover letter
https://patchwork.freedesktop.org/api/1.0/series/19533/revisions/1/mbox/

fi-bdw-5557u     total:252  pass:241  dwarn:0   dfail:0   fail:0   skip:11 
fi-bsw-n3050     total:252  pass:213  dwarn:0   dfail:0   fail:0   skip:39 
fi-bxt-j4205     total:252  pass:233  dwarn:0   dfail:0   fail:0   skip:19 
fi-bxt-t5700     total:83   pass:70   dwarn:0   dfail:0   fail:0   skip:12 
fi-byt-j1900     total:252  pass:225  dwarn:0   dfail:0   fail:0   skip:27 
fi-byt-n2820     total:252  pass:221  dwarn:0   dfail:0   fail:0   skip:31 
fi-hsw-4770      total:252  pass:236  dwarn:0   dfail:0   fail:0   skip:16 
fi-hsw-4770r     total:252  pass:236  dwarn:0   dfail:0   fail:0   skip:16 
fi-ilk-650       total:252  pass:202  dwarn:0   dfail:0   fail:0   skip:50 
fi-ivb-3520m     total:252  pass:234  dwarn:0   dfail:0   fail:0   skip:18 
fi-ivb-3770      total:252  pass:234  dwarn:0   dfail:0   fail:0   skip:18 
fi-kbl-7500u     total:252  pass:234  dwarn:0   dfail:0   fail:0   skip:18 
fi-skl-6260u     total:252  pass:242  dwarn:0   dfail:0   fail:0   skip:10 
fi-skl-6700hq    total:252  pass:235  dwarn:0   dfail:0   fail:0   skip:17 
fi-skl-6700k     total:252  pass:230  dwarn:4   dfail:0   fail:0   skip:18 
fi-skl-6770hq    total:252  pass:242  dwarn:0   dfail:0   fail:0   skip:10 
fi-snb-2520m     total:252  pass:224  dwarn:0   dfail:0   fail:0   skip:28 
fi-snb-2600      total:252  pass:223  dwarn:0   dfail:0   fail:0   skip:29 

f48164e93e75dbbe0bb9dac254779eeea08c0657 drm-tip: 2017y-02m-12d-11h-10m-28s UTC integration manifest
9db795c drm/i915: Handle validation of relocation object with length >=2GB
60e32db mm: Handle prefault for size >= 2GB

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3786/
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 1/2] mm: Handle prefault for size >= 2GB

[Joonas Lahtinen]

On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>

https://kernel.org/doc/html/latest/process/submitting-patches.html#the-
canonical-patch-format

# git config --global user.name "Sagar Arun Kamble"

> i915 gem_exec_reloc subtests *-31 exercise relocation entries object
> that is of size >=2GB. This patch will allow prefaulting >=2GB size.
> 
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>

<SNIP>

> @@ -538,7 +538,8 @@ static inline void wait_on_page_writeback(struct page *page)
>  /*
>   * Fault everything in given userspace address range in.
>   */
> -static inline int fault_in_pages_writeable(char __user *uaddr, int size)
> +static inline int fault_in_pages_writeable(char __user *uaddr,
> +					   unsigned long size)
>  {
>  	char __user *end = uaddr + size - 1;

ptrdiff_t would be more appropriate for both functions. With those and
correct "From:", this is;

Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 1/2] mm: Handle prefault for size >= 2GB

[Joonas Lahtinen]

On ma, 2017-02-13 at 12:48 +0200, Joonas Lahtinen wrote:
> On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> > -static inline int fault_in_pages_writeable(char __user *uaddr, int size)
> > +static inline int fault_in_pages_writeable(char __user *uaddr,
> > +					   unsigned long size)
> >  {
> >  	char __user *end = uaddr + size - 1;
> 
> ptrdiff_t would be more appropriate for both functions. With those and
> correct "From:", this is;

Actually, size_t would be more appropriate, I don't think anybody is
using those negative values.

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Joonas Lahtinen]

On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
> 
> With prefaulting extended to support >=2GB, i915 can handle
> relocation objects >=2GB. type conversion of length from int to
> unsigned long in access_ok was making the validation bail out with
> -EFAULT.
> 
> Testcase: igt/gem_exec_reloc/wc-31
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>

With the correct "From:" and with "size_t" as type, this is;

Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Chris Wilson]

On Mon, Feb 13, 2017 at 12:58:57PM +0200, Joonas Lahtinen wrote:
> On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> > From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
> > 
> > With prefaulting extended to support >=2GB, i915 can handle
> > relocation objects >=2GB. type conversion of length from int to
> > unsigned long in access_ok was making the validation bail out with
> > -EFAULT.
> > 
> > Testcase: igt/gem_exec_reloc/wc-31
> > Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
> 
> With the correct "From:" and with "size_t" as type, this is;
> 
> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

My earlier comments still apply in that this is not how I fixed this
years ago, and I still much prefer my fix.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Joonas Lahtinen]

On ma, 2017-02-13 at 11:02 +0000, Chris Wilson wrote:
> On Mon, Feb 13, 2017 at 12:58:57PM +0200, Joonas Lahtinen wrote:
> > 
> > On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> > > 
> > > From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
> > > 
> > > With prefaulting extended to support >=2GB, i915 can handle
> > > relocation objects >=2GB. type conversion of length from int to
> > > unsigned long in access_ok was making the validation bail out with
> > > -EFAULT.
> > > 
> > > Testcase: igt/gem_exec_reloc/wc-31
> > > Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
> > 
> > With the correct "From:" and with "size_t" as type, this is;
> > 
> > Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> 
> My earlier comments still apply in that this is not how I fixed this
> years ago, and I still much prefer my fix.

