RE: [PATCH 3/7] megaraid_sas : Do not process IOCTLs and SCSI commands during driver removal

[Kashyap Desai <kashyap.desai@avagotech.com>]

> -----Original Message-----
> From: Kashyap Desai [mailto:kashyap.desai@avagotech.com]
> Sent: Wednesday, November 12, 2014 4:39 PM
> To: 'Tomas Henzl'; Sumit Saxena; 'linux-scsi@vger.kernel.org'
> Cc: 'martin.petersen@oracle.com'; 'hch@infradead.org';
> 'jbottomley@parallels.com'; 'aradford@gmail.com'
> Subject: RE: [PATCH 3/7] megaraid_sas : Do not process IOCTLs and SCSI
> commands during driver removal
>
>
>
> > -----Original Message-----
> > From: Tomas Henzl [mailto:thenzl@redhat.com]
> > Sent: Tuesday, November 11, 2014 7:13 PM
> > To: Sumit.Saxena@avagotech.com; linux-scsi@vger.kernel.org
> > Cc: martin.petersen@oracle.com; hch@infradead.org;
> > jbottomley@parallels.com; kashyap.desai@avagotech.com;
> > aradford@gmail.com
> > Subject: Re: [PATCH 3/7] megaraid_sas : Do not process IOCTLs and SCSI
> > commands during driver removal
> >
> > On 11/10/2014 01:21 PM, Sumit.Saxena@avagotech.com wrote:
> > > Do not process any SCSI and IOCTL command further(return them with
> > > appropriate return values to callers), while driver removal is in
> > > progress/PCI
> > shutdown is invoked.
> > >
> > > Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
> > > Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
> > > ---
> > >  drivers/scsi/megaraid/megaraid_sas_base.c |   20 ++++++++++++++++--
> --
> > >  1 files changed, 16 insertions(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c
> > > b/drivers/scsi/megaraid/megaraid_sas_base.c
> > > index f6a69a3..7754eeb 100644
> > > --- a/drivers/scsi/megaraid/megaraid_sas_base.c
> > > +++ b/drivers/scsi/megaraid/megaraid_sas_base.c
> > > @@ -1572,6 +1572,12 @@ megasas_queue_command(struct Scsi_Host
> > *shost, struct scsi_cmnd *scmd)
> > >  	instance = (struct megasas_instance *)
> > >  	    scmd->device->host->hostdata;
> > >
> > > +	if (instance->unload == 1) {
> > > +		scmd->result = DID_NO_CONNECT << 16;
> > > +		scmd->scsi_done(scmd);
> > > +		return 0;
> > > +	}
> > > +
> > >  	if (instance->issuepend_done == 0)
> > >  		return SCSI_MLQUEUE_HOST_BUSY;
> > >
> > > @@ -4957,10 +4963,6 @@ static int megasas_io_attach(struct
> > megasas_instance *instance)
> > >  		return -ENODEV;
> > >  	}
> > >
> > > -	/*
> > > -	 * Trigger SCSI to scan our drives
> > > -	 */
> > > -	scsi_scan_host(host);
> > >  	return 0;
> > >  }
> > >
> > > @@ -5288,6 +5290,10 @@ retry_irq_register:
> > >  		goto fail_io_attach;
> > >
> > >  	instance->unload = 0;
> > > +	/*
> > > +	 * Trigger SCSI to scan our drives
> > > +	 */
> > > +	scsi_scan_host(host);
> > >
> > >  	/*
> > >  	 * Initiate AEN (Asynchronous Event Notification) @@ -6051,6
> > > +6057,12 @@ megasas_mgmt_fw_ioctl(struct megasas_instance
> > *instance,
> > >  	megasas_issue_blocked_cmd(instance, cmd, 0);
> > >  	cmd->sync_cmd = 0;
> >
> > I've expected that you'll not send the command to the card, you first
> > issue blocked command and after that test unload state.
> > Shouldn't it be reversed?
> Actually, there was a two part of this Patch for IOCTL path. One will
not allow
> to send IOCTL while driver removal.
> Another will not use already return IOCTL from FW to avoid illegal
memory
> access on MFI frame.
> Driver wakeup all blocked command in unload path and MFI/MPT frames are
> freed as well, so this check is good to have.
>
> As you suggested, we missed to have another check in IOCTL path, before
> driver takes semaphore. It was planned, but somehow we missed while
> submitting this patch set.
> We will send incremental patch for that part.

Correction - First of all sorry for confusion.
megasas_mgmt_ioctl_fw and megasas_mgmt_ioctl_aen already has a check for
instance->unload. We don't need any changes for this patch.

Tomas -
We are planning to provide resend patch series which will have changes for
below two patch.
[PATCH 1/7] megaraid_sas : Driver version upgrade and
[PATCH 4/7] megaraid_sas : Online Firmware upgrade support for Extended VD
feature


` Kashyap

>
> >
> > >
> > > +	if (instance->unload == 1) {
> > > +		dev_info(&instance->pdev->dev, "we are doing unload so no
> > need"
> > > +			"to submit data to application. This may be force
exit"
> > > +			"from driver\n");
> >
> > I'm not a native speaker, so I'll not comment on the wording, but you
> > should add a space at the line breaks.
>
> Agree. will address this part.
>
> > +		dev_info(&instance->pdev->dev, "we are doing unload so no
> > need "
> > +			"to submit data to application. This may be force
exit
> > "
> > +			"from driver\n");
> >
> > > +		goto out;
> > > +	}
> > >  	/*
> > >  	 * copy out the kernel buffers to user buffers
> > >  	 */
> >
> > Aren't there other ioctl paths too - for example
> > megasas_mgmt_ioctl_aen - isn't a block needed there too?
> This path will be cleanup from megasas_shutdown_controller() call.   AEN
> and MAP SYNC command will be aborted explicitly.
>
> >