Matthew Brost [Thu, 9 Sep 2021 16:47:32 +0000 (09:47 -0700)]
drm/i915/selftests: Add initial GuC selftest for scrubbing lost G2H
While debugging an issue with full GT resets I went down a rabbit hole
thinking the scrubbing of lost G2H wasn't working correctly. This proved
to be incorrect as this was working just fine but this chase inspired me
to write a selftest to prove that this works. This simple selftest
injects errors dropping various G2H and then issues a full GT reset
proving that the scrubbing of these G2H doesn't blow up.
v2:
(Daniel Vetter)
- Use ifdef instead of macros for selftests
v3:
(Checkpatch)
- A space after 'switch' statement
v4:
(Daniele)
- A comment saying GT won't idle if G2H are lost
Matthew Brost [Thu, 9 Sep 2021 16:47:31 +0000 (09:47 -0700)]
drm/i915/guc: Copy whole golden context, set engine state size of subset
When the GuC does a media reset, it copies a golden context state back
into the corrupted context's state. The address of the golden context
and the size of the engine state restore are passed in via the GuC ADS.
The i915 had a bug where it passed in the whole size of the golden
context, not the size of the engine state to restore resulting in a
memory corruption.
Also copy the entire golden context on init rather than just the engine
state that is restored.
v2 (Daniele): use defines to avoid duplicated const variables (John).
Fixes: 5d0ff731bbbe ("drm/i915/guc: Add golden context to GuC ADS") Signed-off-by: Matthew Brost <matthew.brost@intel.com> Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: John Harrison <John.C.Harrison@Intel.com> Reviewed-by: John Harrison <John.C.Harrison@Intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-11-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:29 +0000 (09:47 -0700)]
drm/i915/guc: Kick tasklet after queuing a request
Kick tasklet after queuing a request so it submitted in a timely manner.
Fixes: 750a049f8ce3 ("drm/i915/guc: Implement GuC context operations for new inteface") Signed-off-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-9-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:28 +0000 (09:47 -0700)]
Revert "drm/i915/gt: Propagate change in error status to children on unhold"
Propagating errors to dependent fences is broken and can lead to errors
from one client ending up in another. In commit 901f31c70d8d ("Revert
"drm/i915: Propagate errors on awaiting already signaled fences""), we
attempted to get rid of fence error propagation but missed the case
added in commit f59fbc86c96e ("drm/i915/gt: Propagate change in error
status to children on unhold"). Revert that one too. This error was
found by an up-and-coming selftest which triggers a reset during
request cancellation and verifies that subsequent requests complete
successfully.
Matthew Brost [Thu, 9 Sep 2021 16:47:27 +0000 (09:47 -0700)]
drm/i915/guc: Workaround reset G2H is received after schedule done G2H
If the context is reset as a result of the request cancellation the
context reset G2H is received after schedule disable done G2H which is
the wrong order. The schedule disable done G2H release the waiting
request cancellation code which resubmits the context. This races
with the context reset G2H which also wants to resubmit the context but
in this case it really should be a NOP as request cancellation code owns
the resubmit. Use some clever tricks of checking the context state to
seal this race until the GuC firmware is fixed.
v2:
(Checkpatch)
- Fix typos
v3:
(Daniele)
- State that is a bug in the GuC firmware
Fixes: 94d4cdfc27c9 ("drm/i915/guc: Support request cancellation") Signed-off-by: Matthew Brost <matthew.brost@intel.com> Cc: <stable@vger.kernel.org> Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-7-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:26 +0000 (09:47 -0700)]
drm/i915/guc: Process all G2H message at once in work queue
Rather than processing 1 G2H at a time and re-queuing the work queue if
more messages exist, process all the G2H in a single pass of the work
queue.
Signed-off-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-6-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:25 +0000 (09:47 -0700)]
drm/i915/guc: Don't drop ce->guc_active.lock when unwinding context
Don't drop ce->guc_active.lock when unwinding a context after reset.
At one point we had to drop this because of a lock inversion but that is
no longer the case. It is much safer to hold the lock so let's do that.
Fixes: 81ab2f1fdb42 ("drm/i915/guc: Reset implementation for new GuC interface") Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: Matthew Brost <matthew.brost@intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-5-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:24 +0000 (09:47 -0700)]
drm/i915/guc: Unwind context requests in reverse order
When unwinding requests on a reset context, if other requests in the
context are in the priority list the requests could be resubmitted out
of seqno order. Traverse the list of active requests in reverse and
append to the head of the priority list to fix this.
Fixes: 81ab2f1fdb42 ("drm/i915/guc: Reset implementation for new GuC interface") Signed-off-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-4-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:23 +0000 (09:47 -0700)]
drm/i915/guc: Fix outstanding G2H accounting
A small race that could result in incorrect accounting of the number
of outstanding G2H. Basically prior to this patch we did not increment
the number of outstanding G2H if we encoutered a GT reset while sending
a H2G. This was incorrect as the context state had already been updated
to anticipate a G2H response thus the counter should be incremented.
As part of this change we remove a legacy (now unused) path that was the
last caller requiring a G2H response that was not guaranteed to loop.
This allows us to simplify the accounting as we don't need to handle the
case where the send fails due to the channel being busy.
Also always use helper when decrementing this value.
v2 (Daniele): update GEM_BUG_ON check, pull in dead code removal from
later patch, remove loop param from context_deregister.
Fixes: 6f53e7ffdf42 ("drm/i915/guc: Ensure G2H response has space in buffer") Signed-off-by: Matthew Brost <matthew.brost@intel.com> Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: <stable@vger.kernel.org> Reviewed-by: John Harrison <John.C.Harrison@Intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210909164744.31249-3-matthew.brost@intel.com
Matthew Brost [Thu, 9 Sep 2021 16:47:22 +0000 (09:47 -0700)]
drm/i915/guc: Fix blocked context accounting
Prior to this patch the blocked context counter was cleared on
init_sched_state (used during registering a context & resets) which is
incorrect. This state needs to be persistent or the counter can read the
incorrect value resulting in scheduling never getting enabled again.
drm/i915: Use Transparent Hugepages when IOMMU is enabled
Usage of Transparent Hugepages was disabled in 1a80e9cbf9f8
("drm/i915: Disable THP until we have a GPU read BW W/A"), but since it
appears majority of performance regressions reported with an enabled IOMMU
can be almost eliminated by turning them on, lets just do that.
