* does DRM_GPUSVM require hugepages?
@ 2026-05-03 0:03 Randy Dunlap
2026-05-06 1:42 ` Dave Airlie
0 siblings, 1 reply; 3+ messages in thread
From: Randy Dunlap @ 2026-05-03 0:03 UTC (permalink / raw)
To: dri-devel@lists.freedesktop.org, David Airlie, Simona Vetter,
intel-xe@lists.freedesktop.org
Cc: linux-next
Hi,
In linux-next-20260430, I'm seeing build errors on some ARCHes when
DRM_XE=m (probably just "is set") but HAVE_ARCH_TRANSPARENT_HUGEPAGE
is not set (not selected) by the ARCH.
Does drm_gpusvm.c require hugepages?
Or possibly DRM_XE needs some kind of restriction?
I see build failures on ARCH=alpha, ARCH=openrisc, ARCH=mips (32-bit
or 64-bit), and ARCH=arm (without ARM_LPAE set).
Example:
../drivers/gpu/drm/drm_gpusvm.c: In function 'drm_gpusvm_get_pages':
./../include/linux/compiler_types.h:699:45: error: call to '__compiletime_assert_361' declared with attribute error: BUILD_BUG failed
699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^
./../include/linux/compiler_types.h:680:25: note: in definition of macro '__compiletime_assert'
680 | prefix ## suffix(); \
| ^~~~~~
./../include/linux/compiler_types.h:699:9: note: in expansion of macro '_compiletime_assert'
699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
| ^~~~~~~~~~~~~~~~~~~
../include/linux/build_bug.h:40:37: note: in expansion of macro 'compiletime_assert'
40 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
| ^~~~~~~~~~~~~~~~~~
../include/linux/build_bug.h:60:21: note: in expansion of macro 'BUILD_BUG_ON_MSG'
60 | #define BUILD_BUG() BUILD_BUG_ON_MSG(1, "BUILD_BUG failed")
| ^~~~~~~~~~~~~~~~
../include/linux/huge_mm.h:113:28: note: in expansion of macro 'BUILD_BUG'
113 | #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; })
| ^~~~~~~~~
../include/linux/huge_mm.h:120:35: note: in expansion of macro 'HPAGE_PMD_SHIFT'
120 | #define HPAGE_PMD_SIZE ((1UL) << HPAGE_PMD_SHIFT)
| ^~~~~~~~~~~~~~~
../drivers/gpu/drm/drm_gpusvm.c:1560:52: note: in expansion of macro 'HPAGE_PMD_SIZE'
1560 | HPAGE_PMD_SIZE ?
| ^~~~~~~~~~~~~~
make[6]: *** [../scripts/Makefile.build:289: drivers/gpu/drm/drm_gpusvm.o] Error 1
where include/linux/huge_mm.h:113: says:
#ifdef CONFIG_PGTABLE_HAS_HUGE_LEAVES
#define HPAGE_PMD_SHIFT PMD_SHIFT
#define HPAGE_PUD_SHIFT PUD_SHIFT
#else
#define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; }) <<<<<<<<<<<<<<<<<< line 113
#define HPAGE_PUD_SHIFT ({ BUILD_BUG(); 0; })
#endif
--
~Randy
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: does DRM_GPUSVM require hugepages?
2026-05-03 0:03 does DRM_GPUSVM require hugepages? Randy Dunlap
@ 2026-05-06 1:42 ` Dave Airlie
2026-05-06 2:39 ` Matthew Brost
0 siblings, 1 reply; 3+ messages in thread
From: Dave Airlie @ 2026-05-06 1:42 UTC (permalink / raw)
To: Randy Dunlap, Matthew Brost, Thomas Hellström
Cc: dri-devel@lists.freedesktop.org, Simona Vetter,
intel-xe@lists.freedesktop.org, linux-next
On Sun, 3 May 2026 at 10:03, Randy Dunlap <rdunlap@infradead.org> wrote:
>
> Hi,
>
> In linux-next-20260430, I'm seeing build errors on some ARCHes when
> DRM_XE=m (probably just "is set") but HAVE_ARCH_TRANSPARENT_HUGEPAGE
> is not set (not selected) by the ARCH.
I'm seeing this locally on my arm32 builds, Matthew, Thomas?
Dave.
>
> Does drm_gpusvm.c require hugepages?
> Or possibly DRM_XE needs some kind of restriction?
>
> I see build failures on ARCH=alpha, ARCH=openrisc, ARCH=mips (32-bit
> or 64-bit), and ARCH=arm (without ARM_LPAE set).
>
> Example:
> ../drivers/gpu/drm/drm_gpusvm.c: In function 'drm_gpusvm_get_pages':
> ./../include/linux/compiler_types.h:699:45: error: call to '__compiletime_assert_361' declared with attribute error: BUILD_BUG failed
> 699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> | ^
> ./../include/linux/compiler_types.h:680:25: note: in definition of macro '__compiletime_assert'
> 680 | prefix ## suffix(); \
> | ^~~~~~
> ./../include/linux/compiler_types.h:699:9: note: in expansion of macro '_compiletime_assert'
> 699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> | ^~~~~~~~~~~~~~~~~~~
> ../include/linux/build_bug.h:40:37: note: in expansion of macro 'compiletime_assert'
> 40 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
> | ^~~~~~~~~~~~~~~~~~
> ../include/linux/build_bug.h:60:21: note: in expansion of macro 'BUILD_BUG_ON_MSG'
> 60 | #define BUILD_BUG() BUILD_BUG_ON_MSG(1, "BUILD_BUG failed")
> | ^~~~~~~~~~~~~~~~
> ../include/linux/huge_mm.h:113:28: note: in expansion of macro 'BUILD_BUG'
> 113 | #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; })
> | ^~~~~~~~~
> ../include/linux/huge_mm.h:120:35: note: in expansion of macro 'HPAGE_PMD_SHIFT'
> 120 | #define HPAGE_PMD_SIZE ((1UL) << HPAGE_PMD_SHIFT)
> | ^~~~~~~~~~~~~~~
> ../drivers/gpu/drm/drm_gpusvm.c:1560:52: note: in expansion of macro 'HPAGE_PMD_SIZE'
> 1560 | HPAGE_PMD_SIZE ?
