| Age | Commit message (Collapse) | Author |
|---|
|
The internal design of RDMA/core ensures that there dealloc ucontext will
be called only if alloc_ucontext succeeded, hence there is no need to
manage internal variable to mark validity of ucontext.
As part of this change, remove redundant memeset too.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The following build warning was produced for the TID RDMA READ
patch ("IB/hfi1: Enable TID RDMA READ protocol"):
drivers/infiniband/hw/hfi1/qp.c: In function 'hfi1_setup_wqe':
drivers/infiniband/hw/hfi1/qp.c:328:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
hfi1_setup_tid_rdma_wqe(qp, wqe);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/infiniband/hw/hfi1/qp.c:329:2: note: here
case IB_QPT_UC:
^~~~
This patch will fix the issue by adding the "fall through" comment.
Fixes: f1ab4efa6d32 ("IB/hfi1: Enable TID RDMA READ protocol")
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Kaike Wan <kaike.wan@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The new output_written block was wrongly placed before the ret=0, causing
the error code to be lost. uverbs_output_written is not expected to fail,
and even if it does fail it has no significant impact on the userspace
flow.
Reported-by: Bart Van Assche <bvanassche@acm.org>
Fixes: d6f4a21f309d ("RDMA/uverbs: Mark ioctl responses with UVERBS_ATTR_F_VALID_OUTPUT")
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
|
|
Now when we have the udata passed to all the ib_xxx object creation APIs
and the additional macro 'rdma_udata_to_drv_context' to get the
ib_ucontext from ib_udata stored in uverbs_attr_bundle, we can finally
start to remove the dependency of the drivers in the
ib_xxx->uobject->context.
Signed-off-by: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Helper function to get driver's context out of ib_udata wrapped in
uverbs_attr_bundle for user objects or NULL for kernel objects.
Signed-off-by: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Add ib_ucontext to the uverbs_attr_bundle sent down the iocl and cmd flows
as soon as the flow has ib_uobject.
In addition, remove rdma_get_ucontext helper function that is only used by
ib_umem_get.
Signed-off-by: Shamir Rabinovitch <shamir.rabinovitch@oracle.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Changed debug statements to use %s and __func__ instead
of hard-coded function's name.
Signed-off-by: Erez Alfasi <ereza@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/infiniband/core/iwpm_util.c: In function 'iwpm_send_hello':
drivers/infiniband/core/iwpm_util.c:811:6: warning:
variable 'msg_seq' set but not used [-Wunused-but-set-variable]
It never used since introduction in commit b0bad9ad514f ("RDMA/IWPM:
Support no port mapping requirements")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
This patch adds new device capability for IB_DEVICE_MEM_MGT_EXTENSIONS to
indicate device support for the following features:
1. Fast register memory region.
2. send with remote invalidate by frmr
3. local invalidate memory regsion
As well as adds the max depth of frmr page list len.
Signed-off-by: Yangyang Li <liyangyang20@huawei.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Current all messages printed for aeq subtype event are wrong. Thus,
delete them and only the value of subtype event is printed.
Signed-off-by: Yixian Liu <liuyixian@huawei.com>
Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The memory allocated for wrid should be initialized to zero.
Signed-off-by: Yixian Liu <liuyixian@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The state of mr after reregister operation should be set to valid
state. Otherwise, it will keep the same as the state before reregistered.
Signed-off-by: Yixian Liu <liuyixian@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
This patch modifies the minimum CQ depth specification of hip08 and is
consistent with the processing of hip06.
Signed-off-by: chenglang <chenglang@huawei.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Driver now supports new link modes: 50Gbps per lane support for
50G/100G/200G. This patch reads the correct field (legacy vs. extended)
based on a FW indication bit, and adds a translation function (link
modes to IB width and speed) to the new link modes.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
This patch exposes new link modes (including 50Gbps per lane), and ext_*
fields which describes the new link modes in Port Type and Speed
register (PTYS).
Access functions, translation functions (speed <-> HW bits) and
link max speed function were modified.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Register Port Type and Speed (PTYS) introduces three new fields
extending the speed/protocols the can be reported and configured.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
This patch fascicles queries to speed related fields in Port Type and
Speed register (PTYS) into a single API. I addition, this patch
refactors functions which serves only Ethernet driver: remove the
protocol type as an input parameter, move code from 'core' directory
into 'en' directory and add 'eth' prefix to the function's name. The
patch also encapsulates functions that are not used outside the Ethernet
driver removes redundant include files.