To err on the side of safety we keep the current default in cases where
IOMMU is not active, and only when it is default to the "huge=within_size"
mode. Although there probably would be wins to enable them throughout,
more extensive testing across benchmarks and platforms would need to be
done.
With the patch and IOMMU enabled my local testing on a small Skylake part
shows OglVSTangent regression being reduced from ~14% (IOMMU on versus
IOMMU off) to ~2% (same comparison but with THP on).
More detailed testing done in the below referenced Gitlab issue by Eero:
Matthew Auld [Mon, 6 Sep 2021 09:17:29 +0000 (10:17 +0100)]
drm/i915/selftests: fixup igt_shrink_thp
Since the object might still be active here, the shrink_all will simply
ignore it, which blows up in the test, since the pages will still be
there. Currently THP is disabled which should result in the test being
skipped, but if we ever re-enable THP we might start seeing the failure.
Fix this by forcing I915_SHRINK_ACTIVE.
v2: Some machine in the shard runs doesn't seem to have any available
swap when running this test. Try to handle this.
Matthew Auld [Fri, 3 Sep 2021 15:53:17 +0000 (16:53 +0100)]
drm/i915/gtt: add some flushing for the 64K GTT path
If we need to mark the PDE as operating in 64K GTT mode, we should be
paranoid and flush the extra writes, like we already do for the PTEs. On
some platforms the clflush can apparently add the just the right amount
of magical delay to force the GPU to see the updated entry.
drm/i915/gt: Add separate MOCS table for Gen12 devices other than TGL/RKL
MOCS table of TGL/RKL has MOCS[1] set to L3_UC.
While for other gen12 devices we need to set MOCS[1] as L3_WB,
So adding a new MOCS table for other gen 12 devices eg. ADL.
Fixes: 3fc55445bd1b ("drm/i915/gt: Initialize unused MOCS entries with device specific values") Cc: Matt Roper <matthew.d.roper@intel.com> Signed-off-by: Ayaz A Siddiqui <ayaz.siddiqui@intel.com> Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
[mattrope: fix whitespace error] Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210907171639.1221287-1-ayaz.siddiqui@intel.com
Removing force probe protection from ADLS platform. Did
not observe warnings, errors, flickering or any visual
defects while doing ordinary tasks like browsing and
editing documents in a two monitor setup.
For more info drm-tip idle run results :
https://intel-gfx-ci.01.org/tree/drm-tip/bat-all.html?
Daniel Vetter [Thu, 2 Sep 2021 14:20:57 +0000 (16:20 +0200)]
drm/i915: Stop rcu support for i915_address_space
The full audit is quite a bit of work:
- i915_dpt has very simple lifetime (somehow we create a display pagetable vm
per object, so its _very_ simple, there's only ever a single vma in there),
and uses i915_vm_close(), which internally does a i915_vm_put(). No rcu.
Aside: wtf is i915_dpt doing in the intel_display.c garbage collector as a new
feature, instead of added as a separate file with some clean-ish interface.
Also, i915_dpt unfortunately re-introduces some coding patterns from
pre-dma_resv_lock conversion times.
- i915_gem_proto_ctx is fully refcounted and no rcu, all protected by
fpriv->proto_context_lock.
- i915_gem_context is itself rcu protected, and that might leak to anything it
points at. Before
drm/i915/gem: Don't allow changing the VM on running contexts (v4)
Note that we drop the vm refcount before the final release of the gem context
refcount, so this is all very dangerous even without rcu. Note that aside from
later on creating new engines (a defunct feature) and debug output we're never
looked at gem_ctx->vm for anything functional, hence why this is ok.
Fingers crossed.
Preceeding patches removed all vestiges of rcu use from gem_ctx->vm
derferencing to make it clear it's really not used.
The commit message is somewhat entertaining because it fails to
mention this fact completely, and compensates that by an in-commit
changelog entry that claims that ctx->vm is protected by ctx->mutex.
Which was the case _before_ this commit, but no longer after it.
- intel_context holds a full reference. Unfortunately intel_context is also rcu
protected and the reference to the ->vm is dropped before the
rcu barrier - only the kfree is delayed. So again we need to check
whether that leaks anywhere on the intel_context->vm. RCU is only
used to protect intel_context sitting on the breadcrumb lists, which
don't look at the vm anywhere, so we are fine.
Nothing else relies on rcu protection of intel_context and hence is
fully protected by the kref refcount alone, which protects
intel_context->vm in turn.
given some credence to my claim that I've actually caught them all.
- drm_i915_gem_object's shares_resv_from pointer has a full refcount to the
dma_resv, which is a sub-refcount that's released after the final
i915_vm_put() has been called. Safe.
Aside: Maybe we should have a struct dma_resv_shared which is just dma_resv +
kref as a stand-alone thing. It's a pretty useful pattern which other drivers
might want to copy.
- the fpriv->vm_xa was relying on rcu_read_lock for lookup, but that
was updated in a prep patch too to just be a spinlock-protected
lookup.
- intel_gt->vm is set at driver load in intel_gt_init() and released
in intel_gt_driver_release(). There seems to be some issue that
in some error paths this is called twice, but otherwise no rcu to be
found anywhere. This was added in the below commit, which
unfortunately doesn't explain why this complication exists.
The proper fix most likely for this is to start using drmm_ at large
scale, but that's also huge amounts of work.
- i915_vma->vm is some real pain, because rcu is rcu protected, at
least in the vma lookup in the context lookup cache in
eb_lookup_vma(). This was added in
drm/i915/gem: Avoid gem_context->mutex for simple vma lookup
the locking was changed from dev->struct_mutex to rcu, which added
the requirement to rcu protect i915_vma. Somehow this was missed in
review (or I'm completely blind).
Irrespective of all that the vma lookup cache rcu_read_lock grabs a
full reference of the vma and the rcu doesn't leak further. So no
impact on i915_address_space from that.
I have not found any other rcu use for i915_vma, but given that it
seems broken I also didn't bother to do a careful in-depth audit.
Alltogether there's nothing left in-tree anymore which requires that a
pointer deref to an i915_address_space is safe undre rcu_read_lock
only.
rcu protection of i915_address_space was introduced in
drm/i915/gtt: Defer address space cleanup to an RCU worker
by mixing up a bugfixing (i915_address_space needs to be released from
a worker) with enabling rcu support. The commit message also seems
somewhat confused, because it talks about cleanup of WC pages
requiring sleep, while the code and linked bugzilla are about a
requirement to take dev->struct_mutex (which yes sleeps but it's a
much more specific problem). Since final kref_put can be called from
pretty much anywhere (including hardirq context through the
scheduler's i915_active cleanup) we need a worker here. Hence that
part must be kept.
