linux-toradex.git/drivers/mailbox/mailbox.c, branch v4.17

mailbox: reset txdone_method TXDONE_BY_POLL if client knows_txdone

2017-11-14T14:51:28+00:00

Currently the mailbox framework sets txdone_method to TXDONE_BY_POLL if
the controller sets txdone_by_poll. However some clients can have a
mechanism to do TXDONE_BY_ACK which they can specify by knows_txdone.
However, we endup setting both TXDONE_BY_POLL and TXDONE_BY_ACK in that
case. In such scenario, we may end up with below warnings as the tx
ticker is run both by mailbox framework and the client.

WARNING: CPU: 1 PID: 0 at kernel/time/hrtimer.c:805 hrtimer_forward+0x88/0xd8
CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.12.0-rc5 #242
Hardware name: ARM LTD ARM Juno Development Platform
task: ffff8009768ca700 task.stack: ffff8009768f8000
PC is at hrtimer_forward+0x88/0xd8
LR is at txdone_hrtimer+0xd4/0xf8
Call trace:
 hrtimer_forward+0x88/0xd8
 __hrtimer_run_queues+0xe4/0x158
 hrtimer_interrupt+0xa4/0x220
 arch_timer_handler_phys+0x30/0x40
 handle_percpu_devid_irq+0x78/0x130
 generic_handle_irq+0x24/0x38
 __handle_domain_irq+0x5c/0xb8
 gic_handle_irq+0x54/0xa8

This patch fixes the issue by resetting TXDONE_BY_POLL if client has set
knows_txdone.

Cc: Alexey Klimov 
Signed-off-by: Sudeep Holla 
Signed-off-by: Jassi Brar

mailbox: Make startup and shutdown ops optional

2017-06-14T12:17:09+00:00

Some mailbox hardware doesn't have to perform any additional operations
on startup of shutdown, so make these optional.

Signed-off-by: Bjorn Andersson 
Signed-off-by: Jassi Brar

mailbox: handle empty message in tx_tick

2017-04-27T10:50:04+00:00

We already check if the message is empty before calling the client
tx_done callback. Calling completion on a wait event is also invalid
if the message is empty.

This patch moves the existing empty message check earlier.

Fixes: 2b6d83e2b8b7 ("mailbox: Introduce framework for mailbox")
Signed-off-by: Sudeep Holla 
Signed-off-by: Jassi Brar

mailbox: skip complete wait event if timer expired

2017-04-27T10:50:04+00:00

If a wait_for_completion_timeout() call returns due to a timeout,
complete() can get called after returning from the wait which is
incorrect and can cause subsequent transmissions on a channel to fail.
Since the wait_for_completion_timeout() sees the completion variable
is non-zero caused by the erroneous/spurious complete() call, and
it immediately returns without waiting for the time as expected by the
client.

This patch fixes the issue by skipping complete() call for the timer
expiry.

Fixes: 2b6d83e2b8b7 ("mailbox: Introduce framework for mailbox")
Reported-by: Alexey Klimov 
Signed-off-by: Sudeep Holla 
Signed-off-by: Jassi Brar

mailbox: always wait in mbox_send_message for blocking Tx mode

2017-04-27T10:50:04+00:00

There exists a race when msg_submit return immediately as there was an
active request being processed which may have completed just before it's
checked again in mbox_send_message. This will result in return to the
caller without waiting in mbox_send_message even when it's blocking Tx.

This patch fixes the issue by waiting for the completion always if Tx
is in blocking mode.

Fixes: 2b6d83e2b8b7 ("mailbox: Introduce framework for mailbox")
Reported-by: Alexey Klimov 
Signed-off-by: Sudeep Holla 
Reviewed-by: Alexey Klimov 
Signed-off-by: Jassi Brar

mailbox: check ->last_tx_done for NULL in case of timer-based polling

2017-03-28T18:22:02+00:00

It is allowed by code to register mailbox controller that sets txdone_poll
flag to request timer-based polling with missed ->last_tx_done() method.
If such thing happens and since presence of last_tx_done() is not checked
it will fail in hrtimer callback function txdone_hrtimer() when first
message will be transmitted.

This patch adds check for this method and logging of error on
registration of mailbox controller if it requested timer-based polling.

Signed-off-by: Alexey Klimov 
Acked-by: Sudeep Holla 
Signed-off-by: Jassi Brar

ktime: Cleanup ktime_set() usage

2016-12-25T16:21:22+00:00

ktime_set(S,N) was required for the timespec storage type and is still
useful for situations where a Seconds and Nanoseconds part of a time value
needs to be converted. For anything where the Seconds argument is 0, this
is pointless and can be replaced with a simple assignment.

Signed-off-by: Thomas Gleixner 
Cc: Peter Zijlstra

mailbox: Stop using ENOSYS for anything other than unimplemented syscalls

2016-04-12T07:58:32+00:00

In accordance with e15f431fe2d5 ("errno.h: Improve ENOSYS's comment") and
91c9afaf97ee ("checkpatch.pl: new instances of ENOSYS are errors") we're
converting from the old meaning of: ENOSYS "Function not implemented" to
a more standard EINVAL.

Reported-by: Seraphin Bonnaffe 
Signed-off-by: Lee Jones 
Signed-off-by: Jassi Brar

mailbox: switch to hrtimer for tx_complete polling

2015-08-10T08:59:27+00:00

The mailbox core uses jiffy based timer to handle polling for the
transmit completion. If the client/protocol have/support notification
of the last packet transmit completion via ACK packet, then we tick the
Tx state machine immediately in the callback. However if the client
doesn't support that mechanism we might end-up waiting for atleast a
jiffy even though the remote is ready to receive the next request.

This patch switches the timer used for that polling from jiffy-based
to hrtimer-based so that we can support polling at much higher time
resolution.

Reported-and-suggested-by: Juri Lelli 
Signed-off-by: Sudeep Holla 
Signed-off-by: Jassi Brar

mailbox: Add ability for clients to request channels by name

2015-06-11T16:49:45+00:00

This patch supplies a new framework API; mbox_request_channel_byname().

It works by supplying the usual client pointer as the first argument and
a string as the second.  The API will search the client's node for a
'mbox-names' property then request a channel in the normal way using the
requested string's index as the expected second 'index' argument.

Signed-off-by: Lee Jones 
Signed-off-by: Jassi Brar