sd: Ensure we correctly disable devices with unknown protection type
Hi Martin,
The patch fe542396da73: "[SCSI] sd: Ensure we correctly disable devices with unknown protection type" from Sep 21, 2012, leads to the following warning:include/scsi/scsi_host.h:879 scsi_host_dif_capable()
warn: buffer overflow 'cap' 4 <= 4
838 enum scsi_host_prot_capabilities {
839 SHOST_DIF_TYPE1_PROTECTION = 1 << 0, /* T10 DIF Type 1 */
840 SHOST_DIF_TYPE2_PROTECTION = 1 << 1, /* T10 DIF Type 2 */
841 SHOST_DIF_TYPE3_PROTECTION = 1 << 2, /* T10 DIF Type 3 */
842
843 SHOST_DIX_TYPE0_PROTECTION = 1 << 3, /* DIX between OS and HBA only */
844 SHOST_DIX_TYPE1_PROTECTION = 1 << 4, /* DIX with DIF Type 1 */
845 SHOST_DIX_TYPE2_PROTECTION = 1 << 5, /* DIX with DIF Type 2 */
846 SHOST_DIX_TYPE3_PROTECTION = 1 << 6, /* DIX with DIF Type 3 */
847 }; 869 static inline unsigned int scsi_host_dif_capable(struct Scsi_Host *shost, unsigned int target_type) 870 {
871 static unsigned char cap[] = { 0,
872 SHOST_DIF_TYPE1_PROTECTION,
873 SHOST_DIF_TYPE2_PROTECTION,
874 SHOST_DIF_TYPE3_PROTECTION };
875
876 if (target_type > SHOST_DIF_TYPE3_PROTECTION)
877 return 0;
878
879 return shost->prot_capabilities & cap[target_type] ? target_type : 0;
^^^^^^^^^^^^^^^^It looks like we're doing cap[1 << 2] here not cap[2].
880 }
882 static inline unsigned int scsi_host_dix_capable(struct Scsi_Host *shost, unsigned int target_type)
883 {
884 #if defined(CONFIG_BLK_DEV_INTEGRITY)
885 static unsigned char cap[] = { SHOST_DIX_TYPE0_PROTECTION,
886 SHOST_DIX_TYPE1_PROTECTION,
887 SHOST_DIX_TYPE2_PROTECTION,
888 SHOST_DIX_TYPE3_PROTECTION };
889
890 if (target_type > SHOST_DIX_TYPE3_PROTECTION)
891 return 0;
892
893 return shost->prot_capabilities & cap[target_type];
^^^^^^^^^^^^^^^^^We're doing cap[1 << 6] here, not cap[6];
894 #endif
895 return 0;
896 }
regards,
dan carpenter
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Dan" == Dan Carpenter <[email protected]> writes:
Dan,
Dan> warn: buffer overflow 'cap' 4 <= 4
Argh, yes. Type 3 is 4 because it's a bitmask.
Signature · 11 lines
Martin K. Petersen Oracle Linux Engineering SCSI: Fix range check in scsi_host.h The range checking from fe542396 was bad. We would still end up walking beyond the array as Type 3 is defined to be 4 in the protection bitmask. Instead use ARRAY_SIZE() for the range check. Reported-by: Dan Carpenter <[email protected]> Signed-off-by: Martin K. Petersen <[email protected]>
include/scsi/scsi_host.h +2 -2
@@ -873,7 +873,7 @@ static inline unsigned int scsi_host_dif_capable(struct Scsi_Host *shost, unsignSHOST_DIF_TYPE2_PROTECTION , SHOST_DIF_TYPE3_PROTECTION }; -if (target_type >SHOST_DIF_TYPE3_PROTECTION ) +if (target_type >=ARRAY_SIZE (cap))return 0 ;return shost->prot_capabilities & cap[target_type] ? target_type :0 ;
@@ -887,7 +887,7 @@ static inline unsigned int scsi_host_dix_capable(struct Scsi_Host *shost, unsignSHOST_DIX_TYPE2_PROTECTION , SHOST_DIX_TYPE3_PROTECTION }; -if (target_type >SHOST_DIX_TYPE3_PROTECTION ) +if (target_type >=ARRAY_SIZE (cap))return 0 ;return shost->prot_capabilities & cap[target_type];
This patch wasn't applied. Probably because it didn't have [PATCH] in the email subject.
regards,
dan carpenter
Quoted 43 lines · Martin K. Petersen
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Dan" == Dan Carpenter <[email protected]> writes:
Dan,
Dan> warn: buffer overflow 'cap' 4 <= 4
Argh, yes. Type 3 is 4 because it's a bitmask.
Signature · 11 lines
Martin K. Petersen Oracle Linux Engineering SCSI: Fix range check in scsi_host.h The range checking from fe542396 was bad. We would still end up walking beyond the array as Type 3 is defined to be 4 in the protection bitmask. Instead use ARRAY_SIZE() for the range check. Reported-by: Dan Carpenter <[email protected]> Signed-off-by: Martin K. Petersen <[email protected]>include/scsi/scsi_host.h +2
-2@@ -873,7 +873,7 @@ static inline unsigned int scsi_host_dif_capable(struct Scsi_Host *shost, unsignSHOST_DIF_TYPE2_PROTECTION , SHOST_DIF_TYPE3_PROTECTION }; -if (target_type >SHOST_DIF_TYPE3_PROTECTION ) +if (target_type >=ARRAY_SIZE (cap))return 0 ;return shost->prot_capabilities & cap[target_type] ? target_type :0 ;@@ -887,7 +887,7 @@ static inline unsigned int scsi_host_dix_capable(struct Scsi_Host *shost, unsignSHOST_DIX_TYPE2_PROTECTION , SHOST_DIX_TYPE3_PROTECTION }; -if (target_type >SHOST_DIX_TYPE3_PROTECTION ) +if (target_type >=ARRAY_SIZE (cap))return 0 ;return shost->prot_capabilities & cap[target_type];
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Dan" == Dan Carpenter <[email protected]> writes:
Dan> This patch wasn't applied.
Nope, but it hasn't been forgotten. I have it in my resubmit queue with a couple of other stable fixes.
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering