RDMA/efa: Fix wrong resources deallocation order

When trying to destroy QP or CQ, we first decrease the refcount and
potentially free memory regions allocated for the object and then
request the dev

RDMA/efa: Fix wrong resources deallocation order

When trying to destroy QP or CQ, we first decrease the refcount and
potentially free memory regions allocated for the object and then
request the device to destroy the object. If the device fails, the
object isn't fully destroyed so the user/IB core can try to destroy the
object again which will lead to underflow when trying to decrease an
already zeroed refcount.

Deallocate resources in reverse order of allocating them to safely free
them.

Fixes: ff6629f88c52 ("RDMA/efa: Do not delay freeing of DMA pages")
Reviewed-by: Michael Margolin <mrgolin@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Yonatan Nachum <ynachum@amazon.com>
Link: https://lore.kernel.org/r/20230822082725.31719-1-ynachum@amazon.com
Signed-off-by: Leon Romanovsky <leon@kernel.org>

show more ...

RDMA/efa: Add RDMA write HW statistics counters

Update device API and request RDMA write counters if RDMA write is
supported by device. Expose newly added counters through ib core
counters mechanism

RDMA/efa: Add RDMA write HW statistics counters

Update device API and request RDMA write counters if RDMA write is
supported by device. Expose newly added counters through ib core
counters mechanism.

Reviewed-by: Daniel Kranzdorf <dkkranzd@amazon.com>
Reviewed-by: Yonatan Nachum <ynachum@amazon.com>
Signed-off-by: Michael Margolin <mrgolin@amazon.com>
Link: https://lore.kernel.org/r/20230703153404.30877-1-mrgolin@amazon.com
Reviewed-by: Gal Pressman <gal.pressman@linux.dev>
Signed-off-by: Leon Romanovsky <leon@kernel.org>

show more ...

RDMA/efa: Fix unsupported page sizes in device

Device uses 4KB size blocks for user pages indirect list while the
driver creates those blocks with the size of PAGE_SIZE of the kernel. On
kernels wit

RDMA/efa: Fix unsupported page sizes in device

Device uses 4KB size blocks for user pages indirect list while the
driver creates those blocks with the size of PAGE_SIZE of the kernel. On
kernels with PAGE_SIZE different than 4KB (ARM RHEL), this leads to a
failure on register MR with indirect list because of the miss
communication between driver and device.

Fixes: 40909f664d27 ("RDMA/efa: Add EFA verbs implementation")
Link: https://lore.kernel.org/r/20230511115103.13876-1-ynachum@amazon.com
Reviewed-by: Firas Jahjah <firasj@amazon.com>
Reviewed-by: Michael Margolin <mrgolin@amazon.com>
Signed-off-by: Yonatan Nachum <ynachum@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Add rdma write capability to device caps

Add rdma write capability that is propagated from the device to rdma-core.
Enable MR creation with remote write permissions according to this devic

RDMA/efa: Add rdma write capability to device caps

Add rdma write capability that is propagated from the device to rdma-core.
Enable MR creation with remote write permissions according to this device
capability.

Link: https://lore.kernel.org/r/20230404154313.35194-1-ynachum@amazon.com
Reviewed-by: Firas Jahjah <firasj@amazon.com>
Reviewed-by: Michael Margolin <mrgolin@amazon.com>
Signed-off-by: Yonatan Nachum <ynachum@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Add data polling capability feature bit

Add feature bit to existing device caps field. EFA supports data polling
of 128 bytes blocks.

The flag indicates that the NIC guarentees that a 128

RDMA/efa: Add data polling capability feature bit

Add feature bit to existing device caps field. EFA supports data polling
of 128 bytes blocks.

The flag indicates that the NIC guarentees that a 128 byte aligned block
is written in order, ie that observing the last 8 bits of the block mean
the prior 127 bytes are also written.

It is useful for "last data polling" acceleration techniques.

Link: https://lore.kernel.org/r/20230219081328.10419-1-mrgolin@amazon.com
Reviewed-by: Yehuda Yitschak <yehuday@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Yonatan Nachum <ynachum@amazon.com>
Signed-off-by: Michael Margolin <mrgolin@amazon.com>
Acked-by: Gal Pressman <gal.pressman@linux.dev>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Support CQ receive entries with source GID

Add a parameter for create CQ admin command to set source address on
receive completion descriptors. Report capability for this feature
through q

RDMA/efa: Support CQ receive entries with source GID

Add a parameter for create CQ admin command to set source address on
receive completion descriptors. Report capability for this feature
through query device verb.