Ideally all these reclaim workers should have some kind of integration
with our shrinkers, but for some of these it's rather tricky. Anyway,
that's a preexisting condition in the codeebase that we wont fix in
this patch here.
We also remove the rcu_barrier in ggtt_cleanup_hw added in
Daniel Vetter [Thu, 2 Sep 2021 14:20:56 +0000 (16:20 +0200)]
drm/i915: use xa_lock/unlock for fpriv->vm_xa lookups
We don't need the absolute speed of rcu for this. And
i915_address_space in general dont need rcu protection anywhere else,
after we've made gem contexts and engines a lot more immutable.
drm/i915/gem: Don't allow changing the VM on running contexts (v4)
this just completes the deed. I've tried to split out prep work for
more careful review as much as possible, this is what's left:
- get_ppgtt gets simplified since we don't need to grab a temporary
reference - we can rely on the temporary reference for the gem_ctx
while we inspect the vm. The new vm_id still needs a full
i915_vm_open ofc. This also removes the final caller of context_get_vm_rcu
- A pile of selftests can now just look at ctx->vm instead of
rcu_dereference_protected( , true) or similar things.
- All callers of i915_gem_context_vm also disappear.
- I've changed the hugepage selftest to set scrub_64K without any
locking, because when we inspect that setting we're also not taking
any locks either. It works because it's a selftests that's careful
(single threaded gives you nice ordering) and not a live driver
where races can happen from anywhere.
These can only be split up further if we have some intermediate state
with a bunch more rcu_dereference_protected(ctx->vm, true), just to
shut up lockdep and sparse.
Note that we're not breaking the actual bugfix in there: The real
bugfix is pushing the i915_vm_relase onto a separate worker, to avoid
locking inversion issues. The rcu conversion was just thrown in for
entertainment value on top (no vm lookup isn't even close to anything
that's a hotpath where removing the single spinlock can be measured).
v2: Rebase over the change to move the i915_vm_put() into
i915_gem_context_release().
v3: Trivial conflict against repainted shed.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-9-daniel.vetter@ffwll.ch
drm/i915/gem: Don't allow changing the VM on running contexts (v4)
the gem_ctx->vm can't change anymore. Plus we always set the
intel_context->vm, so might as well use the helper we have for that.
This makes it very clear that we always overwrite intel_context->vm
for userspace contexts, since the default is gt->vm, which is
explicitly reserved for kernel context use. It would be good to split
things up a bit further and avoid any possibility for an accident
where we run kernel stuff in userspace vm or the other way round.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-8-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:53 +0000 (16:20 +0200)]
drm/i915: Add i915_gem_context_is_full_ppgtt
And use it anywhere we have open-coded checks for ctx->vm that really
only check for full ppgtt.
Plus for paranoia add a GEM_BUG_ON that checks it's really only set
when we have full ppgtt, just in case. gem_context->vm is different
since it's NULL in ggtt mode, unlike intel_context->vm or gt->vm,
which is always set.
v2: 0day found a testcase that I missed.
v3: Repaint shed (Jon, Tvrtko)
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-7-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:52 +0000 (16:20 +0200)]
drm/i915: Use i915_gem_context_get_eb_vm in ctx_getparam
Consolidates the "which is the vm my execbuf runs in" code a bit. We
do some get/put which isn't really required, but all the other users
want the refcounting, and I figured doing a function just for this
getparam to avoid 2 atomis is a bit much.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-6-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:51 +0000 (16:20 +0200)]
drm/i915: Rename i915_gem_context_get_vm_rcu to i915_gem_context_get_eb_vm
The important part isn't so much that this does an rcu lookup - that's
more an implementation detail, which will also be removed.
The thing that makes this different from other functions is that it's
gettting you the vm that batchbuffers will run in for that gem
context, which is either a full ppgtt stored in gem->ctx, or the ggtt.
We'll make more use of this function later on.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-5-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:50 +0000 (16:20 +0200)]
drm/i915: Drop code to handle set-vm races from execbuf
Changing the vm from a finalized gem ctx is no longer possible, which
means we don't have to check for that anymore.
I was pondering whether to keep the check as a WARN_ON, but things go
boom real bad real fast if the vm of a vma is wrong. Plus we'd need to
also get the ggtt vm for !full-ppgtt platforms. Ditching it all seemed
like a better idea.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
References: 51c998b7960b ("drm/i915/gem: Don't allow changing the VM on running contexts (v4)") Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-4-daniel.vetter@ffwll.ch
drm/i915/gem: Safely acquire the ctx->vm when copying
suggested that i915_address_space were at least intended to be managed
through SLAB_TYPESAFE_BY_RCU:
* This ppgtt may have be reallocated between
* the read and the kref, and reassigned to a third
* context. In order to avoid inadvertent sharing
* of this ppgtt with that third context (and not
* src), we have to confirm that we have the same
* ppgtt after passing through the strong memory
* barrier implied by a successful
* kref_get_unless_zero().
But extensive git history search has not brough any such reuse to
light.
(yes this commit is earlier) the final i915_vma_put call has been
moved from i915_gem_context_free (now called _release) to
context_close, which means it's not actually safe anymore to access
the ctx->vm pointer without lock helds, because it might disappear at
any moment. Note that superficially things all still work, because the
i915_address_space is RCU protected since
drm/i915/gtt: Defer address space cleanup to an RCU worker
except the very clever macro above (which is designed to protected
against object reuse due to SLAB_TYPESAFE_BY_RCU or similar tricks)
results in an endless loop if the refcount of the ctx->vm ever
permanently drops to 0. Which it totally now can.
Fix that by moving the final i915_vm_put to where it should be.
Note that i915_gem_context is rcu protected, but _only_ the final
kfree. This means anyone who chases a pointer to a gem ctx solely
under the protection can pretty only call kref_get_unless_zero(). This
seems to be pretty much the case, aside from a bunch of cases that
consult the scheduling information without any further protection.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: "Thomas Hellström" <thomas.hellstrom@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Fixes: 587eefbe9921 ("drm/i915: Pull i915_vma_pin under the vm->mutex") Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-3-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:48 +0000 (16:20 +0200)]
drm/i915: Release ctx->syncobj on final put, not on ctx close
gem context refcounting is another exercise in least locking design it
seems, where most things get destroyed upon context closure (which can
race with anything really). Only the actual memory allocation and the
locks survive while holding a reference.
This tripped up Jason when reimplementing the single timeline feature
in
drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)
We could fix the bug by holding ctx->mutex in execbuf and clear the
pointer (again while holding the mutex) context_close, but it's
cleaner to just make the context object actually invariant over its
_entire_ lifetime. This way any other ioctl that's potentially racing,
but holding a full reference, can still rely on ctx->syncobj being
an immutable pointer. Which without this change, is not the case.
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Fixes: 346397d9927e ("drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)") Cc: Jason Ekstrand <jason@jlekstrand.net> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: "Thomas Hellström" <thomas.hellstrom@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-2-daniel.vetter@ffwll.ch
Daniel Vetter [Thu, 2 Sep 2021 14:20:47 +0000 (16:20 +0200)]
drm/i915: Release i915_gem_context from a worker
The only reason for this really is the i915_gem_engines->fence
callback engines_notify(), which exists purely as a fairly funky
reference counting scheme for that. Otherwise all other callers are
from process context, and generally fairly benign locking context.
Unfortunately untangling that requires some major surgery, and we have
a few i915_gem_context reference counting bugs that need fixing, and
they blow in the current hardirq calling context, so we need a
stop-gap measure.
Put a FIXME comment in when this should be removable again.
v2: Fix mock_context(), noticed by intel-gfx-ci.
Acked-by: Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210902142057.929669-1-daniel.vetter@ffwll.ch
Initialize the L3CC table as part of mocs initialization to program
LNCFCMOCSx registers so that the mocs settings are available for
selection for subsequent memory transactions in the driver load path.
We need to keep L3CC initialization in intel_mocs_init_engine() also
so that in execlists submission, these registers can be rewritten
during engine reset.
drm/i915/gt: Initialize unused MOCS entries with device specific values
Historically we've initialized all undefined/reserved entries in
a platform's MOCS table to the contents of table entry #1 (i.e.,
I915_MOCS_PTE).
Going forward, we can't assume that table entry #1 will always
contain suitable values to use for undefined/reserved table
indices. We'll allow a platform-specific table index to be
selected at table initialization time in these cases.
This new mechanism to select L3 WB entry will be applicable for
all the Gen12+ platforms except TGL and RKL.
Since TGL and RLK are already in production so their mocs settings
are intact to avoid ABI break.
Blitter commands which do not have MOCS fields rely on
cacheability of BlitterCacheControlRegister which was mapped
to index 0 by default.Once we changed the MOCS value of
index 0 to L3 WB, tests like gem_linear_blits started failing
due to a change in cacheability from UC to WB.
Program and place the BlitterCacheControlRegister in
build_aux_regs().
Cache-control registers for Command Stream(CMD_CCTL) are used
to set catchability for memory writes and reads outputted by
Command Streamers on Gen12 onward platforms.
These registers need to point un-cached(UC) MOCS index.
Now there are lots of Command and registers that require mocs index
programming.
So propagating mocs_index from mocs to gt so that it can be
used directly without having platform-specific checks.
V2:
Changed 'i915_mocs_index_gt' to anonymous structure.
Thomas Hellström [Tue, 31 Aug 2021 12:29:31 +0000 (14:29 +0200)]
drm/i915/gem: Fix the mman selftest
Using the I915_MMAP_TYPE_FIXED mmap type requires the TTM backend, so
for that mmap type, use __i915_gem_object_create_user() instead of
i915_gem_object_create_internal(), as we really want to tests objects
mmap-able by user-space.
This also means that the out-of-space error happens at object creation
and returns -ENXIO rather than -ENOSPC, so fix the code up to expect
that on out-of-offset-space errors.
Finally only use I915_MMAP_TYPE_FIXED for LMEM and SMEM for now if
testing on LMEM-capable devices. For stolen LMEM, we still take the
same path as for integrated, as that haven't been moved over to TTM yet,
and user-space should not be able to create out of stolen LMEM anyway.
v2:
- Check the presence of the obj->ops->mmap_offset callback rather than
hardcoding the supported mmap regions in can_mmap() (Maarten Lankhorst)
The function is only used from within GEM_BUG_ON(), which is causing
warnings with Wunneeded-internal-declaration in some builds. Since the
function is a simple wrapper around a CT function, we can just call the
CT function directly instead.
Fixes: d7ec7f773782 ("drm/i915/guc: skip disabling CTBs before sanitizing the GuC") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: John Harrison <John.C.Harrison@Intel.com> Reviewed-by: Matthew Brost <matthew.brost@intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210823163137.19770-1-daniele.ceraolospurio@intel.com
Daniel Vetter [Fri, 20 Aug 2021 15:49:32 +0000 (17:49 +0200)]
drm/i915: Actually delete gpu reloc selftests
In commit dd0a9abbf861 ("drm/i915: delete gpu reloc code") I deleted
the gpu relocation code and the selftest include and enabling, but
accidentally forgot about the selftest source code.
Fix this oversight.
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210820154932.296628-1-daniel.vetter@ffwll.ch
Matt Roper [Thu, 26 Aug 2021 03:35:59 +0000 (20:35 -0700)]
drm/i915: Ensure wa_init_finish() is called for ctx workaround list
A recent restructuring of our context workaround list initialization
added an early return for non-render engines; this caused us to
potentially miss the wa_init_finish() call at the end of the function.
The mistake is pretty harmless --- the only impact is that non-render
engines on graphics version 12.50+ platforms we don't trim down the
workaround list to reclaim some memory, and we don't print the usual
"Initialized 1 context workaround" message in dmesg. Let's change the
early return to a jump down to the wa_init_finish() call at the bottom
of the function.
Thomas Hellström [Fri, 13 Aug 2021 14:43:31 +0000 (16:43 +0200)]
drm/ttm, drm/i915: Update ttm_move_memcpy for async use
The buffer object argument to ttm_move_memcpy was only used to
determine whether the destination memory should be cleared only
or whether we should copy data. Replace it with a "clear" bool, and
update the callers.
The intention here is to be able to use ttm_move_memcpy() async under
a dma-fence as a fallback if an accelerated blit fails in a security-
critical path where data might leak if the blit is not properly
performed. For that purpose the bo is an unsuitable argument since
its relevant members might already have changed at call time.
Finally, update the ttm_move_memcpy kerneldoc that seems to have
ended up with a stale version.
