net: ipa: add HW_PARAM_4 GSI register

Starting at IPA v5.0, the number of event rings per EE is defined
in a field in a new HW_PARAM_4 GSI register rather than HW_PARAM_2.
Define this new register and its fields, and update the code that
checks the number of rings supported by hardware to use the proper
field based on IPA version.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: support different event ring encoding

Starting with IPA v5.0, a channel's event ring index is encoded in
a field in the CH_C_CNTXT_1 GSI register rather than CH_C_CNTXT_0.
Define a new field ID for the former register and encode the event
ring in the appropriate register.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: avoid setting an undefined field

The GSI channel protocol field in the CH_C_CNTXT_0 GSI register is
widened starting IPA v5.0, making the CHTYPE_PROTOCOL_MSB field
added in IPA v4.5 unnecessary. Update the code to reflect this.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: kill ev_ch_e_cntxt_1_length_encode()

Now that we explicitly define each register field width there is no
need to have a special encoding function for the event ring length.
Add a field for this to the EV_CH_E_CNTXT_1 GSI register, and use it
in place of ev_ch_e_cntxt_1_length_encode() (which can be removed).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: kill gsi->virt_raw

Starting at IPA v4.5, almost all GSI registers had their offsets
changed by a fixed amount (shifted downward by 0xd000). Rather than
defining offsets for all those registers dependent on version, an
adjustment was applied for most register accesses. This was
implemented in commit cdeee49f3ef7f ("net: ipa: adjust GSI register
addresses"). It was later modified to be a bit more obvious about
the adjusment, in commit 571b1e7e58ad3 ("net: ipa: use a separate
pointer for adjusted GSI memory").

We now are able to define every GSI register with its own offset, so
there's no need to implement this special adjustment.

So get rid of the "virt_raw" pointer, and just maintain "virt" as
the (non-adjusted) base address of I/O mapped GSI register memory.

Redefine the offsets of all GSI registers (other than the INTER_EE
ones, which were not subject to the adjustment) for IPA v4.5+,
subtracting 0xd000 from their defined offsets instead.

Move the ERROR_LOG and ERROR_LOG_CLR definitions further down in the
register definition files so all registers are defined in order of
their offset.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: fix an incorrect assignment

I spotted an error in a patch posted this week, unfortunately just
after it got accepted. The effect of the bug is that time-based
interrupt moderation is disabled. This is not technically a bug,
but it is not what is intended. The problem is that a |= assignment
got implemented as a simple assignment, so the previously assigned
value was ignored.

Fixes: edc6158b18af ("net: ipa: define fields for event-ring related registers")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

net: ipa: define fields for remaining GSI registers

Define field IDs for the remaining GSI registers, and populate the
register definition files accordingly. Use the reg_*() functions to
access field values for those regiters, and get rid of the previous
field definition constants.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define fields for event-ring related registers

Define field IDs for the EV_CH_E_CNTXT_0 and EV_CH_E_CNTXT_8 GSI
registers, and populate the register definition files accordingly.
Use the reg_*() functions to access field values for those regiters,
and get rid of the previous field definition constants.

The remaining EV_CH_E_CNTXT_* registers are written with full 32-bit
values (and have no fields).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define more fields for GSI registers

Beyond the CH_C_QOS register, two other registers whose offset is
related to channel number have fields within them.

Define the fields within the CH_C_CNTXT_0 GSI register, using an
enumerated type to identify the register's fields, and define an
array of field masks to use for that register's reg structure.

For the CH_C_CNTXT_1 GSI register, ch_c_cntxt_1_length_encode()
previously hid the difference in bit width in the channel ring
length field. Instead, define a new field CH_R_LENGTH and encode
the ring size with reg_encode().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define GSI CH_C_QOS register fields

Define the fields within the CH_C_QOS GSI register using an array of
field masks in that register's reg structure. Use the reg functions
for encoding values in those fields.

One field in the register is present for IPA v4.0-4.2 only, two
others are present starting at IPA v4.5, and one more is there
starting at IPA v4.9.

Drop the "GSI_" prefix in symbols defined in the gsi_prefetch_mode
enumerated type, and define their values using decimal rather than
hexidecimal values.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define IPA remaining GSI register offsets

Add the remaining GSI register offset definitions. Use gsi_reg()
rather than the corresponding GSI_*_OFFSET() macros to get the
offsets for these registers, and get rid of the macros.

Note that we are now defining information for the HW_PARAM_2
register, and that doesn't appear until IPA v3.5.1.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: define IPA v3.1 GSI interrupt register offsets

Add definitions of the offsets for IRQ-related GSI registers. Use
gsi_reg() rather than the corresponding GSI_CNTXT_*_OFFSET() macros
to get the offsets for these registers, and get rid of the macros.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: define IPA v3.1 GSI event ring register offsets

Add definitions of the offsets and strides for registers whose
offset depends on an event ring ID, and use gsi_reg() and its
returned value to determine offsets for these registers. Get
rid of the corresponding GSI_EV_CH_E_*_OFFSET() macros.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: add more GSI register definitions

Continue populating with GSI register definitions, adding remaining
registers whose offset depends on a channel ID. Use gsi_reg() and
reg_n_offset() to determine offsets for those registers, and get rid
of the corresponding GSI_CH_C_*_OFFSET() macros.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: start creating GSI register definitions

Create a new register definition file in the "reg" subdirectory,
and begin populating it with GSI register definitions based on IPA
version. The GSI registers haven't changed much, so several IPA
versions can share the same GSI register definitions.

As with IPA registers, an array of pointers indexed by GSI register ID
refers to these register definitions, and a new "regs" field in the
GSI structure is initialized in gsi_reg_init() to refer to register
information based on the IPA version (though for now there's only
one). The new function gsi_reg() returns register information for
a given GSI register, and the result can be used to look up that
register's offset.

This patch is meant only to put the infrastructure in place, so only
eon register (CH_C_QOS) is defined for each version, and only the
offset and stride are defined for that register. Use new function
gsi_reg() to look up that register's information to get its offset,
This makes the GSI_CH_C_QOS_OFFSET() unnecessary, so get rid of it.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: introduce gsi_reg_init()

Create a new source file "gsi_reg.c", and in it, introduce a new
function to encapsulate initializing GSI registers, including
looking up and I/O mapping their memory.

Create gsi_reg_exit() as the inverse of the init function.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: GSI register cleanup

Move some static inline function definitions out of "gsi_reg.h" and
into "gsi.c", which is the only place they're used. Rename them so
their names identify the register they're associated with.

Move the gsi_channel_type enumerated type definition below the
offset and field definitions for the CH_C_CNTXT_0 register where
it's used.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use bitmasks for GSI IRQ values

There are seven GSI interrupt types that can be signaled by a single
GSI IRQ. These are represented in a bitmask, and the gsi_irq_type_id
enumerated type defines what each bit position represents.

Similarly, the global and general GSI interrupt types each has a set
of conditions it signals, and both types have an enumerated type
that defines which bit that represents each condition.

When used, these enumerated values are passed as an argument to BIT()
in *all* cases. So clean up the code a little bit by defining the
enumerated type values as one-bit masks rather than bit positions.

Rename gsi_general_id to be gsi_general_irq_id for consistency.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: generic command param fix

Starting at IPA v4.11, the GSI_GENERIC_COMMAND GSI register got a
new PARAMS field. The code that encodes a value into that field
sets it unconditionally, which is wrong.

We currently only provide 0 as the field's value, so this error has
no real effect. Still, it's a bug, so let's fix it.

Fix an (unrelated) incorrect comment as well. Fields in the
ERROR_LOG GSI register actually *are* defined for IPA versions
prior to v3.5.1.

Fixes: fe68c43ce388 ("net: ipa: support enhanced channel flow control")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: update copyrights

Some source files state copyright dates that are earlier than the
last modification of the file. Change the copyright year to 2022 in
all such cases.

Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20220930224549.3503434-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: update comments

This patch just updates comments throughout the IPA code.