Link: https://lore.kernel.org/r/20220818140449.414-1-mrgolin@amazon.com
Reviewed-by: Firas Jahjah <firasj@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Daniel Kranzdorf <dkkranzd@amazon.com>
Signed-off-by: Michael Margolin <mrgolin@amazon.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>

show more ...

RDMA/efa: Add support for dmabuf memory regions

Implement a dmabuf importer for the EFA driver. As ODP is not supported,
the pinned dmabuf are used to prevent the move_notify callback from being
cal

RDMA/efa: Add support for dmabuf memory regions

Implement a dmabuf importer for the EFA driver. As ODP is not supported,
the pinned dmabuf are used to prevent the move_notify callback from being
called.

Link: https://lore.kernel.org/r/20211012120903.96933-4-galpress@amazon.com
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/counter: Add a descriptor in struct rdma_hw_stats

Add a counter statistic descriptor structure in rdma_hw_stats. In addition
to the counter name, more meta-information will be added. This code

RDMA/counter: Add a descriptor in struct rdma_hw_stats

Add a counter statistic descriptor structure in rdma_hw_stats. In addition
to the counter name, more meta-information will be added. This code
extension is needed for optional-counter support in the following patches.

Link: https://lore.kernel.org/r/20211008122439.166063-4-markzhang@nvidia.com
Signed-off-by: Aharon Landau <aharonl@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Mark Zhang <markzhang@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: CQ notifications

This patch adds support for CQ notifications through the standard verbs
api.

In order to achieve that, a new event queue (EQ) object is introduced,
which is in charge of

RDMA/efa: CQ notifications

This patch adds support for CQ notifications through the standard verbs
api.

In order to achieve that, a new event queue (EQ) object is introduced,
which is in charge of reporting completion events to the driver. On
driver load, EQs are allocated and their affinity is set to a single
cpu. When a user app creates a CQ with a completion channel, the
completion vector number is converted to a EQ number, which is in charge
of reporting the CQ events.

In addition, the CQ creation admin command now returns an offset for the
CQ doorbell, which is mapped to the userspace provider and is used to arm
the CQ when requested by the user.

The EQs use a single doorbell (located on the registers BAR), which
encodes the EQ number and arm as part of the doorbell value. The EQs are
polled by the driver on each new EQE, and arm it when the poll is
completed.

Link: https://lore.kernel.org/r/20211003105605.29222-1-galpress@amazon.com
Reviewed-by: Firas JahJah <firasj@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Globally allocate and release QP memory

Convert QP object to follow IB/core general allocation scheme. That
change allows us to make sure that restrack properly kref the memory.

Link: https:

RDMA: Globally allocate and release QP memory

Convert QP object to follow IB/core general allocation scheme. That
change allows us to make sure that restrack properly kref the memory.

Link: https://lore.kernel.org/r/48e767124758aeecc433360ddd85eaa6325b34d9.1627040189.git.leonro@nvidia.com
Reviewed-by: Gal Pressman <galpress@amazon.com> #efa
Tested-by: Gal Pressman <galpress@amazon.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> #rdma and core
Tested-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Tested-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Remove double QP type assignment

The QP type is set by the IB/core and shouldn't be set in the driver.

Fixes: 40909f664d27 ("RDMA/efa: Add EFA verbs implementation")
Link: https://lore.ke

RDMA/efa: Remove double QP type assignment

The QP type is set by the IB/core and shouldn't be set in the driver.

Fixes: 40909f664d27 ("RDMA/efa: Add EFA verbs implementation")
Link: https://lore.kernel.org/r/838c40134c1590167b888ca06ad51071139ff2ae.1627040189.git.leonro@nvidia.com
Acked-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Split hardware stats to device and port stats

The hardware stats API distinguishes between device and port statistics,
split the EFA stats accordingly instead of always dumping everything.

RDMA/efa: Split hardware stats to device and port stats

The hardware stats API distinguishes between device and port statistics,
split the EFA stats accordingly instead of always dumping everything.

Link: https://lore.kernel.org/r/20210712105923.17389-1-galpress@amazon.com
Reviewed-by: Firas JahJah <firasj@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Split the alloc_hw_stats() ops to port and device variants

This is being used to implement both the port and device global stats,
which is causing some confusion in the drivers. For instance E

RDMA: Split the alloc_hw_stats() ops to port and device variants

This is being used to implement both the port and device global stats,
which is causing some confusion in the drivers. For instance EFA and i40iw
both seem to be misusing the device stats.

