| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Now some hwspinlock controllers did not have the requirement to implement
the PM runtime, but drivers must enable the PM runtime to comply with the
hwspinlock core.
Thus we can change the PM runtime support to be optional by validating
the -EACCES error number which means the PM runtime is not enabled, and
removing the return value validating of pm_runtime_put(). So that we
can remove some redundant PM runtime code in drivers.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
The original code use BUG_ON() to validate the parameters when locking
or unlocking one hardware lock, but we should not crash the whole kernel
though the hwlock parameters are incorrect, instead we can return
the error number for users and give some warning.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
Add the 'in_atomic' mode which can be called from an atomic context.
This mode relies on the existing 'raw' mode (no lock, no preemption/irq
disabling) with the difference that the timeout is not based on jiffies
(jiffies won't increase when irq are disabled) but handled with
busy-waiting udelay() calls.
Signed-off-by: Fabien Dessenne <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
Do not wait for hwspinlock device registration if it is not available
for use.
Acked-by: Suman Anna <[email protected]>
Signed-off-by: Fabien Dessenne <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
The commit 4f1acd758b08 ("hwspinlock: Add devm_xxx() APIs to request/free
hwlock") introduces one bug, that will return one error pointer if failed
to request one hwlock, but we expect NULL pointer on error for consumers.
This patch will fix this issue.
Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
Fix one comment mistake with correct function names.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
This patch introduces devm_hwspin_lock_register() and devm_hwspin_lock_unregister()
interfaces to help to register or unregister one hardware spinlock controller, that
will help to simplify the cleanup code for hwspinlock drivers.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
This patch introduces some devm_xxx() APIs to help to request or free
the hwlocks, which will help to simplify the cleanup code for drivers
requesting one hwlock, ensuring that the hwlock is automatically freed
whenever the device is unbound.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
The hardware spinlock binding already supplied the 'hwlock-names' property
to match and get a specific hwlock, but did not supply one API for users
to get a specific hwlock by the hwlock name. So this patch introduces one
API to support this requirement.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
Use the appropriate SPDX license identifier in the Hwspinlock core
driver source files and drop the previous boilerplate license text.
Signed-off-by: Suman Anna <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
In some scenarios, user need do some time-consuming or sleepable
operations under the hardware spinlock protection for synchronization
between the multiple subsystems.
For example, there is one PMIC efuse on Spreadtrum platform, which
need to be accessed under one hardware lock. But during the hardware
lock protection, the efuse operation is time-consuming to almost 5 ms,
so we can not disable the interrupts or preemption so long in this case.
Thus we can introduce one new mode to indicate that we just acquire the
hardware lock and do not disable interrupts or preemption, meanwhile we
should force user to protect the hardware lock with mutex or spinlock to
avoid dead-lock.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
We have different hwspinlock modes to select, thus it will be more
readable to handle different modes with using 'switch' statement
instead of 'if' statement.
Signed-off-by: Baolin Wang <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
|
|
radix_tree_is_indirect_ptr() is an internal API. The correct call to
use is radix_tree_deref_retry() which has the appropriate unlikely()
annotation.
Fixes: c6400ba7e13a ("drivers/hwspinlock: fix race between radix tree insertion and lookup")
Signed-off-by: Matthew Wilcox <[email protected]>
Cc: Konstantin Khlebnikov <[email protected]>
Cc: Kirill Shutemov <[email protected]>
Cc: Jan Kara <[email protected]>
Cc: Neil Brown <[email protected]>
Cc: Ross Zwisler <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
of_hwspin_lock_get_id() is protected by the RCU lock, which means that
insertions can occur simultaneously with the lookup. If the radix tree
transitions from a height of 0, we can see a slot with the indirect_ptr
bit set, which will cause us to at least read random memory, and could
cause other havoc.
Fix this by using the newly introduced radix_tree_iter_retry().
Signed-off-by: Matthew Wilcox <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Ohad Ben-Cohen <[email protected]>
Cc: Konstantin Khlebnikov <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
This patch adds a new OF-friendly API of_hwspin_lock_get_id()
for hwspinlock clients to use/request locks from a hwspinlock
device instantiated through a device-tree blob. This new API
can be used by hwspinlock clients to get the id for a specific
lock using the phandle + args specifier, so that it can be
requested using the available hwspin_lock_request_specific()
API.
Signed-off-by: Suman Anna <[email protected]>
Reviewed-by: Bjorn Andersson <[email protected]>
[small comment clarification]
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
Even in failed case of pm_runtime_get_sync, the usage_count
is incremented. In order to keep the usage_count with correct
value and runtime power management to behave correctly, call
pm_runtime_put_noidle in such case.
In __hwspin_lock_request, module_put is also called before
return in pm_runtime_get_sync failed case.
