In ib_mad.h, ib_mad_snoop_handler uses send_buf rather than send_wr
Signed-off-by: Hal Rosenstock
---
Change since v2:
Changed title to use "higher" language
Change since v1:
Fixed typo in patch description
diff --git a/include/rdma/ib_mad.h b/include/rdma/ib_mad.h
index
On Fri, Dec 25, 2015 at 06:46:07PM +0200, Liran Liss wrote:
> > From: Jason Gunthorpe
>
> > > >fill mr->key by the lkey or rkey based on that and everything will
> > > >work fine.
> > >
> > > But the ULP *can* register a memory buffer with local and remote
> > >
> -Original Message-
> From: linux-rdma-ow...@vger.kernel.org
> [mailto:linux-rdma-ow...@vger.kernel.org] On Behalf Of Doug Ledford
> Sent: Wednesday, December 23, 2015 9:19 PM
> To: J. Bruce Fields; Chuck Lever
> Cc: Christoph Hellwig; Anna Schumaker; Jason Gunthorpe;
>
On Tue, Jan 5, 2016 at 6:46 PM, Steve Wise wrote:
> Hey Doug, I don't see this branch. Which branch has the accepted device attr
> change?
k.o/for-4.5 on Doug's kernel.org tree
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of
> In ib_mad.h, ib_mad_snoop_handler uses send_buf rather than send_wr
The MAD snooping should be removed from the mad stack. There are no in tree
users and the only attempt at adding one was rejected.
Let the target core check task existence instead of the SRP target
driver.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 54 ++-
1 file changed, 2 insertions(+), 52
The only allowed channel state changes are those that change
the channel state into a state with a higher numerical value.
This allows to merge the functions srpt_set_ch_state() and
srpt_test_and_set_ch_state() into a single function.
Signed-off-by: Bart Van Assche
The target core guarantees that shutdown_session() is only invoked
once per session. This means that the ib_srpt target driver doesn't
have to track whether or not shutdown_session() has been called.
Additionally, ensure that target_sess_cmd_list_set_waiting() is
called before
Although sizeof is an operator and hence in many cases parentheses can
be left out, the recommended kernel coding style is to surround the
sizeof argument with parentheses. This patch does not change any
functionality. This patch has been generated by running the following
shell command:
sed -i
srpt_sdev_name() is too trivial to keep it as a separate function.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 14 ++
1 file changed, 2 insertions(+), 12 deletions(-)
diff --git
Keep the ib_srpt session as long as srpt_close_session() may
access it.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 4
1 file changed, 4 insertions(+)
diff --git
srpt_abort_cmd() must not be called in state SRPT_STATE_DATA_IN. Issue
a warning if this occurs.
srpt_abort_cmd() must not invoke target_put_sess_cmd() for commands
in state SRPT_STATE_DONE because the srpt_abort_cmd() callers already
do this when necessary. Hence remove this call.
If an RDMA
In the CM REQ message handler, store the channel pointer in
cm_id->context such that the function srpt_find_channel() is no
longer needed. Additionally, make the CM event messages more
informative.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
The target core function that should be called if target_submit_cmd()
fails is target_put_sess_cmd(). Additionally, change the return type
of srpt_handle_cmd() from int into void.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
Avoid that srpt_close_session() waits if it doesn't have to wait.
Additionally, increase the time during which srpt_close_session()
waits until closing a session has finished. This makes it easier
to detect session shutdown bugs.
Signed-off-by: Bart Van Assche
Cc:
The callers of srpt_get_ch_state() can access ch->state safely without
using locking. Hence inline this function.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 42 ++-
1
The Last WQE Reached event is only generated after one or more work
requests have been queued on the QP associated with a session. Since
session shutdown can start before any work requests have been queued,
use a zero-length RDMA write to wait until a QP has been drained.
Signed-off-by: Bart Van
The only allowed return values for the write_pending() callback
function are 0, -EAGAIN and -ENOMEM. Since attempting to perform
RDMA over a disconnecting channel will result in an IB error
completion anyway, remove the code that checks the channel state
from srpt_write_pending().