Split it into two ops so drivers that don't support one or the other can
leave the op NULL'd, making the calling code a little simpler to
understand.

Link: https://lore.kernel.org/r/1955c154197b2a159adc2dc97266ddc74afe420c.1623427137.git.leonro@nvidia.com
Tested-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Support more than 255 rdma ports

Current code uses many different types when dealing with a port of a RDMA
device: u8, unsigned int and u32. Switch to u32 to clean up the logic.

This allows u

RDMA: Support more than 255 rdma ports

Current code uses many different types when dealing with a port of a RDMA
device: u8, unsigned int and u32. Switch to u32 to clean up the logic.

This allows us to make (at least) the core view consistent and use the
same type. Unfortunately not all places can be converted. Many uverbs
functions expect port to be u8 so keep those places in order not to break
UAPIs. HW/Spec defined values must also not be changed.

With the switch to u32 we now can support devices with more than 255
ports. U32_MAX is reserved to make control logic a bit easier to deal
with. As a device with U32_MAX ports probably isn't going to happen any
time soon this seems like a non issue.

When a device with more than 255 ports is created uverbs will report the
RDMA device as having 255 ports as this is the max currently supported.

The verbs interface is not changed yet because the IBTA spec limits the
port size in too many places to be u8 and all applications that relies in
verbs won't be able to cope with this change. At this stage, we are
extending the interfaces that are using vendor channel solely

Once the limitation is lifted mlx5 in switchdev mode will be able to have
thousands of SFs created by the device. As the only instance of an RDMA
device that reports more than 255 ports will be a representor device and
it exposes itself as a RAW Ethernet only device CM/MAD/IPoIB and other
ULPs aren't effected by this change and their sysfs/interfaces that are
exposes to userspace can remain unchanged.

While here cleanup some alignment issues and remove unneeded sanity
checks (mainly in rdmavt),

Link: https://lore.kernel.org/r/20210301070420.439400-1-leon@kernel.org
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Check flags during create_cq

Each driver should check that the CQ attrs is supported. Unfortuantely
when flags was added to the CQ attrs the drivers were not updated,
uverbs_ex_cmd_mask was us

RDMA: Check flags during create_cq

Each driver should check that the CQ attrs is supported. Unfortuantely
when flags was added to the CQ attrs the drivers were not updated,
uverbs_ex_cmd_mask was used to block it. This was missed when create CQ
was converted to ioctl, so non-zero flags could have been passed into
drivers.

Check that flags is zero in all drivers that don't use it, remove
IB_USER_VERBS_EX_CMD_CREATE_CQ from uverbs_ex_cmd_mask.

Fixes: 41b2a71fc848 ("IB/uverbs: Move ioctl path of create_cq and destroy_cq to a new file")
Link: https://lore.kernel.org/r/7-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Check attr_mask during modify_qp

Each driver should check that it can support the provided attr_mask during
modify_qp. IB_USER_VERBS_EX_CMD_MODIFY_QP was being used to block
modify_qp_ex becau

RDMA: Check attr_mask during modify_qp

Each driver should check that it can support the provided attr_mask during
modify_qp. IB_USER_VERBS_EX_CMD_MODIFY_QP was being used to block
modify_qp_ex because the driver didn't check RATE_LIMIT.

Link: https://lore.kernel.org/r/6-v1-caa70ba3d1ab+1436e-ucmd_mask_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Remove double QP type assignment

[ Upstream commit f9193d266347fe9bed5c173e7a1bf96268142a79 ]

The QP type is set by the IB/core and shouldn't be set in the driver.

Fixes: 40909f664d27 ("

RDMA/efa: Remove double QP type assignment

[ Upstream commit f9193d266347fe9bed5c173e7a1bf96268142a79 ]

The QP type is set by the IB/core and shouldn't be set in the driver.

Fixes: 40909f664d27 ("RDMA/efa: Add EFA verbs implementation")
Link: https://lore.kernel.org/r/838c40134c1590167b888ca06ad51071139ff2ae.1627040189.git.leonro@nvidia.com
Acked-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

show more ...

RDMA/efa: Use the correct current and new states in modify QP

The local variables cur_state and new_state hold the state that should be
used for the modify QP operation instead of the ones in the ib

RDMA/efa: Use the correct current and new states in modify QP

The local variables cur_state and new_state hold the state that should be
used for the modify QP operation instead of the ones in the ib_qp_attr
struct.

Fixes: 40909f664d27 ("RDMA/efa: Add EFA verbs implementation")
Link: https://lore.kernel.org/r/20201201091724.37016-1-galpress@amazon.com
Reviewed-by: Firas JahJah <firasj@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Drop double zeroing for sg_init_table()

sg_init_table() zeroes its first argument, so the allocation of that
argument doesn't have to.

The semantic patch that makes this change is as foll

RDMA/efa: Drop double zeroing for sg_init_table()

sg_init_table() zeroes its first argument, so the allocation of that
argument doesn't have to.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression x,n,flags;
@@

x =
- kcalloc
+ kmalloc_array
(n,sizeof(*x),flags)
...
sg_init_table(x,n)
// </smpl>

Link: https://lore.kernel.org/r/1600601186-7420-6-git-send-email-Julia.Lawall@inria.fr
Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Acked-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Add messages and RDMA read work requests HW stats

Add separate stats types for send messages and RDMA read work requests.

Link: https://lore.kernel.org/r/20200915141449.8428-3-galpress@am

RDMA/efa: Add messages and RDMA read work requests HW stats

Add separate stats types for send messages and RDMA read work requests.

Link: https://lore.kernel.org/r/20200915141449.8428-3-galpress@amazon.com
Signed-off-by: Daniel Kranzdorf <dkkranzd@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Group keep alive received counter with other SW stats

The keep alive received counter is a software stat, keep it grouped with
all other software stats. Since all stored stats are softwar

RDMA/efa: Group keep alive received counter with other SW stats

The keep alive received counter is a software stat, keep it grouped with
all other software stats. Since all stored stats are software stats,
remove the efa_sw_stats struct and use efa_stats instead.

Link: https://lore.kernel.org/r/20200915141449.8428-2-galpress@amazon.com
Reviewed-by: Daniel Kranzdorf <dkkranzd@amazon.com>
Reviewed-by: Yossi Leybovich <sleybo@amazon.com>
Signed-off-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/efa: Use ib_umem_num_dma_pages()

If ib_umem_find_best_pgsz() returns > PAGE_SIZE then the equation here is
not correct. 'start' should be 'virt'. Change it to use the core code for
page_num and

RDMA/efa: Use ib_umem_num_dma_pages()

If ib_umem_find_best_pgsz() returns > PAGE_SIZE then the equation here is
not correct. 'start' should be 'virt'. Change it to use the core code for
page_num and the canonical calculation of page_shift.

Fixes: 40ddb3f02083 ("RDMA/efa: Use API to get contiguous memory blocks aligned to device supported page size")
Link: https://lore.kernel.org/r/7-v2-270386b7e60b+28f4-umem_1_jgg@nvidia.com
Tested-by: Gal Pressman <galpress@amazon.com>
Acked-by: Gal Pressman <galpress@amazon.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA/umem: Add rdma_umem_for_each_dma_block()

This helper does the same as rdma_for_each_block(), except it works on a
umem. This simplifies most of the call sites.

Link: https://lore.kernel.org/r/

RDMA/umem: Add rdma_umem_for_each_dma_block()

This helper does the same as rdma_for_each_block(), except it works on a
umem. This simplifies most of the call sites.

Link: https://lore.kernel.org/r/4-v2-270386b7e60b+28f4-umem_1_jgg@nvidia.com
Acked-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Acked-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Allow fail of destroy CQ

Like any other verbs objects, CQ shouldn't fail during destroy, but
mlx5_ib didn't follow this contract with mixed IB verbs objects with
DEVX. Such mix causes to the s

RDMA: Allow fail of destroy CQ

Like any other verbs objects, CQ shouldn't fail during destroy, but
mlx5_ib didn't follow this contract with mixed IB verbs objects with
DEVX. Such mix causes to the situation where FW and kernel are fully
interdependent on the reference counting of each side.

Kernel verbs and drivers that don't have DEVX flows shouldn't fail.

Fixes: e39afe3d6dbd ("RDMA: Convert CQ allocations to be under core responsibility")
Link: https://lore.kernel.org/r/20200907120921.476363-7-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...

RDMA: Restore ability to fail on AH destroy

Like any other IB verbs objects, AH are refcounted by ib_core. The release
of those objects are controlled by ib_core with promise that AH destroy
can't f

RDMA: Restore ability to fail on AH destroy

Like any other IB verbs objects, AH are refcounted by ib_core. The release
of those objects are controlled by ib_core with promise that AH destroy
can't fail.

Being SW object for now, this change makes dealloc_ah() to behave like any
other destroy IB flows.

Fixes: d345691471b4 ("RDMA: Handle AH allocations by IB/core")
Link: https://lore.kernel.org/r/20200907120921.476363-3-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

show more ...