Re: [openib-general] RDMA_CM_EVENT_UNREACHABLE(-ETIMEDOUT)
Sean Hefty wrote: Or Gerlitz wrote: Conceptually, do we agree that it would be better not to expose IB reject code to the CMA consumers? that is in the spirit of the CMA being a framework for doing connection management in RDMA transport independent fashion, etc. My concern is that I do not want to mask the real reason for the reject in a way that prevents the user from understanding what's needed to establish the connection. A different way to view this is that the event provides the generic information, and the status detailed. So you are fine with the CMA consumer being aware to the RDMA transport up to the extent of having a per transport reject codes handler? The CMA does return **errno** values on the status field for some events (eg with UNREACHABLE event as of REQ/REP timeout, as in the case that started this thread...), so we need to decide a clearer approach here. We can provide two status values with an event, one that maps to an errno, and another that maps to a transport specific reason. This sounds much better. Or. ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [openfabrics-ewg] OFED 1.1 planning meeting - summary
We received our DDN equipment today and have started setting it up. Scott Weitzenkamp SQA and Release Manager Server Virtualization Business Unit Cisco Systems -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Shawn Hansen (shahanse) Sent: Tuesday, July 25, 2006 1:39 PM To: Tziporet Koren; Matt Leininger Cc: [EMAIL PROTECTED]; openib Subject: Re: [openib-general] [openfabrics-ewg] OFED 1.1 planning meeting - summary Yes, Cisco plans to test OFED on a DDN SRP target. --Shawn -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Tziporet Koren Sent: Tuesday, July 25, 2006 8:40 AM To: Matt Leininger Cc: [EMAIL PROTECTED]; openib Subject: Re: [openfabrics-ewg] OFED 1.1 planning meeting - summary Matt Leininger wrote: 5. SRP: - GA quality - DM (Device Mapper) - for high availability - Basic failover/failback testing with daemon+srp+XVM/MPP and Engenio target Tziporet, Are there any plans to test with the DDN SRP target? Several DoE sites are testing/using the DDN IB based storage. Mellanox does not have DDN SRP target. We will be happy to test it of DDN will loan us a system. Another option is that DDN will take OFED 1.1 RCs and test it in their labs. Can you approach them and ask this. If yes then I can cc them on the RCs mails so they can do it. Is there any other vendor who has DDN SRP target, and going to test OFED with it? Tziporet ___ openfabrics-ewg mailing list [EMAIL PROTECTED] http://openib.org/mailman/listinfo/openfabrics-ewg ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
[openib-general] [PATCH] mthca: make IB_SEND_FENCE work
IB/mthca: make IB_SEND_FENCE work
Fence bit must be set in the doorbell, not only in WQE
Signed-off-by: Michael S. Tsirkin [EMAIL PROTECTED]
diff --git a/drivers/infiniband/hw/mthca/mthca_doorbell.h
b/drivers/infiniband/hw/mthca/mthca_doorbell.h
index dd9a44d..e5f0ad6 100644
--- a/drivers/infiniband/hw/mthca/mthca_doorbell.h
+++ b/drivers/infiniband/hw/mthca/mthca_doorbell.h
@@ -42,6 +42,10 @@ #define MTHCA_RECEIVE_DOORBELL 0x18
#define MTHCA_CQ_DOORBELL 0x20
#define MTHCA_EQ_DOORBELL 0x28
+enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
+
#if BITS_PER_LONG == 64
/*
* Assume that we can just write a 64-bit doorbell atomically. s390
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c
b/drivers/infiniband/hw/mthca/mthca_qp.c
index cd8b672..6efba4a 100644
--- a/drivers/infiniband/hw/mthca/mthca_qp.c
+++ b/drivers/infiniband/hw/mthca/mthca_qp.c
@@ -1502,7 +1502,7 @@ int mthca_tavor_post_send(struct ib_qp *
int i;
int size;
int size0 = 0;
- u32 f0 = 0;
+ u32 f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
int ind;
u8 op0 = 0;
@@ -1843,7 +1843,7 @@ int mthca_arbel_post_send(struct ib_qp *
int i;
int size;
int size0 = 0;
- u32 f0 = 0;
+ u32 f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
int ind;
u8 op0 = 0;
@@ -1864,6 +1864,7 @@ int mthca_arbel_post_send(struct ib_qp *
qp-sq.head += MTHCA_ARBEL_MAX_WQES_PER_SEND_DB;
size0 = 0;
+ f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
/*
* Make sure that descriptors are written before
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
[openib-general] libmthca: make fence work
libmthca: fence bit must be set in doorbell, not only in WQE
Signed-off-by: Michael S. Tsirkin [EMAIL PROTECTED]
Index: openib/src/userspace/libmthca/src/qp.c
===
--- openib/src/userspace/libmthca/src/qp.c (revision 8841)
+++ openib/src/userspace/libmthca/src/qp.c (working copy)
@@ -106,7 +106,8 @@ int mthca_tavor_post_send(struct ibv_qp
int ret = 0;
int size, size0 = 0;
int i;
- uint32_t f0 = 0, op0 = 0;
+ uint32_t f0 = (wr-send_flags IBV_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
+ uint32_t op0 = 0;
pthread_spin_lock(qp-sq.lock);
@@ -436,7 +437,8 @@ int mthca_arbel_post_send(struct ibv_qp
int ret = 0;
int size, size0 = 0;
int i;
- uint32_t f0 = 0, op0 = 0;
+ uint32_t f0 = (wr-send_flags IBV_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
+ uint32_t op0 = 0;
pthread_spin_lock(qp-sq.lock);
@@ -469,6 +471,7 @@ int mthca_arbel_post_send(struct ibv_qp
mthca_write64(doorbell, to_mctx(ibqp-context),
MTHCA_SEND_DOORBELL);
size0 = 0;
+ f0 = (wr-send_flags IBV_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
}
if (wq_overflow(qp-sq, nreq, to_mcq(qp-ibv_qp.send_cq))) {
Index: openib/src/userspace/libmthca/src/wqe.h
===
--- openib/src/userspace/libmthca/src/wqe.h (revision 8841)
+++ openib/src/userspace/libmthca/src/wqe.h (working copy)
@@ -42,6 +42,10 @@ enum {
};
enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
+
+enum {
MTHCA_NEXT_DBD = 1 7,
MTHCA_NEXT_FENCE = 1 6,
MTHCA_NEXT_CQ_UPDATE = 1 3,
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
[openib-general] \x97T\x95\x9F\x82\xC8\x8A\x84\x90\xD8\x8A\xF3\x96]\x90l\x8D\xC8\x83p\x81[\x83g\x83i\x81[\x8F\xD0\x89
\x82\xA8\x90\xA2\x98b\x82\xC9\x82\xC8\x82\xC1\x82\xC4\x82\xA2\x82\xDC\x82\xB7m(_
_)m
\x92S\x93\x96\x82\xCC\x8AF\x8C\xB4\x82\xC5\x82\xB7\x81B
\x96{\x93\xFA\x82\xE0\x97D\x97\xC7VIP\x8F\x97\x90\xAB\x82\xAA\x91\xBD\x90\x94\x82\xB2\x93o\x98^\x82\xC9\x82\xC8\x82\xE8\x82\xDC\x82\xB5\x82\xBD\x82\xCC\x82\xC5\x82\xA8\x92m\x82\xE7\x82\xB9\x92v\x82\xB5\x82\xDC\x82\xB5\x82\xBD\x81B
