Signed-off-by: Sean Hefty
---
Patch is dependent on proposed kernel changes.
include/rdma/rdma_cma.h |1 +
src/rsocket.c | 23 +++
2 files changed, 24 insertions(+), 0 deletions(-)
diff --git a/include/rdma/rdma_cma.h b/include/rdma/rdma_cma.h
index c0f83b1..
Provide an option for the user to specify that listens should
only accept connections where the incoming address family
matches that of the locally bound address. This is used to
support the equivalent of IPV6_V6ONLY socket option, which
allows an app to only accept connection requests directed
to
The rdma_cm maps IPv4 and IPv6 addresses to the same service
ID. This prevents apps from listening only for IPv4 or IPv6
addresses. It also results in an app binding to an IPv4
address receiving connection requests for an IPv6 address.
Match socket behavior. Restrict listens on IPv4 addresses
t
The rdma cm uses a single port space for all associated (tcp, udp,
etc.) port bindings, regardless of the address family that the
user binds to. The result is that if a user binds to AF_INET,
but does not specify an IP address, the bind will occur for AF_INET6.
This causes an attempt to bind to th
> Yes, good point. If the other side does not have rsockets then it is not that
> straightforward.
>
> Some thoughts:
>
> 1. The best option might be if we exchanged an option during
> connection setup. This tells the peers if the other side is capable of RDMA.
> If it is then one can switch to s
> Traditional sockets based applications wanting high throughput could use
> rsockets Since it is layered on top of uverbs we expected to see good
> throughput numbers.
> So, we started to run netperf and iperf. We observed that it tops off at
> about 20Gb/s with QDR adapters. A quick "perf top" re
> > Match socket behavior. Restrict listens on IPv4 addresses
> > to only IPv4 addresses. If a listen is on an IPv6 address,
> > allow it to receive either IPv4 or IPv6 addresses.
>
> Can you match the IP stack and incorporate /proc/sys/net/ipv6/bindv6only
I'll check
> and socket option IPV6_V
The rdma cm uses a single port space for all associated (tcp, udp,
etc.) port bindings, regardless of the address family that the
user binds to. The result is that if a user binds to AF_INET,
but does not specify an IP address, the bind will occur for AF_INET6.
This causes an attempt to bind to th
Provide an option for the user to specify that listens should
only accept connections where the incoming address family
matches that of the locally bound address. This is used to
support the equivalent of IPV6_V6ONLY socket option, which
allows an app to only accept connection requests directed
to
The rdma_cm maps IPv4 and IPv6 addresses to the same service
ID. This prevents apps from listening only for IPv4 or IPv6
addresses. It also results in an app binding to an IPv4
address receiving connection requests for an IPv6 address.
Match socket behavior. Restrict listens on IPv4 addresses
t
> In a parallel universe, struct ibv_wc would have a bitmap field indicating
> which others fields are valid. In this part of the multiverse, a more
> complete documentation would be welcome. If libmlx4/libmthca behavior is
> the compliant one, I can provide an updated man page.
The best documenta
> Do you have something more specific re how to actually align (say) the
> RSS QP group into the framework used by XRC?
Not really. That was more of a conceptual statement regarding the design. The
operation, particularly on the receive side, just seems similar.
I don't think we should force a
> Though one can consider the fall-back in reverse order i.e. if the
> rdma connection fails continue with the already established connection (over
> the normal inet socket).
When I consider fallback, one of the issues is handling the case where one of
the two sides is not using rsockets. This
> But to map standard networking applications to rsockets we will run into the
> above problem i.e. fork() will not work. It would be very useful to
> allow for the standard networking paradigm of: bind()->listen()->accept()-
> ->fork(), and then the server goes back to accept(). That would allow
> Unfortunately this introduces another issue.
> This causes netperf to hang in recv() after shutdown(SHUT_WR)
> on the data socket.
Bah. Thanks for testing. I was hoping for a simple work-around, rather than
expand rsocket states. Oh well. Let me figure out the proper way to handle
partial
Sridhar Samudrala reported an error (EOPNOTSUPP)
after calling select().
The issue is that rshutdown(SHUT_WR) was called before select().
As part of shutdown, rsockets switches the underlying fd from
nonblocking to blocking to ensure that previously sent data has
completed. shutdown(SHUT_WR) ind
Give an administrator control over the default settings
used by rsockets. Use files under %sysconfig%/rdma/rsocket as shown:
mem_default - default size of receive buffer(s)
wmem_default - default size of send buffer(s)
sqsize_default - default size of send queue
rqsize_default - default size of r
The latency cost of blocking is significant compared to round
trip ping-pong time. Spin briefly on rsockets before calling
into the kernel and blocking.
The time to spin before blocking is read from an rsocket
configuration file %sysconfig%/rdma/rsocket/polling_time. This
is user adjustable.
As
netperf uses the TCP_MAXSEG socket option. Add support for it.
Problem reported by Sridhar Samudrala
getsockopt returns the path MTU as the TCP_MAXSEG. setsockopt
currently ignores the value.
Signed-off-by: Sean Hefty
---
src/rsocket.c |9 +
1 files changed, 9 insertions(+), 0 de
Problem and cause reported by Hal Rosenstock
Signed-off-by: Sean Hefty
---
src/acme.c | 14 +-
1 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/src/acme.c b/src/acme.c
index e6ae188..0e1d4ed 100644
--- a/src/acme.c
+++ b/src/acme.c
@@ -495,7 +495,8 @@ static int reso
If the user specifies a DLID or DGID as part of a path record lookup,
automatically select a local port. This allows a user to query an SA
without needing to specify the local SLID or SGID.
Signed-off-by: Sean Hefty
---
src/acm.c |6 +-
1 files changed, 5 insertions(+), 1 deletions(-)
> Still, the 1st and most thing to handle here is feedback on the QP
> groups concept suggested by this patch set to support TSS/RSS over
> verbs. The plan is for this concept to (with little help from a
> framework for verbs extension) apply to user space RSS as well,
> for both UD and RAW QPs.
C
> In my test case to evaluate fetch-and-add, I spawn multiple threads, each
> owning its own QP inside the same PD and context and sending fetch-and-add
> requests without any inner contention (no lock, etc.). I quickly reach a
> ceiling of about 900KOPS with 5/6 threads, and I have a hard time fig
thanks! - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> diff --git a/src/acme.c b/src/acme.c
> index d3f8174..533588c 100644
> --- a/src/acme.c
> +++ b/src/acme.c
> @@ -618,12 +618,18 @@ static void resolve(char *svc)
> ret = resolve_name(&path);
> break;
> case 'l':
> -
thanks! - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Thanks for the feedback.
> 1. netperf: get_transport_info: getsockopt: errno 95
> This failure is due to the missing TCP_MAXSEG socket option support. May
> be this is OK as this option
> doesn't make much sense when using RDMA. Or we could return a reasonable
> value.
Missing socket options are
Add setsockopt options that allow the user to specify the desired
size of the underlying QP. The provided sizes are used as the
maximum size when creating the QP. The actual sizes of the QP
are the smaller of the user provided maximum and the maximum
sizes supported by the underlying hardware.
A
Allow a user to control some of the RDMA related attributes
of an rsocket through setsockopt/getsockopt. A user specifies
that the rsocket should be modified through SOL_RDMA level.
This patch provides the initial framework. Subsequent patches
will add the configurable parameters.
Signed-off-by
Allow the user to override the default inline data size.
We still require a minimum size in order to transfer receive
buffer update message.
Signed-off-by: Sean Hefty
---
We can eliminate the need for inline entirely by reserving some of the
send buffer space for control messages, but that work i
Allow the user to specify the size of the send/receive queues
and inline data size through environment variables:
RS_SQ_SIZE, RS_RQ_SIZE, and RS_INLINE.
Signed-off-by: Sean Hefty
---
src/preload.c | 39 +++
1 files changed, 39 insertions(+), 0 deletions(-)
Signed-off-by: Sean Hefty
---
src/rsocket.c | 26 +++---
1 files changed, 11 insertions(+), 15 deletions(-)
diff --git a/src/rsocket.c b/src/rsocket.c
index ef070a8..b89ef42 100644
--- a/src/rsocket.c
+++ b/src/rsocket.c
@@ -126,6 +126,9 @@ union rs_wr_id {
};
};
Avoid possible buffer overrun.
