It is possible for some waves in a workgroup to finish their save sequence before the group leader has had time to capture the workgroup barrier state. When this happens, having those waves exit do impact the barrier state. As a consequence, the state captured by the group leader is invalid, and is eventually incorrectly restored.
This patch proposes to have all waves in a workgroup wait for each other at the end of their save sequence (just before calling s_endpgm_saved).
This is a cherry-pick. The cwsr_trap_handler.h part of the original part was valid and applied cleanly. The part of the patch that applied to cwsr_trap_handler_gfx12.asm did not apply cleanly since 80ae55e6115ef "drm/amdkfd: Move gfx12 trap handler to separate file" is not part of this branch. Instead, I ported the change to cwsr_trap_handler_gfx10.asm, and guarded it with "ASIC_FAMILY >= CHIP_GFX12".
Signed-off-by: Lancelot SIX lancelot.six@amd.com Reviewed-by: Jay Cornwall jay.cornwall@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com Cc: stable@vger.kernel.org # 6.12.x (cherry picked from commit d584198a6fe4c51f4aa88ad72f258f8961a0f11c) Signed-off-by: Lancelot SIX lancelot.six@amd.com --- drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h | 3 ++- drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm | 6 ++++++ 2 files changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h index 02f7ba8c93cd4..7062f12b5b751 100644 --- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h +++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler.h @@ -4117,7 +4117,8 @@ static const uint32_t cwsr_trap_gfx12_hex[] = { 0x0000ffff, 0x8bfe7e7e, 0x8bea6a6a, 0xb97af804, 0xbe804ec2, 0xbf94fffe, - 0xbe804a6c, 0xbfb10000, + 0xbe804a6c, 0xbe804ec2, + 0xbf94fffe, 0xbfb10000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0xbf9f0000, 0x00000000, diff --git a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm index 44772eec9ef4d..3ef294ead56c5 100644 --- a/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm +++ b/drivers/gpu/drm/amd/amdkfd/cwsr_trap_handler_gfx10.asm @@ -1463,6 +1463,12 @@ L_RETURN_WITHOUT_PRIV: s_rfe_b64 s_restore_pc_lo //Return to the main shader program and resume execution
L_END_PGM: +#if ASIC_FAMILY >= CHIP_GFX12 + // Make sure that no wave of the workgroup can exit the trap handler + // before the workgroup barrier state is saved. + s_barrier_signal -2 + s_barrier_wait -2 +#endif s_endpgm_saved end
base-commit: 984391de59a1d6918ac9ba63c095decbcfc85c71