On Thu, 2007-01-18 at 18:14 +0200, Michael S. Tsirkin wrote:
Quoting John W. Marland [EMAIL PROTECTED]:
Subject: Re: [PATCH] IB/core - ib_umad can cause address alignment fault on
ia64
Michael S. Tsirkin wrote:
Quoting Ralph Campbell [EMAIL PROTECTED]:
Subject: [PATCH] IB/core
On Thu, 2007-01-18 at 19:50 +0200, Michael S. Tsirkin wrote:
The int - long change is needed because struct ib_umad_packet
includes struct ib_user_mad (which has 4 byte alignment) but
is then cast to struct ib_mad_hdr which has 8 byte alignment.
But I thought it is the data field which is
for load/stores.
This patch fixes the problem by changing the offset on which
struct ib_user_mad is defined within struct ib_umad_packet.
Thanks go to John W. Marland [EMAIL PROTECTED] for finding this.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r b1128b48dc99 drivers/infiniband/core
On Fri, 2006-11-10 at 10:34 -0500, James Lentini wrote:
Ralph,
Thanks for writing this! It will be a big help to ULP authors.
You are welcome.
A few suggestions below:
On Thu, 9 Nov 2006, Ralph Campbell wrote:
This patch adds a new file to the kernel infiniband documentation
Here is the updated documention with changes suggested by
James Lentini [EMAIL PROTECTED].
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r b9d92097f918 Documentation/infiniband/memory_regions.txt
--- /dev/null Thu Jan 01 00:00:00 1970 +
+++ b/Documentation/infiniband
On Fri, 2006-11-10 at 14:58 -0500, James Lentini wrote:
Replies below:
On Fri, 10 Nov 2006, Ralph Campbell wrote:
snip
+ Remote processes should use the same address for 'remote_addr'
+ as the local kernel's address as returned by the mapping functions
+ listed above
On Fri, 2006-11-10 at 17:14 -0500, James Lentini wrote:
On Fri, 10 Nov 2006, Ralph Campbell wrote:
+ ib_reg_phys_mr() This function returns a pointer to struct
ib_mr.
+ It takes an array of device DMA addresses and lengths which are
used
+ to describe
This patch adds a new file to the kernel infiniband documentation
directory to briefly describe how to use memory regions.
Note: I will be on vacation from Nov. 11 through Nov. 26.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r b9d92097f918 Documentation/infiniband/memory_regions.txt
On Sun, 2006-11-05 at 17:22 +0200, Michael S. Tsirkin wrote:
Quoting r. Ralph Campbell [EMAIL PROTECTED]:
diff -r f37bd0e41fec drivers/infiniband/ulp/ipoib/ipoib_ib.c
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c Thu Oct 26 21:44:41
2006 +0700
+++ b/drivers/infiniband/ulp/ipoib
Ralph Campbell wrote:
IB/ipoib - Use the new verbs DMA mapping functions
This patch converts IPoIB to use the new DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec drivers/infiniband/ulp/ipoib/ipoib_ib.c
--- a/drivers
On Fri, 2006-11-03 at 12:54 -0500, James Lentini wrote:
Ralph,
Could you add documentation to Documentation/infiniband/ explainin
when a ULP needs to use the new verbs DMA mapping functions (what
structures/function parameters must be specified using this API)?
Good idea. I will.
Did
On Fri, 2006-11-03 at 12:08 -0600, Steve Wise wrote:
But how?
Wait, work request posting functions actually get a virtual
address and a key, not a dma address.
Work requests posted using ib_post_send/recv are specified using a dma
address obtained using the appropriate Linux
the mappings created by dma_map_single(),
dma_map_sg(), etc.
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec include/rdma/ib_verbs.h
--- a/include/rdma/ib_verbs.h Thu Oct 26 21:44:41 2006 +0700
+++ b/include/rdma/ib_verbs.h Thu Oct 26 16:10:04 2006 -0800
@@ -43,6 +43,8 @@
#include
IB/ipath - Implement new verbs DMA mapping functions
This patch implements the interposing DMA mapping functions to allow
support for IOMMUs and remove the dependence on phys_to_virt().
