One of the costs of superioctl has been the need to perform relocations
inside the kernel. The cost of mapping the buffers to the CPU and writing
data is fairly high, especially if those buffers have been mapped and read
by the GPU.
If we assume that buffers don't move around very often, we can have the
client compute the relocations itself using the previous GPU address. When
that object doesn't move, the kernel can skip computing and writing the
updated data.
Here's a patch which adds a new field to struct drm_bo_info_req called
'presumed_offset', and a new DRM_BO_HINT_PRESUMED_OFFSET that is set when
this field has been filled in by the client.
There are two separate optimizations performed when the presumed_offset is
correct:
1. i915_exec_reloc checks to see if all previous buffer offsets were guessed
correctly. If so, there's no need for it to look at *any* of the
relocations for a buffer. When this happens, it skips the whole
relocation process, simply returning success.
2. i915_apply_reloc checks to see if the target buffer offset was guessed
correctly. If so, it skips mapping the relocatee, computing the
relocation and writing the value. If no relocations are needed, the
relocatee should never be mapped to the CPU, and so the kernel shouldn't
need to wait for any fences to pass.
Add suspend/resume support to the i915 driver. Moves some of the
initialization into the driver load routine, and fixes up places where we
assumed no dev_private existed in some of the cleanup paths. This allows
us to suspend/resume properly even if X isn't running.
This adds the initial i915 superioctl interface. The interface should be
sufficent even if the implementation may needs fixes/optimisations internally
in the drm wrt caching etc.
I should not have renamed this field.
I should not have renamed this field.
I should not have renamed this field.
On the plus side, it was at least binary compatible.
This mod makes the SAREA track plane to pipe mappings and corrects the name of
the plane info variables (they were mislabeled as pipe info since until now all
code assumed a direct mapping between planes and pipes).
It also updates the flip ioctl argument to take a set of planes rather than
pipes, since planes are flipped while pipes generate vblank events.
The data is now in kernel space, copied in/out as appropriate according to the
This results in DRM_COPY_{TO,FROM}_USER going away, and error paths to deal
with those failures. This also means that XFree86 4.2.0 support for i810 DRM
is lost.
As a fallout, replace filp storage with file_priv storage for "unique
identifier of a client" all over the DRM. There is a 1:1 mapping, so this
should be a noop. This could be a minor performance improvement, as everything
on Linux dereferenced filp to get file_priv anyway, while only the mmap ioctls
went the other direction.
This was used to make all ioctl handlers return -errno on linux and errno on
*BSD. Instead, just return -errno in shared code, and flip sign on return from
shared code to *BSD code.
These require that the status page be referenced by a pointer in GTT, rather
than phsyical memory. So, we have the X Server allocate that memory and tell
us the address, instead.
i915_driver_irq_postinstall was forcing vblank interrupts to pipe A when
called with vblank interrupts disabled. This caused vblank interrupts to be
accidentally re-enabled when VT switching the X server. Instead, start the
driver with vblank interrupts enabled on pipe A to support older X servers,
but then leave control over the state to the X server if it is able to do so.
The MI_WAIT_FOR_EVENT instruction does not support waiting for several events
at once, so this should fix the lockups with page flipping when both pipes are
enabled.
Always use dev_priv->sarea_priv->pf_current_page directly. This allows clients
to modify it as well while they hold the HW lock, e.g. in order to sync pages
between pipes.
The assumption is that synchronous flips are not isolated usually, and waiting
for all of them could result in stalling the pipeline for long periods of time.
Also use i915_emit_mi_flush() instead of an old-fashioned way to achieve the
same effect.
Unfortunately, emitting asynchronous flips during vertical blank results in
tearing. So we have to wait for the previous vertical blank and emit a
synchronous flip.
Leave it to the client to wait for the flip to complete when necessary,
but wait for a previous flip to complete before emitting another one. This
should help avoid unnecessary stalling of the ring due to pending flips.
Call i915_do_cleanup_pageflip() unconditionally in preclose.
This means the loop will wait up to ~10ms for ring buffer space to become
available, rather than just however long it takes to check the space 10000
times. This matches other drivers' behavior when waiting for ring buffer/fifo
space.
This uses the core facility to schedule a driver callback that will be called
ASAP after the given vertical blank interrupt with the HW lock held.
(cherry picked from 257771fa29 commit)
0x00 EXE fence. Signals when command stream interpreter has reached the point
where the fence was emitted.
0x01 FLUSH fence. Signals when command stream interpreter has reached the point
where the fence was emitted, and all previous drawing operations have been
completed and flushed.
Implements busy wait (for fastest response time / high CPU) and
lazy wait (User interrupt or timer driven).
i915 vblanks can be generated from either pipe a or b, however a disabled
pipe generates no interrupts. This change allows the X server to select
which pipe generates vblank interrupts.
that particular file. Its contents have changed a good bit since the
original sis code, and the original sis code didn't care much about
attribution since it routinely disclaims Precision Insight/VA Linux
from responsibility. Also, adjust formatting around license headers
(have a comment open immediately before the "Copyright" line, not as a
runon of any previous comments) for automatic processing into FreeBSD,
where /*- is used to signal the beginning of license headers for
automatic compilation of license lists.
understandable: preinit -> load postinit -> (removed) presetup ->
firstopen postsetup -> (removed) open_helper -> open prerelease ->
preclose free_filp_priv -> postclose pretakedown -> lastclose
postcleanup -> unload release -> reclaim_buffers_locked version ->
(removed)
postinit and version were replaced with generic code in the Linux DRM
(drivers now set their version numbers and description in the driver
structure, like on BSD). postsetup wasn't used at all. Fixes the savage
hooks for initializing and tearing down mappings at the right times.
Testing involved at least starting X, running glxgears, killing
glxgears, exiting X, and repeating.
Tested on: FreeBSD (g200, g400, r200, r128) Linux (r200, savage4)
that a device absolutely is, absolutely is not, or may or may not be
AGP. Modify the i915 DRM to use this to force all i9x5 devices to be
"AGP" (even the PCI-e devices).
Reported by: Lukas Hejtmanek
with IOMMUs and such. There is one usage of the forbidden vtophys()
left in drm_scatter.c which will be fixed up soon. This required a KPI
change for drm_pci_alloc/free() to return/use a drm_dma_handle_t that
keeps track of os-specific bits, rather than just passing around the
vaddr/busaddr/size.
Submitted by: Tonnerre Lombard (partially) Tested on: FreeBSD: Rage128
AGP/PCI Linux: Savage4 AGP/PCI
code. Remove the "drv" from sisdrv, as it's unnecessary. Use the
drm_pci functions in i915 instead of per-os implementations of the
same. Avoid whitespace within fields in drm_pciids.txt (one of the r300
definitions), since it breaks the bsd pciids script. Tested on sis,
mga, r128. i915 needs more work.
Keith Packard
Kristian Høgsberg
Robert Noland
Robert Noland
Jiri Slaby
Dave Airlie
Dave Airlie
Eric Anholt
Robert Noland
Dave Airlie
Jesse Barnes
Thomas Hellstrom
Dave Airlie
Michel Dänzer
Wang Zhenyu
Jung-uk Kim
Keith Packard
Nian Wu
Nian Wu
Haihao Xiang
Zou Nan hai
Alan Hourihane
Alan Hourihane
Eric Anholt
Jon Smirl
Ian Romanick