linux

mirror of https://github.com/torvalds/linux synced 2024-10-24 04:06:04 +00:00

History

Chris Wilson d18b961903 drm/i915: Fix incoherence with fence updates on Sandybridge+ This hopefully fixes the root cause behind the workaround added in commit `25ff1195f8` Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Thu Apr 4 21:31:03 2013 +0100 drm/i915: Workaround incoherence between fences and LLC across multiple CPUs Thanks to further investigation by Jon Bloomfield, he realised that the 64-bit register might be broken up by the hardware into two 32-bit writes (a problem we have encountered elsewhere). This non-atomicity would then cause an issue where a second thread would see an intermediate register state (new high dword, old low dword), and this register would randomly be used in preference to its own thread register. This would cause the second thread to read from and write into a fairly random tiled location. Breaking the operation into 3 explicit 32-bit updates (first disable the fence, poke the upper bits, then poke the lower bits and enable) ensures that, given proper serialisation between the 32-bit register write and the memory transfer, that the fence value is always consistent. Armed with this knowledge, we can explain how the previous workaround work. The key to the corruption is that a second thread sees an erroneous fence register that conflicts and overrides its own. By serialising the fence update across all CPUs, we have a small window where no GTT access is occurring and so hide the potential corruption. This also leads to the conclusion that the earlier workaround was incomplete. v2: Be overly paranoid about the order in which fence updates become visible to the GPU to make really sure that we turn the fence off before doing the update, and then only switch the fence on afterwards. Signed-off-by: Jon Bloomfield <jon.bloomfield@intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Carsten Emde <C.Emde@osadl.org> Cc: stable@vger.kernel.org Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>		2013-07-10 14:41:46 +02:00
..
ast	drm/ast: inline reservations	2013-06-28 12:04:04 +10:00
cirrus	drm/cirrus: inline reservations	2013-06-28 12:04:05 +10:00
exynos	drm: Drop all the stub gamma_get, gamma_set, load_lut functions from drivers	2013-06-17 19:42:47 +10:00
gma500	Merge branch 'gma500-fixes' of git://github.com/patjak/drm-gma500 into drm-fixes	2013-06-11 08:16:10 +10:00
i2c
i810
i915	drm/i915: Fix incoherence with fence updates on Sandybridge+	2013-07-10 14:41:46 +02:00
mga
mgag200	drm/mgag200: inline reservations	2013-06-28 12:04:06 +10:00
nouveau	drm/ttm: convert to the reservation api	2013-06-28 12:04:01 +10:00
omapdrm	drm/gem: add mutex lock when using drm_gem_mmap_obj	2013-06-28 12:30:15 +10:00
qxl	drm/ttm: convert to the reservation api	2013-06-28 12:04:01 +10:00
r128
radeon	drm/radeon: get rid of ttm_bo_is_reserved usage	2013-06-28 12:04:12 +10:00
rcar-du	drm: Renesas R-Car Display Unit DRM driver	2013-06-27 10:08:04 +10:00
savage
shmobile	drm/shmobile: Enable compilation on all ARM platforms	2013-06-20 10:07:14 +02:00
sis
tdfx
tilcdc	drm/tilcdc: Clear bits of register we're going to set.	2013-06-28 09:13:00 +10:00
ttm	drm/ttm: get rid of ttm_bo_is_reserved usage	2013-06-28 12:04:10 +10:00
udl	drm: Drop all the stub gamma_get, gamma_set, load_lut functions from drivers	2013-06-17 19:42:47 +10:00
via
vmwgfx	drm/vmwgfx: get rid of ttm_bo_is_reserved usage	2013-06-28 12:04:14 +10:00
ati_pcigart.c
drm_agpsupport.c
drm_auth.c
drm_buffer.c
drm_bufs.c	drm: Don't leak phys_wc "handles" to userspace	2013-05-31 13:37:39 +10:00
drm_cache.c
drm_context.c
drm_crtc.c	drm: add hotspot support for cursors.	2013-06-28 09:13:39 +10:00
drm_crtc_helper.c	Revert "drm: kms_helper: don't lose hotplug event"	2013-06-28 20:31:34 +10:00
drm_debugfs.c
drm_dma.c
drm_dp_helper.c
drm_drv.c	drm: add hotspot support for cursors.	2013-06-28 09:13:39 +10:00
drm_edid.c	drm/edid: Don't print messages regarding stereo or csync by default	2013-06-27 21:15:44 +10:00
drm_edid_load.c	drm: Cocci spatch "memdup.spatch"	2013-06-11 08:29:22 +10:00
drm_encoder_slave.c
drm_fb_cma_helper.c
drm_fb_helper.c	drm/fb-helper: Make load_lut and gamma_set/gamma_get hooks optional	2013-06-17 19:42:47 +10:00
drm_fops.c
drm_gem.c	drm/gem: fix not to assign error value to gem name	2013-06-28 12:31:23 +10:00
drm_gem_cma_helper.c	drm/gem: add mutex lock when using drm_gem_mmap_obj	2013-06-28 12:30:15 +10:00
drm_global.c
drm_hashtab.c
drm_info.c
drm_ioc32.c
drm_ioctl.c	drm: Don't leak phys_wc "handles" to userspace	2013-05-31 13:37:39 +10:00
drm_irq.c	drm: fix a use-after-free when GPU acceleration disabled	2013-06-03 19:12:04 +10:00
drm_lock.c
drm_memory.c
drm_mm.c
drm_modes.c	drm: Sort connector modes based on vrefresh	2013-06-11 08:35:51 +10:00
drm_pci.c
drm_platform.c
drm_prime.c	drm/prime: add return check for dma_buf_fd	2013-06-28 14:43:04 +10:00
drm_proc.c
drm_rect.c
drm_scatter.c
drm_stub.c	drm: drm_stub: Fixing return value if driver master_set call failed	2013-06-27 21:03:16 +10:00
drm_sysfs.c
drm_trace.h
drm_trace_points.c
drm_usb.c
drm_vm.c	drm_vm: drop explicit VM_IO setting	2013-06-24 06:26:50 +10:00
Kconfig	drm/nouveau: always select ACPI_VIDEO if ACPI is enabled.	2013-06-28 11:56:22 +10:00
Makefile	drm: Renesas R-Car Display Unit DRM driver	2013-06-27 10:08:04 +10:00
README.drm