Transaction state is now tracked using indexes into an array rather
than linked lists, and a few comments refer to the "old way" of
doing things. The description of how transactions are used was
changed to refer to "operations" rather than "commands", to
(hopefully) remove a possible ambiguity.

IPA register offsets and fields are now handled differently as well,
and the register documentation is updated to better describe the
code.

A few minor updates to comments were made (e.g., adding a missing
word, fixing a typo or punctuation, etc.).

Finally, the local macro atomic_dec_not_zero() is no longer used, so
it is deleted.

Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20220930224527.3503404-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: drop the weight argument from netif_napi_add

We tell driver developers to always pass NAPI_POLL_WEIGHT
as the weight to netif_napi_add(). This may be confusing
to newcomers, drop the weight argument, those who really
need to tweak the weight can use netif_napi_add_weight().

Acked-by: Marc Kleine-Budde <mkl@pengutronix.de> # for CAN
Link: https://lore.kernel.org/r/20220927132753.750069-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't have gsi_channel_update() return a value

If it finds no completed transactions, gsi_channel_trans_complete()
calls gsi_channel_update() to check hardware. If new transactions
have completed, gsi_channel_update() records that, then calls
gsi_channel_trans_complete() to return the first of those found.
This recursion won't go any further, but can be avoided if we
have gsi_channel_update() only be responsible for updating state
after accessing hardware.

Change gsi_channel_update() so it simply checks for and handles
new completions, without returning a value. If it needs to call
that function, have gsi_channel_trans_complete() determine whether
there are new transactions available after the update.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: update channel in gsi_channel_trans_complete()

Have gsi_channel_trans_complete() update the known state from
hardware rather than doing so in gsi_channel_poll_one().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: further simplify gsi_channel_trans_last()

Do a little more refactoring in gsi_channel_trans_last() to simplify
it further. The resulting code should behave exactly as before.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: simplify gsi_channel_trans_last()

Using a little logic we can simplify gsi_channel_trans_last().

The first condition in that function looks like this:
if (trans_info->allocated_id != trans_info->free_id)
And if that's false, we proceed to the next one:
if (trans_info->committed_id != trans_info->allocated_id)

Failure of the first test implies:
trans_info->allocated_id == trans_info->free_id
And therefore, the second one can be rewritten this way:
if (trans_info->committed_id != trans_info->free_id)

Substituting free_id for allocated_id and committed_id can also be
done in the code blocks executed when these conditions yield true.
The net result is that all three blocks for TX endpoints can be
consolidated into just one.

The two blocks of code at the end of that function (used for both TX
and RX channels) can be similarly consolidated into a single block.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use IDs exclusively for last transaction

Always use transaction IDs when finding the "last" transaction to
await when quiescing a channel. This basically extends what was
done in the previous patch to all other transaction state IDs.

As a result we are no longer updating any transaction lists inside
gsi_channel_trans_last(), so there's no need to take the spinlock.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use IDs for last allocated transaction

Use the allocated and free transaction IDs to determine whether the
"last" transaction used for quiescing a channel is in allocated
state. The last allocated transaction that has not been committed
(if any) immediately precedes the first free transaction in the
transaction array.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: rework last transaction determination

When quiescing a channel, we find the "last" transaction, which is
the latest one to have been allocated. (New transaction allocation
will have been prevented by the time this is called.)

Currently we do this by looking for the first non-empty transaction
list in each state, then return the last entry from that last.
Instead, determine the last entry in each list (if any) and return
that entry if found.

Temporarily (locally) introduce list_last_entry_or_null() as a
helper for this, mirroring list_first_entry_or_null(). This macro
definition will be removed by an upcoming patch.

Remove the temporary warnings added by the previous commit.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: add a transaction committed list

We currently put a transaction on the pending list when it has
been committed. But until the channel's doorbell rings, these
transactions aren't actually "owned" by the hardware yet.

Add a new "committed" state (and list), to represent transactions
that have been committed but not yet sent to hardware. Define
"pending" to mean committed transactions that have been sent
to hardware but have not yet completed.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: initialize ring indexes to 0

When a GSI channel is initially allocated, and after it has been
reset, the hardware assumes its ring index is 0. And although we
do initialize channels this way, the comments in the IPA code don't
really explain this. For event rings, it doesn't matter what value
we use initially, so using 0 is just fine.

Add some information about the assumptions made by hardware above
the definition of the gsi_ring structure in "gsi.h".

Zero the index field for all rings (channel and event) when the ring
is allocated. As a result, that function initializes all fields in
the structure.

Stop zeroing the index the top of gsi_channel_program(). Initially
we'll use the index value set when the channel ring was allocated.
And we'll explicitly zero the index value in gsi_channel_reset()
before programming the hardware, adding a comment explaining why
it's required.

For event rings, use the index initialized by gsi_ring_alloc()
rather than 0 when ringing the doorbell in gsi_evt_ring_program().
(It'll still be zero, but we won't assume that to be the case.)

Use a local variable in gsi_evt_ring_program() that represents the
address of the event ring's ring structure.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: move more code out of gsi_channel_update()

Move the processing done for TX channels in gsi_channel_update()
into gsi_evt_ring_rx_update(). The called function is called for
both RX and TX channels, so rename it to be gsi_evt_ring_update().
As a result, this code no longer assumes events in an event ring are
associated with just one channel.

Because all events in a ring are handled in that function, we can
move the call to gsi_trans_move_complete() there, and can ring the
event ring doorbell there as well after all new events in the ring
have been processed.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: call gsi_evt_ring_rx_update() unconditionally

When an RX transaction completes, we update the trans->len field to
contain the actual number of bytes received. This is done in a loop
in gsi_evt_ring_rx_update().

Change that function so it checks the data transfer direction
recorded in the transaction, and only updates trans->len for RX
transfers.

Then call it unconditionally. This means events for TX endpoints
will run through the loop without otherwise doing anything, but
this will change shortly.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: pass GSI pointer to gsi_evt_ring_rx_update()

The only reason the event ring's channel pointer is needed in
gsi_evt_ring_rx_update() is so we can get at its GSI pointer.

We can pass the GSI pointer as an argument, along with the event
ring ID, and thereby avoid using the event ring channel pointer.
This is another step toward no longer assuming an event ring
services a single channel.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't assume one channel per event ring

In gsi_evt_ring_rx_update(), use gsi_event_trans() repeatedly
to find the transaction associated with an event, rather than
assuming consecutive events are associated with the same channel.
This removes the only caller of gsi_trans_pool_next(), so get rid
of it.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: rework gsi_channel_tx_update()

Rename gsi_channel_tx_update() to be gsi_trans_tx_completed(), and
pass it just the transaction pointer, deriving the channel from the
transaction. Update the comments above the function to provide a
more concise description of how statistics for TX endpoints are
maintained and used.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: stop counting total RX bytes and transactions

In gsi_evt_ring_rx_update(), we update each transaction so its len
field reflects the actual number of bytes received. In the process,
the total number of transactions and bytes processed on the channel
are summed, and added to a running total for the channel.

But we don't actually use those running totals for RX endpoints.
They're maintained for TX channels to support CoDel when they are
associated with a "real" network device.

So stop maintaining these totals for RX endpoints, and update the
comment where the fields are defined to make it clear they're only
valid for TX channels.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: simplify TX completion statistics

When a TX request is issued, its channel's accumulated byte and
transaction counts are recorded. This currently does *not* take
into account the transaction being committed.

Later, when the transaction completes, the number of bytes and
transactions that have completed since the transaction was committed
are reported to the network stack. The transaction and its byte
count are accounted for at that time.

Instead, record the transaction and its bytes in the counts recorded
at commit time. This avoids the need to do so when the transaction
completes, and provides a (small) simplification of that code.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: introduce gsi_trans_tx_committed()

Create a new function that encapsulates recording information needed
for TX channel statistics when a transaction is committed.

