On Thu, 2018-11-29 at 02:01 +-0100, David Disseldorp wrote: +AD4 - strncpy(+ACY-dev-+AD4-t10+AF8-wwn.vendor+AFs-0+AF0, +ACI-LIO-ORG+ACI, 8)+ADs +AD4 +- strncpy(+ACY-dev-+AD4-t10+AF8-wwn.vendor+AFs-0+AF0, +ACI-LIO-ORG+ACI, INQUIRY+AF8-VENDOR+AF8-LEN)+ADs +AD4 +- dev-+AD4-t10+AF8-wwn.vendor+AFs-INQUIRY+AF8-VENDOR+AF8-LEN+AF0 +AD0 '+AFw-0'+ADs This looks weird to me. Have you considered to use strlcpy() instead of strncpy() and explicit '+AFw-0'-termination? +AD4 strncpy(+ACY-dev-+AD4-t10+AF8-wwn.model+AFs-0+AF0, +AD4 dev-+AD4-transport-+AD4-inquiry+AF8-prod, 16)+ADs +AD4 strncpy(+ACY-dev-+AD4-t10+AF8-wwn.revision+AFs-0+AF0, +AD4 diff --git a/drivers/target/target+AF8-core+AF8-pscsi.c b/drivers/target/target+AF8-core+AF8-pscsi.c +AD4 index 47d76c862014..ee65b5bb674c 100644 +AD4 --- a/drivers/target/target+AF8-core+AF8-pscsi.c +AD4 +-+-+- b/drivers/target/target+AF8-core+AF8-pscsi.c +AD4 +AEAAQA -190,7 +-190,9 +AEAAQA pscsi+AF8-set+AF8-inquiry+AF8-info(struct scsi+AF8-device +ACo-sdev, struct t10+AF8-wwn +ACo-wwn) +AD4 /+ACo +AD4 +ACo Use sdev-+AD4-inquiry from drivers/scsi/scsi+AF8-scan.c:scsi+AF8-alloc+AF8-sdev() +AD4 +ACo-/ +AD4 - memcpy(+ACY-wwn-+AD4-vendor+AFs-0+AF0, +ACY-buf+AFs-8+AF0, sizeof(wwn-+AD4-vendor))+ADs +AD4 +- BUILD+AF8-BUG+AF8-ON(sizeof(wwn-+AD4-vendor) +ACEAPQ INQUIRY+AF8-VENDOR+AF8-LEN +- 1)+ADs +AD4 +- memcpy(+ACY-wwn-+AD4-vendor+AFs-0+AF0, +ACY-buf+AFs-8+AF0, INQUIRY+AF8-VENDOR+AF8-LEN)+ADs +AD4 +- wwn-+AD4-vendor+AFs-INQUIRY+AF8-VENDOR+AF8-LEN+AF0 +AD0 '+AFw-0'+ADs Have you considered to use snprintf(..., +ACIAJQ.+ACo-s+ACI, ...) instead of memcpy() followed by explicit '+AFw-0'-termination? I think that would result in code that is easier to read. strlcpy() can't be used here because it is not guaranteed that the input buffer is '+AFw-0'-terminated. Thanks, Bart.
