[SYCL] Fix ICE when -g and -ffile-prefix-map flags are set

[zahiraam]

Set the source file name when footer is present and option fmacro-prefix-map is used when generating debug information.

This patch is to fix a crash in the compiler when generating debug information with the option -ffile-prefix-map.
Example:
$ pwd
/tmp/tmp.jocQk9WVa0
$
$ cat builtin.cpp
void a(__builtin_va_list);
using ::a;
$
$ clang -c -g -fsycl /tmp/tmp.jocQk9WVa0/builtin.cpp
$
$ clang -g -fsycl -ffile-prefix-map=/tmp/= /tmp/tmp.jocQk9WVa0/builtin.cpp
Expected must be checked before access or destruction.
Unchecked Expected contained error:
No such file or directoryPLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace, preprocessed source, and associated run script.
Stack dump:
0. Program arguments:
…………………/tmp/tmp.jocQk9WVa0/builtin.cpp

1. parser at end of file
2. /tmp/tmp.jocQk9WVa0/builtin.cpp:2:9: LLVM IR generation of declaration 'a

https://github.com/intel/llvm/blob/sycl/clang/lib/CodeGen/CGDebugInfo.cpp#L397 is returning a error because the FileName is wrong.

[elizabethandrews]

Can you describe what this change is doing in more detail in PR description?

[zahiraam]

@bwyma Thanks for reviewing and checking the debug information. I have fixed the issue you raised about the directory path. Also added in the other LIT tests a check line for the directory to make sure it's correct (instead of using regular expression). Would you please check if this new patch fixes the issue? Thanks.

[bwyma]

LGTM.

[zahiraam]

LGTM.

Thanks.

[elizabethandrews]

I'm not very familiar with debug information that needs to be generated but I trust @bwyma's approval!

[zahiraam]

@intel/dpcpp-devops-reviewers I have been re-basing and re-testing this PR for a week. Can someone help with the latest fail please? Thanks.
It says:
"This request was automatically failed because there were no enabled runners online to process the request for more than 1 days."

[aelovikov-intel]

@apstasen , could you take a look please? We've seen similar issue with not getting AWS CUDA runner picked up in at least one another PR.

[bader]

@intel/dpcpp-devops-reviewers I have been re-basing and re-testing this PR for a week. Can someone help with the latest fail please? Thanks. It says: "This request was automatically failed because there were no enabled runners online to process the request for more than 1 days."

I hope #8567 fixes this issue.

[zahiraam]

@intel/dpcpp-devops-reviewers I have been re-basing and re-testing this PR for a week. Can someone help with the latest fail please? Thanks. It says: "This request was automatically failed because there were no enabled runners online to process the request for more than 1 days."

I hope #8567 fixes this issue.

The testing started 14h 14m 8s ago. Unlless I am mistaken this is still hanging?
"Requested labels: aws-cuda_4318684912-4
Job defined at: intel/llvm/.github/workflows/sycl_linux_build_and_test.yml@refs/heads/sycl
Reusable workflow chain:
intel/llvm/.github/workflows/sycl_precommit.yml@refs/heads/sycl (fe12cab)
-> intel/llvm/.github/workflows/sycl_linux_build_and_test.yml@refs/heads/sycl (fe12cab)
Waiting for a runner to pick up this job..."
I seem to have no way of re-starting the test again.
@bader, @stasenko, Alexander P?

[bader]

@zahiraam, you need to merge the fix to your DBI branch.

[zahiraam]

@bader Your fix seems to have worked. Thanks.
Can we now merge in this PR?