fabtests: Do not require FI_TAGGED for fi_av_xfer test #10828
Conversation
This test does not require tagged messaging Signed-off-by: Sai Sunku <sunkusa@amazon.com>
| @@ -233,8 +233,7 @@ int main(int argc, char **argv) | |||
| if (optind < argc) | |||
| opts.dst_addr = argv[optind]; | |||
|
|
|||
| hints->caps = hints->ep_attr->type == FI_EP_RDM ? | |||
| FI_TAGGED : FI_MSG; | |||
| hints->caps = FI_MSG; | |||
There was a problem hiding this comment.
Can we make it handled by -o? https://github.com/ofiwg/libfabric/blob/main/fabtests/common/shared.c#L3709-L3732
There was a problem hiding this comment.
But why? The test itself isn't testing messaging or tagged messaging. I don't know why FI_TAGGED was requested in the first place.
There was a problem hiding this comment.
I don't know why it was requesting FI_TAGGED when the test was added, especially for RDM ep type.... I'd like to know if there is any risks to make this test silently move from TAGGED to MSG, if there was a special motivation @shefty
There was a problem hiding this comment.
I think it's fine to choose one method and not make it a command line argument. My best guess on why it was set to tagged in the first place is because if you're using FI_EP_RDM then you're mostly likely using tag matching so it matches the applications more. I don't have any issues with switching it but I'm curious why you want to change it? Is there something that switching it would allow you to test that isn't possible right now?
There was a problem hiding this comment.
Yes, we have a code path that doesn't support TAGGED but support MSG.
There was a problem hiding this comment.
Gotcha, then yeah I think it's fine to switch to FI_MSG. @j-xiong is that ok with you?
This test does not require tagged messaging