[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <6cd6b923-4231-fed2-f691-5ff5d55bc64a@linux.ibm.com>
Date: Fri, 4 Dec 2020 16:20:35 -0800
From: Tyrel Datwyler <tyreld@...ux.ibm.com>
To: Brian King <brking@...ux.vnet.ibm.com>,
james.bottomley@...senpartnership.com
Cc: martin.petersen@...cle.com, linux-scsi@...r.kernel.org,
linuxppc-dev@...ts.ozlabs.org, linux-kernel@...r.kernel.org,
brking@...ux.ibm.com
Subject: Re: [PATCH v3 18/18] ibmvfc: drop host lock when completing commands
in CRQ
On 12/4/20 1:35 PM, Brian King wrote:
> On 12/2/20 8:08 PM, Tyrel Datwyler wrote:
>> The legacy CRQ holds the host lock the even while completing commands.
>> This presents a problem when in legacy single queue mode and
>> nr_hw_queues is greater than one since calling scsi_done() introduces
>> the potential for deadlock.
>>
>> If nr_hw_queues is greater than one drop the hostlock in the legacy CRQ
>> path when completing a command.
>>
>> Signed-off-by: Tyrel Datwyler <tyreld@...ux.ibm.com>
>> ---
>> drivers/scsi/ibmvscsi/ibmvfc.c | 8 +++++++-
>> 1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/scsi/ibmvscsi/ibmvfc.c b/drivers/scsi/ibmvscsi/ibmvfc.c
>> index e499599662ec..e2200bdff2be 100644
>> --- a/drivers/scsi/ibmvscsi/ibmvfc.c
>> +++ b/drivers/scsi/ibmvscsi/ibmvfc.c
>> @@ -2969,6 +2969,7 @@ static void ibmvfc_handle_crq(struct ibmvfc_crq *crq, struct ibmvfc_host *vhost)
>> {
>> long rc;
>> struct ibmvfc_event *evt = (struct ibmvfc_event *)be64_to_cpu(crq->ioba);
>> + unsigned long flags;
>>
>> switch (crq->valid) {
>> case IBMVFC_CRQ_INIT_RSP:
>> @@ -3039,7 +3040,12 @@ static void ibmvfc_handle_crq(struct ibmvfc_crq *crq, struct ibmvfc_host *vhost)
>> del_timer(&evt->timer);
>> list_del(&evt->queue);
>> ibmvfc_trc_end(evt);
>> - evt->done(evt);
>> + if (nr_scsi_hw_queues > 1) {
>> + spin_unlock_irqrestore(vhost->host->host_lock, flags);
>> + evt->done(evt);
>> + spin_lock_irqsave(vhost->host->host_lock, flags);
>> + } else
>> + evt->done(evt);
>
> Similar comment here as previously. The flags parameter is an output for
> spin_lock_irqsave but an input for spin_unlock_irqrestore. You'll need
> to rethink the locking here. You could just do a spin_unlock_irq / spin_lock_irq
> here and that would probably be OK, but probably isn't the best.
>
Yeah, this will also get its own lock and flags saved in the per-queue struct in
the next spin.
-Tyrel
>> }
>>
>> /**
>>
>
>
Powered by blists - more mailing lists