linux.git/drivers/gpu/drm/vmwgfx, branch v5.17

drm/vmwgfx: Fix stale file descriptors on failed usercopy

2022-01-27T15:55:20+00:00

A failing usercopy of the fence_rep object will lead to a stale entry in
the file descriptor table as put_unused_fd() won't release it. This
enables userland to refer to a dangling 'file' object through that still
valid file descriptor, leading to all kinds of use-after-free
exploitation scenarios.

Fix this by deferring the call to fd_install() until after the usercopy
has succeeded.

Fixes: c906965dee22 ("drm/vmwgfx: Add export fence to file descriptor support")
Signed-off-by: Mathias Krause 
Signed-off-by: Zack Rusin 
Signed-off-by: Dave Airlie 
Signed-off-by: Linus Torvalds

drm/vmwgfx: Fix possible usage of an uninitialized variable

2021-12-17T15:35:30+00:00

vmw_user_bo_lookup can fail to lookup user buffers, especially because
the buffer handles come from the userspace. The return value has
to be checked before the buffers are put back.

This was spotted by Dan's Smatch statick checker:
    drivers/gpu/drm/vmwgfx/vmwgfx_bo.c:574 vmw_user_bo_synccpu_release()
	error: uninitialized symbol 'vmw_bo'.

Signed-off-by: Zack Rusin 
Reported-by: Dan Carpenter 
Fixes: 8afa13a0583f ("drm/vmwgfx: Implement DRIVER_GEM")
Reviewed-by: Martin Krastev 
Link: https://patchwork.freedesktop.org/patch/msgid/20211215200224.3693345-1-zack@kde.org
(cherry picked from commit 60c9ecd705be3a28f79d70ea21c3939db668bf46)
Signed-off-by: Thomas Zimmermann

drm/vmwgfx: Remove unused compile options

2021-12-17T15:35:02+00:00

Before the driver had screen targets support we had to disable explicit
bringup of its infrastructure because it was breaking screen objects
support.
Since the implementation of screen targets landed there hasn't been a
reason to explicitly disable it and the options were never used.
Remove of all that unused code.

Signed-off-by: Zack Rusin 
Fixes: d80efd5cb3de ("drm/vmwgfx: Initial DX support")
Reviewed-by: Martin Krastev 
Link: https://patchwork.freedesktop.org/patch/msgid/20211215184147.3688785-3-zack@kde.org
(cherry picked from commit 11343099d5ae6c7411da1425b6b162c89fb5bf10)
Signed-off-by: Thomas Zimmermann

drm/vmwgfx: Remove explicit transparent hugepages support

2021-12-17T15:34:27+00:00

Old versions of the svga device used to export virtual vram, handling of
which was optimized on top of transparent hugepages support. Only very
old devices (OpenGL 2.1 support and earlier) used this code and at this
point performance differences are negligible.

Because the code requires very old hardware versions to run it has
been largely untested and unused for a long time.

Furthermore removal of the ttm hugepages support in:
commit 0d979509539e ("drm/ttm: remove ttm_bo_vm_insert_huge()")
broke the coherency mode in vmwgfx when running with hugepages.

Fixes: 0d979509539e ("drm/ttm: remove ttm_bo_vm_insert_huge()")
Signed-off-by: Zack Rusin 
Cc: Jason Gunthorpe 
Cc: Thomas Hellström 
Cc: Christian König 
Cc: Daniel Vetter 
Reviewed-by: Martin Krastev 
Reviewed-by: Maaz Mombasawala 
Link: https://patchwork.freedesktop.org/patch/msgid/20211215184147.3688785-2-zack@kde.org
(cherry picked from commit 49d535d64d52945e2c874f380705675e20a02b6a)
Signed-off-by: Thomas Zimmermann

drm/vmwgfx: Fix a size_t/long int format specifier mismatch

2021-12-17T15:34:02+00:00

On i386 size_t is of course 32bits and using long int throws warnings,
trivially fix it by using the dedicated size_t format.