Signed-off-by: Aya Levin <ayal@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
When dealing with the offloads mode initialization, driver refers to
the number of VFs and add magic number one (1) to take account of the
uplink. This is not clear and will make the code less readable after
adding other vports (e.g. host PF). As these are special vports
compared to VF vports, add a helper macro to denote such special
vports and eliminate the use of magic number.
Moreover, when creating offloads flow table and groups, the driver
reserves two more slots for UC and MC miss rules. Replace this magic
number with a helper macro as well.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
These are two macros in the driver general header which deal with the
number of total vports and if a vport is vport manager. Such macros
are vport entities, better to place them at the vport header file.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Driver used to name uplink vport as FDB_UPLINK_VPORT, it's hard to
comply with the same naming convention along with the introduction of
other vports. Use MLX5_VPORT as the prefix for such vports and
relocate the uplink vport definition to public header file for the
benefits of both net and IB drivers.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
ECPF doesn't support SR-IOV, but an ECPF E-Switch manager shall know
the max VFs supported by its peer host PF in order to control those
VF vports.
The current driver implementation uses the total vfs quantity as
provided by the pci sub-system for an upper bound of the VF vports
the e-switch code needs to deal with. This obviously can't work as
is on ECPF e-switch manager. For now, we use a hard coded value of
128 on such systems.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
In Embedded CPU (EC) configurations, the EC driver needs to know when
the number of virtual functions change on the corresponding PF at the
host side. This is required so the EC driver can create or destroy
representor net devices that represent the VFs ports.
Whenever a change in the number of VFs occurs, firmware will generate an
event towards the EC which will trigger a work to complete the rest of
the handling. The specifics of the handling will be introduced in a
downstream patch.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
The QUERY_HOST_PARAMS command is used by an Embedded CPU Physical
Function (ECPF) driver to identify and retrieve information about the
PF on the host side. E.g, number of virtual functions and PCI BDF.
The number of VFs can be changed on the fly, a function is added to
query current number of VFs and will be used in downstream patches.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
With the introduction of ECPF, we require that the ECPF driver will
aways call enable/disable HCA for that PF in the same way a PF does
this for its VFs. The PF is still responsible for calling enable and
disable HCA for its VFs.
To distinguish between the ECPF executing enable/disable HCA for
itself or for the PF, it sets the embedded CPU function bit in the
input params struct of these commands. When the bit is cleared and
function ID is zero, it refers to the peer PF.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Mellanox's SmartNIC combines embedded CPU(e.g, ARM) processing power
with advanced network offloads to accelerate a multitude of security,
networking and storage applications.
With the introduction of the SmartNIC, there is a new PCI function
called Embedded CPU Physical Function(ECPF). And it's possible for a
PF to get its ICM pages from the ECPF PCI function. Driver shall
identify if it is running on such a function by reading a bit in
the initialization segment.
When firmware asks for pages, it would issue a page request event
specifying how many pages it requests and for which function. That
driver responds with a manage_pages command providing the requested
pages along with an indication for which function it is providing these
pages.
The encoding before this patch was as follows:
function_id == 0: pages are requested for the function receiving
the EQE.
function_id != 0: pages are requested for VF identified by the
function_id value
A new one bit field in the EQE identifies that pages are requested for
the ECPF.
The notion of page_supplier can be introduced here and to support that,
manage pages and query pages were modified so firmware can distinguish
the following cases:
1. Function provides pages for itself
2. PF provides pages for its VF
3. ECPF provides pages to itself
4. ECPF provides pages for another function
This distinction is possible through the introduction of the bit
"embedded_cpu_function" in query_pages, manage_pages and page request
EQE.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
IB driver maintains different registration and load function calls
for uplink and VF vports. This is not necessary as they only differ
with each other on their profiles.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Use u16 for vport number, which matches how hardware refers to this
argument throughout commands.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Better to use void * and avoid unnecessary casts.
This patch doesn't change any functionality.