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec drivers/infiniband/hw/ipath/Makefile
IB/ipoib - Use the new verbs DMA mapping functions
This patch converts IPoIB to use the new DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec drivers/infiniband/ulp/ipoib/ipoib_ib.c
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c Thu
IB/iser - Use the new verbs DMA mapping functions
This patch converts iser to use the new verbs DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec drivers/infiniband/ulp/iser/iser_memory.c
--- a/drivers/infiniband/ulp/iser/iser_memory.c
IB/rds - Use the new verbs DMA mapping functions
This patch converts RDS to use the new DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/ulp/rds/rds_buf.c
IB/sdp - Use the new verbs DMA mapping functions
This patch converts SDP to use the new DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/ulp/sdp/sdp_bcopy.c
IB/srp - Use new verbs IB DMA mapping functions
This patch converts SRP to use the new verbs DMA mapping functions
for kernel verbs consumers.
From: Ralph Campbell [EMAIL PROTECTED]
diff -r f37bd0e41fec drivers/infiniband/ulp/srp/ib_srp.c
--- a/drivers/infiniband/ulp/srp/ib_srp.c Thu Oct
On Fri, 2006-11-03 at 01:44 +0200, Michael S. Tsirkin wrote:
Quoting r. Roland Dreier [EMAIL PROTECTED]:
Subject: Re: [PATCH 1/7] IB/core - Add DMA mapping functions to allow
device drivers to interpose
However, this means that the API must give the HCA the choice of
what to keep
Yes, I am the author.
On Fri, 2006-09-29 at 14:46 -0700, Roland Dreier wrote:
Thanks, applied (I assumed Ralph was the author when merging, please
let me know if that was wrong)
___
openib-general mailing list
openib-general@openib.org
In either case, if we want to support testing switches and HCAs,
we should have a command line option to change the tests
as appropriate for each.
On Wed, 2006-09-20 at 00:39 -0400, Hal Rosenstock wrote:
On Tue, 2006-09-19 at 21:16, Robert Walsh wrote:
Hal Rosenstock wrote:
On Tue,
On Sun, 2006-09-17 at 12:57 +0300, Or Gerlitz wrote:
Ralph Campbell wrote:
Here is my thinking so far:
The driver is passed an LKEY/RKEY plus an address.
For ib_get_dma_mr(), the address is currently from
dma_map_single(), dma_map_page(), or dma_map_sg().
With the ib_dma_*() routines
On Thu, 2006-09-14 at 09:36 +0300, Or Gerlitz wrote:
Ralph Campbell wrote:
On Wed, 2006-09-13 at 12:00 +0300, Or Gerlitz wrote:
Ralph Campbell wrote:
Well, the other parts of the kernel might not need a kernel virtual
address but the ib_ipath driver still does.
So you agree
On Thu, 2006-09-14 at 13:51 +0300, Or Gerlitz wrote:
Ralph Campbell wrote:
+static inline dma_addr_t ib_dma_map_sg(struct ib_device *dev,
+ struct scatterlist *sg, int nents,
+ enum dma_data_direction direction
On Tue, 2006-09-12 at 21:10 -0600, Jason Gunthorpe wrote:
On Tue, Sep 12, 2006 at 05:40:10PM -0700, Ralph Campbell wrote:
The ib_ipath driver needs kernel virtual addresses in order to be able
to copy data to/from the posted work requests since it does not
use HW DMA. It currently relies
On Tue, 2006-09-12 at 20:15 -0700, Roland Dreier wrote:
My current proposal is to provide wrapper routines for the
dma_*() routines which only the IB kernel code would use.
These ib_dma_*() variants would allow a device driver to interpose
on the call and do appropriate code to convert
On Wed, 2006-09-13 at 12:00 +0300, Or Gerlitz wrote:
Ralph Campbell wrote:
Problem:
The IB kernel to IB device driver interface uses dma_map_single()
and dma_map_sg() to allocate device bus addresses for HW DMA.
These bus addresses are passed to the IB device driver via ib_post_send
Problem:
The IB kernel to IB device driver interface uses dma_map_single()
and dma_map_sg() to allocate device bus addresses for HW DMA.