> | ^~~~~~~~~~~~~~
> make[6]: *** [../scripts/Makefile.build:289: drivers/gpu/drm/drm_gpusvm.o] Error 1
>
> where include/linux/huge_mm.h:113: says:
>
> #ifdef CONFIG_PGTABLE_HAS_HUGE_LEAVES
> #define HPAGE_PMD_SHIFT PMD_SHIFT
> #define HPAGE_PUD_SHIFT PUD_SHIFT
> #else
> #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; }) <<<<<<<<<<<<<<<<<< line 113
> #define HPAGE_PUD_SHIFT ({ BUILD_BUG(); 0; })
> #endif
>
> --
> ~Randy
>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: does DRM_GPUSVM require hugepages?
2026-05-06 1:42 ` Dave Airlie
@ 2026-05-06 2:39 ` Matthew Brost
0 siblings, 0 replies; 3+ messages in thread
From: Matthew Brost @ 2026-05-06 2:39 UTC (permalink / raw)
To: Dave Airlie
Cc: Randy Dunlap, Thomas Hellström,
dri-devel@lists.freedesktop.org, Simona Vetter,
intel-xe@lists.freedesktop.org, linux-next
On Wed, May 06, 2026 at 11:42:13AM +1000, Dave Airlie wrote:
> On Sun, 3 May 2026 at 10:03, Randy Dunlap <rdunlap@infradead.org> wrote:
> >
> > Hi,
> >
> > In linux-next-20260430, I'm seeing build errors on some ARCHes when
> > DRM_XE=m (probably just "is set") but HAVE_ARCH_TRANSPARENT_HUGEPAGE
> > is not set (not selected) by the ARCH.
>
> I'm seeing this locally on my arm32 builds, Matthew, Thomas?
>
Thanks. Yes, this looks like an issue we fix in this cycle.
Matt
> Dave.
>
> >
> > Does drm_gpusvm.c require hugepages?
> > Or possibly DRM_XE needs some kind of restriction?
> >
> > I see build failures on ARCH=alpha, ARCH=openrisc, ARCH=mips (32-bit
> > or 64-bit), and ARCH=arm (without ARM_LPAE set).
> >
> > Example:
> > ../drivers/gpu/drm/drm_gpusvm.c: In function 'drm_gpusvm_get_pages':
> > ./../include/linux/compiler_types.h:699:45: error: call to '__compiletime_assert_361' declared with attribute error: BUILD_BUG failed
> > 699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> > | ^
> > ./../include/linux/compiler_types.h:680:25: note: in definition of macro '__compiletime_assert'
> > 680 | prefix ## suffix(); \
> > | ^~~~~~
> > ./../include/linux/compiler_types.h:699:9: note: in expansion of macro '_compiletime_assert'
> > 699 | _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> > | ^~~~~~~~~~~~~~~~~~~
> > ../include/linux/build_bug.h:40:37: note: in expansion of macro 'compiletime_assert'
> > 40 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
> > | ^~~~~~~~~~~~~~~~~~
> > ../include/linux/build_bug.h:60:21: note: in expansion of macro 'BUILD_BUG_ON_MSG'
> > 60 | #define BUILD_BUG() BUILD_BUG_ON_MSG(1, "BUILD_BUG failed")
> > | ^~~~~~~~~~~~~~~~
> > ../include/linux/huge_mm.h:113:28: note: in expansion of macro 'BUILD_BUG'
> > 113 | #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; })
> > | ^~~~~~~~~
> > ../include/linux/huge_mm.h:120:35: note: in expansion of macro 'HPAGE_PMD_SHIFT'
> > 120 | #define HPAGE_PMD_SIZE ((1UL) << HPAGE_PMD_SHIFT)
> > | ^~~~~~~~~~~~~~~
> > ../drivers/gpu/drm/drm_gpusvm.c:1560:52: note: in expansion of macro 'HPAGE_PMD_SIZE'
> > 1560 | HPAGE_PMD_SIZE ?
> > | ^~~~~~~~~~~~~~
> > make[6]: *** [../scripts/Makefile.build:289: drivers/gpu/drm/drm_gpusvm.o] Error 1
> >
> > where include/linux/huge_mm.h:113: says:
> >
> > #ifdef CONFIG_PGTABLE_HAS_HUGE_LEAVES
> > #define HPAGE_PMD_SHIFT PMD_SHIFT
> > #define HPAGE_PUD_SHIFT PUD_SHIFT
> > #else
> > #define HPAGE_PMD_SHIFT ({ BUILD_BUG(); 0; }) <<<<<<<<<<<<<<<<<< line 113
> > #define HPAGE_PUD_SHIFT ({ BUILD_BUG(); 0; })
> > #endif
> >
> > --
> > ~Randy
> >
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-05-06 2:39 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-05-03 0:03 does DRM_GPUSVM require hugepages? Randy Dunlap
2026-05-06 1:42 ` Dave Airlie
2026-05-06 2:39 ` Matthew Brost
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox