linux-kernel - Re: [PATCH v2] ata: libata-scsi: Make __ata_scsi

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Wed, 5 Jan 2022 09:07:13 +0000
From:   Damien Le Moal <Damien.LeMoal@....com>
To:     "linux-ide@...r.kernel.org" <linux-ide@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "haowenchao@...wei.com" <haowenchao@...wei.com>,
        "damien.lemoal@...nsource.wdc.com" <damien.lemoal@...nsource.wdc.com>
CC:     "liuzhiqiang26@...wei.com" <liuzhiqiang26@...wei.com>
Subject: Re: [PATCH v2] ata: libata-scsi: Make __ata_scsi_queuecmd()
 parameters check more clearly

On Wed, 2022-01-05 at 16:48 +0800, Wenchao Hao wrote:
> On 2022/1/5 15:39, Damien Le Moal wrote:
> > On Wed, 2022-01-05 at 15:27 -0500, Wenchao Hao wrote:
> > > This is just a clean code. Since each branch of "if" state would check
> > > scmd->cmd_len, so move the check of scmd->cmd_len out of "if" state to
> > > simplify input parameters check.
> > > 
> > > And remove redundant init of xlat_func at hand
> > > 
> > > The patch do not change origin function logic.
> > > 
> > > Signed-off-by: Wenchao Hao <haowenchao@...wei.com>
> > > ---
> > >   drivers/ata/libata-scsi.c | 9 ++++-----
> > >   1 file changed, 4 insertions(+), 5 deletions(-)
> > > 
> > > diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
> > > index 313e9475507b..ab8a2833dfec 100644
> > > --- a/drivers/ata/libata-scsi.c
> > > +++ b/drivers/ata/libata-scsi.c
> > > @@ -4023,16 +4023,15 @@ int __ata_scsi_queuecmd(struct scsi_cmnd *scmd, struct ata_device *dev)
> > >   	ata_xlat_func_t xlat_func;
> > >   	int rc = 0;
> > >   
> > > +	if (unlikely(!scmd->cmd_len))
> > > +		goto bad_cdb_len;
> > > +
> > >   	if (dev->class == ATA_DEV_ATA || dev->class == ATA_DEV_ZAC) {
> > > -		if (unlikely(!scmd->cmd_len || scmd->cmd_len > dev->cdb_len))
> > > +		if (unlikely(scmd->cmd_len > dev->cdb_len))
> > >   			goto bad_cdb_len;
> > >   
> > >   		xlat_func = ata_get_xlat_func(dev, scsi_op);
> > >   	} else {
> > > -		if (unlikely(!scmd->cmd_len))
> > > -			goto bad_cdb_len;
> > > -
> > > -		xlat_func = NULL;
> > >   		if (likely((scsi_op != ATA_16) || !atapi_passthru16)) {
> > >   			/* relay SCSI command to ATAPI device */
> > >   			int len = COMMAND_SIZE(scsi_op);
> > 
> > Did you miss my reply ?
> > This change is OK, but while at it, let's cleanup this function further.
> > I suggested something like this, which includes your changes.
> > 
> 
> Maybe I misunderstood your previous reply. I think you ask me to change 
> prefix.

Yes, I asked that too. It is OK now.

> 
> > diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
> > index a16ef0030667..ed8be585a98f 100644
> > --- a/drivers/ata/libata-scsi.c
> > +++ b/drivers/ata/libata-scsi.c
> > @@ -3958,42 +3958,39 @@ int __ata_scsi_queuecmd(struct scsi_cmnd *scmd,
> > struct ata_device *dev)
> >   {
> >   	u8 scsi_op = scmd->cmnd[0];
> >   	ata_xlat_func_t xlat_func;
> > -	int rc = 0;
> > +
> > +	if (unlikely(!scmd->cmd_len))
> > +		goto bad_cdb_len;
> > 
> >   	if (dev->class == ATA_DEV_ATA || dev->class == ATA_DEV_ZAC) {
> > -		if (unlikely(!scmd->cmd_len || scmd->cmd_len > dev->cdb_len))
> > +		if (unlikely(scmd->cmd_len > dev->cdb_len))
> >   			goto bad_cdb_len;
> > 
> >   		xlat_func = ata_get_xlat_func(dev, scsi_op);
> > -	} else {
> > -		if (unlikely(!scmd->cmd_len))
> > -			goto bad_cdb_len;
> > +	} else if (scsi_op != ATA_16 || !atapi_passthru16) {
> > +		/* relay SCSI command to ATAPI device */
> > +		int len = COMMAND_SIZE(scsi_op);
> > 
> > -		xlat_func = NULL;
> > -		if (likely((scsi_op != ATA_16) || !atapi_passthru16)) {
> > -			/* relay SCSI command to ATAPI device */
> > -			int len = COMMAND_SIZE(scsi_op);
> > -			if (unlikely(len > scmd->cmd_len ||
> > -				     len > dev->cdb_len ||
> > -				     scmd->cmd_len > ATAPI_CDB_LEN))
> > -				goto bad_cdb_len;
> > +		if (unlikely(len > scmd->cmd_len ||
> > +			     len > dev->cdb_len ||
> > +			     scmd->cmd_len > ATAPI_CDB_LEN))
> > +			goto bad_cdb_len;
> > 
> > -			xlat_func = atapi_xlat;
> > -		} else {
> > -			/* ATA_16 passthru, treat as an ATA command */
> > -			if (unlikely(scmd->cmd_len > 16))
> > -				goto bad_cdb_len;
> > +		xlat_func = atapi_xlat;
> > +	} else {
> > +		/* ATA_16 passthru, treat as an ATA command */
> > +		if (unlikely(scmd->cmd_len > 16))
> > +			goto bad_cdb_len;
> > 
> > -			xlat_func = ata_get_xlat_func(dev, scsi_op);
> > -		}
> > +		xlat_func = ata_get_xlat_func(dev, scsi_op);
> >   	}
> > 
> >   	if (xlat_func)
> > -		rc = ata_scsi_translate(dev, scmd, xlat_func);
> > -	else
> > -		ata_scsi_simulate(dev, scmd);
> > +		return ata_scsi_translate(dev, scmd, xlat_func);
> > 
> > -	return rc;
> > +	ata_scsi_simulate(dev, scmd);
> > +
> > +	return 0;
> > 
> >    bad_cdb_len:
> >   	scmd->result = DID_ERROR << 16;
> > 
> > Do you see any problem with this change ?
> > 
> 
> This change looks good to me. Should I include this change in next 
> patch? Or you would do this by youself?

Just include the change in your v3, that's fine. But test please. I did test
what I proposed and it worked fine but I prefer more tests.

And given that the patch does more that just changing the command checks, maybe
change the patch title to something like:

ata: libata-scsi: simplify __ata_scsi_queuecmd()

-- 
Damien Le Moal
Western Digital Research