A later change will make list_for_each_entry() cache the next element
before entering the loop body. for_each_fw_node() intentionally appends
newly discovered child nodes to the temporary walk list while the list is
being traversed.
Keep the loop open-coded so the next node is looked up only after
children have been appended. This preserves the current breadth-first
traversal semantics and prepares the code for the list iterator update.
Signed-off-by: Kaitao Cheng <chengkaitao@kylinos.cn>
Link: https://lore.kernel.org/r/20260609061347.93688-3-kaitao.cheng@linux.dev
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
It is preferable for the critical section to be as small as possible.
Current implementation of iso_resource_auto_work() function uses a
spinlock to control concurrent access to members of fw_card, fw_device,
iso_resource_auto structures, however the locking duration could be
reduced.
This commit refactors to shorten that duration.
Link: https://lore.kernel.org/r/20260501135823.241940-2-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Unlike FW_CDEV_IOC_ALLOCATE_ISO_RESOURCE operation, the operations of
FW_CDEV_IOC_[DE]ALLOCATE_ISO_RESOURCE_ONCE require no client resource,
thus they keeps no handle value.
This commit adds the series of functions to separate these operations,
according to divide-and-conquer methodology.
Link: https://lore.kernel.org/r/20260429093449.160545-5-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The add_client_resource() function checks the type of client resource
every time to be called. If the type is for iso_resource, it schedules
work item.
However, the iso_resource client resource is only added by the call of
init_iso_resource(). There is no need to check the type every time adding
any client resource.
Link: https://lore.kernel.org/r/20260429093449.160545-3-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
This is the exact same thing as the 'alloc_obj()' version, only much
smaller because there are a lot fewer users of the *alloc_flex()
interface.
As with alloc_obj() version, this was done entirely with mindless brute
force, using the same script, except using 'flex' in the pattern rather
than 'objs*'.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
Pull SCSI updates from James Bottomley:
"Usual driver updates (qla2xxx, mpi3mr, mpt3sas, ufs) plus assorted
cleanups and fixes.
The biggest core change is the massive code motion in the sd driver to
remove forward declarations and the most significant change is to
enumify the queuecommand return"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (78 commits)
scsi: csiostor: Fix dereference of null pointer rn
scsi: buslogic: Reduce stack usage
scsi: ufs: host: mediatek: Require CONFIG_PM
scsi: ufs: mediatek: Fix page faults in ufs_mtk_clk_scale() trace event
scsi: smartpqi: Fix memory leak in pqi_report_phys_luns()
scsi: mpi3mr: Make driver probing asynchronous
scsi: ufs: core: Flush exception handling work when RPM level is zero
scsi: efct: Use IRQF_ONESHOT and default primary handler
scsi: ufs: core: Use a host-wide tagset in SDB mode
scsi: qla2xxx: target: Add WQ_PERCPU to alloc_workqueue() users
scsi: qla2xxx: Add WQ_PERCPU to alloc_workqueue() users
scsi: qla4xxx: Add WQ_PERCPU to alloc_workqueue() users
scsi: mpi3mr: Driver version update to 8.17.0.3.50
scsi: mpi3mr: Fixed the W=1 compilation warning
scsi: mpi3mr: Record and report controller firmware faults
scsi: mpi3mr: Update MPI Headers to revision 39
scsi: mpi3mr: Use negotiated link rate from DevicePage0
scsi: mpi3mr: Avoid redundant diag-fault resets
scsi: mpi3mr: Rename log data save helper to reflect threaded/BH context
scsi: mpi3mr: Add module parameter to control threaded IRQ polling
...
Pull firewire updates from Takashi Sakamoto:
- Refactor page allocation dedicated to 1394 OHCI IR/IT/AR DMA contexts
Although 1394 OHCI specification does not impose any restriction on
the memory size dedicated to these DMA contexts, 1394 OHCI PCI driver
allocates pages for convenience when mapping them into either kernel
space or userspace VMA. The driver previously used dma_alloc_pages()
for both page allocation and mapping creation, even though this
kernel API is rarely used. Following discussions questioning the
page-oriented kernel API in the DMA layer, the driver has been
refactored to avoid using this API. In addition, the use of private
members in the allocated pages has been removed following
long-standing concern.
- Allocate variable-sized buffer for isochronous context header
1394 OHCI PCI driver previously allocated a single page for
isochronous context header. As a result, the buffer size for the
header was fixed to PAGE_SIZE, which imposed a limitation on IEC
61883-1/6 packet streaming engine. Consequently, the ALSA PCM devices
provided by drivers for audio and music units in IEEE 1394 bus were
constrained in the maximum size of buffer period (64 ms in most
cases). This limitation is resolved by dynamically allocating the
header buffer with an arbitrary size.
* tag 'firewire-updates-7.0' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
ALSA: firewire: remove PCM buffer size constraint from isoc context header
firewire: core: add fw_iso_context_create() variant with header storage size
firewire: core: provide isoc header buffer size outside card driver
firewire: ohci: allocate isoc context header by kvmalloc()
firewire: core: add flags member for isochronous context structure
firewire: ohci: use cleanup helper for isoc context header allocation
firewire: ohci: code refactoring to use union for isoc multiple channel state
firewire: ohci: refactor isoc single-channel state using a union
firewire: core: add function variants for isochronous context creation
firewire: ohci: fix index of pages for dma address to 1394 OHCI IT context
firewire: ohci: stop using page private to store DMA mapping address
firewire: ohci: split page allocation from dma mapping
firewire: ohci: use MAX macro to guarantee minimum count of pages for AR contexts
firewire: core: stop using page private to store DMA mapping address
firewire: core: use common kernel API to allocate and release a batch of pages
firewire: core: code refactoring with cleanup function for isoc pages
firewire: core: use mutex instead of spinlock for client isochronous context
firewire: core: move private function declaration from public header to internal header
In clang version 21.1 and later the -Wimplicit-enum-enum-cast warning
option has been introduced. This warning is enabled by default and can
be used to catch .queuecommand() implementations that return another
value than 0 or one of the SCSI_MLQUEUE_* constants. Hence this patch
that changes the return type of the .queuecommand() implementations from
'int' into 'enum scsi_qc_status'. No functionality has been changed.
Cc: Damien Le Moal <dlemoal@kernel.org>
Cc: John Garry <john.g.garry@oracle.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Link: https://patch.msgid.link/20260115210357.2501991-6-bvanassche@acm.org
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For single-channel isochronous contexts, the header storage size is
hard-coded to PAGE_SIZE. which is inconvenient for protocol
implementations requiring more space.
This commit refactors the code to obtain the header storage size outside
the 1394 OHCI driver.
Link: https://lore.kernel.org/r/20260117142823.440811-8-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
The fw_iso_callback union was added by a commit ebe4560ed5 ("firewire:
Remove function callback casts") to remove function pointer cast.
That change affected the cdev layer of the core code, but it is more
convenient for fw_iso_context_create() to accept the union directly.
This commit renames and changes the existing function to take the union
argument, and add static inline wrapper functions as variants.
Link: https://lore.kernel.org/r/20260117142823.440811-2-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
1394 OHCI PCI driver had long been the only user of dma_alloc_pages().
Although tee subsystem recently started using it, they are still a few
users of the infrequently-used function.
In the discussion for dma-mapping function, Jason Gunthorpe shows his
opinion about the design of public API for the function. According to it,
the users provide physical address to the function, then receive DMA
mapping address, regardless of the location of mapping target.
With the above aspects, this commit eliminates the use of
dma_alloc_pages() from this subsystem.
Link: https://lore.kernel.org/lkml/20250905174324.GI616306@nvidia.com/
Link: https://lore.kernel.org/r/20260110013911.19160-8-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
There is a restriction that the userspace client associated to a file
descriptor can hold one isochronous context. The client-level spinning
lock is used to guarantee it, however the lock is also used for
multi-purposes. Additionally, there is no need to use this type of lock,
and the mutex is available, instead.
Link: https://lore.kernel.org/r/20260110013911.19160-3-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Pull firewire updates from Takashi Sakamoto:
"This includes two changes for core functions, which affects all use
cases of this subsystem:
- Handle per-device interoperability quirks
Some devices have quirks affecting interoperability. To identify
such quirks at an early stages of device detection, the step for
reading the configuration ROM contents has been changed. As a side
effect, the entire detection process is now performed at the basic
transaction speed (S100), without a trial to probe higher
supported speeds.
With this change, the following devices should now work with fewer
issues:
- TASCAM FW-1884, FW-1804, and FW-1082
- MOTU Audio Express
- Safer removals of host card
There was a race condition between host card removal and handling
of bus reset events in the workqueue. This appears to be a long
standing issue, and recent changes to use more workqueues escalate
it.
To solve it, a new callback has been added to the 1394 OHCI PCI
driver to unregister the interrupt sources and wait for workqueue
completions when removing a card instance"
* tag 'firewire-updates-6.19' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394:
firewire: core: abort pending transactions at card removal
firewire: core: add WQ_UNBOUND to alloc_workqueue users
firewire: core: clear sources of hardware interrupt at card removal
firewire: core: code refactoring to find and pop transaction entry
firewire: core: code refactoring to remove transaction entry
firewire: core: use cleanup function to release cached configuration ROM
ALSA: firewire-tascam: reserve resources for transferred isochronous packets at S400
firewire: core: handle device quirk of TASCAM FW-1884/FW-1804/FW-1082
firewire: core: determine transaction speed after detecting quirks
firewire: core: code refactoring to compute transaction speed
firewire: core: handle device quirk of MOTU Audio Express
firewire: core: detect device quirk when reading configuration ROM
IEEE 1394 defines the split, concatenated, and unified transaction.
To support the split transaction, core function uses linked list to
maintain the transactions waiting for acknowledge packet. After clearing
sources of hardware interrupts, the acknowledge packet is no longer
handled, therefore it is required to abort the pending transactions.
This commit executes callback with RCODE_CANCELLED for the pending
transactions at card removal.
Link: https://lore.kernel.org/r/20251111223834.311287-1-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Currently if a user enqueues a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.
This lack of consistency cannot be addressed without refactoring the API.
alloc_workqueue() treats all queues as per-CPU by default, while unbound
workqueues must opt-in via WQ_UNBOUND.
This default is suboptimal: most workloads benefit from unbound queues,
allowing the scheduler to place worker threads where they’re needed and
reducing noise when CPUs are isolated.
This continues the effort to refactor workqueue APIs, which began with
the introduction of new workqueues and a new alloc_workqueue flag in:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
This change adds the WQ_UNBOUND flag to explicitly request
alloc_workqueue() to be unbound, because this specific workload has no
benefit being per-cpu.
With the introduction of the WQ_PERCPU flag (equivalent to !WQ_UNBOUND),
any alloc_workqueue() caller that doesn’t explicitly specify WQ_UNBOUND
must now use WQ_PERCPU.
Once migration is complete, WQ_UNBOUND can be removed and unbound will
become the implicit default.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Link: https://lore.kernel.org/r/20251112120125.124578-1-marco.crivellari@suse.com
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Due to the factors external to the system, hardware events may still be
handled while a card instance is being removed. The sources of hardware
IRQs should be cleared during card removal so that workqueues can be safely
destroyed.
This commit adds a disable callback to the underlying driver operations.
After this callback returns, the underlying driver guarantees that it
will no longer handle hardware events.
Link: https://lore.kernel.org/r/20251109065525.163464-1-o-takashi@sakamocchi.jp
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
