mercury

mercury copied to clipboard

NA OFI: Memory leak in _verbs_init_and_alloc_context

2

I am using mercury in my application for RPC. When I exit the application, cleanup of mercury is done as per documentation but I am getting a couple of memory...

lokesharo

**Is your feature request related to a problem? Please describe.** There are some code paths in Mercury that will allocate extra memory if needed (for example, to satisfy a request...

carns

**Is your feature request related to a problem? Please describe.** Currently DAOS is collecting information directly from libfabric to understand how to assign fabric interfaces appropriately to client processes. Now...

kjacque

We are encountering a problem in DAOS with how to properly finalize/shutdown servers using APIs that mercury provides today. The root of the problem is that in order to cleanly...

frostedcmos

HG: safe mechanism to deregister an RPC while handles for that RPC are still in use

3

**Is your feature request related to a problem? Please describe.** Imagine a hypothetical scenario in which a service is periodically receiving a particular RPC type. The service then begins to...

carns

Warning when registering an RPC with an already used id

2

It would be useful if Mercury could print a warning if `HG_Register` is called with an id that is already used. That way we would catch (1) programming errors when...

mdorier

HG_Get_input_extra_buf giving back size = 0

2

I'm not sure if this is a bug or a misunderstanding on my part of what `HG_Get_input_extra_buf` does (the same behavior happens with `HG_Get_ouput_extra_buf`), but I can't seem to be...

mdorier

**Describe the bug** I have a simple Mercury client and server (code available at the end of this issue). The client sends RPCs to the server indefinitely and the server...

mdorier

OFI can generate cancel events without users calling cancel resulting in callbacks returning an NA_CANCELED event. This is confusing for users and we should return an appropriate error code instead.

soumagne

HG util: cpu_spinwait in atomic_queue has bad interaction w cpu binding

1

Original message from Chuck: > I have found a bad feature interaction between mercury's atomic queue data structure, NA_NO_BLOCK, and thread CPU binding. > The impact of this issue varies...

soumagne