Signed-off-by: Sean Hefty
---
examples/rstream.c | 38 +-
1 files changed, 21 insertions(+), 17 deletions(-)
diff --git a/examples/rstream.c b/examples/rstream.c
index df36e34..054d11e 100644
--- a/examples/rstream.c
+++ b/exa
Simple latency/bandwidth tests using rstream showed minimal
difference in performance between using a QP sized to 384
entries versus 512. Reduce the overhead of a default rsocket
by using 384 entries. A user can request a larger size by
calling rsetsockopt.
Signed-off-by: Sean Hefty
---
src/rs
Signed-off-by: Sean Hefty
---
examples/rstream.c | 34 +-
man/rstream.1 |5 -
2 files changed, 17 insertions(+), 22 deletions(-)
diff --git a/examples/rstream.c b/examples/rstream.c
index c440f04..df36e34 100644
--- a/examples/rstream.c
+++ b/exampl
When ACM support is enabled in the librdmacm, it will attempt to
establish a socket connection to the ACM daemon. When the rsocket
preload library is in use, this can result in a recursive call
to socket() that results in the library hanging. The resulting
call stack is:
socket() -> rsocket() ->
When porting rsockets to iwarp, it was discovered that the default
QP size (512) was larger than that supported by the hardware.
Decrease the size of the QP if the default size is larger than
the maximum supported by the hardware.
Signed-off-by: Sean Hefty
---
src/cma.c | 10
src/
socket() can return 0 as a valid socket. This can happen
when using a daemon that closes stdin/out/err.
Signed-off-by: Sean Hefty
---
src/acm.c |8
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/acm.c b/src/acm.c
index bcf11da..9c65919 100755
--- a/src/acm.c
+++
Avoid creating a connection to the ACM service when
it's not needed. For example, if the user of the librdmacm
is a server application, it will not use ACM services.
Signed-off-by: Sean Hefty
---
src/acm.c | 22 +-
src/cma.c |2 --
src/cma.h |2 --
3 files changed,
If a user calls rrecv() after a blocking rsocket has been disconnected,
it will hang. This problem and the cause was reported by Sridhar Samudrala
. It can be reproduced by running netserver -f -D
using the rs-preload library. A similar issue exists with rsend().
Fix this by not blocking on a C
Optimize each connection for either latency or bandwidth
results. This improves small message latency under 384
bytes by .5 - 1 us, while increasing bandwidth by
1 - 1.5 Gbps.
Signed-off-by: Sean Hefty
---
examples/rstream.c | 146 ++--
1 files c
Signed-off-by: Sean Hefty
---
examples/rstream.c | 14 +-
1 files changed, 13 insertions(+), 1 deletions(-)
diff --git a/examples/rstream.c b/examples/rstream.c
index 104b318..c440f04 100644
--- a/examples/rstream.c
+++ b/examples/rstream.c
@@ -448,7 +448,8 @@ static int client_co
The following patch set provides fixes, adds user configurability,
and optimizes rsockets based on the needs and results of performance analysis
and wider testing.
Additional optimizations will follow, but I wanted to go ahead and
push these changes out.
Signed-off-by: Sean Hefty
--
To unsubscr
ucma_destroy_cqs() destroys both the send and recv CQs if they
are non-null. If the two CQs are actually the same one, this
results in a crash when trying to destroy the second CQ. Check
that the CQs are different before destroying the second CQ.
This fixes a crash when using rsockets, which set
> Did you try out the patches? was it helpful to address the problem
> you're facing?
I have not had time to test it yet
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/maj
> We are trying to figure out the cause for RDMA_CM_EVENT_ROUTE_ERROR
> errors after a failover event of the bonding driver.
> The event status returned is -EINVAL. To gather further information on
> when this EINVAL is returned,
> I added some debug which showed 3 for mad_hdr.status in the below
>
> 1. Limit max number of wqes per QP reported when querying the device,
> so that ib_create_qp will never fail due to any additional headroom WQEs
> allocated.
>
> 2. Limit qp resources accepted for ib_create_qp() to the limits
> reported in ib_query_device(). In kernel space, make sure that
> the
> netperf
> - By default netserver forks a child process for each netperf client. As
> rsockets
>doesn't support fork() yet, this doesn't work.
Btw, fork() is unlikely to work anytime soon, if ever...
