linux-scsi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Don Brace <don.brace@pmcs.com>
To: scott.teel@pmcs.com, Kevin.Barnett@pmcs.com,
	scott.benesh@pmcs.com, james.bottomley@parallels.com,
	hch@infradead.org, Justin.Lindley@pmcs.com, elliott@hpe.com
Cc: linux-scsi@vger.kernel.org
Subject: [PATCH 1 23/25] hpsa: fix multiple issues in path_info_show
Date: Wed, 28 Oct 2015 17:06:50 -0500	[thread overview]
Message-ID: <20151028220650.5323.11242.stgit@brunhilda> (raw)
In-Reply-To: <20151028215206.5323.84194.stgit@brunhilda>

From: Rasmus Villemoes <linux@rasmusvillemoes.dk>

path_info_show() seems to be broken in multiple ways.

First, there's

  817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
  818       path[0], path[1], path[2], path[3],
  819       path[4], path[5], path[6], path[7]);

so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have

  767      output_len = snprintf(path[i],
  768                       PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
  769                       h->scsi_host->host_no,
  770                       hdev->bus, hdev->target, hdev->lun,
  771                       scsi_device_type(hdev->devtype));

so output_len at best contains the length of the last string printed.

Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:

  802                  output_len += snprintf(path[i] + output_len,
  803                          PATH_STRING_LEN,
  804                          "BOX: %hhu BAY: %hhu %s\n",
  805                          box, bay, active);

An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.

We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.

Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
---
 drivers/scsi/hpsa.c |   38 +++++++++++++++++---------------------
 1 file changed, 17 insertions(+), 21 deletions(-)

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index b2418c3..56526312 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -735,7 +735,6 @@ static ssize_t host_show_hp_ssd_smart_path_enabled(struct device *dev,
 }
 
 #define MAX_PATHS 8
-#define PATH_STRING_LEN 50
 
 static ssize_t path_info_show(struct device *dev,
 	     struct device_attribute *attr, char *buf)
@@ -751,9 +750,7 @@ static ssize_t path_info_show(struct device *dev,
 	u8 path_map_index = 0;
 	char *active;
 	unsigned char phys_connector[2];
-	unsigned char path[MAX_PATHS][PATH_STRING_LEN];
 
-	memset(path, 0, MAX_PATHS * PATH_STRING_LEN);
 	sdev = to_scsi_device(dev);
 	h = sdev_to_hba(sdev);
 	spin_lock_irqsave(&h->devlock, flags);
@@ -773,8 +770,9 @@ static ssize_t path_info_show(struct device *dev,
 		else
 			continue;
 
-		output_len = snprintf(path[i],
-				PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
+		output_len += scnprintf(buf + output_len,
+				PAGE_SIZE - output_len,
+				"[%d:%d:%d:%d] %20.20s ",
 				h->scsi_host->host_no,
 				hdev->bus, hdev->target, hdev->lun,
 				scsi_device_type(hdev->devtype));
@@ -782,9 +780,9 @@ static ssize_t path_info_show(struct device *dev,
 		if (hdev->external ||
 			hdev->devtype == TYPE_RAID ||
 			is_logical_device(hdev)) {
-			output_len += snprintf(path[i] + output_len,
-						PATH_STRING_LEN, "%s\n",
-						active);
+			output_len += snprintf(buf + output_len,
+						PAGE_SIZE - output_len,
+						"%s\n", active);
 			continue;
 		}
 
@@ -796,35 +794,33 @@ static ssize_t path_info_show(struct device *dev,
 		if (phys_connector[1] < '0')
 			phys_connector[1] = '0';
 		if (hdev->phys_connector[i] > 0)
-			output_len += snprintf(path[i] + output_len,
-				PATH_STRING_LEN,
+			output_len += snprintf(buf + output_len,
+				PAGE_SIZE - output_len,
 				"PORT: %.2s ",
 				phys_connector);
 		if (hdev->devtype == TYPE_DISK && hdev->expose_device) {
 			if (box == 0 || box == 0xFF) {
-				output_len += snprintf(path[i] + output_len,
-					PATH_STRING_LEN,
+				output_len += snprintf(buf + output_len,
+					PAGE_SIZE - output_len,
 					"BAY: %hhu %s\n",
 					bay, active);
 			} else {
-				output_len += snprintf(path[i] + output_len,
-					PATH_STRING_LEN,
+				output_len += snprintf(buf + output_len,
+					PAGE_SIZE - output_len,
 					"BOX: %hhu BAY: %hhu %s\n",
 					box, bay, active);
 			}
 		} else if (box != 0 && box != 0xFF) {
-			output_len += snprintf(path[i] + output_len,
-				PATH_STRING_LEN, "BOX: %hhu %s\n",
+			output_len += snprintf(buf + output_len,
+				PAGE_SIZE - output_len, "BOX: %hhu %s\n",
 				box, active);
 		} else
-			output_len += snprintf(path[i] + output_len,
-				PATH_STRING_LEN, "%s\n", active);
+			output_len += snprintf(buf + output_len,
+				PAGE_SIZE - output_len, "%s\n", active);
 	}
 
