From: Sagi Grimberg <sagig@dev.mellanox.co.il>
To: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
Tom Talpey <tom@talpey.com>
Cc: "'Christoph Hellwig'" <hch@infradead.org>,
Steve Wise <swise@opengridcomputing.com>,
dledford@redhat.com, sagig@mellanox.com, ogerlitz@mellanox.com,
roid@mellanox.com, linux-rdma@vger.kernel.org, eli@mellanox.com,
target-devel@vger.kernel.org, linux-nfs@vger.kernel.org,
trond.myklebust@primarydata.com, bfields@fieldses.org,
Oren Duer <oren@mellanox.com>
Subject: Re: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
Date: Tue, 14 Jul 2015 12:10:36 +0300 [thread overview]
Message-ID: <55A4D20C.2000904@dev.mellanox.co.il> (raw)
In-Reply-To: <20150713201538.GA11681@obsidianresearch.com>
On 7/13/2015 11:15 PM, Jason Gunthorpe wrote:
> On Mon, Jul 13, 2015 at 03:36:44PM -0400, Tom Talpey wrote:
>> On 7/11/2015 6:25 AM, 'Christoph Hellwig' wrote:
>>> I think what we need to support for now is FRMR as the primary target,
>>> and FMR as a secondar[y].
>>
>> FMR is a *very* bad choice, for several reasons.
>
> If an API can transparently support FMR, then I think it can also
> transparently support ib_get_phys_mr as an alternative, they look
> pretty similar... ?
Having an API that does FRMR/FMR/PHYS_MR is even worse from the ULP
PoV. If you expose an API that might schedule (PHYS_MR) it limits the
context that the caller is allowed to call in.
I'm 100% against an registration API that *might* schedule.
WARNING: multiple messages have this Message-ID (diff)
From: Sagi Grimberg <sagig-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
To: Jason Gunthorpe
<jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>,
Tom Talpey <tom-CLs1Zie5N5HQT0dZR+AlfA@public.gmane.org>
Cc: 'Christoph Hellwig' <hch-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>,
Steve Wise
<swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>,
dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org,
sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
roid-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
eli-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
target-devel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
trond.myklebust-7I+n7zu2hftEKMMhf/gKZA@public.gmane.org,
bfields-uC3wQj2KruNg9hUCZPvPmw@public.gmane.org,
Oren Duer <oren-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Subject: Re: [PATCH V3 1/5] RDMA/core: Transport-independent access flags
Date: Tue, 14 Jul 2015 12:10:36 +0300 [thread overview]
Message-ID: <55A4D20C.2000904@dev.mellanox.co.il> (raw)
In-Reply-To: <20150713201538.GA11681-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
On 7/13/2015 11:15 PM, Jason Gunthorpe wrote:
> On Mon, Jul 13, 2015 at 03:36:44PM -0400, Tom Talpey wrote:
>> On 7/11/2015 6:25 AM, 'Christoph Hellwig' wrote:
>>> I think what we need to support for now is FRMR as the primary target,
>>> and FMR as a secondar[y].
>>
>> FMR is a *very* bad choice, for several reasons.
>
> If an API can transparently support FMR, then I think it can also
> transparently support ib_get_phys_mr as an alternative, they look
> pretty similar... ?
Having an API that does FRMR/FMR/PHYS_MR is even worse from the ULP
PoV. If you expose an API that might schedule (PHYS_MR) it limits the
context that the caller is allowed to call in.