Thomas Hellström [Fri, 13 Aug 2021 14:43:30 +0000 (16:43 +0200)]
drm/i915/ttm: Reorganize the ttm move code somewhat
In order to make the code a bit more readable and to facilitate
async memcpy moves, reorganize the move code a little. Determine
at an early stage whether to copy or to clear.
v2:
- Don't set up the memcpy iterators unless we are actually going to memcpy.
New LRI register offsets were introduced for DG2, this patch adds
those extra registers, and create new register table for setting offsets
to compare with HW generated context image - especially for gt_lrc test.
Also updates general purpose register with scratch offset for DG2, in
order to use it for live_lrc_fixed selftest.
Matthew Brost [Tue, 27 Jul 2021 03:17:03 +0000 (20:17 -0700)]
drm/i915/selftests: Increase timeout in i915_gem_contexts selftests
Like in the case of several other selftests, generating lots of requests
in a loop takes a bit longer with GuC submission. Increase a timeout in
i915_gem_contexts selftest to take this into account.
drm/i915/selftest: Fix use of err in igt_reset_{fail, nop}_engine()
Clang warns:
In file included from drivers/gpu/drm/i915/gt/intel_reset.c:1514:
drivers/gpu/drm/i915/gt/selftest_hangcheck.c:465:62: warning: variable
'err' is uninitialized when used here [-Wuninitialized]
pr_err("[%s] Create context failed: %d!\n", engine->name, err);
^~~
...
drivers/gpu/drm/i915/gt/selftest_hangcheck.c:580:62: warning: variable
'err' is uninitialized when used here [-Wuninitialized]
pr_err("[%s] Create context failed: %d!\n", engine->name, err);
^~~
...
2 warnings generated.
This appears to be a copy and paste issue. Use ce directly using the %pe
specifier to pretty print the error code so that err is not used
uninitialized in these functions.
Lucas De Marchi [Thu, 19 Aug 2021 21:03:49 +0000 (14:03 -0700)]
drm/i915/dg1: remove __maybe_unused leftover
This was added in commit ca11bfeaec6b ("drm/i915/dg1: add initial DG-1
definitions") so we could continue to add support for DG1 without
risk to expose a broken UAPI. Now that we added DG1 to the PCI ID list
i915 may bind to, remove the leftover.
Matthew Auld [Thu, 19 Aug 2021 09:34:18 +0000 (10:34 +0100)]
drm/i915/buddy: add some pretty printing
Implement the debug hook for the buddy resource manager. For this we
want to print out the status of the memory manager, including how much
memory is still allocatable, what page sizes we have etc. This will be
triggered when TTM is unable to fulfil an allocation request for device
local-memory.
Thomas Hellström [Mon, 16 Aug 2021 17:14:44 +0000 (19:14 +0200)]
drm/i915: Ditch the i915_gem_ww_ctx loop member
It's only used by the for_i915_gem_ww() macro and we can use
the (typically) on-stack _err variable in its place.
v2:
- Don't clear the _err variable when entering the loop
(Matthew Auld, Maarten Lankhorst).
- Use parentheses around the _err macro argument.
- Fix up comment.
Matthew Brost [Fri, 30 Jul 2021 19:53:42 +0000 (12:53 -0700)]
drm/i915: Fix syncmap memory leak
A small race exists between intel_gt_retire_requests_timeout and
intel_timeline_exit which could result in the syncmap not getting
free'd. Rather than work to hard to seal this race, simply cleanup the
syncmap on fini.
Signed-off-by: Matthew Brost <matthew.brost@intel.com> Fixes: 6e35a2ddc1f3 ("drm/i915/gt: Track timeline activeness in enter/exit") Cc: <stable@vger.kernel.org> Reviewed-by: John Harrison <John.C.Harrison@Intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210730195342.110234-1-matthew.brost@intel.com
For tgl+, the per-context setting of MI_MODE[12] determines whether
the bits of a nested MI_BATCH_BUFFER_START instruction should be
interpreted in the traditional manner or whether they should
instead use a new tgl+ meaning that breaks backward compatibility, but
allows nesting into 3rd-level batchbuffers. For previous platforms,
the hardware default for this register bit is to maintain
backward-compatible behavior unless a context intentionally opts into
the new behavior; however Xe_HPG flips the hardware default behavior.
From a SW perspective, we want to maintain the backward-compatible
behavior for userspace, so we'll apply a fake workaround to set it back
to the legacy behavior on platforms where the hardware default is to
break compatibility. At the moment there is no Linux userspace that
utilizes third-level batchbuffers, so this will avoid userspace from
needing to make any changes. using the legacy meaning is the correct
thing to do. If/when we have userspace consumers that want to utilize
third-level batch nesting, we can provide a context parameter to allow
them to opt-in.
Kees Cook [Tue, 17 Aug 2021 23:33:57 +0000 (16:33 -0700)]
drm/i915: Use designated initializers for init/exit table
The kernel builds with -Werror=designated-init, and __designated_init
is used by CONFIG_GCC_PLUGIN_RANDSTRUCT for automatically selected (all
function pointer) structures. Include the field names in the init/exit
table. Avoids warnings like:
drivers/gpu/drm/i915/i915_module.c:59:4: error: positional initialization of field in 'struct' declared with 'designated_init' attribute [-Werror=designated-init]
Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: David Airlie <airlied@linux.ie> Cc: intel-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Fixes: 8d8e1a0ece53 ("drm/i915: Use a table for i915_init/exit (v2)") Signed-off-by: Kees Cook <keescook@chromium.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Link: https://patchwork.freedesktop.org/patch/msgid/20210817233357.2379455-1-keescook@chromium.org
DG1 has support for local memory, which requires the usage of the
lmem placement extension for creating bo's, and memregion queries
to obtain the size. Because of this, those parts of the uapi are
no longer guarded behind FAKE_LMEM.
According to the pull request referenced below, mesa should be mostly
ready for DG1. VK_EXT_memory_budget is not hooked up yet, but we
should definitely just enable the uapi parts by default.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
References: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11584 Cc: Jordan Justen <jordan.l.justen@intel.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210812124452.622233-2-maarten.lankhorst@linux.intel.com Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> Acked-by: Jason Ekstrand <jason@jlekstrand.net>
drm/i915/gem: Utilize rcu iteration of context engines
Note that the other use of __context_engines_await have disappeard in
the following commits:
51c998b7960b ("drm/i915/gem: Don't allow changing the VM on running contexts (v4)") 022e404d5a40 ("drm/i915: Drop getparam support for I915_CONTEXT_PARAM_ENGINES") de411643c8a1 ("drm/i915: Drop the CONTEXT_CLONE API (v2)")
None of these have any business to optimize their engine lookup with
rcu, unless extremely convincing benchmark data and a solid analysis
why we can't make that workload (whatever it is that does) faster with
a proper design fix.