Signed-off-by: Bodong Wang <bodong@mellanox.com>
Signed-off-by: Eli Cohen <eli@mellanox.com>
Reviewed-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Update newly introduced function to be aligned to netdev coding style.
Fixes: 46861e3e88be ("net/mlx5: Set ODP SRQ support in firmware")
Reported-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
rdmavt expects a uniform size on all umem SGEs which is currently at
PAGE_SIZE.
Adapt to a umem API change which could return non-uniform sized SGEs due
to combining contiguous PAGE_SIZE regions into an SGE. Use
for_each_sg_page variant to unfold the larger SGEs into a list of
PAGE_SIZE elements.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Reviewed-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The null check on an allocation failure on pd is currently checking
if pd is non-null rather than null. Fix this by adding the missing !
operator.
Fixes: 21a428a019c9 ("RDMA: Handle PD allocations by IB/core")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma into for-next
I had merged the hfi1-tid code into my local copy of for-next, but was
waiting on 0day testing before pushing it (I pushed it to my wip
branch). Having waited several days for 0day testing to show up, I'm
finally just going to push it out. In the meantime, though, Jason
pushed other stuff to for-next, so I needed to merge up the branches
before pushing.
Signed-off-by: Doug Ledford <dledford@redhat.com>
|
|
mlx5_eq_cq_get() is called in IRQ handler, the spinlock inside
gets a lot of contentions when we test some heavy workload
with 60 RX queues and 80 CPU's, and it is clearly shown in the
flame graph.
In fact, radix_tree_lookup() is perfectly fine with RCU read lock,
we don't have to take a spinlock on this hot path. This is pretty
much similar to commit 291c566a2891
("net/mlx4_core: Fix racy CQ (Completion Queue) free"). Slow paths
are still serialized with the spinlock, and with synchronize_irq()
it should be safe to just move the fast path to RCU read lock.
This patch itself reduces the latency by about 50% for our memcached
workload on a 4.14 kernel we test. In upstream, as pointed out by Saeed,
this spinlock gets some rework in commit 02d92f790364
("net/mlx5: CQ Database per EQ"), so the difference could be smaller.
Cc: Saeed Mahameed <saeedm@mellanox.com>
Cc: Tariq Toukan <tariqt@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
The struct member comp_mask has not been initialized however a bit
pattern is being bitwise or'd into the member and hence other bit
fields in comp_mask may contain any garbage from the stack. Fix this
by making the bitwise or into an assignment.
Fixes: 95b86d1c91ad ("RDMA/bnxt_re: Update kernel user abi to pass chip context")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Devesh Sharma <devesh.sharma@broadcom.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Make sure the IB device is freed on failure.
Fixes: b5ca15ad7e61 ("IB/mlx5: Add proper representors support")
Signed-off-by: Mark Bloch <markb@mellanox.com>
Reviewed-by: Bodong Wang <bodong@mellanox.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Fix bad flow upon DEVX mkey creation to prevent deleting the indirect mkey
from the radix tree in case there was a previous failure to insert it.
Fixes: 534fd7aac56a ("IB/mlx5: Manage indirection mkey upon DEVX flow for ODP")
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Reviewed-by: Artemy Kovalyov <artemyko@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
The driver walks the umem SGL assuming a 1:1 mapping between SGE and
system page. Update to use the for_each_sg_page iterator to get individual
pages contained in the SGEs. This is a pre-requisite before adding page
combining into SGEs while building the scatter table in IB core.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Acked-by: Michal Kalderon <michal.kalderon@marvell.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Acked-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Acked-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Use the for_each_sg_dma_page iterator variant to walk the umem DMA-mapped
SGL and get the page DMA address. This avoids the extra loop to iterate
pages in the SGE when for_each_sg iterator is used.
Additionally, purge umem->page_shift usage in the driver as its only
relevant for ODP MRs. Use system page size and shift instead.
Signed-off-by: Shiraz, Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Commit 2db76d7c3c6d ("lib/scatterlist: sg_page_iter: support sg lists w/o
backing pages") introduced the sg_page_iter_dma_address() function without
providing a way to use it in the general case. If the sg_dma_len() is not
equal to the sg length callers cannot safely use the
for_each_sg_page/sg_page_iter_dma_address combination.