I'm 100% against an registration API that *might* schedule.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2015-07-14 9:10 UTC|newest]
Thread overview: 223+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-05 23:21 [PATCH V3 0/5] Transport-independent MRs Steve Wise
2015-07-05 23:22 ` [PATCH V3 1/5] RDMA/core: Transport-independent access flags Steve Wise
2015-07-06 5:25 ` Christoph Hellwig
2015-07-06 14:23 ` Steve Wise
2015-07-06 14:23 ` Steve Wise
2015-07-07 8:58 ` 'Christoph Hellwig'
2015-07-06 7:09 ` Haggai Eran
2015-07-06 7:09 ` Haggai Eran
2015-07-06 14:29 ` Steve Wise
2015-07-06 14:29 ` Steve Wise
2015-07-07 14:17 ` Steve Wise
2015-07-07 14:17 ` Steve Wise
2015-07-07 14:34 ` Haggai Eran
2015-07-07 14:34 ` Haggai Eran
2015-07-07 14:46 ` Steve Wise
2015-07-07 14:46 ` Steve Wise
2015-07-07 15:07 ` Haggai Eran
2015-07-07 15:07 ` Haggai Eran
2015-07-06 7:53 ` Sagi Grimberg
2015-07-06 7:53 ` Sagi Grimberg
2015-07-06 14:37 ` Steve Wise
2015-07-06 14:37 ` Steve Wise
2015-07-06 16:17 ` Sagi Grimberg
2015-07-06 16:17 ` Sagi Grimberg
2015-07-06 16:55 ` Steve Wise
2015-07-06 16:55 ` Steve Wise
2015-07-07 9:00 ` Christoph Hellwig
2015-07-07 9:14 ` Sagi Grimberg
2015-07-07 9:14 ` Sagi Grimberg
2015-07-07 14:05 ` Steve Wise
2015-07-07 14:05 ` Steve Wise
2015-07-07 16:17 ` Jason Gunthorpe
2015-07-07 16:17 ` Jason Gunthorpe
2015-07-07 16:27 ` Sagi Grimberg
2015-07-07 16:27 ` Sagi Grimberg
2015-07-07 21:36 ` Jason Gunthorpe
2015-07-07 21:36 ` Jason Gunthorpe
2015-07-08 7:29 ` Sagi Grimberg
2015-07-08 8:13 ` 'Christoph Hellwig'
2015-07-08 8:13 ` 'Christoph Hellwig'
2015-07-08 10:05 ` Sagi Grimberg
2015-07-08 10:05 ` Sagi Grimberg
2015-07-08 10:20 ` 'Christoph Hellwig'
2015-07-08 10:20 ` 'Christoph Hellwig'
2015-07-08 11:08 ` Sagi Grimberg
2015-07-08 11:08 ` Sagi Grimberg
2015-07-08 17:14 ` Hefty, Sean
2015-07-08 17:14 ` Hefty, Sean
2015-07-09 8:46 ` Sagi Grimberg
2015-07-09 13:52 ` Chuck Lever
2015-07-10 19:34 ` Christoph Hellwig
2015-07-12 7:49 ` Sagi Grimberg
2015-07-13 16:50 ` Jason Gunthorpe
2015-07-14 8:06 ` Sagi Grimberg
2015-07-14 12:24 ` Tom Talpey
2015-07-14 12:24 ` Tom Talpey
2015-07-14 13:21 ` Sagi Grimberg
2015-07-14 13:21 ` Sagi Grimberg
2015-07-23 0:43 ` Hefty, Sean
2015-07-23 0:43 ` Hefty, Sean
2015-07-08 19:08 ` Jason Gunthorpe
2015-07-08 20:32 ` 'Christoph Hellwig'
2015-07-08 20:32 ` 'Christoph Hellwig'
2015-07-08 20:37 ` 'Christoph Hellwig'
2015-07-08 20:37 ` 'Christoph Hellwig'
2015-07-09 0:03 ` Jason Gunthorpe
2015-07-09 8:00 ` 'Christoph Hellwig'
2015-07-09 8:00 ` 'Christoph Hellwig'
2015-07-09 8:58 ` Sagi Grimberg
2015-07-09 8:58 ` Sagi Grimberg
2015-07-09 22:18 ` Doug Ledford
2015-07-09 22:18 ` Doug Ledford
2015-07-09 22:53 ` Jason Gunthorpe
2015-07-09 22:53 ` Jason Gunthorpe
2015-07-10 13:22 ` Tom Talpey
2015-07-10 16:11 ` Jason Gunthorpe
2015-07-10 16:11 ` Jason Gunthorpe
2015-07-10 17:56 ` Doug Ledford
2015-07-10 18:34 ` Chuck Lever
2015-07-10 18:42 ` Tom Talpey
2015-07-10 19:54 ` Jason Gunthorpe
2015-07-10 19:54 ` Jason Gunthorpe
2015-07-10 20:48 ` Jason Gunthorpe
2015-07-10 22:33 ` Doug Ledford