Also since there's only one caller of context_apply_all left and it's
really just a loop, inline it and then inline the lopp body too. This
is how all other callers that take the engine lock loop over engines,
it's much simpler.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Jason Ekstrand <jason@jlekstrand.net> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Matthew Brost <matthew.brost@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210810130523.1972031-1-daniel.vetter@ffwll.ch
Stuart Summers [Fri, 6 Aug 2021 17:29:01 +0000 (10:29 -0700)]
drm/i915/xehpsdv: Add compute DSS type
Starting in XeHP, the concept of slice has been removed in favor of
DSS (Dual-Subslice) masks for various workload types. These workloads have
been divided into those enabled for geometry and those enabled for compute.
i915 currently maintains a single set of S/SS/EU masks for the device.
The goal of this patch set is to minimize the amount of impact to prior
generations while still giving the user maximum flexibility.
v2:
- Generalize a comment about uapi access to geometry/compute masks; the
proposed uapi has changed since the comment was first written, and
will show up in a future series once the userspace code is published.
(Lucas)
v3:
- Eliminate unnecessary has_compute_dss flag. (Lucas)
- Drop unwanted comment change in uapi header. (Lucas)
Bspec: 33117, 33118, 20376 Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Matt Roper <matthew.d.roper@intel.com> Cc: Lucas De Marchi <lucas.demarchi@intel.com> Signed-off-by: Stuart Summers <stuart.summers@intel.com> Signed-off-by: Steve Hampson <steven.t.hampson@intel.com> Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210806172901.1049133-1-matthew.d.roper@intel.com
Matt Roper [Thu, 5 Aug 2021 16:36:41 +0000 (09:36 -0700)]
drm/i915/dg2: Report INSTDONE_GEOM values in error state
Xe_HPG adds some additional INSTDONE_GEOM debug registers; the Mesa team
has indicated that having these reported in the error state would be
useful for debugging GPU hangs. These registers are replicated per-DSS
with gslice steering.
Matt Roper [Thu, 5 Aug 2021 16:36:40 +0000 (09:36 -0700)]
drm/i915/xehp: Loop over all gslices for INSTDONE processing
We no longer have traditional slices on Xe_HP platforms, but the
INSTDONE registers are replicated according to gslice representation
which is similar. We can mostly re-use the existing instdone code with
just a few modifications:
* Create an alternate instdone loop macro that will iterate over the
flat DSS space, but still provide the gslice/dss steering values for
compatibility with the legacy code.
* We should allocate INSTDONE storage space according to the maximum
number of gslices rather than the maximum number of legacy slices to
ensure we have enough storage space to hold all of the values. XeHP
design has 8 gslices, whereas older platforms never had more than 3
slices.
Daniel Vetter [Tue, 10 Aug 2021 14:27:48 +0000 (16:27 +0200)]
drm/doc/rfc: drop lmem uapi section
We still have quite a bit more work to do with overall reworking of
the ttm-based dg1 code, but the uapi stuff is now finalized with the
latest pull. So remove that.
This also fixes kerneldoc build warnings because we've included the
same headers in two places, resulting in sphinx complaining about
duplicated symbols. This regression has been created when we moved the
uapi definitions to the real include/uapi/ folder in 9c8411eb9561
("drm/doc/rfc: drop the i915_gem_lmem.h header")
v2: Fix a few references that I missed, the htmldocs build took
forever.
Acked-by: Jason Ekstrand <jason@jlekstrand.net> Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Tested-by Stephen Rothwell <sfr@canb.auug.org.au> (v1)
References: https://lore.kernel.org/dri-devel/20210603193242.1ce99344@canb.auug.org.au/ Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Fixes: 9c8411eb9561 ("drm/doc/rfc: drop the i915_gem_lmem.h header") Cc: Matthew Auld <matthew.auld@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210810142748.1983271-1-daniel.vetter@ffwll.ch
Matt Roper [Thu, 29 Jul 2021 05:41:18 +0000 (22:41 -0700)]
drm/i915/xehp: Xe_HP shadowed registers are a strict superset of gen12
The list of shadowed registers on XeHP is identical to the set for
earlier gen12 platforms, with additional ranges added for the new VCS
and VECS engines. Since those register ranges were reserved on earlier
gen12 platforms, it's safe to consolidate to a single gen12 table
rather than tracking Xe_HP separately.
Matt Roper [Thu, 29 Jul 2021 05:41:17 +0000 (22:41 -0700)]
drm/i915/gen12: Update shadowed register table
The bspec lists many shadowed registers (i.e., registers for which we
don't need to grab forcewake when writing) that we weren't tracking in
the driver. Although we may not actually use all of these registers
right now, it's best to just match the bspec list exactly.
Note that the bspec also lists registers that are shadowed for various
HW-internal accesses; we can ignore those and just list the ones that
are shadowed for accesses from the IA/CPU.
Matt Roper [Thu, 29 Jul 2021 05:41:16 +0000 (22:41 -0700)]
drm/i915/gen11: Update shadowed register table
The bspec lists many shadowed registers (i.e., registers for which we
don't need to grab forcewake when writing) that we weren't tracking in
the driver. Although we may not actually use all of these registers
right now, it's best to just match the bspec list exactly.
Note that the bspec also lists registers that are shadowed for various
HW-internal accesses; we can ignore those and just list the ones that
are shadowed for accesses from the IA/CPU.
Matt Roper [Thu, 29 Jul 2021 15:21:58 +0000 (08:21 -0700)]
drm/i915: Make shadow tables range-based
Rather than defining our shadow tables as a list of individual
registers, provide them as a list of register ranges; we'll have some
ranges of multiple registers being added soon (and we already have a
couple adjacent registers that we can squash into a single range now).
This change also defines the table with hex literal values rather than
symbolic register names; since that's how the tables are defined in the
bspec, this change will make it easier to review the tables overall.
v2:
- Force signed comparison on range overlap sanity check
Matt Roper [Thu, 29 Jul 2021 05:41:14 +0000 (22:41 -0700)]
drm/i915: Re-use gen11 forcewake read functions on gen12
The forcewake read logic is identical between gen11 and gen12, only the
forcewake table data (which is tracked separately) differs; there's no
need to generate a separate set of gen12 read functions when the gen11
functions will work just as well.