These bus addresses are passed to the IB device driver via ib_post_send()
and ib_post_recv().
The ib_ipath driver needs kernel virtual addresses in order to be
A minor change to fix RPM builds for libipathverbs.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/Makefile.am
===
--- src/userspace/libipathverbs/Makefile.am (revision 9347)
+++ src/userspace
On Wed, 2006-08-23 at 09:47 -0700, Caitlin Bestler wrote:
[EMAIL PROTECTED] wrote:
Quoting r. john t [EMAIL PROTECTED]:
Subject: basic IB doubt
Hi
I have a very basic doubt. Suppose Host A is doing RDMA write (say 8
MB) to Host B. When data is copied into Host B's local
buffer,
On Wed, 2006-08-23 at 12:28 -0500, Tang, Changqing wrote:
Actually, A knows the data is in B's memory when A gets the
completion notice. B can't rely on anything unless A uses the
RDMA write with immediate which puts a completion event in B's CQ.
Ralph:
Can you give a few more words
On Wed, 2006-08-23 at 14:00 -0500, Tang, Changqing wrote:
Thanks for all your replies. So my general question is, why only 4bytes
immediate data can
Generate completion event on B side, Why RDMA-write with any data size
does not generate
A completion event on B side? basic there are the same
On Wed, 2006-08-23 at 14:50 -0700, Roland Dreier wrote:
I applied this, but I'm wondering if this:
+int ipath_resize_cq(struct ibv_cq *ibcq, int cqe)
{
+ struct ipath_cq*cq = to_icq(ibcq);
+ struct ibv_resize_cqcmd;
+ struct ipath_resize_cq_resp
This patch makes libipathverbs backward compatible with old ib_ipath kernel
drivers.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/verbs.c
===
--- src/userspace/libipathverbs/src/verbs.c
I plan to be on vacation starting this evening and back on
Sunday night. If anyone has further comments or questions
about the ipath patches, I will answer when I get back to
work on Monday. I will be around until about 3:30 today
if there are any comments.
On Wed, 2006-08-16 at 13:56 -0700, Roland Dreier wrote:
ralphc A #define won't help the plug-in know what parameters to
ralphc pass, only a function name change will work if the
ralphc semantics change.
I don't follow:
#ifdef OLD_FUNCTION_HAS_NEW_PARAMS
On Wed, 2006-08-16 at 14:18 -0700, Roland Dreier wrote:
Ralph Was I misunderstanding your goal? The above works if you
Ralph are only trying to have one source code which can be
Ralph compiled to work with either version of libibverbs.
Yes, that's right. I just want to be able
On Mon, 2006-08-14 at 12:37 -0700, Roland Dreier wrote:
ralphc This doesn't break compatibility. uverbs_cmd.c
ralphc ib_uverbs_resize_cq() allocates a struct
ralphc ib_uverbs_resize_cq_resp on the stack but only reads the
ralphc first element in. The structure change isn't
posted earlier for review.
This posting should be considered ready for inclusion.
Allow the driver plug-in library to return additional data in the response
from ibv_cmd_resize_cq().
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libibverbs/include/infiniband/driver.h
MTHCA changes to correspond to the libibverbs changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libmthca/src/verbs.c
===
--- src/userspace/libmthca/src/verbs.c (revision 8843)
+++ src/userspace/libmthca
Improve the performance of the userspace verbs ibv_poll_cq(),
ibv_post_recv(), and ibv_post_srq_recv().
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/verbs.c
===
--- src/userspace
Allow the driver plug-in library to return additional data in the response
from ibv_cmd_resize_cq(). Also, allow the user library to pass additional
information to ib_modify_qp() and ib_modify_srq().
