Because of removing endpoint delay when the stop channel failed case
observing some junk packets received on a pipe. To avoid junk packets,
add change to avoid endpoint delay if stopping GSI channel fails.
Change-Id: I4d3d8d9a0de823229d780fe31c8d7f1162e20668
Signed-off-by: Ashok Vuyyuru <avuyyuru@codeaurora.org>
Add api to write to mpm wakeup registers to configure a gpio
as wakeup capable.
Change-Id: Ia11a22160394e76969200691cfaf8fb1505bf5c1
Signed-off-by: Rishabh Bhatnagar <rishabhb@codeaurora.org>
Signed-off-by: Elliot Berman <eberman@codeaurora.org>
GSI QoS register upgrade for GSI v2.9, introduce the
new field of db_in_bytes.
Change-Id: I1d0ff4381a6b6ac1f5119d0be2efea29a49a5dcc
Signed-off-by: Bojun Pan <bojunp@codeaurora.org>
There is a race for clients that open sockets before the control port
is bound. If a client gets an idr that was allocated before the control
port is bound, there is a chance the previous address owner sent lookup
packets to the control port. The new address owner will get residual
responses to this the lookup packets.
Change the idr_alloc to idr_alloc_cyclic so new idr's are allocated
instead of trying to reuse the freed idrs.
Change-Id: Ie1bda7a818309503f80542e739bac646327296f7
Signed-off-by: Chris Lew <clew@codeaurora.org>
There is a chance for a remote processor to become unresponsive while
QRTR is trying to broadcast a control message. This results in the node
lock being held for an extended period of time and prevents node lookup
which is needed to send messages.
Change the mutex to a rw_semaphore to allow concurrency during node
lookups and broadcasts.
Change-Id: I2e3fdde22edaac64f164cf08900e9d09b16c380d
Signed-off-by: Chris Lew <clew@codeaurora.org>
A remote endpoint should not need to know when a client socket is freed
if the socket never established commnication with the endpoint. Add a
mode to keep track of which endpoints a socket communicates with.
There are three modes a socket can be in:
INIT - Socket has not sent anything or only local messages,
only send client close to local services.
SINGLE - Socket has sent messages to a single ept, send event
to this single ept.
MULTI - Socket has sent messages to multiple epts, broadcast
release of this socket.
Server state changes should be broadcast throughout the system. Change
the ipc state of a port when it sends a NEW SERVER control packet. This
ensures the DEL CLIENT control packet is propagated correctly for
servers.
This change squashes the following commits from msm-4.14:
commit 3373eb77cd6d ("net: qrtr: Add socket mode optimization")
commit d984a2d6f150 ("net: qrtr: Make servers multi state ports")
In addition, fix minor format issues.
Change-Id: I416742cff598bf7aa9da2baa2e86a53b18367870
Signed-off-by: Chris Lew <clew@codeaurora.org>
There is a race where broadcast packets can be sent to a node that has
not sent the hello message to the remote processor. This breaks the
protocol expectation. Add a status variable to track when the hello
packet has been sent.
An alternative solution attempted was to remove the nodes from the
broadcast list until the hello packet is sent. This is not a valid
solution because hello messages are broadcasted if the ns is restarted
or started late. There needs to be a status variable separate from the
broadcast list.
This change squashes the following commits from msm-4.14:
commit abd3961d05be ("qrtr: Only broadcast to initialized nodes")
commit 90cb0e473226 ("net: qrtr: Do not send packets before Hello
negotiation")
commit 07abd1abe803 ("net: qrtr: Fix hello packet drop memory leak")
In addition, fix minor format issues.
Change-Id: I2a55e587f324e64e2bde4ea634de66008f190de1
Signed-off-by: Chris Lew <clew@codeaurora.org>
The for_each_[available]_child_of_node iterators implicitly
increment the reference count on the current OF node at the
start of the iteration, and decrements it at the end of the
iteration.
However, if the control flow is such that the loop
can be broken out of prior to finishing an iteration, then
the reference count must be decremented in the subsequent
code path.
Thus, fix the error path in the MSM ION code to decrement the
reference on the current OF node when prematurely exiting
an of_node loop.
Change-Id: I5c2ec02a57eff416c04b1fe63c80d6e573a692d4
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Island IO can be assigned to Apps TLMM dynamically
if that island is not using the pin for any connectivity.
Add support for enabling EGPIO feature if the pin can
support it. EGPIO_PRESENT bit shows if EGPIO feature
is present or not for particular pin. Once EGPIO_ENABLE
bit is set the pin is now controlled by Apps TLMM.
When pinctrl properties are getting initialized by device
driver framework, pin functions select would call to
msm_pinmux_set_mux function to select pin function and
enable EGPIO.
Change-Id: I64514904096286f2a5df351880cd56d753f0658f
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
Signed-off-by: Rishabh Bhatnagar <rishabhb@codeaurora.org>
Signed-off-by: Elliot Berman <eberman@codeaurora.org>
Lahaina uses different pipe endpoint mappings and group resources.