README.drm

************************************************************
* For the very latest on DRI development, please see:      *
*     http://dri.freedesktop.org/                          *
************************************************************

The Direct Rendering Manager (drm) is a device-independent kernel-level
device driver that provides support for the XFree86 Direct Rendering
Infrastructure (DRI).

The DRM supports the Direct Rendering Infrastructure (DRI) in four major
ways:

    1. The DRM provides synchronized access to the graphics hardware via
       the use of an optimized two-tiered lock.

    2. The DRM enforces the DRI security policy for access to the graphics
       hardware by only allowing authenticated X11 clients access to
       restricted regions of memory.

    3. The DRM provides a generic DMA engine, complete with multiple
       queues and the ability to detect the need for an OpenGL context
       switch.

    4. The DRM is extensible via the use of small device-specific modules
       that rely extensively on the API exported by the DRM module.


Documentation on the DRI is available from:
    http://dri.freedesktop.org/wiki/Documentation
    http://sourceforge.net/project/showfiles.php?group_id=387
    http://dri.sourceforge.net/doc/

For specific information about kernel-level support, see:

    The Direct Rendering Manager, Kernel Support for the Direct Rendering
    Infrastructure
    http://dri.sourceforge.net/doc/drm_low_level.html

    Hardware Locking for the Direct Rendering Infrastructure
    http://dri.sourceforge.net/doc/hardware_locking_low_level.html

    A Security Analysis of the Direct Rendering Infrastructure
    http://dri.sourceforge.net/doc/security_low_level.html