Re: [openib-general] Re: RFC: ipath ioctls and their replacements

From: Sean Hefty <mshefty@xxxxxxxxxxxxxxxx>
Date: Thu, 19 Jan 2006 16:02:16 -0800

Bryan O'Sullivan wrote:

We are not really in the RDMA camp.  Our facility looks more like "when
this kind of message comes in, be sure that it shows up at this point in
my address space", which does not match RDMA semantics.

A lot of people mean QP-like semantics when they talk about "RDMA", rather than the RDMA operation itself. I.e. pre-posted receive buffers associated with a particular user-space process.

That aside, conceptually, I see little difference between RDMA semantics versus the facility that you describe. The main difference is the complexity of the header and the checks done against it.

- Sean
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

References:
- RFC: ipath ioctls and their replacements
  - From: Bryan O'Sullivan
- Re: RFC: ipath ioctls and their replacements
  - From: Greg KH
- Re: RFC: ipath ioctls and their replacements
  - From: Bryan O'Sullivan
- Re: RFC: ipath ioctls and their replacements
  - From: Greg KH
- Re: RFC: ipath ioctls and their replacements
  - From: Bryan O'Sullivan
- Re: RFC: ipath ioctls and their replacements
  - From: Greg KH
- Re: RFC: ipath ioctls and their replacements
  - From: Bryan O'Sullivan

Prev by Date: Re: [PATCH 000 of 5] md: Introduction
Next by Date: [PATCH] shrink_list: Use of && instead || leads to unintended writing of pages
Previous by thread: Re: RFC: ipath ioctls and their replacements
Next by thread: Re: RFC: ipath ioctls and their replacements
Index(es):
- Date
- Thread

Relevant Pages

Re: [openib-general] Re: [PATCH repost] PROT_DONTCOPY: ifiniband uverbs fork support
... > that uncleanliness - but both of these would be in mmap only, ... They are both relying on the way user allocates memory for RDMA. ... send the line "unsubscribe linux-kernel" in ...
(Linux-Kernel)
Re: [openib-general] Re: [PATCH repost] PROT_DONTCOPY: ifiniband uverbs fork support
... > They are both relying on the way user allocates memory for RDMA. ... > kernel that the pages are used for RDMA and it is not safe to copy them now. ... Your stack example is a good one: if we end up setting VM_DONTCOPY on ... send the line "unsubscribe linux-kernel" in ...
(Linux-Kernel)
Re: chasing the four level page table
... I'm working on InfiniBand drivers, which will be doing RDMA. ... wants to do something like mlock() to keep any pages to be DMAed ... send the line "unsubscribe linux-kernel" in ...
(Linux-Kernel)
Re: [openib-general] Re: [PATCH repost] PROT_DONTCOPY: ifiniband uverbs fork support
... >> They are both relying on the way user allocates memory for RDMA. ... >> kernel that the pages are used for RDMA and it is not safe to copy them now. ... but I prefer child SIGSEGV than silent data corruption. ... send the line "unsubscribe linux-kernel" in ...
(Linux-Kernel)
Re: [PATCH][RFC][0/4] InfiniBand userspace verbs implementation
... The whole point behind RDMA is that these APIs exist and are being ... Staff Software Engineer ... send the line "unsubscribe linux-kernel" in ...
(Linux-Kernel)