14 Dec, 2020
2 commits
-
To i.MX8QM/QXP partitioned M4 case, M4 will do self recovery when
crashed. It not need Linux to reload firmware. So add a flag to indicate
there is no need to request firmware when recovery.Acked-by: Richard Zhu
Signed-off-by: Peng Fan -
To arm64, "dc zva, dst" is used in memset.
Per ARM DDI 0487A.j, chapter C5.3.8 DC ZVA, Data Cache Zero by VA,"If the memory region being zeroed is any type of Device memory,
this instruction can give an alignment fault which is prioritized
in the same way as other alignment faults that are determined
by the memory type."On i.MX platforms, when elf is loaded to onchip TCM area, the region
is ioremapped, so "dc zva, dst" will trigger abort. And ioremap_wc()
on i.MX not able to write correct data to TCM area.So we need to use io helpers, and extend the elf loader to support
platform specific memory functions.Acked-by: Richard Zhu
Signed-off-by: Peng Fan
14 Oct, 2020
1 commit
-
Currently "default" configuration option means coredumps are
enabled. To avoid confusion rename the "default" configuration
option to "enabled" and disable collection of dumps by default
as doing so makes sense for production devices.Signed-off-by: Rishabh Bhatnagar
Link: https://lore.kernel.org/r/1601662144-5964-2-git-send-email-rishabhb@codeaurora.org
Signed-off-by: Bjorn Andersson
05 Aug, 2020
1 commit
-
Add the character device interface into remoteproc framework.
This interface can be used in order to boot/shutdown remote
subsystems and provides a basic ioctl based interface to implement
supplementary functionality. An ioctl call is implemented to enable
the shutdown on release feature which will allow remote processors to
be shutdown when the controlling userspace application crashes or hangs.Reviewed-by: Bjorn Andersson
Reviewed-by: Mathieu Poirier
Signed-off-by: Rishabh Bhatnagar
Signed-off-by: Siddharth Gupta
Link: https://lore.kernel.org/r/1596044401-22083-2-git-send-email-sidgup@codeaurora.org
[bjorn: s/int32_t/s32/ per checkpatch]
Signed-off-by: Bjorn Andersson
22 Jul, 2020
2 commits
-
The current coredump implementation uses vmalloc area to copy
all the segments. But this might put strain on low memory targets
as the firmware size sometimes is in tens of MBs. The situation
becomes worse if there are multiple remote processors undergoing
recovery at the same time. This patch adds inline coredump
functionality that avoids extra memory usage. This requires
recovery to be halted until data is read by userspace and free
function is called.Reviewed-by: Bjorn Andersson
Reviewed-by: Sibi Sankar
Reviewed-by: Mathieu Poirier
Signed-off-by: Rishabh Bhatnagar
Tested-by: Sibi Sankar
Link: https://lore.kernel.org/r/1594938035-7327-5-git-send-email-rishabhb@codeaurora.org
Signed-off-by: Bjorn Andersson -
Change the segment dump API signature to include size and offset
arguments. Refactor the qcom_q6v5_mss driver to use these
arguments while copying the segment. Doing this lays the ground
work for "inline" coredump functionality being added in the next
patch.Tested-by: Sibi Sankar
Reviewed-by: Bjorn Andersson
Reviewed-by: Sibi Sankar
Reviewed-by: Mathieu Poirier
Signed-off-by: Rishabh Bhatnagar
Link: https://lore.kernel.org/r/1594938035-7327-4-git-send-email-rishabhb@codeaurora.org
Signed-off-by: Bjorn Andersson
18 Jul, 2020
1 commit
-
Make function rproc_resource_cleanup() public so that it can be
used by platform drivers when allocating resources to be used by
a detached remote processor.Acked-by: Arnaud Pouliquen
Signed-off-by: Mathieu Poirier
Link: https://lore.kernel.org/r/20200714200445.1427257-8-mathieu.poirier@linaro.org
Signed-off-by: Bjorn Andersson
17 Jul, 2020
3 commits
-
This patch prevents the firmware image name from being displayed when
the remoteproc core is attaching to a remote processor. This is needed
needed since there is no guarantee about the nature of the firmware
image that is loaded by the external entity.Signed-off-by: Mathieu Poirier
Reviewed-by: Arnaud Pouliquen
Reviewed-by: Bjorn Andersson
Tested-by: Arnaud Pouliquen
Link: https://lore.kernel.org/r/20200714195035.1426873-10-mathieu.poirier@linaro.org
Signed-off-by: Bjorn Andersson -
Add an new attach() operation in order to properly deal with
scenarios where the remoteproc core needs to attach to a
remote processor that has been booted by another entity.Signed-off-by: Mathieu Poirier
Reviewed-by: Bjorn Andersson
Tested-by: Arnaud Pouliquen
Link: https://lore.kernel.org/r/20200714195035.1426873-3-mathieu.poirier@linaro.org
Signed-off-by: Bjorn Andersson -
Add a new RPROC_DETACHED state to take into account scenarios
where the remoteproc core needs to attach to a remote processor
that is booted by another entity.Signed-off-by: Mathieu Poirier
Reviewed-by: Bjorn Andersson
Tested-by: Arnaud Pouliquen
Link: https://lore.kernel.org/r/20200714195035.1426873-2-mathieu.poirier@linaro.org
Signed-off-by: Bjorn Andersson
13 May, 2020
1 commit
-
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:struct foo {
int stuff;
struct boo array[];
};By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.Also, notice that, dynamic memory allocations won't be affected by
this change:"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")Signed-off-by: Gustavo A. R. Silva
Link: https://lore.kernel.org/r/20200507191943.GA16033@embeddedor
Signed-off-by: Bjorn Andersson
23 Apr, 2020
1 commit
-
On some SoC architecture, it is needed to enable HW like
clock, bus, regulator, memory region... before loading
co-processor firmware.This patch introduces prepare and unprepare ops to execute
platform specific function before firmware loading and after
stop execution.Signed-off-by: Loic Pallardy
Signed-off-by: Suman Anna
Reviewed-by: Bjorn Andersson
Reviewed-by: Mathieu Poirier
Link: https://lore.kernel.org/r/20200417002036.24359-2-s-anna@ti.com
Signed-off-by: Bjorn Andersson
21 Apr, 2020
2 commits
-
Add API functions devm_rproc_alloc() and devm_rproc_add(), which behave
like rproc_alloc() and rproc_add() respectively, but register their
respective cleanup function to be called on driver detach.Reviewed-by: Bjorn Andersson
Signed-off-by: Paul Cercueil
Link: https://lore.kernel.org/r/20200417170040.174319-2-paul@crapouillou.net
Signed-off-by: Bjorn Andersson -
For cases where @firmware is declared "const char *", use function
kstrdup_const() to avoid needlessly creating another copy on the
heap.Suggested-by: Bjorn Andersson
Signed-off-by: Mathieu Poirier
Reviewed-by: Alex Elder
Reviewed-by: Bjorn Andersson
Link: https://lore.kernel.org/r/20200420231601.16781-2-mathieu.poirier@linaro.org
Signed-off-by: Bjorn Andersson
20 Apr, 2020
1 commit
-
This function allows drivers to correctly setup the coredump output
elf information.Reviewed-by: Bjorn Andersson
Reviewed-by: Mathieu Poirier
Signed-off-by: Clement Leger
Link: https://lore.kernel.org/r/20200410102433.2672-2-cleger@kalray.eu
Signed-off-by: Bjorn Andersson
26 Mar, 2020
5 commits
-
Introduce generic support for handling kernel panics in remoteproc
drivers, in order to allow operations needed for aiding in post mortem
system debugging, such as flushing caches etc.The function can return a number of milliseconds needed by the remote to
"settle" and the core will wait the longest returned duration before
returning from the panic handler.Reviewed-by: Mathieu Poirier
Link: https://lore.kernel.org/r/20200324052904.738594-3-bjorn.andersson@linaro.org
Signed-off-by: Bjorn Andersson -
Now that remoteproc can load an elf64, coredump elf class should be
the same as the loaded elf class. In order to do that, add a
elf_class field to rproc with default values. If an elf is loaded
successfully, this field will be updated with the loaded elf class.
