platform/kernel/linux-exynos.git
7 years agoMerge branches 'chelsio', 'debug-cleanup', 'hns' and 'i40iw' into merge-test
Doug Ledford [Wed, 14 Dec 2016 19:43:14 +0000 (14:43 -0500)]
Merge branches 'chelsio', 'debug-cleanup', 'hns' and 'i40iw' into merge-test

7 years agoi40iw: Reorganize structures to align with HW capabilities
Henry Orosco [Tue, 6 Dec 2016 22:16:20 +0000 (16:16 -0600)]
i40iw: Reorganize structures to align with HW capabilities

Some resources are incorrectly organized and at odds with
HW capabilities. Specifically, ILQ, IEQ, QPs, MSS, QOS
and statistics belong in a VSI.

Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix incorrect check for error
Mustafa Ismail [Tue, 6 Dec 2016 21:49:35 +0000 (15:49 -0600)]
i40iw: Fix incorrect check for error

In i40iw_ieq_handle_partial() the check for !status is incorrect.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Assign MSS only when it is a new MTU
Mustafa Ismail [Tue, 6 Dec 2016 21:49:34 +0000 (15:49 -0600)]
i40iw: Assign MSS only when it is a new MTU

Currently we are changing the MSS regardless of whether
there is a change or not in MTU. Fix to make the
assignment of MSS dependent on an MTU change.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix race condition in terminate timer's handler
Shiraz Saleem [Tue, 6 Dec 2016 21:49:33 +0000 (15:49 -0600)]
i40iw: Fix race condition in terminate timer's handler

Add a QP reference when terminate timer is started to ensure
the destroy QP doesn't race ahead to free the QP while it is being
referenced in the terminate timer's handler.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix memory leak in CQP destroy when in reset
Mustafa Ismail [Tue, 6 Dec 2016 21:49:32 +0000 (15:49 -0600)]
i40iw: Fix memory leak in CQP destroy when in reset

On a device close, the control QP (CQP) is destroyed by calling
cqp_destroy which destroys the CQP and frees its SD buffer memory.
However, if the reset flag is true, cqp_destroy is never called and
leads to a memory leak on SD buffer memory. Fix this by always calling
cqp_destroy, on device close, regardless of reset. The exception to this
when CQP create fails. In this case, the SD buffer memory is already
freed on an error check and there is no need to call cqp_destroy.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix QP flush to not hang on empty queues or failure
Shiraz Saleem [Tue, 6 Dec 2016 21:49:31 +0000 (15:49 -0600)]
i40iw: Fix QP flush to not hang on empty queues or failure

When flush QP and there are no pending work requests, signal completion
to unblock i40iw_drain_sq and i40iw_drain_rq which are waiting on
completion for iwqp->sq_drained and iwqp->sq_drained respectively.
Also, signal completion if flush QP fails to prevent the drain SQ or RQ
from being blocked indefintely.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix double free of QP
Mustafa Ismail [Tue, 6 Dec 2016 21:49:30 +0000 (15:49 -0600)]
i40iw: Fix double free of QP

A QP can be double freed if i40iw_cm_disconn() is
called while it is currently being freed by
i40iw_rem_ref(). The fix in i40iw_cm_disconn() will
first check if the QP is already freed before
making another request for the QP to be freed.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Use correct src address in memcpy to rdma stats counters
Shiraz Saleem [Fri, 11 Nov 2016 16:55:41 +0000 (10:55 -0600)]
i40iw: Use correct src address in memcpy to rdma stats counters

hw_stats is a pointer to i40_iw_dev_stats struct in i40iw_get_hw_stats().
Use hw_stats and not &hw_stats in the memcpy to copy the i40iw device stats
data into rdma_hw_stats counters.

Fixes: b40f4757daa1 ("IB/core: Make device counter infrastructure dynamic")

Cc: stable@vger.kernel.org # 4.7+
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove macros I40IW_STAG_KEY_FROM_STAG and I40IW_STAG_INDEX_FROM_STAG
Thomas Huth [Wed, 5 Oct 2016 11:55:38 +0000 (13:55 +0200)]
i40iw: Remove macros I40IW_STAG_KEY_FROM_STAG and I40IW_STAG_INDEX_FROM_STAG

The macros I40IW_STAG_KEY_FROM_STAG and I40IW_STAG_INDEX_FROM_STAG are
apparently bad - they are using the logical "&&" operation which
does not make sense here. It should have been a bitwise "&" instead.
Since the macros seem to be completely unused, let's simply remove
them so that nobody accidentially uses them in the future. And while
we're at it, also remove the unused macro I40IW_CREATE_STAG.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Acked-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoiw_cxgb4: Fix error return code in c4iw_rdev_open()
Wei Yongjun [Sat, 17 Sep 2016 00:41:37 +0000 (00:41 +0000)]
iw_cxgb4: Fix error return code in c4iw_rdev_open()

Fix to return error code -ENOMEM from the __get_free_page() error
handling case instead of 0, as done elsewhere in this function.