Cc: [email protected]
Signed-off-by Liu Chuansheng <[email protected]>
Signed-off-by: Li Fei <[email protected]>
[edit commit log]
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
The dereference should be moved below the NULL test.
spatch with a semantic match is used to found this.
(http://coccinelle.lip6.fr/)
Signed-off-by: Wei Yongjun <[email protected]>
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
Commit 300bab9770 (hwspinlock/core: register a bank of hwspinlocks in a
single API call, 2011-09-06) introduced 'hwspin_lock_register_single()'
to register numerous (a bank of) hwspinlock instances in a single API,
'hwspin_lock_register()'.
At which time, 'hwspin_lock_register()' accidentally passes 'local IDs'
to 'hwspin_lock_register_single()', despite that ..._single() requires
'global IDs' to register hwspinlocks.
We have to convert into global IDs by supplying the missing 'base_id'.
Cc: stable <[email protected]>
Signed-off-by: Shinya Kuribayashi <[email protected]>
[ohad: fix error path of hwspin_lock_register, too]
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
Hardware Spinlock devices usually contain numerous locks (known
devices today support between 32 to 256 locks).
Originally hwspinlock core required drivers to register (and later,
when needed, unregister) each lock separately.
That worked, but required hwspinlocks drivers to do a bit extra work
when they were probed/removed.
This patch changes hwspin_lock_{un}register() to allow a bank of
hwspinlocks to be {un}registered in a single invocation.
A new 'struct hwspinlock_device', which contains an array of 'struct
hwspinlock's is now being passed to the core upon registration (so
instead of wrapping each struct hwspinlock, a priv member has been added
to allow drivers to piggyback their private data with each hwspinlock).
While at it, several per-lock members were moved to be per-device:
1. struct device *dev
2. struct hwspinlock_ops *ops
In addition, now that the array of locks is handled by the core,
there's no reason to maintain a per-lock 'int id' member: the id of the
lock anyway equals to its index in the bank's array plus the bank's
base_id.
Remove this per-lock id member too, and instead use a simple pointers
arithmetic to derive it.
As a result of this change, hwspinlocks drivers are now simpler and smaller
(about %20 code reduction) and the memory footprint of the hwspinlock
framework is reduced.
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
Since we're using non-atomic radix tree allocations, we
should be protecting the tree using a mutex and not a
spinlock.
Non-atomic allocations and process context locking is good enough,
as the tree is manipulated only when locks are registered/
unregistered/requested/freed.
The locks themselves are still protected by spinlocks of course,
and mutexes are not involved in the locking/unlocking paths.
Cc: <[email protected]>
Signed-off-by: Juan Gutierrez <[email protected]>
[[email protected]: rewrite the commit log, #include mutex.h, add minor
commentary]
[[email protected]: update register/unregister parts in hwspinlock.txt]
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
hwspinlock devices provide system-wide hardware locks that are used
by remote processors that have no other way to achieve synchronization.
To achieve that, each physical lock must have a system-wide id number
that is agreed upon, otherwise remote processors can't possibly assume
they're using the same hardware lock.
Usually boards have a single hwspinlock device, which provides several
hwspinlocks, and in this case, they can be trivially numbered 0 to
(num-of-locks - 1).
In case boards have several hwspinlocks devices, a different base id
should be used for each hwspinlock device (they can't all use 0 as
a starting id!).
While this is certainly not common, it's just plain wrong to just
silently use 0 as a base id whenever the hwspinlock driver is probed.
This patch provides a hwspinlock_pdata structure, that boards can use
to set a different base id for each of the hwspinlock devices they may
have, and demonstrates how to use it with the omap hwspinlock driver.
While we're at it, make sure the hwspinlock core prints an explicit
error message in case an hwspinlock is registered with an id number
that already exists; this will help users catch such base id issues.
Reported-by: Arnd Bergmann <[email protected]>
Signed-off-by: Ohad Ben-Cohen <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
|
|
Use struct device_driver's owner member instead of asking drivers to
explicitly pass the owner again.
This simplifies drivers and also save some memory, since there's no
point now in maintaining a separate owner pointer per hwspinlock.
Signed-off-by: Ohad Ben-Cohen <[email protected]>
|
|
Add a platform-independent hwspinlock framework.
Hardware spinlock devices are needed, e.g., in order to access data
that is shared between remote processors, that otherwise have no
alternative mechanism to accomplish synchronization and mutual exclusion
operations.
Signed-off-by: Ohad Ben-Cohen <[email protected]>
Cc: Hari Kanigeri <[email protected]>
Cc: Benoit Cousson <[email protected]>
Cc: Kevin Hilman <[email protected]>
Cc: Grant Likely <[email protected]>
Cc: Paul Walmsley <[email protected]>
Cc: Russell King <[email protected]>
Acked-by: Arnd Bergmann <[email protected]>
Signed-off-by: Tony Lindgren <[email protected]>
|