As all vdev resources are allocated before we boot the remote processor
we no longer need to support modifying the resource table while the
remote is running.

This saves us from the table_ptr dance, but more importantly allow the
remote processor to enable security lock down of the loaded table memory
region.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

The vdev handler is now just another resource allocator, so handle all
resource types in a single pass.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Vrings are now allocated as we parse the resource table, before we
boot the rproc or register any virtio devices, so it's safe to bump
max_notifyid as part of this process.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Represent the virtio device part of the vdev resources as remoteproc
subdevices to finalize the decoupling of the virtio resource and device
handling.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Tie the vdev (and hence vring) life cycle to the resource parsing and
resource cleanup operations, allowing us to safely register and
unregister virtio devices on the go.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Instead of having the vrings being allocated and freed as they are
requested by the virtio device tie their life cycle to the vdev
resource. This allows us to decouple the vdev resource management from
the virtio device management.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

No functional change
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Allow the wcnss smd edge to be described as a child of the wcnss
remoteproc node and make the edge life cycle follow the running state of
the remoteproc.

This bond is necessary to clean up the smd state when the remote
processor is suddenly removed, and in some cases even when it shut down
in a controlled fasion.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

The edge registration functions is to be used from a remoteproc driver
to register and unregister an edge as the remote processor comes and
goes.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

A subdevice is an abstract entity that can be used to tie actions to the
booting and shutting down of a remote processor. The subdevice object is
expected to be embedded in concrete implementations, allowing for a
variety of use cases to be implemented.
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Fix to return a negative error code from the st_rproc_state() error
handling case instead of 0, as done elsewhere in this function.

Fixes: 63edb031 ("remoteproc: Supply controller driver for ST's Remote Processors")
Signed-off-by: NWei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Since there is now an always available state file in sysfs with the same
function as this one in debugfs, remove the redundant entry.
Signed-off-by: NMatt Redfearn <matt.redfearn@imgtec.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

This patch adds a sysfs interface to rproc allowing the firmware name
and processor state to be changed dynamically.

State was previously available in debugfs, and is replicated here. The
firmware file allows retrieval of the running firmware name, and a new
one to be specified at run time, so long as the remote processor has
been stopped.
Signed-off-by: NMatt Redfearn <matt.redfearn@imgtec.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

If the driver is built as a module, autoload won't work because the module
alias information is not filled. So user-space can't match the registered
device with the corresponding module.

Export the module alias information using the MODULE_DEVICE_TABLE() macro.

Before this patch:

$ modinfo drivers/remoteproc/qcom_wcnss_iris.ko | grep alias
$

After this patch:

$ modinfo drivers/remoteproc/qcom_wcnss_iris.ko | grep alias
alias:          of:N*T*Cqcom,wcn3680C*
alias:          of:N*T*Cqcom,wcn3680
alias:          of:N*T*Cqcom,wcn3660C*
alias:          of:N*T*Cqcom,wcn3660
alias:          of:N*T*Cqcom,wcn3620C*
alias:          of:N*T*Cqcom,wcn3620
Signed-off-by: NJavier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

If the driver is built as a module, autoload won't work because the module
alias information is not filled. So user-space can't match the registered
device with the corresponding module.

Export the module alias information using the MODULE_DEVICE_TABLE() macro.

Before this patch:

$ modinfo drivers/remoteproc/qcom_wcnss.ko | grep alias
$

After this patch:

$ modinfo drivers/remoteproc/qcom_wcnss.ko | grep alias
alias:          of:N*T*Cqcom,pronto-v2-pilC*
alias:          of:N*T*Cqcom,pronto-v2-pil
alias:          of:N*T*Cqcom,pronto-v1-pilC*
alias:          of:N*T*Cqcom,pronto-v1-pil
alias:          of:N*T*Cqcom,riva-pilC*
alias:          of:N*T*Cqcom,riva-pil
Signed-off-by: NJavier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

If the driver is built as a module, autoload won't work because the module
alias information is not filled. So user-space can't match the registered
device with the corresponding module.