Record the accumulated length in the transaction before the call
(for both RX and TX), so it can be used when updating TX statistics.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: derive channel from transaction

In gsi_channel_tx_queued(), we report when a transaction gets passed
to hardware. Change that function so it takes transaction rather
than a channel as its argument, and derive the channel from the
transaction. Rename the function accordingly.

Delete the header comments above the function definition; the ones
above the declaration in "gsi_private.h" should suffice. In
addition, the comments above gsi_channel_tx_update() do a fine job
of explaining what's going on.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: determine channel from event

Each event in an event ring describes the TRE whose completion
caused the event. Currently, every event ring is dedicated to a
single channel, so the channel is easily derived from the event
ring.

An event ring can actually be shared by more than one channel
though, and to distinguish events for one channel from another, the
event structure contains a field indicating which channel the event
is associated with.

In gsi_event_trans(), use the channel ID in an event to determine
which channel the event is for. This makes the channel pointer now
passed to that function irrelevant; pass the GSI pointer to that
function instead.

And although it shouldn't happen, warn if an event arrives that
records a channel ID that's not in use, or if the event does not
have a transaction associated with it.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: rename channel->tlv_count

Each GSI channel has a TLV FIFO of a certain size, specified in the
configuration data for an AP channel. That size dictates the
maximum number of TREs that are allowed in a single transaction.

The only way that value is used after initialization is as a limit
on the number of TREs in a transaction; calling it "tlv_count"
isn't helpful, and in fact gsi_channel_trans_tre_max() exists to
sort of abstract it.

Instead, rename the channel->tlv_count field trans_tre_max, and get
rid of the helper function. Update a couple of comments as well.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: verify command channel TLV count

In commit 8797972afff3d ("net: ipa: remove command info pool"), the
maximum number of IPA commands that would be sent in a single
transaction was defined. That number can't exceed the size of the
TLV FIFO on the command channel, and we can check that at runtime.

To add this check, pass a new flag to gsi_channel_data_valid() to
indicate the channel being checked is being used for IPA commands.
Knowing that we can also verify the channel direction is correct.

Use a new local variable that refers to the command-specific portion
of the data being checked.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: rename a GSI error code

The CHANNEL_NOT_RUNNING error condition has been generalized, so
rename it to be INCORRECT_CHANNEL_STATE.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: drop an unneeded transaction reference

In gsi_channel_update(), a reference count is taken on the last
completed transaction "to keep it from completing" before we give
the event back to the hardware. Completion processing for that
transaction (and any other "new" ones) will not occur until after
this function returns, so there's no risk it completing early. So
there's no need to take and drop the additional transaction
reference.

Use local variables in the call to gsi_evt_ring_doorbell().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: switch to netif_napi_add_tx()

Switch net callers to the new API not requiring
the NAPI_POLL_WEIGHT argument.

Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Acked-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Acked-by: Alexandra Winter <wintera@linux.ibm.com>
Link: https://lore.kernel.org/r/20220504163725.550782-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: record proper RX transaction count

Each time we are notified that some number of transactions on an RX
channel has completed, we record the number of bytes that have been
transferred since the previous notification. We also track the
number of transactions completed, but that is not currently being
calculated correctly; we're currently counting the number of such
notifications, but each notification can represent many transaction
completions. Fix this.

Fixes: 650d1603825d8 ("soc: qcom: ipa: the generic software interface")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: support enhanced channel flow control

IPA v4.2 introduced GSI channel flow control, used instead of IPA
endpoint DELAY mode to prevent a TX channel from injecting packets
into the IPA core. It used a new FLOW_CONTROLLED channel state
which could be entered using GSI generic commands.

IPA v4.11 extended the channel flow control model. Rather than
having a distinct FLOW_CONTROLLED channel state, each channel has a
"flow control" property that can be enabled or not--independent of
the channel state. The AP (or modem) can modify this property using
the same GSI generic commands as before.

The AP only uses channel flow control on modem TX channels, and only
when recovering from a modem crash. The AP has no way to discover
the state of a modem channel, so the fact that (starting with IPA
v4.11) flow control no longer uses a distinct channel state is
invisible to the AP. So enhanced flow control generally does not
change the way AP uses flow control.

There are a few small differences, however:
- There is a notion of "primary" or "secondary" flow control, and
when enabling or disabling flow control that must be specified
in a new field in the GSI generic command register. For now, we
always specify 0 (meaning "primary").
- When disabling flow control, it's possible a request will need
to be retried. We retry up to 5 times in this case.
- Another new generic command allows the current flow control
state to be queried. We do not use this.

Other than the need for retries, the code essentially works the same
way as before.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: introduce channel flow control

One quirk for certain versions of IPA is that endpoint DELAY mode
does not work properly. IPA DELAY mode prevents any packets from
being delivered to the IPA core for processing on a TX endpoint.
The AP uses DELAY mode when the modem crashes, to prevent modem TX
endpoints from generating traffic during crash recovery. Without
this, there is a chance the hardware will stall during recovery from
a modem crash.

To achieve a similar effect, a GSI FLOW_CONTROLLED channel state
was created. A STARTED TX channel can be placed in FLOW_CONTROLLED
state, which prevents the transfer of any more packets. A channel
in FLOW_CONTROLLED state can be either returned to STARTED state, or
can be transitioned to STOPPED state.

Because this operates on GSI channels, two generic commands were
added to allow the AP to control this state for modem channels
(similar to the ALLOCATE and HALT channel commands).

Previously the code assumed this quirk only applied to IPA v4.2.
In fact, channel flow control (rather than endpoint DELAY mode)
should be used for all versions *starting* with IPA v4.2.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: GSI only needs one completion

A mutex ensures we never submit more than one GSI command of any
kind at once. This means the per-channel and per-event ring
completion structures provide no benefit. Instead, just use the
single (existing) GSI completion to signal the completion of GSI
commands of all types.

This makes gsi_evt_ring_init() a trivial function with no inverse,
so open-code it in its sole caller and get rid of the function.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: disable GSI interrupts while suspended

Introduce new functions gsi_suspend() and gsi_resume(), which will
disable the GSI interrupt handler after all endpoints are suspended
and re-enable it before endpoints are resumed. This will ensure no
GSI interrupt handler will fire when the hardware is suspended.

Here's a little further explanation. There are seven GSI interrupt
types, and most are disabled except when needed.
- These two are not used (never enabled):
GSI_INTER_EE_CH_CTRL
GSI_INTER_EE_EV_CTRL
- These two are only used to implement channel and event ring
commands, and are only enabled while a command is underway:
GSI_CH_CTRL
GSI_EV_CTRL
- The IEOB interrupt signals I/O completion. It will not fire
when a channel is stopped (or "suspended").
GSI_IEOB
- This interrupt is used to allocate or halt modem channels,
and is only enabled while such a command is underway.
GSI_GLOB_EE
However it also is used to signal certain errors, and this could
occur at any time.
- The general interrupt signals general errors, and could occur at
any time.
GSI_GENERAL

The purpose for this change is to ensure no global or general
interrupts fire due to errors while the hardware is suspended.
We enable the clock on resume, and at that time we can "handle"
(at least report) these error conditions.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: move gsi_irq_init() code into setup

The GSI IRQ handler could be triggered as soon as it is registered
with request_irq(). The handler function, gsi_isr(), touches
hardware, meaning the IPA clock must be operational. The IPA clock
is not operating when the handler is registered (in gsi_irq_init()),
so this is a problem.

Move the call to request_irq() for the GSI interrupt handler into
gsi_irq_setup(), which is called when the IPA clock is known to be
operational (and furthermore, the GSI firmware will have been
loaded). Request the IRQ at the end of that function, after all
interrupt types have been disabled and masked.

Move the matching free_irq() call into gsi_irq_teardown(), and get
rid of the now empty gsi_irq_exit(),

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: have gsi_irq_setup() return an error code

Change gsi_irq_setup() so it returns an error value, and introduce
gsi_irq_teardown() as its inverse. Set the interrupt type (IRQ
rather than MSI) in gsi_irq_setup().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: move some GSI setup functions

Move gsi_irq_setup() and gsi_ring_setup() so they're defined right
above gsi_setup() where they're called. This is a trivial movement
of code to prepare for upcoming patches.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: move version check for channel suspend/resume

Change the Boolean flags passed to __gsi_channel_start() and
__gsi_channel_stop() so they represent whether the request is being
made to implement suspend (versus stop) or resume (versus start).