> - netserver has a -f option to disable forking a child and handle 1
> netperf client at a
Older versions of acm.h do not include the resolve_data or
perf_data fields in struct acm_msg. If we're using an older
version of the acm.h header file, use an internal definition
of struct acm_msg.
Signed-off-by: Sean Hefty
---
configure.in |5 +
src/acm.c| 16 ++--
2
The RDMA CM fail calls to set REUSEADDR on an rdma_cm_id if
it is not in the idle state. As a result, this causes a failure
in NetPipe when run with socket calls intercepted by rsockets.
Fix this by returning success when REUSEADDR is set on an rsocket
that has already been connected. When runnin
If asynchronous use is specified (use of poll/select), set the
rsocket to nonblocking. This matches the common usage case for
asynchronous sockets.
When asynchronous support is enabled, the nonblocking/blocking
test option determines whether the poll/select call will block,
or if rstream will spi
The NODELAY option is coupled with whether the socket is blocking
or nonblocking. Remove this coupling and always set the NODELAY
option.
NODELAY currently has no effect on rsockets.
Signed-off-by: Sean Hefty
---
examples/rstream.c | 11 ++-
1 files changed, 2 insertions(+), 9 deleti
The base set of rstream tests want nonblocking rsockets, but don't
actually set the rsocket to nonblocking. It instead relies on the
MSG_DONTWAIT flag. Make the code match the expected behavior and
set the rsocket to nonblocking and make nonblocking the default.
Provide a test option to switch i
The RDMA CM fail calls to set REUSEADDR on an rdma_cm_id if
it is not in the idle state. As a result, this causes a failure
in NetPipe when run with socket calls intercepted by rsockets.
Fix this by returning success when REUSEADDR is set on an rsocket
that has already been connected. When runnin
> > A test that acquired and released a lock 2 billion times reported that
> > the custom lock was roughly 20% faster than using the mutex.
> > 26.6 seconds versus 33.0 seconds.
>
> I think you are measuring the fact your call is inlined and pthreads
> has an indirect jump - because internally pth
Hotspot performance analysis using VTune showed pthread_mutex_unlock()
as the most significant hotspot when transferring small messages using
rstream. To reduce the impact of using pthread mutexes, replace it
with a custom lock built using an atomic variable and a semaphore.
When there's no conten
> Are you sure that only one thread at a time will invoke a CM callback ? As
> far as I can see cm_recv_handler() queues work without checking whether
> any other work is ongoing. From drivers/infiniband/core/cm.c:
All callbacks for a single ID should be serialized. (I think the listen ID is
an
> That makes me wonder how it is prevented that two CM callbacks for the
> same CM ID run concurrently on different CPUs ?
The callback code ends up looking like this:
ret = atomic_inc_and_test(&cm_id_priv->work_count);
if (!ret)
list_add_tail(&work->list, &cm_id_p
> If I interpret the source code in drivers/infiniband/core/cm.c correctly
> ib_destroy_cm_id() can return before an ongoing cm_id callback has
> finished. Is this on purpose ? If not, isn't there a
> flush_workqueue(cm.wq) call missing in cm_destroy_id() ?
ib_destroy_cm_id() will block while ther
> Hal/Sean, I defer to you on whether you think we should add this change to
> ibstat. If you all recommend against it,
> we'll advise customers to use ibv_devinfo, which is included with libibverbs
> and is required for iwarp user apps. But
> it seems a minimal change to add, and previously it d
> Users seem to expect ibstat to show all rdma devices...
They why not change ibstat to use ibverbs or have it gather its data directly?
What other functionality does umad provide for RNICs?
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to maj
> I noticed that doing ibstat, none of the iWARP RNIC adapters were showing
> up. I have attached a patch to address the issue (libibumad.patch).
>
> SYS_NODE_TYPE for iWARP RNIC is 4 and is_ib_type only checked to 3.
I don't think libibumad should support RNICs. Users can use ibv_devinfo
inst
The following lockdep problem was reported by Or Gerlitz
.
=
[ INFO: possible recursive locking detected ]
3.3.0-32035-g1b2649e-dirty #4 Not tainted
-
kworker/5:1/418 is trying to acquire lock:
(&id_priv->hand
> The usual fix is to use mutex_lock_nested, cf "Exception: Nested data
> dependencies leading to nested locking" in
> Documentation/lockdep-design.txt. I haven't looked at how easy that
> is to do in this case.
