[PATCH 3/13] IB-mgmt: Add SA MAD definitions to umad

Add a new header file, umad_sa.h, to libibumad that contains basic SA MAD definitions. Only the SA attributes values, new methods, and base SA MAD format are given. SA attribute structure definitions are left for future development. Because the SM key is an unaligned 64-bit value (the only one t

[PATCH 0/13] IB-mgmt: Port madeye to userspace

This series ports the kernel madeye debug utility to userspace. It depends on the kernel MAD snooping functionality patches recently submitted to this list. The series adds the following to the ib-mgmt tree: * Adds the ability to snoop MADs to libibumad. * Adds new header files to libibumad that

[PATCH 13/13] IB-mgmt: Define data type to indicate values are in big-endian

The ib management interfaces require that the user specify certain values in network-byte order. To clarify which values should be in big endian format, introduce a new data type at the libibumad level. Byte arrays which are in big endian format are simply marked with a comment. The new data typ

[PATCH 7/13] IB-mgmt: Update mad.h to use umad_types.h

Update mad.h to include umad_types.h and use the umad definitions. The data type names in mad.h are kept for backwards compatibility, until all applications update to using the umad names. Signed-off-by: Sean Hefty --- libibmad/include/infiniband/mad.h | 173 +++

[PATCH 5/13] IB-mgmt: Add CM definitions to umad

To support the madeye diag and ibacm, add CM MAD definitions to umad. Only the attributes values are added for now. Additional structure definitions will be added later. Signed-off-by: Sean Hefty --- libibumad/include/infiniband/umad_cm.h | 64 1 files change

[PATCH 11/13] IB-mgmt: Move ntohll/htonll from mad.h to umad.h

Users of umad require ntohll/htonll to set/extract data from MADs. Move the definition from libibmad to libibumad. Signed-off-by: Sean Hefty --- libibmad/include/infiniband/mad.h | 29 + libibmad/include/infiniband/mad_osd.h |2 -- libibumad/include/infinib

[PATCH 8/13] IB-mgmt: Fix trace table response

A response to a get trace table request is a get table response. There is no 'get trace table response'. See IB spec 1.2.1, section 15.4.9. Signed-off-by: Sean Hefty --- Submitting as a separate patch, so the change doesn't get lost in the noise. libibmad/include/infiniband/mad.h |2 +- 1

[PATCH 12/13] IB-mgmt: Add utility to display snooped mads

Port the kernel madeye utility to user space. Use the new MAD snooping ability exported from the kernel to view captured MADs and display them. Signed-off-by: Sean Hefty --- infiniband-diags/Makefile.am |6 infiniband-diags/configure.in |3 infiniband-diags/infi

[PATCH 10/13] IB-mgmt: Remove duplicate definition of IB_MAD_STATUS_CLASS_MASK

osmtest includes a duplicate definition. Remove it. Signed-off-by: Sean Hefty --- opensm/osmtest/osmtest.c |2 -- 1 files changed, 0 insertions(+), 2 deletions(-) mode change 100644 => 100755 opensm/osmtest/osmtest.c diff --git a/opensm/osmtest/osmtest.c b/opensm/osmtest/osmtest.c old mod

[PATCH 6/13] IB-mgmt: Add new umad header files to release

Signed-off-by: Sean Hefty --- libibumad/Makefile.am |6 +- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/libibumad/Makefile.am b/libibumad/Makefile.am index 6dc2955..1d9ca5c 100644 --- a/libibumad/Makefile.am +++ b/libibumad/Makefile.am @@ -35,7 +35,11 @@ libibumad_la_DEP

[PATCH 1/13] IB-mgmt: Add API to register for MAD snooping

Add a new call to libibumad that allows a user to register for snooping MADs. Snooped MADs are returned to the user through the umad_recv routine. Signed-off-by: Sean Hefty --- libibumad/Makefile.am |1 + libibumad/include/infiniband/umad.h | 10 ++ libibumad/man/umad_re

