[PATCH V2] xfsprogs: check for size parsing errors in xfs_quota

[Eric Sandeen]

Doing something like 

# xfs_quota -x -c 'limit -u bhard=1.2g ...

will cause cvtnum to fail and return a value of -1LL (because it
cannot parse the decimal), but the quota caller doesn't check
for this error value, casts it to U64, shifts right, and we end
up with an answer of 16 petabytes rather than erroring out.
Fix this.

Reported-by: James Lawrie <james at jdlawrie.co.uk>
Signed-off-by: Eric Sandeen <sandeen at redhat.com>
---

V2: Fix mysterious change in shift size, shorten long line

diff --git a/quota/edit.c b/quota/edit.c
index b704e63..cad3aee 100644
--- a/quota/edit.c
+++ b/quota/edit.c
@@ -226,13 +226,19 @@ extractb(
 	uint		sectorsize,
 	__uint64_t	*value)
 {
-	__uint64_t	v;
+	long long	v;
 	char		*s = string;
 
 	if (strncmp(string, prefix, length) == 0) {
 		s = string + length + 1;
-		v = (__uint64_t)cvtnum(blocksize, sectorsize, s);
-		*value = v >> 9;	/* syscalls use basic blocks */
+		v = cvtnum(blocksize, sectorsize, s);
+		if (v == -1LL) {
+			fprintf(stderr,
+				_("%s: Error: could not parse size %s.\n"),
+				progname, s);
+			return 0;
+		}
+		*value = (__uint64_t)v >> 9;	/* syscalls use basic blocks */
 		if (v > 0 && *value == 0)
 			fprintf(stderr, _("%s: Warning: `%s' in quota blocks is 0 (unlimited).\n"), progname, s);
 		return 1;