Re: [PATCH v2] firmware: stratix10-svc: add PSCI secondary CPU offline on warm reboot for agilex and stratix10

["Ng, Adrian Ho Yin" <adrian.ho.yin.ng@altera.com>]

On 9/3/2026 7:40 PM, Dinh Nguyen wrote:
> 
> 
> On 7/14/26 17:16, Adrian Ng Ho Yin wrote:
>> On Agilex7 and Stratix10 SoCs, secondary CPUs must be offlined before
>> a warm reboot to ensure a clean PSCI state. Register a reboot notifier
>> that calls remove_cpu() on each secondary CPU when SYS_RESTART is
>> triggered in REBOOT_WARM mode.
>>
>> A snapshot of cpu_online_mask is taken before the loop since remove_cpu()
>> modifies the mask as CPUs go down. CPU 0 is excluded as it is typically
>> not hotpluggable. Failures are logged and the loop continues so the
>> reboot is not blocked by a single CPU that refuses to offline.
>> The of_device_id table carries a per-compatible pdata flag to enable the
>> notifier only on platforms that need it; Agilex5 leaves the flag unset.
>>
>> The notifier is unregistered at the start of remove(), before async
>> channel teardown, to prevent a reboot notification from racing with
>> driver shutdown.
>>
>> Signed-off-by: Adrian Ng Ho Yin <adrian.ho.yin.ng@altera.com>
>> ---
>> changelog:
>> v1 -> v2:
>> - Drop workqueue fan-out: remove_cpu() serialises on device_hotplug_lock
>>    internally so concurrent workers provided no parallelism. Replace with
>>    a single sequential loop over a cpumask snapshot.
>> - Fix CPU identity bug: get_cpu()/put_cpu() did not prevent task 
>> migration
>>    and the notifier can run on any CPU before migrate_to_reboot_cpu().
>>    Exclude CPU 0 (typically not hotpluggable) and offline everything 
>> else.
>> - Fix pre-count race: the two-loop atomic pre-count had a window where a
>>    concurrent hotplug event could cause dec_and_test() to fire 
>> prematurely.
>>    Eliminated entirely along with the completion and timeout.
>> - Replace of_device_is_compatible() checks with a stratix10_svc_pdata
>>    struct carried in of_device_id.data, retrieved via 
>> of_device_get_match_data().
>> - Move psci_cpu_off_teardown() to the top of stratix10_svc_drv_remove(),
>>    before async channel teardown, to close the race where a warm-reboot
>>    notification could fire while the driver is being dismantled.
>> ---
>>   drivers/firmware/stratix10-svc.c | 88 ++++++++++++++++++++++++++++++--
>>   1 file changed, 85 insertions(+), 3 deletions(-)
> 
> I had a patch conflict when I applied it on top of v7.3-rc1. I fixed it 
> up, please take a look at socfpga_firmware_for_v7.4 to verify the fix.
> 
> A few comments:
> 
> 
>>
>> diff --git a/drivers/firmware/stratix10-svc.c b/drivers/firmware/ 
>> stratix10-svc.c
>> index de938ab2db0b..e10550237ea1 100644
>> --- a/drivers/firmware/stratix10-svc.c
>> +++ b/drivers/firmware/stratix10-svc.c
>> @@ -24,6 +24,10 @@
>>   #include <linux/firmware/intel/stratix10-smc.h>
>>   #include <linux/firmware/intel/stratix10-svc-client.h>
>>   #include <linux/types.h>
>> +#include <linux/cpu.h>
>> +#include <linux/cpumask.h>
>> +#include <linux/of_device.h>
>> +#include <linux/reboot.h>
>>   /**
>>    * SVC_NUM_DATA_IN_FIFO - number of struct stratix10_svc_data in the 
>> FIFO
>> @@ -96,6 +100,14 @@
>>   #define STRATIX10_GET_SDM_STATUS_CODE(status) \
>>       (FIELD_GET(STRATIX10_SDM_STATUS_MASK, status))
>> +struct stratix10_svc_pdata {
>> +    bool needs_psci_cpu_off;
>> +};
>> +
>> +static const struct stratix10_svc_pdata psci_cpu_off_pdata = {
>> +    .needs_psci_cpu_off = true,
>> +};
>> +
>>   typedef void (svc_invoke_fn)(unsigned long, unsigned long, unsigned 
>> long,
