[ewg] RE: Please make sure your document is updated for OFED 1.3
The version is inside the opensm_release_notes.txt file: @@ -1,4 +1,4 @@ -OpenSM Release Notes 3.1.9 +OpenSM Release Notes 3.1.10 = Version: OpenFabrics Enterprise Distribution (OFED) 1.3 @@ -11,7 +11,7 @@ Date:February 2008 This document describes the contents of the OpenSM OFED 1.3 release. OpenSM is an InfiniBand compliant Subnet Manager and Administration, and runs on top of OpenIB. The OpenSM version for this release -is openib-3.1.9 +is openib-3.1.10 This document includes the following sections: 1 This Overview section (describing new features and software -- Perhaps it would be a good idea to remove the version from the text? Regards, Yevgeny Kliteynik Mellanox Technologies LTD Tel: +972-4-909-7200 ext: 394 Fax: +972-4-959-3245 P.O. Box 586 Yokneam 20692 ISRAEL -Original Message- From: Sasha Khapyorsky [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 26, 2008 3:58 AM To: Yevgeny Kliteynik Cc: Tziporet Koren; ewg@lists.openfabrics.org; Sean Hefty; Davis, Arlin R; Jeremy Brown; Pavel Shamis; Jeff Squyres (jsquyres); Vu Pham; Oren Kladnitsky; Oren Meron; Eli Cohen; Olaf Kirch; Ira Weiny; Moni Shoua; Jim Mott; Ralph Campbell; Jonathan L. Perkins; Johann George; Jack Morgenstein; Vladimir Sokolovsky; Ishai Rabinovitz; Hoang-Nam Nguyen; Glenn Streiff Subject: Re: Please make sure your document is updated for OFED 1.3 On 00:02 Tue 26 Feb , Yevgeny Kliteynik wrote: > > opensm_release_notes.txt - OpenSM version should be 3.1.10. > If you prefer, I can provide a patch. The content is same. Under OFED docs it is w/out version prefix... Sasha ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] maximum size of ibv_post_send
Hi. Matthew Small wrote: Does anyone know what is the maximum size of a transfer using the verbs library? I have a program which sends increasing amounts of data using ibv_post_send and it seems to be having trouble when I try and send > ~2MB. Obviously, I can split this up into several communications, but I am unaware if there is a limit (or is it only limited by the amount of memory you can register?). You can send using InfiniBand up to 2 GB in a message (WR). Every IB port has an attribute: max_msg_sz which define the maximum message size that it supports. The mentioned value can be limited by the amount of physical memory in your host and by the amount of memory that your process can register. Dotan ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Don't miss to see my pic
Hello! I am tired tonight. I am nice girl that would like to chat with you. Email me at [EMAIL PROTECTED] only, because I am using my friend's email to write this. You will see some of my private pics. ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Re: Please make sure your document is updated for OFED 1.3
On 00:02 Tue 26 Feb , Yevgeny Kliteynik wrote: > > opensm_release_notes.txt - OpenSM version should be 3.1.10. > If you prefer, I can provide a patch. The content is same. Under OFED docs it is w/out version prefix... Sasha ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] maximum size of ibv_post_send
Does anyone know what is the maximum size of a transfer using the verbs library? I have a program which sends increasing amounts of data using ibv_post_send and it seems to be having trouble when I try and send > ~2MB. Obviously, I can split this up into several communications, but I am unaware if there is a limit (or is it only limited by the amount of memory you can register?). -Matt ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] RE: Please make sure your document is updated for OFED 1.3