[PATCH 4/13] IB-mgmt: Add basic SM definitions to umad

Add a new header file, umad_sm.h, to libibumad that contains basic SM MAD definitions. Only the SM attributes values, new methods, and base SM MAD format are given. SM attribute structure definitions are left for future development. Signed-off-by: Sean Hefty --- libibumad/include/infiniband/um

[PATCH 9/13] IB-mgmt: Convert ib_types.h to use umad types

Several data type definitions have been added to umad. Update ib_types.h to make use of these definitions. For backwards compatibility, the existing names in ib_types.h are kept until they can be deprecated. Signed-off-by: Sean Hefty --- opensm/include/iba/ib_types.h | 186

[PATCH 2/13] IB-mgmt: Provide MAD definitions with libibumad

The only definitions for MAD data types are in ib_types.h, which is an opensm file. The libibumad interface is essentially useless without these data types. Add new header files to libibumad with basic MAD data types. The definitions and structures are based on those defined in the linux kernel.

[PATCH 1/2] ib/mad: Simplify snooping interface

In preparation for exporting the kernel mad snooping capability to user space, remove all code originally inserted as place holders and simplify the mad snooping interface. For performance reasons, we want to filter which mads are reported to clients of the snooping interface at the lowest level,

[PATCH 2/2] IB/umad: Export mad snooping capability to userspace

The kernel mad interface allows a client to view all sent and received MADs. This has proven to be a useful debugging techinque when used with an external module. However, external kernel modules are not easily supported and do not easily allow for more complex filtering or analysis of MAD traffic

RE: [PATCH v2] Add exponential backoff + random delay to MADs when retrying after timeout.

Okay - I see that; but it's independent of the code I've patched. Basically, if an RMPP is requested it won't be affected by this change, it will continue to use the existing algorithm - which appears to use a maximum 2 second time out for the first try of each segment (mad_rmpp.c). Reviewing

RE: [PATCH v2] Add exponential backoff + random delay to MADs when retrying after timeout.

> Hal asked the same thing - and I'm confused because I thought > that if receiving an RMPP response times out, the entire transaction is > aborted. RMPP still uses retries. If the user specifies a timeout of 1 second, with 3 retries, _each_ RMPP window will be retried up to 3 times, waiting for

RE: [PATCH v2] Add exponential backoff + random delay to MADs when retrying after timeout.

Sean said: > When you consider RMPP, the timeout/retry values specified by > the user are not straightforward in their meaning. I haven't > look at this patch in detail yet, but how do the timeout > changes work with RMPP MADs? Is the timeout reset to the > minimum after an ACK is received?

RE: [PATCH V2 0/3] New RAW_PACKET QP type

> The following patches add a new QP type named RAW_PACKET. > > This type of QP is used in Ethernet environment and intended for > creation of a whole packet, including L2 headers, RAW_FRAME seems like a better name to me. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma"

[PATCH V2 3/3] Add RAW_PACKET to ib_attach/detach mcast calls

Added RAW_PACKET to ib_attach/detach calls Add iWARP transport to attach/detach mcast. Differentiate between transport types when dealing with multicast. Signed-off-by: Aleksey Senin --- drivers/infiniband/core/verbs.c | 24 1 files changed, 20 insertions(+), 4 delet

[PATCH V2 2/3] Security check on QP type

Check if user process has permission to create RAW_PACKET QP. In order to prevent regular user from creating malicious packets, at the time of QP creation, kernel will check if the process has necessary permissions. This examination areis done in uverbs layer and is general for all low level drive

[PATCH V2 1/3] RAW_PACKET QP type definition

The following patch add a new QP type named RAW_PACKET. This type of QP is used in Ethernet environment and intended for creation of a whole packet, including L2 headers, from userspace. Reserve a place for future XRC patch in order to be sure that OFED kernel ABI will not broken when XRC and RAW

[PATCH V2 0/3] New RAW_PACKET QP type

The following patches add a new QP type named RAW_PACKET. This type of QP is used in Ethernet environment and intended for creation of a whole packet, including L2 headers, from userspace. In order to prevent regular user from creating malicious packets, at the time of QP creation, kernel will che