Then stop or start the channel for suspend/resume requests only if
the hardware version indicates it should be done.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use gsi->version for channel suspend/resume

The GSI layer has the IPA version now, so there's no need for
version-specific flags to be passed from IPA. One instance of
this is in gsi_channel_suspend() and gsi_channel_resume(), which
indicate whether or not the endpoint suspend is implemented by
GSI stopping the channel. We can make that determination based
on gsi->version, eliminating the need for a Boolean flag in those
functions.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: kill the remaining conditional validation code

There are only a few remaining spots that validate IPA code
conditional on whether a symbol is defined at compile time.
The checks are not expensive, so just build them always.

This completes the removal of all CONFIG_VALIDATE/CONFIG_VALIDATION
IPA code.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: introduce gsi_ring_setup()

Prior to IPA v3.5.1, there is no HW_PARAM_2 GSI register, which we
use to determine the number of channels and endpoints per execution
environment. In that case, we will just assume the number supported
is the maximum supported by the driver.

Introduce gsi_ring_setup() to encapsulate the code that determines
the number of channels and endpoints.

Update GSI_EVT_RING_COUNT_MAX so it is big enough to handle any
available channel for all supported hardware (IPA v4.9 can have 23
channels and 24 event rings).

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: AngeloGioacchino Del Regno
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: inter-EE interrupts aren't always available

The GSI inter-EE interrupts are not supported prior to IPA v3.5.
Don't attempt to initialize them in gsi_irq_setup() for hardware
that does not support them.

Originally proposed by AngeloGioacchino Del Regno.

Link: https://lore.kernel.org/netdev/20210211175015.200772-4-angelogioacchino.delregno@somainline.org
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: AngeloGioacchino Del Regno
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: fix inter-EE IRQ register definitions

In gsi_irq_setup(), two registers are written with the intention of
disabling inter-EE channel and event IRQs.

But the wrong registers are used (and defined); the ones used are
read-only registers that indicate whether the interrupt condition is
present.

Define the mask registers instead of the status registers, and use
them to disable the inter-EE interrupt types.

Fixes: 46f748ccaf01 ("net: ipa: explicitly disallow inter-EE interrupts")
Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20210505223636.232527-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: get rid of empty GSI functions

There are place holder functions in the GSI code that do nothing.
Remove these, knowing we can add something back in their place if
they're really needed someday.

Some of these are inverse functions (such as teardown to match setup).
Explicitly comment that there is no inverse in these cases.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: DMA addresses are nicely aligned

A recent patch avoided doing 64-bit modulo operations by checking
the alignment of some DMA allocations using only the lower 32 bits
of the address.

David Laight pointed out (after the fix was committed) that DMA
allocations might already satisfy the alignment requirements. And
he was right.

Remove the alignment checks that occur after DMA allocation requests,
and update comments to explain why the constraint is satisfied. The
only place IPA_TABLE_ALIGN was used was to check the alignment; it is
therefore no longer needed, so get rid of it.

Add comments where GSI_RING_ELEMENT_SIZE and the tre_count and
event_count channel data fields are defined to make explicit they
are required to be powers of 2.

Revise a comment in gsi_trans_pool_init_dma(), taking into account
that dma_alloc_coherent() guarantees its result is aligned to a page
size (or order thereof).

Don't bother printing an error if a DMA allocation fails.

Suggested-by: David Laight <David.Laight@ACULAB.COM>
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: expand GSI channel types

IPA v4.5 (GSI v2.5) supports a larger set of channel protocols, and
adds an additional field to hold the most-significant bits of the
protocol identifier on a channel.

Add an inline function that encodes the protocol (including the
extra bits for newer versions of IPA), and define some additional
protocols. At this point we still use only GPI protocol.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: update GSI ring size registers

Each GSI channel has a CNTXT_1 register that encodes the size of its
ring buffer. The size of the field that records that is increased
starting at IPA v4.9. Replace the use of a fixed-size field mask
with a new inline function that encodes that size value.

Similarly, the size of GSI event rings can be larger starting with
IPA v4.9, so create a function to encode that as well.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: reduce IPA version assumptions

Modify conditional tests throughout the IPA code so they do not
assume that IPA v3.5.1 is the oldest version supported. Also remove
assumptions that IPA v4.5 is the newest version of IPA supported.

Augment versions in comments with "+", to be clearer that the
comment applies to a version and subsequent versions. (E.g.,
"present for IPA v4.2+" instead of just "present for v4.2".)

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: avoid 64-bit modulus

It is possible for a 32 bit x86 build to use a 64 bit DMA address.

There are two remaining spots where the IPA driver does a modulo
operation to check alignment of a DMA address, and under certain
conditions this can lead to a build error on i386 (at least).

The alignment checks we're doing are for power-of-2 values, and this
means the lower 32 bits of the DMA address can be used. This ensures
both operands to the modulo operator are 32 bits wide.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use upper_32_bits()

Use upper_32_bits() to extract the high-order 32 bits of a DMA
address. This avoids doing a 32-position shift on a DMA address
if it happens not to be 64 bits wide. Use lower_32_bits() to
extract the low-order 32 bits (because that's what it's for).

Suggested-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: Remove useless error message

Fix the following coccicheck report:

drivers/net/ipa/gsi.c:1341:2-9:
line 1341 is redundant because platform_get_irq() already prints an error

Remove dev_err() messages after platform_get_irq_byname() failures.

Signed-off-by: Zihao Tang <tangzihao1@hisilicon.com>
Signed-off-by: Jay Fang <f.fangjian@huawei.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: introduce gsi_channel_initialized()

Create a simple helper function that indicates whether a channel has
been initialized. This abstacts/hides the details of how this is
determined.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: use a separate pointer for adjusted GSI memory

This patch actually fixes a bug, though it doesn't affect the two
platforms supported currently. The fix implements GSI memory
pointers a bit differently.

For IPA version 4.5 and above, the address space for almost all GSI
registers is adjusted downward by a fixed amount. This is currently
handled by adjusting the I/O virtual address pointer after it has
been mapped. The bug is that the pointer is not "de-adjusted" as it
should be when it's unmapped.

This patch fixes that error, but it does so by maintaining one "raw"
pointer for the mapped memory range. This is assigned when the
memory is mapped and used to unmap the memory. This pointer is also
used to access the two registers that do *not* sit in the "adjusted"
memory space.

Rather than adjusting *that* pointer, we maintain a separate pointer
that's an adjusted copy of the "raw" pointer, and that is used for
most GSI register accesses.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: do not cache event ring state

An event ring's state only needs to be known when it is allocated,
reset, or deallocated. We check an event ring's state both before
and after performing an event ring control command that changes
its state. These are only issued at startup and shutdown, so there
is very little value in caching the state.

Stop recording a copy of the channel's last known state, and instead
fetch the true state from hardware whenever it's needed. In such
cases, *do* record the state in a local variable, in case an error
message reports it (so the value reported is the value seen).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: synchronize NAPI only for suspend

When stopping a channel, gsi_channel_stop() will ensure NAPI
polling is complete when it calls napi_disable(). So there is no
need to call napi_synchronize() in that case.

Move the call to napi_synchronize() out of __gsi_channel_stop()
and into gsi_channel_suspend(), so it's only used where needed.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: move mutex calls into __gsi_channel_stop()

Move the mutex calls out of gsi_channel_stop_retry() and into
__gsi_channel_stop(), to make the latter more semantically similar
to __gsi_channel_start().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: expand last transaction check

Transactions to send data for a network device can be allocated at
any time up until the point the TX queue is stopped. It is possible
for ipa_start_xmit() to be called in one context just before a
the transmit queue is stopped in another.