Fixes: 05eb23893c2c ("cxgb4/iw_cxgb4: Doorbell Drop Avoidance Bug Fixes")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Acked-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add request for reset on CQP timeout
Henry Orosco [Wed, 30 Nov 2016 21:14:15 +0000 (15:14 -0600)]
i40iw: Add request for reset on CQP timeout

When CQP times out, send a request to LAN driver for reset.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Code cleanup, remove check of PBLE pages
Henry Orosco [Wed, 30 Nov 2016 21:13:47 +0000 (15:13 -0600)]
i40iw: Code cleanup, remove check of PBLE pages

Remove check for zero 'pages' of unallocated pbles calculated in
add_pble_pool(); as it can never be true.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Correctly fail loopback connection if no listener
Shiraz Saleem [Wed, 30 Nov 2016 21:12:35 +0000 (15:12 -0600)]
i40iw: Correctly fail loopback connection if no listener

Fail the connect and return the proper error code if a client
is started with local IP address and there is no corresponding
loopback listener.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fill in IRD value when on connect request
Shiraz Saleem [Wed, 30 Nov 2016 21:12:11 +0000 (15:12 -0600)]
i40iw: Fill in IRD value when on connect request

IRD is not populated on connect request and application is
getting 0 for the value. Fill in the correct value on
connect request.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Set TOS field in IP header
Shiraz Saleem [Wed, 30 Nov 2016 21:09:34 +0000 (15:09 -0600)]
i40iw: Set TOS field in IP header

Set the TOS field in IP header with the value passed in
from application. If there is mismatch between the remote
client's TOS and listener, set the listener Tos to the higher
of the two values.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add NULL check for ibqp event handler
Shiraz Saleem [Wed, 30 Nov 2016 21:09:07 +0000 (15:09 -0600)]
i40iw: Add NULL check for ibqp event handler

Add NULL check for ibqp event handler before calling it to report
QP events, as it might not initialized.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Replace list_for_each_entry macro with safe version
Mustafa Ismail [Wed, 30 Nov 2016 21:08:34 +0000 (15:08 -0600)]
i40iw: Replace list_for_each_entry macro with safe version

Use list_for_each_entry_safe macro for the IPv6 addr list
as IPv6 addresses can be deleted while going through the
list.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add IP addr handling on netdev events
Mustafa Ismail [Wed, 30 Nov 2016 21:07:30 +0000 (15:07 -0600)]
i40iw: Add IP addr handling on netdev events

Disable listeners and disconnect all connected QPs on
a netdev interface down event. On an interface up event,
the listeners are re-enabled.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add missing cleanup on device close
Mustafa Ismail [Wed, 30 Nov 2016 20:59:26 +0000 (14:59 -0600)]
i40iw: Add missing cleanup on device close

On i40iw device close, disconnect all connected QPs by moving
them to error state; and block further QPs, PDs and CQs from
being created. Additionally, make sure all resources have been
freed before deallocating the ibdev as part of the device close.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add 2MB page support
Henry Orosco [Wed, 30 Nov 2016 20:57:40 +0000 (14:57 -0600)]
i40iw: Add 2MB page support

Add support to allow each independent memory region to
be configured for 2MB page size in addition to 4KB
page size.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Utilize physically mapped memory regions
Henry Orosco [Wed, 30 Nov 2016 20:56:14 +0000 (14:56 -0600)]
i40iw: Utilize physically mapped memory regions

Add support to use physically mapped WQ's and MR's if determined
that the OS registered user-memory for the region is physically
contiguous. This feature will eliminate the need for unnecessarily
setting up and using PBL's when not required.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix incorrect assignment of SQ head
Henry Orosco [Tue, 22 Nov 2016 15:44:20 +0000 (09:44 -0600)]
i40iw: Fix incorrect assignment of SQ head

The SQ head is incorrectly incremented when the number
of WQEs required is greater than the number available.
The fix is to use the I40IW_RING_MOV_HEAD_BY_COUNT
macro. This checks for the SQ full condition first and
only if SQ has room for the request, then we move the
head appropriately.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove variable flush_code and check to set qp->sq_flush
Henry Orosco [Thu, 10 Nov 2016 04:20:31 +0000 (22:20 -0600)]
i40iw: Remove variable flush_code and check to set qp->sq_flush

The flush_code variable in i40iw_bld_terminate_hdr() is obsolete and
the check to set qp->sq_flush is unreachable. Currently flush code is
populated in setup_term_hdr() and both SQ and RQ are flushed always
as part of the tear down flow.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove check on return from device_init_pestat()
Henry Orosco [Thu, 10 Nov 2016 03:42:26 +0000 (21:42 -0600)]
i40iw: Remove check on return from device_init_pestat()

Remove unnecessary check for return code from
device_init_pestat() and change func to void.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Use runtime check for IS_ENABLED(CONFIG_IPV6)
Henry Orosco [Thu, 10 Nov 2016 03:34:02 +0000 (21:34 -0600)]
i40iw: Use runtime check for IS_ENABLED(CONFIG_IPV6)

To be consistent, use the runtime check instead of
conditional compile.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Use actual page size
Henry Orosco [Thu, 10 Nov 2016 03:33:32 +0000 (21:33 -0600)]
i40iw: Use actual page size

In i40iw_post_send, use the actual page size instead of
encoded page size. This is to be consistent with the
rest of the file.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove NULL check for cm_node->iwdev
Henry Orosco [Thu, 10 Nov 2016 03:32:25 +0000 (21:32 -0600)]
i40iw: Remove NULL check for cm_node->iwdev

It is not necessary to check cm_node->iwdev in
i40iw_rem_ref_cm_node() as it can never be NULL after
a successful call out of i40iw_make_cm_node().

