RE: [RFC PATCH] scsi: storvsc: Fix validation for unsolicited incoming packets

[Haiyang Zhang]

> -----Original Message-----
> From: Haiyang Zhang
> Sent: Tuesday, September 28, 2021 6:06 PM
> To: Andrea Parri (Microsoft) <parri.andrea@xxxxxxxxx>; linux-
> kernel@xxxxxxxxxxxxxxx; linux-hyperv@xxxxxxxxxxxxxxx; linux-
> scsi@xxxxxxxxxxxxxxx
> Cc: KY Srinivasan <kys@xxxxxxxxxxxxx>; Stephen Hemminger
> <sthemmin@xxxxxxxxxxxxx>; Wei Liu <wei.liu@xxxxxxxxxx>; James E . J .
> Bottomley <jejb@xxxxxxxxxxxxx>; Martin K . Petersen
> <martin.petersen@xxxxxxxxxx>; Michael Kelley <mikelley@xxxxxxxxxxxxx>;
> Dexuan Cui <decui@xxxxxxxxxxxxx>
> Subject: RE: [RFC PATCH] scsi: storvsc: Fix validation for unsolicited
> incoming packets
> 
> 
> 
> > -----Original Message-----
> > From: Andrea Parri (Microsoft) <parri.andrea@xxxxxxxxx>
> > Sent: Tuesday, September 28, 2021 12:38 PM
> > To: linux-kernel@xxxxxxxxxxxxxxx; linux-hyperv@xxxxxxxxxxxxxxx; linux-
> > scsi@xxxxxxxxxxxxxxx
> > Cc: KY Srinivasan <kys@xxxxxxxxxxxxx>; Haiyang Zhang
> > <haiyangz@xxxxxxxxxxxxx>; Stephen Hemminger <sthemmin@xxxxxxxxxxxxx>;
> > Wei Liu <wei.liu@xxxxxxxxxx>; James E . J . Bottomley
> > <jejb@xxxxxxxxxxxxx>; Martin K . Petersen
> > <martin.petersen@xxxxxxxxxx>; Michael Kelley <mikelley@xxxxxxxxxxxxx>;
> > Andrea Parri (Microsoft) <parri.andrea@xxxxxxxxx>; Dexuan Cui
> > <decui@xxxxxxxxxxxxx>
> > Subject: [RFC PATCH] scsi: storvsc: Fix validation for unsolicited
> > incoming packets
> >
> > The validation on the length of incoming packets performed in
> > storvsc_on_channel_callback() does not apply to "unsolicited"
> > packets with ID of 0 sent by Hyper-V.  Adjust the validation by
> > handling such unsolicited packets separately.
> >
> > Fixes: 91b1b640b834b2 ("scsi: storvsc: Validate length of incoming
> > packet in storvsc_on_channel_callback()")
> > Reported-by: Dexuan Cui <decui@xxxxxxxxxxxxx>
> > Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@xxxxxxxxx>
> > ---
> > The (new) bound, VSTOR_MIN_UNSOL_PKT_SIZE, was "empirically derived"
> > based on testing and code auditing.  This explains the RFC tag...
> >
> >  drivers/scsi/storvsc_drv.c | 17 ++++++++++++++---
> >  1 file changed, 14 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c
> > index ebbbc1299c625..a9bbcbbfb54ee 100644
> > --- a/drivers/scsi/storvsc_drv.c
> > +++ b/drivers/scsi/storvsc_drv.c
> > @@ -292,6 +292,9 @@ struct vmstorage_protocol_version {
> >  #define STORAGE_CHANNEL_REMOVABLE_FLAG		0x1
> >  #define STORAGE_CHANNEL_EMULATED_IDE_FLAG	0x2
> >
> > +/* Lower bound on the size of unsolicited packets with ID of 0 */
> > +#define VSTOR_MIN_UNSOL_PKT_SIZE		48
> > +
> >  struct vstor_packet {
> >  	/* Requested operation type */
> >  	enum vstor_packet_operation operation; @@ -1285,11 +1288,13 @@
> > static void storvsc_on_channel_callback(void
> > *context)
> >  	foreach_vmbus_pkt(desc, channel) {
> >  		struct vstor_packet *packet = hv_pkt_data(desc);
> >  		struct storvsc_cmd_request *request = NULL;
> > +		u32 pktlen = hv_pkt_datalen(desc);
> >  		u64 rqst_id = desc->trans_id;
> >
> > -		if (hv_pkt_datalen(desc) < sizeof(struct vstor_packet) -
> > +		/* Unsolicited packets with ID of 0 are validated separately
> > below */
> > +		if (rqst_id != 0 && pktlen < sizeof(struct vstor_packet) -
> >  				stor_device->vmscsi_size_delta) {
> > -			dev_err(&device->device, "Invalid packet len\n");
> > +			dev_err(&device->device, "Invalid packet: length=%u\n",
> > pktlen);
> >  			continue;
> >  		}
> >
> > @@ -1298,8 +1303,14 @@ static void storvsc_on_channel_callback(void
> > *context)
> >  		} else if (rqst_id == VMBUS_RQST_RESET) {
> >  			request = &stor_device->reset_request;
> >  		} else {
> > -			/* Hyper-V can send an unsolicited message with ID of 0
> > */
> >  			if (rqst_id == 0) {
> > +				if (pktlen < VSTOR_MIN_UNSOL_PKT_SIZE) {
> > +					dev_err(&device->device,
> > +						"Invalid packet with ID of 0:
> > length=%u\n",
> > +						pktlen);
> > +					continue;
> > +				}
> > +
> >  				/*
> >  				 * storvsc_on_receive() looks at the vstor_packet
> in the message
> >  				 * from the ring buffer.  If the operation in the
> vstor_packet is
> 
> The patch looks good. But for readability, I'd suggested put the length
> checks together like this:
> 
> 	u32 minlen = rqst_id ? sizeof(struct vstor_packet) -
> 		stor_device->vmscsi_size_delta : VSTOR_MIN_UNSOL_PKT_SIZE;
> 
> 	if (pktlen < minlen) {
> 		dev_err(&device->device,
> 			   "Invalid pkt: id=%llu, len=%u, minlen=%u\n",
> 			   rqst_id, pktlen, minlen);
> 		continue;
> 	}
> 
> Thanks.
> 
> Signed-off-by: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>
The tag was meant to be:
Reviewed-by: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>

Thanks.