Export the module alias information using the MODULE_DEVICE_TABLE() macro.

Before this patch:

$ modinfo drivers/remoteproc/qcom_q6v5_pil.ko | grep alias
$

After this patch:

$ modinfo drivers/remoteproc/qcom_q6v5_pil.ko | grep alias
alias:          of:N*T*Cqcom,q6v5-pilC*
alias:          of:N*T*Cqcom,q6v5-pil
Signed-off-by: NJavier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Storage of the firmware name was inconsistent, either storing a pointer
to a name stored with unknown ownership, or a variable length tacked
onto the end of the struct proc allocated in rproc_alloc.

In preparation for allowing the firmware of an already allocated struct
rproc to be changed, instead always keep a locally maintained copy of
the firmware name.
Signed-off-by: NMatt Redfearn <matt.redfearn@imgtec.com>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>

Add error handling support.
Register ib device with ib stack.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for GSI over light L2.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add light L2 interface for RoCE.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Implement fastpath verbs like ib_send_post, ib_post_recv and ib_poll_cq.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for user, dma and memory regions registration.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for Queue Pair verbs which adds, deletes,
modifies and queries Queue Pairs.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for protection domain and completion queue verbs.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Add support for ucontext, query port, add and del gid verbs.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Allocate and setup RoCE resources, interrupts and completion queues.
Adds device attributes.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

Adds a skeletal implementation of the qed* RoCE driver -
basically the ability to communicate with the qede driver and
receive notifications from it regarding various init/exit events.
Signed-off-by: NRajesh Borundia <rajesh.borundia@cavium.com>
Signed-off-by: NRam Amrani <Ram.Amrani@cavium.com>
Signed-off-by: NDoug Ledford <dledford@redhat.com>

The desired_perf is an abstract performance number. Its value should
be in the range of [lowest perf, highest perf] of CPPC.
The correct calculation is
  desired_perf = freq * cppc_highest_perf / cppc_dmi_max_khz

And cppc_cpufreq_set_target() returns if desired_perf is exactly
the same with the old perf.
Signed-off-by: NHoan Tran <hotran@apm.com>
Reviewed-by: NPrashanth Prakash <pprakash@codeaurora.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

The dev parameter passed to __axienet_device_reset() is not used inside
the function, so remove it.
Signed-off-by: NTobias Klauser <tklauser@distanz.ch>
Reviewed-by: NMichal Simek <michal.simek@xilinx.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is supposed to loop 1000 times and then give up.  The problem is
it's a post-op and after the loop we test if "loop" is zero when really
it would be -1.  Fix this by making it a pre-op.

Fixes: 1b7c55c4 ("liquidio: CN23XX queue manipulation")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The phy_start() is used to indicate the PHY is now ready to do its
work. The state is changed, normally to PHY_UP which means that both
the MAC and the PHY are ready.

If the phy driver is using polling, when the next poll happens, the
state machine notices the PHY is now in PHY_UP, and kicks off
auto-negotiation, if needed.

If however, the PHY is using interrupts, there is no polling. The phy
is stuck in PHY_UP until the next interrupt comes along. And there is
no reason for the PHY to interrupt.

Have phy_start() schedule the state machine to run, which both speeds
up the polling use case, and makes the interrupt use case actually
work.

This problems exists whenever there is a state change which will not
cause an interrupt. Trigger the state machine in these cases,
e.g. phy_error().
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Cc: Kyle Roeschley <kyle.roeschley@ni.com>
Tested-by: NKyle Roeschley <kyle.roeschley@ni.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Based upon v2 of Stephen's patch.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Wiht the latest rework of the xen-netback driver, we get a warning
on ARM about the types passed into min():

drivers/net/xen-netback/rx.c: In function 'xenvif_rx_next_chunk':
include/linux/kernel.h:739:16: error: comparison of distinct pointer types lacks a cast [-Werror]

The reason is that XEN_PAGE_SIZE is not size_t here. There
is no actual bug, and we can easily avoid the warning using the
min_t() macro instead of min().