Tziporet, opensm_release_notes.txt - OpenSM version should be 3.1.10. If you prefer, I can provide a patch. Regards, Yevgeny Kliteynik Mellanox Technologies LTD Tel: +972-4-909-7200 ext: 394 Fax: +972-4-959-3245 P.O. Box 586 Yokneam 20692 ISRAEL -Original Message- From: Tziporet Koren Sent: Monday, February 25, 2008 6:03 PM To: ewg@lists.openfabrics.org Cc: Sean Hefty; Davis, Arlin R; Jeremy Brown; Sasha Khapyorsky; Pavel Shamis; Jeff Squyres (jsquyres); Vu Pham; Oren Kladnitsky; Oren Meron; Eli Cohen; Olaf Kirch; Ira Weiny; Moni Shoua; Jim Mott; Ralph Campbell; Jonathan L. Perkins; Johann George; Jack Morgenstein; Yevgeny Kliteynik; Vladimir Sokolovsky; Ishai Rabinovitz; Hoang-Nam Nguyen; Glenn Streiff Subject: Please make sure your document is updated for OFED 1.3 Hi Attached all OFED 1.3 documents Each owner - please make sure the component you own is updated I did my best merging all changes I got but mistakes always happened :-( Thanks Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] ***SPAM*** Re: [ofa-general] Agenda for OFED meeting today
Shirley Ma wrote: Hello Tziporet, >fix 4k MTU with IPoIB (bug 951) According to the bug report: In RC5, IPoIB does not work on the QLogic 7140 SDR card. This needs to be fixed for the release. I didn't see any evidence this is a 4K MTU bug. How 4K MTU patch impact 7140 SDR card? And more inside? The bug was in ipath driver (not IPoIB). Qlogic can provide more data Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] OFED 1.3 RC6 release is available
OFED 1.3 RC6 release is available on http://www.openfabrics.org/builds/ofed-1.3/release/OFED-1.3-rc6.tgz To get BUILD_ID run ofed_info Please report any issues in bugzilla https://bugs.openfabrics.org/ The GA release is expected on February 28 No more code changes should be applied for the GA release unless a new blocker is found. Tziporet & Vlad Release information: Linux Operating Systems: - RedHat EL4 up4: 2.6.9-42.ELsmp - RedHat EL4 up5: 2.6.9-55.ELsmp - RedHat EL4 up6: 2.6.9-67.ELsmp * - RedHat EL5: 2.6.18-8.el5 - RedHat EL5 up1: 2.6.18-53.el5 - Fedora C6:2.6.18-8.fc6* - SLES10: 2.6.16.21-0.8-smp - SLES10 SP1: 2.6.16.46-0.12-smp - SLES10 SP1 up1: 2.6.16.53-0.16-smp - OpenSuSE 10.3:2.6.22-*-* * - kernel.org: 2.6.23 and 2.6.24 * OSes that are partially tested Systems: * x86_64 * x86 * ia64 * ppc64 List of changes from OFED 1.3-RC5: == Fixed 5 bugs (see attachment) Kernel: --- 1. cm: flush workqueue when removing device (from Sean) 2. ipoib: set child MTU as the parent's (bug 927) 3. RDS: widen the RDMA user token to 64bits (from Olaf) 4. hw/nes: Update from Glenn (Neteffect) 5. hw/ipath: reset retry counter if a good response is seen (bug 874) and fix 4k MTU with IPoIB (bug 951) and general fixes from Ralph. User space: --- 1. srp_daemon: Added SRP_DAEMON_ENABLE parameter to the /etc/infiniband/openib.conf. Used to execute srp_daemon in case that HA is not required. 2. libibverbs: Fix minor issues to be consistent with the libibverbs trunk 3. Opensm: rename OpenSM startup script to opensmd and fixing location of generated opensm.init script (bug 939) Note: In the attached tgz file you can find git-diff of all changes. Tasks that should be completed for the GA package: == 1. Review and update all documents ofed-1.3_rc6-rc5.diff.tgz Description: ofed-1.3_rc6-rc5.diff.tgz rc6-fixed-bugs.csv Description: rc6-fixed-bugs.csv ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] Interoperability testing between OFED 1.3 and OFED 1.2.5
