Add the QRTR gunyah transport that facilitates sending and receiving
IPC Router messages between virtual machines that use the haven
hypervisor.
Change-Id: I805ee5b144c5a8fc92e4a045061770d06495f63c
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
To support high priority clients like sensor use cases in all system
conditions elevating the router reader threads to RT priorities based
on configuration.
Change-Id: I6f46097f6678f2e00c3ec8872b208d773c377224
Signed-off-by: Arun Kumar Neelakantam <aneela@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Add support for QRTR to forward messages between network clusters. The
network subnet id's are attached to the QRTR subnodes in the MHI and
RPMSG/GLINK nodes.
The current forwarding decisions are done based on the net ids and are
wrapped in a single function for easier maintenance of forwarding
decisions later on. The function will return true once it determines
a subnet has no connection to another subnet.
The NEW_SERVER, DEL_SERVER, and DEL_CLIENT control messages should be
forwarded while the DATA and RESUME_TX commands should be passed along
to their destination node.
The nameservice is expected to send NEW_SERVER commands for it's entire
database instead of just the local service database with these changes.
This is to make sure new nodes get service notifications that came
earlier and were meant to be forwarded to the new node.
This change squashes the following commits from msm-4.14:
commit e5f6a6769ceb ("net: qrtr: Add forwarding support based on net id")
commit 4516ab18718d ("net: qrtr: avoid deadlock in case of recursive
lock for node list")
commit dc2d47cd12d3 ("net: qrtr: Only update pkts with broadcast node
ID")
commit 4277aebeff91 ("net: qrtr: Free skb if qrtr_node_lookup fails
during qrtr_fwd_pkt")
In addition fix minor format issues and update copyright.
Change-Id: Iaf01dd212b07dfbb13ec087ea0aae980cda7c2fd
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
There will be non-atomic work associated with receiving the tx resume
packet from the remote. Move the handling to work function context in
preparation for the non-blocking socket blocking improvement.
In system heavy load cases the rx work queued in global worker queue is
getting delayed and causing clients request timeouts.
Create and use separate worker thread to process qrtr rx packets.
Change-Id: I2389bdcffba068258650bfbb579f8cb98a0ac928
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
synchronize_rcu() is slow and can create delays on creating/removing
sockets.
Some clients have latency requirements for spinning off a
thread that creates a client and lookup sockets for some sessions. They
close the lookup socket during this sequence which causes delays. Each
call to synchronize_rcu takes 20-40 ms which is too long.
Remove the rcu logic and change qrtr_port_lock to a spinlock so it can
be used in qrtr_port_lookup which may be used in atomic context.
Change-Id: I3244e31d3ca0599df794d3936033735bcdd3e972
Signed-off-by: Jay Jayanna <jayanna@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Give control port access to processes that have root user priveleges,
AID_VENDOR_QRTR user priveleges, or NET_ADMIN capabilities.
Remove the need for SYS_ADMIN priveleges because they will not be
granted on any QTI targets.
This change squashes the following commits from msm-4.14:
commit 17da4da93c96 ("qrtr: Allow net bind service capabilities")
commit 56faadfd27b4 ("qrtr: Fix NET_BIND_SERVICE logic")
commit 82ea58d60c46 ("net: qrtr: Allow control port access to
AID_VENDOR_QRTR")
commit 9db4f5803ce1 ("net: qrtr: Allow Root guid process to bind")
In addition fix minor format issues.
Change-Id: I1441cbf53eb62880333a1031334a5331c1f54a77
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
If a port tries to send a message before the name service binds to the
control port, then that message should succeed. This will allow clients
to operate normally until the name service comes online and causes a
net reset to restart all the ports.
Change-Id: Ica0a5e45df0e51f282e5bf426d11ff8e41b339c4
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
A dl callback can be received anytime after mhi_prepare_for_transfer
has been called. There is a window where the callback may happen
before the probe initializes the qrtr_mhi_dev state. Move
mhi_prepare_for_transfer after the registering the endpoint.
Once moved, the reverse can happen where qrtr will try to send a packet
before the channels are prepared. Add a wait in the sending path to
ensure the channels are prepared before trying to do a ul transfer.
