The name conditions isn't quite correct for the case statement. This CL
updates the code to use selectors instead of conditions.
Change-Id: I98b8050b11e2328f97e4443469572ab47d7c1555
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22520
Reviewed-by: David Neto <dneto@google.com>
In this case, a basic block is a continue target
for both an outer loop and is also for itself as a
single-block loop.
Bug: Tint:3
Change-Id: If361004a4b871966674ca972922c45e712ce7c9c
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22420
Reviewed-by: dan sinclair <dsinclair@google.com>
This Cl adds preliminary support for generating Kill commands. This may
change in the future if the semantics of kill are set in WGSL to not
match OpKill.
Bug: tint:5
Change-Id: Id466a1b26b37867a26373b93f56d627a0491f47f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22240
Reviewed-by: David Neto <dneto@google.com>
This CL updates a few names, addeds a return_stmt method and re-orders
some code to closer match the current WGSL specification.
Change-Id: I388be1c22d5d10229fdfcdb2ff929c410f5ae638
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22305
Reviewed-by: David Neto <dneto@google.com>
The intrinsic methods were removed from the WGSL grammar and are treated
as builtin functions. This Cl updates Tint to match.
Bug: tint:41
Change-Id: I3f9ff6c17f1ca57ad159d883fd5a966657caeb4f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22301
Reviewed-by: David Neto <dneto@google.com>
This CL removes the builtin_decoration from the parser in favour of
using IDENT tokens for builtins. We still convert to an enum in the
parser and validate the value provided.
Bug: tint:41
Change-Id: If5dc3844e3325c75951e7b320c123cf66cb8e106
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22300
Reviewed-by: David Neto <dneto@google.com>
The NOP command was added for complete SPIR-V bijectivity which is no
longer a goal of WGSL. The NOP command has been removed from the spec,
so remove from Tint.
Change-Id: Ic640d9d3b987a19668dc169ff549444921c5bbb8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/22160
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Do so systmatically. Before we had tested some as a side effect
of other objectives.
Fix the error message for when we have a bad exit from a loop construct
that bypasses not only the continue construct but the loop merge block
itself.
Bug: tint:3
Change-Id: Iaf8fc9bcd3162002aa906efa90a244ef5f439911
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21580
Reviewed-by: dan sinclair <dsinclair@google.com>
Remove TOOD about special code for finding badly nested selections.
This is already adequately covered by the dominance-violation checks,
and related tests:
ClassifyCFGEdges_DomViolation_BeforeIfToSelectionInterior
ClassifyCFGEdges_DomViolation_BeforeSwitchToSelectionInterior
ClassifyCFGEdges_DomViolation_BeforeLoopToLoopBodyInterior
ClassifyCFGEdges_DomViolation_BeforeContinueToContinueInterior
ClassifyCFGEdges_DomViolation_AfterContinueToContinueInterior
Bug: tint:3
Change-Id: I8c547474d1422a54078d748b39a7d22a500d5d1d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21482
Reviewed-by: dan sinclair <dsinclair@google.com>
ClassifyCFGEdges_Forward_LoopToContinue is already covered via
ClassifyCFGEdges_LoopContinue_LoopBodyToContinue and ...ConditionalFromNestedIf
and other LoopContinue cases
Bug: tint:3
Change-Id: I4a89a82c0c9a08939036a20f13932c18e1f90df6
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21560
Reviewed-by: dan sinclair <dsinclair@google.com>
Finds the "then", the "else", and "premerge" nodes.
The premerge node, if it exists, is the first block where
the normal forward flow of the "then" and "else" clauses
converge, but before the merge block.
Finds error case where there a block has both an if-break
edge and a forward-to-premerge. There is no good way
to model that in a high level language.
Bug: tint:3
Change-Id: I759fc539f3480e38d091041db6a9abd15f3df769
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21240
Reviewed-by: dan sinclair <dsinclair@google.com>
Push constants won't be supported by WebGPU. This CL removes them from
Tint.
For the SPIR-V reader case, this means the push constant will cause a
unknown storage class error.
Bug: tint:19
Change-Id: I10b09ce589b6b370fc828af0cd56e6213c5bc694
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21180
Reviewed-by: David Neto <dneto@google.com>
Distinguish between selections constructs starting with with
OpBranchConditional and those starting with OpSwitch.
We'll use this in a followup CL to track break from a switch.
Bug: tint:3
Change-Id: I8d000cb42325535a4937c84f83a83c98a9b8d4c5
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21080
Reviewed-by: dan sinclair <dsinclair@google.com>
Classify CFG edges:
- loop backedge
- a structured exit:
- loop break
- loop continue
- selection break
- fallthrough
- forward (any of the rest)
Also error out when there should have been a merge instruction.
(More than one unique fallthrough or forward edge).
Includes lots of tests.
Bug: tint:3
Change-Id: I70f27680bdf098213056522abf04ac58a6b478ab
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20481
Reviewed-by: dan sinclair <dsinclair@google.com>
In SPIR-V a compute entry point requires a WorkGroup Size constant or a
LocalSize execution mode. This CL adds a 1,1,1 execution mode as a
starting point.
Bug: tint:74
Change-Id: Ie5bf639472033dca9ffe333548e7f31d4d318768
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/21020
Reviewed-by: David Neto <dneto@google.com>
The WGSL spec requires initializers for Output, Private and Function
variables. This CL adds initializers where needed.
Bug: tint:75
Change-Id: Id97f85a67ead2ffc41d6bdd1b71bf7034b04502a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20980
Reviewed-by: David Neto <dneto@google.com>
This CL adds the first simple pass at adding interface variables into
the entry point command. It simply lists all Input/Output module scoped
variables onto all entry points.
Bug: tint:28
Change-Id: I962462d783f3b97bb3da32fd9890ceb90808942b
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20963
Reviewed-by: David Neto <dneto@google.com>
This CL fixes the IntLiteral name to contain the type of the literal.
This keeps i32 and u32 from fighting over a given name. Now, the name
ends up being __int__i32_0 instead of __int_0.
Change-Id: Ifb9f0516139d25f34312c75c77318eccbe076ef8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20941
Reviewed-by: David Neto <dneto@google.com>
In some cases the array access must generate a load if there is a
following swizzle. This CL adds the needed load.
Bug: tint:71
Change-Id: Ib6ca6284af993962d9d573b323d54be7ce0e726c
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20920
Reviewed-by: David Neto <dneto@google.com>
If an array accessor index is a variable it needs to be loaded before we
can use it to access the array. This CL adds the call to generate the
load for the index.
Bug: tint:5
Change-Id: I74143d2eb746727e857218d54126e1f5fd785bdf
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20780
Reviewed-by: David Neto <dneto@google.com>
This CL adds assignment tests which require evaluating an access chain
to get the thing being written into.
Bug: tint:5
Change-Id: I475173b7b4ef4223de6f4258b8fa920760196b12
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/20626
Reviewed-by: David Neto <dneto@google.com>