Signed-off-by: Chien Tin Tung <chien.tin.tung@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove checks for more than 48 bytes inline data
Henry Orosco [Thu, 10 Nov 2016 03:32:03 +0000 (21:32 -0600)]
i40iw: Remove checks for more than 48 bytes inline data

Remove dead code, which isn't executed because we
return error if the data size is greater than 48 bytes.

Inline data size greater than 48 bytes isn't supported
and the maximum WQE size is 64 bytes.

Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Query device accounts for internal rsrc
Henry Orosco [Thu, 10 Nov 2016 03:30:28 +0000 (21:30 -0600)]
i40iw: Query device accounts for internal rsrc

Some resources are consumed internally and not available to the user.
After hw is initialized, figure out how many resources are consumed
and subtract those numbers from the initial max device capability in
i40iw_query_device().

Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Optimize inline data copy
Henry Orosco [Thu, 10 Nov 2016 03:28:02 +0000 (21:28 -0600)]
i40iw: Optimize inline data copy

Use memcpy for inline data copy in sends
and writes instead of byte by byte copy.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Fix for LAN handler removal
Henry Orosco [Thu, 10 Nov 2016 03:27:02 +0000 (21:27 -0600)]
i40iw: Fix for LAN handler removal

If i40iw_open() fails for any reason, the LAN handler
is not being removed. Modify i40iw_deinit_device()
to always remove the handler.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Correct values for max_recv_sge, max_send_sge
Henry Orosco [Thu, 10 Nov 2016 03:26:39 +0000 (21:26 -0600)]
i40iw: Correct values for max_recv_sge, max_send_sge

When creating QPs, ensure init_attr->cap.max_recv_sge
is clipped to MAX_FRAG_COUNT.

Expose MAX_FRAG_COUNT for max_recv_sge and max_send_sge in
i40iw_query_qp().

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Reviewed-By: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Use vector when creating CQs
Henry Orosco [Thu, 10 Nov 2016 03:24:48 +0000 (21:24 -0600)]
i40iw: Use vector when creating CQs

Assign each CEQ vector to a different CPU when possible, then
when creating a CQ, use the vector for the CEQ id. This
allows completion work to be distributed over multiple cores.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Convert page_size to encoded value
Henry Orosco [Sun, 20 Nov 2016 02:26:25 +0000 (20:26 -0600)]
i40iw: Convert page_size to encoded value

Passed in page_size was used as encoded value for writing
the WQE and passed in value was usually 4096. This was
working out since bit 0 was 0 and implies 4KB pages,
but would not work for other page sizes.

Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Set MAX IRD, MAX ORD size to max supported value
Henry Orosco [Wed, 19 Oct 2016 20:33:58 +0000 (15:33 -0500)]
i40iw: Set MAX IRD, MAX ORD size to max supported value

Set the MAX_IRD and MAX_ORD size negotiated to the maximum
supported values.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Remove workaround for pre-production errata
Henry Orosco [Wed, 19 Oct 2016 20:33:32 +0000 (15:33 -0500)]
i40iw: Remove workaround for pre-production errata

Pre-production silicon incorrectly truncates 4 bytes of the MPA
packet in UDP loopback case. Remove the workaround as it is no
longer necessary.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Enable message packing
Henry Orosco [Wed, 19 Oct 2016 20:32:53 +0000 (15:32 -0500)]
i40iw: Enable message packing

Remove the parameter to disable message packing and
always enable it.

Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoi40iw: Add Quality of Service support
Henry Orosco [Tue, 11 Oct 2016 02:12:10 +0000 (21:12 -0500)]
i40iw: Add Quality of Service support

Add support for QoS on QPs. Upon device initialization,
a map is created from user priority to queue set
handles. On QP creation, use ToS to look up the queue
set handle for use with the QP.

Signed-off-by: Faisal Latif <faisal.latif@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Henry Orosco <henry.orosco@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Move HNS RoCE user vendor structures
Leon Romanovsky [Wed, 19 Oct 2016 17:13:07 +0000 (20:13 +0300)]
IB/hns: Move HNS RoCE user vendor structures

This patch moves HNS vendor's specific structures to
common UAPI folder which will be visible to all consumers.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the IB device name
Lijun Ou [Tue, 29 Nov 2016 23:10:30 +0000 (23:10 +0000)]
IB/hns: Fix the IB device name

This patch mainly fix the name for IB device in order
to match with libhns.

Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the bug when free cq
Shaobo Xu [Tue, 29 Nov 2016 23:10:29 +0000 (23:10 +0000)]
IB/hns: Fix the bug when free cq

If the resources of cq are freed while executing the user case, hardware
can not been notified in hip06 SoC. Then hardware will hold on when it
writes the cq buffer which has been released.

In order to slove this problem, RoCE driver checks the CQE counter, and
ensure that the outstanding CQE have been written. Then the cq buffer
can be released.

Signed-off-by: Shaobo Xu <xushaobo2@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Delete the redundant memset operation
Wei Hu (Xavier) [Tue, 29 Nov 2016 23:10:28 +0000 (23:10 +0000)]
IB/hns: Delete the redundant memset operation

It deleted the redundant memset operation because the memory allocated
by ib_alloc_device has been set zero.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the bug of setting port mtu
Wei Hu (Xavier) [Tue, 29 Nov 2016 23:10:27 +0000 (23:10 +0000)]
IB/hns: Fix the bug of setting port mtu

In hns_roce driver, we need not call iboe_get_mtu to reduce
IB headers from effective IBoE MTU because hr_dev->caps.max_mtu
has already been reduced.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the bug when free mr
Shaobo Xu [Tue, 29 Nov 2016 23:10:26 +0000 (23:10 +0000)]
IB/hns: Fix the bug when free mr

If the resources of mr are freed while executing the user case, hardware
can not been notified in hip06 SoC. Then hardware will hold on when it
reads the payload by the PA which has been released.

In order to slove this problem, RoCE driver creates 8 reserved loopback
QPs to ensure zero wqe when free mr. When the mac address is reset, in
order to avoid loopback failure, we need to release the reserved loopback
QPs and recreate them.

Signed-off-by: Shaobo Xu <xushaobo2@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the bug when destroy qp
Wei Hu (Xavier) [Tue, 29 Nov 2016 23:10:25 +0000 (23:10 +0000)]
IB/hns: Fix the bug when destroy qp

If send queue is still working when qp is in reset state by modify qp
in destroy qp function, hardware will hold on and don't work in hip06
SoC. In current codes, RoCE driver check hardware pointer of sending and
hardware pointer of processing to ensure that hardware has processed all
the dbs of this qp. But while the environment of wire becomes not good,
The checking time maybe too long.

In order to solve this problem, RoCE driver created a workqueue at probe
function. If there is a timeout when checking the status of qp, driver
initialize work entry and push it into the workqueue, Work function will
finish checking and release the related resources later.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Dongdong Huang(Donald) <hdd.huang@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix for Checkpatch.pl comment style errors
Salil [Wed, 23 Nov 2016 19:41:09 +0000 (19:41 +0000)]
IB/hns: Fix for Checkpatch.pl comment style errors

This patch correct the comment style errors caught by
checkpatch.pl script

Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Implement the add_gid/del_gid and optimize the GIDs management
Shaobo Xu [Wed, 23 Nov 2016 19:41:08 +0000 (19:41 +0000)]
IB/hns: Implement the add_gid/del_gid and optimize the GIDs management

IB core has implemented the calculation of GIDs and the management
of GID tables, and it is now responsible to supply query function
for GIDs. So the calculation of GIDs and the management of GID
tables in the RoCE driver is redundant.

The patch is to implement the add_gid/del_gid to set the GIDs in
the RoCE driver, remove the redundant calculation and management of
GIDs in the notifier call of the net device and the inet, and
update the query_gid.

Signed-off-by: Shaobo Xu <xushaobo2@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Change qpn allocation to round-robin mode.
Wei Hu (Xavier) [Wed, 23 Nov 2016 19:41:07 +0000 (19:41 +0000)]
IB/hns: Change qpn allocation to round-robin mode.

When using CM to establish connections, qp number that was freed
just now will be rejected by ib core. To fix these problem, We
change qpn allocation to round-robin mode. We added the round-robin
mode for allocating resources using bitmap. We use round-robin mode
for qp number and non round-robing mode for other resources like
cq number, pd number etc.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Modify query info named port_num when querying RC QP
Wei Hu (Xavier) [Wed, 23 Nov 2016 19:41:06 +0000 (19:41 +0000)]
IB/hns: Modify query info named port_num when querying RC QP

This patch modified the output query info qp_attr->port_num
to fix bug in hip06.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Modify the macro for the timeout when cmd process
Wei Hu (Xavier) [Wed, 23 Nov 2016 19:41:05 +0000 (19:41 +0000)]
IB/hns: Modify the macro for the timeout when cmd process

This patch modified the macro for the timeout when cmd is
processing as follows:
Before modification:
 enum {
HNS_ROCE_CMD_TIME_CLASS_A       = 10000,
HNS_ROCE_CMD_TIME_CLASS_B       = 10000,
HNS_ROCE_CMD_TIME_CLASS_C       = 10000,
 };
After modification:
 #define HNS_ROCE_CMD_TIMEOUT_MSECS 10000

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Fix the bug for qp state in hns_roce_v1_m_qp()
Lijun Ou [Wed, 23 Nov 2016 19:41:04 +0000 (19:41 +0000)]
IB/hns: Fix the bug for qp state in hns_roce_v1_m_qp()

In old code, the value of qp state from qpc was assigned for
attr->qp_state. The value may be an error while attr_mask &
IB_QP_STATE is zero.

Signed-off-by: Lijun Ou <oulijun@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Modify the condition of notifying hardware loopback
Lijun Ou [Wed, 23 Nov 2016 19:41:03 +0000 (19:41 +0000)]
IB/hns: Modify the condition of notifying hardware loopback

This patch modified the condition of notifying hardware loopback.

In hip06, RoCE Engine has several ports, one QP is related
to one port. hardware only support loopback in the same port,
not in the different ports.

So, If QP related to port N, the dmac in the QP context equals
the smac of the local port N or the loop_idc is 1, we should
set loopback bit in QP context to notify hardware.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: add self loopback for CM
Lijun Ou [Wed, 23 Nov 2016 19:41:02 +0000 (19:41 +0000)]
IB/hns: add self loopback for CM

This patch mainly adds self loopback support for CM.

Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Peter Chen <luck.chen@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Optimize the logic of allocating memory using APIs
Wei Hu (Xavier) [Wed, 23 Nov 2016 19:41:01 +0000 (19:41 +0000)]
IB/hns: Optimize the logic of allocating memory using APIs

This patch modified the logic of allocating memory using APIs in
hns RoCE driver. We used kcalloc instead of kmalloc_array and
bitmap_zero. And When kcalloc failed, call vzalloc to alloc
memory.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Ping Zhang <zhangping5@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Add code for refreshing CQ CI using TPTR
Wei Hu (Xavier) [Wed, 23 Nov 2016 19:41:00 +0000 (19:41 +0000)]
IB/hns: Add code for refreshing CQ CI using TPTR

This patch added the code for refreshing CQ CI using TPTR in hip06
SoC.

We will send a doorbell to hardware for refreshing CQ CI when user
succeed to poll a cqe. But it will be failed if the doorbell has
been blocked. So hardware will read a special buffer called TPTR
to get the lastest CI value when the cq is almost full.

This patch support the special CI buffer as follows:
a) Alloc the memory for TPTR in the hns_roce_tptr_init function and
   free it in hns_roce_tptr_free function, these two functions will
   be called in probe function and in the remove function.
b) Add the code for computing offset(every cq need 2 bytes) and
   write the dma addr to every cq context to notice hardware in the
   function named hns_roce_v1_write_cqc.
c) Add code for mapping TPTR buffer to user space in function named
   hns_roce_mmap. The mapping distinguish TPTR and UAR of user mode
   by vm_pgoff(0: UAR, 1: TPTR, others:invaild) in hip06.
d) Alloc the code for refreshing CQ CI using TPTR in the function
   named hns_roce_v1_poll_cq.
e) Add some variable definitions to the related structure.

Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Dongdong Huang(Donald) <hdd.huang@huawei.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hns: Add the interface for querying QP1
Lijun Ou [Wed, 23 Nov 2016 19:40:59 +0000 (19:40 +0000)]
IB/hns: Add the interface for querying QP1

In old code, It only added the interface for querying non-specific
QP. This patch mainly adds an interface for querying QP1.

Signed-off-by: Lijun Ou <oulijun@huawei.com>
Reviewed-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoinfiniband: remove WARN that is not kernel bug
Leon Romanovsky [Mon, 21 Nov 2016 17:38:20 +0000 (19:38 +0200)]
infiniband: remove WARN that is not kernel bug

On Mon, Nov 21, 2016 at 09:52:53AM -0700, Jason Gunthorpe wrote:
> On Mon, Nov 21, 2016 at 02:14:08PM +0200, Leon Romanovsky wrote:
> > >
> > > In ib_ucm_write function there is a wrong prefix:
> > >
> > > + pr_err_once("ucm_write: process %d (%s) tried to do something hinky\n",
> >
> > I did it intentionally to have the same errors for all flows.
>
> Lets actually use a good message too please?
>
>  pr_err_once("ucm_write: process %d (%s) changed security contexts after opening FD, this is not allowed.\n",
>
> Jason

>From 70f95b2d35aea42e5b97e7d27ab2f4e8effcbe67 Mon Sep 17 00:00:00 2001
From: Leon Romanovsky <leonro@mellanox.com>
Date: Mon, 21 Nov 2016 13:30:59 +0200
Subject: [PATCH rdma-next V2] IB/{core, qib}: Remove WARN that is not kernel bug

WARNINGs mean kernel bugs, in this case, they are placed
to mark programming errors and/or malicious attempts.

BUG/WARNs that are not kernel bugs hinder automated testing efforts.

Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/ipoib: Remove and fix debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:25 +0000 (16:44 +0200)]
IB/ipoib: Remove and fix debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/isert: Remove and fix debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:24 +0000 (16:44 +0200)]
IB/isert: Remove and fix debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Acked-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/rxe: Remove and fix debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:23 +0000 (16:44 +0200)]
IB/rxe: Remove and fix debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/ocrdma: Remove and fix debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:22 +0000 (16:44 +0200)]
IB/ocrdma: Remove and fix debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/usninc: Remove and fix debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:21 +0000 (16:44 +0200)]
IB/usninc: Remove and fix debug prints after allocation failure

This patch removes unneeded prints after allocation failure
and moves one debug print into the appropriate place.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/mthca: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:20 +0000 (16:44 +0200)]
IB/mthca: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/nes: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:19 +0000 (16:44 +0200)]
IB/nes: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/qib: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:18 +0000 (16:44 +0200)]
IB/qib: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/i40iw: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:17 +0000 (16:44 +0200)]
IB/i40iw: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/cxgb4: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:16 +0000 (16:44 +0200)]
IB/cxgb4: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/cxgb3: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:15 +0000 (16:44 +0200)]
IB/cxgb3: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/hfi1: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:14 +0000 (16:44 +0200)]
IB/hfi1: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/mlx5: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:13 +0000 (16:44 +0200)]
IB/mlx5: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/mlx4: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:12 +0000 (16:44 +0200)]
IB/mlx4: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/core: Release allocated memory in cache setup failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:11 +0000 (16:44 +0200)]
IB/core: Release allocated memory in cache setup failure

The failure in ib_cache_setup_one function during
ib_register_device will leave leaked allocated memory.