Change-Id: Ib5bf34ec8c1fa56a8ce29c89449d062de7133846
Signed-off-by: Bhaumik Bhatt <bbhatt@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Some transports need to know qrtr packet to read complete packet
from underlying transport.
Add API support to get the incoming packet size form qrtr header.
Change-Id: I4e4e4ab2e4c1fe0e1e1261af85a8b8618ce65bb3
Signed-off-by: Arun Kumar Neelakantam <aneela@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Log packets processed by ns using ipc logging framework.
Change-Id: I08e9b8551efb4a58cf899835cfac25c1952bd4b8
Signed-off-by: Jay Jayanna <jayanna@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Error prints for commands that are broadcasted may spam the kernel logs
if a client is misbehaving and stops reading from it's socket.
Change these prints to ratelimited in order to prevent performance
issues when a client socket is full.
Change-Id: I5aa5fa1f693f9865a3d28f28de38f870a42fbc8d
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Add pr_fmt to the ns file for easier error log matching.
Change-Id: I711735ee60c1fd9fde25850b8e625bb44c04f4c4
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Ignore the ENODEV failures returned by kernel_sendmsg(). These errors
mean either the local port has closed or the remote has gone down.
Neither of these scenarios are fatal and will eventually be handled
through packets that are later queued on the control port.
Also improve the logging messages to print the error code returned by
kernel_sendmsg().
Change-Id: Ibcebd1f44a3fbc87febaa84b11e05675663da5a0
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Increase the socket recvbuf for the ns so control packets are not
dropped in heavy traffic situations. This change will not
actually set the recvbuf to INT_MAX but to the system defined rmem_max
for sockets.
Change-Id: I860cc1052080cc6e683b44bef1dd9f8fc1899f31
Signed-off-by: Jay Jayanna <jayanna@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
If worker is not processing packets on control port fast enough,
socket buffer may get full and result in drop of control packets.
Creating dedicated worker thread to process control packets.
CRs-Fixed: 2763561
Change-Id: I4585015cef2a507a9ba13014cd93f40e60f8a25a
Signed-off-by: Deepak Kumar Singh <deesin@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
By enabling this flag init_boot.img will be generated if
the header version is 4.
Change-Id: I8a01d10cee8142ef900376d7e3d29f1b49e80d00
Signed-off-by: Gokul krishna Krishnakumar <quic_gokukris@quicinc.com>
Adding Bootconfig to kernel command line to support
the version change.
Change-Id: I24b667e494e5bbf14397f16d9ef7219634f3f17b
Signed-off-by: Gokul krishna Krishnakumar <quic_gokukris@quicinc.com>
sched_trace_rd_span is removed upstream. Restore the function locally.
Change-Id: I066be161e1da49b11de10f034e8f46a4c1510391
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
sched_switch tracehook has been changed upstream. Update functions
registered with tracehook accordingly.
Change-Id: Ifba53f905634b476c60a8c69cfd87c6ad0c9cff7
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
group_first_cpu() is removed upstream. Expand the function in all the
places that it is used.
Change-Id: I04f4f9e2b37fca897ee45165b1976e1bec83f5bd
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
The rq_relock() function is removed upstream. Replace all instances of it
by expanding the function.
Change-Id: I7b99d46d552695ae0afdf75f3a6033b80376fe01
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
Replace HK_FLAG_TIMER with HK_TYPE_TIMER because HK_FLAG_TIMER is
removed upstream.
Change-Id: I6432556b6148030c8c6621cd0790d8906e0c94ec
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
Due to reorganization of sched header files upstream, include missing
header files.
Change-Id: Iab3cda49a71b6a0f3f98194f975031716df47736
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
The Android Common Kernel has updated the enqueue,dequeue tracehooks.
Update the tracehooks accordingly.
Change-Id: Idb2efdb2f70288721c9100850e9cb344328a2c42
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
The Android Common Kernel has replaced the irq accounting tracehook.
Update the tracehook accordingly.
Change-Id: I4d0d2e7a4d5a6d567ac2c1edaae4d050b1c82ac0
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
Due to reorganization of tracehooks upstream, topology.h does not exist
anymore. Remove the associated header.
