Re: [PATCH 4.19 075/276] scsi: qla2xxx: Fix a qla24xx_enable_msix() error path

[Bart Van Assche]

On 5/31/19 1:59 AM, Pavel Machek wrote:
> 
> On Wed 2019-05-29 20:03:53, Greg Kroah-Hartman wrote:
>> [ Upstream commit 24afabdbd0b3553963a2bbf465895492b14d1107 ]
>>
>> Make sure that the allocated interrupts are freed if allocating memory for
>> the msix_entries array fails.
>>
>> Cc: Himanshu Madhani <hmadhani@xxxxxxxxxxx>
>> Cc: Giridhar Malavali <gmalavali@xxxxxxxxxxx>
>> Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
>> Acked-by: Himanshu Madhani <hmadhani@xxxxxxxxxxx>
>> Signed-off-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx>
>> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> 
>> --- a/drivers/scsi/qla2xxx/qla_isr.c
>> +++ b/drivers/scsi/qla2xxx/qla_isr.c
>> @@ -3449,7 +3449,7 @@ qla24xx_enable_msix(struct qla_hw_data *ha, struct rsp_que *rsp)
>>  		ql_log(ql_log_fatal, vha, 0x00c8,
>>  		    "Failed to allocate memory for ha->msix_entries.\n");
>>  		ret = -ENOMEM;
>> -		goto msix_out;
>> +		goto free_irqs;
> 
> Could we just do > +     pci_free_irq_vectors(ha->pdev); return
> -ENOMEM; here? Going through two gotos just does not feel right.
> 
> And yes, I'd replace msix_out with direct returns, too. gotos have
> value when there's cleanup to be done, but we are not doing any here.

That would have been an appropriate comment for the original patch. This
e-mail thread is about backporting an upstream commit to the stable
tree. If you want to change this code you need to send a patch to the
linux-scsi mailing list.

Bart.