I think in this case, it's easier to just delay the call to rdma_destroy_id()
until
> Specifically, I see that your commit 186834b5de69 "RDMA/ucma: Fix
> AB-BA deadlock" deals only with the ucma and this trace doesn't
> have ucma symbols, so is that a different race? or false alarm?
Maybe I'm overlooking something, but this looks like a false match.
> ==
thanks - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
thanks - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> It's been a few days since I posted this question,
> but I've had no responses so far.
Can I use the IB netlink infrastructure to produce actual RDMA data transfers?
no
If this is not possible, what is the actual entry point in the IB kernel code
that results in a call to get_user_pages()?
I
thanks - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> We evaluated the verbs extension framework and we noticed that two new
> attributes in the structures:
> ibv_device
> ibv_context
>
> contain an attribute called "private", which is a reserved work in C++.
> Changing those attributes to private_data seems like a good idea.
>
> Do you wa
I've committed the rsocket implementation in my librdmacm git tree, but I'll be
fairly open about wire protocol changes until an actual release.
I'd like to start a discussion on the best way to support direct placement of
data into an application's buffer with rsockets. I spoke with many peopl
I'm a little slow writing this up, but for anyone interested, see below for
details on the wire protocol.
> +#define RS_QP_CTRL_SIZE 4
4 entries on the send queue are reserved for control messages. (At least 1 is
needed to avoid deadlock.)
User data is only transferred if there is an availabl
ibacm release 1.0.6 is now available from:
https://www.openfabrics.org/downloads/rdmacm/ibacm-1.0.6.tar.gz
The git shortlog from 1.0.5 is:
Dotan Barak (1):
After allocation of dynamic memory blocks, check the allocation
Hal Rosenstock (4):
ib_acme: Fix typo
ib_acme: Use IPv4 r
> Interesting work. Regarding direct data placement: have the io_submit()
> and io_getevents() system calls been considered ? Those are the
> foundation of libaio.
I did look at aio and other interfaces, but I didn't try to move beyond the
idea phase. With aio, I didn't see how the calls could e
rcopy will copy files from a source system to a specified remote
server. It's essentially a really dumb FTP type program that can
be used to quickly transfer files between systems, which can be
useful to verify data integrity.
(It was easier to create this program than modify an existing FTP
clie
The following patch set contains an initial implementation of rsockets as
presented at the 2012 OpenFabrics Workshop. A copy of that presentation
is available at:
https://www.openfabrics.org/downloads/rdmacm/rsockets-ofa12.pptx
and a video of the presentation can be found here:
rstream provides an example that uses either rsocket or socket
APIs. The latter allows rstream to be used to verify rsocket
behavior compared to socket.
Signed-off-by: Sean Hefty
---
Makefile.am|5
examples/rstream.c | 570
man/
Intercept socket calls and convert TCP socket operation to
streaming over RDMA.
Allow falling back from rsockets to normal sockets on error
or when trying to bind/connect to a reserved port. This is
needed to handle MPI job startup, where MPI should use rsockets,
but mpiexect needs to communicate
ibacm: Fixes to ACM package to support distros
From: Sean Hefty
Set of changes to fixup the ibacm package for inclusion into RedHat 6.
Changes are based on feedback from Doug Ledford .
These are primarily changes to the build files, along with name changes
to the man pages and sample configurati
Thanks - I combined this patch with your other LID format change patch.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
> > I create a .tar.gz package using 'make dist', copy it to another
> > system, then install it using 'configure && make install'. When I do
> > that, sysconfdir defaults to /usr/local/etc, sbindir /usr/local/sbin,
> > and bindir to /usr/local/bin. I added /usr/local to PATH, so that
> > the ini
> In this instance you need to have configure variable substitute your
> init script as well. You should not manipulate PATH from an init.d
> script, and you should use an absolute path to the daemon to avoid
> contamination from the invoking user's environment.
This is exactly what I wanted to do
applied
Thanks for the changes!
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
thanks - applied
I will add a separate patch to support IPv6, since librdmacm should be updated
as well.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.htm
Thanks - applied
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
I've incorporated most of the changes, but see below.