We'll keep the separate write functions for now since the generated code
directly references different shadow tables between the two platforms.
Matt Roper [Thu, 29 Jul 2021 05:41:13 +0000 (22:41 -0700)]
drm/i915: correct name of GT forcewake domain in error messages
For historical reasons, the GT forcewake domain used to be referred to
as the "blitter" domain; that name is no longer accurate since the GT
domain contains a lot of additional registers and functionality besides
just the blitter. Although we renamed the domain in the driver in
commit c35d05453808 ("drm/i915: Rename FORCEWAKE_BLITTER to
FORCEWAKE_GT"), we neglected to update the string that gets printed in
driver error messages; let's do that now to avoid confusion.
Matt Roper [Thu, 29 Jul 2021 16:59:55 +0000 (09:59 -0700)]
drm/i915/dg2: Add SQIDI steering
Although DG2_G10 platforms will always have all SQIDI's present and
don't need steering for registers in a SQIDI MMIO range, this isn't true
for DG2_G11 platforms; only SQIDI's 2 and 3 can be used on those.
We handle SQIDI ranges a bit differently from other types of explicit
steering. The SQIDI ranges belong to either the MCFG unit or the SF
unit, both of which have their own dedicated steering registers and do
not use the typical 0xFDC steering control that all other types of
ranges use. Thus we only need to worry about picking a valid initial
value for the MCFG and SF steering registers (0xFD0 and 0xFD8
respectively) at driver init; they won't change after we set them up so
we don't need to worry about re-steering them explicitly at runtime.
Given that any SQIDI value should work fine for DG2-G10 and XeHP SDV,
while only values of 2 and 3 are valid for DG2-G11, we'll just
initialize the MCFG and SF steering registers to a constant value of "2"
for all XeHP-based platforms for simplicity --- that will work in all
cases.
Matt Roper [Thu, 29 Jul 2021 16:59:54 +0000 (09:59 -0700)]
drm/i915/dg2: Update steering tables
DG2's replicated register ranges are almost the same at XeHP SDV with
the exception of one LNCF sub-range that switches to gslice steering.
We can re-use the XeHP SDV mslice steering table and just provide a
DG2-specific LNCF steering table.
Xe_HP is more modular than its predecessors and as a consequence it has
more types of replicated registers. As with l3bank regions on previous
platforms, we may need to explicitly re-steer accesses to these new
types of ranges at runtime if we can't find a single default steering
value that satisfies the fusing of all types.
v2:
- Add a local 'i915' variable to reduce gt->i915 usage. (Caz)
- Drop unused 'intel_gt_read_register' prototype. (Caz)
v3:
- Drop unnecessary comment text. (Lucas)
- Drop unused register bit definition. (Lucas)
Bspec: 66534 Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: Caz Yokoyama <caz.yokoyama@intel.com> Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210729170008.2836648-2-matthew.d.roper@intel.com
Chris Wilson [Fri, 23 Jul 2021 11:34:05 +0000 (12:34 +0100)]
drm/i915/userptr: Probe existence of backing struct pages upon creation
Jason Ekstrand requested a more efficient method than userptr+set-domain
to determine if the userptr object was backed by a complete set of pages
upon creation. To be more efficient than simply populating the userptr
using get_user_pages() (as done by the call to set-domain or execbuf),
we can walk the tree of vm_area_struct and check for gaps or vma not
backed by struct page (VM_PFNMAP). The question is how to handle
VM_MIXEDMAP which may be either struct page or pfn backed...
With discrete we are going to drop support for set_domain(), so offering
a way to probe the pages, without having to resort to dummy batches has
been requested.
v2:
- add new query param for the PROBE flag, so userspace can easily
check if the kernel supports it(Jason).
- use mmap_read_{lock, unlock}.
- add some kernel-doc.
v3:
- In the docs also mention that PROBE doesn't guarantee that the pages
will remain valid by the time they are actually used(Tvrtko).
- Add a small comment for the hole finding logic(Jason).
- Move the param next to all the other params which just return true.
Testcase: igt/gem_userptr_blits/probe Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Signed-off-by: Matthew Auld <matthew.auld@intel.com> Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Cc: Jordan Justen <jordan.l.justen@intel.com> Cc: Kenneth Graunke <kenneth@whitecape.org> Cc: Jason Ekstrand <jason@jlekstrand.net> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: Ramalingam C <ramalingam.c@intel.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Acked-by: Kenneth Graunke <kenneth@whitecape.org> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210723113405.427004-1-matthew.auld@intel.com
This code just disables gpu relocations, leaving the garbage
collection for later patches and more importantly, much less confusing
diff. Also given how much headaches this code has caused in the past,
letting this soak for a bit seems justified.
Acked-by: Dave Airlie <airlied@redhat.com> Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Reviewed-by: Jason Ekstrand <jason@jlekstrand.net> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Cc: Jon Bloomfield <jon.bloomfield@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Daniel Vetter <daniel.vetter@ffwll.ch> Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com> Cc: Matthew Auld <matthew.auld@intel.com> Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Jason Ekstrand <jason@jlekstrand.net> Link: https://patchwork.freedesktop.org/patch/msgid/20210803124833.3817354-1-daniel.vetter@ffwll.ch
Matthew Auld [Thu, 29 Jul 2021 16:59:58 +0000 (09:59 -0700)]
drm/i915/xehp: Changes to ss/eu definitions
Xe_HP no longer has "slices" in the same way that old platforms did.
There are new concepts (gslices, cslices, mslices) that apply in various
contexts, but for the purposes of fusing slices no longer exist and we
just have one large pool of dual-subslices (DSS) to work with.
Furthermore, the meaning of the DSS fuse is inverted compared to past
platforms --- it now specifies which DSS are enabled rather than which
ones are disabled.