2015-07-10 22:33 ` Doug Ledford
2015-07-11 10:17 ` 'Christoph Hellwig'
2015-07-11 10:17 ` 'Christoph Hellwig'
2015-07-13 16:57 ` Jason Gunthorpe
2015-07-13 16:57 ` Jason Gunthorpe
2015-07-14 7:25 ` 'Christoph Hellwig'
2015-07-14 9:05 ` Sagi Grimberg
2015-07-14 15:35 ` 'Christoph Hellwig'
2015-07-14 17:26 ` Jason Gunthorpe
2015-07-15 7:10 ` Sagi Grimberg
2015-07-15 7:10 ` Sagi Grimberg
2015-07-10 22:30 ` Doug Ledford
2015-07-10 22:30 ` Doug Ledford
2015-07-10 20:57 ` Jason Gunthorpe
2015-07-10 22:27 ` Doug Ledford
2015-07-10 22:27 ` Doug Ledford
[not found] ` <20150710233417.GA8919@obsidianresearch.com>
2015-07-11 3:10 ` Doug Ledford
2015-07-11 3:10 ` Doug Ledford
2015-07-13 17:18 ` Jason Gunthorpe
2015-07-13 17:18 ` Jason Gunthorpe
2015-07-13 22:23 ` Tom Talpey
2015-07-11 16:37 ` Steve Wise
2015-07-12 10:46 ` Sagi Grimberg
2015-07-12 10:46 ` Sagi Grimberg
2015-07-14 19:25 ` Steve Wise
2015-07-14 19:25 ` Steve Wise
2015-07-14 19:29 ` Jason Gunthorpe
2015-07-14 19:32 ` Steve Wise
2015-07-14 19:32 ` Steve Wise
2015-07-14 19:37 ` Jason Gunthorpe
2015-07-14 19:55 ` 'Christoph Hellwig'
2015-07-14 19:55 ` 'Christoph Hellwig'
2015-07-14 20:10 ` Steve Wise
2015-07-14 20:10 ` Steve Wise
2015-07-14 20:29 ` Jason Gunthorpe
2015-07-14 20:29 ` Jason Gunthorpe
2015-07-14 20:40 ` Steve Wise
2015-07-14 20:40 ` Steve Wise
2015-07-14 20:44 ` Jason Gunthorpe
2015-07-14 20:44 ` Jason Gunthorpe
2015-07-14 20:54 ` Steve Wise
2015-07-14 20:54 ` Steve Wise
2015-07-14 20:59 ` Jason Gunthorpe
2015-07-14 20:59 ` Jason Gunthorpe
2015-07-14 20:50 ` Tom Talpey
2015-07-14 20:50 ` Tom Talpey
2015-07-15 6:50 ` 'Christoph Hellwig'
2015-07-15 19:12 ` Jason Gunthorpe
2015-07-15 19:12 ` Jason Gunthorpe
2015-07-16 6:41 ` Jason Gunthorpe
2015-07-16 6:41 ` Jason Gunthorpe
2015-07-16 8:04 ` 'Christoph Hellwig'
2015-07-16 8:04 ` 'Christoph Hellwig'
2015-07-16 16:13 ` Jason Gunthorpe
2015-07-16 16:13 ` Jason Gunthorpe
2015-07-15 8:47 ` Sagi Grimberg
2015-07-15 8:47 ` Sagi Grimberg
2015-07-15 12:19 ` 'Christoph Hellwig'
2015-07-15 12:19 ` 'Christoph Hellwig'
2015-07-15 19:17 ` Jason Gunthorpe
2015-07-15 19:17 ` Jason Gunthorpe
2015-07-14 20:46 ` Tom Talpey
2015-07-14 19:45 ` 'Christoph Hellwig'
2015-07-14 19:57 ` Jason Gunthorpe
2015-07-14 19:57 ` Jason Gunthorpe
2015-07-14 19:58 ` Steve Wise
2015-07-14 19:58 ` Steve Wise
2015-07-14 20:41 ` Jason Gunthorpe
2015-07-14 20:41 ` Jason Gunthorpe
2015-07-14 20:51 ` Steve Wise
2015-07-14 20:51 ` Steve Wise
2015-07-14 21:01 ` Steve Wise
2015-07-14 21:01 ` Steve Wise
2015-07-14 21:14 ` Jason Gunthorpe
2015-07-14 21:14 ` Jason Gunthorpe
2015-07-23 18:53 ` Hefty, Sean
2015-07-23 18:53 ` Hefty, Sean
2015-07-23 19:03 ` Steve Wise
2015-07-23 19:03 ` Steve Wise
2015-07-23 23:30 ` Hefty, Sean
2015-07-23 23:30 ` Hefty, Sean
2015-07-23 23:53 ` Jason Gunthorpe
2015-07-23 23:53 ` Jason Gunthorpe
2015-07-24 0:18 ` Hefty, Sean
2015-07-24 0:18 ` Hefty, Sean
2015-07-24 4:46 ` Jason Gunthorpe
2015-07-09 8:47 ` Sagi Grimberg
2015-07-09 8:47 ` Sagi Grimberg
2015-07-08 21:38 ` Tom Talpey
2015-07-08 23:36 ` Jason Gunthorpe