Woodruff, Robert J wrote: There was a question in today's meeting about if anyone had tested ipoib between 1.2.5 and 1.3. I just tried this and it seems to work fine, at least on RedHat EL4/rocks cluster. I tested it using Intel MPI over IPoIB and also Intel MPI over uDAPL, both interoperate fine between 1.3 and 1.2.5. Many thanks Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Interoperability testing between OFED 1.3 and OFED 1.2.5
There was a question in today's meeting about if anyone had tested ipoib between 1.2.5 and 1.3. I just tried this and it seems to work fine, at least on RedHat EL4/rocks cluster. I tested it using Intel MPI over IPoIB and also Intel MPI over uDAPL, both interoperate fine between 1.3 and 1.2.5. woody ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] ***SPAM*** Re: [ofa-general] Agenda for OFED meeting today
Hello Tziporet, >fix 4k MTU with IPoIB (bug 951) According to the bug report: In RC5, IPoIB does not work on the QLogic 7140 SDR card. This needs to be fixed for the release. I didn't see any evidence this is a 4K MTU bug. How 4K MTU patch impact 7140 SDR card? And more inside? Thanks Shirley "Tziporet Koren" .co.il>cc Sent by: [EMAIL PROTECTED] general-b Subject [EMAIL PROTECTED] [ofa-general] Agenda for OFED meeting today sts.openf abrics.or g 02/25/08 07:56 AM Agenda for the OFED meeting today on OFED 1.3 GA readiness: 1. We creating RC6 with the latest critical issues fixed (see list of changes at the end of this mail). It will be available later today. To have a release this week this must be the GOLD RC and no more code changes should be allowed. Proposed schedule: RC6 - today Feb 25 GA - Thursday Feb 28 2. Status update - all 3. bugs review (blocker & critical only): 917 cri P1 SLES [EMAIL PROTECTED] ipath build error on ppc64 895 cri P3 Othe [EMAIL PROTECTED] kernel panic while running multiple test on sdp I think we should not hold the release for both bugs 4. Open discussion List of changes since RC5: == Kernel: 1. cm: flush workqueue when removing device (from Sean) 2. ipoib: set child MTU as the parent's (bug 927) 3. RDS: widen the RDMA user token to 64bits (from Olaf) 4. hw/nes: Update from Glenn (Neteffect) 5. hw/ipath: reset retry counter if a good response is seen (bug 874) and fix 4k MTU with IPoIB (bug 951) and general fixes from Ralph. User space: 1. srp_daemon: Added SRP_DAEMON_ENABLE parameter to the /etc/infiniband/openib.conf. Used to execute srp_daemon in case that HA is not required. 2. libibverbs: Fix minor issues to be consistent with the libibverbs trunk 3. Opensm: rename OpenSM startup script to opensmd and fixing location of generated opensm.init script (bug 939) ___ general mailing list [EMAIL PROTECTED] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general <><><>___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Re: [PATCH] opensm: rename OpenSM startup script to opensmd
On 08:05 Mon 25 Feb , Hal Rosenstock wrote: > On Sun, 2008-02-24 at 14:15 +, Sasha Khapyorsky wrote: > > Rename OpenSM startup script /etc/init.d/opensm -> /etc/init.d/opensmd. > > Does any OFED README or build or install script need updating due to the > name change ? Not that I know. In OFED-1.2 script was named with 'd'. Sasha ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Re: [PATCH] opensm: rename OpenSM startup script to opensmd