Fixes: 03db3a2d81e6 ("IB/core: Add RoCE GID table management")
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/core: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:10 +0000 (16:44 +0200)]
IB/core: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agoIB/mad: Remove debug prints after allocation failure
Leon Romanovsky [Thu, 3 Nov 2016 14:44:09 +0000 (16:44 +0200)]
IB/mad: Remove debug prints after allocation failure

The prints after [k|v][m|z|c]alloc() functions are not needed,
because in case of failure, allocator will print their internal
error prints anyway.

Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Doug Ledford <dledford@redhat.com>
7 years agonet/mlx5e: Add tc support for FWD rule with counter
Mark Bloch [Tue, 20 Sep 2016 07:58:29 +0000 (07:58 +0000)]
net/mlx5e: Add tc support for FWD rule with counter

When creating a FWD rule using tc create also a HW counter
for this rule.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Add option to add fwd rule with counter
Mark Bloch [Tue, 20 Sep 2016 07:59:05 +0000 (07:59 +0000)]
net/mlx5: Add option to add fwd rule with counter

Currently the code supports only drop rules to possess counters,
add that ability also for fwd rules.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Add multi dest support
Mark Bloch [Wed, 31 Aug 2016 11:24:25 +0000 (11:24 +0000)]
net/mlx5: Add multi dest support

Currently when calling mlx5_add_flow_rule we accept
only one flow destination, this commit allows to pass
multiple destinations.

This change forces us to change the return structure to a more
flexible one. We introduce a flow handle (struct mlx5_flow_handle),
it holds internally the number for rules created and holds an array
where each cell points the to a flow rule.

From the consumers (of mlx5_add_flow_rule) point of view this
change is only cosmetic and requires only to change the type
of the returned value they store.

From the core point of view, we now need to use a loop when
allocating and deleting rules (e.g given to us a flow handler).

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Group similer rules under the same fte
Mark Bloch [Sun, 11 Sep 2016 13:03:06 +0000 (13:03 +0000)]
net/mlx5: Group similer rules under the same fte

When adding a new rule, if we can match it with compare_match_value and
flow tag we might be able to insert the rule to the same fte.
In order to do that, there must be an overlap between the actions of the
fte and the new rule.

When updating the action of an existing fte, we must tell the firmware
we are doing so.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Refactor find_flow_rule
Mark Bloch [Fri, 16 Sep 2016 07:00:10 +0000 (07:00 +0000)]
net/mlx5: Refactor find_flow_rule

The way we compare between two dests will need to be used in other
places in the future, so we factor out the comparison logic
between two dests into a separate function.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Use fte status to decide on firmware command
Mark Bloch [Sun, 11 Sep 2016 12:54:50 +0000 (12:54 +0000)]
net/mlx5: Use fte status to decide on firmware command

An fte status becomes FS_FTE_STATUS_EXISTING only after it was
created in HW. We can use this in order to simplify the logic on
what firmware command to use. If the status isn't FS_FTE_STATUS_EXISTING
we need to create the fte, otherwise we need only to update it.

Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Don't unlock fte while still using it
Mark Bloch [Mon, 5 Sep 2016 10:58:04 +0000 (10:58 +0000)]
net/mlx5: Don't unlock fte while still using it

When adding a new rule to an fte, we need to hold the fte lock
until we add that rule to the fte and increase the fte ref count.

Fixes: 0c56b97503fd ("net/mlx5_core: Introduce flow steering API")
Signed-off-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Add SRIOV VF max rate configuration support
Mohamad Haj Yahia [Thu, 11 Aug 2016 08:28:21 +0000 (11:28 +0300)]
net/mlx5: Add SRIOV VF max rate configuration support

Implement the vf set rate ndo by modifying the TSAR vport rate limit.

Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Introduce E-switch QoS management
Mohamad Haj Yahia [Thu, 11 Aug 2016 08:26:36 +0000 (11:26 +0300)]
net/mlx5: Introduce E-switch QoS management

Add TSAR to the eswitch which will act as the vports rate limiter.
Create/Destroy TSAR on Enable/Dsiable SRIOV.
Attach/Detach vport to eswitch TSAR on Enable/Disable vport.

Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Introduce TSAR manipulation firmware commands
Mohamad Haj Yahia [Thu, 11 Aug 2016 08:21:39 +0000 (11:21 +0300)]
net/mlx5: Introduce TSAR manipulation firmware commands

TSAR (stands for Transmit Scheduling ARbiter) is a hardware component
that is responsible for selecting the next entity to serve on the
transmit path.
The arbitration defines the QoS policy between the agents connected to
the TSAR.
The TSAR is a consist two main features:
1) BW Allocation between agents:
The TSAR implements a defecit weighted round robin between the agents.
Each agent attached to the TSAR is assigned with a weight and it is
awarded transmission tokens according to this weight.
2) Rate limer per agent:
Each agent attached to the TSAR is (optionally) assigned with a rate
limit.
TSAR will not allow scheduling for an agent exceeding its defined rate
limit.

In this patch we implement the API of manipulating the TSAR.

Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Add ConnectX-5 PCIe 4.0 VF device ID
Saeed Mahameed [Tue, 18 Oct 2016 18:35:35 +0000 (21:35 +0300)]
net/mlx5: Add ConnectX-5 PCIe 4.0 VF device ID

For the mlx5 driver to support ConnectX-5 PCIe 4.0 VFs, we add the
device ID "0x101a" to mlx5_core_pci_table.

Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Fix length of async_event_mask
Eugenia Emantayev [Wed, 28 Sep 2016 14:29:41 +0000 (17:29 +0300)]
net/mlx5: Fix length of async_event_mask

According to PRM async_event_mask have to be 64 bits long.

Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Ensure SRQ physical address structure endianness
Artemy Kovalyov [Wed, 31 Aug 2016 05:29:58 +0000 (05:29 +0000)]
net/mlx5: Ensure SRQ physical address structure endianness

SRQ physical address structure field should be in big-endian format.

Signed-off-by: Artemy Kovalyov <artemyko@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agonet/mlx5: Update struct mlx5_ifc_xrqc_bits
Artemy Kovalyov [Wed, 31 Aug 2016 05:17:54 +0000 (05:17 +0000)]
net/mlx5: Update struct mlx5_ifc_xrqc_bits

Update struct mlx5_ifc_xrqc_bits according to last specification

Signed-off-by: Artemy Kovalyov <artemyko@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agoIB/mlx5: Skip handling unknown events
Saeed Mahameed [Thu, 29 Sep 2016 16:35:38 +0000 (19:35 +0300)]
IB/mlx5: Skip handling unknown events

Do not dispatch unknown mlx5 core events on mlx5_ib_event.

Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
7 years agoLinux 4.9-rc3 v4.9-rc3
Linus Torvalds [Sat, 29 Oct 2016 20:52:02 +0000 (13:52 -0700)]
Linux 4.9-rc3

7 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 29 Oct 2016 20:42:44 +0000 (13:42 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 bugfix from Thomas Gleixner:
 "A single bugfix for the recent changes related to registering the boot
  cpu when this has not happened before prefill_possible_map().

  The main problem with this change got fixed already, but we missed the
  case where the local APIC is not yet mapped, when prefill_possible_map()
  is invoked, so the registration of the boot cpu which has the APIC bit
  set in CPUID will explode.

  I should have seen that issue earlier, but all I can do now is feeling
  embarassed"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/smpboot: Init apic mapping before usage

7 years agoMerge tag 'upstream-4.9-rc3' of git://git.infradead.org/linux-ubifs
Linus Torvalds [Sat, 29 Oct 2016 20:15:24 +0000 (13:15 -0700)]
Merge tag 'upstream-4.9-rc3' of git://git.infradead.org/linux-ubifs

Pull ubi/ubifs fixes from Richard Weinberger:
 "This contains fixes for issues in both UBI and UBIFS:

   - A regression wrt overlayfs, introduced in -rc2.
   - An UBI issue, found by Dan Carpenter's static checker"

* tag 'upstream-4.9-rc3' of git://git.infradead.org/linux-ubifs:
  ubifs: Fix regression in ubifs_readdir()
  ubi: fastmap: Fix add_vol() return value test in ubi_attach_fastmap()

7 years agoMerge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Linus Torvalds [Sat, 29 Oct 2016 19:07:29 +0000 (12:07 -0700)]
Merge tag 'armsoc-fixes' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "We haven't seen a whole lot of fixes for the first two weeks since the
  merge window, but here is the batch that we have at the moment.

  Nothing sticks out as particularly bad or scary, it's mostly a handful
  of smaller fixes to several platforms. The Uniphier reset controller
  changes could probably have been delayed to 4.10, but they're not
  scary and just plumbing up driver changes that went in during the
  merge window.

  We're also adding another maintainer to Marvell Berlin platforms, to
  help out when Sebastian is too busy. Yay teamwork!"

* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031
  ARM: dts: vf610: fix IRQ flag of global timer
  ARM: imx: gpc: Fix the imx_gpc_genpd_init() error path
  ARM: imx: gpc: Initialize all power domains
  arm64: dts: Updated NAND DT properties for NS2 SVK
  arm64: dts: uniphier: change MIO node to SD control node
  ARM: dts: uniphier: change MIO node to SD control node
  reset: uniphier: rename MIO reset to SD reset for Pro5, PXs2, LD20 SoCs
  arm64: uniphier: select ARCH_HAS_RESET_CONTROLLER
  ARM: uniphier: select ARCH_HAS_RESET_CONTROLLER
  arm64: dts: Add timer erratum property for LS2080A and LS1043A
  arm64: dts: rockchip: remove the abuse of keep-power-in-suspend
  ARM: multi_v7_defconfig: Enable Intel e1000e driver
  MAINTAINERS: add myself as Marvell berlin SoC maintainer
  bus: qcom-ebi2: depend on ARCH_QCOM or COMPILE_TEST
  ARM: dts: fix the SD card on the Snowball
  arm64: dts: rockchip: remove always-on and boot-on from vcc_sd
  arm64: dts: marvell: fix clocksource for CP110 master SPI0
  ARM: mvebu: Select corediv clk for all mvebu v7 SoC

7 years agoMerge tag 'char-misc-4.9-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 29 Oct 2016 18:19:02 +0000 (11:19 -0700)]
Merge tag 'char-misc-4.9-rc3' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes from Greg KH:
 "Here are a few small char/misc driver fixes for reported issues.