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: Lucas De Marchi <lucas.demarchi@intel.com> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Signed-off-by: Matthew Auld <matthew.auld@intel.com> Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Signed-off-by: Radhakrishna Sripada <radhakrishna.sripada@intel.com> Signed-off-by: Stuart Summers <stuart.summers@intel.com> Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Signed-off-by: Prasad Nallani <prasad.nallani@intel.com> Reviewed-by: José Roberto de Souza <jose.souza@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210729170008.2836648-9-matthew.d.roper@intel.com
Matt Roper [Thu, 29 Jul 2021 16:59:53 +0000 (09:59 -0700)]
drm/i915/dg2: Add forcewake table
The DG2 forcewake table is very similar to the one used by XeHP SDV (and
both platforms are even presented as a single table in the bspec). For
the most part DG2 starts using a few additional ranges that were
'reserved' on XeHP SDV and stops using some others. However there is a
single range (0xd800-0xd87f) that needs to be handled differently
between the two platforms (it needs GT wake on XeHP SDV, but render wake
on DG2) so unless we want to wake both domains (which could waste power)
or define new types of forcewake domains for this special case we need
to have separate tables for the two platforms. Let's define the ranges
for both platforms with a parameterized macro so that we don't actually
need to duplicate everything in the code.
It should be fine for DG2 to re-use the Xe_HP shadow register list so we
can continue to use the 'xehpsdv' MMIO write functions and don't need to
spin up a separate DG2 instance.
This feature hands over the control of HW RC6 to the GuC.
GuC decides when to put HW into RC6 based on it's internal
busyness algorithms.
GuCRC needs GuC submission to be enabled, and only
supported on Gen12+ for now.
When GuCRC is enabled, do not set HW RC6. Use a H2G message
to tell GuC to enable GuCRC. When disabling RC6, tell GuC to
revert RC6 control back to KMD. KMD is still responsible for
enabling everything related to Coarse Power Gating though.
v2: Address comments (Michal W)
v3: Don't set hysterisis values when GuCRC is used (Matt Roper)
v4: checkpatch()
Update the get/set min/max freq hooks to work for
SLPC case as well. Consolidate helpers for requested/min/max
frequency get/set to intel_rps where the proper action can
be taken depending on whether SLPC is enabled.
v2: Add wrappers for getting rp0/1/n frequencies, update
softlimits in set min/max SLPC functions. Also check for
boundary conditions before setting them.
v3: Address review comments (Michal W)
v4: Add helper for host part of intel_rps_set_freq helpers (Michal W)
v5: checkpatch()
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Acked-by: Michal Wajdeczko <michal.wajdeczko@intel.com> Signed-off-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> Signed-off-by: Sujaritha Sundaresan <sujaritha.sundaresan@intel.com> Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20210730202119.23810-13-vinay.belgaumkar@intel.com
drm/i915/guc/slpc: Cache platform frequency limits
Cache rp0, rp1 and rpn platform limits into SLPC structure
for range checking while setting min/max frequencies.
Also add "soft" limits which keep track of frequency changes
made from userland. These are initially set to platform min
and max.
v2: Address review comments (Michal W)
v3: Formatting (Michal W)
v4: Add separate function to parse rp values (Michal W)
v5: Perform range checking for set min/max (Michal W)
v6: checkpatch() and rename static functions (Michal W)
v7: check ret code while setting SLPC limits (Michal W)
This interrupt is enabled during RPS initialization, and
now needs to be done by SLPC code. It allows ARAT timer
expiry interrupts to get forwarded to GuC.
Add helpers to read the min/max frequency being used
by SLPC. This is done by send a H2G command which forces
SLPC to update the shared data struct which can then be
read. These helpers will be used in a sysfs patch later
on.
v2: Address review comments (Michal W)
v3: Return err in case of query failure (Michal W)
v4: Move decode_min/max_freq to this patch
drm/i915/guc/slpc: Remove BUG_ON in guc_submission_disable
The assumption when it was added was that GT would not be
holding any gt_pm references. However, uc_init is called
from gt_init_hw, which holds a forcewake ref. If SLPC
enable fails, we will still be holding this ref, which will
result in the BUG_ON.
drm/i915/guc/slpc: Enable SLPC and add related H2G events
Add methods for interacting with GuC for enabling SLPC. Enable
SLPC after GuC submission has been established. GuC load will
fail if SLPC cannot be successfully initialized. Add various
helper methods to set/unset the parameters for SLPC. They can
be set using H2G calls or directly setting bits in the shared
data structure.
v2: Address several review comments, add new helpers for
decoding the SLPC min/max frequencies. Use masks instead of hardcoded
constants. (Michal W)
v3: Split global_state_to_string function, and check for positive
non-zero return value from intel_guc_send() (Michal W)
v4: Optimize the stringify function and other comments (Michal W)
v5: Enable slpc as well before declaring GuC submission status (Michal W)
drm/i915/guc/slpc: Allocate, initialize and release SLPC
Allocate data structures for SLPC and functions for
initializing on host side.
v2: Address review comments (Michal W)
v3: Remove unnecessary header includes (Michal W)
v4: Rebase
v5: Move allocation of shared data into slpc_init() (Michal W)
drm/i915/guc/slpc: Gate Host RPS when SLPC is enabled
Also ensure uc_init is called before we initialize RPS so that we
can check for SLPC support. We do not need to enable up/down
interrupts when SLPC is enabled. However, we still need the ARAT
interrupt, which will be enabled separately later.
v2: Explicitly return from intel_rps_enable with slpc check (Matthew B)
Lucas De Marchi [Wed, 28 Jul 2021 22:03:26 +0000 (15:03 -0700)]
drm/i915/gt: remove GRAPHICS_VER == 10
Replace all remaining handling of GRAPHICS_VER {==,>=} 10 with
{==,>=} 11. With the removal of CNL, there is no platform with graphics
version equals 10.
Lucas De Marchi [Wed, 28 Jul 2021 22:03:24 +0000 (15:03 -0700)]
drm/i915/gt: remove explicit CNL handling from intel_sseu.c
CNL is the only platform with GRAPHICS_VER == 10. With its removal we
don't need to handle that version anymore.
Also we can now reduce the max number of slices: the call to
intel_sseu_set_info() with the highest number of slices comes from SKL
and BDW with 3 slices. Recent platforms actually increase the
number of subslices so the number of slices remain 1.
Daniel Vetter [Tue, 27 Jul 2021 12:10:37 +0000 (14:10 +0200)]
drm/i915: Extract i915_module.c
The module init code is somewhat misplaced in i915_pci.c, since it
needs to pull in init/exit functions from every part of the driver and
pollutes the include list a lot.
Extract an i915_module.c file which pulls all the bits together, and
allows us to massively trim the include list of i915_pci.c.
The downside is that have to drop the error path check Jason added to
catch when we set up the pci driver too early. I think that risk is
acceptable for this pretty nice include.
projects / kernel.git / log