\x82\xDC\x82\xBD\x81A\x8Ae\x8F\x97\x90\xAB\x82\xCC\x95\xFB\x82\xA9\x82\xE7\x82\xCC\x8B\xA6\x97\xCD\x97\xBF\x8B\xE0\x82\xCD\x90\xE6\x82\xC9\x82\xC5\x82\xE0\x82\xA8\x93n\x82\xB5\x8Fo\x97\x88\x82\xDC\x82\xB7\x81B
\x91S\x82\xC4\x82\xCC\x8F\x97\x90\xAB\x82\xCD\x90\xE6\x92\x85\x8F\x87\x82\xC9\x82\xC8\x82\xC1\x82\xC4\x82\xA2\x82\xDC\x82\xB7\x82\xCC\x82\xC5\x81A\x82\xA8\x91\x81\x82\xDF\x82\xCC\x82\xB2\x98A\x97\x8D\x82\xF0\x82\xA8\x91\xD2\x82\xBF\x82\xB5\x82\xC4\x82\xA2\x82\xDC\x82\xB7\x81B
http://dsdsyyydsds.h1x.com:112/hito7/
\x93\x96\x91R\x96\xB3\x97\xBF\x8F\xD0\x89\xEE\x81\x95\x83\x81\x81[\x83\x8B\x82\xCC\x82\xE2\x82\xE8\x8E\xE6\x82\xE8\x82\xE0\x96\xB3\x97\xBF\x81I\x8A\xAE\x91S\x96\xB3\x97\xBF\x82\xC5\x97V\x82\xD7\x82\xE9\x81\xF4
\x83t\x83\x8A\x81[\x83\x81\x81[\x83\x8B\x82\xC5\x82n\x82j\x82\xC5\x82\xB7\x81I
\x84\xAC\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAD
[EMAIL PROTECTED]
[EMAIL
PROTECTED]@\x8D\xA1\x98A\x97\x8D\x91\xD2\x82\xBF\x82\xCC\x8F\x97\x90\xAB\x82\xCD\x82\xB1\x82\xBF\x82\xE7\x82\xCC\x95\xFB\x82\xC6\x82\xC8\x82\xC1\x82\xC4\x82\xDC\x82\xB7\x81I
[EMAIL
PROTECTED]@\x81\x99\x96\xBC\x91O\x81F\x97\x9D\x8Cb\x82\xB3\x82\xF1\x81i\x82R\x82T\x8D\xCE\x81j
[EMAIL PROTECTED]@\x81\x99\x8E\xFB\x93\xFC\x81F\x8C\x8E\x82V\x82O\x96\x9C\x89~
\x81@ [EMAIL PROTECTED]@
\x81@
\x81\xA6\x82\xB1\x82\xCC\x93x\x81A\x8F\xE3\x8BLVIP\x8F\x97\x90\xAB\x82\xB3\x82\xDC\x82\xA9\x82\xE7\x82\xCD\x8Am\x94F\x82\xF0\x8E\xE6\x82\xE7\x82\xB9\x82\xC4\x92\xB8\x82\xAB\x91\xA6\x93\xFA\x82\xA8\x90U\x8D\x9E\x82\xDD\x82\xB3\x82\xB9\x82\xC4\x92\xB8\x82\xAB\x82\xDC\x82\xB7\x81B
[EMAIL PROTECTED]@http://dsdsyyydsds.h1x.com:112/hito7/
\x84\xAF\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAA\x81\\x84\xAE
\x82\xDC\x82\xBD\x91\xBD\x90\x94VIP\x8F\x97\x90\xAB\x82\xAA\x8BM\x95\xFB\x97l\x82\xF0\x82\xB8\x82\xC1\x82\xC6\x82\xA8\x91\xD2\x82\xBF\x82\xC9\x82\xC8\x82\xC1\x82\xC4\x82\xA8\x82\xE8\x82\xDC\x82\xB7\x82\xCC\x82\xC5\x8BX\x82\xB5\x82\xAD\x82\xA8\x8A\xE8\x82\xA2\x92v\x82\xB5\x82\xDC\x82\xB7\x81B
\x90\xE6\x92\x85\x8F\x87\x82\xC9\x82\xC8\x82\xC1\x82\xC4\x82\xA2\x82\xDC\x82\xB7\x82\xCC\x82\xC5\x81A\x8F\x97\x90\xAB\x82\xCD\x90\x8F\x8E\x9E\x8DX\x90V\x82\xC6\x82\xC8\x82\xE8\x82\xDC\x82\xB7\x81B
[EMAIL PROTECTED]
\x81\xA6\x8Dw\x93\xC7\x89\xF0\x8F\x9C\x82\xCC\x95\xFB\x82\xCD
http://dsdsyyydsds.h1x.com:112/hito7/
\x82\xCC\x8D\xC5\x89\xBA\x92i\x82\xA9\x82\xE7\x82\xB2\x93o\x98^\x82\xAD\x82\xBE\x82\xB3\x82\xA2
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH] mthca: make IB_SEND_FENCE work
On Wed, Aug 09, 2006 at 10:04:35AM +0300, Michael S. Tsirkin wrote:
diff --git a/drivers/infiniband/hw/mthca/mthca_doorbell.h
b/drivers/infiniband/hw/mthca/mthca_doorbell.h
index dd9a44d..e5f0ad6 100644
--- a/drivers/infiniband/hw/mthca/mthca_doorbell.h
+++ b/drivers/infiniband/hw/mthca/mthca_doorbell.h
@@ -42,6 +42,10 @@ #define MTHCA_RECEIVE_DOORBELL 0x18
#define MTHCA_CQ_DOORBELL 0x20
#define MTHCA_EQ_DOORBELL 0x28
+enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
Does anonymous enum have any benefit over define for this?
Cheers,
Muli
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH] mthca: make IB_SEND_FENCE work
Quoting r. Muli Ben-Yehuda [EMAIL PROTECTED]:
+enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
Does anonymous enum have any benefit over define for this?
Not really, no.