  The "biggest" are two binder fixes for reported issues that have been
  shipping in Android phones for a while now, the others are various
  fixes for reported problems.

  And there's a MAINTAINERS update for good measure.

  All have been in linux-next with no reported issues"

* tag 'char-misc-4.9-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  MAINTAINERS: Add entry for genwqe driver
  VMCI: Doorbell create and destroy fixes
  GenWQE: Fix bad page access during abort of resource allocation
  vme: vme_get_size potentially returning incorrect value on failure
  extcon: qcom-spmi-misc: Sync the extcon state on interrupt
  hv: do not lose pending heartbeat vmbus packets
  mei: txe: don't clean an unprocessed interrupt cause.
  ANDROID: binder: Clear binder and cookie when setting handle in flat binder struct
  ANDROID: binder: Add strong ref checks

7 years agoMerge tag 'v4.9-rockchip-dts64-fixes1' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Sat, 29 Oct 2016 18:09:37 +0000 (11:09 -0700)]
Merge tag 'v4.9-rockchip-dts64-fixes1' of git://git./linux/kernel/git/mmind/linux-rockchip into fixes

Correct regulator handling on Rockchip arm64 boards to make
bind/unbind calls work correctly and remove a sdio-only
property from non-sdio mmc hosts, that accidentially was
added there.

* tag 'v4.9-rockchip-dts64-fixes1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
  arm64: dts: rockchip: remove the abuse of keep-power-in-suspend
  arm64: dts: rockchip: remove always-on and boot-on from vcc_sd

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'arm-soc/for-4.9/devicetree-arm64-fixes' of http://github.com/Broadcom...
Olof Johansson [Sat, 29 Oct 2016 18:09:11 +0000 (11:09 -0700)]
Merge tag 'arm-soc/for-4.9/devicetree-arm64-fixes' of github.com/Broadcom/stblinux into fixes

This pull request contains a single fix for Broadcom ARM64-based SoCs:

- Ray adds the required bus width and OOB sector size properties to the
  Northstar 2 SVK reference board in order for the NAND controller to work
  properly

* tag 'arm-soc/for-4.9/devicetree-arm64-fixes' of http://github.com/Broadcom/stblinux:
  arm64: dts: Updated NAND DT properties for NS2 SVK

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'imx-fixes-4.9' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo...
Olof Johansson [Sat, 29 Oct 2016 18:08:50 +0000 (11:08 -0700)]
Merge tag 'imx-fixes-4.9' of git://git./linux/kernel/git/shawnguo/linux into fixes

The i.MX fixes for 4.9:
 - A couple of patches from Fabio to fix the GPC power domain regression
   which is caused by PM Domain core change 0159ec670763dd
   ("PM / Domains: Verify the PM domain is present when adding a
   provider"), and a related kernel crash seen with multi_v7_defconfig
   build.
 - Correct the PHY ID mask for AR8031 to match phy driver code.
 - Apply new added timer erratum A008585 for LS1043A and LS2080A SoC.
 - Correct vf610 global timer IRQ flag to avoid warning from gic driver
   after commit 992345a58e0c ("irqchip/gic: WARN if setting the
   interrupt type for a PPI fails").

* tag 'imx-fixes-4.9' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
  ARM: imx: mach-imx6q: Fix the PHY ID mask for AR8031
  ARM: dts: vf610: fix IRQ flag of global timer
  ARM: imx: gpc: Fix the imx_gpc_genpd_init() error path
  ARM: imx: gpc: Initialize all power domains
  arm64: dts: Add timer erratum property for LS2080A and LS1043A

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'uniphier-fixes-v4.9' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Sat, 29 Oct 2016 18:05:49 +0000 (11:05 -0700)]
Merge tag 'uniphier-fixes-v4.9' of git://git./linux/kernel/git/masahiroy/linux-uniphier into fixes

UniPhier ARM SoC fixes for v4.9

- Add "select ARCH_HAS_RESET_CONTROLLER" in Kconfig
- Rename wrongly-named mioctrl to sdctrl

* tag 'uniphier-fixes-v4.9' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-uniphier:
  arm64: dts: uniphier: change MIO node to SD control node
  ARM: dts: uniphier: change MIO node to SD control node
  reset: uniphier: rename MIO reset to SD reset for Pro5, PXs2, LD20 SoCs
  arm64: uniphier: select ARCH_HAS_RESET_CONTROLLER
  ARM: uniphier: select ARCH_HAS_RESET_CONTROLLER

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'driver-core-4.9-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 29 Oct 2016 17:57:40 +0000 (10:57 -0700)]
Merge tag 'driver-core-4.9-rc3' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are two small driver core / kernfs fixes for 4.9-rc3.

  One makes the Kconfig entry for DEBUG_TEST_DRIVER_REMOVE a bit more
  explicit that this is a crazy thing to enable for a distro kernel
  (thanks for trying Fedora!), the other resolves an issue with vim
  opening kernfs files (sysfs, configfs, etc.)

  Both have been in linux-next with no reported issues"

* tag 'driver-core-4.9-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  driver core: Make Kconfig text for DEBUG_TEST_DRIVER_REMOVE stronger
  kernfs: Add noop_fsync to supported kernfs_file_fops