From: Roland Dreier <rdreier@cisco.com>
To: Joachim Fenkes <fenkes@de.ibm.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
LinuxPPC-Dev <linuxppc-dev@ozlabs.org>,
Christoph Raisch <raisch@de.ibm.com>,
OF-General <general@lists.openfabrics.org>,
Stefan Roscher <stefan.roscher@de.ibm.com>
Subject: Re: [PATCH 01/10] IB/ehca: Support for multiple event queues
Date: Mon, 16 Jul 2007 09:04:26 -0700 [thread overview]
Message-ID: <adalkdgz65x.fsf@cisco.com> (raw)
In-Reply-To: <200707121746.36763.fenkes@de.ibm.com> (Joachim Fenkes's message of "Thu, 12 Jul 2007 17:46:35 +0200")
> The eHCA driver can now handle multiple event queues (read: interrupt
> sources) instead of one. The number of available EQs is selected via the
> nr_eqs module parameter.
> CQs are either assigned to the EQs based on the comp_vector index or, if the
> dist_eqs module parameter is supplied, using a round-robin scheme.
Do you have any data on how well this round-robin assignment works?
It seems not quite right to me for the driver to advertise nr_eqs
completion vectors, but then if round-robin is turned on to ignore the
consumer's decision about which vector to use.
Maybe if round-robin is turned on you should report 0 as the number of
completion vectors? Or maybe we should allow well-known values for
the completion vector passed to ib_create_cq to allow consumers to
specify a policy (like round robin) instead of a particular vector?
Maybe the whole interface is broken and we should only be exposing
policies to consumers instead of the specific vector?
I think I would rather hold off on multiple EQs for this merge window
and plan on having something really solid and thought-out for 2.6.24.
- R.
WARNING: multiple messages have this Message-ID (diff)
From: Roland Dreier <rdreier@cisco.com>
To: Joachim Fenkes <fenkes@de.ibm.com>
Cc: "LinuxPPC-Dev" <linuxppc-dev@ozlabs.org>,
LKML <linux-kernel@vger.kernel.org>,
"OF-General" <general@lists.openfabrics.org>,
Roland Dreier <rolandd@cisco.com>,
"Hoang-Nam Nguyen" <hnguyen@de.ibm.com>,
Christoph Raisch <raisch@de.ibm.com>,
Stefan Roscher <stefan.roscher@de.ibm.com>
Subject: Re: [PATCH 01/10] IB/ehca: Support for multiple event queues
Date: Mon, 16 Jul 2007 09:04:26 -0700 [thread overview]
Message-ID: <adalkdgz65x.fsf@cisco.com> (raw)
In-Reply-To: <200707121746.36763.fenkes@de.ibm.com> (Joachim Fenkes's message of "Thu, 12 Jul 2007 17:46:35 +0200")
> The eHCA driver can now handle multiple event queues (read: interrupt
> sources) instead of one. The number of available EQs is selected via the
> nr_eqs module parameter.
> CQs are either assigned to the EQs based on the comp_vector index or, if the
> dist_eqs module parameter is supplied, using a round-robin scheme.
Do you have any data on how well this round-robin assignment works?
It seems not quite right to me for the driver to advertise nr_eqs
completion vectors, but then if round-robin is turned on to ignore the
consumer's decision about which vector to use.
Maybe if round-robin is turned on you should report 0 as the number of
completion vectors? Or maybe we should allow well-known values for
the completion vector passed to ib_create_cq to allow consumers to
specify a policy (like round robin) instead of a particular vector?
Maybe the whole interface is broken and we should only be exposing
policies to consumers instead of the specific vector?
I think I would rather hold off on multiple EQs for this merge window
and plan on having something really solid and thought-out for 2.6.24.