Then, the coredump core code has been modified to use the generic elf
macro in order to create an elf file with correct class.Reviewed-by: Mathieu Poirier
Reviewed-by: Bjorn Andersson
Signed-off-by: Clement Leger
Link: https://lore.kernel.org/r/20200302093902.27849-9-cleger@kalray.eu
Signed-off-by: Bjorn Andersson -
elf64 entry is defined as a u64. Since boot_addr is used to store the
elf entry point, change boot_addr type to u64 to support both elf32
and elf64. In the same time, fix users that were using this variable.Reviewed-by: Bjorn Andersson
Signed-off-by: Clement Leger
Link: https://lore.kernel.org/r/20200302093902.27849-4-cleger@kalray.eu
[bjorn: Fixes up return type of rproc_get_boot_addr()]
Signed-off-by: Bjorn Andersson -
Now that rproc_da_to_va uses a size_t for length, use a size_t for len
field of rproc_mem_entry. Function used to create such structures now
takes a size_t instead of int to allow full size range to be handled.Reviewed-by: Bjorn Andersson
Reviewed-by: Mathieu Poirier
Signed-off-by: Clement Leger
Link: https://lore.kernel.org/r/20200302093902.27849-3-cleger@kalray.eu
Signed-off-by: Bjorn Andersson -
With upcoming changes in elf loader for elf64 support, section size will
be a u64. When used with da_to_va, this will potentially lead to
overflow if using the current "int" type for len argument. Change
da_to_va prototype to use a size_t for len and fix all users of this
function.Reviewed-by: Bjorn Andersson
Reviewed-by: Mathieu Poirier
Signed-off-by: Clement Leger
Link: https://lore.kernel.org/r/20200302093902.27849-2-cleger@kalray.eu
Signed-off-by: Bjorn Andersson
30 Jun, 2019
1 commit
-
In order to allow rproc backend to handle vendor resources such as in
OpenAMP, add a handle_rsc hook. This hook allow the rproc backends to
handle vendor resources as they like. The hook will be called only for
vendor resources and should return RSC_HANDLED on successful resource
handling, RSC_IGNORED if resource was ignored, or a negative value on
error.Signed-off-by: Clement Leger
Signed-off-by: Bjorn Andersson
21 Feb, 2019
2 commits
-
Commit 7e83cab824a87e83cab824a8 ("remoteproc: Modify recovery path
to use rproc_{start,stop}()") replaces rproc_{shutdown,boot}() with
rproc_{stop,start}(), which skips destroy the virtio device at stop
but re-initializes it again at start.Issue is that struct virtio_dev is not correctly reinitialized like done
at initial allocation thanks to kzalloc() and kobject is considered as
already initialized by kernel. That is due to the fact struct virtio_dev
is allocated and released at vdev resource handling level managed and
virtio device is registered and unregistered at rproc subdevices level.Moreover kernel documentation mentions that device struct must be
zero initialized before calling device_initialize().This patch disentangles struct virtio_dev from struct rproc_vdev as
the two struct don't have the same life-cycle.struct virtio_dev is now allocated on rproc_start() and released
on rproc_stop().This patch applies on top of patch
remoteproc: create vdev subdevice with specific dma memory pool [1][1]: https://patchwork.kernel.org/patch/10755781/
Fixes: 7e83cab824a8 ("remoteproc: Modify recovery path to use rproc_{start,stop}()")
Reported-by: Xiang Xiao
Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson -
This patch creates a dedicated vdev subdevice for each vdev declared
in firmware resource table and associates carveout named "vdev%dbuffer"
(with %d vdev index in resource table) if any as dma coherent memory pool.Then vdev subdevice is used as parent for virtio device.
Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson
07 Jan, 2019
1 commit
-
Fix the kernel-doc comment for "parse_fw" and fix a typo.
Signed-off-by: Fabien Dessenne
Signed-off-by: Bjorn Andersson
20 Oct, 2018
2 commits
-
This patch adds a mechanism for assigning each rproc dump segment with
a custom dump function and private data. The dump function is to be
called for each rproc segment during coredump if assigned.Signed-off-by: Sibi Sankar
[bjorn: reordred arguments to rproc_coredump_add_custom_segment()]
Signed-off-by: Bjorn Andersson -
Introduce custom dump function and private data per remoteproc dump
segment. The dump function is responsible for filling the device memory
segment associated with coredumpSigned-off-by: Sibi Sankar
Signed-off-by: Bjorn Andersson
15 Oct, 2018
1 commit
-
Current version of rproc_alloc_vring function supports only dynamic vring
allocation.This patch allows to allocate vrings based on memory region declatation.
Vrings are now manage like memory carveouts, to communize memory management
code in rproc_alloc_registered_carveouts().Allocated buffer is retrieved in rp_find_vq() thanks to
rproc_find_carveout_by_name() functions for.This patch sets vrings names to vdev"x"vring"y" with x vdev index in
resource table and y vring index in vdev. This will be updated when
name will be associated to vdev in firmware resource table.Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson
10 Oct, 2018
6 commits
-
This patch introduces rproc_res_mem_entry_init() helper function to
allocate a rproc_mem_entry structure from a reserved memory region.
In that case, rproc_mem_entry structure has no alloc and release ops.
It will be used to assigned the specified reserved memory to any
rproc sub device.
Relation between rproc_mem_entry and rproc sub device will be done
by name.Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson -
Memory entry could be allocated in different ways (ioremap,
dma_alloc_coherent, internal RAM allocator...).
This patch introduces an alloc ops in rproc_mem_entry structure
to associate dedicated allocation mechanism to each memory entry
descriptor in order to do remote core agnostic from memory allocators.The introduction of this ops allows to perform allocation of all registered
carveout at the same time, just before calling rproc_start().
It simplifies and makes uniform carveout management whatever origin.Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson -
This patch introduces a new API to allow platform driver to register
platform specific carveout regions.Signed-off-by: Loic Pallardy
Acked-by: Bjorn Andersson
Signed-off-by: Bjorn Andersson -
This patch introduces rproc_mem_entry_init helper function to
simplify rproc_mem_entry structure allocation and filling by
client.Signed-off-by: Loic Pallardy
Signed-off-by: Bjorn Andersson -
Add name field in struct rproc_mem_entry.
This new field will be used to match memory area
requested in resource table with pre-registered carveout.Signed-off-by: Loic Pallardy
Acked-by: Bjorn Andersson
Signed-off-by: Bjorn Andersson -
Memory entry could be allocated in different ways (ioremap,
dma_alloc_coherent, internal RAM allocator...).
This patch introduces a release ops in rproc_mem_entry structure
to associate dedicated release mechanism to each memory entry descriptor
in order to keep remoteproc core generic.Signed-off-by: Loic Pallardy
Acked-by: Bjorn Andersson
Signed-off-by: Bjorn Andersson
06 Oct, 2018
1 commit
-
The commit ddf711872c9d ("remoteproc: Introduce auto-boot flag")
introduced the auto-boot flag but missed adding the corresponding
kernel-doc comment. Add the same.Signed-off-by: Suman Anna
Signed-off-by: Bjorn Andersson
27 Jun, 2018
3 commits
-
On rare occasions a subdevice might need to prepare some hardware
resources before a remote processor is booted, and clean up some
state after it has been shut down.One such example is the IP Accelerator found in various Qualcomm
platforms, which is accessed directly from both the modem remoteproc
and the application subsystem and requires an intricate lockstep
process when bringing the modem up and down.Tested-by: Fabien Dessenne
Signed-off-by: Bjorn Andersson
[elder@linaro.org: minor description and comment edits]
Signed-off-by Alex Elder
Signed-off-by: Bjorn Andersson -
In preparation of adding the additional prepare and unprepare operations
make the client responsible for filling out the function pointers of the
rproc_subdev. This makes the arguments to rproc_add_subdev() more
manageable, in particular when some of the functions are left out.Tested-by: Fabien Dessenne
Signed-off-by: Bjorn Andersson
[elder@linaro.org: added comment about assigning function pointers]
Signed-off-by Alex Elder
Signed-off-by: Bjorn Andersson -
"start" and "stop" are more suitable names for how these two operations
are used, and they fit better with the upcoming introduction of two
additional operations in the struct.Tested-by: Fabien Dessenne
Signed-off-by: Bjorn Andersson
[elder@linaro.org: minor comment edits]
Signed-off-by Alex Elder
Signed-off-by: Bjorn Andersson
26 Apr, 2018
1 commit
-
Fix rproc_add_subdev parameter name and inverse the crashed logic.
Fixes: 880f5b388252 ("remoteproc: Pass type of shutdown to subdev remove")
Reviewed-by: Alex Elder
Signed-off-by: Arnaud Pouliquen
Signed-off-by: Bjorn Andersson
13 Feb, 2018
2 commits
-
remoteproc instances can be stopped either by invoking shutdown or by an
attempt to recover from a crash. For some subdev types it's expected to
clean up gracefully during a shutdown, but are unable to do so during a
crash - so pass this information to the subdev remove functions.Acked-By: Chris Lew
Signed-off-by: Bjorn Andersson -
The resource table is just one possible source of information that can
be extracted from the firmware file. Generalize this interface to allow
drivers to override this with parsers of other types of information.Signed-off-by: Bjorn Andersson