On Sun, 2008-02-24 at 14:15 +, Sasha Khapyorsky wrote: > Rename OpenSM startup script /etc/init.d/opensm -> /etc/init.d/opensmd. Does any OFED README or build or install script need updating due to the name change ? -- Hal > Signed-off-by: Sasha Khapyorsky <[EMAIL PROTECTED]> [snip...] ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Agenda for OFED meeting today
Agenda for the OFED meeting today on OFED 1.3 GA readiness: 1. We creating RC6 with the latest critical issues fixed (see list of changes at the end of this mail). It will be available later today. To have a release this week this must be the GOLD RC and no more code changes should be allowed. Proposed schedule: RC6 - today Feb 25 GA - Thursday Feb 28 2. Status update - all 3. bugs review (blocker & critical only): 917 cri P1 SLES [EMAIL PROTECTED] ipath build error on ppc64 895 cri P3 Othe [EMAIL PROTECTED] kernel panic while running multiple test on sdp I think we should not hold the release for both bugs 4. Open discussion List of changes since RC5: == Kernel: 1. cm: flush workqueue when removing device (from Sean) 2. ipoib: set child MTU as the parent's (bug 927) 3. RDS: widen the RDMA user token to 64bits (from Olaf) 4. hw/nes: Update from Glenn (Neteffect) 5. hw/ipath: reset retry counter if a good response is seen (bug 874) and fix 4k MTU with IPoIB (bug 951) and general fixes from Ralph. User space: 1. srp_daemon: Added SRP_DAEMON_ENABLE parameter to the /etc/infiniband/openib.conf. Used to execute srp_daemon in case that HA is not required. 2. libibverbs: Fix minor issues to be consistent with the libibverbs trunk 3. Opensm: rename OpenSM startup script to opensmd and fixing location of generated opensm.init script (bug 939) ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Some nuts for your Mac kNOX
Save 80% on \/ista. http://jodiesantillanps.blogspot.com Cheap \/ista for everyone. http://amaliabowdenkt.blogspot.com Dicount offers for ultimate users! ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] Reminder: OFED teleconference today
Friendly reminder: the OFED teleconference is today (Monday, 25 February, 2008). Next few teleconferences: - All are at noon US eastern / 9am US Pacific / 7pm Israel 1. Monday, Feb 25, code 210020028 Dial-in information: US/Canada: +1.866.432.9903 India: +91.80.4103.3979 Israel: +972.9.892.7026 Others: http://cisco.com/en/US/about/doing_business/conferencing/ Tziporet (for Jeff Squyres) ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] Do you plan to support ib-bonding on RHEL-5 up1?
Olga Shern wrote: Tziporet, sorry for the late response. Yes, we support bonding on RH5 UP1 Thanks, I updated ib-bonding doc with this info Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ewg] Do you plan to support ib-bonding on RHEL-5 up1?
Tziporet, sorry for the late response. Yes, we support bonding on RH5 UP1 On 2/18/08, Tziporet Koren <[EMAIL PROTECTED]> wrote: > > Thanks, > Tziporet > ___ > ewg mailing list > ewg@lists.openfabrics.org > http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg > ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] [PATCH 2/2] libibverbs/man: add XRC support
Add XRC (eXtended Reliable Connection) support to all of the relevant man pages and add new man pages to new XRC verbs. Signed-off-by: Dotan Barak <[EMAIL PROTECTED]> --- diff --git a/Makefile.am b/Makefile.am index 45914d3..1126416 100644 --- a/Makefile.am +++ b/Makefile.am @@ -44,14 +44,17 @@ man_MANS = man/ibv_asyncwatch.1 man/ibv_devices.1 man/ibv_devinfo.1 \ man/ibv_alloc_pd.3 man/ibv_attach_mcast.3 man/ibv_create_ah.3 \ man/ibv_create_ah_from_wc.3 man/ibv_create_comp_channel.3 \ man/ibv_create_cq.3 man/ibv_create_qp.3 man/ibv_create_srq.3 \ +man/ibv_create_xrc_rcv_qp.3\ man/ibv_fork_init.3 man/ibv_get_async_event.3 \ man/ibv_get_cq_event.3 man/ibv_get_device_guid.3 \ man/ibv_get_device_list.3 man/ibv_get_device_name.3 \ -man/ibv_modify_qp.3 man/ibv_modify_srq.3 man/ibv_open_device.3 \ +man/ibv_modify_qp.3 man/ibv_modify_srq.3 man/ibv_modify_xrc_rcv_qp.3 \ +man/ibv_open_device.3 man/ibv_open_xrc_domain.3\ man/ibv_poll_cq.3 man/ibv_post_recv.3 man/ibv_post_send.3 \ man/ibv_post_srq_recv.3 man/ibv_query_device.3 man/ibv_query_gid.3 \ man/ibv_query_pkey.3 man/ibv_query_port.3 man/ibv_query_qp.3 \ -man/ibv_query_srq.3 man/ibv_rate_to_mult.3 man/ibv_reg_mr.3 \ +man/ibv_query_srq.3 man/ibv_query_xrc_rcv_qp.3 \ +man/ibv_rate_to_mult.3 man/ibv_reg_mr.3 man/ibv_reg_xrc_rcv_qp.3 \ man/ibv_req_notify_cq.3 man/ibv_resize_cq.3 man/verbs.7 DEBIAN = debian/changelog debian/compat debian/control debian/copyright \ @@ -73,6 +76,8 @@ install-data-hook: $(RM) ibv_ack_async_event.3 && \ $(RM) ibv_ack_cq_events.3 && \ $(RM) ibv_close_device.3 && \ + $(RM) ibv_close_xrc_domain.3 && \ + $(RM) ibv_create_xrc_srq.3 && \ $(RM) ibv_dealloc_pd.3 && \ $(RM) ibv_dereg_mr.3 && \ $(RM) ibv_destroy_ah.3 && \ @@ -83,10 +88,13 @@ install-data-hook: $(RM) ibv_detach_mcast.3 && \ $(RM) ibv_free_device_list.3 && \ $(RM) ibv_init_ah_from_wc.3 && \ + $(RM) ibv_unreg_xrc_rcv_qp.3 && \ $(RM) mult_to_ibv_rate.3 && \ $(LN_S) ibv_get_async_event.3 ibv_ack_async_event.3 && \ $(LN_S) ibv_get_cq_event.3 ibv_ack_cq_events.3 && \ $(LN_S) ibv_open_device.3 ibv_close_device.3 && \ + $(LN_S) ibv_open_xrc_domain.3 ibv_close_xrc_domain.3 && \ + $(LN_S) ibv_create_srq.3 ibv_create_xrc_srq.3 && \ $(LN_S) ibv_alloc_pd.3 ibv_dealloc_pd.3 && \ $(LN_S) ibv_reg_mr.3 ibv_dereg_mr.3 && \ $(LN_S) ibv_create_ah.3 ibv_destroy_ah.3 && \ @@ -97,5 +105,6 @@ install-data-hook: $(LN_S) ibv_attach_mcast.3 ibv_detach_mcast.3 && \ $(LN_S) ibv_get_device_list.3 ibv_free_device_list.3 && \ $(LN_S) ibv_create_ah_from_wc.3 ibv_init_ah_from_wc.3 && \ + $(LN_S) ibv_reg_xrc_rcv_qp.3 ibv_unreg_xrc_rcv_qp.3 && \ $(LN_S) ibv_rate_to_mult.3 mult_to_ibv_rate.3 diff --git a/man/ibv_create_qp.3 b/man/ibv_create_qp.3 index abd5449..fb6f041 100644 --- a/man/ibv_create_qp.3 +++ b/man/ibv_create_qp.3 @@ -28,8 +28,9 @@ struct ibv_cq *send_cq;/* CQ to be associated with the Send Que struct ibv_cq *recv_cq;/* CQ to be associated with the Receive Queue (RQ) */ struct ibv_srq *srq;/* SRQ handle if QP is to be associated with an SRQ, otherwise NULL */ struct ibv_qp_cap cap;/* QP capabilities */ -enum ibv_qp_typeqp_type;/* QP Transport Service Type: IBV_QPT_RC, IBV_QPT_UC, or IBV_QPT_UD */ +enum ibv_qp_typeqp_type;/* QP Transport Service Type: IBV_QPT_RC, IBV_QPT_UC, IBV_QPT_UD or IBV_QPT_XRC */ int sq_sig_all; /* If set, each Work Request (WR) submitted to the SQ generates a completion entry */ +struct ibv_xrc_domain *xrc_domain; /* XRC domain the QP will be associated with (valid only for IBV_QPT_XRC QP), otherwise NULL */ .in -8 }; .sp diff --git a/man/ibv_create_srq.3 b/man/ibv_create_srq.3 index ff33736..3fe545d 100644 --- a/man/ibv_create_srq.3 +++ b/man/ibv_create_srq.3 @@ -10,12 +10,26 @@ ibv_create_srq, ibv_destroy_srq \- create or destroy a shared receive queue (SRQ .BI "struct ibv_srq *ibv_create_srq(struct ibv_pd " "*pd" ", struct " .BI " ibv_srq_init_attr " "*srq_init_attr" ); .nl +.BI "struct ibv_srq *ibv_create_xrc_srq(struct ibv_pd " "*pd" ", +.BI " struct ibv_xrc_domain " "*xrc_domain" ", +.BI " struct ibv_cq " "*xrc_cq" ", +.BI " struct ibv_srq_init_attr " "*srq_init_attr" ); +.nl .BI "int ibv_destroy_srq(struct ibv_srq " "*srq" ); .fi .SH "DESCRIPTION" .B ibv_create_srq() creates a shared receive queue (SRQ) associated with th