> > -EXTRA_DIST = src/acm_mad.h src/libacm.h \
> > -linux/osd.h linux/dlist.h ibacm.spec.in $(man_MANS) acm_opts.cfg \
> > -acm_addr.cfg
> > +EXTRA_DIST = src/acm_mad.h src/libacm.h ibacm.init \
> > +linux/osd.h linux/dl
If the ibacm header file is available, automatically have the
librdmacm configured to use it. This removes the --with-ib_acm
configure option.
Signed-off-by: Sean Hefty
---
configure.in | 16 ++--
1 files changed, 2 insertions(+), 14 deletions(-)
diff --git a/configure.in b/confi
Set of changes to fixup the ibacm package for inclusion into RedHat 6.
Changes are based on feedback from Doug Ledford .
These are primarily changes to the build files, along with name changes
to the man pages and sample configuration files.
Rename the ib_acm service to match the package name, iba
> By default ibacm expects to find its configuration files in /etc/ibacm.
> This adds to the proliferation of directories in /etc/ needlessly. We
> already have a number of RDMA related directories to choose from
> depending on your install (OFED == /etc/ofed or /etc/openib in the old
> days, RHE
> We noticed the following interesting scenario:
>
> One host sent a CM request to a remote host.
>
> The remote host, which doesn't have any CM support, performed the following
> steps:
>
>
> * Replaced the sLID with the dLID
> * Added an indication that this is a response MAD
> *
> On 4/2/2012 7:35 PM, Hal Rosenstock wrote:
> > Rather than always overwriting active_speed in this case, wouldn't it
> > be better to only do that for invalid values?
>
> Yes, I have thought about that, however, spotting invalid values would
> make the code a bit ugly, so I took this approach, R
> - async event handling.
Can you briefly explain how the code synchronizes between async event reporting
and the user trying to free the same object?
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majord...@vger.kernel.org
More majordomo info
> +static int ocrdma_inet6addr_event(struct notifier_block *,
> + unsigned long, void *);
> +
> +static struct notifier_block ocrdma_inet6addr_notifier = {
> + .notifier_call = ocrdma_inet6addr_event
> +};
> +
> +int ocrdma_get_instance(void)
> +{
> + int insta
> +static inline void *ocrdma_get_eqe(struct ocrdma_eq *eq)
> +{
> + return (u8 *)eq->q.va + (eq->q.tail * sizeof(struct ocrdma_eqe));
> +}
casts from (void *) to (u8 *) are not needed. This occurs in multiple places.
> +enum ib_qp_state get_ibqp_state(enum ocrdma_qp_state qps)
> +{
> +
> +struct ocrdma_alloc_ucontext_resp {
> + u32 dev_id;
> + u32 wqe_size;
> + u32 max_inline_data;
> + u32 dpp_wqe_size;
> + u64 ah_tbl_page;
> + u32 ah_tbl_len;
> + u32 rsvd;
> + u8 fw_ver[32];
> + u32 rqe_size;
> + u64 rsvd1;
> +} __packed;
Is there some re
> +struct ocrdma_cq {
> + struct ib_cq ibcq;
> + struct ocrdma_dev *dev;
nit: There are several structures where you store ocrdma_dev *. You can remove
these and use the struct ib_* to reach it as well.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body
> What mkey model is being proposed here ? It looks to me like it is a
> single mkey for all ports in the subnet which is the simplest but least
> flexible model. If so, I think we need something more flexible as IBA
> allows each port to have it's own different mkey.
Is something more needed than
> On Mon, Feb 27, 2012 at 2:22 PM, Hefty, Sean wrote:
> >> > --- a/include/linux/socket.h
> >> > +++ b/include/linux/socket.h
> >> > @@ -184,6 +184,7 @@ struct ucred {
> >> > #define AF_PPPOX 24 /* PPPoX sockets */
&g
When we destroy a cm_id, we must purge associated events from the
event queue. If the cm_id is for a listen request, we also purge
corresponding pending connect requests. This requires destroying
the cm_id's associated with the connect requests by calling
rdma_destroy_id(). rdma_destroy_id() blo
> I run into the below ucma related warnings from the kernel with 3.3-rc5
> when I stepped over crash of process as of wrong libs/etc (not the point
> here...). Do you see here a real bug? basically the process was exiting
> and the cleanup code in the kernel was running rdma_destroy_id when a
> ca
701 - 800 of 1448 matches
Mail list logo
