* [patch] [SCSI] nsp32: fix an off by one
@ 2014-04-03 7:28 Dan Carpenter
0 siblings, 0 replies; only message in thread
From: Dan Carpenter @ 2014-04-03 7:28 UTC (permalink / raw)
To: GOTO Masanori
Cc: YOKOTA Hiroshi, James E.J. Bottomley, linux-scsi, kernel-janitors
This off by one bug will cause memory corruption when we use every
sglun.
The problem is that NSP32_SG_TABLE_SIZE is too small. Here is how
it's caculated in the original code:
sizeof(nsp32_sgtable) * NSP32_SG_SIZE * MAX_TARGET * MAX_LUN
But the problem is that the NSP32_SG_SIZE should be (NSP32_SG_SIZE + 1)
like this:
sizeof(nsp32_sgtable) * (NSP32_SG_SIZE + 1) * MAX_TARGET * MAX_LUN
Or another more intuitive way to write it would be:
sizeof(nsp32_sglun) * MAX_TARGET * MAX_LUN
We use NSP32_SG_TABLE_SIZE to allocate data->sg_list which is an array
of nsp32_sglun structs. The array has MAX_TARGET * MAX_LUN elements.
MAX_TARGET (8) * MAX_LUN (8) = 64 elements. The missing "+ 1" means
that we only allocate enough space for 63 elements. In terms of bytes,
the old code allocated 65536 bytes and now it allocates 66048 bytes.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
Static checker stuff.
diff --git a/drivers/scsi/nsp32.h b/drivers/scsi/nsp32.h
index c022182..d94f4cd 100644
--- a/drivers/scsi/nsp32.h
+++ b/drivers/scsi/nsp32.h
@@ -454,7 +454,7 @@ typedef struct _nsp32_sgtable {
typedef struct _nsp32_sglun {
nsp32_sgtable sgt[NSP32_SG_SIZE+1]; /* SG table */
} __attribute__ ((packed)) nsp32_sglun;
-#define NSP32_SG_TABLE_SIZE (sizeof(nsp32_sgtable) * NSP32_SG_SIZE * MAX_TARGET * MAX_LUN)
+#define NSP32_SG_TABLE_SIZE (sizeof(nsp32_sglun) * MAX_TARGET * MAX_LUN)
/* Auto parameter mode memory map. */
/* All values must be little endian. */
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2014-04-03 7:28 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-04-03 7:28 [patch] [SCSI] nsp32: fix an off by one Dan Carpenter
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).