RE: [PATCH 4/5] megaraid_sas: TM command refire leads to controller firmware crash

From: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
To: Hannes Reinecke <hare@suse.de>, linux-scsi@vger.kernel.org
Cc: Kashyap Desai <kashyap.desai@broadcom.com>,
	Sumit Saxena <sumit.saxena@broadcom.com>,
	Kiran Kumar Kasturi <kiran-kumar.kasturi@broadcom.com>,
	Sankar Patra <sankar.patra@broadcom.com>,
	Sasikumar PC <sasikumar.pc@broadcom.com>,
	Shivasharan Srikanteshwara 
	<shivasharan.srikanteshwara@broadcom.com>,
	Anand Lodnoor <anand.lodnoor@broadcom.com>,
	stable@vger.kernel.org
Subject: RE: [PATCH 4/5] megaraid_sas: TM command refire leads to controller firmware crash
Date: Sat, 16 May 2020 12:26:51 +0530	[thread overview]
Message-ID: <8e37b956275b2837da4ba71773033c7d@mail.gmail.com> (raw)
In-Reply-To: <b007f7ae-0a9d-2bfd-6da8-e72576c40353@suse.de>

> > Issue: When TM command times-out driver invokes the controller reset.
> > Post reset, driver re-fires pended TM commands which leads to firmware
> > crash.
> >
> > Fix: Post controller reset, return pended TM commands back to OS.
> >
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
> > Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
> > ---
> >   drivers/scsi/megaraid/megaraid_sas_fusion.c | 7 ++++++-
> >   1 file changed, 6 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/scsi/megaraid/megaraid_sas_fusion.c
> > b/drivers/scsi/megaraid/megaraid_sas_fusion.c
> > index 87f91a38..319f241 100644
> > --- a/drivers/scsi/megaraid/megaraid_sas_fusion.c
> > +++ b/drivers/scsi/megaraid/megaraid_sas_fusion.c
> > @@ -4180,6 +4180,7 @@ static void megasas_refire_mgmt_cmd(struct
> megasas_instance *instance,
> >   	struct fusion_context *fusion;
> >   	struct megasas_cmd *cmd_mfi;
> >   	union MEGASAS_REQUEST_DESCRIPTOR_UNION *req_desc;
> > +	struct MPI2_RAID_SCSI_IO_REQUEST *scsi_io_req;
> >   	u16 smid;
> >   	bool refire_cmd = false;
> >   	u8 result;
> > @@ -4247,6 +4248,11 @@ static void megasas_refire_mgmt_cmd(struct
> megasas_instance *instance,
> >   			result = COMPLETE_CMD;
> >   		}
> >
> > +		scsi_io_req = (struct MPI2_RAID_SCSI_IO_REQUEST *)
> > +				cmd_fusion->io_request;
> > +		if (scsi_io_req->Function == MPI2_FUNCTION_SCSI_TASK_MGMT)
> > +			result = RETURN_CMD;
> > +
> >   		switch (result) {
> >   		case REFIRE_CMD:
> >   			megasas_fire_cmd_fusion(instance, req_desc); @@ -4475,7
> +4481,6
> > @@ megasas_issue_tm(struct megasas_instance *instance, u16
> > device_handle,
> >   	if (!timeleft) {
> >   		dev_err(&instance->pdev->dev,
> >   			"task mgmt type 0x%x timed out\n", type);
> > -		cmd_mfi->flags |= DRV_DCMD_SKIP_REFIRE;
> >   		mutex_unlock(&instance->reset_mutex);
> >   		rc = megasas_reset_fusion(instance->host, MFI_IO_TIMEOUT_OCR);
> >   		mutex_lock(&instance->reset_mutex);
> >
> Why didn't the 'DRV_DCMD_SKIP_REFIRE' work?
> And if it doesn't work, can't it be removed completely?

Re-fire logic doesn't check  'DRV_DCMD_SKIP_REFIRE'  flag for TM commands
but it will check the flag for DCMDs
Hence, 'DRV_DCMD_SKIP_REFIRE' flag is only removed for TM commands.

-Chandrakanth Patil