Signed-off-by:
Use the function target_reverse_dma_direction() instead of
reimplementing it.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c | 17 ++---
1 file changed, 2 insertions(+), 15 deletions(-)
diff
Just like other target drivers, use scsilun_to_int() to unpack SCSI
LUN numbers. This patch only changes the behavior of ib_srpt for LUN
numbers >= 16384.
Signed-off-by: Bart Van Assche
Cc: Christoph Hellwig
---
drivers/infiniband/ulp/srpt/ib_srpt.c |
> > There are no in tree users and the only attempt at adding one was
> rejected.
>
> There are no in tree users of this but there is your madeye tool (which
> is out of tree). This is still a useful debug tool for MADs and there
> are people who still use that.
It's an out of tree tool.
This patchset adds support for RDMA cgroup by addressing review comments
of [1] by implementing published RFC [2].
Overview:
Currently user space applications can easily take away all the rdma
device specific resources such as AH, CQ, QP, MR etc. Due to which other
applications in other cgroup or
Adds RDMA controller to limit the number of RDMA resources that can be
consumed by processes of a rdma cgroup.
RDMA resources are global resource that can be exhauasted without
reaching any kmemcg or other policy. RDMA cgroup implementation allows
limiting RDMA/IB well defined resources to be
Added function pointer table to store resource pool specific
operation for each resource type (verb and hw).
Added list node to link device to rdma cgroup so that it can
participate in resource accounting and limit configuration.
Signed-off-by: Parav Pandit
---
In ib_mad.h, ib_mad_snoop_handler uses send_buf rather than send_wr
Signed-off-by: Hal Rosenstock
---
Change since v3:
Fixed title to not include function name
Change since v2:
Changed title to use "higher" language
Change since v1:
Fixed typo in patch description
diff
Added rdma cgroup header file which defines its APIs to perform
charing/uncharing functionality.
Signed-off-by: Parav Pandit
---
include/linux/cgroup_rdma.h | 91 +
1 file changed, 91 insertions(+)
create mode 100644
On 1/5/2016 12:38 PM, Hefty, Sean wrote:
>> In ib_mad.h, ib_mad_snoop_handler uses send_buf rather than send_wr
>
> The MAD snooping should be removed from the mad stack.
This last discussed on linux-rdma list back in late September when Ira
posted a partial RFC patch to do this.
> There are
On Mon, Jan 04, 2016 at 10:23:46PM +0200, Leon Romanovsky wrote:
> On Mon, Jan 04, 2016 at 11:21:19AM -0500, Jubin John wrote:
> > From: Jim Snow
> >
> > } else {
> > + /* Auto activate link on non-SC15 packet receive */
> > +
It uses charge API to perform resource charing before allocating low
level resource. It continues to link the resource to the owning
thread group leader task.
It uncharges the resource after successful deallocation of resource.
Signed-off-by: Parav Pandit
---
It defines verb RDMA resources that will be registered with
RDMA cgroup. It defines new APIs to register device with
RDMA cgroup and defines resource token table access interface.
Signed-off-by: Parav Pandit
---
drivers/infiniband/core/Makefile| 1 +
Added documentation for rdma controller to use in legacy mode and
using new unified hirerchy.
Signed-off-by: Parav Pandit
---
Documentation/cgroup-legacy/rdma.txt | 129 +++
Documentation/cgroup.txt | 79 +
From: Jim Snow
The link state will transition from ARMED to ACTIVE when a non-SC15
packet arrives, but the driver might not notice the change. With this
fix, if the slowpath receive interrupt handler sees a non-SC15 packet
while in the ARMED state, we queue work to call
On Tue, Jan 5, 2016 at 9:00 PM, Hefty, Sean wrote:
>> > There are no in tree users and the only attempt at adding one was
>> rejected.