[ewg] [PATCH 1/2] libibverbs/man: minor fixes to be consistent with libibverbs
Fix minor issues to be consistent with the libibverbs trunk. Signed-off-by: Dotan Barak <[EMAIL PROTECTED]> --- diff --git a/man/ibv_get_cq_event.3 b/man/ibv_get_cq_event.3 index 430ffd0..27dca6f 100644 --- a/man/ibv_get_cq_event.3 +++ b/man/ibv_get_cq_event.3 @@ -20,9 +20,9 @@ waits for the next completion event in the completion event channel .I channel\fR. Fills the arguments .I cq -with the CQ that got the event and +with the CQ that got the event and .I cq_context -with it's context\fR. +with the CQ's context\fR. .PP .B ibv_ack_cq_events() acknowledges diff --git a/man/ibv_post_recv.3 b/man/ibv_post_recv.3 index 8efa9d3..46a630b 100644 --- a/man/ibv_post_recv.3 +++ b/man/ibv_post_recv.3 @@ -60,8 +60,12 @@ and not .B ibv_post_recv()\fR, since the QP's own receive queue will not be used. .PP -If a WR is being posted to a UD QP, the Global Routing Header (GRH) will be placed -in the first 40 bytes of the buffer (whether or not GRH is actually being used by the QP). +If a WR is being posted to a UD QP, the Global Routing Header (GRH) of +the incoming message will be placed in the first 40 bytes of the +buffer(s) in the scatter list. If no GRH is present in the incoming +message, then the first bytes will be undefined. This means that in +all cases, the actual data of the incoming message will start at an +offset of 40 bytes into the buffer(s) in the scatter list. .SH "SEE ALSO" .BR ibv_create_qp (3), .BR ibv_post_send (3), diff --git a/man/ibv_post_srq_recv.3 b/man/ibv_post_srq_recv.3 index 2663a6d..65877cd 100644 --- a/man/ibv_post_srq_recv.3 +++ b/man/ibv_post_srq_recv.3 @@ -52,8 +52,12 @@ The buffers used by a WR can only be safely reused after WR the request is fully executed and a work completion has been retrieved from the corresponding completion queue (CQ). .PP -If a WR is being consumed by a UD QP, the Global Routing Header (GRH) will be placed -in the first 40 bytes of the buffer (whether or not GRH is actually being used by the QP). +If a WR is being posted to a UD QP, the Global Routing Header (GRH) of +the incoming message will be placed in the first 40 bytes of the +buffer(s) in the scatter list. If no GRH is present in the incoming +message, then the first bytes will be undefined. This means that in +all cases, the actual data of the incoming message will start at an +offset of 40 bytes into the buffer(s) in the scatter list. .SH "SEE ALSO" .BR ibv_create_qp (3), .BR ibv_post_send (3), diff --git a/man/ibv_query_pkey.3 b/man/ibv_query_pkey.3 index f6d37c9..fcdfe6c 100644 --- a/man/ibv_query_pkey.3 +++ b/man/ibv_query_pkey.3 @@ -22,7 +22,7 @@ through the pointer .I pkey\fR. .SH "RETURN VALUE" .B ibv_query_pkey() -returns 0 on success, and -1 on error. +returns 0 on success, and \-1 on error. .SH "SEE ALSO" .BR ibv_open_device (3), .BR ibv_query_device (3), ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] What are you up to?
