Re:RE: [PATCH v3] counter: rz-mtu3-cnt: put runtime PM ref if counter init fails

[李佑鸿]

Hi Biju,





At 2026-11-02 15:40:31, "Biju Das" <biju.das.jz@xxxxxxxxxxxxxx> wrote:
>Hi Li Youhong,
>
>> -----Original Message-----
>> From: Li Youhong <dayou5941@xxxxxxx>
>> Sent: 01 November 2026 13:08
>> Subject: [PATCH v3] counter: rz-mtu3-cnt: put runtime PM ref if counter init fails
>> 
>> From: Li Youhong <liyouhong@xxxxxxxxxx>
>> 
>> Counter enable takes a runtime PM reference before initialization but only puts it on disable. Drop the
>> reference if initialization fails.
>> 
>> Fixes: 0be8907359df ("counter: Add Renesas RZ/G2L MTU3a counter driver")
>> Cc: stable@xxxxxxxxxxxxxxx
>> Signed-off-by: Li Youhong <liyouhong@xxxxxxxxxx>
>> ---
>> v3:
>> - Set count_is_enabled from enable after either branch succeeds.
>> - v2: https://lore.kernel.org/all/20261031142813.2522444-1-dayou5941@xxxxxxx/
>> 
>> v2:
>> - Add Cc: stable@xxxxxxxxxxxxxxx.
>> - v1: https://lore.kernel.org/all/20261008020506.211566-1-dayou5941@xxxxxxx/
>> 
>>  drivers/counter/rz-mtu3-cnt.c | 9 ++++++---
>>  1 file changed, 6 insertions(+), 3 deletions(-)
>> 
>> diff --git a/drivers/counter/rz-mtu3-cnt.c b/drivers/counter/rz-mtu3-cnt.c index
>> 7bfb6979193c..e7fe8bf45164 100644
>> --- a/drivers/counter/rz-mtu3-cnt.c
>> +++ b/drivers/counter/rz-mtu3-cnt.c
>> @@ -506,14 +506,17 @@ static int rz_mtu3_count_enable_write(struct counter_device *counter,
>>  	if (enable) {
>>  		pm_runtime_get_sync(counter->parent);
>>  		ret = rz_mtu3_initialize_counter(counter, count->id);
>
>Reviewed-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
>
>Now days bots started complaining if pm_runtime_get_sync() fails
>Then rz_mtu3_initialize_counter() will lead to kernel crash(sysnchronous abort)
>as it is accessing registers without the clk.
>
>Since you have added fixes tag, not sure this to be fixed in this patch??

>

The runtime PM get checks are patch 2 of this series:
https://lore.kernel.org/all/20260903090719.3498955-1-dayou5941@xxxxxxx/

We already discussed that here:
https://lore.kernel.org/all/66575d8d.214d.1a1193fd2ba.Coremail.dayou5941@xxxxxxx/

So I don't think the pm_runtime_get_sync() return value needs to be
handled in this patch. Cosmin's patch series covers that.

Thanks,
Li

>Cheers,
>Biju
>
>> -		if (ret == 0)
>> -			priv->count_is_enabled[count->id] = true;
>> +		if (ret < 0) {
>> +			pm_runtime_put(counter->parent);
>> +			goto exit;
>> +		}
>>  	} else {
>>  		rz_mtu3_terminate_counter(counter, count->id);
>> -		priv->count_is_enabled[count->id] = false;
>>  		pm_runtime_put(counter->parent);
>>  	}
>> 
>> +	priv->count_is_enabled[count->id] = enable;
>> +
>>  exit:
>>  	mutex_unlock(&priv->lock);
>>