Update gsi_channel_trans_last() so that for TX channels the
allocated and pending transaction lists are checked--in addition
to the completed and polled lists--to determine the "last"
transaction. This means any transaction that has been allocated
before the TX queue is stopped will be allowed to complete before
we conclude the channel is quiesced.

Rework the function a bit to use a list pointer and gotos.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't disable interrupt on suspend

No completion interrupts will occur while an endpoint is suspended,
nor when a channel has been stopped for suspend. So there's no need
to disable the interrupt during suspend and re-enable it when
resuming. Without any interrupts occurring, there is no need to
disable/re-enable NAPI for channel suspend/resume either.

We'll only enable NAPI and the interrupt when we first start the
channel, and disable it again only when it's "really" stopped.

To accomplish this, move the enable/disable calls out of
__gsi_channel_start() and __gsi_channel_stop(), and into
gsi_channel_start() and gsi_channel_stop() instead.

Add a call to napi_synchronize() to gsi_channel_suspend(), to ensure
NAPI polling is done before moving on.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: disable interrupt and NAPI after channel stop

Disable both the I/O completion interrupt and NAPI polling on a
channel *after* we successfully stop it rather than before. This
ensures a completion occurring just before the channel is stopped
gets processed.

Enable NAPI polling and the interrupt *before* starting a channel
rather than after, to be symmetric. A stopped channel won't
generate any completion interrupts anyway.

Enable NAPI before the interrupt and disable it afterward.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: kill gsi_channel_freeze() and gsi_channel_thaw()

Open-code gsi_channel_freeze() and gsi_channel_thaw() in all callers
and get rid of these two functions. This is part of reworking the
sequence of things done during channel suspend/resume and start/stop.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: introduce __gsi_channel_start()

Create a new function that does most of the work of starting a
channel. What's different is that it takes a flag indicating
whether the channel should really be started or not. Create
another new function __gsi_channel_stop() that behaves similarly.

IPA v3.5.1 implements suspend using a special SUSPEND endpoint
setting. If the endpoint is suspended when an I/O completes on the
underlying GSI channel, a SUSPEND interrupt is generated.

Newer versions of IPA do not implement the SUSPEND endpoint mode.
Instead, endpoint suspend is implemented by simply stopping the
underlying GSI channel. In this case, a completing I/O on a
*stopped* channel causes the SUSPEND interrupt condition.

These new functions put all activity related to starting or stopping
a channel (including "thawing/freezing" the channel) in one place,
whether or not the channel is actually started or stopped.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: introduce gsi_channel_stop_retry()

Create a new helper function that encapsulates issuing a set of
channel stop commands, retrying if appropriate, with a short delay
between attempts.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't thaw channel if error starting

If an error occurs starting a channel, don't "thaw" it.
We should assume the channel remains in a non-started state.

Update the comment in gsi_channel_stop(); calls to this function
are no longer retried.

Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: disable IEOB interrupts before clearing

Currently in gsi_isr_ieob(), event ring IEOB interrupts are disabled
one at a time. The loop disables the IEOB interrupt for all event
rings represented in the event mask. Instead, just disable them all
at once.

Disable them all *before* clearing the interrupt condition. This
guarantees we'll schedule NAPI for each event once, before another
IEOB interrupt could be signaled.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: repurpose gsi_irq_ieob_disable()

Rename gsi_irq_ieob_disable() to be gsi_irq_ieob_disable_one().

Introduce a new function gsi_irq_ieob_disable() that takes a mask of
events to disable rather than a single event id. This will be used
in the next patch.

Rename gsi_irq_ieob_enable() to be gsi_irq_ieob_enable_one() to be
consistent.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: have gsi_channel_update() return a value

Have gsi_channel_update() return the first transaction in the
updated completed transaction list, or NULL if no new transactions
have been added.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: heed napi_complete() return value

Pay attention to the return value of napi_complete(), completing
polling only if it returns true.

Just use napi rather than &channel->napi as the argument passed to
napi_complete().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: count actual work done in gsi_channel_poll()

There is an off-by-one problem in gsi_channel_poll(). The count of
transactions completed is incremented each time through the loop
*before* determining whether there is any more work to do. As a
result, if we exit the loop early the counter its value is one more
than the number of transactions actually processed.

Instead, increment the count after processing, to ensure it reflects
the number of processed transactions. The result is more naturally
described as a for loop rather than a while loop, so change that.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: retry TX channel stop commands

For RX channels we issue a stop command more than once if necessary
to allow it to stop. It turns out that TX channels could also
require retries.

Retry channel stop commands if necessary regardless of the channel
direction. Rename the symbol defining the retry count so it's not
RX-specific.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: change stop channel retry delay

If a GSI stop channel command leaves the channel in STOP_IN_PROC
state, we retry the stop command after a 1-2 millisecond delay.

I have been told that a 3-5 millisecond delay is a better choice.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: change GSI command timeout

The GSI command timeout is currently 5 seconds, which is much higher
than it should be.

Express the timeout in milliseconds rather than seconds, and reduce
it to 50 milliseconds.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use usleep_range()
65;6003;1c
The use of msleep() for small periods (less than 20 milliseconds) is
not recommended because the actual delay can be much different than
expected.

We use msleep(1) in several places in the IPA driver to insert short
delays. Replace them with usleep_range calls, which should reliably
delay a period in the range requested.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: introduce some interrupt helpers

Create a new function gsi_irq_ev_ctrl_enable() that encapsulates
enabling the event ring control GSI interrupt type, and enables a
single event ring to signal that interrupt. When an event ring
changes state as a result of an event ring command, it triggers this
interrupt.

Create an inverse function gsi_irq_ev_ctrl_disable() as well.
Because only one event ring at a time is enabled for this interrupt,
we can simply disable the interrupt for *all* channels.

Create a pair of helpers that serve the same purpose for channel
commands.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: a few simple renames

The return value of gsi_command() is true if successful or false if
we time out waiting for a completion interrupt.

Rename the variables in the three callers of gsi_command() to be
"timeout", to make it more obvious that's the only reason for
failure.

In addition, add a "gsi_" prefix to evt_ring_command() so its name
is consistent with the convention used for GSI channel and generic
commands.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: set error code in gsi_channel_setup()

In gsi_channel_setup(), we check to see if the configuration data
contains any information about channels that are not supported by
the hardware. If one is found, we abort the setup process, but
the error code (ret) is not set in this case. Fix this bug.

Fixes: 650d1603825d8 ("soc: qcom: ipa: the generic software interface")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20210204010655.15619-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: add a missing __iomem attribute

The virt local variable in gsi_channel_state() does not have an
__iomem attribute but should. Fix this.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Amy Parker <enbyamy@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: pass correct dma_handle to dma_free_coherent()

The "ring->addr = addr;" assignment is done a few lines later so we
can't use "ring->addr" yet. The correct dma_handle is "addr".

Fixes: 650d1603825d ("soc: qcom: ipa: the generic software interface")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/YBjpTU2oejkNIULT@mwanda
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't return a value from evt_ring_command()

Callers of evt_ring_command() no longer care whether the command
times out, and don't use what evt_ring_command() returns. Redefine
that function to have void return type.

Reported-by: kernel test robot <lkp@intel.com>
Fixes: 428b448ee764a ("net: ipa: use state to determine event ring command success")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't return a value from gsi_channel_command()

Callers of gsi_channel_command() no longer care whether the command
times out, and don't use what gsi_channel_command() returns. Redefine
that function to have void return type.

Reported-by: kernel test robot <lkp@intel.com>
Fixes: 6ffddf3b3d182 ("net: ipa: use state to determine channel command success")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use state to determine event ring command success

This patch implements the same basic fix for event rings as the
previous one does for channels.

The result of issuing an event ring control command should be that
the event ring changes state. If enabled, a completion interrupt
signals that the event ring state has changed. This interrupt is
enabled by gsi_evt_ring_command() and disabled again after the
command has completed (or we time out).