- R.
next prev parent reply other threads:[~2007-07-16 16:04 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-07-12 15:45 [PATCH 00/10] IB/ehca: Multiple Event Queues, MR/MW rework, large page MRs, fixes Joachim Fenkes
2007-07-12 15:45 ` Joachim Fenkes
2007-07-12 15:46 ` [PATCH 01/10] IB/ehca: Support for multiple event queues Joachim Fenkes
2007-07-12 15:46 ` Joachim Fenkes
2007-07-16 16:04 ` Roland Dreier [this message]
2007-07-16 16:04 ` Roland Dreier
2007-07-16 20:34 ` Joachim Fenkes
2007-07-16 20:34 ` Joachim Fenkes
2007-07-16 20:37 ` Hoang-Nam Nguyen
2007-07-16 20:37 ` Hoang-Nam Nguyen
2007-07-17 3:48 ` Roland Dreier
2007-07-17 3:48 ` Roland Dreier
2007-07-17 4:37 ` Michael S. Tsirkin
2007-07-17 4:37 ` Michael S. Tsirkin
2007-07-17 17:52 ` Roland Dreier
2007-07-17 17:52 ` Roland Dreier
2007-07-17 21:32 ` Michael S. Tsirkin
2007-07-17 21:32 ` Michael S. Tsirkin
2007-07-19 15:03 ` Hoang-Nam Nguyen
2007-07-19 15:03 ` Hoang-Nam Nguyen
2007-07-17 5:57 ` [ofa-general] " Shirley Ma
2007-07-12 15:47 ` [PATCH 02/10] IB/ehca: Fix HW level autodetection Joachim Fenkes
2007-07-12 15:47 ` Joachim Fenkes
2007-07-12 15:48 ` [PATCH 03/10] IB/ehca: fix memory leak in error path of ehca_get_dma_mr() Joachim Fenkes
2007-07-12 15:48 ` Joachim Fenkes
2007-07-12 15:49 ` [PATCH 04/10] IB/ehca: use common error code mapping instead of specific ones Joachim Fenkes
2007-07-12 15:49 ` Joachim Fenkes
2007-07-16 17:14 ` Roland Dreier
2007-07-16 17:14 ` Roland Dreier
2007-07-16 20:35 ` Joachim Fenkes
2007-07-16 20:35 ` Joachim Fenkes
2007-07-12 15:51 ` [PATCH 05/10] IB/ehca: use #define for "pages per register_rpage" instead of hardcoded value Joachim Fenkes
2007-07-12 15:51 ` Joachim Fenkes
2007-07-12 15:51 ` [PATCH 06/10] IB/ehca: use macro to calculate number of chunks in a mem block Joachim Fenkes
2007-07-12 15:51 ` Joachim Fenkes
2007-07-12 15:52 ` [PATCH 07/10] IB/ehca: MR/MW structure refactoring Joachim Fenkes
2007-07-12 15:52 ` Joachim Fenkes
2007-07-12 15:53 ` [PATCH 08/10] IB/ehca: Restructure ehca_set_pagebuf() Joachim Fenkes
2007-07-12 15:53 ` Joachim Fenkes
2007-07-12 15:53 ` [PATCH 09/10] IB/ehca: Fix warnings issued by checkpatch.pl Joachim Fenkes
2007-07-12 15:53 ` Joachim Fenkes
2007-07-12 15:54 ` [PATCH 10/10] IB/ehca: Support large page MRs Joachim Fenkes
2007-07-12 15:54 ` Joachim Fenkes
2007-07-16 17:37 ` Roland Dreier
2007-07-16 17:37 ` Roland Dreier
2007-07-16 21:11 ` Joachim Fenkes
2007-07-16 21:11 ` Joachim Fenkes
2007-07-17 3:50 ` Roland Dreier
2007-07-17 3:50 ` Roland Dreier
2007-07-17 6:29 ` Joachim Fenkes
2007-07-17 6:29 ` Joachim Fenkes
2007-07-12 17:15 ` [PATCH 00/10] IB/ehca: Multiple Event Queues, MR/MW rework, large page MRs, fixes Roland Dreier
2007-07-12 17:15 ` Roland Dreier
2007-07-13 8:26 ` Joachim Fenkes
2007-07-13 8:26 ` Joachim Fenkes
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=adalkdgz65x.fsf@cisco.com \
--to=rdreier@cisco.com \
--cc=fenkes@de.ibm.com \
--cc=general@lists.openfabrics.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@ozlabs.org \
--cc=raisch@de.ibm.com \
--cc=stefan.roscher@de.ibm.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.