2015-07-09 11:02 ` Sagi Grimberg
2015-07-09 17:01 ` Jason Gunthorpe
2015-07-09 17:01 ` Jason Gunthorpe
2015-07-09 20:00 ` Tom Talpey
2015-07-09 21:16 ` Jason Gunthorpe
[not found] ` <20150709170142.GA21921-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2015-07-10 8:55 ` kernel memory registration (was: RDMA/core: Transport-independent access flags) Sagi Grimberg
[not found] ` <559F8881.7070308-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-07-10 16:35 ` Jason Gunthorpe
2015-07-11 10:31 ` 'Christoph Hellwig'
[not found] ` <20150711103153.GC14741-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
2015-07-13 16:46 ` Jason Gunthorpe
[not found] ` <20150713164652.GC23832-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2015-07-14 8:24 ` kernel memory registration Sagi Grimberg
[not found] ` <55A4C73A.7080001-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-07-14 17:24 ` Jason Gunthorpe
2015-07-11 10:25 ` [PATCH V3 1/5] RDMA/core: Transport-independent access flags 'Christoph Hellwig'
2015-07-13 16:35 ` Jason Gunthorpe
2015-07-13 19:36 ` Tom Talpey
2015-07-13 20:15 ` Jason Gunthorpe
2015-07-14 9:10 ` Sagi Grimberg [this message]
2015-07-14 9:10 ` Sagi Grimberg
2015-07-14 15:36 ` 'Christoph Hellwig'
2015-07-14 15:36 ` 'Christoph Hellwig'
2015-07-14 15:47 ` Tom Talpey
2015-07-14 15:47 ` Tom Talpey
2015-07-14 16:22 ` Jason Gunthorpe
2015-07-14 16:22 ` Jason Gunthorpe
2015-07-14 7:37 ` 'Christoph Hellwig'
2015-07-14 9:22 ` Sagi Grimberg
2015-07-14 12:12 ` Tom Talpey
2015-07-14 12:12 ` Tom Talpey
2015-07-14 13:23 ` Sagi Grimberg
2015-07-14 14:45 ` Steve Wise
2015-07-14 14:45 ` Steve Wise
2015-07-14 15:40 ` 'Christoph Hellwig'
2015-07-14 15:40 ` 'Christoph Hellwig'
2015-07-08 8:11 ` 'Christoph Hellwig'
2015-07-06 7:58 ` Sagi Grimberg
2015-07-06 7:58 ` Sagi Grimberg
2015-07-06 14:39 ` Steve Wise
2015-07-06 14:39 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 2/5] RDMA/iser: Use transport independent MR allocation Steve Wise
2015-07-05 23:22 ` [PATCH V3 3/5] RDMA/isert: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 4/5] svcrdma: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-05 23:22 ` [PATCH V3 5/5] xprtrdma: " Steve Wise
2015-07-05 23:22 ` Steve Wise
2015-07-06 5:25 ` [PATCH V3 0/5] Transport-independent MRs Christoph Hellwig
2015-07-06 5:25 ` Christoph Hellwig
2015-07-06 14:24 ` Steve Wise
2015-07-06 14:24 ` Steve Wise
2015-07-07 9:01 ` 'Christoph Hellwig'
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55A4D20C.2000904@dev.mellanox.co.il \
--to=sagig@dev.mellanox.co.il \
--cc=bfields@fieldses.org \
--cc=dledford@redhat.com \
--cc=eli@mellanox.com \
--cc=hch@infradead.org \
--cc=jgunthorpe@obsidianresearch.com \
--cc=linux-nfs@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=ogerlitz@mellanox.com \
--cc=oren@mellanox.com \
--cc=roid@mellanox.com \
--cc=sagig@mellanox.com \
--cc=swise@opengridcomputing.com \
--cc=target-devel@vger.kernel.org \
--cc=tom@talpey.com \
--cc=trond.myklebust@primarydata.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.