There is a window of time during which the command could complete
successfully without interrupting. This would cause the event ring
to transition to the desired new state.

So whether a event ring command ends via completion interrupt or
timeout, we can consider the command successful if the event ring
has entered the desired state (and a failure if it has not,
regardless of the cause).

Fixes: b4175f8731f78 ("net: ipa: only enable GSI event control IRQs when needed")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use state to determine channel command success

The result of issuing a channel control command should be that the
channel changes state. If enabled, a completion interrupt signals
that the channel state has changed. This interrupt is enabled by
gsi_channel_command() and disabled again after the command has
completed (or we time out).

There is a window of time--after the completion interrupt is disabled
but before the channel state is read--during which the command could
complete successfully without interrupting. This would cause the
channel to transition to the desired new state.

So whether a channel command ends via completion interrupt or
timeout, we can consider the command successful if the channel
has entered the desired state (and a failure if it has not,
regardless of the cause).

Fixes: d6c9e3f506ae8 ("net: ipa: only enable generic command completion IRQ when needed");
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: clear pending interrupts before enabling

We enable the completion interrupt for channel or event ring
commands only when we issue them. The interrupt is disabled after
the interrupt has fired, or after we have timed out waiting for it.

If we time out, the command could complete after the interrupt has
been disabled, causing a state change in the channel or event ring.
The interrupt associated with that state change would be delivered
the next time the completion interrupt is enabled.

To avoid previous command completions interfering with new commands,
clear all pending completion interrupts before re-enabling them for
a new command.

Fixes: b4175f8731f78 ("net: ipa: only enable GSI event control IRQs when needed")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: adjust GSI register addresses

The offsets for almost all GSI registers we use have different
offsets starting at IPA version 4.5. Only two registers remain
in their original location.

In a way though, the new register locations are not *that*
different. The entire group of affected registers has simply
been shifted down in memory by a fixed amount (0xd000). So for
example, the channel context 0 register that has a base offset of
0x0001c000 for "older" hardware now has a base offset of 0x0000f000.

This patch aims to add support for IPA v4.5 registers at their new
offets in a way that minimizes the amount of code that needs to
change. It is not ideal, but it avoids the need to maintain
a nearly complete set of additional register offset definitions.

The approach takes advantage of the fact that when accessing GSI
registers we do not access any of memory at lower end of the "gsi"
memory range (with two exceptions already noted). In particular,
we do not access anything within the bottom 0xd000 bytes of the
GSI memory range.

For IPA version 4.5, after we map the GSI memory, we adjust the
virtual memory pointer downward by the fixed amount (0xd000).
That way, register accesses using the offsets defined by the
existing GSI_REG_*() macros will resolve to the proper locations
for IPA version 4.5.

The two registers *not* affected by this offset are accessed only
in gsi_irq_setup(). There, for IPA version 4.5, we undo the general
register adjustment by adding the fixed amount back to the virtual
address to access these registers.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: update gsi registers for IPA v4.5

Very few GSI register definitions change for IPA v4.5, however
as a group their position in memory shifts a constant amount
(handled by the next commit).

Add definitions and update comments to the set of GSI registers to
support changes that come with IPA v4.5.

Update the logic in gsi_channel_program() to accommodate the new
(expanded) PREFETCH_MODE field in the CH_C_QOS register.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: reverse logic on escape buffer use

Starting with IPA v4.2 there is a GSI channel option to use an
"escape buffer" instead of prefetch buffers. This should be used
for all channels *except* the AP command TX channel. The logic
that implements this has it backwards; fix this bug.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: support retries on generic GSI commands

When stopping an AP RX channel, there can be a transient period
while the channel enters STOP_IN_PROC state before reaching the
final STOPPED state. In that case we make another attempt to stop
the channel.

Similarly, when stopping a modem channel (using a GSI generic
command issued from the AP), it's possible that multiple attempts
will be required before the channel reaches STOPPED state.

Add a field to the GSI structure to record an errno representing the
result code provided when a generic command completes. If the
result learned in gsi_isr_gp_int1() is RETRY, record -EAGAIN in the
result code, otherwise record 0 for success, or -EIO for any other
result.

If we time out nf gsi_generic_command() waiting for the command to
complete, return -ETIMEDOUT (as before). Otherwise return the
result stashed by gsi_isr_gp_int1().

Add a loop in gsi_modem_channel_halt() to reissue the HALT command
if the result code indicates -EAGAIN. Limit this to 10 retries
(after the initial attempt).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: ignore CHANNEL_NOT_RUNNING errors

IPA v4.2 has a hardware quirk that requires the AP to allocate GSI
channels for the modem to use. It is recommended that these modem
channels get stopped (with a HALT generic command) by the AP when
its IPA driver gets removed.

The AP has no way of knowing the current state of a modem channel.
So when the IPA driver issues a HALT command it's possible the
channel is not running, and in that case we get an error indication.
This error simply means we didn't need to stop the channel, so we
can ignore it.

This patch adds an explanation for this situation, and arranges for
this condition to *not* report an error message.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: don't reset an ALLOCATED channel

If the rmnet_ipa0 network device has not been opened at the time
we remove or shut down the IPA driver, its underlying TX and RX
GSI channels will not have been started, and they will still be
in ALLOCATED state.

The RESET command on a channel is meant to return a channel to
ALLOCATED state after it's been stopped. But if it was never
started, its state will still be ALLOCATED, the RESET command
is not required.

Quietly skip doing the reset without printing an error message if a
channel is already in ALLOCATED state when we request it be reset.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: print channel/event ring number on error

When a GSI command is used to change the state of a channel or event
ring we check the state before and after the command to ensure it is
as expected. If not, we print an error message, but it does not
include the channel or event ring id, and it easily can. Add the
channel or event ring id to these error messages.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use enumerated types for GSI field values

Replace constants defined with an "_FVAL" suffix with values defined
in enumerated types, to be consistent with other usage in the driver.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: move GSI command opcode values into "gsi_reg.h"

The gsi_ch_cmd_opcode, gsi_evt_cmd_opcode, and gsi_generic_cmd_opcode
enumerated types are values that fields in the GSI command registers
can take on. Move their definitions out of "gsi.c" and into "gsi_reg.h",
alongside the definition of registers they are associated with.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: move GSI error values into "gsi_reg.h"

The gsi_err_code and gsi_err_type enumerated types are values that
fields in the GSI ERROR_LOG register can take on. Move their
definitions out of "gsi.c" and into "gsi_reg.h", alongside the
definition of the ERROR_LOG register offset and field symbols.

Drop the "_ERR" suffix in the names of the gsi_err_code members.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: move channel type values into "gsi_reg.h"

The gsi_channel_type enumerated type define values used for the
channel type/protocol for event rings and channels. Move its
definition out of "gsi.c" and into "gsi_reg.h", alongside the
definition of the CH_C_CNTXT_0 register offset and its fields.
Add a comment near the definition of the EV_CH_E_CNTXT_0 register
indicating this type is used for its EV_CHTYPE field.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use common value for channel type and protocol

The numeric values that represent the event ring channel type are
identical to the values that represent the matching protocol used
for a channel. Use a new gsi_channel_type enumerated type to
represent the values programmed for both cases, using "CHANNEL_TYPE"
in member names in place of "EVT_CHTYPE" and "CHANNEL_PROTOCOL".

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define GSI interrupt types with enums

Define the GSI global interrupt types with an enumerated type whose
values are the bit positions representing the global interrupt types.

Similarly, define the GSI general interrupt types with an enumerated
type whose values are the bit positions of general interrupt types.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: drop an error message

There is no need for gsi_modem_channel_halt() to report an error,
because gsi_generic_command() will already have done that if the
command times out. So get rid of the extra message.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: pass a value to gsi_irq_type_update()

Now that all of the GSI interrupts are handled uniformly,
change gsi_irq_type_update() so it takes a value. Have the
function assign that value to the cached mask of enabled GSI
IRQ types before writing it to hardware.