This is enough to fix the following warning found by the kernel test
robot:
   drivers/gpu/drm/vmwgfx/vmwgfx_gem.c: In function 'vmw_bo_print_info':
>> drivers/gpu/drm/vmwgfx/vmwgfx_gem.c:230:33: warning: format '%ld'
expects argument of type 'long int', but argument 4 has type 'size_t'
{aka 'unsigned int'} [-Wformat=]
     230 |  seq_printf(m, "\t\t0x%08x: %12ld bytes %s, type = %s",
         |                             ~~~~^
         |                                 |
         |                                 long int
         |                             %12d
     231 |      id, bo->base.base.size, placement, type);
         |          ~~~~~~~~~~~~~~~~~~
         |                       |

Reported-by: kernel test robot 
Signed-off-by: Zack Rusin 
Fixes: 8afa13a0583f ("drm/vmwgfx: Implement DRIVER_GEM")
Reviewed-by: Martin Krastev 
Reviewed-by: Maaz Mombasawala 
Link: https://patchwork.freedesktop.org/patch/msgid/20211215184147.3688785-1-zack@kde.org
(cherry picked from commit 72345114c96b8211ec693b1ce59f95a83a613f1d)
Signed-off-by: Thomas Zimmermann

drm/vmwgfx: Bump the minor version

2021-12-09T18:16:35+00:00

v2: Old userspace doesn't like 3.x.x and we'd like to keep it working,
so lets just bump the minor version until we have no choice.

With GEM, GL4.3, stats and removal of a lot of old code it's time to bump
the minor version of the driver.

Signed-off-by: Zack Rusin 
Reviewed-by: Martin Krastev 
Link: https://patchwork.freedesktop.org/patch/msgid/20211209024924.3298137-1-zack@kde.org

drm/vmwgfx: Remove usage of MOBFMT_RANGE

2021-12-09T18:16:34+00:00

Using MOBFMT_RANGE in the early days of guest backed objects was a major
performance win but that has changed a lot since. There's no more
a performance reason to use MOBFMT_RANGE. The device can/will still
profit from the pages being contiguous but marking them as MOBFMT_RANGE
no longer matters.
Benchmarks (e.g. heaven, valley) show that creating page tables
for mob memory is actually faster than using mobfmt ranges.

Signed-off-by: Zack Rusin 
Reviewed-by: Martin Krastev 
Link: https://patchwork.freedesktop.org/patch/msgid/20211206172620.3139754-12-zack@kde.org

drm/vmwgfx: add support for updating only offsets of constant buffers

2021-12-09T18:16:30+00:00

This adds support for the
SVGA_3D_CMD_DX_SET_VS/PS/GS/HS/DS/CS_CONSTANT_BUFFER_OFFSET commands (which only update
the offset, but don't rebind the buffer), which saves some overhead.

Signed-off-by: Roland Scheidegger 
Reviewed-by: Charmaine Lee 
Reviewed-by: Martin Krastev 
Signed-off-by: Zack Rusin 
Link: https://patchwork.freedesktop.org/patch/msgid/20211206172620.3139754-11-zack@kde.org

drm/vmwgfx: Allow checking for gl43 contexts

2021-12-09T18:16:29+00:00

To make sure we're running on top of hardware that can support
GL4.3 we need to add a way of querying for those capabilities.
DRM_VMW_PARAM_GL43 allows userspace to check for presence of
GL4.3 capable contexts.

Signed-off-by: Zack Rusin 
Reviewed-by: Charmaine Lee 
Link: https://patchwork.freedesktop.org/patch/msgid/20211206172620.3139754-10-zack@kde.org

drm/vmwgfx: support 64 UAVs

2021-12-09T18:16:28+00:00

If the host supports SVGA3D_DEVCAP_GL43, we can handle 64 instead of
just 8 UAVs.
Based on a patch from Roland Scheidegger .

Signed-off-by: Roland Scheidegger 
Signed-off-by: Zack Rusin 
Reviewed-by: Martin Krastev 
Link: https://patchwork.freedesktop.org/patch/msgid/20211206172620.3139754-9-zack@kde.org