This change captures these configurations.
Change-Id: Ib45848bb52d9a6680b38b4305cbab38c0b00e7d2
Acked-by: Priyadarshini Rajagopal <prajagop@qti.qualcomm.com>
Signed-off-by: Bojun Pan <bojunp@codeaurora.org>
Remote IPC Router drivers expect the received packet to be aligned.
This case is only exposed using the MHI transport because it does not
ensure the packet will be received as one chunk like RPMSG GLINK.
Change-Id: I0e5c27e3c5896e6721c8bbbfcda29551f0968626
Signed-off-by: Chris Lew <clew@codeaurora.org>
Rather than enqueuing messages and scheduling a worker to deliver them
to the individual sockets we can now, thanks to the previous work, move
this directly into the endpoint callback.
This saves us a context switch per incoming message and removes the
possibility of an opportunistic suspend to happen between the message is
coming from the endpoint until it ends up in the socket's receive
buffer.
Change-Id: I7177fa36f9d7cc7e5de5f24975f15284ef1fcc80
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Patch-mainline: linux-arm-kernel @ 07/05/19, 23:06
Signed-off-by: Chris Lew <clew@codeaurora.org>
sk_user_data can be protected with RCU instead of calling the callback
rwlock. RCU will ensure the pointer obtained in qmi_data_ready is
formed properly and cleanup can synchronize before continuing in handle
release.
Change-Id: I6a687bd7b311fe169a89cb959ae755730b81a9f1
Signed-off-by: Chris Lew <clew@codeaurora.org>
Add a Kconfig flag to differentiate between a GKI and an Non GKI build.
CONFIG_QGKI is enabled if the build is a Non GKI variant (QGKI or debug).
This flag doesn't differentiate between QGKI and debug variants.
Change-Id: I6b53b1e0e4c071fc7d9b365ff374ffe4e8b3e8ba
Signed-off-by: Prakruthi Deepak Heragu <pheragu@codeaurora.org>
Add debugfs node for ion which prints out the debug_show
of all the heaps within ion.
Change-Id: I8499967c8f202b079b6ccbe1c31cb38ee6a51374
Signed-off-by: Sudarshan Rajagopalan <sudaraja@codeaurora.org>
[isaacm@codeaurora.org: Rework to move logic out of ION core]
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Allow the system heap to expose debug information about the pools
through debugfs.
Change-Id: Id11706648b32260f19dff8d3f5cced1733a05978
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Allow ION clients to be able to determine the VM information associated
with an ION buffer, such as the VMIDs it is assigned to as well as the
VM permissions.
Change-Id: Id0be0b505bc84beff373e4e2ae16b71ec4f09d43
Signed-off-by: Liam Mark <lmark@codeaurora.org>
[isaacm@codeaurora.org: Relocate changes to new header locations]
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Remove support for passing parameters via platform_data now that
relevant targets use devicetree.
Change-Id: I2d8398163b5a4f358e257f416fd0e0c325664154
Signed-off-by: Patrick Daly <pdaly@codeaurora.org>
[isaacm@codeaurora.org: Relocated changes to correct file]
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
If a device cannot be created for a child node, or the ID cannot
be retrieved for a heap, then msm_ion_parse_dt() will return ret.
However, under these circumstances, ret has not been updated to
reflect these failures, and as such, msm_ion_probe will erroneously
continue as if the call to msm_ion_parse_dt() had been successful.
Thus, ensure that ret is set before following the error path in
msm_ion_parse_dt().
Change-Id: I241805e2c169fb15bda68c2a375def04790aa20e
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Upgrade the ION ABI version from to 2 to 3 to help ION
clients decide what they need to do based on the target
version.
Change-Id: I3d37fb9abb237e432283eaa4e7ea52419c08a2df
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Rename msm_ion_of.[c/h] to msm_ion.[c/h], as the OF portion is
a bit of a misnomer, considering that the file supports
more than just the OF parsing machinery.
Change-Id: I013cbdb78cb031ab7521f55873ed9fb031cbbed5
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
ION clients that reside in the kernel may need information that is
present in the UAPI header, such as flags to specify memory
attributes, or IDs to allocate from different heaps, so include
the UAPI header in the kernel space header. Also, ensure that
drivers that use this header can still compile even when
CONFIG_ION_MSM_HEAPS is not set.
Change-Id: If030a58ce62b73b66b337b6dfc6a607bbc0ce193
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
The APIs fdt_path_offset() and fdt_getprop_w() are not exported.
As a result, drivers that are compiled as modules cannot directly
use them. Hence, switch to of_find_node_by_path() and
of_property_read_u32() to achive the same functionality while also
being module friendly.
Change-Id: I90b299c60b60a006d0932442549c5d5a20cf2ea0
Signed-off-by: Raghavendra Rao Ananta <rananta@codeaurora.org>
Ensure that the secure buffer device is set only if dma_set_mask()
occurs successfully.
Change-Id: I9b84692b2860bafd8fe645b95e05d6436698d840
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