Resolve this API mistake by providing a DMA specific iterator,
for_each_sg_dma_page(), that uses the right length so
sg_page_iter_dma_address() works as expected with all sglists.
A new iterator type is introduced to provide compile-time safety against
wrongly mixing accessors and iterators.
Acked-by: Christoph Hellwig <hch@lst.de> (for scatterlist)
Acked-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com> (ipu3-cio2)
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>
|
|
Due to concurrent work by myself and Jason, a normal fast forward merge
was not possible. This brings in a number of hfi1 changes, mainly the
hfi1 TID RDMA support (roughly 10,000 LOC change), which was reviewed
and integrated over a period of days.
Signed-off-by: Doug Ledford <dledford@redhat.com>
|
|
Omni-Path TID RDMA Feature
Intel Omni-Path (OPA) TID RDMA support is a feature that accelerates
data movement between two OPA nodes through the IB Verbs interface. It
improves RDMA READ/WRITE performance by delivering the data payload to a
user buffer directly without any software copying.
Architecture
=============
The TID RDMA protocol is implemented on the hfi1 driver level and is
therefore transparent to the ULPs. It is designed to facilitate the data
transactions for two specific RDMA requests:
- RDMA READ;
- RDMA WRITE.
Previously, when a verbs data packet is received at the destination
(requester side for RDMA READ and responder side for RDMA WRITE), the
data payload is copied to the user buffer by software, which slows down
the performance significantly for large requests.
Internally, hfi1 converts qualified RDMA READ/WRITE requests into TID
RDMA READ/WRITE requests when the requests are post sent to the hfi1
driver. Non-qualified RDMA requests are handled by normal RDMA protocol.
For TID RDMA requests, hardware resources (hardware flow and TID entries)
are allocated on the destination side (the requester side for TID RDMA
READ and the responder side for TID RDMA WRITE). The information for
these resources is conveyed to the data source side (the responder side
for TID RDMA READ and the requester side for TID RDMA WRITE) and embedded
in data packets. When data packets are received by the destination,
hardware will deliver the data payload to the destination buffer without
involving software and therefore improve the performance.
Details
=======
RDMA READ/WRITE requests are qualified by the following:
- Total data length >= 256k;
- Totoal data length is a multiple of 4K pages.
Additional qualifications are enforced for the destination buffers:
For RDMA RAED:
- Each destination sge buffer is 4K aligned;
- Each destination sge buffer is a multiple of 4K pages.
For RDMA WRITE:
- The destination number is 4K aligned.
In addition, in an OPA fabric, some nodes may support TID RDMA while
others may not. As such, it is important for two transaction nodes to
exchange the information about the features they support. This discovery
mechanism is called OPA Feature Negotion (OPFN) and is described in
details in the patch series. Through OPFN, two nodes can find whether
they both support TID RDMA and subsequently convert RDMA requests into
TID RDMA requests.
* hfi1-tid: (46 commits)
IB/hfi1: Prioritize the sending of ACK packets
IB/hfi1: Add static trace for TID RDMA WRITE protocol
IB/hfi1: Enable TID RDMA WRITE protocol
IB/hfi1: Add interlock between TID RDMA WRITE and other requests
IB/hfi1: Add TID RDMA WRITE functionality into RDMA verbs
IB/hfi1: Add the dual leg code
IB/hfi1: Add the TID second leg ACK packet builder
IB/hfi1: Add the TID second leg send packet builder
IB/hfi1: Resend the TID RDMA WRITE DATA packets
IB/hfi1: Add a function to receive TID RDMA RESYNC packet
IB/hfi1: Add a function to build TID RDMA RESYNC packet
IB/hfi1: Add TID RDMA retry timer
IB/hfi1: Add a function to receive TID RDMA ACK packet
IB/hfi1: Add a function to build TID RDMA ACK packet
IB/hfi1: Add a function to receive TID RDMA WRITE DATA packet
IB/hfi1: Add a function to build TID RDMA WRITE DATA packet
IB/hfi1: Add a function to receive TID RDMA WRITE response
IB/hfi1: Add TID resource timer
IB/hfi1: Add a function to build TID RDMA WRITE response
IB/hfi1: Add functions to receive TID RDMA WRITE request
...
Signed-off-by: Doug Ledford <dledford@redhat.com>
|