Change-Id: Ib4c7e630dfbefd6965524f91101784a57327af39
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
struct task_struct does not contain a member to store task cpu
information anymore. Instead, use task_thread_info() to obtain that
information.
Change-Id: I8fc9ef4ebf11f6fccd6ecbe68d768ac48d2928c5
Signed-off-by: Sai Harshini Nimmala <quic_snimmala@quicinc.com>
Add a tracepoint to the whitelist to be used in the UBWC-P heap.
Change-Id: I333c00519101b45840a32f340824450ebc385d28
Signed-off-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
Move sys heap allocation code into the system_qcom_sg_buffer_alloc()
function to allow core allocation code to be used from other modules.
Additionally, move the allocation of the struct qcom_sg_buffer
out of system_qcom_sg_buffer_alloc(), so that other modules can
allocate their own buffer structs with qcom_sg_buffer as an embedded
member.
Each other module will then have dmabuf->priv pointing to the
qcom_sg_buffer member of their custom buffer struct. This is needed
to allow the qcom_sg_ops.c functions to continue working as is.
Change-Id: Id11ca188163e5c3ceba28852c2cfe8f0acde4591
Signed-off-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
Add the declarations needed for the DMA-BUF heap driver to call into
the UBWC-P driver.
Change-Id: Id1bb19410025b858a2154eae39ac0c28e4d6235c
Signed-off-by: Chris Goldsworthy <quic_cgoldswo@quicinc.com>
When lookup fails, print the service path along with the
service name since the service name might not be unique.
Change-Id: I614d6c0a8d6cb8a83b47795315811fbf47bb6e48
Signed-off-by: Siddharth Gupta <sidgup@codeaurora.org>
Signed-off-by: Gokul krishna Krishnakumar <quic_gokukris@quicinc.com>
When a del client message is received from remote, the corresponding
server port is deleted. And then, a del server message is announced
for this server. As part of handling this message, ns will try to
delete this server port again and returns a -ENOENT error. Return
value from server_del is returned back to qrtr_ns_worker and any
non-zero value is printed to the serial port. This is flooding the
serial port display. Returinng 0 in this case as this is not an error.
Change-Id: Icd2594875f55313baa6e3313b6aa0c62facc5b2d
Signed-off-by: Jay Jayanna <jayanna@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
In external soc usecases, it is possible to receive control
messages for a node that is not the src node. The APPS on the
external soc will forward control messages from the modem.
Change-Id: Ibbf8f27fa64b18e387e40e655a13b5251190a00b
Signed-off-by: Tony Truong <truong@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
There is a use after free scenario while iterating through the servers
radix tree despite the ns being a single threaded process. This can
happen when the radix tree APIs are not synchronized with the
rcu_read_lock() APIs.
Convert the radix tree for servers to xarray to take advantage of the
built in rcu lock usage provided by xarray.
Change-Id: I1d9b017da4efba9d8fc72e4666253060cc7b87e3
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
There is a use after free scenario while iterating through the servers
radix tree despite the ns being a single threaded process. This can
happen when the radix tree APIs are not synchronized with the
rcu_read_lock() APIs.
Convert the radix tree for nodes to xarray to take advantage of the
built in rcu lock usage provided by xarray.
Change-Id: If2f735abf911b7b47cd7cb07224751114a2bf943
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
The destination port value in the IPCR control buffer on older
targets is 0xFFFF. Handle the same by updating the dst_port to
QRTR_PORT_CTRL.
Change-Id: Ia70ce1c078ea84f0de47240f6fc3e764f4ae7a6f
Signed-off-by: Ajay Agarwal <ajaya@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Socket client's should be able to block on recv without blocking other
socket operations such as sending. The qrtr core logic should also be
able to operate on the sock while a client is blocked in recv. Remove
the sock lock from around the recv_datagram to allow execution to
continue while a client is blocked in recv.
Change-Id: I6adb09b8e44c5ff3542b88db87b376a3ae5cb5ba
Signed-off-by: Jay Jayanna <jayanna@codeaurora.org>
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