Where are the earlier comments? I must have missed them.

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Chris Wilson]

On Mon, Feb 13, 2017 at 02:25:29PM +0200, Joonas Lahtinen wrote:
> On ma, 2017-02-13 at 11:02 +0000, Chris Wilson wrote:
> > On Mon, Feb 13, 2017 at 12:58:57PM +0200, Joonas Lahtinen wrote:
> > > 
> > > On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
> > > > 
> > > > From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
> > > > 
> > > > With prefaulting extended to support >=2GB, i915 can handle
> > > > relocation objects >=2GB. type conversion of length from int to
> > > > unsigned long in access_ok was making the validation bail out with
> > > > -EFAULT.
> > > > 
> > > > Testcase: igt/gem_exec_reloc/wc-31
> > > > Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> > > > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > > > Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
> > > 
> > > With the correct "From:" and with "size_t" as type, this is;
> > > 
> > > Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> > 
> > My earlier comments still apply in that this is not how I fixed this
> > years ago, and I still much prefer my fix.
> 
> Where are the earlier comments? I must have missed them.

I replied to the first time this patch landed in my inbox, which might
have been trybotted.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 1/2] mm: Handle prefault for size >= 2GB

[Kamble, Sagar A]

On 2/13/2017 4:18 PM, Joonas Lahtinen wrote:
> On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
>> From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
> https://kernel.org/doc/html/latest/process/submitting-patches.html#the-
> canonical-patch-format
>
> # git config --global user.name "Sagar Arun Kamble"
Thanks Joonas. Even though ~/.gitconfig has user.name set properly 
"From" is not getting set properly.
Will find out the problem.
>
>> i915 gem_exec_reloc subtests *-31 exercise relocation entries object
>> that is of size >=2GB. This patch will allow prefaulting >=2GB size.
>>
>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Cc: Chris Wilson <chris@chris-wilson.co.uk>
>> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
> <SNIP>
>
>> @@ -538,7 +538,8 @@ static inline void wait_on_page_writeback(struct page *page)
>>   /*
>>    * Fault everything in given userspace address range in.
>>    */
>> -static inline int fault_in_pages_writeable(char __user *uaddr, int size)
>> +static inline int fault_in_pages_writeable(char __user *uaddr,
>> +					   unsigned long size)
>>   {
>>   	char __user *end = uaddr + size - 1;
> ptrdiff_t would be more appropriate for both functions. With those and
> correct "From:", this is;
>
> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
>
> Regards, Joonas

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 2/2] drm/i915: Handle validation of relocation object with length >=2GB

[Kamble, Sagar A]

[-- Attachment #1.1: Type: text/plain, Size: 1607 bytes --]



On 2/13/2017 6:03 PM, Chris Wilson wrote:
> On Mon, Feb 13, 2017 at 02:25:29PM +0200, Joonas Lahtinen wrote:
>> On ma, 2017-02-13 at 11:02 +0000, Chris Wilson wrote:
>>> On Mon, Feb 13, 2017 at 12:58:57PM +0200, Joonas Lahtinen wrote:
>>>> On ma, 2017-02-13 at 12:14 +0530, Sagar Arun Kamble wrote:
>>>>> From: "sagar.a.kamble@intel.com" <sagar.a.kamble@intel.com>
>>>>>
>>>>> With prefaulting extended to support >=2GB, i915 can handle
>>>>> relocation objects >=2GB. type conversion of length from int to
>>>>> unsigned long in access_ok was making the validation bail out with
>>>>> -EFAULT.
>>>>>
>>>>> Testcase: igt/gem_exec_reloc/wc-31
>>>>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>>>>> Cc: Chris Wilson <chris@chris-wilson.co.uk>
>>>>> Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
>>>> With the correct "From:" and with "size_t" as type, this is;
>>>>
>>>> Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
>>> My earlier comments still apply in that this is not how I fixed this
>>> years ago, and I still much prefer my fix.
>> Where are the earlier comments? I must have missed them.
> I replied to the first time this patch landed in my inbox, which might
> have been trybotted.
> -Chris
Chris, I thought those comments were about relocs_total and relocs_max 
values.
Pasting below inputs from trybot.
------------

The abi is limited to 32b max relocs.

Please see
https://cgit.freedesktop.org/~ickle/linux-2.6/commit/?h=prescheduler&id=bcf3bca07f21a4c0c754ccecca81228c28437906
(a patch that is years overdue) in which this just evaporates.
-Chris
---------




[-- Attachment #1.2: Type: text/html, Size: 3250 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx