On 14.09.2023 17:29, Jan Beulich wrote:
> On 14.09.2023 17:19, Jarkko Sakkinen wrote:
>> On Thu Sep 14, 2023 at 5:28 PM EEST, Jan Beulich wrote:
>>> tpm_tis_core_init() may fail before tpm_tis_probe_irq_single() is
>>> called, in which case tpm_tis_remove() unconditionally calling
>>> flush_work() is triggering a warning for .func still being NULL.
>>>
>>> Fixes: 481c2d14627d ("tpm,tpm_tis: Disable interrupts after 1000 unhandled
>>> IRQs")
>>> Signed-off-by: Jan Beulich <[email protected]>
>>> ---
>>> An alternative would be to move INIT_WORK(), but where to put it is far
>>> more difficult to tell for an outsider than simply making the flush call
>>> conditional.
>>>
>>> --- a/drivers/char/tpm/tpm_tis_core.c
>>> +++ b/drivers/char/tpm/tpm_tis_core.c
>>> @@ -1022,7 +1022,8 @@ void tpm_tis_remove(struct tpm_chip *chi
>>> interrupt = 0;
>>>
>>> tpm_tis_write32(priv, reg, ~TPM_GLOBAL_INT_ENABLE & interrupt);
>>> - flush_work(&priv->free_irq_work);
>>> + if (priv->free_irq_work.func)
>>> + flush_work(&priv->free_irq_work);
>>>
>>> tpm_tis_clkrun_enable(chip, false);
>>>
>>
>> Reviewed-by: Jarkko Sakkinen <[email protected]>
>>
>> Jan, I'm about to leave to vacation but will be back after next week.
>>
>> Do you think that the fix can hold up unti that?
>
> There's no rush from my pov, as I have helped myself.
Has this possibly fallen through the cracks?
Jan