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/include/rdma
EHCA changes to correspond to the ib_core module changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/hw/ehca/ehca_qp.c
===
--- src/linux-kernel/infiniband/hw/ehca/ehca_qp.c (revision
MTHCA changes to correspond to the ib_core module changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/hw/mthca/mthca_srq.c
===
--- src/linux-kernel/infiniband/hw/mthca/mthca_srq.c(revision
are backward compatible
with the old libipathverbs.so.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r dcc321d1340a drivers/infiniband/hw/ipath/Makefile
--- a/drivers/infiniband/hw/ipath/Makefile Sun Aug 06 19:00:05 2006 +
+++ b/drivers/infiniband/hw/ipath/Makefile Fri Aug 11
This is the sort of thing that needs to be clearly documented
for the library. We currently don't have a useable set of
documents (some collection of reading the code and the IB spec.).
On Tue, 2006-08-01 at 16:50 -0700, Roland Dreier wrote:
I guess this is pretty uncontroversial, so I'll check
This patch fixes a problem where certain error packets are passed
to the Infiniband layer for processing even though the packet
actually was received with an error.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r 1b8d63e34819 drivers/infiniband/hw/ipath/ipath_driver.c
--- a/drivers
I am still working on a proposal to remove the phys_to_virt() calls
in the ib_ipath driver. In the mean time, this patch allows SRP
to work by fixing the Rkey check and conversion from IB address
to kernel virtual address. It also returns the correct page size
for FMRs.
Signed-off-by: Ralph
never used.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
diff -r 62120593e7d6 drivers/infiniband/hw/ipath/ipath_verbs.c
--- a/drivers/infiniband/hw/ipath/ipath_verbs.c Mon Jul 17 17:36:13 2006 -0700
+++ b/drivers/infiniband/hw/ipath/ipath_verbs.c Mon Jul 17 17:42:52 2006 -0700
@@ -191,10 +191,6
static void i2c_wait_for_writes(struct ipath_devdata *dd)
{
+ mb();
(void)ipath_read_kreg32(dd, dd-ipath_kregs-kr_scratch);
}
That's a bit weird. I wouldn't have expected the compiler to muck around
with a readl().
I never liked this patch. The last time it came up
On Thu, 2006-07-13 at 08:46 +0300, Muli Ben-Yehuda wrote:
On Wed, Jul 12, 2006 at 05:40:13PM -0700, David Miller wrote:
From: Roland Dreier [EMAIL PROTECTED]
Date: Wed, 12 Jul 2006 17:11:26 -0700
A cleaner solution would be to make the dma_ API really use the device
it's passed
On Fri, 2006-07-14 at 15:35 -0700, David Miller wrote:
From: Ralph Campbell [EMAIL PROTECTED]
Date: Fri, 14 Jul 2006 15:27:07 -0700
Note that the current design only supports one IOMMU type in a system.
This could support multiple IOMMU types at the same time.
This is not true
Thanks to all for the pointers and suggestions.
It will probably take me a while to follow up on these
and make another proposal.
On Thu, 2006-07-13 at 09:02 -0700, Roland Dreier wrote:
A cleaner solution would be to make the dma_ API really use the device
it's passed anyway, and allow
I have been looking at how to eliminate the bus_to_virt() and
phys_to_virt() calls used by the ib_ipath driver.
I am looking for suggestions on how to proceed.
The current IB core to IB device driver interface relies
on a kernel module being able to call ib_get_dma_mr() to allocate
a memory
This is intended to be an architecture specific function
so if the CPU does support HW dma to the CPU's L2 cache, the
architecture specific version of memcpy_nc() would not replace
the default definition which maps memcpy_nc() to memcpy().
For CPUs like the vast majority currently available,
The libibverbs.so.2 has a different device plug-in module
intialization convention from libibverbs.so.1. This patch
updates the InfiniPath libipathverbs module to conform
to the new convention.
Signed-off-by Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/ipathverbs.map
This patch adds support for ibv_query_qp() and ibv_query_srq()
to libipathverbs which are new in libibverbs.so.2.
