From: Sagi Grimberg <sagig@mellanox.com>
To: "santosh.shilimkar@oracle.com" <santosh.shilimkar@oracle.com>,
"Sagi Grimberg" <sagig@dev.mellanox.co.il>,
<linux-rdma@vger.kernel.org>
Cc: <linux-nfs@vger.kernel.org>,
"Nicholas A. Bellinger" <nab@linux-iscsi.org>
Subject: Re: [PATCH v1 00/24] New fast registration API
Date: Tue, 22 Sep 2015 10:19:06 +0300 [thread overview]
Message-ID: <560100EA.8080302@mellanox.com> (raw)
In-Reply-To: <5600928A.7070202@oracle.com>
>
> As mentioned earlier, I have a WIP RDS fastreg branch [3]
> which is functional (at least I can RDMA messages across
> nodes ;-)).
Nice!
> So merging [2] and [3], I created [4] and applied
> a delta change based on your other patches. I saw ib_post_send
> failure with my HCA driver returning '-EINVAL'. I didn't
> debug it further but at least opcode and num_sge were set
> correctly so I shouldn't have seen it. So I did memset()
> on reg_wr which seems to have helped to fix the ib_post_send()
> failure.
Yep - that was my fault. When converting the ULPs I optimized by
removing the memset but I forgot to set reg_wr.wr.next = NULL when
the ULP needed. This caused the driver to read a second bogus work
request. Steve just reported this as well so I'll fix that in v2.
>
> But I got into remote access errors which tells me that I
> have messed up setup(rkey, sge setup or access flags)
One thing that pops is that in the old API the MR was registered
with iova_start = 0 (which is probably what was sent to the peer),
but the new API the iova is implicitly sg_dma_address(&sg[0]).
The registered MR holds these attributes in:
mr->rkey
mr->iova
mr->length
These should be passed to a peer to perform rdma.
Hope this helps,
Sagi.
WARNING: multiple messages have this Message-ID (diff)
From: Sagi Grimberg <sagig-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
To: "santosh.shilimkar-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org"
<santosh.shilimkar-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>,
Sagi Grimberg
<sagig-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>,
linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Cc: linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
"Nicholas A. Bellinger"
<nab-IzHhD5pYlfBP7FQvKIMDCQ@public.gmane.org>
Subject: Re: [PATCH v1 00/24] New fast registration API
Date: Tue, 22 Sep 2015 10:19:06 +0300 [thread overview]
Message-ID: <560100EA.8080302@mellanox.com> (raw)
In-Reply-To: <5600928A.7070202-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
>
> As mentioned earlier, I have a WIP RDS fastreg branch [3]
> which is functional (at least I can RDMA messages across
> nodes ;-)).
Nice!
> So merging [2] and [3], I created [4] and applied
> a delta change based on your other patches. I saw ib_post_send
> failure with my HCA driver returning '-EINVAL'. I didn't
> debug it further but at least opcode and num_sge were set
> correctly so I shouldn't have seen it. So I did memset()
> on reg_wr which seems to have helped to fix the ib_post_send()
> failure.
Yep - that was my fault. When converting the ULPs I optimized by
removing the memset but I forgot to set reg_wr.wr.next = NULL when
the ULP needed. This caused the driver to read a second bogus work
request. Steve just reported this as well so I'll fix that in v2.
>
> But I got into remote access errors which tells me that I
> have messed up setup(rkey, sge setup or access flags)
One thing that pops is that in the old API the MR was registered
with iova_start = 0 (which is probably what was sent to the peer),
but the new API the iova is implicitly sg_dma_address(&sg[0]).
The registered MR holds these attributes in:
mr->rkey
mr->iova
mr->length
These should be passed to a peer to perform rdma.
Hope this helps,
Sagi.