 	spin_unlock_irqrestore(&h->devlock, flags);
-	return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
-		path[0], path[1], path[2], path[3],
-		path[4], path[5], path[6], path[7]);
+	return output_len;
 }
 
 static DEVICE_ATTR(raid_level, S_IRUGO, raid_level_show, NULL);


  parent reply	other threads:[~2015-10-28 22:10 UTC|newest]

Thread overview: 107+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-28 22:04 [PATCH 1 00/25] hpsa updates Don Brace
2015-10-28 22:04 ` [PATCH 1 01/25] hpsa: stop zeroing reset_cmds_out and ioaccel_cmds_out during rescan Don Brace
2015-10-29 13:25   ` Tomas Henzl
2015-10-30  7:45   ` Hannes Reinecke
2015-10-28 22:04 ` [PATCH 1 02/25] hpsa: remove unused hpsa_tag_discard_error_bits Don Brace
2015-10-29 13:26   ` Tomas Henzl
2015-10-29 14:37   ` Manoj Kumar
2015-10-29 14:49     ` Don Brace
2015-10-30  7:46   ` Hannes Reinecke
2015-10-28 22:04 ` [PATCH 1 03/25] hpsa: check for null arguments to dev_printk Don Brace
2015-10-29 13:41   ` Tomas Henzl
2015-10-29 14:41   ` Manoj Kumar
2015-10-30  7:47   ` Hannes Reinecke
2015-10-30 14:16     ` Don Brace
2015-10-28 22:04 ` [PATCH 1 04/25] hpsa: fix null device issues Don Brace
2015-10-29 14:06   ` Tomas Henzl
2015-10-30  7:49   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 05/25] hpsa: allow driver requested rescans Don Brace
2015-10-30  7:51   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 06/25] hpsa: abandon rescans on memory alloaction failures Don Brace
2015-10-30  7:53   ` Hannes Reinecke
2015-10-30 20:44     ` Don Brace
2015-10-28 22:05 ` [PATCH 1 07/25] hpsa: correct transfer length for 6 byte read/write commands Don Brace
2015-10-30  7:54   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 08/25] hpsa: fix hpsa_adjust_hpsa_scsi_table Don Brace
2015-10-29 14:23   ` Tomas Henzl
2015-10-30  7:57   ` Hannes Reinecke
2015-10-30 20:46     ` Don Brace
2015-10-28 22:05 ` [PATCH 1 09/25] hpsa: fix physical target reset Don Brace
2015-10-29 14:30   ` Tomas Henzl
2015-10-29 15:29     ` Don Brace
2015-10-29 15:52       ` Tomas Henzl
2015-10-30  7:59   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 10/25] hpsa: correct check for non-disk devices Don Brace
2015-10-29 14:37   ` Tomas Henzl
2015-10-30  8:01   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 11/25] hpsa: correct ioaccel2 sg chain len Don Brace
2015-10-29 15:01   ` Tomas Henzl
2015-10-30  8:01   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 12/25] hpsa: simplify check for device exposure Don Brace
2015-10-29 15:03   ` Tomas Henzl
2015-10-30  8:04   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 13/25] hpsa: simplify update scsi devices Don Brace
2015-10-29 15:53   ` Tomas Henzl
2015-10-29 16:43   ` Matthew R. Ochs
2015-10-29 19:01     ` Don Brace
2015-10-29 20:28       ` Matthew R. Ochs
2015-10-30  8:05   ` Hannes Reinecke
2015-10-28 22:05 ` [PATCH 1 14/25] hpsa: add function is_logical_device Don Brace
2015-10-29 15:53   ` Tomas Henzl
2015-10-29 16:46   ` Matthew R. Ochs
2015-10-30  8:05   ` Hannes Reinecke
2015-10-28 22:06 ` [PATCH 1 15/25] hpsa: enhance hpsa_get_device_id Don Brace
2015-10-29 16:04   ` Tomas Henzl
2015-10-29 17:04   ` Matthew R. Ochs
2015-10-30  8:08   ` Hannes Reinecke
2015-10-30 20:59     ` Don Brace
2015-10-28 22:06 ` [PATCH 1 16/25] hpsa: refactor hpsa_figure_bus_target_lun Don Brace
2015-10-29 16:27   ` Tomas Henzl
2015-10-30  8:09   ` Hannes Reinecke
2015-10-28 22:06 ` [PATCH 1 17/25] hpsa: move scsi_add_device and scsi_remove_device calls to new function Don Brace
2015-10-29 16:37   ` Tomas Henzl
2015-10-29 17:21   ` Matthew R. Ochs
2015-10-29 20:30     ` Don Brace
2015-10-30 15:56       ` Matthew R. Ochs
2015-10-30  8:09   ` Hannes Reinecke
2015-10-28 22:06 ` [PATCH 1 18/25] External array LUNs must use target and lun numbers assigned by the Don Brace
2015-10-29 19:41   ` Matthew R. Ochs
2015-10-30  8:11   ` Hannes Reinecke
2015-10-30 14:11   ` Tomas Henzl
2015-10-28 22:06 ` [PATCH 1 19/25] hpsa: eliminate fake lun0 enclosures Don Brace
2015-10-29 20:05   ` Matthew R. Ochs
2015-10-30  8:12   ` Hannes Reinecke
2015-10-30 14:12   ` Tomas Henzl
2015-10-28 22:06 ` [PATCH 1 20/25] hpsa: add discovery polling for PT RAID devices Don Brace
2015-10-29 20:20   ` Matthew R. Ochs
     [not found]     ` <563286B7.8070200@pmcs.com>
2015-10-29 20:59       ` Matthew R. Ochs
2015-10-30 14:08         ` Don Brace
2015-10-30 15:58           ` Matthew R. Ochs
2015-10-30  8:15   ` Hannes Reinecke
2015-10-28 22:06 ` [PATCH 1 21/25] hpsa: disable report lun data caching Don Brace
2015-10-30  8:16   ` Hannes Reinecke
2015-10-30 14:25   ` Tomas Henzl
2015-10-30 21:18     ` Don Brace
2015-10-30 16:27   ` Matthew R. Ochs
2015-10-28 22:06 ` [PATCH 1 22/25] hpsa: enhance device messages Don Brace
2015-10-30  8:16   ` Hannes Reinecke
2015-10-30 14:32   ` Tomas Henzl
2015-11-02 16:54     ` Don Brace
2015-11-03 13:12       ` Tomas Henzl
2015-10-30 16:53   ` Matthew R. Ochs
2015-10-28 22:06 ` Don Brace [this message]
2015-10-30  8:18   ` [PATCH 1 23/25] hpsa: fix multiple issues in path_info_show Hannes Reinecke
2015-10-30 14:33   ` Tomas Henzl
2015-10-30 17:07   ` Matthew R. Ochs
2015-10-28 22:06 ` [PATCH 1 24/25] hpsa: add in sas transport class Don Brace
2015-10-30  8:21   ` Hannes Reinecke
2015-10-30 14:40   ` Tomas Henzl
2015-10-30 20:07   ` Matthew R. Ochs
2015-10-30 22:00     ` Don Brace
2015-10-30 23:23       ` Matthew R. Ochs
2015-11-03  0:40   ` kbuild test robot
2015-10-28 22:07 ` [PATCH 1 25/25] hpsa: bump the driver version Don Brace
2015-10-30  8:22   ` Hannes Reinecke
2015-10-30 14:44   ` Tomas Henzl
2015-10-30 20:08   ` Matthew R. Ochs
2015-11-03  4:49 ` [PATCH 1 00/25] hpsa updates Martin K. Petersen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20151028220650.5323.11242.stgit@brunhilda \
    --to=don.brace@pmcs.com \
    --cc=Justin.Lindley@pmcs.com \
    --cc=Kevin.Barnett@pmcs.com \
    --cc=elliott@hpe.com \
    --cc=hch@infradead.org \
    --cc=james.bottomley@parallels.com \
    --cc=linux-scsi@vger.kernel.org \
    --cc=scott.benesh@pmcs.com \
    --cc=scott.teel@pmcs.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).