Note that it layers on top of my previous patch.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/ipathverbs.h
This patch just corrects some white space code conventions.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/ipathverbs.h
===
--- src/userspace/libipathverbs/src/ipathverbs.h(old)
+++ src
.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libibverbs/include/infiniband/driver.h
===
--- src/userspace/libibverbs/include/infiniband/driver.h(revision 8021)
+++ src/userspace/libibverbs/include/infiniband
This patch contains the mthca and ehca specific changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libmthca/src/verbs.c
===
--- src/userspace/libmthca/src/verbs.c (revision 8021)
+++ src/userspace
This patch contains the libipathverbs specific changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/verbs.c
===
--- src/userspace/libipathverbs/src/verbs.c (revision 8021)
+++ src
This patch contains the ib_ipath kernel driver specific changes.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/hw/ipath/ipath_qp.c
===
--- src/linux-kernel/infiniband/hw/ipath/ipath_qp.c
succeeds, the old mmap is invalid, then there isn't
an error recovery issue. All user level threads lock before
using the queue address so the change of address is protected.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib
I get a NULL pointer panic when trying to use the current trunk
SVN (rev 8088). I traced it down to ib_uverbs_create_ah()
failing to initialize the ib_uobject.object pointer.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/core/uverbs_cmd.c
-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/ipathverbs.c
===
--- src/userspace/libipathverbs/src/ipathverbs.c(revision 8089)
+++ src/userspace/libipathverbs/src/ipathverbs.c(working copy
The patch I just sent left out a minor change so please
ignore the previous patch and apply this one instead.
(I forgot to include the change to the map file)
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/libipathverbs/src/ipathverbs.c
the trunk to run again
when you sent your request.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org/mailman/listinfo
--
qlogic is definitely going to be the maintainer of that code.
- R.
Thanks. I was just trying to be complete in case anyone was
applying the patch themselves before SVN was updated.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
. This patch adds those generic code changes.
A subsequent patch will add the InfiniPath specific changes.
Note that I didn't include matching changes to ehca since
I don't have HW to test with but I can try to make a patch
that allows it compile if requested to.
Signed-off-by: Ralph Campbell
the QP's receive queue.
- R.
--
Ralph Campbell [EMAIL PROTECTED]
___
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
corruption when responses go
off the end of a buffer.
Or the new kernel driver needs to handle the old way and the new way.
- R.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo
ipath_post_srq_recv(struct ibv_srq *srq, struct ibv_recv_wr *wr,
+ struct ibv_recv_wr **bad_wr);
struct ibv_ah *ipath_create_ah(struct ibv_pd *pd, struct ibv_ah_attr *attr);
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing
ib_grh));
+ wc.wr_id = qp-r_wr_id;
wc.status = IB_WC_SUCCESS;
wc.opcode = IB_WC_RECV;
wc.vendor_err = 0;
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman
There is no need to limit the MTU to 1K on Pathscale HCAs.
This patch fixes the problem.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/mpi/mvapich-gen2/mpid/ch_gen2/ibverbs_const.h
===
--- src/userspace/mpi
This patch fixes a core dump problem if debugging is on and
umad_register() fails.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/management/libibumad/src/umad.c
===
--- src/userspace/management/libibumad/src
I noticed that perfquery wasn't working and tracked it down to
madrpc_init() using the wrong rmpp version.
This patch fixes the problem.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/userspace/management/libibmad/src/rpc.c
The code to display local_link_integrity_errors and
excessive_buffer_overrun_errors in
/sys/class/infiniband/hca/ports/n/counters/
had a bug extracting the 4 bit values.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/core/sysfs.c
, sdp_inet_ioctl() calls
dev_ioctl() which is no longer exported in 2.6.16.
(it's defined in net/core/dev.c).
I'm not sure what the right fix is so I haven't included a patch.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: openib/src/linux-kernel/infiniband/core/cma.c
___
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
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general
The set_current_state() macro should be used instead of setting the
task state directly.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/ulp/sdp/sdp_conn.c
===
--- src/linux-kernel/infiniband/ulp
This patch changes the hardwired MTU limit of 1024 in SDP
into a module parameter so it can be disabled for HCAs
without the RC performance problem.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/ulp/sdp/sdp_actv.c
On Fri, 2006-02-03 at 17:06 -0800, Roland Dreier wrote:
I think both of these places can use __set_current_state().
- R.
Good point. Here is the updated patch.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: src/linux-kernel/infiniband/ulp/sdp/sdp_conn.c
SDP doesn't remove cached SA path records if the device is removed.
Minor nit: flush_workqueue() doesn't need to be called before
destroy_workqueue().
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: sdp_proto.h
switch
statement into a new pingpong.c file...)
- R.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
http://openib.org/mailman/listinfo/openib-general
To unsubscribe, please visit http://openib.org
This patch adds a new -m mtu_size option to the ping pong
programs which have a path MTU parameter to ib_modify_qp().
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: libibverbs/examples/rc_pingpong.c
===
--- libibverbs
This patch fixes a minor misspelling in SDP.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: ulp/sdp/sdp_inet.c
===
--- ulp/sdp/sdp_inet.c (revision 5055)
+++ ulp/sdp/sdp_inet.c (working copy)
@@ -836,7 +836,7
OK. Here is a much simplified patch which fixes the problem of a
directed route SMP with a with beginning or ending LID routed part.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: core/mad.c
===
--- core/mad.c (revision
I tried getting the latest OpenIB tree from
trunk/src/linux-kernel/infiniband with my 2.6.15 kernel.
Everything seems to be working OK except that I can't
establish a SDP connection. I get:
ib_sdp WARN: 0 2100 Path record completion error -101
Any ideas what is happening?
--
Ralph Campbell
from
trunk/src/linux-kernel/infiniband with my 2.6.15 kernel.
Everything seems to be working OK except that I can't
establish a SDP connection. I get:
ib_sdp WARN: 0 2100 Path record completion error -101
Any ideas what is happening?
--
Ralph Campbell [EMAIL PROTECTED
to start.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: libibverbs/examples/rc_pingpong.c
===
--- libibverbs/examples/rc_pingpong.c (revision 5031)
+++ libibverbs/examples/rc_pingpong.c (working copy)
@@ -568,6 +568,12
Zahavi
-Original Message-
From: [EMAIL PROTECTED] [mailto:openib-general-
[EMAIL PROTECTED] On Behalf Of Ralph Campbell
Sent: Tuesday, January 17, 2006 12:00 AM
To: Eitan Zahavi
Cc: openib-general@openib.org
Subject: RE: [openib-general] SDP status?
I can ping via IPoIB
.
On Fri, 2006-01-13 at 09:51 -0800, Sean Hefty wrote:
Ralph Campbell wrote:
struct ib_mad_send_buf * ib_create_send_mad(struct ib_mad_agent *mad_agent,
u32 remote_qpn, u16 pkey_index,
- int rmpp_active
,
- port_priv-port_num))
+ if (!smi_check_local_smp(recv-mad.smp, port_priv-device))
goto out;
}
--
Ralph Campbell [EMAIL PROTECTED]
___
openib-general mailing list
openib-general@openib.org
http://openib.org
still think its at the beginning of the destination LID
routed part and not decrement hop_ptr.
I think there is a similar problem when sending requests
where the initial part of the path is LID routed.
--
Ralph Campbell [EMAIL PROTECTED]
___
openib
I have only done basic testing (i.e., the link comes up and IPoIB still works)
but I think this fixes the problem.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: include/rdma/ib_mad.h
===
--- include/rdma/ib_mad.h
I tested this and it works for me.
On Tue, 2006-01-10 at 17:51, Hal Rosenstock wrote:
Hi Ralph,
On Tue, 2006-01-10 at 17:47, Ralph Campbell wrote:
I understand. Maybe it should be the first active, if none, then the
first UP, and if none, the first !disabled.
Exactly. I think
and it exits.
The following patch changes the port selection so that the first
active port is selected, and if none, the first non-disabled port.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: umad.c
===
--- umad.c (revision
wrote:
Hi Ralph,
On Tue, 2006-01-10 at 13:31, Ralph Campbell wrote:
If opensm is started with no arguments, the default algorithm
for finding a port to bind to will skip ports which are present
but the link is DOWN. If there is only one port in the system,
no port is selected and opensm
Here is a patch to add the pkeys option to smpquery
to display the Pkey table.
Signed-off-by: Ralph Campbell [EMAIL PROTECTED]
Index: diags/src/smpquery.c
===
--- diags/src/smpquery.c(revision 4808)
+++ diags/src/smpquery.c
1 - 100 of 108 matches
Mail list logo