--
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:[~2015-09-22 7:32 UTC|newest]
Thread overview: 134+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-17 9:42 [PATCH v1 00/24] New fast registration API Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 01/24] IB/core: Introduce new " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-22 21:21 ` Bart Van Assche
2015-09-22 21:21 ` Bart Van Assche
2015-09-24 7:37 ` Sagi Grimberg
2015-09-24 7:37 ` Sagi Grimberg
2015-09-28 20:57 ` Bart Van Assche
2015-09-28 20:57 ` Bart Van Assche
2015-09-29 5:59 ` Christoph Hellwig
2015-09-29 5:59 ` Christoph Hellwig
2015-09-29 6:47 ` Sagi Grimberg
2015-09-29 6:47 ` Sagi Grimberg
2015-09-29 6:49 ` Sagi Grimberg
2015-09-29 6:49 ` Sagi Grimberg
2015-09-29 6:42 ` Sagi Grimberg
2015-09-29 6:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 02/24] IB/mlx5: Remove dead fmr code Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-22 21:24 ` Bart Van Assche
2015-09-22 21:24 ` Bart Van Assche
2015-09-17 9:42 ` [PATCH v1 03/24] IB/mlx5: Support the new memory registration API Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-22 21:27 ` Bart Van Assche
2015-09-22 21:27 ` Bart Van Assche
2015-09-24 7:39 ` Sagi Grimberg
2015-09-24 7:39 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 04/24] IB/mlx4: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 05/24] RDMA/ocrdma: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 06/24] RDMA/cxgb3: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 07/24] iw_cxgb4: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 08/24] IB/qib: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 09/24] RDMA/nes: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 10/24] IB/iser: Port to new fast " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 11/24] iser-target: Port to new memory " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 12/24] xprtrdma: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 13/24] svcrdma: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 14/24] RDS/IW: Convert " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 15/24] IB/srp: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-22 21:58 ` Bart Van Assche
2015-09-22 21:58 ` Bart Van Assche
2015-09-24 9:06 ` Sagi Grimberg
2015-09-24 9:06 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 16/24] IB/mlx5: Remove old FRWR API support Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 17/24] IB/mlx4: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 18/24] RDMA/ocrdma: Remove old FRWR API Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 19/24] RDMA/cxgb3: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 20/24] iw_cxgb4: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 21/24] IB/qib: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 22/24] RDMA/nes: " Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 23/24] IB/hfi1: Remove Old fast registraion API support Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-17 9:42 ` [PATCH v1 24/24] IB/core: Remove old fast registration API Sagi Grimberg
2015-09-17 9:42 ` Sagi Grimberg
2015-09-19 22:45 ` [PATCH v1 00/24] New " Christoph Hellwig
2015-09-19 22:45 ` Christoph Hellwig
2015-09-24 6:53 ` Sagi Grimberg
2015-09-24 6:53 ` Sagi Grimberg
2015-09-24 13:39 ` Christoph Hellwig
2015-09-24 13:39 ` Christoph Hellwig
2015-09-19 23:20 ` santosh.shilimkar
2015-09-19 23:20 ` santosh.shilimkar-QHcLZuEGTsvQT0dZR+AlfA
2015-09-20 9:36 ` Sagi Grimberg
2015-09-20 9:36 ` Sagi Grimberg
2015-09-21 23:28 ` santosh.shilimkar
2015-09-21 23:28 ` santosh.shilimkar-QHcLZuEGTsvQT0dZR+AlfA
2015-09-22 7:19 ` Sagi Grimberg [this message]
2015-09-22 7:19 ` Sagi Grimberg
2015-09-22 7:56 ` Sagi Grimberg
2015-09-22 7:56 ` Sagi Grimberg
2015-09-22 18:23 ` santosh shilimkar
2015-09-22 18:23 ` santosh shilimkar
2015-09-22 21:22 ` Bart Van Assche
2015-09-22 21:22 ` Bart Van Assche
2015-09-24 7:40 ` Sagi Grimberg
2015-09-24 7:40 ` Sagi Grimberg
2015-09-29 19:03 ` Bart Van Assche
2015-09-29 19:03 ` Bart Van Assche
2015-09-29 20:58 ` Sagi Grimberg
2015-09-29 20:58 ` Sagi Grimberg
2015-09-30 6:47 ` Sagi Grimberg
2015-09-30 6:47 ` Sagi Grimberg
2015-09-30 18:59 ` Bart Van Assche
2015-09-30 18:59 ` Bart Van Assche
2015-09-30 20:15 ` Bart Van Assche
2015-09-30 20:15 ` Bart Van Assche
2015-10-01 7:16 ` Sagi Grimberg
2015-10-01 7:16 ` Sagi Grimberg
2015-10-01 17:53 ` Bart Van Assche
2015-10-01 17:53 ` Bart Van Assche
2015-10-01 20:58 ` Bart Van Assche
2015-10-01 20:58 ` Bart Van Assche
[not found] ` <35618B90-4D6E-4036-A69B-4405F020D440@dev.mellanox.co.il>
2015-10-02 15:37 ` Bart Van Assche
2015-10-02 15:37 ` Bart Van Assche
2015-10-06 8:37 ` Sagi Grimberg
2015-10-06 8:37 ` Sagi Grimberg
2015-10-06 18:49 ` Bart Van Assche
2015-10-06 18:49 ` Bart Van Assche
2015-10-07 6:42 ` Sagi Grimberg
2015-10-07 6:42 ` Sagi Grimberg
2015-10-07 15:46 ` Bart Van Assche
2015-10-07 15:46 ` Bart Van Assche
2015-10-07 15:48 ` Sagi Grimberg
2015-10-07 15:48 ` Sagi Grimberg
2015-10-07 9:20 ` Christoph Hellwig
2015-10-07 9:20 ` Christoph Hellwig
2015-10-07 9:25 ` Sagi Grimberg
2015-10-07 9:25 ` Sagi Grimberg
2015-10-07 9:36 ` Christoph Hellwig
2015-10-07 9:36 ` Christoph Hellwig
2015-10-07 10:00 ` Sagi Grimberg
2015-10-07 10:00 ` Sagi Grimberg
2015-10-07 16:30 ` Bart Van Assche
2015-10-07 16:30 ` Bart Van Assche
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=560100EA.8080302@mellanox.com \
--to=sagig@mellanox.com \
--cc=linux-nfs@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=nab@linux-iscsi.org \
--cc=sagig@dev.mellanox.co.il \
--cc=santosh.shilimkar@oracle.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.