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH] mthca: make IB_SEND_FENCE work
On Wed, Aug 09, 2006 at 10:37:20AM +0300, Michael S. Tsirkin wrote:
Quoting r. Muli Ben-Yehuda [EMAIL PROTECTED]:
+enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
Does anonymous enum have any benefit over define for this?
Not really, no.
Good, because I think #define is the idiomatic way to do this...
Cheers,
Muli
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH] mthca: make IB_SEND_FENCE work
Quoting r. Muli Ben-Yehuda [EMAIL PROTECTED]:
Subject: Re: [PATCH] mthca: make IB_SEND_FENCE work
On Wed, Aug 09, 2006 at 10:37:20AM +0300, Michael S. Tsirkin wrote:
Quoting r. Muli Ben-Yehuda [EMAIL PROTECTED]:
+enum {
+ MTHCA_SEND_DOORBELL_FENCE = 1 5
+};
Does anonymous enum have any benefit over define for this?
Not really, no.
Good, because I think #define is the idiomatic way to do this...
Fine with me too.
--
IB/mthca: make IB_SEND_FENCE work
Fence bit must be set in the doorbell, not only in WQE
Signed-off-by: Michael S. Tsirkin [EMAIL PROTECTED]
diff --git a/drivers/infiniband/hw/mthca/mthca_doorbell.h
b/drivers/infiniband/hw/mthca/mthca_doorbell.h
index dd9a44d..e5f0ad6 100644
--- a/drivers/infiniband/hw/mthca/mthca_doorbell.h
+++ b/drivers/infiniband/hw/mthca/mthca_doorbell.h
@@ -42,6 +42,8 @@ #define MTHCA_RECEIVE_DOORBELL 0x18
#define MTHCA_CQ_DOORBELL 0x20
#define MTHCA_EQ_DOORBELL 0x28
+#define MTHCA_SEND_DOORBELL_FENCE (1 5)
+
#if BITS_PER_LONG == 64
/*
* Assume that we can just write a 64-bit doorbell atomically. s390
diff --git a/drivers/infiniband/hw/mthca/mthca_qp.c
b/drivers/infiniband/hw/mthca/mthca_qp.c
index cd8b672..6efba4a 100644
--- a/drivers/infiniband/hw/mthca/mthca_qp.c
+++ b/drivers/infiniband/hw/mthca/mthca_qp.c
@@ -1502,7 +1502,7 @@ int mthca_tavor_post_send(struct ib_qp *
int i;
int size;
int size0 = 0;
- u32 f0 = 0;
+ u32 f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
int ind;
u8 op0 = 0;
@@ -1843,7 +1843,7 @@ int mthca_arbel_post_send(struct ib_qp *
int i;
int size;
int size0 = 0;
- u32 f0 = 0;
+ u32 f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
int ind;
u8 op0 = 0;
@@ -1864,6 +1864,7 @@ int mthca_arbel_post_send(struct ib_qp *
qp-sq.head += MTHCA_ARBEL_MAX_WQES_PER_SEND_DB;
size0 = 0;
+ f0 = unlikely(wr-send_flags IB_SEND_FENCE) ?
MTHCA_SEND_DOORBELL_FENCE : 0;
/*
* Make sure that descriptors are written before
--
MST
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH v3 0/6] Tranport Neutral Verbs Proposal.
