[PATCH v3, 07/16] xfsprogs: metadump: use pointers in generate_obfuscated_name()

[PATCH v3, 07/16] xfsprogs: metadump: use pointers in generate_obfuscated_name()

[Alex Elder]

Switch from using array references to using pointers to refer to the
pathname characters as they get generated.  Also limit the scope of
a few automatic variables.

Signed-off-by: Alex Elder <aelder@sgi.com>

The only update since the last post version is the addition of an
ASSERT() in the loop generating the last 5 bytes.

---
 db/metadump.c |   35 +++++++++++++++++++----------------
 1 file changed, 19 insertions(+), 16 deletions(-)

Index: b/db/metadump.c
===================================================================
--- a/db/metadump.c
+++ b/db/metadump.c
@@ -437,11 +437,8 @@ generate_obfuscated_name(
 {
 	xfs_dahash_t		hash;
 	name_ent_t		*p;
-	int			i;
 	int			dup;
-	xfs_dahash_t		newhash;
 	uchar_t			newname[NAME_MAX];
-	uchar_t			*newp = &newname[0];
 
 	/*
 	 * Our obfuscation algorithm requires at least 5-character
@@ -472,7 +469,12 @@ generate_obfuscated_name(
 
 	hash = libxfs_da_hashname(name, namelen);
 	do {
-		uchar_t	high_bit;
+		int		i;
+		xfs_dahash_t	newhash = 0;
+		uchar_t		*newp = &newname[0];
+		uchar_t		*first;
+		uchar_t		high_bit;
+		int		shift;
 
 		dup = 0;
 
@@ -482,10 +484,10 @@ generate_obfuscated_name(
 		 * characters.  Accumulate its new hash value as we
 		 * go.
 		 */
-		newhash = 0;
 		for (i = 0; i < namelen - 5; i++) {
-			newp[i] = random_filename_char();
-			newhash = newp[i] ^ rol32(newhash, 7);
+			*newp = random_filename_char();
+			newhash = *newp ^ rol32(newhash, 7);
+			newp++;
 		}
 
 		/*
@@ -500,16 +502,17 @@ generate_obfuscated_name(
 		 */
 		newhash = rol32(newhash, 3) ^ hash;
 
+		first = newp;
 		high_bit = 0;
-		for (i = 5; i > 0; i--) {
-		    	int shift = (i - 1) * 7;
-
-			newp[namelen - i] = ((newhash >> shift) & 0x7f) ^ high_bit;
-			if (is_invalid_char(newp[namelen - i])) {
-				newp[namelen - i] ^= 1;
+		for (shift = 28; shift >= 0; shift -= 7) {
+			*newp = (newhash >> shift & 0x7f) ^ high_bit;
+			if (is_invalid_char(*newp)) {
+				*newp ^= 1;
 				high_bit = 0x80;
 			} else
 				high_bit = 0;
+			ASSERT(!is_invalid_char(*newp));
+			newp++;
 		}
 
 		/*
@@ -522,15 +525,15 @@ generate_obfuscated_name(
 		 * worry about it becoming invalid as a result.
 		 */
 		if (high_bit) {
-			newp[namelen - 5] ^= 0x10;
-			ASSERT(!is_invalid_char(newp[namelen - 5]));
+			*first ^= 0x10;
+			ASSERT(!is_invalid_char(*first));
 		}
 
 		ASSERT(libxfs_da_hashname(newname, namelen) == hash);
 
 		for (p = nametable[hash % NAME_TABLE_SIZE]; p; p = p->next) {
 			if (p->hash == hash && p->namelen == namelen &&
-					memcmp(p->name, newname, namelen) == 0){
+					!memcmp(p->name, newname, namelen)) {
 				dup = 1;
 				break;
 			}

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: [PATCH v3, 07/16] xfsprogs: metadump: use pointers in generate_obfuscated_name()

[Dave Chinner]

On Fri, Feb 18, 2011 at 03:21:01PM -0600, Alex Elder wrote:
> Switch from using array references to using pointers to refer to the
> pathname characters as they get generated.  Also limit the scope of
> a few automatic variables.
> 
> Signed-off-by: Alex Elder <aelder@sgi.com>
> 
> The only update since the last post version is the addition of an
> ASSERT() in the loop generating the last 5 bytes.

One small comment below, otherwise:

Reviewed-by: Dave Chinner <dchinner@redhat.com>

> @@ -500,16 +502,17 @@ generate_obfuscated_name(
>  		 */
>  		newhash = rol32(newhash, 3) ^ hash;
>  
> +		first = newp;
>  		high_bit = 0;
> -		for (i = 5; i > 0; i--) {
> -		    	int shift = (i - 1) * 7;
> -
> -			newp[namelen - i] = ((newhash >> shift) & 0x7f) ^ high_bit;
> -			if (is_invalid_char(newp[namelen - i])) {
> -				newp[namelen - i] ^= 1;
> +		for (shift = 28; shift >= 0; shift -= 7) {
> +			*newp = (newhash >> shift & 0x7f) ^ high_bit;

I'd prefer not to have to refer to my C bible to remind myself what
the precedence of ">>" vs "&" is, so perhaps leaving the second set
of () in this statement would be a good idea.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

[PATCH v4, 07/16] xfsprogs: metadump: use pointers in generate_obfuscated_name()

[Alex Elder]

. . .
> One small comment below, otherwise:
> 
> Reviewed-by: Dave Chinner <dchinner@redhat.com>
> 
> > @@ -500,16 +502,17 @@ generate_obfuscated_name(
> >  		 */
> >  		newhash = rol32(newhash, 3) ^ hash;
> >  
> > +		first = newp;
> >  		high_bit = 0;
> > -		for (i = 5; i > 0; i--) {
> > -		    	int shift = (i - 1) * 7;
> > -
> > -			newp[namelen - i] = ((newhash >> shift) & 0x7f) ^ high_bit;
> > -			if (is_invalid_char(newp[namelen - i])) {
> > -				newp[namelen - i] ^= 1;
> > +		for (shift = 28; shift >= 0; shift -= 7) {
> > +			*newp = (newhash >> shift & 0x7f) ^ high_bit;
> 
> I'd prefer not to have to refer to my C bible to remind myself what
> the precedence of ">>" vs "&" is, so perhaps leaving the second set
> of () in this statement would be a good idea.

Done.




_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs