[NFC][AMDGPULowerModuleLDSPass] Use shorter APIs in markUsedByKernel

* Use shorter versions of the LLVM API

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D155589

[NFC][AMDGPULowerModuleLDSPass] Cleanup of getTableLookupKernelIndex

* Do a single lookup when querying the map
* Use shorter versions of the LLVM API

Reviewed By: JonChesterfield

Differential Rev

[NFC][AMDGPULowerModuleLDSPass] Cleanup of getTableLookupKernelIndex

* Do a single lookup when querying the map
* Use shorter versions of the LLVM API

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D155588

show more ...

[amdgpu] Accept an optional max to amdgpu-lds-size attribute for use in PromoteAlloca

[amdgpu] Delete elide-module-lds attribute

Requires D155190

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D155238

[amdgpu][lds] Remove recalculation of LDS frame from backend

Do the LDS frame calculation once, in the IR pass, instead of repeating the work in the backend.

Prior to this patch:
The IR lowering pa

[amdgpu][lds] Remove recalculation of LDS frame from backend

Do the LDS frame calculation once, in the IR pass, instead of repeating the work in the backend.

Prior to this patch:
The IR lowering pass sets up a per-kernel LDS frame and annotates the variables with absolute_symbol
metadata so that the assembler can build lookup tables out of it. There is a fragile association between
kernel functions and named structs which is used to recompute the frame layout in the backend, with
fatal_errors catching inconsistencies in the second calculation.

After this patch:
The IR lowering pass additionally sets a frame size attribute on kernels. The backend uses the same
absolute_symbol metadata that the assembler uses to place objects within that frame size.

Deleted the now dead allocation code from the backend. Left for a later cleanup:
- enabling lowering for anonymous functions
- removing the elide-module-lds attribute (test churn, it's not used by llc any more)
- adjusting the dynamic alignment check to not use symbol names

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D155190

show more ...

[amdgpu][lds] Raise an explicit unimplemented error on absolute address LDS variables

These aren't implemented. They could be at moderate implementation
complexity. Raising an error is better than s

[amdgpu][lds] Raise an explicit unimplemented error on absolute address LDS variables

These aren't implemented. They could be at moderate implementation
complexity. Raising an error is better than silently miscompiling.

Patching now because the patch at D155125 is a step towards using this metadata
more extensively as part of the lowering path and that will interact badly with
input variables with this annotation.

Lowering user defined variables at specific addresses would drop this error,
put them at the requested position in the frame during this pass, and then
use the same codegen that will be used for the kernel specific struct shortly.

Reviewed By: jmmartinez

Differential Revision: https://reviews.llvm.org/D155132

show more ...

[NFC][AMDGPULowerModuleLDSPass] Fix buildbot santizier failed to compile

It seems that the sanitizer-x86_64-linux-android wasn't able to deduce
the template argument:

AMDGPULowerModuleLDSPass.cpp

[NFC][AMDGPULowerModuleLDSPass] Fix buildbot santizier failed to compile

It seems that the sanitizer-x86_64-linux-android wasn't able to deduce
the template argument:

AMDGPULowerModuleLDSPass.cpp:1192:53: error: no viable constructor or
deduction guide for deduction of template arguments of 'vector'
auto TableLookupVariablesOrdered = sortByName(std::vector(

This patch makes the template argument explicit.

show more ...

Reland "[NFC][AMDGPULowerModuleLDSPass] Factorize repetead sort code"

Fixed compilation error and reudndant copy warning

Differential Revision: https://reviews.llvm.org/D154977

[amdgpu][lds] Fix missing markUsedByKernel calls and undef lookup table elements

More robust association between the kernels and lds struct.

Use poison instead of value() for lookup table elements

[amdgpu][lds] Fix missing markUsedByKernel calls and undef lookup table elements

More robust association between the kernels and lds struct.

Use poison instead of value() for lookup table elements introduced by dynamic lds lowering.

Extracted from D154946, new test from there verbatim. Segv fixed.

Fixes issues/63338

Fixes SWDEV-404491

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D154972

show more ...

Revert "[NFC][AMDGPULowerModuleLDSPass] Factorize repetead sort code"

This reverts commit 125b90749a98d6dc6b492883c9617f9e91ab60e0.

[NFC][AMDGPULowerModuleLDSPass] Factorize repetead sort code

Reviewed By: JonChesterfield

Differential Revision: https://reviews.llvm.org/D154970

[NFC][AMDGPULowerModuleLDSPass] Add const to some variables/parameters

Moving out some changes not related to the bugfix in https://reviews.llvm.org/D154946

Reviewed By: JonChesterfield, arsenm

Di

[NFC][AMDGPULowerModuleLDSPass] Add const to some variables/parameters

Moving out some changes not related to the bugfix in https://reviews.llvm.org/D154946

Reviewed By: JonChesterfield, arsenm

Differential Revision: https://reviews.llvm.org/D154959

show more ...

[NFC][AMDGPULowerModuleLDSPass] Remove dead variable

[amdgpu][nfc] Update comments on LDS lowering

[amdgpu] Implement dynamic LDS accesses from non-kernel functions

The premise here is to allow non-kernel functions to locate external LDS variables without using LDS or extra magic SGPRs to do so.

[amdgpu] Implement dynamic LDS accesses from non-kernel functions

The premise here is to allow non-kernel functions to locate external LDS variables without using LDS or extra magic SGPRs to do so.

1/ First it crawls the callgraph to work out which external LDS variables are reachable from a given kernel
2/ Then it creates a new `extern char[0]` variable for each kernel, which will alias all the other extern LDS variables because that's the documented behaviour of these variables
3/ The address of that variable is written to a lookup table. The global variable is tagged with metadata to track what address it was allocated at by codegen
4/ The assembler builds the lookup table using the metadata
5/ Any non-kernel functions use the same magic intrinsic used by table lookups of non-dynamic LDS variables to find the address to use

Heavy overlap with the code paths taken for other lowering, in particular the same intrinsic is used to pass the dynamic scope information through the same sgpr as for table lookups of static LDS.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D144233

show more ...

[amdgpu][nfc] Refactor prior to D144233 to remove noise from diff

[amdgpu][nfc] clang-format AMDGPULowerModuleLDS for easier merging

[amdgpu][nfc] Extract more functions in LowerModuleLDS, mark more methods static

[amdgpu][nfc] Comment and extract two functions in LowerModuleLDS

[amdgpu][nfc] Replace ad hoc LDS frame recalculation with absolute_symbol MD

Post ISel, LDS variables are absolute values. Representing them as
such is simpler than the frame recalculation currently

[amdgpu][nfc] Replace ad hoc LDS frame recalculation with absolute_symbol MD

Post ISel, LDS variables are absolute values. Representing them as
such is simpler than the frame recalculation currently used to build assembler
tables from their addresses.

This is a precursor to lowering dynamic/external LDS accesses from non-kernel
functions.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D144221

show more ...

[ReplaceConstant] Extract code for expanding users of constant (NFC)

AMDGPU implements some handy code for expanding all constexpr
users of LDS globals. Extract the core logic into ReplaceConstant,

[ReplaceConstant] Extract code for expanding users of constant (NFC)

AMDGPU implements some handy code for expanding all constexpr
users of LDS globals. Extract the core logic into ReplaceConstant,
so that it can be reused elsewhere.

show more ...

[amdgpu] Change LDS lowering default to hybrid

Postponed from D139433 until the bug fixed by D139874 could be resolved.

Reviewed By: arsenm

Differential Revision: https://reviews.llvm.org/D141852

Revert D141134 "[NFC] Only expose getXXXSize functions in TypeSize"

The patch should be discussed further.

This reverts commit dd56e1c92b0e6e6be249f2d2dd40894e0417223f.

[NFC] Only expose getXXXSize functions in TypeSize

Currently 'TypeSize' exposes two functions that serve the same purpose:
- getFixedSize / getFixedValue
- getKnownMinSize / getKnownMinValue

sour

[NFC] Only expose getXXXSize functions in TypeSize

Currently 'TypeSize' exposes two functions that serve the same purpose:
- getFixedSize / getFixedValue
- getKnownMinSize / getKnownMinValue

source : https://github.com/llvm/llvm-project/blob/bf82070ea465969e9ae86a31dfcbf94c2a7b4c4c/llvm/include/llvm/Support/TypeSize.h#L337-L338

This patch offers to remove one of the two and stick to a single function in the code base.

Differential Revision: https://reviews.llvm.org/D141134

show more ...

AMDGPU: Diagnose which LDS global failed to lower

Also lowercase the message to start since that seems to be the
prevailing convention for error messages.