-
Notifications
You must be signed in to change notification settings - Fork 721
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Replace CPU APIs on x86 #9421
Replace CPU APIs on x86 #9421
Conversation
1d1984e
to
b1da9d6
Compare
f1b777d
to
b636af0
Compare
I believe this needs to be merged in tandem with eclipse/omr#5151 right? |
Yes |
58439ee
to
126b6ac
Compare
126b6ac
to
cfb6657
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Throughout this commit I would replace all the asserts that will go away with TR_ASSERT_FATAL.
3d8642e
to
02f50d0
Compare
02f50d0
to
9bd8f42
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
TravisCI seems to have some problems:
|
This is because this change depends on eclipse/omr#5151. It's expected. |
It would be nice if this could be merged without breaking the OpenJ9 builds while waiting for OMR promotion. If the OMR changes can be merged first and OpenJ9 still builds, then nothing else is required. Otherwise, the OMR change should be merged first, and then an OMR promotion PR build can be run like "test sanity jdk8,jdk11 depends eclipse/omr#master". If this build passes or only has known failures, the tested OMR level can be pushed from master to the openj9 branch at the same time this PR is merged. Some preliminary testing should be done even before the OMR change is merged, as OMR promotion will be blocked waiting on the successful promotion build. |
The OMR changes will be merged first and they won't break OpenJ9. |
9bd8f42
to
7c48505
Compare
Jenkins test sanity xlinux,win jdk8 |
This change depends on eclipse/omr#5151 btw. |
Jenkins test sanity xlinux,win jdk8 depends eclipse/omr#5151 |
Seeing one failure on windows
Tried a test run without my change last night and saw this failure as well, therefore don't think it's caused by this change. |
That failure seems to be because of #9707. |
Caught a zlinux crash with this one. Looking into it. |
How is that possible if the code here only affects x86 codegen? |
The problem is caused by this line in OMR: (This PR exposed this issue)
Can't query options in |
TR::CPU
OMR::Z::CPU::detect(OMRPortLibrary * const omrPortLib)
{
if (omrPortLib == NULL)
return TR::CPU();
OMRPORT_ACCESS_FROM_OMRPORT(omrPortLib);
OMRProcessorDesc processorDescription;
omrsysinfo_get_processor_description(&processorDescription);
if (processorDescription.processor >= OMR_PROCESSOR_S390_Z10 && TR::Options::getCmdLineOptions()->getOption(TR_DisableZ10))
processorDescription.processor = OMR_PROCESSOR_S390_FIRST; |
7c48505
to
faa7bec
Compare
May still affect other platforms. Let me verify, and possibly will try to limit it to x86 only. |
Actually all we need to do is to merge #9431 first. (I'm going to verify whether #9431 breaks anything). |
- Replace old processor APIs with new APIs from the CPU class - Replace TR::Compiler target with per comp target depends on: eclipse/omr#5151 Signed-off-by: Harry Yu <harryyu1994@gmail.com>
faa7bec
to
163729e
Compare
@fjeremic Hey Filip, I have verified that this one works on all platforms. I think it's ready to be merged. |
Jenkins test sanity xlinux,win,osx jdk8,jdk11 |
Seeing 1 intermittent failure (Most likely not related to my change).
|
Jenkins test sanity xlinux jdk11 |
depends on: eclipse/omr#5151
depends on: eclipse/omr#5153
Signed-off-by: Harry Yu harryyu1994@gmail.com