We configure the minimum and maximum number of various types of IPA
resources in ipa_resource_config(). It iterates over resource types
in the configuration data and assigns resource limits to each
resource group for each type.
Unfortunately, we are repeatedly initializing the resource data for
the first type, rather than initializing each of the types whose
limits are specified.
Fix this bug.
Fixes: 4a0d7579d4 ("net: ipa: avoid going past end of resource group array")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
Introduce a new function to abstract the knowledge of whether hashed
routing and filter tables are supported for a given IPA instance.
IPA v4.2 is the only one that doesn't support hashed tables (now
and for the foreseeable future), but the name of the helper function
is better for explaining what's going on.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
In ipa_cmd_register_write_valid() we verify that values we will
supply to a REGISTER_WRITE IPA immediate command will fit in
the fields that need to hold them. This patch fixes some issues
in that function and ipa_cmd_register_write_offset_valid().
The dev_err() call in ipa_cmd_register_write_offset_valid() has
some printf format errors:
- The name of the register (corresponding to the string format
specifier) was not supplied.
- The IPA base offset and offset need to be supplied separately to
match the other format specifiers.
Also make the ~0 constant used there to compute the maximum
supported offset value explicitly unsigned.
There are two other issues in ipa_cmd_register_write_valid():
- There's no need to check the hash flush register for platforms
(like IPA v4.2) that do not support hashed tables
- The highest possible endpoint number, whose status register
offset is computed, is COUNT - 1, not COUNT.
Fix these problems, and add some additional commentary.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
When initializing the IPA core clock and interconnects, it's
possible we'll get an EPROBE_DEFER error. This isn't really an
error, it's just means we need to be re-probed later.
Use dev_err_probe() to report the error rather than dev_err().
This avoids polluting the log with these "error" messages.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
It's possible that the length passed to ipa_header_size_encoded()
is larger than what can be represented by the HDR_LEN field alone
(starting with IPA v4.5). If we attempted that, u32_encode_bits()
would trigger a build-time error.
Avoid this problem by masking off high-order bits of the value
encoded as the lower portion of the header length.
The same sort of problem exists in ipa_metadata_offset_encoded(),
so implement the same fix there.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is a build-time check that the packet status structure is a
multiple of 4 bytes in size. It's not clear where that constraint
comes from, but the structure defines what hardware provides so its
definition won't change. Get rid of the check; it adds no value.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The count argument to ipa_endpoint_replenish() is only ever 0 or 1,
and always will be (because we always handle each receive buffer in
a single transaction). Rename the argument to be add_one and change
it to be Boolean.
Update the function description to reflect the current code.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We do not support inter-EE channel or event ring commands. Inter-EE
interrupts are disabled (and never re-enabled) for all channels and
event rings, so we have no need for the GSI registers that clear
those interrupt conditions. So remove their definitions.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
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>
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>
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: 650d160382 ("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>
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>
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>
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>
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>
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>
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>
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>
Fix two format specifiers that used %lu for a size_t in "ipa_mem.c".
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When extracting the destination endpoint ID from the status in
ipa_endpoint_status_skip(), u32_get_bits() is used. This happens to
work, but it's wrong: the structure field is only 8 bits wide
instead of 32.
Fix this by using u8_get_bits() to get the destination endpoint ID.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Sparse warns that the assignment of the metadata mask for a QMAP
endpoint in ipa_endpoint_init_hdr_metadata_mask() is a bad
assignment. We know we want the mask value to be big endian, even
though the value we write is in host byte order. Use a __force
tag to indicate we really mean it.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
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: 650d160382 ("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>
The only time we transfer data (rather than issuing a command) out
of the AP->command TX endpoint is when we're clearing the hardware
pipeline. All that's needed is a "small" data buffer, and its
contents aren't even important.
For convenience, we just transfer a command structure in this case
(it's already mapped for DMA). The TRE is added to a transaction
using ipa_cmd_ip_tag_status_add(), but we ignore the size value
provided to that function. So just get rid of the size argument.
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We only send a tagged packet from the AP->command TX endpoint when
we're clearing the hardware pipeline. And when we receive the
tagged packet we don't care what the actual tag value is.
Stop passing a tag value to ipa_cmd_ip_tag_status_add(), and just
encode 0 as the tag sent. Fix the function that encodes the tag so
it uses the proper byte ordering.
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There are times, such as when the modem crashes, when we issue
commands to clear the IPA hardware pipeline. These commands include
a data transfer command that delivers a small packet directly to the
default (AP<-LAN RX) endpoint.
The places that do this wait for the transactions that contain these
commands to complete, but the pipeline can't be assumed clear until
the sent packet has been *received*.
The small transfer will be delivered with a status structure, and
that status will indicate its tag is valid. This is the only place
we send a tagged packet, so we use the tag to determine when the
pipeline clear packet has arrived.
Add a completion to the IPA structure to to be used to signal
the receipt of a pipeline clear packet. Create a new function
ipa_cmd_pipeline_clear_wait() that will wait for that completion.
Reinitialize the completion whenever pipeline clear commands are
added to a transaction. Extend ipa_endpoint_status_tag() to check
whether a packet whose status contains a valid tag was sent from the
AP->command TX endpoint, and if so, signal the new IPA completion.
Have all callers of ipa_cmd_pipeline_clear_add() wait for the
pipeline clear indication after the transaction that clears the
pipeline has completed.
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Introduce ipa_endpoint_status_tag(), which returns true if received
status indicates its tag field is valid. The endpoint parameter is
not yet used.
Call this from ipa_status_drop_packet(), and drop the packet if the
status indicates the tag was valid. Pass the endpoint pointer to
ipa_status_drop_packet(), and rename it ipa_endpoint_status_drop().
The endpoint will be used in the next patch.
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Rearrange some comments and assignments made when handling a packet
that is received with status, aiming to improve understandability.
Use DIV_ROUND_CLOSEST() to get a better per-packet true size estimate.
Signed-off-by: Alex Elder <elder@linaro.org>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is a set of functions and symbols related to performing
"tag_process" immediate commands to clear the IPA pipeline. The
name is related to one of the commands issued when doing this, but
it doesn't really convey the overall purpose of taking this action.
The purpose is to take some steps to "clear out" the hardware
pipeline, and to wait until that process completes, to ensure the
IPA hardware is in a well-defined state.
Rename these symbols to use "pipeline_clear" in their names instead.
Add some comments to explain a bit more about what's going 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>
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>
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>
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>
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>
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>
The IPA driver currently requires a DT property to be defined whose
value is the phandle for the modem subsystem. This was needed to
look up a remoteproc structure pointer used when registering for
notifications in the original IPA notification mechanism.
Remoteproc provides a more generic SSR notifier system, and the IPA
driver switched over to it last summer, but this remoteproc phandle
dependency was not removed at that time.
Get rid of the IPA remoteproc pointer and stop requiring the phandle
be specified.
This avoids a link error (rproc_put() not defined) for certain
configurations.
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently we assume that the IPA hardware has exactly three
interconnects. But that won't be guaranteed for all platforms,
so allow any number of interconnects to be specified in the
configuration data.
For each platform, define an array of interconnect data entries
(still associated with the IPA clock structure), and record the
number of entries initialized in that array.
Loop over all entries in this array when initializing, enabling,
disabling, or tearing down the set of interconnects.
With this change we no longer need the ipa_interconnect_id
enumerated type, so get rid of it.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pass an the address of an IPA interconnect structure and its
configuration data to ipa_interconnect_init_one() and have that
function initialize all the structure's fields. Change the function
to simply return an error code.
Introduce ipa_interconnect_exit_one() to encapsulate the cleanup of
an IPA interconnect structure.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add the name to the configuration data for each interconnect. Use
this information rather than a constant string during initialization.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add fields in the ipa_interconnect structure to hold the average and
peak bandwidth values for the interconnect. Pass the configuring
data for interconnects to ipa_interconnect_init() so these values
can be recorded, and use them when enabling the interconnects.
There's no longer any need to keep a copy of the interconnect data
after initialization.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Rather than having separate pointers for the memory, imem, and
config interconnect paths, maintain an array of ipa_interconnect
structures each of which contains a pointer to a path.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If disabling interconnects fails there's not a lot we can do. The
only two callers of ipa_interconnect_disable() ignore the return
value, so just give the function a void return type.
Print an error message if disabling any of the interconnects is not
successful. Return (and print) only the first error seen.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use "bandwidth" rather than "rate" in describing the average and
peak values to use for IPA interconnects. They should have been
named that way to begin with.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