>>
>> There are no in tree users of this but there is your madeye tool (which
>> is out of tree). This is still a useful debug tool for MADs
The following series of patches is what I came up with while testing the
most recent version of my SCSI target patch series (see also
http://thread.gmane.org/gmane.linux.scsi.target.devel/10905):
0001-IB-srpt-Add-parentheses-around-sizeof-argument.patch
0002-IB-srpt-Inline-srpt_sdev_name.patch
Looks good,
Reviewed-by: Christoph Hellwig
--
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
On Tue, Jan 05, 2016 at 03:22:46PM +0100, Bart Van Assche wrote:
> Just like other target drivers, use scsilun_to_int() to unpack SCSI
> LUN numbers. This patch only changes the behavior of ib_srpt for LUN
> numbers >= 16384.
>
> Signed-off-by: Bart Van Assche
> Cc:
On Tue, Jan 05, 2016 at 03:23:14PM +0100, Bart Van Assche wrote:
> Let the target core check task existence instead of the SRP target
> driver.
Looks good,
Reviewed-by: Christoph Hellwig
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a
On Tue, Jan 05, 2016 at 03:20:50PM +0100, Bart Van Assche wrote:
> Although sizeof is an operator and hence in many cases parentheses can
> be left out, the recommended kernel coding style is to surround the
> sizeof argument with parentheses. This patch does not change any
> functionality. This
On Tue, Jan 05, 2016 at 03:21:53PM +0100, Bart Van Assche wrote:
> The callers of srpt_get_ch_state() can access ch->state safely without
> using locking. Hence inline this function.
Looks good,
Reviewed-by: Christoph Hellwig
--
To unsubscribe from this list: send the line
> pr_debug("Aborting cmd with state %d and tag %lld\n", state,
>ioctx->cmd.tag);
>
> @@ -1299,14 +1291,16 @@ static int srpt_abort_cmd(struct srpt_send_ioctx
> *ioctx)
> case SRPT_STATE_NEW:
> case SRPT_STATE_DATA_IN:
> case SRPT_STATE_MGMT:
> + case
On Tue, Jan 05, 2016 at 03:24:49PM +0100, Bart Van Assche wrote:
> Avoid that srpt_close_session() waits if it doesn't have to wait.
> Additionally, increase the time during which srpt_close_session()
> waits until closing a session has finished. This makes it easier
> to detect session shutdown
On Tue, Jan 05, 2016 at 03:24:15PM +0100, Bart Van Assche wrote:
> The target core guarantees that shutdown_session() is only invoked
> once per session. This means that the ib_srpt target driver doesn't
> have to track whether or not shutdown_session() has been called.
> Additionally, ensure that
On Tue, Jan 05, 2016 at 03:25:13PM +0100, Bart Van Assche wrote:
> The target core function that should be called if target_submit_cmd()
> fails is target_put_sess_cmd(). Additionally, change the return type
> of srpt_handle_cmd() from int into void.
I actually ran into this bug a long time ago
On Tue, Jan 05, 2016 at 03:23:45PM +0100, Bart Van Assche wrote:
> The only allowed channel state changes are those that change
> the channel state into a state with a higher numerical value.
> This allows to merge the functions srpt_set_ch_state() and
> srpt_test_and_set_ch_state() into a single
On Tue, Jan 05, 2016 at 03:26:22PM +0100, Bart Van Assche wrote:
> In the CM REQ message handler, store the channel pointer in
> cm_id->context such that the function srpt_find_channel() is no
> longer needed. Additionally, make the CM event messages more
> informative.
Looks fine,
Reviewed-by:
On Tue, Jan 05, 2016 at 10:46:36AM -0700, Jason Gunthorpe wrote:
> > ULPs are *already* using the same registrations for both local and
> > remote access.
>
> Where? Out of tree?
I haven't found anything in-tree for sure.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma"
Remove the local workqueue to process mad completions and use the CQ API
instead.
Signed-off-by: Christoph Hellwig
Reviewed-by: Hal Rosenstock
Reviewed-by: Ira Weiny
---
drivers/infiniband/core/mad.c | 162
Hi Doug,
I am trying to rping server, but it fails when bound to any address other then
IF_ANY.
