Skip to contentSkip to conversation

ethtool: Can't parse ints with stroul()

PATCH In review Jeff Kirsher 2 participants 4 messages Dec 8, 2009 – Dec 23, 2009

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ethtool: Can't parse ints with stroul()

From: PJ Waskiewicz <[email protected]>

A recent change to how int's were being parsed from the command line had them being read in with an unsigned int string operator. This didn't allow signed numbers from being read in correctly. This patch adds a get_uint() routine, and fixes the get_int() routine to read in signed values.

---

 ethtool.c |   21 ++++++++++++++++++++-
 1 files changed, 20 insertions(+), 1 deletions(-)
ethtool.c +20 -1
index 10dfc80..415ee77 100644
@@ -324,6 +324,7 @@ typedef enum {
 	CMDL_NONE,
 	CMDL_BOOL,
 	CMDL_INT,
+	CMDL_UINT,
 	CMDL_STR,
 } cmdline_type_t;
 
@@ -404,6 +405,20 @@ static struct cmdline_info cmdline_coalesce[] = {
 
 static int get_int(char *str, int base)
 {
+	long v;
+	char *endp;
+
+	if (!str)
+		show_usage(1);
+	errno = 0;
+	v = strtol(str, &endp, base);
+	if ( errno || *endp || v > INT_MAX)
+		show_usage(1);
+	return (int)v;
+}
+
+static int get_uint(char *str, int base)
+{
 	unsigned long v;
 	char *endp;
 
@@ -413,7 +428,7 @@ static int get_int(char *str, int base)
 	v = strtoul(str, &endp, base);
 	if ( errno || *endp || v > INT_MAX)
 		show_usage(1);
-	return (int)v;
+	return v;
 }
 
 static void parse_generic_cmdline(int argc, char **argp,
@@ -447,6 +462,10 @@ static void parse_generic_cmdline(int argc, char **argp,
 					*p = get_int(argp[i],0);
 					break;
 				}
+				case CMDL_UINT: {
+					*p = get_uint(argp[i],0);
+					break;
+				}
 				case CMDL_STR: {
 					char **s = info[idx].wanted_val;
 					*s = strdup(argp[i]);
#1 lore raw
Quoted 8 lines · Jeff Kirsher

From: PJ Waskiewicz <[email protected]>

A recent change to how int's were being parsed from the command line had them being read in with an unsigned int string operator. This didn't allow signed numbers from being read in correctly. This patch adds a get_uint() routine, and fixes the get_int() routine to read in signed values.

Quoted 4 lines

---

Jeff G. - hold off on this patch. We have a version 2 in the works, because we found if the sign bit was set on a uint, this would fail parsing.

Signature · 2 lines
Cheers,
Jeff
#2 lore raw
Quoted 9 lines · Jeff Kirsher
Quoted 8 lines · Jeff Kirsher

From: PJ Waskiewicz<[email protected]>

A recent change to how int's were being parsed from the command line had them being read in with an unsigned int string operator. This didn't allow signed numbers from being read in correctly. This patch adds a get_uint() routine, and fixes the get_int() routine to read in signed values.

Quoted 8 lines
Quoted 4 lines

---

Jeff G. - hold off on this patch. We have a version 2 in the works, because we found if the sign bit was set on a uint, this would fail parsing.

Was version 2 ever posted? I never saw it...

Jeff

#3 lore raw
Quoted 18 lines · Jeff Garzik
Quoted 17 lines · Jeff Kirsher
Quoted 13 lines · Jeff Kirsher

From: PJ Waskiewicz<[email protected]>

A recent change to how int's were being parsed from the command line had them being read in with an unsigned int string operator. This didn't allow signed numbers from being read in correctly. This patch adds a get_uint() routine, and fixes the get_int() routine to read in signed values.

---

Quoted 8 lines
Quoted 3 lines

Jeff G. - hold off on this patch.  We have a version 2 in the works, because we found if the sign bit was set on a uint, this would fail parsing.

Was version 2 ever posted?  I never saw it...

       Jeff

Not yet, version 2 of the patch is in test right now. I should have it sent out either later tonight or tomorrow.

Signature · 2 lines
Cheers,
Jeff