Hi James, Sorry for the late response, my system was down and I just got it fixed. Is there a benefit to having rdmav_create_qp() take generic parameters if the application needs to understand the type of QP (IB, iWARP, etc.) created and the transport specific communication manager calls that are needed to manipulate it? Would it make more sense if the QP create command was also transport specific? My opinion is that the create_qp taking generic parameters is correct, only subsequent calls may need to use transport specific calls/arguments. Infact rdma_create_qp uses the ibv_create_qp (now changed to rdmav_create_qp) call internally. PS : What is the opinion on this patchset ? Thanks, - KK ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [openfabrics-ewg] OFED 1.1-rc1 is available
Hi, Tziporet, On 8/8/06, Tziporet Koren [EMAIL PROTECTED] wrote: o iSER: - Stability - Testing more platforms (e.g. ppc64 and ia64) - Performance improvements Only number two above is in the scope of OFED from our perspective, so we prefer to have it listed alone. 2. iSER support in install script for SLES 10 is missing We have a fix for that and it will be part of RC2 -- Moni ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] OFED 1.0 - Official Release (Tziporet Koren)
Dear Sir, Can I test the new release fixed for large SDP connections? tks zhu --- Tziporet Koren [EMAIL PROTECTED] wrote: zhu shi song wrote: Dear Sir, what's your progress on sdp connection? I'm waiting for the new release to test. tks zhu Progress is very good, and we succeeded to run Polygraph with 800 connections for few days. RC1 is expected this week so you will be able to test it yourself Tziporet __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] OFED 1.0 - Official Release (Tziporet Koren)
Things work well on 64 bit systems. Quoting r. zhu shi song [EMAIL PROTECTED]: Subject: Re: OFED 1.0 - Official Release (Tziporet Koren) Dear Sir, Can I test the new release fixed for large SDP connections? tks zhu --- Tziporet Koren [EMAIL PROTECTED] wrote: zhu shi song wrote: Dear Sir, what's your progress on sdp connection? I'm waiting for the new release to test. tks zhu Progress is very good, and we succeeded to run Polygraph with 800 connections for few days. RC1 is expected this week so you will be able to test it yourself Tziporet __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general -- MST ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [patch 02/45] IB/mthca: restore missing PCI registers after reset
On Tue, Aug 08, 2006 at 02:39:37AM +0300, Michael S. Tsirkin wrote: Quoting r. Michael S. Tsirkin [EMAIL PROTECTED]: Subject: Re: [patch 02/45] IB/mthca: restore missing PCI registers after reset Quoting r. Adrian Bunk [EMAIL PROTECTED]: Thanks for this information, I've applied it. BTW, is there a git tree to see what you are cooking? Never mind, I found linux/kernel/git/stable/linux-2.6.16.y.git. It says owner Greg Kroah-Hartman which is what confused me. Thanks for the note, I've fixed this. MST cu Adrian -- Is there not promise of rain? Ling Tan asked suddenly out of the darkness. There had been need of rain for many days. Only a promise, Lao Er said. Pearl S. Buck - Dragon Seed ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [PATCH] libibumad: nit on short mad read
Hi Eitan,
On Tue, 2006-08-01 at 06:28, Eitan Zahavi wrote:
Hi Hal
This was reported to me by Ishai R.
Consider function umad_recv line 810:
if ((n = read(port-dev_fd, umad, sizeof *mad + *length)) =
sizeof *mad + *length) {
DEBUG(mad received by agent %d length %d, mad-agent_id, n);
*length = n - sizeof *mad;
return mad-agent_id;
}
if (n == -EWOULDBLOCK) {
if (!errno)
errno = EWOULDBLOCK;
return n;
}
Seems that umad.c umad_recv would never go through the second if
as if the read return n 0 it will be cought by the first if.
Then I have noticed that a wrap around of the returned length is also
possible.
The patch fixes these issue.
Eitan
Good catches. Thanks. Applied to both trunk and 1.1 branch.
-- Hal
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
[openib-general] howto get copy of https://openib.org/svn repository
Hi, Maybe somebody know. How is it possible to download exact copy (not SVK mirrors) of https://openib.org/svn repository? I like to test some svn to git converter, but would prefer to not overload the network. Thanks, Sasha ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] return error when rdma_listen fails
Pete Wyckoff wrote: Calling rdma_listen() on a cm_id bound to INADDR_ANY can fail, e.g. with EADDRINUSE, but report no error back to the user. This patch fixes that by propagating the error. Success occurs only if at least one of the possibly multiple devices in the system was able to listen. In the case of multiple devices reporting errors on listen, only the first error value is returned. iwarp branch. There's a problem if the listen is done before any devices have been added to the system. In this case, the listen should succeed. - Sean ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] [openfabrics-ewg] OFED 1.1 planning meeting - summary
Can we include librdmacm and dapl in the basic installation option? Also, it would be nice to have rdma_ucm and rdma_cm load on boot by default. Thanks, -arlin This is a small change in the OFED scripts. I suggest that if we go for this change we will do it for the HPC install and not for the basic install (which includes only the verbs and IPoIB). If there is no objection from anyone we will go for this change. Tziporet I don't see this change in OFED 1.1 RC1. Please add librdmacm and dapl into the HPC install and make sure rdma_ucma and rdma_cma gets loaded during boot by default in RC2. Thanks, -arlin ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
Re: [openib-general] howto get copy of https://openib.org/svn repository
search for svn-mirror ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
[openib-general] Outstanding RDMA operations
Hi, Some time ago, I thought there were some thoughts about making the outstanding RDMA count to be configurable via a patch to the mthca driver (the 'rdb_per_qp' parameter in mthca_main.c). Just curious if the patch was going to go in at some point. Thanks, Manpreet. ___ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
