cpufreq: qcom-hw: Fix probable nested interrupt handling
authorVladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Thu, 11 Nov 2021 15:48:07 +0000 (17:48 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 27 Jan 2022 10:03:15 +0000 (11:03 +0100)
[ Upstream commit e0e27c3d4e20dab861566f1c348ae44e4b498630 ]

Re-enabling an interrupt from its own interrupt handler may cause
an interrupt storm, if there is a pending interrupt and because its
handling is disabled due to already done entrance into the handler
above in the stack.

Also, apparently it is improper to lock a mutex in an interrupt contex.

Fixes: 275157b367f4 ("cpufreq: qcom-cpufreq-hw: Add dcvs interrupt support")
Signed-off-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Reviewed-by: Matthias Kaehlcke <mka@chromium.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/cpufreq/qcom-cpufreq-hw.c

index 0138b2e..35d9336 100644 (file)
@@ -343,9 +343,9 @@ static irqreturn_t qcom_lmh_dcvs_handle_irq(int irq, void *data)
 
        /* Disable interrupt and enable polling */
        disable_irq_nosync(c_data->throttle_irq);
-       qcom_lmh_dcvs_notify(c_data);
+       schedule_delayed_work(&c_data->throttle_work, 0);
 
-       return 0;
+       return IRQ_HANDLED;
 }
 
 static const struct qcom_cpufreq_soc_data qcom_soc_data = {