The following circular locking dependency was reported when running
cpus online/offline test on an arm64 system.
[ 84.195923] Chain exists of:
dmc620_pmu_irqs_lock --> cpu_hotplug_lock --> cpuhp_state-down
[ 84.207305] Possible unsafe locking scenario:
[ 84.213212] CPU0 CPU1
[ 84.217729] ---- ----
[ 84.222247] lock(cpuhp_state-down);
[ 84.225899] lock(cpu_hotplug_lock);
[ 84.232068] lock(cpuhp_state-down);
[ 84.238237] lock(dmc620_pmu_irqs_lock);
[ 84.242236]
*** DEADLOCK ***
The following locking order happens when dmc620_pmu_get_irq() calls
cpuhp_state_add_instance_nocalls().
lock(dmc620_pmu_irqs_lock) --> lock(cpu_hotplug_lock)
On the other hand, the calling sequence
cpuhp_thread_fun()
=> cpuhp_invoke_callback()
=> dmc620_pmu_cpu_teardown()
leads to the locking sequence
lock(cpuhp_state-down) => lock(dmc620_pmu_irqs_lock)
Here dmc620_pmu_irqs_lock protects both the dmc620_pmu_irqs and the
pmus_node lists in various dmc620_pmu instances. dmc620_pmu_get_irq()
requires protected access to dmc620_pmu_irqs whereas
dmc620_pmu_cpu_teardown() needs protection to the pmus_node lists.
Break this circular locking dependency by using two separate locks to
protect dmc620_pmu_irqs list and the pmus_node lists respectively.
Suggested-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Waiman Long <longman@redhat.com>
Link: https://lore.kernel.org/r/20230812235549.494174-1-longman@redhat.com
Signed-off-by: Will Deacon <will@kernel.org>
Add a cpumask for the DMC620 PMU. As it is an uncore PMU, perf
userspace tool only needs to open a single counter on the CPU
specified by the CPU mask for each event on a given DMC620 device.
Signed-off-by: Xin Yang <xin.yang@arm.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/20230613013423.2078397-1-xin.yang@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
According to commit 890cc39a87 ("drivers: provide
devm_platform_get_and_ioremap_resource()"), convert
platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Link: https://lore.kernel.org/r/20230315023017.35789-1-yang.lee@linux.alibaba.com
Signed-off-by: Will Deacon <will@kernel.org>
dmc620_pmu_init() won't remove the callback added by
cpuhp_setup_state_multi() when platform_driver_register() failed. Remove
the callback by cpuhp_remove_multi_state() in fail path.
Similar to the handling of arm_ccn_init() in commit 26242b3300 ("bus:
arm-ccn: Prevent hotplug callback leak")
Fixes: 53c218da22 ("driver/perf: Add PMU driver for the ARM DMC-620 memory controller")
Signed-off-by: Shang XiaoJing <shangxiaojing@huawei.com>
Reviewed-by: Punit Agrawal <punit.agrawal@bytedance.com>
Link: https://lore.kernel.org/r/20221115115540.6245-2-shangxiaojing@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
The driver uses irq_set_affinity_hint() to set the affinity for the PMU
interrupts, which relies on the undocumented side effect that this function
actually sets the affinity under the hood.
Setting an hint is clearly not a guarantee and for these PMU interrupts an
affinity hint, which is supposed to guide userspace for setting affinity,
is beyond pointless, because the affinity of these interrupts cannot be
modified from user space.
Aside of that the error checks are bogus because the only error which is
returned from irq_set_affinity_hint() is when there is no irq descriptor
for the interrupt number, but not when the affinity set fails. That's on
purpose because the hint can point to an offline CPU.
Replace the mindless abuse with irq_set_affinity().
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Will Deacon <will@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Acked-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/20210518093118.395086573@linutronix.de
Signed-off-by: Will Deacon <will@kernel.org>
sprintf does not know the PAGE_SIZE maximum of the temporary buffer
used for sysfs content and it's possible to overrun the buffer length.
Use sysfs_emit() function to ensures that no overrun is done.
Signed-off-by: Qi Liu <liuqi115@huawei.com>
Link: https://lore.kernel.org/r/1616148273-16374-4-git-send-email-liuqi115@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
Fix to return negative error code -ENOMEM from the error handling
case instead of 0, as done elsewhere in this function.
Fixes: 53c218da22 ("driver/perf: Add PMU driver for the ARM DMC-620 memory controller")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Link: https://lore.kernel.org/r/20210312080421.277562-1-weiyongjun1@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
Set "suppress_bind_attrs" to true, so that bind/unbind can be
disabled via sysfs and prevent unbinding ARM_DMC620_PMU drivers
during perf sampling.
Signed-off-by: Qi Liu <liuqi115@huawei.com>
Link: https://lore.kernel.org/r/1612252686-50329-1-git-send-email-liuqi115@huawei.com
Signed-off-by: Will Deacon <will@kernel.org>
The only usage is to put their addresses in an array of pointers to
const struct attribute group. Make them const to allow the compiler
to put them in read-only memory.
Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Link: https://lore.kernel.org/r/20210117212847.21319-5-rikard.falkeborn@gmail.com
Signed-off-by: Will Deacon <will@kernel.org>
DMC-620 PMU supports total 10 counters which each is
independently programmable to different events and can
be started and stopped individually.
Currently, it only supports ACPI. Other platforms feel free to test and add
support for device tree.
Usage example:
#perf stat -e arm_dmc620_10008c000/clk_cycle_count/ -C 0
Get perf event for clk_cycle_count counter.
#perf stat -e arm_dmc620_10008c000/clkdiv2_allocate,mask=0x1f,match=0x2f,
incr=2,invert=1/ -C 0
The above example shows how to specify mask, match, incr,
invert parameters for clkdiv2_allocate event.
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Tuan Phan <tuanphan@os.amperecomputing.com>
Link: https://lore.kernel.org/r/1604518246-6198-1-git-send-email-tuanphan@os.amperecomputing.com
Signed-off-by: Will Deacon <will@kernel.org>