>>                    unsigned long, unsigned long, unsigned long,
>>                    unsigned long, unsigned long,
>> @@ -280,6 +292,7 @@ struct stratix10_svc_chan {
>>    * @svc: manages the list of client svc drivers
>>    * @sdm_lock: only allows a single command single response to SDM
>>    * @actrl: async control structure
>> + * @psci_reboot_nb: reboot notifier for PSCI secondary CPU offlining
>>    * @chans: array of service channels
>>    *
>>    * This struct is used to create communication channels for service 
>> clients, to
>> @@ -296,6 +309,7 @@ struct stratix10_svc_controller {
>>       struct stratix10_svc *svc;
>>       struct mutex sdm_lock;
>>       struct stratix10_async_ctrl actrl;
>> +    struct notifier_block psci_reboot_nb;
>>       struct stratix10_svc_chan chans[] __counted_by(num_chans);
>>   };
>> @@ -1981,9 +1995,60 @@ void stratix10_svc_free_memory(struct 
>> stratix10_svc_chan *chan, void *kaddr)
>>   }
>>   EXPORT_SYMBOL_GPL(stratix10_svc_free_memory);
>> +static void psci_offline_secondary_cpus(struct 
>> stratix10_svc_controller *ctrl)
> 
> ctrl is never used in this function.> +{
>> +    cpumask_var_t mask;
>> +    int cpu, ret;
>> +
>> +    if (!alloc_cpumask_var(&mask, GFP_KERNEL))
>> +        return;
>> +
>> +    /*
>> +     * Snapshot cpu_online_mask before the loop; remove_cpu() 
>> modifies it
>> +     * as each CPU is brought down. Always preserve CPU 0 (boot CPU) to
>> +     * run the reboot.
>> +     */
>> +    cpumask_copy(mask, cpu_online_mask);
>> +    cpumask_clear_cpu(0, mask);
>> +
>> +    /*
>> +     * Offlining is best-effort: if a CPU refuses to go down we log the
>> +     * error and continue so the remaining secondaries are still 
>> attempted
>> +     * and the warm reboot can proceed.
>> +     */
>> +    for_each_cpu(cpu, mask) {
>> +        ret = remove_cpu(cpu);
>> +        if (ret)
>> +            pr_err("psci_cpu_off: failed to offline CPU%d: %d\n",
>> +                   cpu, ret);
> 
> Perhaps use dev_err()e so that you can use the ctrl->dev?
> 
Will update to use dev_err instead. >> +    }
>> +
>> +    free_cpumask_var(mask);
>> +}
>> +
>> +static int psci_cpu_off_reboot_notifier(struct notifier_block *nb,
>> +                    unsigned long action, void *data)
>> +{
>> +    struct stratix10_svc_controller *ctrl =
>> +        container_of(nb, struct stratix10_svc_controller, 
>> psci_reboot_nb);
>> +
>> +    if (reboot_mode != REBOOT_WARM)
>> +        return NOTIFY_DONE;
>> +
>> +    if (action == SYS_RESTART)
>> +        psci_offline_secondary_cpus(ctrl);
>> +
>> +    return NOTIFY_OK;
>> +}
>> +
>> +static void psci_cpu_off_teardown(struct stratix10_svc_controller *ctrl)
>> +{
>> +    unregister_reboot_notifier(&ctrl->psci_reboot_nb);
>> +}
>> +
>>   static const struct of_device_id stratix10_svc_drv_match[] = {
>> -    {.compatible = "intel,stratix10-svc"},
>> -    {.compatible = "intel,agilex-svc"},
>> +    { .compatible = "intel,stratix10-svc", .data = 
>> &psci_cpu_off_pdata },
>> +    { .compatible = "intel,agilex-svc",    .data = 
>> &psci_cpu_off_pdata },
> 
> I'm not sure why you need this of_match_data to make the distinction 
> between the 2 platforms when both platforms are performing the same 
> function.
> 
We need to keep of_device_id.data rather than dropping it because both 
intel,stratix10-svc and intel,agilex-svc happen to share the same 
notifier path. A separate patch adds intel,agilex5-svc without .data
so Agilex5 does not run this sequence. Agilex5 warm reset is an 
SDM-owned HPS reset (REBOOT_HPS) that holds secondaries in the 
reset-release / P-channel block until ATF CPURSTRELEASE.