Note that gsi_irq_teardown() will only be called after
gsi_irq_disable(), so it's not necessary for the former
to disable all IRQ types. Get rid of that.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: only enable GSI general IRQs when needed

Most GSI general errors are unrecoverable without a full reset.
Despite that, we want to receive these errors so we can at least
report what happened before whatever undefined behavior ensues.

Explicitly disable all such interrupts in gsi_irq_setup(), then
enable those we want in gsi_irq_enable(). List the interrupt types
we are interested in (everything but breakpoint) explicitly rather
than using GSI_CNTXT_GSI_IRQ_ALL, and remove that symbol's
definition.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: explicitly disallow inter-EE interrupts

It is possible for other execution environments (EEs, like the modem)
to request changes to local (AP) channel or event ring state. We do
not support this feature.

In gsi_irq_setup(), explicitly zero the mask that defines which
channels are permitted to generate inter-EE channel state change
interrupts. Do the same for the event ring mask.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: only enable GSI IEOB IRQs when needed

A GSI channel must be started in order to use it to perform a
transfer data (or command) transaction. And the only time we'll see
an IEOB interrupt is if we send a transaction to a started channel.
Therefore we do not need to have the IEOB interrupt type enabled
until at least one channel has been started. And once the last
started channel has been stopped, we can disable the IEOB interrupt
type again.

We already enable the IEOB interrupt for a particular channel only
when it is started. Extend that by having the IEOB interrupt *type*
be enabled only when at least one channel is in STARTED state.

Disallow all channels from triggering the IEOB interrupt in
gsi_irq_setup(). We only enable an channel's interrupt when
needed, so there is no longer any need to zero the channel mask
in gsi_irq_disable().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: only enable generic command completion IRQ when needed

The completion of a generic EE GSI command is signaled by a global
interrupt of type GP_INT1. The only other used type for a global
interrupt is a hardware error report.

First, disallow all global interrupt types in gsi_irq_setup(). We
want to know about hardware errors, so re-enable the interrupt type
in gsi_irq_enable(), to allow hardware errors to be reported.
Disable that interrupt type again in gsi_irq_disable().

We only issue generic EE commands one at a time, and there's no
reason to keep the completion interrupt enabled when no generic
EE command is pending. We furthermore have no need to enable the
GP_INT2 or GP_INT3 interrupt types (which aren't used).

The change in gsi_irq_enable() makes GSI_CNTXT_GLOB_IRQ_ALL unused,
so get rid of it. Have gsi_generic_command() enable the GP_INT1
interrupt type (in addition to the ERROR_INT type) only while a
generic command is pending.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: only enable GSI event control IRQs when needed

A GSI event ring causes an event control interrupt to fire whenever
its state changes (between NOT_ALLOCATED and ALLOCATED). No event
ring should ever change state except when we request it to.

Currently, we permit *all* events rings to generate event control
interrupts--even those that are never used. And we enable event
control interrupts essentially at all times, from setup to teardown.

Instead, only enable the event control interrupt type for the
duration of an event ring command, and when doing so, only allow
the event ring being operated upon to cause the interrupt to fire.
Disallow all event rings from issuing the event control interrupt
in gsi_irq_setup().

Because an event ring's interrupt is only enabled when needed,
there is no longer any need to zero the event channel mask in
gsi_irq_disable().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: only enable GSI channel control IRQs when needed

A GSI channel causes a channel control interrupt to fire whenever
its state changes (between NOT_ALLOCATED, ALLOCATED, STARTED, etc.).
We do not support inter-EE channel commands (initiated by other EEs),
so no channel should ever change state except when we request it to.

Currently, we permit *all* channels to generate channel control
interrupts--even those that are never used. And we enable channel
control interrupts essentially at all times, from setup to teardown.

Instead, disable all channel control interrupts initially in
gsi_irq_setup(), and only enable the channel control interrupt
type for the duration of a channel command. When doing so, only
allow the channel being operated upon to cause the interrupt to
fire.

Because a channel's interrupt is now enabled only when needed (one
channel at a time), there is no longer any need to zero the channel
mask in gsi_irq_disable().

Add new gsi_irq_type_enable() and gsi_irq_type_disable() as helper
functions to control whether a given GSI interrupt type is enabled.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: cache last-saved GSI IRQ enabled type

Keep track of the set of GSI interrupt types that are currently
enabled by recording the mask value to write (or last written) to
the TYPE_IRQ_MSK register.

Create a new helper function gsi_irq_type_update() to handle
actually writing the register.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: disable all GSI interrupt types initially

Introduce gsi_irq_setup() and gsi_irq_teardown() to disable all
GSI interrupts when first setting up GSI hardware, and to clean
things up when we're done.

Re-enable all GSI interrupt types in gsi_irq_enable(), but do
so only after each of the type-specific interrupt masks has
been configured. Similarly, disable all interrupt types in
gsi_irq_disable()--first--before zeroing out the type-specific
masks.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: define GSI interrupt types with an enum

Define the GSI interrupt types with an enumerated type whose values
are the bit positions representing each interrupt type. Include a
short comment describing how each interrupt type is used.

Build up the enabled interrupt mask explicitly in gsi_irq_enable(),
and get rid of the definition of GSI_CNTXT_TYPE_IRQ_MSK_ALL.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: rename gsi->event_enable_bitmap

Rename the "event_enable_bitmap" field of the GSI structure to be
"ieob_enabled_bitmap". An upcoming patch will cache the last value
stored for another interrupt mask and this is a more direct naming
convention to follow.

Add a few comments to explain the bitmap fields in the GSI structure.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: request GSI IRQ later

Introduce gsi_irq_init() and gsi_irq_exit(), to encapsulate looking
up the GSI IRQ and registering its handler. Call gsi_irq_init() a
little later in gsi_init(), and initialize the completion earlier.
The IRQ handler accesses both the GSI virtual memory pointer and the
completion, and this way these things will have been initialized
before the gsi_irq() can ever be called.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: eliminate legacy arguments

We enable a channel doorbell engine only for IPA v3.5.1, and that is
now handled directly by gsi_channel_program().

When initially setting up a channel, we want that doorbell engine
enabled, and we can request that independent of the IPA version.

Doing that makes the "legacy" argument to gsi_channel_setup_one()
unnecessary. And with that gone we can get rid of the "legacy"
argument to gsi_channel_setup(), and gsi_setup() as well.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use version in gsi_channel_program()

Use the IPA version in gsi_channel_program() to determine whether
we should enable the GSI doorbell engine when requested. This way,
callers only say whether or not it should be enabled if needed,
regardless of hardware version.

Rename the "legacy" argument to gsi_channel_reset(), and have
it indicate whether the doorbell engine should be enabled when
reprogramming following the reset.

Change all callers of gsi_channel_reset() to indicate whether to
enable the doorbell engine after reset, independent of hardware
version.

Rework a little logic in ipa_endpoint_reset() to get rid of the
"legacy" variable previously passed to gsi_channel_reset().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use version in gsi_channel_reset()

A quirk of IPA v3.5.1 requires a channel reset on an RX channel to
be performed twice. Use the IPA version in gsi_channel_reset()
rather than the passed-in legacy flag to determine that.

This is actually a bug fix, because this double reset is supposed
to occur independent of whether we're enabling the doorbell engine.
Now they will be independent.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: use version in gsi_channel_init()

A quirk of IPA v4.2 requires the AP to allocate the GSI channels
that are owned by the modem.

Rather than pass a flag argument to gsi_channel_init(), use the
IPA version directly in that function to determine whether modem
channels need to be allocated.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: record IPA version in GSI structure

Record the IPA version passed to gsi_init() in the GSI structure.
This allows that value to be used directly where needed, rather than
passing and storing certain flag arguments through the code.

In particular, for all but one supported version of IPA, the command
channel is programmed to only use an "escape buffer". By storing
the IPA version, we can do a simple version check in one location,
and avoid storing a flag field in every channel (and passing a flag
along while initializing channels to set that field properly).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: expose IPA version to the GSI layer

Although GSI is integral to IPA, it is a separate hardware component
and the IPA code supporting it has been structured to avoid explicit
dependence on IPA details. An example of this is that gsi_init() is
passed a number of Boolean flags to indicate special behaviors,
whose values are dependent on the IPA hardware version. Looking
ahead, newer hardware versions would require even more such special
behaviors.

For any given version of IPA hardware (like 3.5.1 or 4.2), the GSI
hardware version is fixed (in this case, 1.3 and 2.2, respectively).
So the IPA version *implies* the GSI version, and the IPA version
can be used as effectively the equivalent of the GSI hardware version.

Rather than proliferating new special behavior flags, just provide
the IPA version to the GSI layer when it is initialized. The GSI
code can then use that directly to determine whether special
behaviors are required. The IPA version enumerated type is already
isolated to its own header file, so the exposure of this IPA detail
is very limited.

For now, just change gsi_init() to pass the version rather than the
Boolean flags, and set the flag values internal to that function.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

net: ipa: fix two mild warnings

Fix two spots where a variable "channel_id" is unnecessarily
redefined inside loops in "gsi.c". This is warned about if
"W=2" is added to the build command.

Note that this problem is harmless, so there's no need to backport
it as a bugfix.

Remove a comment in gsi_init() about waking the system; the GSI
interrupt does not wake the system any more.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: share field mask values for GSI general interrupt

The GSI general interrupt is managed by three registers: enable;
status; and clear. The three registers have same set of field bits
at the same locations. Use a common set of field masks for all
three registers to avoid duplication.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: share field mask values for GSI global interrupt

The GSI global interrupt is managed by three registers: enable;
status; and clear. The three registers have same set of field bits
at the same locations. Use a common set of field masks for all
three registers to avoid duplication.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: share field mask values for GSI interrupt type

The GSI interrupt type register and interrupt type mask register
have the same field bits at the same locations. Use a common set of
field masks for both registers rather than essentially duplicating
them. The only place the interrupt mask register uses any of these
is in gsi_irq_enable().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: do not enable GSI interrupt for wakeup

We now trigger a system resume when we receive an IPA SUSPEND
interrupt. We should *not* wake up on GSI interrupts.

Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: fix kerneldoc comments

This commit affects comments (and in one case, whitespace) only.

Throughout the IPA code, return statements are documented using
"@Return:", whereas they should use "Return:" instead. Fix these
mistakes.

In function definitions, some parameters are missing their comment
to describe them. And in structure definitions, some fields are
missing their comment to describe them. Add these missing
descriptions.

Some arguments changed name and type along the way, but their
descriptions were not updated (an endpoint pointer is now used in
many places that previously used an endpoint ID). Fix these
incorrect parameter descriptions.

In the description for the ipa_clock structure, one field had a
semicolon instead of a colon in its description. Fix this.

Add a missing function description for ipa_gsi_endpoint_data_empty().

All of these issues were identified when building with "W=1".

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: standarize more GSI error messages

Make minor updates to error messages reported in "gsi.c":
- Use local variables to reduce multi-line function calls
- Don't use parentheses in messages
- Do some slight rewording in a few cases

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: always report GSI state errors

We check the state of an event ring or channel both before and after
any GSI command issued that will change that state. In most--but
not all--cases, if the state is something different than expected we
report an error message.

Add error messages where missing, so that all unexpected states
provide information about what went wrong. Drop the parentheses
around the state value shown in all cases.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: always check for stopped channel

In gsi_channel_stop(), there's a check to see if the channel might
have entered STOPPED state since a previous call, which might have
timed out before stopping completed.

That check actually belongs in gsi_channel_stop_command(), which is
called repeatedly by gsi_channel_stop() for RX channels.

Fixes: 650d1603825d ("soc: qcom: ipa: the generic software interface")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers: ipa: fix typoes for ipa

Change "transactio" -> "transaction". Also an alignment correction.

Signed-off-by: Wang Wenhu <wenhu.wang@vivo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: do not clear interrupt in gsi_channel_start()

In gsi_channel_start() there is harmless-looking comment "Clear the
channel's event ring interrupt in case it's pending". The intent
was to avoid getting spurious interrupts when first bringing up a
channel.

However we now use channel stop/start to implement suspend and
resume, and an interrupt pending at the time we resume is actually
something we don't want to ignore.

The very first time we bring up the channel we do not expect an
interrupt to be pending, and even if it were, the effect would
simply be to schedule NAPI on that channel, which would find nothing
to do, which is not a problem.

Stop clearing any pending IEOB interrupt in gsi_channel_start().
That leaves one caller of the trivial function gsi_isr_ieob_clear().
Get rid of that function and just open-code it in gsi_isr_ieob()
instead.

This fixes a problem where suspend/resume IPA v4.2 would get stuck
when resuming after a suspend.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: don't be a hog in gsi_channel_poll()

The iteration count value used in gsi_channel_poll() is intended to
limit poll iterations to the budget supplied as an argument. But
it's never updated.

Fix this bug by incrementing the count each time through the loop.

Reported-by: Sharath Chandra Vurukala <sharathv@codeaurora.com>
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: only reset channel twice for IPA v3.5.1

In gsi_channel_reset(), RX channels are subjected to two consecutive
CHANNEL_RESET commands. This workaround should only be used for IPA
version 3.5.1, and for newer hardware "can lead to unwanted behavior."

Only issue the second CHANNEL_RESET command for legacy hardware.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: rename db_enable flag

In several places, a Boolean flag is used in the GSI code to
indicate whether the "doorbell engine" should be enabled or not
when a channel is configured. This is basically done to abstract
this property from the IPA version; the GSI code doesn't otherwise
"know" what the IPA hardware version is. The doorbell engine is
enabled only for IPA v3.5.1, not for IPA v4.0 and later.

The next patch makes another change that affects behavior during
channel reset (which also involves programming the channel). It
also distinguishes IPA v3.5.1 hardware from newer hardware.

Rather than creating another flag whose value matches the "db_enable"
value, just rename "db_enable" to be "legacy" so it can be used to
signal more than just the special doorbell handling.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: do not cache channel state

It is possible for a GSI channel's state to be changed as a result
of an action by a different execution environment. Specifically,
the modem is able to issue a GSI generic command that causes a state
change on a GSI channel associated with the AP.

A channel's state only needs to be known when a channel is allocated
or deallocaed, started or stopped, or reset. So there is little
value in caching the state anyway.

Stop recording a copy of the channel's last known state, and instead
fetch the true state from hardware whenever it's needed. In such
cases, *do* record the state in a local variable, in case an error
message reports it (so the value reported is the value seen).

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: pass channel pointer to gsi_channel_state()

Pass a channel pointer rather than a GSI pointer and channel ID to
gsi_channel_state().

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: zero return code before issuing generic EE command

Zero the result code stored in a field of the scratch 0 register
before issuing a generic EE command. This just guarantees that
the value we read later was actually written as a result of the
command.

Also add the definitions of two more possible result codes that can
be returned when issuing flow control enable or disable commands:
INCORRECT_CHANNEL_STATE: - channel must be in started state
INCORRECT_DIRECTION - flow control is only valid for TX channels

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: ipa: fix an error message in gsi_channel_init_one()

An error message about limiting the number of TREs used prints the
wrong value. Fix this bug.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

soc: qcom: ipa: the generic software interface

This patch includes "gsi.c", which implements the generic software
interface (GSI) for IPA. The generic software interface abstracts
channels, which provide a means of transferring data either from the
AP to the IPA, or from the IPA to the AP. A ring buffer of "transfer
elements" (TREs) is used to describe data transfers to perform. The
AP writes a doorbell register associated with a channel to let it know
it has added new entries (for an AP->IPA channel) or has finished
processing entries (for an IPA->AP channel).

Each channel also has an event ring buffer, used by the IPA to
communicate information about events related to a channel (for
example, the completion of TREs). The IPA writes its own doorbell
register, which triggers an interrupt on the AP, to signal that
new event information has arrived.

Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>