Testsuite Status A01a30020c78: GCC Postcommit CI Report

Hey guys! Let's dive into the latest testsuite status report. This discussion covers the status of the testsuite for commit a01a30020c78ba15d26784a7ec390ea0a3926895. This report, generated by the GCC postcommit CI, gives us a detailed look at the new failures, resolved failures, and unresolved failures across different architectures and configurations.

Summary of Testsuite Status

Alright, let's break down the summary tables. These tables give us a bird's-eye view of what's happening with the GCC, G++, and GFortran compilers across various newlib configurations. We'll check out the new failures, the good news about resolved issues, and the persistent unresolved failures. Knowing this helps us zoom in on the areas needing the most attention. So, let's get started!

New Failures

The following table highlights the new failures observed in the testsuite:

This table indicates that there are new failures specifically in the gcc compiler across different newlib configurations. For example, the newlib: RVA23U64 profile lp64d medlow multilib configuration shows 1 new failure out of 1 test in gcc. These failures need our attention to ensure the compiler is functioning correctly in these environments.

Resolved Failures

Now for some good news! The following table shows the failures that have been resolved:

It’s awesome to see that issues have been resolved in these configurations. The gcc compiler shows improvements, which is always encouraging. Specifically, configurations like newlib: rv32imac ilp32 medlow multilib have seen resolutions, meaning previous failures are now passing. This indicates that fixes and improvements are making their way into the codebase.

Unresolved Failures

Okay, now let's tackle the section on unresolved failures. This is where we see the issues that are still sticking around and need our attention. Looking closely at these tables helps us figure out where the persistent problems lie, so we can strategize on how to knock them out. Let's dive in and see what's still on the list!

As you can see, there are a significant number of unresolved failures, particularly in the linux and newlib configurations. The gcc compiler has a higher number of unresolved failures compared to g++ and gfortran. For instance, the linux: RVA23U64 profile lp64d medlow multilib configuration shows a whopping 385 failures out of 86 tests for gcc. Similarly, the newlib: rv32gcv ilp32d medlow multilib configuration has 609 failures out of 131 tests. These persistent failures indicate that there are underlying issues that need to be addressed to improve the stability and reliability of the compiler.

Architecture Specific New Failures

Let's zoom in on the architecture-specific new failures. This section gives us the nitty-gritty on which specific tests are failing in certain configurations. It's like having a magnifying glass for our problems, so we can see exactly what's going wrong and where. Let's get to it and break down the specifics!

FAIL: gcc.target/riscv/target-attr-06.c   -O0  (test for excess errors)

newlib rv64imafdcv_zicond_zawrs_zbc_zvkng_zvksg_zvbb_zvbc_zicsr_zba_zbb_zbs_zicbom_zicbop_zicboz_zfhmin_zkt lp64d medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -Os  (test for excess errors)

newlib rv32gcv ilp32d medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -Os  (test for excess errors)

newlib rv32imc_zba_zbb_zbc_zbs ilp32 medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -O2  (test for excess errors)

newlib rv64imc_zba_zbb_zbc_zbs lp64 medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -O2  (test for excess errors)

Here, we see that the gcc.target/riscv/target-attr-06.c test is consistently failing across various newlib configurations. The failures occur with different optimization levels (-O0, -Os, -O2), suggesting the issue might be related to how the compiler handles target-specific attributes or optimizations for the RISC-V architecture. It's crucial to investigate this test case further to identify the root cause and implement a fix.

Architecture Specific Resolved Failures

Time for some more good news! Let's check out the architecture-specific failures that have been resolved. This is where we see the direct impact of our efforts to squash bugs and improve the compiler. It's always motivating to see these successes, so let's dive into the details!

FAIL: gcc.target/riscv/target-attr-06.c   -O1  (test for excess errors)
FAIL: gcc.target/riscv/target-attr-06.c   -Os  (test for excess errors)

newlib rv64gc lp64d medlow:

FAIL: gcc.target/riscv/target-attr-06.c   -O2  (test for excess errors)

newlib rv64imafdcv_zicond_zawrs_zbc_zvkng_zvksg_zvbb_zvbc_zicsr_zba_zbb_zbs_zicbom_zicbop_zicboz_zfhmin_zkt lp64d medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -O0  (test for excess errors)

newlib rv32imac_zba_zbb_zbc_zbs ilp32 medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -Os  (test for excess errors)

newlib rv64imc_zba_zbb_zbc_zbs lp64 medlow multilib:

FAIL: gcc.target/riscv/target-attr-06.c   -Os  (test for excess errors)

It's fantastic to see that failures related to gcc.target/riscv/target-attr-06.c have been resolved in several configurations. This indicates that progress has been made in addressing the issues with target-specific attributes for the RISC-V architecture. However, it’s important to note that this test is still causing new failures in other configurations, so further investigation is necessary.

Associated Run

For more detailed information, you can check out the associated run on the GCC postcommit CI: https://github.com/patrick-rivos/gcc-postcommit-ci/actions/runs/19079063362

Conclusion

Alright guys, wrapping things up, we’ve taken a detailed look at the testsuite status for commit a01a30020c78ba15d26784a7ec390ea0a3926895. We've seen a mix of new, resolved, and persistent failures, giving us a clear picture of where the GCC compiler needs some extra love. It's awesome to celebrate the resolved issues, but we can't ignore those unresolved ones. They're like the puzzles we need to solve to make our compiler even better.

Specifically, the recurring failures in gcc.target/riscv/target-attr-06.c across different optimization levels and newlib configurations tell us we need to dig deeper into how the compiler handles RISC-V target attributes. By tackling these persistent issues, we're not just fixing bugs; we're boosting the overall reliability and stability of GCC.

So, let's keep the ball rolling! Your insights and contributions are super important as we work together to make GCC the best it can be. Thanks for tuning in, and let's keep crushing those bugs!