The layer abstracting the building of commands and extracting
responses is currently based on macros that shift and mask the command
fields and requires exposing offset/size values as macro parameters
and makes the code harder to read.
For clarity and maintainability, instead use an implementation based on
mapping the MC command definitions to C structures. These structures
contain the hardware command fields (which are naturally-aligned)
and individual fields are little-endian ordering (the byte ordering
of the hardware).
As such, there is no need to perform the conversion between core and
hardware (LE) endianness in mc_send_command(), but instead each
individual field in a command will be converted separately if needed
by the function building the command or extracting the response.
This patch does not introduce functional changes, both the hardware
ABIs and the APIs exposed for the DPAA2 objects remain the same.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For an MSI domain the hwirq is an arbitrary but unique
id to identify an interrupt. Previously the hwirq was set to
the MSI index of the interrupt, but that only works if there is
one DPRC. Additional DPRCs require an expanded namespace. Use
both the ICID (which is unique per DPRC) and the MSI index to
compose a hwirq value.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When unbinding a dprc from the dprc driver the cleanup of
the resource pools must happen after irq pool cleanup
is done.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
add missing free of the Linux irq when tearing down interrupts
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An mc_io represents a mapped MC portal. Previously, an mc_io was
created for the root dprc in fsl_mc_bus_probe() and for child dprcs
in dprc_probe(). But the free of that data structure happened in the
general bus remove callback. This asymmetry resulted in some bugs due
to unwanted destroys of mc_io object in some scenarios (e.g. vfio).
Fix this bug by making things symmetric-- mc_io created in
fsl_mc_bus_probe() is freed in fsl_mc_bus_remove(). The mc_io created
in dprc_probe() is freed in dprc_remove().
Signed-off-by: Bharat Bhushan <Bharat.Bhushan@nxp.com>
[Stuart: added check for root dprc and reworded commit message]
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
make fsl_mc_is_root_dprc() global so that the dprc driver
can use it
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
some drivers (built as modules) rely on mc_get_version()
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move the definition of fsl_mc_device_id to its proper location in
mod_devicetable.h, and add fsl-mc bus support to devicetable-offsets.c
and file2alias.c to enable device table matching. With this patch udev
based module loading of fsl-mc drivers is supported.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-rename the struct used for fsl-mc device ids to be more
consistent with other busses
-remove the now obsolete and unused version fields
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Replace placeholder code in the uevent callback to properly
set the MODALIAS env variable.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In order to support uevent based module loading implement modalias support
for the fsl-mc bus driver. Aliases are based on vendor and object/device
id and are of the form "fsl-mc:vNdN".
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When updating the irq_chip and msi_domain_ops, the code checkes for
already present functions.
When more then one ITS controller are present in the system,
irq_chip and msi_domain_ops got already set and a warning is invoked.
This patch deletes the warning, as the funtions are just already set to
the needed callbacks.
Signed-off-by: Matthias Brugger <mbrugger@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are some error paths that allow for a NULL new_mc_io and err = 0
return code. Return -EINVAL instead.
Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dpseci objects < 4.0 are not coherent-- in spite of the fact
that the MC reports them to be coherent in certain versions.
Add a special case to set the no shareability flag for dpseci
objects < 4.0.
Signed-off-by: Horia Geanta <horia.geanta@nxp.com>
(Stuart: reworded commit message, updated comment in patch)
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The dprc driver supports dprc version 5.0 and above.
This patch adds the code to check the version.
Signed-off-by: Itai Katz <itai.katz@nxp.com>
(Stuart: resolved merge conflicts, split dpseci quirk into separate patch)
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The root dprc is discovered as a platform device in the device tree. The
version of that dprc was previously set using hardcoded values from the API
header in the kernel). This patch removes the use of the hardcoded version
numbers and instead reads the actual dprc version from the hardware.
Signed-off-by: Itai Katz <itai.katz@nxp.com>
(Stuart: resolved merge conflict, updated commit subject/log)
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some DPAA2 devices have mmio regions that should be mapped as
cacheable by drivers. Set IORESOURCE_CACHEABLE in the region's
flags if applicable.
Signed-off-by: Itai Katz <itai.katz@nxp.com>
[Stuart: update subject and commit message]
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Unless discovered devices have the no shareability flag set,
set up coherent dma ops for them.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Up until now if the object version expected by a driver (in the API header
file) did not match the actual object version in the MC hardware the bus
driver refused to bind the object to the driver or printed out WARN_ON
dumps.
This patch removes those checks, and the responsibility of object version
checking should now be done in the object drivers themselves. If the actual
version discovered is not supported, the driver's probe function should fail.
Drivers should use version checks to support new features and provide
backwards compatibility if at all possible.
This patch also removes the checks that caused bus driver probing to fail
if the overall MC version discovered did not match the firmware version
from the API header...this was too strict. The overall MC version is
informational like a release number, and continues to be printed in the
boot log.
Signed-off-by: Itai Katz <itai.katz@nxp.com>
(Stuart: reworded commit log)
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The meaning of the "status" parameter in dprc_get_irq_status
has changed, and this patch updates the flib and caller
of the API.
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
incorporated feedback from review comments, other misc cleanup/tweaks
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
remove 3 of the remaining TODO items:
-multiple root fsl-mc buses-- done in patch series starting with
commit 14f928054a ("staging: fsl-mc: abstract test for existence
of fsl-mc bus")
-interrupt support-- done in patch series starting with
commit 9b1b282ccd ("irqdomain: Added domain bus token
DOMAIN_BUS_FSL_MC_MSI")
-MC command serialization-- done in commit 63f2be5c3b ("staging:
fsl-mc: Added serialization to mc_send_command()")
Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Acked-by: German Rivera <german.rivera@nxp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove mutex_locked variables which are used to determine whether mutex is
locked, instead add another label to unlock mutex on premature exits due to
an error.
This patch also addresses the folowing warnings reported by coccinelle:
drivers/staging/fsl-mc/bus/mc-allocator.c:237:1-7: preceding lock on line 204
drivers/staging/fsl-mc/bus/mc-allocator.c:89:1-7: preceding lock on line 57
drivers/staging/fsl-mc/bus/mc-allocator.c:157:1-7: preceding lock on line 124
Signed-off-by: Cihangir Akturk <cakturk@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dev_err macros expect const struct device ** as its second
argument, but here the argument we are passing is of typ
struct device **. This patch fixes this error.
Fixes: 454b0ec8bf ("Staging: fsl-mc: Replace pr_err with dev_err")
Cc: Bhumika Goyal <bhumirks@gmail.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Cihangir Akturk <cakturk@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dev_dbg macros expect const struct device ** as its second
argument but here the argument we are passing is of type
struct device ** this patch fixes this error.
Fixes: de71daf5c8 ("Staging: fsl-mc: Replace pr_debug with dev_dbg")
Cc: Bhumika Goyal <bhumirks@gmail.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Cihangir Akturk <cakturk@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A call to irq_find_matching_host was already made and the result
has been stored in mc_msi_domain. mc_msi_domain is again reassigned
using the same function call which is redundant.
irq_find_matching_host returns/locates a domain for a given fwnode.
The domain is identified using device node and bus_token(if several
domains have same device node but different purposes they can be
distinguished using bus-specific token).
http://www.bricktou.com/include/linux/irqdomain_irq_find_matching_host_en.html
Also, of_property_read_bool finds and reads a boolean from a property
device node from which the property value is to be read. It doesn't
alter the device node.
http://lists.infradead.org/pipermail/linux-arm-kernel/2012-February/083698.html
Since, both the function calls have the same device node and bus_token,
the return values shall be the same. Hence, the second call has been
removed.
This was done using Coccinelle:
@r@
idexpression *x;
identifier f;
position p1,p2;
@@
x@p1 = f(...)
... when != x
(
x@p2 = f(...)
)
@script:python@
p1 << r.p1;
p2 << r.p2;
@@
if (p1[0].line == p2[0].line):
cocci.include_match(False)
@@
idexpression *x;
identifier f;
position r.p1,r.p2;
@@
*x@p1 = f(...)
...
*x@p2 = f(...)
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For platform_driver, we don't need to set .owner field as is set by
platform driver core. The semantic patch used here first checks whether
platform_driver struct was actually used in a call to set the .owner
field.
The coccinelle script that generated the patch can be found here:
http://www.spinics.net/lists/kernel/msg2029903.html
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The fsl_mc_allocator_driver_exit() function is marked __exit, but is
called by the error handling code in fsl_mc_allocator_driver_init().
This results in a section mismatch, which in turn could lead to
executing random code.
Remove the __exit annotation to fix this.
Cc: J. German Rivera <German.Rivera@freescale.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This driver uses functionality (MSI IRQ domain) whose symbols aren't
exported, and hence the modular build fails. While arguably there might
be reasons to make these symbols available to modules, that change would
be fairly involved and the set of exported functions should be carefully
auditioned. Fix the build failure for now by marking the driver boolean.
Cc: J. German Rivera <German.Rivera@freescale.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove unneeded parentheses on the right hand side of assignment
statements.
Semantic patch:
@@
expression a, b, c;
@@
(
a = (b == c)
|
a =
- (
b
- )
)
Signed-off-by: Janani Ravichandran <janani.rvchndrn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Initialize/Cleanup ITS-MSI support for the MC bus driver at driver
init/exit time. Associate an MSI domain with each DPAA2 child device.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The interrupt handler for DPRC IRQs is added. DPRC IRQs are
generated for hot plug events related to DPAA2 objects in a given
DPRC. These events include, creating/destroying DPAA2 objects in
the DPRC, changing the "plugged" state of DPAA2 objects and moving
objects between DPRCs.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Destroy mc_io in error path in dprc_probe() only if the mc_io was
created in this function.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
THE MSI domain associated with a root DPRC object is
obtained form the device tree. Child DPRCs inherit
the parent DPRC MSI domain.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Scan the corresponding DPRC container to get total count
of IRQs needed by all its child DPAA2 objects. Then,
preallocate a set of MSI IRQs with the DPRC's ICID
(GIT-ITS device Id) to populate the the DPRC's IRQ pool.
Each child DPAA2 object in the DPRC and the DPRC object itself
will allocate their necessary MSI IRQs from the DPRC's IRQ pool,
in their driver probe function.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The DPRC built-in portal's mc_io is used to send commands to the MC
to program MSIs for MC objects. This is done by the
fsl_mc_msi_write_msg() callback, which is invoked by the generic MSI
layer with interrupts disabled. As a result, the mc_io used in
fsl_mc_msi_write_msg needs to be an atomic mc_io.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
All the IRQs for DPAA2 objects in the same DPRC must use
the ICID of that DPRC, as their device Id in the GIC-ITS.
Thus, all these IRQs must share the same ITT table in the GIC.
As a result, a pool of IRQs with the same device Id must be
preallocated per DPRC (fsl-mc bus instance). So, the fsl-mc
bus object allocator is extended to also provide services
to allocate IRQs to DPAA2 devices, from their parent fsl-mc bus
IRQ pool.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Added platform-specific MSI support layer for FSL-MC devices.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Created an MSI domain for the fsl-mc bus-- including functions
to create a domain, find a domain, alloc/free domain irqs, and
bus specific overrides for domain and irq_chip ops.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
WARNING: drivers/staging/built-in.o(.init.text+0xdc): Section mismatch in reference from the function fsl_mc_bus_driver_init() to the function .exit.text:dprc_driver_exit()
The function __init fsl_mc_bus_driver_init() references
a function __exit dprc_driver_exit().
This is often seen when error handling in the init function
uses functionality in the exit path.
The fix is often to remove the __exit annotation of
dprc_driver_exit() so it may be used outside an exit section.
Signed-off-by: Lijun Pan <Lijun.Pan@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When the same portal is used to call mc_send_command() from two
different threads or a thread and an interrupt handler, serialization
is required, as the MC only supports one outstanding command per MC
portal. Thus, a new command should not be sent to the MC until the
last command sent has been responded by the MC.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Refactored mc_send_command() to support two flavors of polling:
- preemptible (for non-atomic portals), which was already supported.
It calls usleep_range() between polling iterations.
- non-preemptible (for atomic portals), which is needed when
mc_send_command() is called with interrupts disabled.
It calls udelay() between polling iterations.
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Moved wait logic in mc_send_command() to its own function
Signed-off-by: J. German Rivera <German.Rivera@freescale.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