Fixes: eb1723a2 ("xen-netback: refactor guest rx")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NPaul Durrant <paul.durrant@citrix.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch fixes a problem when propagated the
failure of ptp_clock_register to open function.
Signed-off-by: NGiuseppe Cavallaro <peppe.cavallaro@st.com>
Cc: Alexandre TORGUE <alexandre.torgue@st.com>
Cc: Rayagond Kokatanur <rayagond@vayavyalabs.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The gmac 4.x version has not extended descriptors
(that are available on 3.x instead of).
While initializing the PTP module, the advanced PTP was
enabled in case of extended descriptors. This cannot be
applied for 4.x version where only the hardware capability
register has to show if the feature is present.
Patch also adds some extra netdev_(debug/inof) to better
dump the configuration.
Signed-off-by: NGiuseppe Cavallaro <peppe.cavallaro@st.com>
Cc: Alexandre TORGUE <alexandre.torgue@st.com>
Cc: Rayagond Kokatanur <rayagond@vayavyalabs.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The definition of qed_get_rdma_ops() is not a prototype unless
we add 'void' here, as indicated by this W=1 warning:

drivers/net/ethernet/qlogic/qed/qed_roce.c: In function ‘qed_get_rdma_ops’:
drivers/net/ethernet/qlogic/qed/qed_roce.c:2950:28: error: old-style function definition [-Werror=old-style-definition]

Fixes: abd49676 ("qed: Add RoCE ll2 & GSI support")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NYuval Mintz <Yuval.Mintz@caviumnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The software calculation of UDP checksum in Netvsc driver was
only handling IPv4 case. By using skb_checksum_help() instead
all protocols can be handled. Rearrange code to eliminate goto
and look like other drivers.

This is a temporary solution; recent versions of Window Server etc
do support UDP checksum offload, just need to do the appropriate negotiation
with host to validate before using. This will be done in later patch.

Please queue this for -stable as well.
Signed-off-by: NStephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a VIF has been ready for rx_stall_timeout (60s by default) and an
Rx ring is drained of all requests an Rx stall will be incorrectly
detected.  When this occurs and the guest Rx queue is empty, the Rx
ring's event index will not be set and the frontend will not raise an
event when new requests are placed on the ring, permanently stalling
the VIF.

This is a regression introduced by eb1723a2 (xen-netback:
refactor guest rx).

Fix this by reinstating the setting of queue->last_rx_time when
placing a packet onto the guest Rx ring.
Signed-off-by: NDavid Vrabel <david.vrabel@citrix.com>
Reviewed-by: NPaul Durrant <paul.durrant@citrix.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The reserved field precise_offset->rsv is not cleared before being
copied to user space, leaking kernel stack memory. Clear the struct
before it's copied.
Signed-off-by: NVlad Tsyrklevich <vlad@tsyrklevich.net>
Acked-by: NRichard Cochran <richardcochran@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

I am hitting this in mlx5:

drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c: In function
reclaim_pages_cmd.clone.0:
drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c:346: error: call
to __compiletime_assert_346 declared with attribute error:
BUILD_BUG_ON failed: __mlx5_bit_off(manage_pages_out, pas[i]) % 64
drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c: In function give_pages:
drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c:291: error: call
to __compiletime_assert_291 declared with attribute error:
BUILD_BUG_ON failed: __mlx5_bit_off(manage_pages_in, pas[i]) % 64

Problem is that this is doing a BUILD_BUG_ON on a non-constant
expression because of trying to take offset of pas[i] in the
structure.

Fix is to create MLX5_ARRAY_SET64 that takes an additional argument
that is the field index to separate between BUILD_BUG_ON on the array
constant field and the indexed field to assign the value to.
There are two callers of MLX5_SET64 that are trying to get a variable
offset, change those to call MLX5_ARRAY_SET64 passing 'pas' and 'i'
as the arguments to use in the offset check and the indexed value
assignment.

Fixes: a533ed5e ("net/mlx5: Pages management commands via mlx5 ifc")
Signed-off-by: NTom Herbert <tom@herbertland.com>
Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>