# rping -s -a 102.1.1.129 -C1 -p -vd
created cm_id 0x23d7800
rdma_bind_addr: No such file or directory
destroy cm_id 0x23d7800
If bound to IF_ANY address, server starts but client fails to
On Mon, Jan 04, 2016 at 07:04:03PM -0500, ira.weiny wrote:
> Sorry I did not catch this before but rather than void * wouldn't it be better
> to use struct ib_cqe?
Sure, I'll fix it up.
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to
Looks good,
Reviewed-by: Christoph Hellwig
--
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
> srpt_disconnect_ch(ch);
>
> + kref_put(>kref, srpt_free_ch);
At some point it might be a good idea to have a srpt_put_ch helper to wrap
this pattern.
Otherwise looks good:
Reviewed-by: Christoph Hellwig
--
To unsubscribe from this list: send the line "unsubscribe
On Tue, Jan 05, 2016 at 03:26:49PM +0100, Bart Van Assche wrote:
> The Last WQE Reached event is only generated after one or more work
> requests have been queued on the QP associated with a session. Since
> session shutdown can start before any work requests have been queued,
> use a zero-length
> > > > >
> > > > > If Doug accepts the library changes, let me know that public git
> > > > > commit
> > > > > and I can pull it into the staging-next branch and you can continue to
> > > > > send me staging patches that way.
> > > >
> > > > Won't this cause a conflict during the merge window?
On Thu, Dec 24, 2015 at 9:46 AM, Matan Barak wrote:
> On Wed, Dec 23, 2015 at 10:04 PM, Doug Ledford wrote:
>> On 10/15/2015 12:58 PM, Hefty, Sean wrote:
>> ib_create_ah_from_wc needs to resolve the DMAC in order to create the
>> AH (this
Hello,
Please consider including mainline commit
50a9eea694ab8e0779069e0a4e0b12e145521468 in the next v4.2.y-ckt
release. It was
included in the mainline tree as of v4.4-rc2. It has been tested and
confirmed to resolve:
http://bugs.launchpad.net/bugs/1528466
commit
On Wed, Jan 06, 2016 at 12:28:02AM +0530, Parav Pandit wrote:
> Added function pointer table to store resource pool specific
> operation for each resource type (verb and hw).
> Added list node to link device to rdma cgroup so that it can
> participate in resource accounting and limit
On Tue, Jan 05, 2016 at 01:52:55PM -0500, Hal Rosenstock wrote:
> In ib_mad.h, ib_mad_snoop_handler uses send_buf rather than send_wr
>
> Signed-off-by: Hal Rosenstock
First off I have to say; this comment is wrong and should be fixed.
Reviewed-by: Ira Weiny
On Tue, 2016-01-05 at 16:01 -0500, Joseph Salisbury wrote:
> Hello,
>
> Please consider including mainline commit
> 50a9eea694ab8e0779069e0a4e0b12e145521468 in the next v4.2.y-ckt
> release. It was
> included in the mainline tree as of v4.4-rc2. It has been tested and
> confirmed to resolve:
>
Hello,
On Wed, Jan 06, 2016 at 12:28:00AM +0530, Parav Pandit wrote:
> Resources are not defined by the RDMA cgroup. Resources are defined
> by RDMA/IB stack & optionally by HCA vendor device drivers.
As I wrote before, I don't think this is a good idea. Drivers will
inevitably add non-sensical
Hello,
On Wed, Jan 06, 2016 at 12:28:06AM +0530, Parav Pandit wrote:
> +5-4-1. RDMA Interface Files
> +
> + rdma.resource.verb.list
> + rdma.resource.verb.limit
> + rdma.resource.verb.usage
> + rdma.resource.verb.failcnt
> + rdma.resource.hw.list
> + rdma.resource.hw.limit
> +
Hello,
On Wed, Jan 06, 2016 at 12:28:03AM +0530, Parav Pandit wrote:
> +/* hash table to keep map of tgid to owner cgroup */
> +DEFINE_HASHTABLE(pid_cg_map_tbl, 7);
> +DEFINE_SPINLOCK(pid_cg_map_lock);/* lock to protect hash table access */
> +
> +/* Keeps mapping of pid to its owning cgroup
61 matches
Mail list logo