Hello! I am bored tonight. I am nice girl that would like to chat with you. Email me at [EMAIL PROTECTED] only, because I am using my friend's email to write this. Would you mind me showing some nice pictures of me? ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
[ewg] [PATCH] perftest README update
Update perftest README Signed-off-by: Oren Meron <[EMAIL PROTECTED]> --- a/README +++ b/README @@ -1,5 +1,5 @@ Open Fabrics Enterprise Distribution (OFED) -Performance Tests README for OFED 1.2 +Performance Tests README for OFED 1.3 February 2008 @@ -9,7 +9,7 @@ Table of Contents === 1. Overview -2. Notes on Testing Methodology +2. Notes on Testing Method 3. Test Descriptions 4. Running Tests @@ -18,30 +18,30 @@ Table of Contents === This is a collection of tests written over uverbs intended for use as a performance micro-benchmark. As an example, the tests can be used for -HW or SW tuning and/or functional testing. +hardware or software tuning and/or functional testing. -Please post results/observations to the openib-general mailing list. +Please post results and observations to the openib-general mailing list. See "Contact Us" at http://openib.org/mailman/listinfo/openib-general and http://www.openib.org. === -2. Notes on Testing Methodology +2. Notes on Testing Method === -- The benchmark used the CPU cycle counter to get time stamps without context - switch. Some CPU architectures (e.g., Intel's 80486 or older PPC) do NOT - have such capability. +- The benchmark uses the CPU cycle counter to get time stamps without a context + switch. Some CPU architectures (e.g., Intel's 80486 or older PPC) do NOT have + such capability. - The benchmark measures round-trip time but reports half of that as one-way latency. This means that it may not be sufficiently accurate for asymmetrical configurations. -- Min/Median/Max result is reported. - The median (vs average) is less sensitive to extreme scores. - Typically, the "Max" value is the first value measured. +- Min/Median/Max results are reported. + The Median (vs average) is less sensitive to extreme scores. + Typically, the Max value is the first value measured. -- Larger samples help marginally only. The default (1000) is pretty good. - Note that an array of cycles_t (typically unsigned long) is allocated +- Larger samples only help marginally. The default (1000) is very satisfactory. + Note that an array of cycles_t (typically an unsigned long) is allocated once to collect samples and again to store the difference between them. Really big sample sizes (e.g., 1 million) might expose other problems with the program. @@ -53,30 +53,31 @@ http://www.openib.org. Architectures tested: i686, x86_64, ia64 - === -4. Test Descriptions +3. Test Descriptions === rdma_lat.c latency test with RDMA write transactions -rdma_bw.c streaming BW test with RDMA write transactions +rdma_bw.c streaming bandwidth test with RDMA write transactions -The following tests are mainly useful for HW/SW benchmarking. +The following tests are mainly useful for hardware/software benchmarking. They are not intended as actual usage examples. send_lat.c latency test with send transactions -send_bw.c BW test with send transactions +send_bw.c bandwidth test with send transactions write_lat.clatency test with RDMA write transactions -write_bw.c BW test with RDMA write transactions +write_bw.c bandwidth test with RDMA write transactions read_lat.c latency test with RDMA read transactions -read_bw.c BW test with RDMA read transactions +read_bw.c bandwidth test with RDMA read transactions + +The executable name of each test starts with the general prefix "ib_"; +for example, ib_write_lat. -The executable name of each test starts with the general prefix "ib_", -e.g., ib_write_lat. -Running Tests -- +=== +4. Running Tests +=== Prerequisites: kernel 2.6 @@ -101,14 +102,20 @@ Common Options to tests: -s, --size=size of message to exchange (default: 1) -a, --allrun sizes from 2 till 2^23 -t, --tx-depth= size of tx queue (default: 50) - -g, --mcgsend messages to multicast group 0xc001 (only available in send-UD) + -g, --mcgsend messages to multicast group 0xc001 + (only available in send-UD) -n, --iters= number of exchanges (at least 100, default: 1000) - -C, --report-cycles report times in cpu cycle units (default: mi
[ewg] Re: Please sned me pathces to release notes for OFED 1.3
Jeremy Brown wrote: Tziporet, I've attached the new version of the release notes as well as a diff against the old one. Please include this when you get a chance. done Tziporet ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
